Towers of Capital: Office Markets & International Financial Services (Real Estate Issues)

Towers of Capital Ofﬁce Markets & International Financial Services Colin Lizieri Professor of Real Estate Finance Schoo...

Author: Colin Lizieri

342 downloads 679 Views 2MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Towers of Capital Ofﬁce Markets & International Financial Services Colin Lizieri Professor of Real Estate Finance School of Real Estate & Planning Henley Business School The University of Reading

This edition ﬁrst published 2009 © 2009 Colin Lizieri Blackwell Publishing was acquired by John Wiley & Sons in February 2007. Blackwell’s publishing programme has been merged with Wiley’s global Scientiﬁc, Technical, and Medical business to form Wiley-Blackwell. Registered ofﬁce John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, United Kingdom Editorial ofﬁces 9600 Garsington Road, Oxford, OX4 2DQ, United Kingdom 2121 State Avenue, Ames, Iowa 50014-8300, USA For details of our global editorial ofﬁces, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com/wiley-blackwell. The right of the author to be identiﬁed as the author of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books. Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought. Library of Congress Cataloging-in-Publication Data Lizieri, Colin. Towers of capital : ofﬁce markets & international ﬁnancial services / Colin Lizieri. – 1st ed. p. cm. – (Real estate issues) Includes bibliographical references and index. ISBN 978-1-4051-5672-1 (alk. paper) 1. Ofﬁces. 2. Ofﬁce buildings. 3. Commercial real estate. 4. Financial institutions, International. 5. International ﬁnance. I. Title. HD1393.55.L59 2009 332′.042–dc22 2008053120 A catalogue record for this book is available from the British Library. Set in 10/13pt TrumpMediaeval by Newgen Imaging Systems (P) Ltd, Chennai Printed in Singapore 1

2009

Books in the series Greenﬁelds, Brownﬁelds & Housing Development Adams & Watkins 9780632063871 Planning, Public Policy & Property Markets Edited by Adams, Watkins & White 9781405124300 Housing & Welfare in Southern Europe Allen, Barlow, Léal, Maloutas & Padovani 9781405103077 Markets and Institutions in Real Estate & Construction Ball 9781405110990 Neighbourhood Renewal and Housing Markets Edited by Beider 9781405134101

Real Estate & the New Economy Dixon, McAllister, Marston & Snow 9781405117784 Economics & Land Use Planning Evans 9781405118613 Economics, Real Estate & the Supply of Land Evans 9781405118620 Development & Developers Guy & Henneberry 9780632058426 The Right to Buy Jones & Murie 9781405131971 Economics of the Mortgage Market Leece 9781405114615

Mortgage Markets Worldwide Ben-Shahar, Leung & Ong 9781405132107

Housing Economics & Public Policy O’Sullivan & Gibb 9780632064618

The Cost of Land Use Decisions Buitelaar 9781405151238

International Real Estate Seabrooke, Kent & How 9781405103084

Urban Regeneration in Europe Couch, Fraser & Percy 9780632058412

British Housebuilders Wellings 9781405149181

Urban Sprawl Couch, Leontidou & Petschel-Held 9781405151238

Mass Appraisal Methods Kauko & d’Amato 9781405180979

Forthcoming Building Cycles & Urban Development Barras 9781405130011

Affordable Housing & the Property Market Monk & Whitehead 9781405147149

Transforming the Private Landlord Crook & Kemp 9781405184151

Property Investment & Finance Newell & Sieracki 9781405151283

Housing Markets & Planning Policy Jones & Watkins 9781405175203

Housing Stock Transfer Taylor 9781405170321

Management of Privatised Housing Gruis, Tsenkova & Nieboer 9781405181884

Real Estate Finance in the New Economic World Tiwari & White 9781405158718

Acknowledgements The ideas presented in Towers of Capital have evolved over a long period and have been shaped by so many conversations, debates and arguments over the years that to single out any individuals who have, wittingly or unwittingly, contributed would almost be unfair. I am very grateful to my colleagues in the School of Real Estate & Planning at Reading, for their ideas, suggestions, and for their patience – and for covering my work while I was on sabbatical completing the text. Particular thanks are due to those who read and commented on drafts of chapters: Phil Allmendinger, Michael Ball, Peter Byrne, Steve Devaney, Franz Fürst, Charles Ward, with apologies to anyone else I have omitted. The book has been shaped, improved and made possible by my many friends and colleagues in the international real estate research community. A number of individuals and ﬁrms have provided data: many thanks to all of them both for the data and for tolerating the queries and questions about it which inevitably followed. I should properly acknowledge three individuals who have made major contributions to the development of my ideas. First, to Geoff Dobilas from GHK International in Toronto: Geoff helped with my early attempts to sketch out the links between international ﬁnance and ofﬁce markets nearly two decades ago and has been there ever since to insult me, distract me and prick pretensions. I owe a great debt, too, to Charles Ward (University of Reading) and to Steve Satchell (University of Cambridge). Both of them over the years, as friends and as co-authors, have forced me to tighten up my ideas and to wrestle with theories and techniques from economics and ﬁnance. My family and friends have borne the brunt of the writing of Towers of Capital: particular thanks to my son Neil and to Ruth Jordan for keeping me on track, I am truly grateful. Most of all, though, I must acknowledge my mother, Marjorie Lizieri, who taught me to ask why and how, not what, to search for explanations and to try to understand how the world is shaped. This book is dedicated to her memory. Colin Lizieri, Reading August 2008

Contents About the Author Introduction

x xi

PART I SYSTEMS OF CITIES AND CITIES OF FINANCE

1

Introduction

2

1

The Urban Hierarchy and Global Cities 1.1 Introduction 1.2 Cities and the urban hierarchy 1.3 World cities, global cities 1.4 Cities, space and ﬁnance

2 International Financial Centres and Global Cities 2.1 Introduction 2.2 Concentration in ﬁnancial centres 2.3 Information technology and cities 2.4 Agglomeration and concentration 2.5 IFCs, concentration and ofﬁce markets 3

A Sense of History: Development and Inertia in the International Financial Hierarchy 3.1 Introduction 3.2 A sense of history 3.3 ‘The age of catastrophe’ 3.4 Asian and Middle Eastern markets 3.5 Lessons of history: a tentative conclusion

5 5 6 18 31 33 33 34 44 52 60

65 65 67 74 76 81

PART II INSIDE THE OFFICE MARKET

87

Introduction

88

4 Occupational Demand, Ofﬁce Space and Rents 4.1 Introduction 4.2 Land value, land rent and ﬁnancial services 4.3 Rent models and rental adjustment 4.4 Empirical evidence of rental adjustment processes in major ofﬁce markets

91 91 94 97 105

viii

Contents

4.5 4.6

Changing business practices and the demand for ofﬁce space Employment, volatility and the use of space

109 117

5

The Supply of Space 5.1 Developers, development cycles and research 5.2 Development, building costs and property value 5.3 Economic models and developer behaviour 5.4 Real options and developer strategies

125 125 129 135 144

6

Investment, Capital Flows and the Ofﬁce Market 6.1 Introduction 6.2 Real estate in the portfolio 6.3 Investment patterns and investor behaviour 6.4 Real estate vehicles and the portfolio 6.5 International real estate investment 6.6 Real estate capital ﬂows 6.7 Investment, capital ﬂows and IFC ofﬁce markets

149 149 150 161 163 165 170 180

PART III

INTERNATIONAL FINANCE, GLOBAL OFFICE MARKETS AND SYSTEMIC RISK

183

Introduction

184

7

Booms, Bubbles, Crises and Contagion 7.1 Introduction 7.2 Financial markets, co-movement and crises 7.3 Booms and bubbles 7.4 Real estate in booms and bubbles 7.5 Bubbles, crashes and IFC ofﬁce markets

187 187 189 201 206 214

8

Globalisation, Ownership and Innovation in Real Estate Markets 8.1 Introduction: innovation in real estate investment and ownership 8.2 Innovation in property investment vehicles 8.3 Global real estate investment 8.4 Ownership patterns 8.5 Conclusions: innovation, ownership and real estate risk in IFCs

9

All Fall Down: Global Financial Centres, Real Estate and Risk 9.1 Introduction 9.2 Retracing the threads

217 217 219 246 251 255

258 258 259

Contents ix

9.3 9.4 9.5

Market integration, globalisation and systemic risk Some preliminary evidence Global ﬁnancial centres and ofﬁce market risk: some policy implications

Epilogue Bibliography Index

270 277 287 297 303 321

About the Author Colin is Professor of Real Estate Finance in the School of Real Estate & Planning at the University of Reading’s Henley Business School. He is Director of Research for Real Estate & Planning and served as Head of Department. Colin has worked in real estate for nearly 30 years. His principal research covers modelling of real estate investment, ofﬁce markets and international real estate. He has carried out research and consultancy for the European Union, the Norwegian Government, Her Majesty’s Treasury, the Department of Communities and Local Government, the Home Ofﬁce and a wide range of public and private sector clients including PREA, the RICS and IPF. He led the Investment Property Forum research projects on commercial property market liquidity and on the pricing of property derivatives. He has published widely in academic and professional journals, with over 100 publications and is the co-author of The Economics of Commercial Property Markets and has been asked to present research and lecture in over 30 countries. In 2004, he was awarded the International Real Estate Society achievement award ‘for outstanding achievement in real estate research, education and practice at the international level’ and in 2005 was elected an honorary fellow of the Society of Property Researchers ‘in recognition of an outstanding contribution in the ﬁeld of property research’.

Introduction For all the many cultural and physical differences that distinguish the developed world’s great cities, a common feature of the majority is the existence of a dense core of ofﬁce space in or near the centre of each city. That ofﬁce space accommodates global ﬁrms, domestic and non-domestic, many of which are international ﬁnance ﬁrms – banks, insurance companies, pension funds, hedge funds, private equity funds, wealth managers and real estate companies – or are business and professional service ﬁrms whose work is linked to global capital ﬂows. From time to time, these ofﬁce markets undergo a dramatic transformation – old buildings are demolished, new larger, higher speciﬁcation buildings, often designed by signature architects, are constructed; the core ofﬁce area spills out into surrounding neighbourhoods, displacing other activities. From time to time, these ofﬁce markets undergo crises: overbuilding creates an oversupply of space, vacancy rates rise, rents and capital values fall, real estate ﬁrms fail. Such crises are often linked to ﬁnancial market crises – and, crucially, often occur simultaneously in cities spread geographically across the regions of the world. Towers of Capital explores the relationship between the evolution of major international ﬁnancial centres (IFCs) as part of the global capital market system, the development of ofﬁce markets in those cities, real estate investment in those ofﬁce markets and the patterns of risk and return that result from the interactions between ﬁnancial ﬂows and ofﬁce markets. Each of these aspects has, individually, been the subject of considerable research from a variety of academic traditions. Here, rather than focusing on just one single aspect of the relationship and conducting micro-analysis, an attempt is made to set out the interconnections between the location of ﬁnancial activity, the processes operating in ofﬁce markets and the volatility of real estate returns. The intention is to create a schematic model of risk in IFC ofﬁce markets that can act as a springboard for subsequent empirical work. The essential thesis examined here starts from the increasing spatial clustering of global ﬁnancial business. The concentration of ﬁnancial activity in a small number of major cities, acting as the coordinating centres for an interlinked international ﬁnancial system has been accompanied by a redevelopment of the core ofﬁce markets of those cities and a growing functional specialisation of activity in those ofﬁces as high-order ﬁnancial services and the business and professional service ﬁrms that supply them have tended to drive out other users. A consequence of this process has been an interlocking of occupation, ownership and ﬁnance: ﬁrms that occupy space are the same ﬁrms that acquire ofﬁces as an investment asset (directly and indirectly) and who provide ﬁnance and funding for the creation of new

xii

Introduction

ofﬁce space. This interlocking creates greater potential volatility in ofﬁce markets, increasing the amplitude of upswings and downswings. Shocks in international ﬁnancial markets are transmitted to occupier, investment and property debt markets and can reinforce any tendency to cyclical behaviour. A number of more recent trends have tended to reinforce this potential volatility: notably the growing globalisation of ownership of real estate and innovation in real estate investment vehicles and debt markets which has fragmented ownership without necessarily diversifying away risk at the overall market level. Aspects of the underlying idea here were sketched out more than 15 years ago in Lizieri (1991, 1992); developed in Lizieri and Finlay (1995) and, more fully, in Lizieri et al. (2000a). The idea of a linkage between real estate and shifts in global ﬁnancial markets and on the importance of real estate processes and real estate ﬁnance in understanding the transformation of world cities is also present to an extent in, for example, Leitner (1994), Fainstein (1993) and Haila (1997). However, over the last two decades, urban research seems to have veered away from a holistic focus on urban development, driven by the apparently increasing gulf between urban social science approaches and those of real estate and urban economics. In particular, the cultural turn in urban geography and urban social science has seen a drift away from critical empirical work to a focus on the symbolic and political implications of urban form and a mode of analysis which often produces opaque, esoteric work that is inaccessible to outsiders. Much of the research in this tradition seems to have moved very far away from the policy-driven critical urban social science that underpinned the early development of the world cities literature and the analyses of urban development processes in the 1980s and early 1990s. From this perspective, traditional urban economic approaches are dismissed for economism and reductionism, generally using a caricature of economic research. In terms of method, the cultural turn has brought an emphasis on textual analysis and small scale interview work, and a focus on individuals – with a consequent rejection of aggregate level analysis. Even where the urban social science focus has returned to economic processes in global cities, there seems a reluctance to deal with real estate processes at anything other than a superﬁcial level. Commentators note the production of real estate in global cities, they note its symbolic importance, they discuss the politics of urban planning – but, with few exceptions, they do not address the production and exchange of that space in sufﬁcient depth as to suggest an understanding of actual processes and the implications of those processes. As an example, Doreen Massey’s recent World City1 presents a powerful critique of much of the global city debate in relation to London, to social inequality and to the relationships between London and 1

Massey (2007).

Introduction xiii

the UK regions. She does deal with real estate: ‘it is these land and property interests that have been the vital mechanism in the concentration of investment in commercial building in the prime central areas of cities and into the central cities such as London’ (Massey, 2007, p. 48). But there is no further analysis – she talks of ‘ﬁnancial land ownership’ citing her own work from fully 30 years ago (Massey & Catalano, 1978), there are a handful of references to urban geography articles from the late 1980s or early 1990s and a brief discussion of real estate values and property taxes. As with Sassen (1991) she emphasises that a world city must produce the means by which ‘command and control’ of the world economy can take place – which must include ofﬁce space. Yet there is almost nothing about the mechanisms by which this takes place, no consideration that these might change over time with consequences for the development of a global urban system and risks within that system. To a large extent, the motivation for this book comes from a frustration about this neglect of the key role played by real estate and by a superﬁcial view of property that ignores change over time and the importance of market structures in determining spatial outcomes in the ofﬁce markets of cities in different countries and regions. I wanted to try to explore the processes that operate in major ofﬁce markets and the dynamics of those markets in more depth – because it seems to me that we cannot truly understand the forces that shape the city unless we have that depth of understanding. That exercise requires a blending of the wider insights of urban social science, international economics and ﬁnance with the more detailed focus on microlevel processes found in real estate and urban economics research, for all the methodological differences that exist between these different literatures and disciplines. There are similar frustrations in drawing on the real estate economics literature. Increasingly, the leading journals in the ﬁeld publish work that is highly quantitative, using theoretical micro-economic modelling and empirical econometric analysis. Wider explanatory frameworks can be squeezed out in the focus on technique. For example, the drive to more advanced quantitative techniques has led to a focus on analysis of public real estate company returns. A number of papers have examined global or regional integration issues.2 Results typically show the existence of a global real estate factor, some convergence, but less integration than national equity indices. Investment in REITs and property companies is, in the long run, a real estate investment, but the results of such analyses tell us little about actual investment in real estate and the extent to which building performance is affected 2

See, for example, papers by Eichholtz and co-workers (e.g. Eichholtz et al., 1998), Quan and Titman (1999), Ling and Naranjo (2002), Bond et al. (2003) and McAllister and Lizieri (2006).

xiv

Introduction

by global rather than local and national factors. As an example, Bardham et al. (2008) examine real estate securities to test explicitly for the implications of global ﬁnancial integration, using an ‘openness’ variable, which is shown to be signiﬁcant (but with little discussion about the validity of the variable itself). The paper, published in Real Estate Economics, reports results from a variety of panel regressions. The highest adjusted R2 was 0.11 which, even given the panel nature of the analysis, calls into question the economic (if not the statistical) signiﬁcance of the global integration result. By implication, the model fails to explain 89% of variation in returns. Comparative international research on direct private markets is less common, hampered by data availability (with short time series at low frequency and with deﬁnitional differences hampering quantitative analysis). Exceptions include Goetzmann and Wachter (1995) and Case et al. (1999) although, revealingly, those much cited papers have not been published in journals to the best of my knowledge. More general work on global property cycles includes Renaud (1997) and Herring and Wachter (1999). Another strand of literature using international data is work on international portfolio diversiﬁcation.3 The last decade has also seen further sophistication in the modelling of ofﬁce market processes in individual cities (reviewed in Part II of this book) but that, too, is reaching a point where sophistication is becoming the enemy of understanding. There have been comparatively few cross-national return or rent models and those that have typically rely on aggregate macro-economic demand variables to explain variation. My aim here, then, is to attempt to pull together the insights, models and ideas that address global capital ﬂows, the evolution of city systems, ofﬁce market processes and real estate ﬁnance, to try to understand and explain the linkages between the evolution of ﬁnancial markets, innovation in commercial real estate markets and the dynamics of the ofﬁce markets in global cities. I have not attempted to conduct detailed econometric or statistical analysis of ofﬁce market performance in IFCs. However, in contrast to much of the recent urban literature, I would argue that an analytic, empirical approach and a focus on the economic is crucial if we wish to understand the mechanisms that shape the urban environment and that connect local outcomes in spatially separate cities. However, such analytic work needs to be, must be, placed in a broad theoretical context, an overall model that explores global processes without resorting to simplistic abstractions of the behaviour of economic agents. Hence, while supporting analyses are presented, the main intention of Towers of Capital is to set out a framework for understanding real estate and the transformation of the built environment in ﬁnancial centres, based both on the development of global capital markets and on micro-level research into the functioning of ofﬁce markets, backed, 3

For reviews, see Sirmans and Worzala (2003) and Hoesli and Lizieri (2007).

Introduction xv

where feasible, with empirical evidence. My hope is that this framework can provide a springboard for the development of a programme of empirical work and theoretical reﬁnement in the future.

Structure The book is divided into three parts, each containing three chapters. Part I, Systems of Cities and Cities of Finance, looks at the evolution of a global hierarchy of cities whose role extends well beyond their national location – often dubbed world cities. In particular, it examines the development of cities that play a key role in coordinating the global ﬁnancial system – IFCs. These major cities have captured a growing market share of international ﬁnancial activity. The role of technology (particularly relating to information exchange and digital capital ﬂows) will be considered: what does this do to spatial patterns and the ‘need for cities’? What are the factors that lead to global ﬁnancial services ﬁrms clustering together and remaining in highcost locations when so much trading activity happens electronically? The chapters draw on the world cities literature – but in a critical way – and on the regional and urban literature of the ‘new economic geography’. They will also look back into economic and urban history to show that the ‘new international ﬁnance system’ and the presence of global ﬁnancial centres have long historic antecedents. Financial activity in global cities creates a demand for space. The centres of these cities contain major ofﬁce markets: supply and demand in those ofﬁce markets is dominated by global ﬁnancial activity. Part II, Inside the Ofﬁce Market, focuses on the mechanics of the ofﬁce market – keeping the focus on prime (class A) ofﬁce space found in the IFCs. The relationship between demand for space, supply of ofﬁce buildings and rent levels will be considered, drawing on the extensive ofﬁce market modelling literature to examine rental adjustment processes. The rent levels that occur in markets derive from the demand for space and the beneﬁts of locating in particular markets – but also reﬂect institutional market structures. Next, the supply side is considered explicitly, looking at the developer’s decision to create new ofﬁce buildings, the role of ﬁnance and funding and the inﬂuence of urban land-use planning. In many major markets, development activity appears to be strongly cyclical: reasons for this tendency to boom and slump are explored. Finally, capital ﬂows into and out of the real estate market are analysed. Why do professional investors acquire real estate as part of their investment portfolios? What barriers do they face in building real estate portfolios and how are they overcome? How have investors sought to build global real estate portfolios? These questions are vital to understanding the mechanics that drive ofﬁce markets in global ﬁnancial centres. This places the ofﬁce market in the urban context introduced in

xvi

Introduction

Part I, with the emphasis on price formation and investment ﬂows as driving factors distinguishing it from much of the urban social science literature on the development process and the restructuring of the built environment on world cities. Part III, International Finance, Global Ofﬁce Markets and Systemic Risk, brings together the themes from Parts I and II to investigate the interaction between ﬁnancial markets and real estate markets and the impact of that interaction on the ofﬁce markets of IFCs. First, the behaviour of international ﬁnancial markets and the tendency for there to be pronounced asset market bubbles and ﬁnancial crises are examined. Crucially, ﬁnancial crises often spread out beyond a single national market, with effects spilling over into surrounding regional and global markets. Real estate markets are affected by ﬁnancial crises but can also have causal effects – the real estate cycle interacting with credit cycles and ﬁnancial booms and busts and contributing to systemic risk. Next, changes in commercial real estate markets that have had a major impact on risk and return in ﬁnancial centre ofﬁce markets are detailed. These changes include innovative debt and equity vehicles in real estate (in particular the growth of private real estate funds and securitised debt products), the globalisation of real estate investment and ﬁnance and the fragmentation of property ownership. All three features have transformed ofﬁce markets in global ﬁnance centres. Finally, the implications of all of these changes are analysed: the way that real estate ﬁnance, investment and development are intertwined in the ofﬁce markets of global cities, how those ofﬁce markets inﬂuence and are inﬂuenced by capital markets and how the factors that drive market performance are increasingly global rather than local. Taken together, this creates systemic risk and major problems for the management of the urban environment in IFCs. The policy implications of this interlocking market are then explored.

Part I Systems of Cities and Cities of Finance

Introduction This ﬁrst part of the book explores the urban system. In particular, it charts the existence of major global cities that are linked together by trading and business activity, skilled labour migration, cultural factors and, above all, by ﬁnancial ﬂows. Starting from basic models of the role of towns and cities and explanations of urban growth, it seeks to explain the evolution of a system of international cities that play a critical role in coordination of global ﬁnancial activity and which capture an increasing share of that activity. Such an understanding not only requires a knowledge of the operation of international ﬁnance but it also demands a historical perspective. It is easy to fall into the trap of assuming that the changes observed today mark a fundamental, a structural break with the past – missing important continuities. It will be argued that the international ﬁnancial centres (IFCs) have their origins in the earlier, pre-1914, era of global capital ﬂows; that their current manifestation is an evolution not a revolution. Nonetheless, the growing concentration of ﬁnancial activity in a small number of global centres has profound implications – not least for the urban structure of those cities themselves. Chapter 1 traces ideas about the development of cities and the evolution of a global urban system. Much of the early research on the growth and location of cities was national or regional in nature. It aimed to explain variations in the size of cities and to understand the spatial distribution of urban settlements. Two strands of theories emerge, one that examines the issue of whether cities should be specialised or diversiﬁed, the other relating the location of cities of different sizes to their role in supplying services to the surrounding catchment – central place theory. In different forms, these two strands appear and reappear in urban social science. Major cities, though, perform functions that transcend national boundaries – acting to coordinate foreign investment and as the base for multinational enterprises, playing global political, cultural and ﬁnancial roles. In a world where barriers to movement of commodities, goods, services and people appear to be reducing, that global role becomes more important. Theories of a world urban system and the growing importance of ‘global cities’ are examined – as are critiques of those theories for being based on a northern trans-Atlantic model of the world that is insensitive to regional differences and to the political economy that a world system of cities implies. Chapter 2 narrows the focus to consider the role of major cities in coordinating a ﬁnancial system that is increasingly global in nature. Financial activity is heavily concentrated nationally and, within nations, by a small

Introduction 3

number of cities. Those cities – IFCs – capture a growing market share of equity trading, bond dealing, foreign currency exchange, derivatives markets, investment banking and wealth management. A small group of IFCs dominates global ﬁnancial activity and are increasingly global in nature with ﬁnancial ﬁrms from many countries having a presence there. How do IFCs evolve and what are the factors that create and sustain their dominance? Why do the major IFCs perform multiple functions and not specialise in particular activities? The chapter sets out possible factors: the need to achieve a critical mass of trading to allow specialisation of activity and attract support services, the need for a liquid international labour market, agglomeration and scale economies that cut across individual ﬁnancial sectors and knowledge spillovers and interactions that contribute to innovation. With the growth of the internet and telecommunications technologies, what does it mean to be ‘in’ an IFC and what types of activity actually need an urban location? Chapter 3 tries to provide some historical perspective. Much of the literature on the ‘New International Financial System’ – which, in turn, underpins much of the literature on the evolution of global cities – argues for a qualitative change in ﬁnancial markets, generally dated to the breakdown of the Bretton Woods agreement in the mid-1970s but driven by technological change that both permits concentration of activity and allows ﬁnancial capital to ﬂow around a system of global ﬁnance with international cities as the nodes in this ﬁnancial network. However, an examination of the history of international ﬁnance shows that global interconnection is no new phenomenon. The early evolution of specialist ﬁnancial cities in Europe was driven by placement of foreign debt and securities and by currency exchange and, by the nineteenth century, international ﬁnance is dominated by a small number of cities whose activities were by no means conﬁned to imperial regions, many of which remain key ﬁnancial centres into the twenty-ﬁrst century. Many of the claimed features of the ‘new international ﬁnancial system’ appear in that earlier era of global capitalism – including ﬁnancial crises that might start in one country but, through contagion, spill over into many other cities, often in different regions. To this extent, it is the 1914– 1970 era that seems the aberration, rather than the 1970s marking a new era of international ﬁnance. Technology changes the dynamics of transmission; but does it fundamentally alter the key features of global ﬁnance? The existence of major IFCs as the key nodes in a system of international ﬁnancial ﬂows has important implications for the structure of those cities. For all the advances in communications technologies, cities of ﬁnance require a physical presence, a base for the banks, equity, bond and derivative dealers, commodity and currency traders, institutional and private investors, the hedge funds and the wealth managers – and all the business and

4

Towers of Capital

professional services that support their activity. That physical presence requires a critical mass of ofﬁce space: and the demands for proximity and interaction generally mean that that ofﬁce space will be centrally located. As IFCs grow, as they become more functionally specialised, there will be a demand for more ofﬁce space, larger space, higher quality space – with profound impacts on the physical structure and development of those cities.

1 The Urban Hierarchy and Global Cities

1.1

Introduction

Ofﬁces exist to provide a business environment for business, professional and ﬁnancial services. Given the spatial distribution of these service sectors, this means that the city is the natural habitat for ofﬁces and the geography of ofﬁce markets is the geography of cities. To understand the distribution of ofﬁces, we need to understand the evolution of the urban system. In this process, there is an internal and an external geography. The internal geography is the structure of cities, the patterns of land-use, the location of activity within the urban area. The external geography is concerned with the distribution of cities and the interrelationship between cities. It is the latter that is the subject of this chapter. In a highly urbanised nation, the distinction between urban and rural is blurred. The idea of the city as a point in space – a central market place, with the centre as the point of maximum accessibility and the location of key service activities, surrounded by a wider hinterland that is both a market and a source of factors of production – is clearly deﬁcient. The costs of crowding and congestion, improvements in communications technology, tax differentials and socio-political issues have created suburban ofﬁce nodes, edge cities and rural ofﬁce clusters. Nonetheless, such developments represent adjustments to the existing system of cities which evolved precisely because of the economic beneﬁts and accessibility advantages of a central location. The pre-existing geography of cities is an inertial force and creates a path dependency in the location of ofﬁce-based services. Further, the establishment of a large ofﬁce market can create agglomeration economies, scale economies that have a key inﬂuence on the competitiveness of a city. In examining the inﬂuence of the spatial impact of cities on ofﬁce markets, then, the starting point must be the tradition models of the urban system based on the idea of the city as a central market place. Such models have been adapted and evolved to deal with more complex economic,

6

Towers of Capital

social and technological contexts but still provide valuable insights about the location decisions of ﬁrms and the distribution of ofﬁce space. Central place models, however, point to the city as being multifunctional. Another strand of urban economics points to the beneﬁts of sectoral (or functional) specialisation. Do cities beneﬁt from being specialised or diversiﬁed? If specialist cities are favoured, what does that imply? Most models of cities have been developed within a national or regional framework of analysis. National economies, though, are not closed and most major cities have been international in nature, linked to other countries and other cities through trade and the ﬁnancing of trade and production. The second half of the chapter examines theories that attempt to explain the global distribution of cities and the position of cities within a world urban hierarchy: World cities, global cities. In turn, this will point to the existence of specialised international ﬁnancial centres, the subject of the next chapter.

1.2

Cities and the urban hierarchy

It is evident that cities within a national system vary considerably in size whether measured in terms of employment or population. There have been many attempts to explain the distribution of city size, building from the work of Auerbach (1913) and Zipf (1949). Zipf’s ‘law’ is based on the idea that the distribution of cities is approximated by a Pareto distribution of the form y = Ax−α where y is the population of a city, A is the size of the largest city and x is the rank of the city. Zipf argued that α should take on the value 1 – which suggests that the size of the nth ranked city should be 1/n times the size of the largest city. Subsequent empirical research has sought to test whether the Zipf rule holds for countries or regions, with the general ﬁnding that α is, on average, slightly greater than one (Soo, 2005). Nonetheless, many countries exhibit a population size distribution similar to that predicted by the Zipf rank size law (see Figure 1.1). The rank size rule explores regularity in the size distribution of cities, but does not provide a theoretical explanation of why such distributions should arise (in principle, they could result from random ﬂuctuations in growth over long periods, hence the Pareto distribution), nor does it provide a prediction of the spatial distribution of cities. As a general explanation, cities beneﬁt from economies of scale, agglomeration economies and other positive externalities, which encourage growth and allow those cities that gain a size advantage to compete successfully against nearby cities; therefore the largest cities in a nation or region are likely to be separate geographically. There are also diseconomies of scale: congestion and transport costs which rise with size and density, pollution and environmental factors, for example,

The Urban Hierarchy and Global Cities

7

(a) 10 000 000

9 000 000 Predicted

8 000 000

Actual

Population

7 000 000 6 000 000 5 000 000 4 000 000 3 000 000 2 000 000 1 000 000 0 1

3

9

7

5

11

13

15

17 19 21 City rank

23

25

27

29

31

33

35

37

(b) 20 000 000

Population

18 000 000 16 000 000

Predicted

14 000 000

Actual

12 000 000 10 000 000 8 000 000 6 000 000 4 000 000 2 000 000 0 1

3

5

7

9

11

13

15 17 19 City rank

21

23

25

27

29

31

Figure 1.1 (a) European Union urban agglomeration: population by rank. (b) India urban agglomeration: population by rank. Source: Calculated from UN urban agglomeration estimates reported in Geltner et al. (2007). Note: Population estimates for EU pre-date the expansion of the European Union.

which serve to limit the concentration of population and the size of cities. This general balancing of forces of concentration and dispersion, however, does not amount to a formal model of the geographical distribution of cities within an economy. Perhaps the most familiar spatial model of the urban system is Christaller’s (1933) central place system – at least in its market format. The starting point for the model has an agricultural population distributed across a plain. This population both markets and consumes goods and there are costs

8

Towers of Capital

associated with distance. To minimise distance and costs, a market should be established that is central to a catchment area. The catchment area will be circular – but will meet with competition from other markets. The nearest form to a circle that ﬁlls the plain leaving neither gaps nor overlaps (that is, that tessellates) is a hexagon and, therefore, the initial market structure consists of a plain of hexagonal market areas, each with a central settlement. The size of the hexagon is determined by the range of the good (the maximum distance that a household will travel to consume that good/that a producer can travel to sell the good) and the threshold of the good: the minimum population (market) size to sustain production and sale. The model then allows for goods differentiation and there may be different scales of production or different transport costs. Some goods are purchased frequently or have a short range (perishable foodstuffs, for example). These will be sold in each of the central settlements. Other goods are purchased infrequently and may have a larger threshold. The provision of these goods, then, must be concentrated in a smaller number of settlements. Positive demand externalities also suggest that the sale of different high-order goods should be concentrated in the same settlements. This creates a higher order of settlements. Distance minimisation again creates a hexagonal pattern of market areas, superimposed on the lower order network. As more high-order goods are introduced, a settlement hierarchy emerges, the standard K = 3 hexagonal structure of central place theory (Figure 1.2). Although the K = 3 structure is the most commonly taught, Christaller allowed for other drivers of centrality – transport-driven and administrative-driven networks that have the same hexagonal form but a different

Figure 1.2

Central place K = 3 lattice.

The Urban Hierarchy and Global Cities

9

distribution of settlements. Lösch (1940) extended the Christaller model by combining these formats. Central place theory has had a major effect on thinking about systems of cities (possibly because it is taught in high school geography!) and has inﬂuenced regional spatial planning policy and retail location models. However, its underlying assumptions – of a dispersed population and of consumer behaviour driving the creation of settlements – do not explicitly deal with the proﬁt-maximising behaviour of ﬁrms within the system, nor the production of goods themselves: in effect it is a model of the dispersed production and central sale of agricultural goods. As such, it might be effective in explaining the origins of settlement patterns, the evolution of the urban system but might be less effective in explaining why those patterns should sustain in an economic system dominated by manufacturing or by services and with a population that is concentrated, not dispersed. Models of the urban hierarchy that are based on Christaller’s ideas imply that cities will be multifunctional. High-order cities offer a wide range of goods and products to consumers, while lower order cities offer less of a range. Such models deﬁne city status in terms of the range of services offered to the residents and the hinterland. An alternative explanation of cities focuses on production and employment structure. Here, the existence and growth of cities depend on the location decisions of ﬁrms and workers (and the supply of space and land by developers). The equilibrium state will depend on the cost structures facing ﬁrms and workers, the production function of ﬁrms, transport and trading costs and the mobility of factors of production. In general, ﬁrms will beneﬁt from economies of scale and from agglomeration – favouring city growth and a force for centralisation – but will experience congestion and crowding costs as cities grow beyond some optimal threshold. Agglomeration economies are considered in more detail in Chapter 3, but some discussion is necessary here. For more complete reviews see Combes et al. (2005) and Duranton and Puga (2000, 2005). Henderson (1974) argued that the agglomeration economies that beneﬁt ﬁrms are Marshallian externalities – that is, they are derived from the proximity of similar types of ﬁrms within a sector (a larger specialised labour pool, expertise, specialist factor inputs and the possibility of vertical and horizontal integration). With the assumption that goods can be freely traded across all cities and that labour is fully mobile, productivity increases with total employment in a sector. However, the congestion and crowding costs associated with population growth limit the optimal size of the city. Henderson’s model allows for proﬁt-oriented developers and the ability to create new urban spaces (an entrepreneurial local government might have a similar effect) – given some credence in a North American context with the development of ‘edge cities’ (Garreau, 1991). With these assumptions, the equilibrium solution is for cities to specialise. Differences in production costs and proﬁts mean that cities will be different sizes.

10

Towers of Capital

The restrictive assumptions of the Henderson model can be relaxed and extended. First, production can be viewed as a multi-stage process where producers of ﬁnal goods obtain services from separate providers of non-tradable goods – producer services. The larger the sector in a city, the more producer services can be supported and the more specialisation can take place. This favours larger cities. Even here, specialisation is favoured. However, another possibility is that agglomeration economies exist that are cross-sectoral in nature. In addition to the specialisation of producer services possible from scale, there are beneﬁts from cross-sectoral learning, technology transfer, the existence of a large multi-skilled workforce and greater interaction, which creates more opportunity for innovation and ﬁrm creation. These urban economies are sometimes labelled Jacobian economies after Jane Jacobs (1960), but similar ideas can be found in the work of Chinitz (1961) inter alia. Jacobian economies point to the advantages of diversiﬁed cities. There is thus a tension between localisation and urbanisation economies, with different urban outcomes depending on which dominates. A further complication comes from transport costs. The early Henderson model assumes that all goods are freely traded across cities. However, as in the Christaller model, it is more appropriate to assume that shipping costs increase with distance. In the Henderson model, diversiﬁcation is costly (since there are lower localisation economies but the same congestion costs) but specialisation is also costly since goods must be imported and exported. The higher the costs, the more likely it will be that cities will be diversiﬁed. The implication is that if there are signiﬁcant falls in transport costs, then cities are more likely to become specialised (Abdel-Rahman, 1996). Allowing for differentiated goods, larger cities beneﬁt by being able to offer variety for consumers: migration creates bigger markets, which may attract more ﬁrms, in a process of cumulative causation. However, if the economy of the city and its hinterland is sufﬁciently large, it may be proﬁtable for developers and ﬁrms to create new smaller cities to provide for local consumption. With multiple production sectors and differential centres, Fujita et al. (1999a) suggest that the outcome is a hierarchy of cities, but not one that necessarily conforms to the regular spatial patterns generated by the Christaller model. A ﬁnal issue to be confronted is economic change in the area containing our system of cities. Sectors grow and decline, the urban system is dynamic. A city that is specialised in a sector that is subject to decline in productivity and products may suffer – as evidenced by the manufacturing restructuring of the 1980s and many examples from history. The specialised city has developed producer services that serve its dominant production sector and consumer services that serve the workforce in the ﬁnal goods production and producer services sectors, so decline has potentially signiﬁcant effects. With full labour mobility, one might expect to see outmigration and a decline in

The Urban Hierarchy and Global Cities

11

the size of the city, although empirically, there appears to be considerable inertia – which can only partially be explained by government intervention. In this sense, a diversiﬁed city diversiﬁes risk from sector shocks in the same way that a diversiﬁed portfolio of investment assets diversiﬁes away speciﬁc risk. Furthermore, the innovative capacity of a city will allow new industries to arrive and new production technologies to be developed. Duranton and Puga (2000) suggest that a long established specialised city may be slower to adopt new technology since ‘learning by doing’ means that the initial productivity of a new technology will be lower than that of the existing technology and hence there is limited incentive to adopt production methods that reduce proﬁts relative to competitors. Innovation can occur either in a new and smaller city (where the lower congestion costs permit the application of initially less productive technologies) or in a diversiﬁed city where shared learning across sectors can occur and where there is a wider range of producer services (from specialisation effects) that can generate and implement new processes. Additionally, given that not all innovations will prove to be effective, the more a city can generate interactions, the more likely it is to produce successful new processes – in effect a version of the incubator hypothesis, but played out across a system of cities rather than within a city-region. It should be stressed that the urban systems originating in Christaller’s and Henderson’s models are normative: that is, they represent a settlement pattern that is efﬁcient given a set of implicit and explicit assumptions. That actually observed settlement patterns do not conform to the theoretically expected patterns do not ‘disprove’ the theories, since the differences may be due to violations of the underlying assumptions. Thus labour is not perfectly mobile (and, indeed, nor necessarily is capital); transportation costs are not a smooth function of distance; there are differences in topography, resource endowment and environment; there are distortions caused by political boundaries and, above all, there is the legacy of the past, the historic pattern of settlements that shapes transportation patterns, population distribution and political jurisdiction and establishes a pattern of land-use and built form. Nonetheless, to have any use and validity, the theoretical models should be translatable into positive models that demonstrate explanatory power in examining the pattern of cities in space.

Empirical evidence What, then, is observable? First, cities in a region are of different sizes and do perform different functions, with high-order retail, service and governmental functions concentrated in the largest cities at the top of the urban hierarchy. Second, in developed economies at least, the distribution and rank

12

Towers of Capital

of cities by size remains remarkably stable over long periods of time, even in the face of substantial economic restructuring ( Eaton & Eckstein, 1997; Duranton & Puga, 2000; Combes et al., 2005). Third, larger cities perform a range of functions, but often take on specialist characteristics. The evidence on the performance of specialist versus diversiﬁed cities is mixed. For North American and UK cities, there is evidence that ﬁrm growth, ﬁrm turnover and innovation is faster in diversiﬁed cities and growth seems more linked to diversity and human capital within cities than to specialised employment structures (e.g. Malpezzi et al., 2004, suggest using shift-share analysis that growth is more closely associated with urban level agglomeration economies than with employment structure). This tends to support a Jacobs view of the city as an arena for innovation and exchange, but the results may be distorted by the restructuring of manufacturing in the 1980s and the ‘deindustrialisation’ of Western economies. Against that, Markusen and Schrock (2006) argue that city growth for second-tier cities in the United States is associated with ‘distinctiveness’ in employment structure, consumption and identity. They suggest this results from competition between cities and public–private attempts to distinguish and market individual cities as the relationship between city and immediate hinterland breaks down. They also present evidence that while some high-level service functions are clustered in the largest cities, others – they cite computing, engineering, life and physical sciences, architecture – appear to favour smaller cities (albeit that some of these specialist clusters appear to be relatively close to cities at the top of urban hierarchy). The idea of specialised cities lies behind attempts to classify cities using clustering techniques. For the United States, for example, Noyelle and Stanback (1984) separate cities into two main categories: ‘command and control’ cities and subordinate cities (see Figure 1.3). The command and control cities have high proportions of headquarters (HQs) buildings, concentrate ﬁnance, high-level producer services and governmental functions. This, clearly, has implications for the built form of cities, in that the presence of such functions demands a mass of high-quality, large ofﬁce space. Command and control cities are further subdivided into diversiﬁed or specialised service centres and then by scale and function: hence national diversiﬁed cities include New York and Los Angeles; Atlanta would be a regional diversiﬁed service city, while Washington, DC would be a specialised government and education city. In turn, subordinate cities are separated into those focused on consumption and those focused on production. The idea of command and control functions being increasingly concentrated in a small number of large, powerful cities is the basis of the ‘global cities’ literature discussed in Section 1.3. Duranton and Puga (2005) extend the concept of city specialisation by noting that increasingly ﬁrms have been separating their management and

The Urban Hierarchy and Global Cities

13

National Diversified service

Regional Sub-regional

Command and control Functional Specialised service

Government Education Education Manufacture

Residential Consumer Resort Retirement Subordinate Manufacturing Production

Industry Military Mining Industrial

Figure 1.3 The US urban hierarchy. Source: Adapted from Noyelle and Stanback (1984) and Ball et al. (1998).

production facilities. The former, beneﬁting from agglomeration economies by co-locating with other management ofﬁces, are increasingly clustered in larger cities, while smaller cities have become more specialised in production. They cite evidence from Kim (1995) that stand-alone HQs increased in the United States by 79% between 1958 and 1987, while employment in those separated HQs rose by 69%.1 Their own analysis of functional specialisation by city size for 1950 and 1990 shows clear evidence of specialist control and management functions up the urban hierarchy (see Figure 1.4). At the same time, there was evidence that sectoral specialisation (at least at two-digit SIC level) was declining although smaller cities

1

For an international perspective, see Markusen (1995).

14

Towers of Capital

Relative management strength

50 40

1990

30

1950

20 10 0 −10 −20 −30 −40 −50 −60

<75k

75k–250k

250k–500k 0.5–1.5 million 1.5–5 million

>5 million

Size of city

Figure 1.4 Functional specialisation in US cities. Source: Estimated from Duranton and Puga (2005), Table 1. Figures show percentage difference in ratio of managers to production workers to national average ratio.

remained more specialised than larger cities. They note that their results have been replicated for Germany. Why might this functional specialisation occur? Increasingly, HQs buildings outsource business service functions and so location decisions will be inﬂuenced by the existence of concentrations of business services and, as noted above, that concentration allows further specialisation in the services offered. There are labour market effects, too, which will be explored in more depth later. This will tend to create a self-reinforcing cycle: the greater the clustering of HQs buildings, the more attractive the city is for business service ﬁrms. Shilton and Stanley (1999) found that 40% of their sample of 5000 HQs in the United States were found in just 20 major urban agglomerations. This clustering alone, though, is not sufﬁcient to explain the separation of HQ and production plant at ﬁrm level. There are costs to the separation: the costs of communicating, monitoring and coordinating across multiple locations. These must be set against the efﬁciency gains from locating the HQ in a core metropolitan area with a good supply of business and producer services. The efﬁciency gains must outweigh the coordination costs and the additional congestion and crowding costs of locating the HQ in a large city. Those congestion and crowding costs preclude locating the production plant in the larger city: these will tend to locate in lower cost locations and also (in general) in relatively specialist cities to capture localisation economies and reduce the costs of production. Since there appears to have been a historical shift, the most likely explanation is that there has been a reduction in the costs of coordination over the last quarter of the twentieth century: through

The Urban Hierarchy and Global Cities

15

increased speed (and reduced real cost) of individual travel, through changes in management practices but, above all, through improvements in communications technology that have dramatically reduced the costs and difﬁculties of passing information and instructions around the ﬁrm. These shifts in the nature of activities in cities have direct implications for land-use and property. With cities as the base for manufacturing activity, ofﬁces play a role as the base for the internal and administrative services of those organisations (and, separately, to coordinate trade between ﬁrms). As manufacturing activity and other primary functions move out of city centres, the ofﬁce activities often remained – with a ﬁrst wave of separation of HQs and production plants following the development of telegraph, telephone and national transport systems and a second wave with twentiethcentury communications, which enabled the decentralisation of lower order ofﬁce and service activities. Thus that economic activity not linked to consumption in the centre of cities became more specialised and more dominated by ofﬁce space.2 Once again, this appears to have signiﬁcant implications for the occupational demand for space: major ofﬁce requirements occur in large diversiﬁed cities (and, as we will see, in the centre of large diversiﬁed cities), while ofﬁce requirements in smaller, specialist production cities will tend to reduce. They will not disappear. In addition to branch ofﬁces, the producers of intermediate goods and services in the production process themselves require business services – accounting, legal, taxation, ﬁnancial services, albeit not at the same scale as demanded by the clusters of HQs buildings. This leaves one unanswered question: if communications technologies have improved so much and have had such a signiﬁcant cost effect, why are the economies of scale and scope from outsourcing business and producer services conditional on location and proximity? Why is it not possible to source such high-level services from multiple locations? That question is addressed later.

Cities outside states: a global hierarchy? The traditional models of the urban hierarchy and of the relationship between population and rank of the city discussed above have generally been conceived in relation to national states or, if across states, to deﬁned regional groupings. Many of the measures used in empirical research are similarly deﬁned by national boundaries (for example, the denominator in location quotients or specialisation indices is usually the national employment share by sector). In an era of strong state control over population, goods and capital ﬂows, this approach might have a logical justiﬁcation. In an era 2

Thanks are due to Steve Devaney for this point.

16

Towers of Capital

characterised by more open economies, rapid ﬂows of capital and goods, ease of international transport and global telecommunications networks, conﬁning analysis to the nation state seems inappropriate. How do urban hierarchy arguments play out on a global scale? United Nations ﬁgures for 2000 (UN, 2004) suggest that 48.3% of the world’s population live in urban agglomerations. In more-developed regions, 74.5% of the population is urban – although, in practice, and given extended commuting patterns, the proportion that is in practice ‘urbanised’ may be higher still (for example, residents of commuter villages travelling into cities for work are clearly urbanised). In the less-developed regions, just 42% of the population is urban – and within those regions, 73% of the population resident in those countries classiﬁed as ‘less-developed’ reside in rural areas. Table 1.1 shows the 20 largest urban agglomerations in 2000 and in 1975 (such tables depend critically on the deﬁnition of agglomeration and urban boundary, where administrative boundaries fail to adapt to growth and

Table 1.1 Largest urban agglomerations, 2000 and 1975. Rank and City (2000)

Population (millions)

Rank and City (1975)

Population (millions)

1. Tokyo

35.3

1. Tokyo

26.6

2. Mexico City

19.0

2. New York – Newark

15.9

3. New York – Newark

18.5

3. Shanghai

11.4

4. Mumbai

18.3

4. Mexico City

10.7

5. Sao Paulo

18.3

5. Osaka Kobe

9.8

6. Delhi

15.3

6. Sao Paulo

9.6

7. Calcutta

14.3

7. Buenos Aires

9.1

8. Buenos Aires

13.3

8 Los Angeles – Laguna Beach

8.9

9. Jakarta

13.2

9. Paris

8.6

10. Shanghai

12.7

10. Beijing

8.5

11. Dhaka

12.6

11. Calcutta

7.9

12. Los Angeles – Laguna Beach

12.1

12. Moscow

7.6

13. Karachi

11.8

13. London

7.6

14. Rio de Janeiro

11.3

14. Mumbai

7.4

15. Osaka-Kobe

11.3

15. Chicago

7.2

16. Lagos

11.1

16. Seoul

6.8

17. Beijing

10.8

17. Rhein-Ruhr

6.4

18. Manila

10.7

18. Cairo

6.4

19. Moscow

10.7

19. Tianjin

6.1

9.9

20. Milan

5.5

20. Paris

Source: Data taken from United Nations (2004), World Urbanization Prospects.

The Urban Hierarchy and Global Cities

17

absorption of surrounding settlements in the urban system). Agglomerations moving into the top 20 are Delhi, Dhaka, Lagos, Manila and Rio de Janeiro – all from developing regions. The UN estimates that by 2015, 54% of the world’s population (and 77% of the developed regions’ population) will live in urban areas. However, these major urban areas by no means contain a high proportion of the world’s population: the largest 20 urban agglomerations account for around 4%, with a further 3% living in cities of between ﬁve and ten million. While many of the cities that are considered to be at the top of the urban hierarchy feature in the UN’s list of the largest agglomerations, it is clear that population alone does not determine global status. PricewaterhouseCoopers (2007) have attempted to estimate the GDP of urban agglomerations. Table 1.2 lists their estimate of the top 30 cities ranked by GDP in US$ at purchasing power parity. Sixteen of the 30 are US cities. Their projections to 2020 have this falling to 12 of 30, with Beijing, Istanbul, Manila, Mumbai and Shanghai entering the list. This ranking is probably closer to most lists of the leading cities in the world economy with some obvious exceptions (Amsterdam, Frankfurt, Singapore, Sydney, Zurich, for example) and emphasises the importance of wealth and earnings in determining global status. Table 1.2 also maps more closely onto investment interest in real estate in general and in ofﬁces in particular. Population alone does not generate

Table 1.2 Cities ranked by GDP. City

GDP $bn @ ppp

City

GDP$bn @ ppp

1. Tokyo

1191

16. Atlanta

236

2. New York

1133

17. Houston

235

3. Los Angeles

639

18. Miami

231

4. Chicago

460

19. Sao Paulo

225

5. Paris

460

20. Seoul

218

6. London

452

21. Toronto

209

7. Osaka-Kobe

341

22. Detroit

203

8. Mexico City

315

23. Madrid

188

9. Philadelphia

312

24. Seattle

186

10. Washington, DC

299

25. Moscow

181

11. Boston

290

26. Sydney

172

12. Dallas–Fort Worth

268

27. Phoenix

156

13. Buenos Aires

245

28. Minneapolis

155

14. Hong Kong

244

29. San Diego

153

15. San Francisco Oakland

242

30. Rio de Janeiro

141

Source: Data taken from PricewaterhouseCoopers (2007).

18

Towers of Capital

demand for large, high-speciﬁcation ofﬁce space – that requires the presence (actual or desired) of major ﬁnancial, business or producer service ﬁrms, the international or regional HQs of major corporations and/or the presence of signiﬁcant government departments. Only such ﬁrms can generate the turnover to be able to afford the rents that underpin ofﬁce market development, ﬁnancing and investment. By implication, this suggests that an understanding of the growth and development of a global real estate market requires an understanding of the evolution of the global urban hierarchy.

1.3

World cities, global cities

Since the 1980s, a body of work has attempted to map and explain the development of a global urban hierarchy and the key role of a limited number of major cities in coordinating and controlling an international network of ﬂows of capital, goods and workers: world or global cities. Although – as will be explored below – the idea of the existence of an elite group of cities with a dominant role in the world economy is by no means new and although the ideas underlying the world city model have been subject to trenchant criticism, an initial review of that literature provides a useful starting point for a consideration of the global urban hierarchy. If such a global hierarchy does exist, then the highest level of producer and business services should be located in those leading cities that are at the pinnacle – with consequences for the development of, and investment and occupational demand for, ofﬁce space therein.

Friedman and Sassen The world city/global city literature generally cites the work of John Friedman and Saskia Sassen (with the early contributions of Anthony King often cited but rarely developed) as the basis for theories of an international system of cities. From Friedman, we have the idea of a ‘world city hierarchy’; from King ‘a global network of cities’, and from Sassen ‘global cities’ and the ‘transnational urban system’. Sassen’s ideas, in particular, have become important in urban debates and feature in governmental reports and even in private consultancy brieﬁngs (see, for example, Mastercard Worldwide, 2007), despite their origins in critical social science and their focus on social polarisation and equity issues in cities. The essential idea, prominent from the late 1980s, is that cities or urban regions form part of a global hierarchy with position determined by the degree of integration into, and inﬂuence on, the global system of capital ﬂows. Friedman’s starting point (Friedmann & Wolff, 1982; Friedmann, 1986) is the idea that major corporate ﬁrms, operating across the globe, have freed

The Urban Hierarchy and Global Cities

19

themselves from national constraints and organise production on a global, not a national, scale. This draws on work on the organisation of multinational ﬁrms (e.g. Hymer, 1979), on ideas of a ‘new international division of labour’ and on the concept of a ‘world system’ found in the work of Wallerstein (1974). Enabled by communications technologies and the erosion of barriers to trade and capital ﬂows, private ﬁrms’ operations are ‘no longer’ conﬁned to national markets but are organised on a multinational basis, with a division of labour and activity expressed geographically: the separation of resourcing extraction, production and assembly, sales and management, coordination and ﬁnancing across national borders. Firms operate internationally by virtue of their size and access to capital and grow organically, or through mergers, acquisitions and strategic alliances. The controlling HQs and ﬁnancial/ownership bases of these transnational corporations (TNCs) are spatially concentrated in powerful and wealthy ‘Northern’ states – Wallerstein’s core states, with a semi-periphery and periphery of states providing labour, resources and market outlets for the TNCs. The key feature here is that within a world economy or world system there is a division of labour that transcends national boundaries. The urban manifestation of the growth of transnational ﬁrms is the concentration of HQs of multinationals – and of service providers for those HQs – in a small number of cities in the core regions/states. The spatial dominance of those cities in terms of control of ﬁnance and production is expressed over a global scale, or over a multinational regional area, not conﬁned to national boundaries. These, then, are ‘world cities’. Friedman argues that this is qualitatively different from the role played by imperial cities in history: the world economy is no longer deﬁned by imperial reach but by a linked set of markets and production units organised and controlled by multinationals, concentrated in cities for economic reasons. Friedmann and Wolff (1982) note that, in addition to acting as a location for transnational HQs, world cities also act as a ‘safe haven’ for capital in the form of real estate investment. Unfortunately, this concept is not developed in their later work, and the real estate investment dimension has largely been ignored in the world city and global city literature with few exceptions – for example, Fainstein (1993) and Haila (1997) both emphasise the production of the built form – with most mentions of ofﬁces treating them as a by-product, a derived demand or as a symbolic representation of the status of the city, rather than as a store of value and as both a cause and a consequence of the ﬂow of capital around the global urban system. Presaging Sassen’s research, Friedman and Wolff note that the development of an urban area as a world city must entail an economic restructuring with the cities needing to accommodate a growing ‘transnational elite’ of management, legal, ﬁnancial and technical professionals to support the HQs operations.

20

Towers of Capital

Although Friedman and Wolff stress the need for empirical research to establish which cities were world cities and to explore the linkages between them, Friedmann (1986) sets out candidates for world cities. The world is divided into Wallerstein’s Core, Semi-periphery and Periphery and, within the Core and Semi-periphery, cities are divided into Primary and Secondary world cities: Core: Primary Core: Secondary Semi-periphery: Primary Semi-periphery: Secondary

New York, Chicago, Los Angeles, London, Paris, Tokyo Toronto, San Francisco, Miami, Houston, Madrid, Milan, Vienna, Sydney, Johannesburg Singapore, Rio de Janeiro and Sao Paulo Mexico City, Caracas, Buenos Aires, Manila, Bangkok, Hong Kong, Taipei, Seoul

These 26 cities (Rio de Janeiro and Sao Paulo are shown as a single city) are drawn with linkages shown between core cities and ‘other linkages’ (see Figure 1.5). Friedman suggests that this creates a ‘linear character of the world city system which connects, on an East–West axis, three distinct subsystems: an Asian subsystem based on the Tokyo–Singapore axis … an American system based on the primary core cities of New York, Chicago and Los Angeles … and a West European subsystem focused on London (and) Paris’. This interpretation relies on the validity of the drawn linkages, which must be open to considerable doubt. Thus, Sydney’s connections to the primary cities of the core region are to Tokyo and Los Angeles (and not to London); Buenos Aires connects to the Brazilian cities and then to New York (not Miami) and to London (not Madrid); there is no link between Hong Kong and London, or between Hong Kong and Toronto, for example. The linkages are asserted and are not supported by evidence. With the beneﬁt of hindsight, it is also striking that ‘the world’ is a constrained place in Friedman’s presentation. India and China are unconnected; to the East of Vienna, three years before the fall of the Berlin Wall, is the terra incognita of the Soviet Union; the Middle East does not feature at all. If nothing else, this demonstrates that the world city system cannot be a static concept, but must evolve with developments in the political economy of the world. This would be consistent with Wallerstein’s approach which sees the world system evolving and changing over long periods of time, starting in the sixteenth century and characterised not only by expansion of inﬂuence over larger spatial areas but also by signiﬁcant changes in power, position and inﬂuence. However, the lack of an evidential base and questions over some of the asserted linkages cast doubt on the robustness of Friedman’s initial analysis. Sassen’s contributions to the debate on global cities (e.g. Sassen, 1991, 2000) begin with a focus on the cities at the very top of the urban

Sydney

Semi-periphery: secondary

Semi-periphery: primary

Core: secondary

Core: primary

Singapore

Manila

Tokyo

Buenos Aires

Mexico City

Houston

Other linkage

Core linkage

Los Angeles

San Francisco

Chicago New York

Rio de Janeiro & Sao Paulo

Caracas

Miami

Toronto

Figure 1.5 Friedman’s hierarchy of world cities. Source: Adapted from Friedmann (1986).

Bangkok

Hong Kong

Taipei

Seoul

Madrid

London

Johannesburg

Milan

Paris Vienna

The Urban Hierarchy and Global Cities 21

22

Towers of Capital

hierarchy: the global cities of London, New York and Tokyo. These three cities play a critical role in the global economy because they function as ‘command and control points’ – as key locations for ﬁnance and for the essential producer services required by TNCs and as sites for the production and consumption of innovations in the organisation and ﬁnancing of economic activity. Two trends drive this process: the globalisation of economic activity and the organisation of the ‘producer services’ industries (which include accounting, advertising, banking and ﬁnance, legal services, management consultancy and marketing). Economic drivers include the shift to services and ﬁnance in the core economies and a focus on control of production downstream; the rapid growth and specialisation of producer services and spatial clustering for agglomeration economies lead to a concentration of command and control activities in a small number of cities of which London, New York and Tokyo are the most important. Sassen produces descriptive evidence of the concentration of producer services within those city-regions. One insight from Sassen’s emphasis on the key role of global producer services is the emphasis on global ﬂows – not so much of goods as of capital, information, people – around the world urban network. This extends the global city concept beyond a mere concentration of multinational HQs in particular cities and provides a link with social theories such as Castells’s ‘network society’. For Castells, the rise of information technology has led to a signiﬁcant shift in the organisation of the economy on a global scale. He envisages three levels; the ﬁrst is the material basis for the ﬂows: this would include global internet linkages (discussed in more detail in Chapter 3), airline and rapid transit systems and so on; the second consists of the ‘nodes and hubs in the space of ﬂows’ – from which would spring the multinational enterprise (MNE) generated system of world cities; the third is the spatial organisation of the managing elite, the geography of the workers in MNEs and their associated business, ﬁnancial and professional services. The emphasis on ﬂows and networks provides a theoretical base for understanding a world urban hierarchy. However, the evidence produced for this is largely anecdotal and informal. Much of the thrust of Sassen’s work has emphasised the social impacts of urban restructuring in global cities. In Sassen’s model, the growing concentration of high-level producer service ﬁrms require space to function and an elite professional and managerial workforce. Manufacturing and lower-level business service ﬁrms are unable to compete for space and are driven from the cities. Urban redevelopment creates new workspaces and new residential areas for the producer service ﬁrms while the presence of the ﬁrms and their staff produce a demand for low-skill, low-wage services, often accompanied by casualisation of the workforce. The result is a rapid social polarisation of global cities, with a hollowing out of middle-income groups and with much

The Urban Hierarchy and Global Cities

23

of the economic gains captured by the ‘new’ elite – a growing inequity in the distribution of income and wealth. Much of the subsequent development of the global cities idea has emphasised this social dimension. This is, in part, an explanation of the neglect of the real estate investment component – in that new development is viewed almost exclusively in terms of planning conﬂict and its symbolic role in the image of the global city, rather than in terms of its role as a store of value, as an investment asset class and as a determinant of location and agglomeration within the city itself. Another theme in the later developments of Sassen’s work concerns the globalisation of major cities in low- and middle-income countries: Sao Paulo, Buenos Aires, Bangkok, Taipei, Shanghai, Manila, Mexico City.

The evidence base for world cities Short et al. (1996) argued that much of the literature on world city was characterised by a lack of evidential base. They concede that the study of the global urban system is hampered by lack of available data but suggest that the literature to that point had done little to confront the problem. As a result, any assertions about relationships or policy suggestions must be, at best, preliminary. Knox (1998, cited in Beaverstock et al., 2000) makes a similar point: ‘... few of the available data reveal anything about the ﬂows and interdependencies that are at the heart of the idea of world cities as basing points for transnational capitalism’. Beaverstock et al. (2000) stress this relational point: the global cities thesis emphasises ﬂows around the world urban system – but most ﬂow data are collected at national level, not at city level. They identify a number of possible direct research avenues but point to the difﬁculties posed by each of them. It is possible to identify proxy measures: airline ﬂights, telephone communications and, more tenuously, internet connections, which provide some indication of the interconnectedness of cities, but there is limited empirical testing of the world cities model. In a series of papers, Taylor and co-workers (grouped around the Globalisation and World Cities (GaWC) project at Loughborough University) have sought to explore the linkages between world cities based on the common presence of major TNC and producer services ﬁrms. Their work – which emphasises ﬂows between cities – is perhaps theoretically closer to Castells’s (1989) ‘space of ﬂows’ than Sassen’s global cities, although much of the empirical work seeks to identify and classify cities, albeit in terms of their interconnectedness. The starting point is a database of leading corporate service providers from four sectors – accountancy, advertising, banking and law. For the ﬁrms in their sample, they seek to identify their presence in different cities (the existence of HQs, regional HQs, branch ofﬁces and afﬁliates) and to assess the signiﬁcance of that presence in each city. It is

24

Towers of Capital

argued that the presence of ofﬁces of major ﬁrms in multiple cities acts as a proxy for information ﬂows between those ofﬁces and, hence, helps measure integration and signiﬁcance in the global urban system. This then allows them to estimate the connectivity and signiﬁcance of each city in terms of corporate services. From the perspective of this book, the approach has the advantage of focusing on ofﬁces – although the actual built form is strikingly absent from the GaWC publications. Beaverstock et al.’s (1999) paper uses the GaWC database to identify a ‘roster of world cities’. Using a relatively simple taxonomic approach, they divide cities into four types: alpha, beta and gamma world cities (which are deﬁned in terms of the relative signiﬁcance of the service presence in each of their four service sectors) with sub-groupings within those types, and cities that ‘show evidence of world city formation’ – cities where there is some indication of the presence of global service providers. The cities included in their estimations are shown in Table 1.3. While there are no major surprises in the alpha group of world cities, it is interesting to see the emergence of Moscow, Prague, Bratislava, Beijing and Shanghai in the table from the terra incognita in Friedman’s original world city hierarchy, along with the emergence of Indian and Middle Eastern cities as showing world city characteristics. Based on the evidence of the GaWC data, there appears to be still greater integration of the global urban system – although it must be emphasised that the presence of a corporate service provider is not evidence of a command function, nor any guarantee of growth of wealth and capital in that city. In subsequent work, the GaWC database has been augmented and reanalysed. Taylor et al. (2002) construct a relational measure that enables them to measure interconnection between cities: the ten most connected cities Table 1.3 World cities by corporate service provision. Alpha world cities

London, Paris, New York, Tokyo Chicago, Frankfurt, Hong Kong, Los Angeles, Milan, Singapore

Beta world cities

San Francisco, Sydney, Toronto, Zurich Brussels, Madrid, Mexico City, Sao Paulo Moscow, Seoul

Gamma world cities

Amsterdam, Boston, Caracas, Dallas, Dusseldorf, Geneva, Houston, Jakarta, Johannesburg, Melbourne, Osaka, Prague, Santiago, Taipei, Washington, DC Bangkok, Beijing, Rome, Stockholm, Warsaw Atlanta, Barcelona, Berlin, Buenos Aires, Budapest, Copenhagen, Hamburg, Istanbul, Kuala Lumpur, Manila, Miami, Minneapolis, Montreal, Munich, Shanghai

Evidence of world city formation (selected cities)

Almaty, Bratislava, Cairo, Dublin, Delhi, Dubai, Luxembourg, Mumbai, Rio de Janeiro, Riyadh

Source: Adapted from Beaverstock et al. (1999).

The Urban Hierarchy and Global Cities

25

are the alpha cities of the Beaverstock et al. paper, although on certain measures Madrid performs strongly. They suggest that the results imply, rather than a hierarchical structure, a complex, interconnected network, with London and New York distinct and central. Washington, DC is shown to be strongly signiﬁcant with respect to legal ﬁrms, but more weakly connected along other dimensions. Taylor and Walker (2001) apply multivariate exploratory analysis to the database, using principal components analysis. Distinct regional groupings emerge (which do not conform clearly to a core– periphery model): a ‘world cities’ component features high factor loadings for New York, London and Los Angeles but not for Tokyo or Paris (which only has high loadings on ‘European’ factors, just as Chicago only loads on North American factors). Derudder et al. (2003) apply a fuzzy clustering technique that produces a hierarchical and regional structure around the core central cities of London and New York. The ‘Band I’ cities – those most connected outside of the core – are divided into three groups: Frankfurt, Hong Kong, Paris (B), Singapore and Tokyo; Chicago, Los Angeles and San Francisco (C); and a larger cluster consisting of Amsterdam, Milan, Madrid, Zurich, Mexico City, Sao Paulo, Buenos Aires, Sydney and Toronto (D). Brussels and Barcelona fall between groups B and D. Derudder and Taylor (2005) apply a connectivity analysis to deﬁne sub-groups or cliques of cities, clusters of cities with strong interconnections. The ﬁndings emphasise the importance of London and New York (the core of all cliques). Other cities in this analysis, notably Frankfurt, Brussels and Washington, DC, appear more important than aggregate service ﬁrm presence might indicate. Tokyo and Hong Kong appear to be strongly linked with other Asian cities while minor US cities have weaker regional and international links than have European cities. Taylor and Aranya (2008) produce yet another GaWC analysis with an updated 2004 database. Compared to 2000 rankings, there is considerable stability: the top six cities are unchanged and the shifts in the top 20 are relatively minor. Toronto advances to seventh while a number of US cities fall in rank. In terms of connectivity, Western European cities rise, US cities fall and there is tentative evidence of an increasing connectivity for Eastern European and Middle Eastern cities. The different GaWC analyses produce distinct results but a common thread of a global core of London and New York and then a structure that is both hierarchical and regional in nature. Of course, all these analyses rest on the data collected: both the validity of the ofﬁce location measure as a proxy for connectivity and the representativeness of the service producers included in the analysis (and speciﬁcally the extent to which, for example, Asian producer service ﬁrms are fully represented in the database). Another line of investigation has been to examine the interconnections between cities based on airline trips. World cities should be characterised

26

Towers of Capital

by a high number of international journeys and linkages with other highorder cities in the global hierarchy. This approach was adopted by Smith and Timberlake (2001) and Alderson and Beckﬁeld (2004). Derudder and Witlox (2005) identify some of the problems faced by the use of airline journey data. First, a focus on international journeys (partly driven by data) misses out key links – for example, New York–Toronto journeys would be included but New York–Chicago or New York–Los Angeles would be excluded. Second, aggregate trip data does not distinguish between business and tourist ﬂights (of course, the facilities that act as attractors for the managerial and professional elite also act as a tourist attractor, so there will be an overlap) – which explains why, for example, Orlando or Mallorca appear much more important than their signiﬁcance in the global economy might suggest. Third, airline hub and spoke models tend to emphasise gateway cities (it is difﬁcult to ﬂy to Germany from the United States without ﬂying ﬁrst to Frankfurt, Milan playing a similar role in Italy). It may be that the airline hubs are, or will become, linked to global city status, but this does represent a data distortion. Derudder and Witlox use a database derived from on-line bookings through global distribution systems (such as Galileo or Sabre) and examine the linkages among over 300 cities. They examine aggregate trips and trips between pairs of cities (see Tables 1.4 and 1.5). The list of leading cities by Table 1.4 Airline trips: leading city pairs. Rank

City pair

Rank

City pair

1

Hong Kong–Taipei

6

Cape Town–Johannesburg

2

Los Angeles–New York

7

Los Angeles–San Francisco

3

London–New York

8

Amsterdam–London

4

Melbourne–Sydney

5

Milan–Rome

9 10

Chicago–New York Bangkok–Hong Kong

Source: Adapted from Derudder and Witlox (2005).

Table 1.5 Rank 1 2 3 4 5 6 7 8 9 10

Airline trips: leading cities, aggregate trips. City

Rank

City

New York London Los Angeles Paris Chicago San Francisco Hong Kong Miami Frankfurt Atlanta

11 12 13 14 15 16 17 18 19 20

Madrid Toronto Washington, DC Rome Bangkok Milan Amsterdam Dallas–Fort Worth Boston Singapore

The Urban Hierarchy and Global Cities

27

aggregate bookings is similar to that produced by Smith and Timberlake (2001) but with some striking differences – notably the higher ranking for the major US cities as a result of the inclusion of intra-national regional trips – with Atlanta, Washington, DC, Dallas and Boston appearing on the Derudder and Witlox top 20 but not in Smith and Timberlake. Of international cities only appearing in the top 20 in one of the rankings, Derudder and Witlox include Rome, Milan and Toronto on their list, Smith and Timberlake include Seoul, Vienna, Dubai, Osaka, Tokyo and Brussels. The map of airline links has some similarities to Friedman’s world city hierarchy, with London dominant within Europe (with Paris acting as a second hub), New York the main centre in North America with a strong New York–Los Angeles internal link, the latter city exhibiting strong links to the Asian cities. The Asian cities seem to be less interconnected than the European and North American cities, although it is not possible to discern whether this is a real phenomenon or a manifestation of the data. Tokyo, regularly cited as one of the three key cities in the global cities literature, is ranked eleventh by Smith and Timberlake and does not feature in the Derudder and Witlox top 20.

Questions, anomalies and critiques There is a sense that the world urban system/global cities thesis lacks a sense of history. What is genuinely new about the world cities models proposed by Friedman and Sassen? Wallerstein’s world systems analysis begins by suggesting that the system has been evolving from the sixteenth century. Before that, empires existed with, at their core, controlling cities – the Alexandrian and Roman empires, the Ottoman empire after its initial expansion, the Chinese expansion all had, at their heart, systems of cities acting to coordinate and control, some more powerful than others. Drawing on Braudel and other authors, it has been suggested that the material difference between those imperial situations and the current world system is that the former were bounded by the limits of state power, while the latter exists independent of states. Yet Wallerstein’s world system includes the period of Pax Britannica and the role of London, superseding Amsterdam – as part of (and ﬁnanced by) a substantial shift in the international division of labour. Wallerstein’s core–periphery structure, adopted by Friedman, has echoes of Myrdal and Latin American dependency theory, which saw the development of urban systems and transport networks in the economic south as being linked to colonial trade routes. Peter Hall (1966) suggests that Patrick Geddes used the term ‘world cities’ in 1915. Hall’s description of world cities – in part reprinted as the prologue to the Global City Reader (Brenner & Keil, 2006) – talks of such cities in terms of political power, trade, wealth, and also of their attraction for

28

Towers of Capital

professional talent, arts, media, culture, consumption and as cosmopolitan places of innovation. The description will have resonances for anyone familiar with the global cities literature. Yet it is applied not just to the cities of the present, but to cities of the past. Hall emphasises continuity – not in the activities in world cities which evolve in line with the shifts in the global economy, but in their power, status and prestige. He notes that world cities may decline – he cites Bruges – but that the most striking feature is continuity. It is hard to reconcile this sense of continuity with the structural break that seems implicit in much of the world cities literature. Descriptions of many of these cities historically stress their cosmopolitan nature: New York is, almost by deﬁnition, a city of immigrants, but waves of immigration and the establishment of businesses by non-nationals shape the development of London, Amsterdam, Paris, Venice and Genoa. So, the idea of the city as a node in a global trading network and as a cosmopolitan market for ideas is hardly a new one. The more recent shift towards service industries in the developed economies undeniably changes the space requirements for cities – with a growing need for ofﬁce space in the centre and, increasingly, at the periphery of cities. As noted above, the growth of a strong business, ﬁnancial and professional service sector has been accompanied by specialisation and outsourcing of activities that would traditionally have been performed within a ﬁrm or corporation. These specialised activities concentrate higher up the urban hierarchy. But is this qualitatively different from the past? What ﬂows between these ﬁrms is capital and information, so any concentration of activity is predicated on the existence of communications networks that are fast and reliable. Technology enables the rapid transmission of commands and capital around the world urban network. Ironically, though, the development of advanced telematic systems would seem to reduce the need for proximity, to free ﬁrms from expensive, congested city locations. We return to this topic, and the importance of agglomeration economies in high-order business and ﬁnancial services activities, in Chapter 3. Another difference asserted is the growing importance of MNEs or TNCs. Again, how new is this? The Dutch East India Company was founded in 1602, the Muscovy Company 50 years before that. Trading companies such as these established trading posts across wide geographical areas, often functioned as the de facto government in their areas of operation and have been argued to have many of the features of multinational companies (Carlos & Nicholas, 1988). More recently, the idea that MNEs and TNCs are instrumental in shaping urban restructuring and in the creation of global cities has been called into question, with much recent research on ‘resurgent cities’ stressing the role of local, regional and national state institutions in enabling and shaping urban change – albeit a role that calls for partnership with private capital and that gives a priority to economic competition over

The Urban Hierarchy and Global Cities

29

social equity (as an example, see the papers in the special issue of Urban Studies devoted to resurgent cities and urban policy – Cheshire (2006)). This suggests that state involvement and public–private partnership produces divergent outcomes that work against a convergence and homogenisation of cities in the global system. The world city/global city model has been accused of having a Eurocentric/ Atlantacist bias – in particular that developments are viewed from a North American and UK frame that sees a decline in the role of the state and national boundaries as an inevitable global trend (e.g. King, 1990; Hill & Kim, 2000; Waley, 2002). This raises questions about the status of Asian cities in the global city model. Friedman included Tokyo in his preliminary list of ﬁrst-tier world cities as a ‘major site for the concentration and accumulation of international capital’ but did express some reservations about whether it ﬁtted with the world hierarchy model. Sassen, in her earlier writings, suggests that Tokyo is unique but that it is likely to converge. However, Tokyo – and other Asian cities including Seoul and Singapore – does appear to be different in character both in the extent of international penetration and in the role of the state. Hill and Kim, in particular, are highly critical of Sassen. They argue that ‘Tokyo departs from the world city paradigm on most salient parameters … [it] has many fewer foreigners and many more factories than do either London or New York … global capital is not a major factor’. They argue that in Tokyo in particular, but also in many other Asian states, ‘developmentalist’ state institutions have played a strong role in urban restructuring and economic restructuring. Within Japan, the strong links between the Ministries of Finance and International Trade & Investment, the keiretsu networks and the interlinkage of Japanese banks, manufacturing ﬁrms and politicians have both ‘protected’ Japan from foreign penetration and enabled the modernisation of the economy and the physical and social restructuring of the Tokyo region. Machimura (1992), similarly emphasises the role and inﬂuence of local and national state institutions in shaping the transformation of urban functions and land-use. Waley (2007), while arguing that private capital plays a more important role in the restructuring of Tokyo than is often suggested, emphasises that this is predominantly Japanese capital, derived from the ability of Japanese corporations to earn proﬁts domestically and internationally, linked to the state by the informal ties and structures that create a distinctiveness that does not ﬁt well with the ‘North Atlantic’ model of a global city. Similar reservations can be expressed for other Asian cities typically included in the tables of world cities. Hill and Kim note the strong manufacturing base of Seoul and the strength of major corporations (Samsung, Daewoo, Hyundai and LG) as sources of capital and reinvestment. Olds and Yeung (2004) seek to distinguish city states like Singapore from cities

30

Towers of Capital

embedded within larger nations: they argue that the enhanced ability of a government that is both the national and the city authority to organise, enable and direct activity (and protect indigenous business) means that global processes have different outcomes. Olds and Yeung include Hong Kong alongside Singapore in their city-state model, but that must be open to question. The historic ties to Britain up to handover and its special status within the People’s Republic of China after handover and, hence, its role as an international entry point to the evolving Chinese market suggest a much greater integration into a world economic system. Massey (2007), in echoing these criticisms, also questions the criteria used to select global cities – speciﬁcally the importance given to ﬁnancial and business services in their role as commanding and controlling production. She points out that other cities are dominant in other areas: Hollywood and Mumbai for ﬁlm, cities as the focus for particular religions3: ‘we need to value and build on the diversity between cities’. She goes on to argue that in creating a very speciﬁc stereotype of a world city – in effect modelled on London and New York – city governments and managers are pressurised into striving for the same thing, which can distort planning policy and economic policy. Nonetheless, it is clear that the world city concept is a powerful one shaping both research and practice in the urban realm. Speciﬁc aspects and predictions concerning social structure within the global cities model have been subject to analysis and critique. Sassen’s model, for example, emphasises growing social polarisation in global cities as middle-ranking occupations are squeezed out and as migrant labour performs low-wage insecure services for the elite ﬁrms and professionals. This has been subject to considerable criticism. For example, Fainstein (2001a) argues that observed inequality in high-order cities is almost entirely attributable to rapid real income growth in the top decile(s) of the income distribution, but that should not be taken to imply that other groups have experienced income growth lower than that found in other cities, nor that the ‘middle classes’ have disappeared. Hamnett (1994) makes a similar point in relation to the employed population of the Randstadt. Hamnett additionally argues that Sassen ignores variations in national and local institutions in mitigating the impact of economic structure. In a similar vein, White (1998) provides contradictory evidence for Paris and Tokyo. While these arguments are tangential to this study of ofﬁce markets, they emphasise that the Friedman and Sassen global city models, while providing valuable insights, do not constitute a fully rigorous and accepted model of urban development.

3

She cites Jerusalem as the sole example, which is perhaps ironic given her insistence on diversity.

The Urban Hierarchy and Global Cities

1.4

31

Cities, space and ﬁnance

If there are problems with the global cities/world urban system model, it nonetheless has value in emphasising the importance of the interconnections between the major trading cities of the world: that their place in the urban hierarchy has less to do with the attributes of their immediate geographical hinterland and is more a function of their role and status in global trading. That has important consequences for the economic activity that creates the demand for occupational space in the commercial property market. Speciﬁcally, the drivers of demand are global drivers, conditioned by a city’s ability to capture a share of the ﬁnancial, legal, professional and business service activity, in competition with other global cities. The competitiveness of the city will be linked to its scale: the size, quality and ﬂexibility of its labour markets, the breadth and depth of its capital markets, its ability to generate agglomeration economies. Competitiveness will be enhanced or constrained by institutional factors: barriers to ﬂows of people, goods, capital and information, the regulatory framework, political stability, transparency and governance. And competitiveness will be determined by the physical environment for business: the communications networks, the transport infrastructure, the availability of quality ofﬁce space. Ofﬁce space in this process is not simply a derived demand, with global business services ﬁrms making location decisions and magically calling forth space. It plays a key role in determining those business decisions. For a global city near the top of the urban hierarchy, a large ofﬁce market is a prerequisite. The scale must be sufﬁciently large to accommodate not only the major global ﬁrms but also all the producer, business and professional ﬁrms that serve those ﬁrms. There needs to be a range of space in terms of size, speciﬁcation and quality to meet those needs. Hence scale is not simply an issue of population or workforce, but also of the built form. That built space is more than just a box for the functioning of global ﬁrms. It is also a store of value. Space is created, through development and redevelopment, requiring capital to ﬁnance and fund the process. Once created, that space has value as an investment asset – the capitalisation of the rental income stream produced by the ﬁrms that occupy that space. That ownership, as we will see later, may be distributed widely and is rarely conﬁned to the city, or even the nation within which the city lies. But, unlike ﬁnancial assets, the physical nature of the real asset – its ‘locational ﬁxity’ – means that the value of the real estate adheres to the city. As cities become more successful in global competition, so this store of value grows. That value, though, is linked critically both to the competitiveness of the city and to the state of the global economy. In seeking greater integration into the global economy, cities tie themselves into a global business cycle.

32

Towers of Capital

This can be seen most clearly in relation to international ﬁnancial services. The global cities literature has focused attention on advanced producer services – the mix of business and professional services that support the global operation of TNCs. The usual list covers accountancy, advertising, banking, executive recruitment, legal services, management consulting, marketing and more. The emphasis on producer services perhaps results from the origins of much of the global city literature in critical social science, where value ultimately must be derived from manufacturing. From this frame, the services and functioning of the global city is to accumulate the value created from globally dispersed manufacturing activity. Yet, from a different perspective, the real deﬁning feature of the cities at the very top of the global urban hierarchy is that they are cities of ﬁnance. It is not ‘banking’ that is their distinctive feature: it is investment, wholesale and merchant banking, bond markets, equity markets, foreign exchange, derivatives markets, asset and wealth management – all the mix of ﬁnancial activities that characterise international ﬁnancial centres. In almost all the urban agglomerations identiﬁed as the major global cities – GaWC’s alpha cities, Friedman’s core primary cities – international ﬁnancial services activity plays a key and critical role in the employment structure and in the generation of wealth. International ﬁnancial services activity plays a key symbolic role as well: New York, London, Frankfurt, Hong Kong, Chicago, Singapore, even Tokyo and Paris are known as ﬁnancial centres. The next chapter focuses on ﬁnancial centres and the development of a small group of dominant cities that act as the key nodes in the network of global ﬁnance and capital ﬂows – the international ﬁnancial centres or IFCs.

2 International Financial Centres and Global Cities

2.1

Introduction

One of the key functions of world cities is to act as a centre for ﬁnancial activity. But what makes a city a ﬁnancial centre, rather than a centre for the provision of services, part of which are ﬁnancial? Kindelberger provides an early but comprehensive deﬁnition of the attributes and functions of a ﬁnancial centre: Financial centers are needed not only to balance through time the savings and investments of individual entrepreneurs and to transfer ﬁnancial capital from savers to investors, but also to effect payments and to transfer savings between places. Banking and ﬁnancial centers perform a medium of exchange function and an inter-spatial store-of-value function. Single payments between separate points in a country are made most efﬁciently through a center and both seasonal and long-run surpluses and deﬁcits of ﬁnancial services are best matched in a center. Furthermore, the specialized functions of international payments and foreign lending or borrowing are typically best performed at one central place that is also [in most instances] the specialized center for domestic interregional payments. (Kindelberger, 1974, p. 6 and cited by Reed, 1980) By implication, not only is high-level ﬁnancial activity concentrated in one city within a country: it is increasingly concentrated in a small number of key cities globally. In a review of ﬁnancial centres, the Economist (1998) noted that centres that thrive will increasingly take business from rival centres … today’s mainly national ﬁnancial centres will be replaced by just a handful

34

Towers of Capital

of international centres. Increasingly, the forces that produced ﬁnancial centres are ignoring national boundaries, gathering up ﬁnancial businesses in even greater concentration. What makes a ﬁnancial centre an ‘international’ ﬁnancial centre? Reed (1980) argued that international centres just evolve from and extend national ﬁnancial centres as ﬁnance becomes more global in nature. An international ﬁnancial centre (IFC) will have a greater concentration of cross-border activity than a domestic ﬁnancial centre, but it is not clear where the boundary lies or whether it is the scale or the proportion of international activity that is critical (for example, Tokyo in the modern era and New York historically were dominated by domestic transactions and capital). It is possible to identify IFC indicators and attributes: global ranking in ﬁnancial activities, headquarters (HQs) of major multinational ﬁnancial ﬁrms, the presence of foreign ﬁnancial services ﬁrms in the city. Research, though, generally leave the deﬁnition open and focuses on the activities and characteristics of the leading cities of ﬁnance, judged by their activity levels: a ﬁnancial hierarchy to mirror the world urban hierarchy. A starting point for analysis, then, is a consideration of the extent to which ﬁnancial activity is concentrated in global cities. Given that much of the literature on the ‘new international ﬁnancial system’ focuses on the importance of linkages through telecommunications and information technology and on the geography of a network of ﬂows, it is also necessary to explain the persistence of IFCs. If it is possible to connect with the international network of capital ﬂows from anywhere with an internet connection, why do ﬁnancial ﬁrms, traders and market makers need to be in large cities, facing high labour, congestion and infrastructure costs? Why do they not seek low-cost locations, in the same way that call centres have been distributed widely both within nations and across nations and continents? To understand the continued concentration of high-level ﬁnancial activity in a small number of cities, we need to consider the nature of agglomeration economies as they apply to ﬁnance and the continuing importance of face-to-face contacts and the development of trust relationships.

2.2

Concentration in ﬁnancial centres

A starting point for deﬁning IFCs is an enumeration of the scale of global ﬁnancial activities, the proportion of the global capital market that is captured by the centre. There are many classiﬁcations and rankings of the leading ﬁnancial centres which use different databases and deﬁnitions. The rankings produced are, nonetheless, remarkably consistent for the top-ranked centres. As an example, Z/Yen (2008) have created a Global

International Financial Centres and Global Cities

Table 2.1

35

Global Financial Centres Index.

Financial centre London New York Hong Kong Singapore Zurich Frankfurt Geneva Chicago Tokyo Sydney Boston San Francisco Dublin Paris Toronto

Index score 795 786 695 675 665 642 640 637 628 621 618 614 613 612 610

Source: Adapted from Z/Yen (2007).

Financial Centres Index (GFCI) initially commissioned by the Corporation of London.1 This is a composite index based on published indices and the results of around 825 interviews. The indices and interview results are clustered into ﬁve factors: 1 People: size, quality and ﬂexibility of the labour market, labour market productivity, business education, the development of human capital; 2 Business Environment: regulation, tax rates, economic freedom, transparency and corruption, the ease of doing business, the presence of global business services; 3 Market Access: the volume and value of trading in equities, bonds and other securities, the level of securitisation, clustering of ﬁnancial service ﬁrms, presence of signiﬁcant global HQs; 4 Infrastructure: the cost and availability of space, transport infrastructure; 5 General Competitiveness: more general factors – price levels, economic sentiment, the city as a living environment. These are combined to create an aggregate ranking of cities which, it is claimed, is an indicative rating of their competitiveness as IFCs: the top 15 cities ranked in 2007 are shown in Table 2.1. They also provide rankings within their ﬁve ‘instrumental factors’ (with London and New York ranked ﬁrst or second on all ﬁve factors). 1

There might be some concern that there is some link between the commissioning by the Corporation of London and the ﬁnding that London is the top-ranked ﬁnancial centre; however, this is a consistent ﬁnding across surveys and rankings.

36

Towers of Capital

Of the other cities ranked, the majority feature prominently in the lists of global and world cities discussed in Chapter 1. There are exceptions: ﬁrst, the list includes a number of ‘offshore’ ﬁnancial centres (including Jersey ranked 16th; Luxembourg, 17th; Guernsey, 19th; Isle of Man, 21st). In a sense it is somewhat misleading to include these cities, as in many instances, the actual presence of ﬁrms is minimal: while companies may be registered there to beneﬁt from favourable tax and regulation, the extent to which they are active may be less evident. Second, the list shows the rise of Middle East ﬁnancial centres: Dubai (24); Bahrain (39), Qatar (47). Third, some cities that appear in many world city lists are lowly ranked: Los Angeles and Miami are not in the top 50, while some of the ‘alpha’ cities are comparatively lowly ranked – Tokyo at 9th, Paris 14th. Z/Yen suggest that Tokyo’s low ranking is linked to issues concerning the regulatory environment, ease of doing business and people factors. The cities are classiﬁed into ﬁve overlapping categories. Global ﬁnancial centres offer ‘a critical mass of ﬁnancial services to act as an intermediary, connecting international, national and regional service participants directly’. It is suggested that only London and New York qualify as global centres. International centres conduct a signiﬁcant volume of cross-border transactions involving at least two locations in different jurisdictions. All the top ten ranked centres qualify as international centres. Niche centres are worldwide leaders in particular centres: for example Zurich for private banking, Hamilton for reinsurance. London and New York qualify here as do Zurich and Geneva in the top ten. Z/Yen do not classify Chicago as a niche centre, but its derivatives and commodities trading would suggest it has a qualifying sectoral specialism. National centres conduct a signiﬁcant proportion of their host country’s ﬁnancial activity (hence Chicago and Geneva in the top ten do not qualify), while Regional centres conduct a substantial amount of the business of their surrounding region (Chicago in the Mid-West of the United States for example). While the categorisations may be debatable, this grouping is useful in pointing to a hierarchy of IFCs which relates both to their geographical location and their share of the global ﬁnancial market. It also links back to the GaWC emphasis on connectivity – in this case, the extent to which they are connected with, and embedded in, the global ﬁnancial system. MasterCard Worldwide (2008) publish an Index of World Centers of Commerce which ranks cities on a range of factors which include dimensions for ﬁnancial strength, ease of doing business and status as a business centre. There are strong similarities between their list and that of Z/Yen, although the MasterCard list omits the small offshore ﬁnancial centres. Seven of the top ten ranked cities are common to both the lists and both place London ﬁrst, and clearly above second-placed New York. For the 36 cities that overlap, there is a 0.75 correlation between the Z/Yen index and

International Financial Centres and Global Cities

37

the MasterCard index scores – although the correlation between the GFCI and scores on MasterCard’s ﬁnancial ﬂows variable is lower at 0.55. There are some striking differences, however. MasterCard’s ﬁnancial ﬂows factor ranks Seoul highly and gives top 20 rankings to Moscow, Mumbai, Shanghai, Sao Paolo and Milan: it gives lower signiﬁcance to the two Swiss centres of Zurich and, in particular, Geneva and gives lesser ratings to San Francisco and Boston. No index or ranking will be completely satisfactory and consensus is unlikely. In general, the Z/Yen ranking will be used in this book, with some reservations. Its scoring system is transparent, it is fully focused on ﬁnancial centres, it blends statistical measures with survey-based evidence and appears to correlate better with other ﬁnancial variables. The major reservations concern the extent to which the Asian and emerging economy ﬁnancial centres are appropriately represented in the index and the extent to which there might be an unwitting linguistic and Euro-centric bias. Wójcik (2007) provides a national complement to these city-based indices, estimating an International Financial Index (IFI) that is based on analysis of data on cross-border lending, international debt securities and the crosslisting of equities, currency trading and over-the-counter (OTC) derivatives activity. Wójcik’s index both conﬁrms the leading position of the United Kingdom in international ﬁnance (and the United Kingdom is largely London in this context) and the comparatively weak position of Japan. The United Kingdom is ranked ﬁrst, just above the United States; there is then a considerable gap to Germany, France, Hong Kong and the Netherlands. Japan is ranked eighth, behind the Cayman Islands. Making the simplifying assumption that the IFI relates to the principal ﬁnancial centre in each country (which is problematic in countries with multiple international centres – the United States with New York, Chicago and other cities; Switzerland with Zurich and Geneva), Wójcik’s index scores have strong positive correlations of 0.8 or more with both the Z/Yen and the MasterCard index scores – suggesting that there is considerable consistency in identifying the main global centres of international ﬁnancial activity. Since Z/Yen use both the Wójcik and MasterCard scores as supporting input, most of the analysis in this book will use the GFCI rankings. Bank for International Settlements (BIS) data for 2007 illustrate the concentration of global activity. Again making the simplifying assumption that international ﬁnancial activity takes place in the principal ﬁnancial centre in a country (as BIS data are national in nature), 85% of foreign exchange trades take place in just ten centres (the BIS survey covers 54 countries), with London taking 34% of the market and New York a further 17% (Figure 2.1). Concentration in OTC derivatives trading is slightly less, with London and the United States (which here must include New York and Chicago) having around a third of the market and the top ten centres 63% (Figure 2.2). Estimating derivatives trading is made more complex by the existence

38

Towers of Capital

1500

Daily turnover, $bn

1250 1000 750 500 250

en Tor ont o Mo sco w Bru sse ls Lux em bou rg Sto ckh olm

hag

urt nkf Co

pen

Fra

Pa ris

ey Syd n

por e Ko ng ng Ho

yo

Sin ga

Tok

Zur ich

wY ork

Ne

Lon

don

0

Figure 2.1 Foreign exchange activity by volume, 2007. Source: Adapted from data in BIS (2007). 1500

Daily turnover, $bn

1250 1000 750 500 250

Figure 2.2 (2007).

am

olm ckh Sto

ls

an

ste rd

Am

Mil

o ont

sse Bru

Tor

en pen

hag

ney Co

ng Ko

Syd

ng

nkf

ich

urt

Ho

Fra

Zur

gap

ore

yo Sin

Tok

ris Pa

rk Yo Ne w

Lon

don

0

OTC derivatives trading by volume, 2007. Source: Adapted from data in BIS

of pan-national electronic trading platforms. International debt issuance ﬁgures by country show a growing concentration over the 20 years between 1987 and 2007. In March 1987, the top ten countries took a 34% share and a 53% share of total global issuance. By 2007, market share of the top ﬁve and ten had risen to, respectively, 62% and 78% (Figure 2.3). Total issuance over that period had risen by 17% a year; the rate of increase of the top ﬁve countries was 21% per annum. The ranking is remarkably stable with a Spearman’s rank correlation coefﬁcient of 0.97 between the March rankings

International Financial Centres and Global Cities

39

90 Top five

Share of total issuance, %

80

Top ten

70 60 50 40 30 20 10

pt Se

Se

em 90 be M r1 a 9 Se rch 91 19 pt em 93 be M r 19 a 94 Se rch 19 pt em 96 be M r 19 ar ch 97 Se 1 pt em 999 be M r 20 ar ch 00 Se 2 pt em 002 be M r2 0 ar ch 03 Se 2 pt em 005 be r2 00 6

8

19

ch

98 M

ar

r1 be

pt

M

em

ar

ch

19

87

0

Figure 2.3 Concentration: international debt issuance. Source: Adapted from data in BIS (2007). 16 000 000 14 000 000

US$ million

12 000 000 10 000 000 8 000 000 6 000 000 4 000 000 2 000 000

N

YS To E N kyo AS D Lo AQ nd Eu o H ron n on e g xt Ko ng D TS eu X ts Sp ch an e Sw ish is O s M X Bo AS rs X a Sh Ital an ia Ko gha re i M a In di um n Jo a N bai ha ati nn on a Sa esb l o ur P g Ta aul iw o an

0

Figure 2.4 WFSE.

Market capitalisation, major stock exchanges, end 2006. Source: Data from

for 1987 and 2007, although Japan’s fall down the rankings from the second largest to the fourteenth largest issuer is striking. World Federation of Stock Exchange (WFSE) data for market capitalisation of its member exchanges show a similar picture of concentration (Figure 2.4). The New York Stock Exchange (NYSE) dominates, with a near 30% share

40

Towers of Capital

of the total market capitalisation of the more than 50 exchanges in the Federation, with National Association of Securities Dealers Automated Quotations (NASDAQ), also New York based, in the third place. The top ﬁve exchanges have a 61% market share, the top ten a 75% share. Given mergers in stock markets, it is hard to provide direct historic comparisons. However, there seems to have been some reduction in concentration since 1991 (when WFSE ﬁgures suggest that the top ten exchanges had a market share of 85%, the top ﬁve 78%), although it is hard to discern how much of this is caused by composition changes and how much by the growing capitalisation of markets in India, China and South America. A key role of ﬁnancial centres is to broker and organise ﬁrm consolidation through mergers and acquisitions (M&A). Amel et al. (2004) provide data on M&A in the major industrial countries across the 1990s (Table 2.2). They report a total of 10 170 deals with a total value of US$2375 billion – around 18% of all M&A deals over the period and 25% of the deals by value. The pace and scale of consolidation increased over the decade, with 57% of deals and 80% of the value of deals taking place between 1996 and 2001. In the 1990–2001 period, there were 246 deals worth more than $1 billion. The United States has the majority of deals, reﬂecting not only its higher GDP but also the consolidation of its fragmented ﬁnancial services sector – and particularly its banking system – as regulatory constraints such as those imposed by the Glass–Steagall Act were relaxed. For example, Cetorelli et al. (2007) show the share of bank assets held by the top four US commercial banks rising from less than 10% in 1990 to over 25% in 2004. They

Table 2.2 Financial sector M&A, 1990–2001 by value, main industrial countries.

Australia Belgium Canada France Germany Italy Japan Netherlands Spain Sweden Switzerland UK USA Total

Banks

Insurance

Other

Total

%

15.6 28.9 16.6 58.4 71.0 99.6 163.5 16.8 37.1 19.7 27.5 147.4 911.5 1613.6 67.9%

4.4 3.7 9.7 23.9 18.9 18.3 15.5 25.2 3.4 2.8 10.9 78.0 218.1 432.8 18.2%

9.6 4.7 13.6 18.8 3.8 4.5 4.5 6.4 2.1 2.8 1.6 42.0 213.9 328.3 13.8%

29.6 37.3 39.9 101.1 93.7 122.4 183.5 48.4 42.6 25.3 40.0 267.4 1343.5 2374.7

1.25 1.57 1.68 4.26 3.95 5.15 7.73 2.04 1.79 1.07 1.68 11.26 56.58

G10 countries plus Switzerland, Australia; figures in $billion. Source: Adapted from Amel et al. (2004).

International Financial Centres and Global Cities

Table 2.3

41

Financial sector concentration, US markets, 2004.

Activity

Market share, top five firms (%)

Concentration trend 1990–2004

Underwriting: Initial public offerings Seasoned offerings Investment-grade bonds High-yield bonds M&A advisory Syndicated loans

60.7 49.2 56.4 56.1 56.8 50.2

Increasing Increasing Decreasing Stable Increasing Decreasing

Secondary market dealing: Treasury bills TIPS Mortgage-backed securities Corporate bonds Federal agency securities

37.6 71.9 58.2 73.6 45.8

Increasing Increasing Stable Decreasing Increasing

Source: Adapted from Cetorelli et al. (2007).

show similar general concentration in underwriting, dealing and other selected ﬁnancial services (Table 2.3). Corroborating evidence is found in Berger et al. (1999), who show that the number of US bank charters fell by 30% between 1988 and 1997, from 13 130 to 9216. M&A activity within the ﬁnancial sector creates concentration within ﬁnancial centres that accompanies the concentration across centres. While the majority of the mergers are intra-national, representing an internal consolidation, there are a signiﬁcant number of international M&A over this period. Cross-border M&A in the ﬁnancial sector almost invariably is linked with a consolidation of HQ functions in major IFCs. As an example, UBS (itself a product of the merger of Union Bank and Swiss Bank Corporation) acquired Dillon Read and Paine Webber in the United States and SG Warburg in the United Kingdom, ING acquired Barings (UK), BBL (Belgium) and Aetna (US), along with an aggressive Asian expansion, Credit Suisse absorbed First Boston Bank and Deutsche Bank acquired Bankers Trust in a series of cross-continental deals. While the nominal HQs of these international ﬁnancial conglomerates may be in the city of their historical origins, generally the effective operational HQ will be based in one of the major ﬁnancial centres. Since both partners in the M&A generally have a presence in a wide range of ﬁnancial centres, it also generally triggers a demand for consolidation of operations with implications for the ofﬁce and labour markets in those cities. Amel et al. (2004) review research on the beneﬁts of bank M&A across developed economies. They suggest that banks and ﬁnancial institutions can make efﬁciency gains from economies of scale, as ﬁxed costs are

42

Towers of Capital

spread over a larger base – but the cost curve is relatively ﬂat and u-shaped, suggesting that there is a size limit beyond which gains end. They suggest that much of the scale economies observed come from larger banks being able to adopt more advanced business technologies. They ﬁnd limited evidence of economies of scope (from the ability to cross-sell products in new markets or apply technological solutions across a range of sub-sectors) and very little evidence of gains for shareholders from M&A (although it may be that beneﬁts are only achievable over a longer period given consolidation costs, which may be hard to detect using conventional event study methods). The motivation, then, may well be more about growth in market share and removal of competitors as short-run cost savings and greater returns. Davis (2007) lists the ten largest banks by assets in US dollars (Table 2.4). The list of the top ten banks varies considerably over time, with volatility caused by bank M&A, by exchange rate ﬂuctuations and by changing economic fortunes at institution and at national level – as evidenced by the shift from Japanese domination in 1985 to the far stronger European presence in 2005. With few exceptions, these leading banks are headquartered in the major global ﬁnancial centres and have a signiﬁcant global presence in other world cities. For example, Barclays website in 2007 listed 42 international ofﬁces for Barclays Global Capital (including ofﬁces in London, Paris, Frankfurt, Zurich, Dubai, Hong Kong, Singapore, Tokyo, Sydney, Los Angeles, Chicago and New York) with ofﬁces for Global Investment in North America, Europe, Asia and the Paciﬁc Rim, private wealth management ofﬁces in the major offshore ﬁnancial centres and retail and commercial banking operations active in 26 countries. Of equal signiﬁcance is their scale. Davis estimates that the assets of the ten largest banks in 2005 are equivalent to 37% of world GDP – the equivalent ﬁgure for 1985

Table 2.4 Largest ten banks, by assets, US$billion. 2005

1 2 3 4 5 6 7 8 9 10

1985

Bank

HQ

Assets

Bank

HQ

Assets

Barclays Mitsubishi UFJ UBS HSBC Citigroup BNP Paribas Credit Agricole RBS Bank of America Mizuho

London Tokyo Zurich London New York Paris Paris Edinburgh Charlotte Tokyo

1587 1585 1563 1499 1494 1484 1380 1334 1294 1268

Citicorp Dai-Ichi Kangyo Fuji Sumitomo Miutsubishi BNP Sanwa Credit Agricole Bank America Credit Lyonnais

New York Tokyo Tokyo Tokyo Tokyo Paris Osaka Paris Los Angeles Paris

167 158 142 136 133 123 123 123 115 111

Source: Adapted from Davis (2007).

International Financial Centres and Global Cities

43

being 26%. This represents a huge and growing concentration of ﬁnancial capital. What motivates this process of consolidation and concentration? Industrial economics and the business management literature have suggested that successful ﬁrms should concentrate on core activities and processes, break down vertical linkages, outsource supply, ‘de-layer’, ‘right size’ and become leaner, smaller and more responsive. Why, then, is the ﬁnancial service industry and its associated business services sector creating ever larger ﬁrms straddling activities and geographical boundaries? The underlying factors behind the consolidation of ﬁnancial services on a global scale come from changes in regulation and technological progress. Regulatory shifts consist of the removal of barriers to international ﬂows of capital, removal of limits non-domestic ownership of ﬁnancial institutions and elimination of domestic market constraints. Technological changes relate mainly to advances in digital telecommunications which allow secure transmission of data across long distances allied to the growth of electronic business and trading platforms. These changes permit ﬁrms to merge or acquire competitors in an attempt to improve efﬁciency or to capture market share by extending the spatial reach of their operations or by expanding the range of products on offer. Investment banks, in particular, are driven towards globalisation strategies by the need to deal with a global clientele and 24-hour geographically dispersed trading. One important driving factor is the need to obtain economies of scale by increasing the size of operation. Combining clients, funds under management, skilled staff and capital not only increases market share in particular sectors, but also allows the ﬁrm to create specialist teams to work in niche areas. By concentrating on larger deals, transaction costs, information gathering costs, management costs and monitoring costs can be minimised, increasing proﬁt margins. Such concentration is facilitated by improvements in technology, which permit the relocation of branch and subsidiary ofﬁces. It is further encouraged by concerns about risk management and reporting lines (many of the well-reported ﬁrm failures, including that of Barings, Société Génerale and Sumitomo were, at least in part, due to poor monitoring of business activities in branch locations). Deregulation has also contributed to mergers, consolidation and concentration. With the erosion of national barriers to trading, investing and capital ﬂows, the need for a formal presence in individual countries is reduced. Thus, within the European Union, it is sufﬁcient to be established in one country to operate in all others. Freed of national restrictions, banks and ﬁnancial ﬁrms can move to the most advantageous locations. In practice, concentration and consolidation are not polar opposites of concentration on core business and outsourcing. Typically, the merged ﬁrm will concentrate on high-value, high-margin activities and downsize

44

Towers of Capital

or shed other operations. This creates opportunities for start-up ﬁrms (often spun off the merged parent) to provide specialist, niche services or to serve a displaced client group. Such ﬁrms are likely to locate as near to the parent ﬁrm as costs permit, to beneﬁt both from prestige and from information linkages. This permits larger centres to offer a greater range of specialised services and to be more adept at innovation and adaptation to changing market environment. Outsourcing of non-core services creates a demand for specialised business and professional services: lawyers, accountants, consultants, research and information services, providers of specialised technology, recruitment ﬁrms and courier services. Again, larger centres are favoured. The business and professional service ﬁrms can beneﬁt from scale economies. Furthermore, the size of markets allows such ﬁrms to specialise and hence provide a tailored service for clients. Following Porter (e.g. Porter, 1990, 1998a,b), then, it is the diversity of markets that provides agglomeration economies and drives further clustering: rather than eroding clustering, IT facilitates it.

2.3

Information technology and cities

Centres in a world of telematics In examining trading ﬁgures for equity, bond and derivatives trading, it is perhaps simplistic to equate activity with the city location of the exchange. Electronic trading platforms mean that trading activity can be taking place remotely; the removal of barriers to capital ﬂows from the 1980s onwards means that the beneﬁcial ownership of the equities and bonds and the interest in the derivatives trades may be widely dispersed. As will be explored, there remain signiﬁcant geographical factors in trading: but equally a Paris derivatives trader may be dealing in Australian stock index futures on behalf of a Middle Eastern client with her middle ofﬁce based in London handling the transaction and settlement process. There has been growing concern from regulators about the existence of ‘dark pools’: private automated crossing systems and trading platforms that match buyers and sellers without publishing bid-offer prices. These allow institutional investors to trade large blocks of stock between each other without causing pricing effects and with anonymity. It was estimated that in the ﬁrst half of 2007 as much as 10–20% of all trades of NYSE securities were through such dark pools although presumably such estimates are speculative. This sense of global dislocation is reinforced by the growing trend for mergers, acquisitions and strategic alliances in securities and derivatives exchanges. Examining just some of the activity in the 2000s, in 2000, Euronext was created which brought together the Amsterdam, Brussels

International Financial Centres and Global Cities

45

and Paris Stock Exchanges, with Lisbon joining in 2001; in 2002, Euronext acquired London International Financial Futures Exchange to provide a stronger derivatives presence; in 2007, Euronext merged with the NYSE to create a major Transatlantic force. OMX – itself a merger of Scandinavian exchanges with a growing presence in the Baltic States was subject to rival bids from NASDAQ and Borse Dubai. In a complex deal, Borse Dubai took a 20% share in NASDAQ, NASDAQ took a 33% share in the Dubai International Financial Centre, sold Dubai its stake in the London Stock Exchange and acquired a controlling interest in OMX. As this unfolded, the Qatar Investment Authority acquired signiﬁcant stakes in London and OMX. Both the DeutscheBó´rse and NASDAQ took stakes in the Mumbai Stock Exchange, while NYSE acquired a 20% ownership of the rival National Stock Exchange of India. Derivatives exchange Eurex (joint owned by DeutscheBó´rse and the Swiss Stock Exchanges) acquired the International Stock Exchange, a North American electronic trading platform (which itself had a strategic alliance with Toronto). The London Stock Exchange (itself a target for DeutscheBó´rse, Euronext, NASDAQ and MacQuarie Bank) bought Borse Italiana at the end of 2007. At a smaller scale the Hellenic exchanges bid for the Slovenian Stock Exchange, as the transitional economies became linked to the international ﬁnancial network. In parallel with European consolidation, regional consolidation can be seen in American markets with NYSE acquiring ArcaEx and the Paciﬁc Coast Exchange, NASDAQ acquiring the Boston and Philadelphia Exchanges, with the merger of Chicago rivals CME and CBOT (which then bid for NYMEX) and with the Toronto Stock Exchange’s proposed acquisition of Montreal as the agreement on the protection of provincial ﬁnancial markets in Canada expired. The strategic alliances between the Mexico, Sao Paolo and Lima exchanges provide a southern echo of the North American activity. Cross-continental strategic alliances include those between the Tokyo and New York Stock Exchanges and between Singapore, the Australian Stock Exchange (Sydney) and the American Stock Exchange (New York). If the international ﬁnancial system is a network, then it needs connections: the literature on the new international ﬁnancial system tends to treat the provision of cables as a given. Physical movement of goods and people requires airline and fast-rail connections, but the major ﬂows between cities are analogue and digital ﬂows of data and information. In exactly the same way as the existence of appropriate ofﬁce space is typically treated as a derived demand – that space will be created to serve the needs of ﬁnancial and producer services ﬁrms – cable technology is typically seen as arising from the needs of ﬁrms which somehow results in the creation of linkages. There is less discussion of the actual provision and ownership of those networks which is, itself, a global industry. There is, however, a literature on the spatial and territorial

46

Towers of Capital

implications of telecommunications networks, going back to Goddard’s work in the 1960s and 1970s (e.g. Goddard, 1973, 1975; Goddard & Pye, 1977), summarised in Graham and Marvin (1996) and Graham (1999) and, more recently, in Rutherford (2005). The linkage of markets is both a product of the waves of ﬁnancial deregulation and globalisation and technological developments in telecommunications (in turn, allied to the privatisation of postal and telephone operators across the 1980s and 1990s). The ﬁrst transoceanic ﬁbre-optic cables were laid in the late 1980s. The early to mid-1990s saw an era of consortia which saw the establishment of major transcontinental cables networks. FLAG (Fibre-Optic Link Around the Globe), for example, consists of four segments, an Atlantic loop from the United States to France via the United Kingdom; FEA, which links the United Kingdom, the Mediterranean countries, the Middle East, the Indian sub-continent and onto South East Asia; Falcon which links Egypt, the Middle East and India; and FNAL which provides an Asian loop for Hong Kong, Taiwan, South Korea and Japan. FEA links to the separate SEA networks, which link Western Europe to the Middle East and onto the Paciﬁc Rim. These networks are vulnerable to disruption; SEA-ME-WE-3 was severed in 2006 by ocean ﬂoor earthquakes; while FALCON and SEA-ME-WE-4 cables were damaged in an accident causing major disruption to Asian internet access in early 2008. The risk of cable disruption creates a need for redundancy in networks to ensure continuation of business activity. This, in turn, creates competitive pressures that led to falling proﬁts, ﬁnancial distress and bankruptcy (e.g. WorldCom, Global Crossing) and reorganisation in the world submarine cable market at the turn of the century. As of 2007, the major global network providers were FLAG/Reliance, VSNL and Global Crossing (Terrabit Consulting, 2007). Their ownership is an indication of the changing international nature of global networking. Reliance Group is a Mumbai-based company controlled by Anil Ambani, which acquired FLAG in 2004, before reﬁnancing with an IPO on the London Stock Exchange. VSNL is part of the Tata Group, also Mumbai based and family controlled. VSNL bought Tyco’s global cable network in 2004 and Teleglobe in 2006. Global Crossing, which has a Bermuda HQs but an administrative HQ in New Jersey, was bought out of Chapter 11 proceedings by Singapore-based ST Telemedia in 2003. The big ﬁve cable providers are Alcatel (Alcatel Lucent, France, whose origins lie in ITT Corporation’s European telecom business, sold in the late 1980s), Tyco (Tata Group, Mumbai), NEC (Tokyo), Fujitsu (Tokyo) and KDDI (Japanese, but with private equity fund Carlyle Group having a significant stake). The trend in cable development has been away from the United States–Europe links towards South and East Asia, with growing cabling linking South American countries and the early stages of offshore cable networks for coastal Africa providing still greater global interconnection.

International Financial Centres and Global Cities

47

Rutherford (2005) argues that the provision of telecommunications networks is inﬂuenced by local and national state actions and regulations and that, hence, technology plays an active role in the formation of world city networks. Cities are shaped by telecommunications, but they shape them too, in a two-way process. He provides the example of Colt Telecommunications’ EuroLAN – a major hub-and-spoke ﬁbre network connecting 32 cities in Europe. The network consists of a series of sub-loops connecting individual cities – these were developed sequentially (starting with London and Frankfurt), based on competitiveness and attractiveness as business propositions. Rutherford contrasts Paris and London. London liberalised its telecommunications industry early, and provides relatively little obstacle to connection. The French market is, by contrast, more heavily regulated and, within Paris, provision of building connections is expensive and cumbersome.2 Hence national and local information strategies, institutions and regulations can affect the extent and cost of connecting businesses to the global networks. Although Rutherford does not make the point explicitly, it also seems evident that a dense network of clients allows for lower unit costs of network provision favouring concentration of activity in large cities. For smaller urban settlements, it may be too unproﬁtable to provide ﬁbre-optic provision for isolated ﬁrms.

The end of geography? Despite such issues about the proﬁtability of connection businesses to the ‘global super highway’, a number of researchers have argued that developments in information and communications technology have transformed the nature of business and removed the raison d’être for agglomeration and a city location. This idea forms the basis of much of the professional literature on changing workplace relations which, in turn, informs much of the discussion in the academic literature about what an appropriate real estate strategy would look like in ‘the information age’. O’Brien (1992) proclaimed ‘the end of geography’, arguing that information technology and globalisation had created a state of affairs within which geographical location no longer matters in ﬁnance. Similarly, Peet (1992) argued that, with electronic trading, there is no persuasive case for a central market place. These ideas resurfaced with the rise of internet trading ﬁrms and ‘day trading’, particularly in US equities markets. If information can be obtained readily anywhere in the world, if trades can be made anywhere in the world, why pay for expensive central business district (CBD) ofﬁce accommodation? 2

Paris places constraints and delays on direct cable laying, but has an extensive sewage network which permits connection: however, the rental of space in that network is very expensive.

48

Towers of Capital

Evidence of decentralisation of ﬁnancial activity away from traditional downtown areas – notably from Wall Street – seemed to give credence to such a view. Webber (1963, 1964) was advancing similar ideas about cities some 30 years before O’Brien, long before the arrival of the personal computer and the ubiquity of the internet. Yet, the most radical, non-spatial, trajectories advanced in Webber’s work have not occurred: cities still appear to retain their importance. The observed impact of technology on ﬁnancial activity, in fact, seems rather different. While the technology means that work can be anywhere, it must be somewhere. The location, then, will depend upon other factors associated with efﬁciency and proﬁtability. For high-level ﬁnancial services, the essential inputs are human capital – skilled labour – and information. This implies not decentralisation but concentration. Successful activities, as Porter, Krugman and others have noticed, increasingly cluster together. Economies of scale, agglomeration economies, information economies, the presence of customers, clients and competitors pull activity to a small number of key locations. In a review of ﬁnancial centres, the Economist (1998) noted that centres that thrive will increasingly take business from rival centres … today’s mainly national ﬁnancial centres will be replaced by just a handful of international centres. Increasingly, the forces that produced ﬁnancial centres are ignoring national boundaries, gathering up ﬁnancial businesses in even greater concentration. In a similar vein, Deutsche Bank (Deutsche Grundbesitz, 1998, cited in Lizieri et al., 2000b) suggested, in a European context, that modern communications technology enables banks who are under increasing pressure from costs to concentrate their international trading activities possibly to a single ﬁnancial centre in the European timezone. It is important here to distinguish between retail and wholesale activity and between high-volume, low-margin and low-volume, high-margin activities. Retail activity may be less likely to concentrate since it relies on customer knowledge, tastes, preferences and local marketing. It is also cost-sensitive and, hence, may be displaced from major centres. The majority of call centres are, of course, for retail ﬁnancial services. High-volume, commoditised, wholesale activities include settlement, clearing and certain forms of trading. The tasks may be relatively standardised, are less reliant on information exchange and innovation, and generate lower proﬁts. As a result, there may be pressure for these activities to decentralise, to seek lower cost locations.

International Financial Centres and Global Cities

49

By contrast, high-value-added, low-volume business – for example, corporate ﬁnance, fund management, raising capital, M&A – relies both on information (from customers, rivals, parallel business and suppliers) and on close client contact. This leads to greater concentration, a concentration further fuelled by the need to access skilled labour. For such activities, as Gehrig (1998) has observed, information from IT sources and from face-toface contacts are complements, not substitutes. Lombard Street Research point to the relative increase in earnings in the City of London as evidence for this ‘upgrading’ of activity. Just as manufacturing, printing, publishing and non-ﬁnancial ofﬁce activities were pushed away from the City in earlier phases, retail ﬁnance and commoditised, high-volume activities are being pushed away now. This change brings, in turn, a change in the nature of ofﬁce requirements and shifts in the required space per worker that counterbalances the ofﬁce intensiﬁcation brought by new working practices. The arguments regarding information technologies, ﬁnance and globalisation are interwoven with debates concerning the ability of the nation state to determine (or even inﬂuence) local economic and social activity. Martin (1994) again cites O’Brien: ‘ﬁnancial market regulators no longer hold full sway over their regulatory frameworks within the nation state’ and cites Castells’s (1989) concept of the space of ﬂows ‘that now dominates and transcends the historically constructed space of places’ (Castells, 1989, p. 254). For Martin, in a post-industrial era, capital and money markets have been separated from industry, money itself commoditised and securitised, ﬁnancial centres become globalised and integrated and national monetary authority eroded. He suggests that the global ﬁnancial system has moved beyond one where international transactions take place to one where ﬁnancial activity is globally integrated and coordinated within ﬁrms that are organised to be, essentially, stateless. If this is the case, then it opens up the possibility that one would expect to see dispersion of the location of ﬁnancial activity, a shift away from spatial concentration: again, this does not seem to be conﬁrmed from empirical observation. Sassen (2002) attempts to explain the apparent contradiction of why, if telematics allow the integration of ﬁrms’ dispersed manufacturing sites, service locations, HQs and research and development facilities, those ﬁrms need to be in cities, why the command and control functions have to be concentrated. She makes the argument, discussed in Chapter 1, that cities represent ‘corporate service complexes’: they exist because specialist ﬁrms sell ﬁnancial, legal, accounting and advertising services to other ﬁrms (in Sassen’s work largely seen as ﬁnal producer ﬁrms). However, such an analysis is weak on mechanisms for agglomeration in a digital era – she cites social connectivity, externalities and intangible agglomeration economies and the importance of M&A, and the location demands of ‘denationalised elites’ of professional workers. However, this, per se, does not provide

50

Towers of Capital

a strong or decisive argument for agglomeration and concentration. In a digital age, producer services could be acquired remotely from multiple sources and locations. Work itself could be decentralised and dispersed to tap into global labour markets. Concentration and agglomeration arguments need to stress the importance of physical (not digital) proximity and the key role played by face-to-face connections between actors in the global ﬁnancial system.

Telematics versus face to face How important are face-to-face communications in a digital world? Firms can communicate via telephone, fax, e-mail and increasingly, more advanced digital formats that would seem to reduce the importance of direct contacts and geographical proximity. Ganesan et al. (2005) note that shared relational ties and norms of trust and reciprocity foster knowledge transfer, which, in turn, provides economic advantage for communication; that face-to-face communication is the optimal way of acquiring knowledge; and that the most valuable knowledge is tacit and not codiﬁed. They suggest that the available literature offers little empirical evidence to support these three propositions. However, tacit, non-codiﬁed knowledge is more nuanced and hence its communication will be less amenable to remote forms of communication. Applying this to business activity, they suggest that knowledge is applied in different ways. For product development, creativity and innovation are vital – and the type of knowledge required here is more likely to be tacit. For implementation in competitive environments, speed is vital and this requires knowledge of processes that is more likely to be codiﬁed. Zaheer and Manrakhan (2001) also discuss remote electronic access and its impact on the location of business activity. Their starting point is Dunning’s (2000) extension of the decision-making process of ﬁrms locating internationally to include the importance of the search for ideas, intellectual capital and strategic advantage. This widens the resources sought by ﬁrms to include managerial skills, knowledge and human capital. Does this, however, imply spatial concentration when telecommunications mean that many of these new factors of production can be accessed ‘remotely’? This will depend on the extent to which the resource can be digitised, the extent to which remote contracts can be deﬁned and enforced and the extent to which knowledge spillovers are localised. They suggest that downstream, customer-focused market-seeking activities ‘are likely to be remotely accessed from centralized locations which have the greatest comparative advantage for their performance, resulting in reduced global dispersion of such activities’ (Zaheer & Manrakhan, 2001, p. 675). However, they also suggest that since spillovers are localised, spatial and temporal proximity will determine where knowledge spreads ﬁrst

International Financial Centres and Global Cities

51

(and there are clearly distance effects, discussed further below). In a neat analogy, they look at the spread of a computer virus, spreading from the Philippines, through Asia, into Europe and the United States, with its effect diminishing through that spread – an effect of proximity (trafﬁc) and temporal effects (time zones). Given the importance of speed of reaction – the capture and, critically, the use of information – major international ﬁrms are likely to develop global portfolios of sites in the lead markets, where ﬁrm-owned branches act as antennae for local knowledge. These locations will vary depending on sector. Financial ﬁrms may be in London, New York, Tokyo and Frankfurt; fashion ﬁrms in London, New York, Paris and Milan. Agnes (2000) provides conﬁrmation of this need for a local presence in his qualitative analysis of the Australian swaps market. His respondents stress the importance of local formal and informal networks that allow traders both to develop relationships (critical where commercially sensitive proprietary information must be exchanged) and to acquire time-sensitive information on the market. One respondent comments ‘Consensus rules in ﬁnancial markets: economic fundamentals matter but consensus rules’: in a segment of the market where speed of decision is vital, it is difﬁcult to gauge that market sentiment remotely and outsiders may not understand all market nuances. Nonetheless, local ofﬁces often play an intermediary role. For example, an Australian client wanting a US dollar swap may call a Sydney broker: but the trade and the hedging will be executed in New York and middle and back ofﬁce settlement and risk control will be centralised, usually in or near a major IFC. The spatial implications of this are that activities within a sector that rely on innovation and new product development are more likely to be clustered and to beneﬁt from agglomeration, while production, processing and routine work beneﬁt less from interaction and hence may seek low cost settings. This is consistent with some of the urban systems arguments discussed earlier. For the ﬁnancial services sector, these ﬁndings suggest a spatial separation of activities with those that require frequent interaction, innovation, knowledge spillovers clustering in large urban agglomerations that are linked into the international ﬁnancial networks. The extent to which more routine, batch production activities can be decentralised will depend on how critical are the downstream links from, for example, broking activities to middle and back ofﬁces. The need to maintain control of these linkages (and the price of breakdowns in control – as evident in the 2008 Société Générale derivative scandal) creates different locational dynamics than those multinational production ﬁrms described, for example, by Buckley and Ghauri (2004) whose hub and spoke structures might include production, assembly and distribution warehouses that are not only geographically remote from the core HQ and Research and Development functions but may also be outsourced or

52

Towers of Capital

at arms length from the parent ﬁrm. In global ﬁnancial services, the need for control may keep data centres and middle ofﬁce functions close to the centre of the major cities (if not necessarily actually in the core CBD area). Similarly disaster recovery centres need to be in close proximity to main ofﬁces, both for key staff access and because of problems of signal attenuation, even over ﬁbre-optic cable networks, since they must mirror trading activity in real time (see Chapter 4 for further discussion). Finally, Charlot and Duranton (2006) provide empirical evidence from France that telecommunications act as a complement rather than as a substitute for face-to-face meetings. Based on extensive survey analysis of worker behaviour, they ﬁnd that workers in cities communicate more than workers in rural and suburban cities – and that workers in larger cities communicate more than those in smaller urban settings. These relationships hold even when corrected for other factors explaining the extent and intensity of communication – size of ﬁrm, industry (with ﬁnancial ﬁrms being positively linked to higher levels of external communication) and education level of the worker. For example, 24% of rural workers give instructions to customers or suppliers, compared to 31% of suburban workers and 50% of workers in the largest urban category. These results hold for all communications media (face to face, telephone, e-mail) except paper communications. The intensity of communication also increases by city size. Far from superseding direct communications, e-mail appears to be a complement for telephone and paper communications and a weak substitute for face-to-face meetings only for workers without higher education. Furthermore, there is no evidence that the gap between urban, suburban and rural areas had narrowed between 1987 and 1997 – while usage of IT had increased, the lead of the largest cities over rural areas was as wide. On the basis of their ﬁndings, they reject the IT-driven ‘demise of cities’ hypothesis. It seems then, that agglomeration economies, knowledge spillovers and the beneﬁts of a city location are still potent – at least for advanced service industries – even with developments in communication technology and electronic trading.

2.4

Agglomeration and concentration

The development of IFCs is strongly inﬂuenced by the agglomeration economies discussed above. This implies considerable path dependency. Nonetheless, IFCs can decline in the face of competition – both nationally and internationally. Relatively recent national examples include Toronto’s growing dominance over Montreal, Sao Paulo surpassing Rio de Janeiro and Johannesburg outstripping Cape Town (Laurenceson & Tang, 2005, citing Porteous, 1995).

International Financial Centres and Global Cities

53

Gordon and McCann (2000) review the literature on clustering, agglomeration and economic growth policies. They note that there is considerable ambiguity in deﬁnitions used in policy, with terms like agglomeration, clustering, new industrial areas and embeddedness used loosely and interchangeably. They suggest that there are distinct theoretical traditions that seek to explain differences in local economic growth in terms of the type of business activity and the relationships between ﬁrms, which have different policy outcomes and require different types of analysis to test for their existence. Their point about imprecise use of terms and concepts ﬁnds an echo in Markusen (1999), who raises similar questions about much of the ‘new’ regional social science, complaining of fuzzy concepts and inadequate standards for veriﬁcation and reliability. She poses a key question for all such research: ‘how do we know if the ﬁndings are true or valid’? To an extent, this is no more than a distinction between ‘good’ and ‘bad’ research. The ﬁrst model that Gordon and McCann identify can be traced back to Marshall’s (1883) ideas of localisation economies and the externalities that ﬁrms in a sector gain from co-location: access to a specialised labour pool, local provision of non-traded inputs and maximum ﬂow of information ideas – the pure agglomeration economies discussed in Chapter 1 on the functions of cities. Labour market effects come from efﬁcient search and matching processes (which enable ﬁrms to identify the best workers for a particular process and for those workers to obtain optimal jobs) and returns to scale from division of labour. Large skilled labour pools aid the development of human capital and skill acquisition. There are scale efﬁciencies in the provision of non-traded goods, too, as service ﬁrms can become more specialised and as competition eliminates inefﬁciency. The localisation economies that apply within a sector and the increasing returns to scale may be complemented across sectors by urbanisation economies which may be linked to information effects (although the mechanism for this information exchange seems underspeciﬁed). Gordon and McCann point out that these beneﬁts do not assume, nor do they arise from, cooperation between ﬁrms, but are largely scale effects. A second strand develops the classic neo-classical industrial location theory of Weber, Isard and Moses to note the spatial behaviour of ﬁrms. Clustering occurs from the decisions of ﬁrms based on production costs and/or costs of marketing and sale. This typically will result in ﬁrms making similar decisions – which may be reinforced by cooperation between ﬁrms up and down the supply chain. With ﬁrms having incurred ﬁxed costs in establishing sites and linkages and with no ﬁrm having complete information, the resultant industrial complexes both generate higher proﬁts relative to a dispersed distribution of ﬁrms and tend to be relatively stable over time. However, there is no great incentive for ﬁrms

54

Towers of Capital

to attract new ﬁrms in. An example of an industrial complex development would be motor manufacturing with just-in-time component delivery systems. The third model emerges as more of a critique of neo-classical economic approaches, and focuses on the relationships between ﬁrms, on contracts between ﬁrms and the fact that contracts are a substitute for trust relationships (Granovetter, 1985). The basic concept is that strong interpersonal relationships in social relationships can transcend ﬁrm boundaries and foster order in inter-ﬁrm relationships. If that exists then ﬁrms will be more willing to undertake joint ventures and to cooperate, without fearing opportunism from competitors. This idea of relationships being ‘socially embedded’ through norms, institutions and common assumptions is central to the ideas of post-Fordist new industrial districts, but it should be noted that there is nothing inherently spatial about the idea (a medieval guild would play a similar role without spatial boundaries) but proximity is likely to result in more frequent interactions. Major questions here include whether the beneﬁts of learning and cooperation in local social networks provide sufﬁcient incentives for ﬁrms to develop trust-based relationships in a world of global competitiveness and whether the economic beneﬁts of such relationships can produce sustainable growth and development. As with pure agglomeration economies, large urban areas produce more opportunities for interaction, a higher probability of successful partnerships being created and better links with national and international networks. These three models point to different processes determining local economic advantage. Gordon and McCann then apply the models to the London region. They identify a number of sectors where concentrations are above national averages. They then use survey data to explore ﬁrms’ attitudes to the advantages of their location. The responses emphasise accessibility factors – general accessibility, access to customers, skilled labour pools – with much less stress given to interaction with or proximity to suppliers, competitors and business services. This implies a general urban agglomeration economy explanation of location. However, those activities clustered in central London – law and, in particular, ﬁnance – do favour proximity to related activities. Financial service ﬁrms seem more inﬂuenced by product competition, knowledge spillover and access to a skilled and developing pool of human capital. For centrally located ﬁnancial ﬁrms, but not for ﬁrms in general, inter-ﬁrm linkages and personal contacts and access to top local intelligence and innovation are important. From this, it seems that high-level ﬁnancial services have most to gain from clustering in large urban agglomerations, favouring the development of IFCs which, as they attain critical mass, will be able to capture greater shares of global ﬁnancial activity.

International Financial Centres and Global Cities

55

Japanese Asia Latin America Africa Eastern Europe Developed countries

North American

European

0

10

20

30

40

50 60 Per cent

70

80

90

100

Figure 2.5 Consolidated external bank claims outside the developed economies. Source: From data in van Rijckeghem and Weder (2003).

Geography and distance Van Rijckeghem and Weder (2003) analyse consolidated interbank nondomestic claims in developing and transitional economies as a measure of bank outward investment. The ﬁgures – for the end of 1998 – show a strong regional focus (Figure 2.5). US bank claims are signiﬁcantly tilted towards Latin America (with 56% of claims); 69% of Japanese bank claims are in Asian markets. European claims are more widely spread, but dominate Eastern European investment. It is also noticeable that European banks have far greater external claims than North American and Japanese banks, with a majority of claims in each region and 63% of the total external claims. In ﬁnance, the ‘home bias’ effect suggests that investors’ portfolios are tilted heavily towards assets (principally equities) that are geographically close to the investor’s base – to a far greater extent than is justiﬁed by risk– return characteristics (Coval & Moskowitz, 1999). This effect – which applies within national boundaries and, much more, across national boundaries – is typically explained with reference to information ﬂows. Investors ﬁnd it harder to obtain information on equities and on companies that are distant from them and to monitor the ﬁrms whose equities they hold. They are thus at a disadvantage compared to local traders. These information asymmetries might result from insider information effects, from culture effects (differences in national business practices and market institutions, in language, in reporting and accounting), and they might have a behavioural component. With electronic trading and telecommunications, though, is distance still a barrier? Do the size of market and the complexity of the market offset these distance effects (such that major ﬁnancial centres can overcome distance and capture a larger proportion of global trade)? Portes and Rey (2005) test the effects of distance on equity ﬂows between countries and suggest that a form of gravity model operates – as in trade

56

Towers of Capital

markets – with the size of the equity markets positively affecting purchase and sale of equities between markets while the distance between those markets negatively affecting ﬂows. Given that trading occurs electronically (the period analysed is 1989–1996), the distance effect is assumed to be linked to information effects. The distance effects are still present when variables measuring alternative information ﬂows are included – telephone trafﬁc and presence of bank branches both have positive effects on equity ﬂows, as does goods trade between the two markets, but, even with these conditioning variables, the distance effect stays signiﬁcant and strong. There is no clear evidence of a regional bloc effect, and distance effects are present in Europe despite the EU. They also note that there are signiﬁcantly higher trades to and from London, New York and Tokyo, even after correction for market size, which implies that those three ﬁnancial centres have information advantages over rival smaller IFCs. Market sophistication is positively linked with equity ﬂows (which again provides a trading advantage for the global ﬁnancial centres). There may also be a portfolio diversiﬁcation effect – however, this is swamped by distance effects (since markets in close proximity tend to have higher correlations and hence lower diversiﬁcation beneﬁts). Choi et al. (2003) also use a gravity model approach to examine bank penetration of ﬁnancial centres. In general, both GNP (positively) and distance (negatively) affect the presence of banks from one ﬁnancial centre in another. Buch (2005) provides further conﬁrmation, modelling bank assets and liabilities for French, German, Italian, UK and US banks for 50 host countries for the period 1983–1999. Her model, conditioned on GDP size, correlation between national GDPs, exchange rate volatility and dummies for the openness of the ﬁnancial system in the host country, ﬁnds consistent signiﬁcantly negative coefﬁcients on distance. She notes that there is little evidence of the coefﬁcients weakening between 1983 and 1999 despite the reduction in communication and information costs from technological progress. Trading levels were strongly signiﬁcantly higher where the host/target country had an IFC although her deﬁnition of an IFC seems a little eccentric. Hau (2001) provides a formal examination of the combined effect of distance and information on equity trading, examining 11 German stocks traded on the Xetra trading platform by traders spread across eight countries. As in the international home bias literature, it is assumed that there are cultural and linguistic barriers that may coincide with national borders. Hau ﬁnds that traders located in ﬁnancial centres outside Germany perform signiﬁcantly worse than traders within Germany (with the differences sufﬁciently large to have real effects), but traders in German-speaking markets (Switzerland, Austria) do comparatively better than those with larger linguistic barriers while still underperforming (that is, there is a separate distance effect). There is tentative evidence that, within Germany, traders who are geographically close to the HQs of the ﬁrms whose shares

International Financial Centres and Global Cities

57

are being traded outperform those from further away but this effect seems conﬁned to short-term trading.

Financial cities and competitiveness As in other sectors, ﬁnancial service centres seek to attract ﬁrms to locate HQs and branch ofﬁces in their cities and to capture a growing market share of international business in those cities. This competition for market share takes the form both of advocacy – extolling the virtues of a particular location for business – and active promotion through planning initiatives and legal/regulatory measures. For ﬁrms, these latter measures must be weighed alongside the immediate economic factors in making locational decisions. For the city and region, ﬁrms provide employment (both directly and through multiplier effects) and, possibly, help to retain capital. Market share, however, does not necessarily bring either employment or wealth. For example, for all the much publicised gains made by the Frankfurt Exchanges over the London International Financial Futures Exchange in trading the new bund futures contracts, it turns out that some 30% of trades were from terminals based in London. Surveys provide an indication of the types of variables that are important in determining the success of a city in attracting ﬁnancial ﬁrms and capturing market. These include capital market size, the labour market and the mix of skills, the regulatory regime, taxation, land and labour costs, transport and accessibility, business culture and the quality of environment for senior staff. However, while some of these are historically contingent, others can be created or changed over time. The list of factors does not explain why some cities are attractive to business and others are less so. Krugman has suggested that the economic success of one region rather than another may, ignoring natural resource endowments, be simply a matter of chance. Once established, ﬁrst-mover advantages enable that city to capture economies of scale, drive down costs, export services and dominate later cities. To an extent, this can be seen in the City of London. The City has always been externally oriented, as Ingham (1984) has demonstrated. Nonetheless, the decline of the British Empire might have been expected to lead to a decline in London’s importance. However, the establishment of the Eurobond and Euromoney markets – themselves a response to the particular nature of US ﬁnancial regulation – enabled a new growth impetus and provided a depth to the capital markets that did not rest on the UK domestic economy. The City is, thus, an exception. While technology has eroded the importance of location, it still plays a role. Thus, time zone is an important factor, despite the growth of 24-hour working. If, as is often suggested, three global centres will dominate the international ﬁnancial market (as is, to an extent, already the case), then

58

Towers of Capital

there will be a major centre in the European time zone. Geo-political factors may also be important in developing new markets – within Europe, opening up the Eastern economies might be an example. As important as geography is accessibility. The presence of one or more major airports is a key element in the attractiveness of a city, while high-speed rail links are, at least in a European context, becoming an important factor. Intra-city accessibility, the existence of a transit system capable of drawing labour into a centre from a wide catchment area is also important. Van Winden et al. (2007) consider the structural characteristics that determine a city’s competitiveness in ‘the knowledge economy’, its ability to acquire, create, disseminate and use codiﬁed and tacit knowledge. There is a danger of anthropomorphism here. Cities do none of those things: it is ﬁrms and individuals within cities that create and use knowledge. Their list of criteria, drawn from a broad literature, however, is useful in pointing towards a typology of competitive advantage. First, they stress the knowledge base and infrastructure: higher education institutions, public and private research and development, the education level of the workforce. Second, they point to industrial structure – the extent to which there is inertia from the presence of traditional industry, the proportion of innovation-based business (this last argument is somewhat circular). To attract and retain talented workers, they point to urban amenities and quality-of-life factors. Accessibility is seen as vital: ‘the knowledge economy is a networked economy’, so international hub airports, fast rail connections and transportation are seen as vital. Urban diversity is cited, drawing on Jacobs (1960) and Glaeser et al. (1995) as is scale. Sheer size is an attractor for companies and knowledge workers and permits specialisation and specialist amenities. Finally, they suggest that social equity is conducive to growth (although no evidence is provided for this). They point to policy implications, to ways in which urban managers might be able to enhance the competitiveness of their cities. However, it is also evident that existing city endowments create a legacy effect, one that is perhaps most marked for leading metropolitan agglomerations. Diversity also features as an important factor in Porter’s discussions of regional growth. Here, clustering of businesses – suppliers, clients, competitors in a range of industrial sectors – create information linkages that foster innovation and reduce production costs. Labour mobility enhances the information linkages. These information agglomeration economies encourage ﬁrms to locate in such centres and permit a higher proportion of ﬁrms to survive and proﬁt in a competitive environment. Such information economies are vital in the ﬁnancial services industry – and particular in high-level international ﬁnancial business activity. The role of government in Porter’s Diamond Model (Figure 2.6) is ‘acting as a catalyst and challenger; it is to encourage – or even push – companies to raise their aspirations and move to higher levels of competitive performance …’.

International Financial Centres and Global Cities

Government

59

Firm strategy, structure and rivalry

Factor conditions

Demand conditions

Related and supporting industries

Figure 2.6

Porter’s Diamond Model of competitive advantage.

Critical mass, then, is not simply a function of the size of the market, it is also a function of diversity – the presence of many ﬁrms in associated business sectors, the existence of specialist service providers serving those sectors and the presence of a large and varied pool of labour. In turn, skilled labour will be attracted to a centre because of its size and range of activities. This will be particularly important for skilled international ﬁnancial service employees. The presence of many ﬁrms both provides the possibility of advancement and also risk of diversiﬁcation in that, in the event of loss of job, there will be many more locally available opportunities. The quality of life in and around the city (cultural amenities, housing, education, environment) will also be important in determining individual decisions – but, I suspect, not as important as some commentators have suggested. It is largely labour market factors and the structure of rewards that determine the decisions of globally mobile professionals. However, do ‘cities’ really compete, and can actions by cities make a fundamental difference to their urban ranking, particularly in relation to ﬁnancial services? Gordon (1999) stresses that a deﬁning factor in ﬁnancial competitiveness is the regulatory environment which is largely national in nature: for example, regulation and governance, taxation, transparency and property rights, rules and regulations regarding labour and migration all enter into the locational decisions of global ﬁnancial services ﬁrms. Similarly, the local rules and norms (as described by Agnes, 2000) that operate and affect competitive advantage are as likely to be national as urban in nature. The efforts of cities alone – whether the local government or public–private growth coalitions – may affect the physical environment for business: Tokyo’s proposed waterfront ﬁnancial district, Frankfurt’s reshaping of its CBD, the rebuilding of the Bund in Shanghai and intra-urban infrastructure, but these may be marginal in the context of urban agglomeration economies. City governments could, however, damage competitiveness, for example, through a strongly anti-development stance that prevented the

60

Towers of Capital

creation of a modern ofﬁce infrastructure as a base for advanced ﬁnancial services. Gordon goes on to suggest that smaller competitive cities may be able to capture niche service markets (particularly in a globally integrated economy where their market reach can extend beyond national boundaries) but that only the larger international cities can support the wide range of specialist business and ﬁnancial services that foster innovation and economies of scope. The role that city governments can play in encouraging ﬁnancial ﬁrms to locate and in contributing to the growth of resident ﬁrms, then, may be limited in comparison to the importance of the economic and business environment. One role is clearly advocacy – raising the proﬁle of the city and ensuring that ﬁrms considering a move are aware of the potential beneﬁts; and arguing for regulatory, resource allocation or taxation changes that beneﬁt the city. Subject to competition laws, it may be possible to provide direct incentives through capital allowances and tax breaks. The most important contribution, however, may well be through land-use planning and the provision of basic infrastructure. Given the importance of accessibility, the establishment and maintenance of an efﬁcient transportation system may be the most critical element. As we will argue later, city governments can also play a major role in determining – or hindering – the development of an appropriate mass and quality of ofﬁce space in which the ﬁnancial ﬁrms may operate.

2.5

IFCs, concentration and ofﬁce markets

The continuing dominant role played by a small number of cities in the global ﬁnancial system – IFCs – can be attributed to a historically contingent and path-dependent process linked to geo-political and economic factors. The ability of a city to compete for a large share of international ﬁnancial market activity can be related to scale – the depth and breadth of its capital and labour markets – and to diversity in the range of specialist ﬁnancial services and associated advanced producer services present in a city. The key role played by knowledge spillovers, by the development of trust relationships and by innovation limits the impact of developments in information technology and telematics in reducing the importance of geography. Models of the location of business activity, both generally and as applied to ﬁnancial services have, typically, assigned a secondary, passive, role to property markets. Economic growth (or decline) in a region is seen to result from locational advantage, the resource endowment of the city or region and from human capital – the skills and ﬂexibility of the labour force. Firms chose to locate, grow and prosper in locations that offer the best combination of factor inputs and costs. Cities and regions compete for ﬁrms and

International Financial Centres and Global Cities

61

ﬁrms compete in a global marketplace. In traditional models, the key factor in ﬁrms’ decision-making (and hence the growth of an area) was transportation cost – the cost of transporting factor inputs to the production site and ﬁnished products to market. As service industries have risen in importance and as information has become a key input and output of economic activity, research has tended to emphasise the importance of agglomeration economies and information spillovers. In both cases, agglomeration results in higher land values and rents in city centres which might be seen to reduce the attraction of a location – or to drive out lower value-added activities in IFCs creating still greater concentration and specialisation. In such analyses, the signiﬁcance of land and property markets tends to be largely peripheral or ignored. Thus, in the standard models of urban economics, land values and property rents result simply from location and transport infrastructure: the issue of supply of space is not considered. Property is seen as a derived demand, with supply responding (albeit not seamlessly) to economic change. Similarly, in the new regional economic literature, networks of ﬁrms arise, compete and exchange information with no consideration of where they do business. When property markets are considered, it tends to be in a critical fashion, with land owners and developers seen as a constraint to economic adjustment. This view of urban change, blind to property markets, is limiting. As D’Arcy and Keogh (1997, 1999) note, real estate market dynamics shape changes in economic and spatial structure. The technological and economic forces that determine the growth and decline of ﬁrms and industrial sectors are mediated through the existing built environment and the property market structures that are in place in particular cities and regions. The same economic forces may, thus, generate quite different outcomes in cities where the built form and real estate market institutions differ markedly, while an inﬂexible real estate market can hamper a city’s competitive ﬁnancial market strategy. Urban economic activity generates a requirement for land and buildings that must, initially, be met from the existing stock, given the necessary time lag between a demand signal and supply of new space. The characteristics of the existing stock of space, then, will help determine how successfully changing patterns of demand can be accommodated. Stock characteristics include the aggregate quantity of space, the quality of that space, the mix of space suitable for a diversity of needs and the ﬂexibility of the existing buildings. Access to that space depends on the structure of the market, which allocates space to different, competing, users and helps determine the price. That same structure will also determine how responsive the market is in the face of new demand: whether supply is encouraged or constrained. Thus, the property market will inﬂuence the ability of an urban economy to accommodate trends and pressures. This, in turn, will feed back into the economic development process. Those cities with more advanced, adaptive

62

Towers of Capital

systems will gain competitive advantage, altering the path of economic development. Examining, ﬁrst, the existing stock of space, it is clear that the total amount of aggregate space must play an important role. The larger the stock, the more ﬁrms and workers that can be accommodated and, hence, the more scale economies and the beneﬁts of agglomeration are likely to arise. The size of the existing stock will, naturally, arise from the historical development of the urban economy. Mere size, however, will be insufﬁcient to maintain competitiveness. Quality of stock is another critical dimension. For large, high-order service ﬁrms, ofﬁce space must be capable of supporting the business: accommodating the technology required for efﬁcient operation, providing a satisfactory working environment for key staff and prestige in the eyes of clients and competitors. Equally important, the space needs to be ﬂexible to permit reconﬁguration and the implementation of new working practices. Thus, an ageing stock, subject to depreciation and functional obsolescence, is ill-suited to a competitive international ﬁnancial services centre. However, it needs to be stressed that this does not imply that a competitive international service centre will only have large, modern, high-speciﬁcation ofﬁces. As Chapter 1 suggested, one key feature in global ﬁnance cities is diversity. This diversity needs to be mirrored in the ofﬁce stock. Major banks and ﬁnancial services ﬁrms may require modern, large ﬂoorplate ofﬁce space; other major professional service ﬁrms may require modern space but with smaller plates to accommodate cellular ofﬁces. Niche product ﬁrms and suppliers need smaller amounts of space either in multi-occupied buildings or in smaller ofﬁces. Around these ﬁrms are a web of smaller suppliers, start-up businesses, outsourced services, information providers and branch ofﬁces of ﬁrms based elsewhere. These require a whole range of property solutions, from serviced ofﬁces and small bespoke dedicated space to space in secondary, class B or C ofﬁces. This is signiﬁcant in terms of competitive city strategies aimed at ‘creating’ IFCs. Constructing large prime ofﬁce buildings to attract major ﬁrms to locate in a city is not, in itself, a sufﬁcient solution if the built environment does not offer the range of space for all the necessary support services and suppliers that those ﬁrms require. The institutional structure of a property market determines, in large measure, its ability to respond to changing occupational requirements. That structure depends on public and private sector actors and on the legal, regulatory and contractual framework within which the market operates. The market structure shapes the processes that channel investment into real estate, create new space and allocate space to ﬁrms. The sophistication of the market determines how freely information ﬂows amongst participants, shaping attitudes to investment and occupation and inﬂuencing key locational decisions. The planning system imposes an obvious constraint on

International Financial Centres and Global Cities

63

the operation of the market. This may well have beneﬁcial impacts – reconciling conﬂicting demands, dampening cyclical volatility and preventing gluts, coordinating land-use and transport infrastructure, preserving environmental quality and ensuring space for public good uses that might otherwise be outbid in a pure market system. Planning, by constraining excess development and preserving some measure of scarcity, may maintain property values and encourage continued investment. However, restrictive planning regimes may hamper the market’s adjustment to changing occupational requirements. For example, a strict conservation policy may prevent replacement of functionally obsolete space; a rigid zoning policy may hamper the provision of space at the fringes of existing ofﬁce areas. Furthermore, the planning system can impose delays in the development process which lead to lags in responding to demand signals and, possibly, to greater market cyclicality. Table 2.5 sums up the link between property market characteristics and urban competitiveness. The competitiveness of IFCs, then, is undoubtedly inﬂuenced by the characteristics of their ofﬁce markets. While a ﬁrm’s decision to locate in Table 2.5 Property markets and IFC competitiveness. Factor

Dimension

Examples

Existing stock

• Quantity • Quality • Diversity

Aggregate stock of space Age; functionality; flexibility Size; configuration; quality; costs

Private market structure

• Ownership pattern

Traditional/modern types; international character; integration to capital markets Lease lengths and terms; tenure; access and barriers; flexibility of formats Property rights; property services; integration with other markets Information flows; asymmetry of information; transactions data; research services Risk/return profile; income security; market liquidity; transparency Diversity in funding modes; diversity in provision; capital access and constraints; integration with capital markets

• Occupational pattern • Market maturity • Market transparency • Investment quality • Finance and funding

Public policy context

• Planning regime • Workplace rules

• Legal structure • Tax regime

Source: Adapted from Lizieri et al. (2000b).

Responsiveness; rigidity; speed of processing; costs; land-use coordination Health and safety; floorspace standards; natural light; facilities; environmental standards; enforcement Property rights; lease terms; enforcement of obligations; security of income and tenure Entry and transaction costs; income and capital taxes; local property taxes; VAT

64

Towers of Capital

a particular city may well be driven primarily by quality of labour markets, presence of customers and competitors and access to capital markets, the cost and quality of real estate will also be a factor. More importantly, the characteristics of the ofﬁce market must affect business efﬁciency – both for incoming ﬁrms and those already based in the city. In a global market with intense competition, these impacts may be critical. Part II of the book considers the structure of ofﬁce markets and the adjustment processes within them, examining how supply and demand interact to generate rental and capital values. Before turning to the ofﬁce market, though, it is worth pausing to consider the extent to which the system of linked IFCs in global cities is truly a ‘new’ phenomenon.

3 A Sense of History: Development and Inertia in the International Financial Hierarchy

3.1

Introduction

In much of the literature on the ‘new international ﬁnancial system’, the development of an interlinked global system of ﬁnance is dated to a series of linked ﬁnancial and geopolitical events. Some start with the development of the Eurodollar markets from the 1960s, others to the breakdown of the Bretton Woods system of ﬁxed exchange rates, to the need to recycle petrodollars from the mid-1970s, or to the wave of ﬁnancial deregulation that took place in the 1980s. Whichever events are included, the implicit conclusion is that the ﬁnancial system of the late twentieth and early twenty-ﬁrst centuries is qualitatively different from the system that preceded. The Eurodollar and Eurobond markets developed to circumvent US ﬁnancial regulation and, in particular, Regulation Q, which restricted the interest payable on US deposit accounts. Dollars began to accumulate outside America and non-US banks began to offer dollar-denominated savings products. OECD reports suggest that the Eurodollar market increased from $9 billion in 1963 to $132 billion in 1973, a compound growth rate of 31% per annum. The Eurodollar markets were quickly followed by Eurobond issues, dollar-denominated ﬁxed interest securities, and Eurocredits, medium-dated dollar bank loans. Together, these created a substantial dollar market that was international in nature and outside US regulation. The City of London played a major role in creating and facilitating the development of the Euromarkets. The Bretton Woods Agreement had attempted to stabilise post-war currencies by re-establishing a system of ﬁxed exchange rates based on convertibility of the US dollar and establishing the International Monetary

66

Towers of Capital

Fund to provide support for currencies that were moving away from dollar parity. With the dollar’s gold convertibility suspended in 1971, the system effectively ceased to function, with currencies ﬂoating from 1973. This not only marked a change in international trade and ﬁnancial relationships, it also created a demand for sophisticated exchange rate and currency risk management services, which were heavily concentrated in the major ﬁnancial centres. The October 1973 decision by OPEC to double oil prices had major impacts on inﬂation but also created a surge in the revenue of the oil-producing nations, particularly, but not exclusively, in the Middle East. Much of this dollar revenue was invested, often in short-term deposits, with the largest European and US banks. This excess of capital needed to be used, to be ‘recycled’. This led to product innovation and fuelled the boom in lending to third-world countries from the 1980s and the growth of third-world debt. This created potential ﬁnancial instability, even though much of the lending was to sovereign governments – as became evident with the Mexican government’s freezing of interest payments in 1982, the ﬁrst of many such events. To this extent, the recycling of petrodollars bound the global system together tightly. This interlinking of ﬁnancial markets was accompanied by a growing trend toward market liberalisation and deregulation from the late 1970s. The United Kingdom abolished exchange control in 1979, while the European Union liberalised capital ﬂows in mainland Europe progressively, culminating in the Single European Market of 1992. From the beginning of the 1980s, Switzerland removed many of the measures that penalised nonresident deposits. The ‘big bang’ in the City of London in 1986 attracted much attention, but was preceded by US market reforms and the Paris Borse’s liberalisation measures of 1984 and 1985. US banking reform was slower; Regulation Q disappeared in 1982, but it was not until 1992 that banks were able to set up branch networks across the whole of the United States and not until the 1999 Financial Modernisation Act that the Glass–Steagall restrictions on multi-service banks were lifted. To these liberalisation measures in the Western economies must be added the fall of the Berlin Wall, the consequent opening up of the Eastern European and Chinese markets and the rapid development of the Brazilian and Indian economies. This transformation of the ﬁnancial landscape was accompanied by a transformation in ﬁnancial products. In part this was facilitated by technology: computer technology that allowed the application of the quantitative methods for pricing and portfolio management that had previously largely been conﬁned to academic journals and from telecommunications developments and electronic trading platforms that permitted new ways of operating and seeking arbitrage proﬁts. This drove the development of new ﬁnancial derivative products and markets, led to the creation of hedge funds

A Sense of History

67

and heavily leveraged private equity funds and saw the growth of assetbacked securitisation and disintermediation as corporations found ways of directly accessing the capital markets. This last is particularly signiﬁcant in real estate markets, as will be explored below. Taken together, these changes appear to represent a major step change in the international ﬁnancial system, as liberalisation and technology combined to create an interlinked global system of capital ﬂows that linked the major developed economies together and penetrated the transitional, emerging and less-developed economies – hence the New International Financial System. In that New International Financial System, a small elite group of cities – the international ﬁnancial centres (IFCs) – play a key role in coordinating ﬂows, providing market liquidity and a location for product innovation. But before accepting this model of a step change, however, it is valuable to step back and gain some historical perspective. An examination of the development of ﬁnance and its concentration in a small number of cities across the eighteenth and nineteenth century, up to the start of the 1914–1918 war reveals the growth and development of an interconnected ﬁnancial system that is recognisably international in nature and has many of the features of the ‘new international ﬁnancial system’ of the post-1970 era.

3.2

A sense of history1

One striking feature of the history of IFCs in the industrial era is continuity. From the end of the sixteenth century to the present, many of the existing world cities have continued to play a major ﬁnancial role throughout and few cities that have acquired signiﬁcant status as ﬁnancial status have relinquished that role. Political events can shape the fortunes of cities, competition can change the rankings, the economic progress of the host nation may vary, but continuity is the norm, not the exception – particularly in Europe. Further back in history, cities playing key roles in trade and ﬁnance have become insigniﬁcant global players – Bruges, Genoa for example. Amsterdam was probably the dominant global ﬁnancial centre at the start of the eighteenth century. It had developed as a specialist ﬁnancial centre in the seventeenth century Dutch hegemony. With the support of the state it emerged as a key centre for ﬁnancing of Dutch trade, the growing Dutch industrial sector and foreign government debt. Rodriguez and Feagin (1986) suggest that the growth of Amsterdam stimulated specialist development 1

This section draws on, inter alia, Cassis (2006) who provides a comprehensive assessment of the development of European and US ﬁnancial centres from the early seventeenth century to the present day.

68

Towers of Capital

of surrounding towns: they cite Leiden as a textile production centre. But it was trade, and the tradition of the Dutch East India Company and overseas activity that created specialist ﬁnancial services: merchants acting as brokers, taking commission rather than proﬁt, the acceptance of bills of exchange, the trading of government securities. In 1820, of 136 stocks listed on the Amsterdam Stock Exchange, 108 were non-Dutch. Foreign banks were active in Amsterdam, too, with bankers from Frankfurt, Berlin, Brussels, Vienna and London all present. Gradually, though, London began to gain dominance in trading activity and the economic strengths gained from industrialisation and from the British Empire, allied to the weaker political status of the Netherlands after the 1795 French occupation eroded Amsterdam’s status. The City of London had emerged as the leading global ﬁnancial city by the start of the nineteenth century, based as much on its trading status as on the industrial revolution (although the City played an important role in moving capital from wealthy agricultural regions to the new manufacturing areas). Much of the City’s institutions and mechanisms were in place before the industrial revolution, with the Bank of England founded in 1694. Trade – with the United States, with the Asian and African colonies, as an intermediary between the colonies and Europe – accumulated wealth in London and created specialist commodity and insurance markets alongside a rapidly developing banking sector that developed expertise in handling bills of exchange and increasingly captured market share in placing foreign government loans. There were already formal and informal international links emerging. Baring Bros rose to prominence in part because of its links to Hope & Co in Amsterdam, foreign bankers arriving in London – the Schröders from Hamburg, the Rothschilds from Frankfurt – created a network of interlinked banks and enabled the simultaneous listing of government loans in London, Frankfurt, Amsterdam and Vienna. There were international markets for government bonds from the 1820s and foreign listings were permitted on the London Stock Exchange from 1823. Paris was a rival to London as the pre-eminent ﬁnancial centre of the ﬁrst half of the nineteenth century, as it recovered from the disruption of the French revolution. While funding of French government debt was a key role, ﬁnancing foreign trade and foreign government debt were key activities. As with London, many foreign banks established a presence in Paris and banks with family links to other emerging ﬁnancial centres existed. France played a critical monetary role with the franc emerging as a major international currency – currency trading accumulating reserves of international capital in Paris. In what is now Germany, there was no one dominant ﬁnancial centre. Berlin had a well-developed banking centre and was active in handling Russian loans. Hamburg beneﬁted from being the largest continental European port

A Sense of History

69

and had strong links with London. Frankfurt had more of a banking and specialist ﬁnance tradition, dating back to 1585, when Frankfurt merchants created a formal market to ﬁx exchange rates, although the true origins lie in the medieval fairs that provided a commercial and trading market from the twelfth century. For much of the early development of the market, the focus was on bills of exchange. From the end of the seventeenth century, there was trading in promissory notes and bonds, but the ﬁrst shares were not listed until 1820 and the ﬁrst foreign share not listed until 1958. Frankfurt was the Rothschild’s home town, but their establishment of banks across Europe helped establish coordination of activity (with the City of London acquiring the lead role). Frankfurt placed loans for Austria, Bavaria, Denmark and Prussia in the early 1800s. Other signiﬁcant European centres included Brussels (capturing Antwerp’s business from the 1820s), Vienna, Amsterdam and Geneva, which played a signiﬁcant role through its links with, and inﬂuence on, French ﬁnancial development, with Swiss bankers playing a prominent role in Paris and helping to place French loans, a role that continued after it had been returned to independence after 1814. Geneva and Zurich subsequently beneﬁted from Swiss neutrality, acting as a safe haven for capital and, from this, developing a specialist wealth management role that continues into the twenty-ﬁrst century. From this, it seems that by 1840 there was already a hierarchy of ﬁnancial centres tied together by complex linkages that were based on politics, on trade and on family ties. The latter should not be overemphasised: family inﬂuence on the banks in each of the centres did not usually survive for more than one or two generations – and there were many linkages that were not familial and which crossed religious boundaries. However, the establishment of family banks certainly facilitated the development of cross-national ﬁnancial services. For example, the Schröders, from their Hamburg origins, established banks in Amsterdam, Bremen, Jakarta, Lima, London, New York, Rio de Janeiro, St Petersburg and Singapore. Alliances were forged, banks in one city acted as correspondent banks for banks in other cities, circumventing listing and membership restrictions. The networks that existed did so in the context of ﬁerce competition between nations, between cities and between businesses to capture market share. The international nature of banking and ﬁnancial activity was to become even more pronounced in the second half of the nineteenth century – a period when international ﬁnance is dominated by the City of London. From the 1840s onwards, the global ﬁnancial system was shaped by the need to ﬁnance the development of the railways and the industries that developed from them. This was both a national and an international activity as railways linked countries and developed the colonies and newly independent countries outside Europe. While much of the early development of railways

70

Towers of Capital

was state funded, increasingly railway companies, mining companies and manufacturing industry were private in nature, growing rapidly in size and demanding large amounts of capital. The growth of businesses consuming capital was matched by the growth of businesses delivering capital and the era saw the arrival of much larger banks through organic growth, mergers and acquisitions. The precise form varied by city but the growth of banks coincided with a growing concentration of ﬁnancing activity with London playing the key role. London’s position relative to Paris strengthened with French defeat to Prussia and the uniﬁcation of Germany in 1871 (which also marked the rise of Berlin and the decline of Frankfurt). That this was an international age can be seen by the emergence of the ‘overseas banks’ in London – which were registered in the City and who used that location for capital raising and placing but whose activities were almost entirely outside Britain. For example, the Hong Kong and Shanghai Bank was founded in 1865. By 1860, Cassis (2006) suggests that there were 15 overseas banks with 132 non-British branches operating out of London; by 1890, 33 banks with 739 foreign branches. German universal banks appear in the 1850s, including Disconto-Gesellschaft and Deutsche Bank, both based in Berlin (Deutsche Bank would eventually absorb Disconto in 1929). According to its company history, at its 1870 founding, Deutsche Bank’s mission was to ‘transact banking business of all kinds, in particular to promote and facilitate trade relations between Germany, other European countries and overseas markets’. Their activities funded railway projects in North and South America, in Turkey and in Iraq, in addition to their European activities. By 1873 it had opened a London branch to facilitate joint ﬁnancing of trade. Inevitably, the activities of ﬁnancial centres in this period were linked to their nations’ empires and spheres of inﬂuence. Thus, Paris funded the development of the French railway system, railways in Mediterranean Europe, Russia and the Austro-Hungarian Empire (aided by its Swiss connections) while London’s focus was on the British Empire and North and South America. However, there was also an active secondary market for railway stock. Trade continued to be a key driver of the development of ﬁnancial markets; the City beneﬁted from British trade dominance but Paris gained from foreign exchange activity. The franc’s role as a key international currency was enhanced by the Latin monetary union (with Belgium, Switzerland and Italy ﬁxing their currency to the French franc and the Banque de France’s metal reserves. There were even serious discussions in the late 1860s about a universal currency system based on the franc, which reached the stage of multilateral meetings before British and Prussian resistance – and the ultimate defeat of France in 1871 – led to the abandonment of the idea. The episode, though, points to the existence of international cooperation and dialogue between national reserve banks in the nineteenth

A Sense of History

71

century. German uniﬁcation led to the decline of Frankfurt (which did not even have the status of regional capital, standing below Wiesbaden) and the strengthening of Berlin. Across the last quarter of the nineteenth century and up to the start of the 1914–1918 war, the growth, development and concentration of the global ﬁnancial system continued with London strengthening its position of world dominance and New York playing an increasingly important role. Both France and Britain retained substantial colonial empires and had informal control of much of the non-imperial world, but ﬁnancial activity was neither bounded nor constrained by the boundaries of empire. Trade and the development of ﬁnancial products and activity were facilitated by improvements in transport (both railways and steam ship) and telecommunications (with the development of the telegraph, the telephone and wireless communication). This permitted simultaneous issuance of stock in multiple centres and the development of search for arbitrage opportunities, providing stronger links between markets and promoting more efﬁcient pricing. This was also an era of protectionism, but this does not seem to have greatly restricted the growth of foreign trade, driven by the strong economic growth of the Americas. In this process, London became increasingly dominant. Cassis (2006) cites UN data on foreign assets held by countries in 1913. Britain’s $18.3 billion represented 42% of total foreign assets, well ahead of France ($8.7 billion, 20%), Germany ($5.6 billion, 13%) and the United States ($3.5 billion, 8%). The source of those foreign assets reﬂected empire and sphere of inﬂuence – 55% of French and 53% of German foreign assets were in Europe, while 34% of Britain’s were in North America and 36% were in the Asian, African and Australasian colonies. These trade ﬂows were matched by the growth of international ﬁrms often headquartered in the IFCs, but with a network of foreign ofﬁces and branches. Although US national income had probably surpassed Britain’s by 1870 and German industry and manufacturing probably had productivity advantages, the City of London’s status as the prime global ﬁnancial market was increasingly based more on international activity than on the economic strength of its host nation. The scale of activity created increasing specialisation, with diversiﬁed banks, major insurance companies, the London Stock Exchange, specialist exchanges and markets for precious metals, commodities, and shipping, the development of investment trusts and the growth of specialist niche ﬁnancial and professional service providers, accountants, lawyers, brokers – the era saw the founding of many of the major law and accounting ﬁrms of the next century: Price Waterhouse in 1849, Cooper Brothers in 1854, Deloitte in 1845, who quickly set up international branches, Price Waterhouse operating in New York from 1890 and Chicago in 1892. The insurance companies were also international in orientation (not only for

72

Towers of Capital

maritime insurance but also for life and ﬁre insurance where there were substantial US revenues) and built up large investment portfolios – with as much as 40% of their investments being outside Britain. Fundamental to the development of the market were the merchant banks who would continue to play a key role in the late twentieth century: Kleinwort, Schröder, Rothschild, JS Morgan (whose son would found JP Morgan in New York). The City hosted many foreign banks (Crédit Lyonnais from 1870, Deutsche Bank from 1873 and 30 branches from 12 countries by 1913) and had sufﬁcient depth and resilience to survive ﬁnancial crises that have resonances with today: notably the 1890 Barings ﬁnancial crisis, where problems with its Argentinean investments – particularly its land mortgage bonds – caused an illiquidity crisis that threatened the ﬁnancial stability of the City and led to a Bank of England-led rescue strategy. Illustrating the international nature of the ﬁnancial system, the Barings Crisis had contagion effects around the world, affecting loan rates across South America and wider interest rates in London and Paris (Mitchener & Weidenmier, 2007). Bordo and Eichengreen (1999) argued that ‘ﬁnancial distress in London and heightened awareness of the risks of foreign lending worsened the capital market access of other “emerging markets” like Australia and New Zealand’, the effects lasting for nearly a decade and having real economic effects. Paris had a similar international orientation, with networks connecting the city to France’s colonies, French overseas banks (for example, the Banque de l’Indo Chine) with over 500 branches outside France and a substantial role in placing foreign government loans, with Paribas a signiﬁcant global player in that market. The Paris Bourse was second only to London by market capitalisation and over half the stocks listed were foreign. Berlin, by contrast, grew as a result of the economic strength of post-uniﬁcation Germany, but had much more of a national orientation. Nonetheless, the period saw the emergence of large universal banks which began to build international branch networks and presence in other cities (Table 3.1). Table 3.1 Rank 1 2 3 4 5 6 7 8 9 10

Largest commercial banks by assets, 1913. Bank

Assets, £million

Crédit Lyonnais (Paris) Deutsche Bank (Berlin) Midland Bank (London) Lloyds Bank (London) Westminster Bank (London) Société Générale (Paris) CNEP (Paris) National Provincial Bank (London) Dresdner Bank (Berlin) Société Générale de Belgique (Brussels)

113 112 109 107 104 95 75 74 72 72

Source: Adapted from Cassis (2006).

A Sense of History

73

Berlin played a signiﬁcant role in foreign exchange activity, too, as gold convertibility made the mark the third global currency after the pound and the French franc. The massive economic growth of the United States created major ﬂows of capital into and across the United States. New York acted as the entry point for capital to North America, importing capital from Europe and exporting it west. It was also the leading North American port, with the development of the railways and the canals to the Great Lakes further enhancing its trading status. The opening up of the country through the railroad and the consequent growth of the manufacturing industry and agri-business created many large businesses whose capital demands, as in London, drove consolidation and growth in the ﬁnancial sector with major banks and ﬁnancial institutions opening. Amongst these JP Morgan had a signiﬁcant role – Morgan, the son of London banker JS Morgan had moved to New York, became a partner in Drexel then acquired and renamed it. JP and JS Morgan retained linkages, as they did with Paris-based Morgan Harjes et Cie. Many of the New York banks developed links with Europe – either through family ties or for business expedience – as with Barings and Kidder Peabody, which had exclusive agency rights from 1886. The growth of the United States and the dominance of New York led to an accumulation of capital, reinforced by its status as the central reserve location for the Federal US system. Large savings and loans banks (including Bankers Trust and Manufacturers Trust) grew up, as did a growing life insurance sector. New York still acted as a conduit for capital to ﬂow to the growing US industries, but it began to develop an independent ﬁnancing role of its own, primarily domestic-focused (New York facilitated the wave of industrial mergers and acquisitions at the turn of the century) but with an international component. The British government placed loans in New York during the Boer War, Japanese government stock was listed in 1905, the Russian government raised capital there and New York funded the development of the Mexican national railway (Markham, 2002). There were fewer foreign banks operating in New York (partly due to restrictions imposed by the State of New York) and their activities were largely conﬁned to foreign trade and foreign exchange. For all the domestic focus and reliance on the City of London for liquidity, stock market and banking crises in New York had systemic effects. For example, the 1907 New York banking crisis which led to the formation of the Federal Reserve System had its origins in Bank of England increases in discount rates to counter the outﬂow of gold in 1906. The consequent stock market falls and runs on banks in New York affecting England (where interest rates rose), France (where the impact of US gold and money demand caused a banking crisis), Italy (where interest rises spiking a speculative stock market and lending bubble) and Japan (where stock market prices fell sharply on US news) and promoting

74

Towers of Capital

curtailment of lending to peripheral countries (Bordo & Eichengreen, 1999; Bordo & Murshid, 2000; Bruner & Carr, 2007). The period up to the start of World War I, then, is characterised by the development of an international ﬁnancial system with ﬂows of capital across nations coordinated from a relatively small number of cities specialising in ﬁnancial activity, facilitated by alliances between banks and by the development of international branch ofﬁces. Better telecommunications created the opportunity for simultaneous ﬂoating of issues in multiple centres – for example, the coordinated placement of Argentinean loans in London, Paris and Berlin by, respectively, Barings, Paribas and Disconto-Gesellschaft. Telecommunications also permitted cross-national secondary trading of stocks. As the major cities accumulated capital, so the ﬁnancial services sectors became increasingly specialised. Some cities could offer particular niche products – Swiss private banking, for example, but only the largest centres: London, Paris and, increasingly, New York could offer the market depth and sophistication to handle the largest deals. The linkage of markets also brought contagion effects and simultaneous ﬁnancial panics. That concentration of capital and ﬁnancial activity and the growing specialisation of the ﬁnancial sector meant an increase in bank, insurance, broking and associated service employment in the major IFCs. That increase in ﬁrms and in employment in ﬁnance brought with it a requirement for working space and it is no coincidence that formal ofﬁce space as we would currently recognise it emerges in the middle of the nineteenth century, as the combination of residence and place of work became increasingly infeasible. And, as ﬁnancial ﬁrms became larger and wealthier, so their property requirements expanded, both in terms of space required and in the speciﬁcation of that space.

3.3

‘The age of catastrophe’

Hobsbawm’s (1995) description of the 1914–1947 period, encompassing the 1914–1918 and 1939–1945 world wars and the great depression of the 1930s, is of an age scarred by economic and political turmoil. It also marks a hiatus in the development of a global ﬁnancial system and a shift in the signiﬁcance of international ﬁnancial service centres: although not as marked or radical a shift as some commentators have suggested. There are signiﬁcant geo-political shifts: notably the decline of European imperial holdings, the rise of communism which blocked the ﬂow of capital into Eastern Europe and, ultimately, into China and the division of Germany. The United States, with New York as its ﬁnancial capital, became the clear leader of the Western economic bloc. Yet, throughout the period, the major cities already identiﬁed continued to play the key roles in global ﬁnance deﬁned in the pre-1914 era.

A Sense of History

75

The onset of the 1914–1918 war provoked bank and stock market crises in most of the European economies with the gold standard abandoned, stock trading suspended, debt repayment moratoria declared and state intervention increased. Where trading continued, notably in Berlin, inﬂationary pressures were released. The major beneﬁciary of this was the United States and speciﬁcally New York, with the dollar increasingly seen as the benchmark currency for foreign trade and with a huge shift in the pattern of capital ﬂows, the United States switching from being a debtor to a creditor nation. US banks – notably JP Morgan – also developed new international roles, placing British and French government debt and coordinating lending syndicates. The ﬁnancial centres of neutral countries also gained relatively – notably Amsterdam and Zurich. New York’s position further strengthened after 1918, as the European centres sought to control inﬂation, rebuild ﬁnancial systems and stabilise currency. The 1920s US economic growth brought with it an accumulation of capital and the growing presence of foreign banks setting up branches, subsidiaries or entering alliance and correspondence arrangements. It also saw rapid growth of the stock markets, mainly through domestic stocks and domestic as well as foreign investment. London retained an important international role, and not just in its dominions. The 1920s saw the emergence of very large banks through a process of mergers – and these banks had an outward as well as a domestic orientation. Barclays acquired a number of overseas banks to build a foreign branch network of over 500 branches and, in all, there were in excess of 2250 UK bank foreign branches at the end of the decade. Paris, once the currency had stabilised, sought to compete directly with London in ﬁnancial markets, with mixed success. Berlin, though, lost status with the consequences of hyperinﬂation, the need to service war reparations and the closing of many of its Eastern European markets. The Wall Street Crash of 1929 and the onset of the great depression mark another break with the 1914 international ﬁnancial system. There was a collapse in world trade, a dramatic rise in unemployment in the industrialised countries, a fall of up to a third in industrial output and widespread bank closures with the United States and Germany particularly badly hit. The contagion impacts of Wall Street and the coincident economic downturns illustrate the extent to which the Western economies and their ﬁnancial markets and centres were interlocked by the early years of the twentieth century. Reactions to the great depression, though, marked a retreat from internationalism. The gold standard was effectively abandoned in Europe in the early 1930s, growing government intervention brought constraints on currency and capital movements, protectionism, nationalisation and other barriers to the operation of global market. The response to the equity and banking crises in the United States brought about the Glass–Steagall Act

76

Towers of Capital

and Regulation Q, for example, with similar restrictions imposed in Europe, most sinisterly in Germany. The 1939–1945 war produced less ﬁnancial disruption than 1914–1918, partly because of the greater domestic focus and closer regulation in place at the onset. The post-war period, marked by rapid economic growth in mainland Europe, was also a period of strong state intervention – for example, France nationalised the Bank of France and a number of leading banks and insurance companies. Currency barriers and ownership restrictions constrained capital ﬂows and ﬁnancial trade. The Bretton Woods exchange rate system established the dollar as the benchmark currency and New York was now clearly the leading IFC – IMF ﬁgures suggest that its share of international issues was four times that of London and ﬁve times that of Zurich between 1955 and 1960. Zurich had beneﬁted from Swiss wartime neutrality, with a growing niche role in private wealth and asset management. The major change in IFCs, however, was the decline of Berlin. It was by no means obvious that Frankfurt should become the new West German ﬁnancial capital. Hamburg was more international, had stronger international trade link and already was home to a strong banking sector – both German (with banks transferring operations from Berlin) and foreign. However, political pressures (particularly from the United States) and the location of the national bank, the antecedent of the Bundesbank, in Frankfurt, tilted the balance. In effect, then, the 1950s and 1960s largely saw a restoration of the 1914 hierarchy of IFCs, with, however, New York now the largest and leading centre and with Berlin replaced by Frankfurt as the German centre. The economic liberalisation of the late 1960s and the move to the ‘New International Financial System’ of the 1980s thus is more a restoration of the 1914 system of ﬁnancial centres than the creation of a new network.

3.4

Asian and Middle Eastern markets

There is a major gap in this history. Reed (1980) notes that ‘most diachronic studies of ﬁnancial centers have focussed almost exclusively on London, the European Continent and the Mediterranean’ with the few studies examining Asian markets looking almost exclusively at Hong Kong and Singapore. He provides an early quantitative analysis of Asian ﬁnancial centres in 17 cities and 10 countries. The analysis is based on a series of banking and ﬁnancial variables that indicate participation in the international ﬁnancial system, including foreign assets and liabilities, headquarters of international banks, presence of branches of foreign banks and representation in foreign capitals. He examines these variables from 1910 to 1975. The results indicate that Tokyo becomes the clear Asian leader in 1960 (with Osaka also highly ranked): the changes in ranking over the period reﬂecting the geo-political

A Sense of History

77

Table 3.2 Ranking of Asian financial centres, 1910–1975. 1910 1. 2. 3. 4.

Hong Kong Shanghai Yokohama Tientsin

1925 1. 2. 3. 4. 5. 6.

Yokohama Hong Kong Shanghai Tientsin Tokyo Osaka

1940 1. 2. 3. 4. 5. 6.

Jakarta Tokyo Hong Kong Yokohama Shanghai Singapore

1960 1. Tokyo 2. Hong Kong 3. Osaka

1975 1. 2. 3. 4.

Tokyo Osaka Hong Kong Singapore

Source: Adapted from Reed (1980).

turmoil in the region more than economic fortunes (Table 3.2). Tokyo’s growing international role was linked to economic growth and the export of capital accumulated in the city from Japan, unlike, for example, the growth of New York and other non-European ﬁnancial centres that emerged in the nineteenth and early twentieth century. Reed suggests that Japan’s total trade in 1975 amounted to $113 billion compared to Singapore’s $14 billion and Hong Kong’s $13 billion. Tokyo’s rise in signiﬁcance as an IFC needs to be placed in the context of Japanese isolationism. Up to 1859 and the development of the Treaty Ports, Japan was effectively closed to foreign ﬁnance, investment and trade and even after the opening of the ports and the Meiji government in 1868, there remained strong barriers to foreign trade. In 1859, the currency and banking system of Japan was weakly developed – in particular, rice remained a real currency into the 1870s, paper money was unbacked and issued by local domains and coins had variable metal content. The beginnings of a more solid currency system emerged from 1868, with the introduction of the yen, the establishment of the Bank of Japan in 1882 and the move to the gold standard in 1897. For much of the early opening up of Japan, foreign trade in the Treaty Ports was largely conducted in Mexican silver dollars (which, according to Tamaki (1995) was the de facto trading currency of East Asia). Tamaki (1995) traces the origins of the Japanese banking system to merchant traders – the most signiﬁcant of which were based in Osaka and Edo (to become Tokyo). These ryogae developed practices based on risk assessment, interest rate management, deposit reserves and discounts that allowed for the rapid development of a more recognisable banking system. Larger banks were created both by organic growth and by merger and also by state and samurai involvement. Many of the large Japanese investment banks that would play a signiﬁcant role in world ﬁnance in the 1980s and 1990s were founded in the years following the Meiji restoration and the early decades of the twentieth century: Daiichi (1873), Mitsui (1876), Mitsubishi (1895), Sumitomo (1895), the Industrial Bank of Japan (1902) and Nomura (1927). As many of these banks were initially based in Osaka as were in Tokyo. The presence of government and the Bank of Japan in Tokyo gave that city

78

Towers of Capital

an advantage which led to differential growth and its increasing dominance over Osaka. These ﬁnancial structures survived the disruptions of war and subsequent occupation and were able to capitalise on the more economically liberal environment that followed the ending of the allied occupation in 1952 and beneﬁt from the expansion of external trading activity. China was a signiﬁcant recipient of international capital in the late nineteenth and early twentieth century both to fund government debt and to ﬁnance the development of railway and infrastructure projects. In this, Chinese banks, particularly in Shanghai, acted as intermediaries between foreign banks and local enterprises. The European overseas banks were actively involved: HSBC organised a public offering of Chinese bonds in 1877, other loans were organised on a partnership or syndicated basis, for example by Barings and Deutsche-Asiatische Bank. By the early twentieth century, Chinese bonds and loans could be purchased denominated in gold, silver, sterling, francs, dollars, yen and local currency. Goetzmann et al. (2007) argue that these debts – particularly for railway building – were accompanied by considerable foreign intervention, often with the management of the railway in non-Chinese hands, and with substantial security, in the form of tax revenues, backing the loans. This hampered the development of domestic ﬁnancial institutions and early attempts to organise domestically ﬁnanced infrastructure development were less successful. This political and economic intervention prompted unrest and contributed to the 1911 revolution. Nonetheless, Shanghai was, until the turmoil that led to the establishment of the People’s Republic of China (PRC), the leading ﬁnancial centre in East Asia. Its position reﬂected its status as the leading trading city in China (surpassing Guangzhou in the 1850s), its industrial development, its resource-rich hinterland and its port, the largest in China. Banking and ﬁnancial services developed over the 1920s, both local and international (aided by the existence of the International Settlement and the French Concession outside formal national regulation (Wu, 1999). Just before its decline in 1949, it contained 24 state banks and over 200 private lenders, trust companies and other ﬁnancial institutions. Its stock exchange was the third largest after New York and London (Laurenceson & Tang, 2005). The ﬁrst exchange was established in 1891 – although this was by foreign businessmen and 90% of the members of the Shanghai Share Brokers Association (who had exclusive trading rights) were non-Chinese. By 1935 over 190 stocks were listed. Capital also ﬂowed from domestic investors to semi-private joint stock companies, through local banking networks, helping to develop a more diversiﬁed and deep indigenous ﬁnancial sector. Financial activity was concentrated in the Bund, which formed an interconnected ﬁnancial district of over 100 buildings on the West of the Huangpu River – and opposite the Lujianzui ﬁnance and trade district established as a commercial central

A Sense of History

79

business district for the Pudong New Area, created as part of Shanghai’s aim to capture a position of prominence in Asian and global ﬁnancial markets. The Chinese revolutionary turmoil, hyperinﬂation and political barriers led to a sharp decline in Shanghai’s status. The principal beneﬁciary of that decline was Hong Kong. Hong Kong was established as a colony in the 1890s to exploit trading links between China and the West, but was dominated by Shanghai as a ﬁnancial centre. As Schenk (2002) notes, Hong Kong’s rise to IFC status was driven by the gradual opening up of the PRC’s markets to trade from the late 1970s. However, Schenk argues that banking, foreign exchange and gold trading played an important role in the immediate post-war period in establishing Hong Kong as a key centre for ﬁnancial activity – and it was the position of Hong Kong in relation to the international ﬁnancial system that drove its development. After 1949, Hong Kong saw the arrival of banks ﬂeeing China, including the regional branches of international banks and branches of Chinese State Banks, as conduits for the politically difﬁcult currency and trade exchanges. These played an important role in establishing the colony as an important centre for ﬁnancial activity in Asian markets. However, Schenck argues that the ‘native banks’ – smaller niche banks operating with minimal control and licensing – played an equally important role, exploiting laissez-faire economic policies and the minimal exchange controls. Under the Bretton Woods system, currency was ﬁxed, but differential inﬂation meant that ‘free market’ exchange rates could differ substantially from the ofﬁcial rates. However, exchange controls and national regulation prevented such adjustments occurring. However, Hong Kong’s position in the Sterling Area, but with the ability to convert currencies into US dollars created arbitrage opportunities which could be exercised through the operation of the native banks. They also played a signiﬁcant role in gold trading (with smuggled gold converted into dollars). While regulations meant that much of this activity was illegal, enforcement was light and hence Hong Kong existed as a just-tolerated gap in the Bretton Woods system. The entrepreneurial trading activities of the native banks allowed some of them to grow to become signiﬁcant players (for example Hang Seng, Wing Lung and Kwong On, the three largest by the end of the 1950s). In addition to foreign exchange and gold trading, they offered commercial loans, mortgages, insurance and own account investment. They also formed formal agreements and alliances with foreign banks and acted as an intermediary between the native banks and US and European clients. For example, Hang Seng acted as an intermediary for Bank of America and used Bank of America’s international network to conduct trade internationally. As regulation became more stringent, mergers and acquisitions activity saw foreign banks (including Japanese banks) gaining control of, or taking stakes in, the local Hong Kong banks or establishing local branch networks. From 1955 to

80

Towers of Capital

1970, Hong Kong ranked in the top six cities by numbers of foreign banks, fourth in 1955, behind London, New York and Paris, passing Paris for third in 1965, with 43 foreign banks operating – well behind London’s 97 and New York’s 67, but two ahead of the French capital – and still sixth in 1970, although passed by Paris, Tokyo and Frankfurt. Also missing from the standard history is the Middle East. A number of early studies of global ﬁnancial centres identify Beirut as a signiﬁcant centre and the dominant centre of the Middle East, until the 1975 civil war. PostOttoman rule, Lebanon was under French mandate and forged links with Europe that persisted until the civil war. Beirut has large port and ‘Swiss style’ banking regulations encourage development of large ﬁnancial sector and fuelled economic growth. On the eve of the civil war, Lebanese per capita income was twice that of Jordan (Addison et al., 2001). Nasr (1978) argues that Lebanon – and Beirut speciﬁcally – was a principal point for economic penetration of the Arab East by the Western economies, but that this role became magniﬁed from the 1950s by the emergence of the Arab oil economy. Similarly, Kauffman argues that Beirut took over from Tangiers as the regional ﬁnancial centre for both the Arabian Peninsula and for North Africa. Migration of the commercial, ﬁnancial and industrial elite of Egypt, Syria and Iraq emphasised its importance. Beirut in the 1970s has many of the characteristics of an international ﬁnancial service centre including a cosmopolitan population and high-quality cultural and leisure amenities. Economically, it had become dominated by services and by an external orientation. By 1970, the share of the tertiary sector of GDP was in excess of 70%. Beirut acted as the regional headquarters of multinational ﬁrms – with Western ﬁnancial ﬁrms making up 75% of the foreign companies in the Lebanese capital. The growth of the ﬁnancial sector in Lebanon led to a growth in domestic banking, but many of these banks were taken over by multinational banking operations to establish Middle Eastern afﬁliates. Beirut thus formed an important node in the global ﬁnancial system. Schenk’s ﬁgures for world cities ranked by number of foreign banks operations show Beirut highly ranked until 1970, when civil war (and its sensitive position in the Arab–Israeli conﬂict) destroyed its chances of being the ﬁnancial capital of the dollar-rich Middle East markets. As an aftermath of the civil war, the population of Beirut fell from 1.2 million to less than 300 000. Attempts to reconstruct Beirut as a ﬁnancial and trading centre have to a large extent been frustrated by continuing religious and political factionalism and its unstable geo-political position despite having one of the most liberal banking regimes in the world, with few restrictions on capital movement. Initially, Bahrain took on Beirut’s role, to be superseded by Dubai which, in turn, is now being challenged by Abu Dhabi. Dubai’s rise is somewhat different in nature to that of other IFCs (Singapore exhibiting some similarities), being based on a very deliberate, government-led

A Sense of History

81

strategy to use accumulated capital to create a critical mass of ﬁnancial services activity.

3.5

Lessons of history: a tentative conclusion

This brief history of the development of IFCs reveals, ﬁrst, a remarkably stable hierarchy of cities, at least at the top of the ﬁnancial centre hierarchy. New ﬁnancial centres emerge and move up the rankings, capturing regional market share and niche elements of the market. Cities, particularly those in the shadow of larger, higher ranking cities, decline. Geo-political turmoil can lead to the collapse of a ﬁnancial centre – but many cities of ﬁnance have been remarkably resilient in coping with crises. But there seems more continuity than change. Second, the ﬁnancial system that evolved in the 50 years preceding World War I was inherently international in nature and contained many of the features that are cited as being characteristic of the modern system of capital ﬂows: large, sophisticated ﬁnancial ﬁrms operating on a multinational basis, signiﬁcant trans-national and trans-continental capital ﬂows, product innovation, the concentration and specialisation of ﬁnancial activity and coordination in a small number of key ﬁnancial centres which were linked together and to other cities across regions and continents by a network of branch ofﬁces, alliances, agents and service providers. This raises a number of linked questions. First, given that the post-1980s ﬁnancial system could be interpreted as a return to 1914 globalisation rather than a newly formed network, what, then, is different about the New International Financial System and its network of connected IFCs? Are there new ways of working or new products or new ﬂows that are qualitatively and fundamentally different from the practices at the turn of the twentieth century? Second, in the light of the apparent stability, what factors help maintain the hierarchy of ﬁnancial centres? Why is it so difﬁcult for a city outside the leading centres to capture market share? One factor cited as marking a signiﬁcant break with the past is the widespread application of digital computing and communications technology. As discussed in Chapter 2, telematics has two broad impacts: product innovation and remote dealing. It permits much more sophisticated and rapid quantitative analysis and data processing. This allows rapid identiﬁcation of arbitrage opportunities across assets and markets, often on an automated basis increasing the linkages in the system. In turn, this drives the development of new products and investment vehicles to seek out pricing anomalies and exploit the enhanced analytic capacity. As an example, exotic derivative products seek to beneﬁt from the temporal relationship in market volatility (periods of high and low volatility in markets and cross-asset linkages in volatility) which can be identiﬁed using statistical and econometric

82

Towers of Capital

techniques such as GARCH analysis2 which could not be used operationally without powerful computing technologies; these products are utilised by newly formed hedge funds as an investment tool. To implement such models and trading strategies is a high-cost activity, requiring very substantial investment in research and development and in technology platforms, which encourages both concentration and specialisation. The second component is the way in which telematics allows trading to take place well away from a central market place. With screen-based electronic trading platforms replacing open outcry markets and exchanges, there is no need to be located in a particular city to trade in assets notionally listed in that city. The combination of remote trading and arbitrage-seeking products locks markets ever more tightly together. However, that alone does not explain the continuing dominant position of the major IFCs. At one level, increasingly automated trading strategies and the ability to operate remotely should weaken the attraction of ﬁnancial centres: ﬁrms can locate anywhere, seeking lower cost locations (Long-Term Capital Management, for example, was based in Greenwich, CT: hardly remote, but well away from Wall Street). However, the need to innovate requires knowledge spillovers; the need to attract highly skilled and specialist staff to implement new vehicles requires access to a very large labour pool; and the need to explain new products and vehicles to clients favours face-to-face meetings. The combination of technology-driven products, communications technology and urban agglomeration economies gives strong advantages to the existing major international ﬁnancial service centres and makes it harder for new centres to compete and capture market share. Of course, once a product or vehicle is established, knowledge of its existence and operation will spread rapidly and it may be implemented elsewhere. However, the nature of arbitrage in markets means that the implementation of an effective arbitrage-seeking tool will remove that arbitrage opportunity, forcing a further round of innovation. It should be noted, though, that international telecommunication is hardly a new phenomenon. Commercial telegraph services were in existence from the 1830s. The ﬁrst cross-channel telegraphic cable was laid in 1850 and by the mid-1850s, there were cable links between London, Paris and Amsterdam. By 1866, the ﬁrst working transatlantic cable had been laid and there were cable links from Britain to India by 1870 and from there onto Australia. Clearly, these copper wire cables had low bandwidth and reliability problems, though provide a communications link across the world, but they were sufﬁcient to allow coordination of stock prices and simultaneous new issues of stocks and bonds across markets in the nineteenth century. The ﬁrst ticker tape service for stock prices was introduced 2

Generalised auto-regressive conditional heteroscedasticity.

A Sense of History

83

in the United States in 1867, providing remote access to near real time price information. Transatlantic radio communication was available from the early 1900s. The ﬁrst transatlantic commercial telephone networks were operational in the 1950s, Telstar, the ﬁrst commercial communications satellite, was launched in 1962 and the ﬁrst packet switching computer network went live in 1969. Thus structures for global ﬁnancial communication have long been in place. It is the extent of the network and the speed and reliability of secure communications that are the striking feature of the current international ﬁnancial networks. Nor is ﬁnancial market product innovation ‘new’. For example, there were quasi-options markets in tulips in sixteenth-century Netherlands, formal rice futures markets in seventeenth-century Japan, and Chicago, the ﬁrst formal commodities exchange and derivatives market, was formed in the mid-nineteenth century. There is also growing evidence of the sophistication of medieval ﬁnancial markets, with innovative techniques developed both to hedge trading risk and to evade or avoid laws outlawing usury (some akin to developments in contemporary Islamic ﬁnance). Bell et al. (2007) analyse the forward contracts for the delivery of wool between English monasteries and Italian merchants in the thirteenth and fourteenth century, demonstrating plausible discount rates and broadly efﬁcient prices of these international derivative contracts. As another example, Cassis (2006) cites the example of the development of secondary markets in Swiss tontines3 – collective life annuities where the last survivor is the principal beneﬁciary – which were linked to reinsurance plans and instalment payments to create a highly speculative investment product that spread rapidly from Geneva to other Huguenot cities and beyond in the late eighteenth century, creating a bubble that only burst with the French revolution. Another argument advanced for the newness of the New International Financial System is that it is a truly global, stateless system, which links together cities, markets and actors across the world in an unprecedented fashion. The liberalisation measures of the 1980s, permitting the free ﬂow of capital, allied to technological change and the collapse of Eastern European communism (and the opening up of the Chinese economy), it is suggested, have resulted in a qualitatively different system. The international ﬁnancial and trading links of the nineteenth century are seen as extensions of the imperial reach of the major Western powers and, as such, cannot truly be considered as a global ﬁnancial system. From the evidence presented above, this view seems somewhat myopic. While it is undoubtedly true that empires and spheres of inﬂuence inﬂuenced both trading relations and ﬁnancial activity, economic beneﬁt frequently 3

Bizarrely, tontines feature in an episode of The Simpsons … as well as in Robert Louis Stevenson’s The Wrong Box.

84

Towers of Capital

outweighed political considerations – most obviously with the continuing trade between Britain and the United States after the American Revolution. In times of war, the national debt of combatants might not be listed on their opponents’ markets, but government and private debt and equity were readily traded in the ﬁnancial markets of rivals and banking networks and alliances straddled national and imperial boundaries. There were marked differences in the extent of state intervention in the operation of ﬁnancial markets and in the restrictions placed on foreign operation in markets but there is also evidence of cross-national cooperation in maintaining stability and facilitating trade and investment at both private and government levels. Financial markets were linked across the world, and included South America, China, Russia, the Middle East and beyond. And while ﬁnancial markets on the periphery did largely act as intermediaries for the developed economies, local ﬁnancial institutions developed alongside foreign banks and created niche roles that were sensitive to their domestic economy and institutional structure. Knowledge spillovers are evident too, as product innovations in one market are quickly adopted and adapted in other markets. A more plausible argument relates to the scale of activity in the international ﬁnancial system. The total volume of ﬁnancial activity relative to world GDP has grown signiﬁcantly – in particular with the growth of ﬁnancial derivatives markets (in particular, those concerning exchange rate, interest rate and equity index) and the growth of global investment and wealth management. The size and growth of capital markets seems to be dissociated from the size and growth of world production. Further, the concentration of that ﬁnancial activity into a small number of cities – as set out in Chapter 2 – breaks down any relationship between the capital accumulated in IFCs and the wealth and economic activity of their surrounding economies and nations. This was, to an extent true of London as early as the late nineteenth century, but is now a more general phenomenon. The concentration of activity is also a concentration of specialist labour with the increases in ﬁnancial sector employment, in turn, having signiﬁcant real estate market implications. Figure 3.1 shows the Greater London Authority’s (GLA’s) estimates of City of London ﬁnancial and business service employees between 1971 and 2006. The ﬁgures suggest an increase of 41% over that period, a relatively modest compound increase of around 1% per annum to a total of around 273 000. However, there are considerable ofﬁce space implications. As a ﬁrst pass estimate, an extra 79 000 workers at the national average of 16 square metres per worker implies the need for an additional 1.26 million square metres – 13.6 million square feet – of ofﬁce space for the ﬁnancial and business service sector. The City’s own land-use survey puts the average ﬂoorspace per worker at around 23 square metres, which equates to additional demand of 1.82 square metres, or over 19.5 million square feet,

A Sense of History

85

Employment, year end

300 000

250 000

200 000

150 000

05

03

20

01

20

99

20

97

19

95

19

93

19

91

19

89

19

87

19

85

19

83

19

81

19

79

19

77

19

75

19

73

19

19

19

71

100 000

Figure 3.1 Financial and business services growth, City of London, 1971–2006. Source: Adapted from GLA employment estimates.

of ofﬁces. Some notes of caution are necessary here. If the City has become more specialised in ﬁnancial service, some of that growth can be accommodated in the space vacated by, for example, head quarters buildings. Second, that additional ﬂoorspace estimate rests on the assumption of static ﬂoorspace per worker. Third, the analysis focuses only on the administrative area of the City of London and hence does not take into account the restructuring of the ﬁnancial services industry itself, the decentralisation of retail ﬁnancial services and batch and back ofﬁce functions out of the City, nor the construction of the ﬁnancial services cluster in Docklands. These issues will be considered more fully in Part III of the book. Nonetheless it is evident that only a city with the capacity to create, maintain and enhance a major prime ofﬁce market cluster can compete as a leading international ﬁnancial services centre. The ﬂows of capital required to complete that amount of space, and the store of value the completed ofﬁce stock represents, should not be forgotten. Can this be created from scratch? Dubai is attempting to do so, Frankfurt with its Finanzplatz master plan attempted to do so, but the critical mass of space required to host a global ﬁnance centre is so vast as to create a major obstacle to competitor cities. The interlocking system of specialist ﬁnance centres coordinating and linking the global ﬂows of capital thus has long antecedents. Cities rise and fall, but there is more continuity than change. There is considerable path dependency in the world ﬁnancial hierarchy. The last quarter of the twentieth century saw both a growth in the importance of ﬁnancial markets in the global economy and an intensiﬁcation of the concentration of capital in the

86

Towers of Capital

major cities of ﬁnance. The fall of the Berlin wall and the liberalisation of China reopened the Soviet and Chinese investment and capital markets and further extended the global network of ﬁnance. Advances in information technology increased the speed of interaction between market, permitted remote trading (with contradictory results in terms of spatial concentration) and facilitated waves of innovation in investment vehicles and ﬁnancial products. But these changes were largely quantitative rather than qualitative. The idea of a fundamental shift in ﬁnancial markets seems misplaced or overstated. The preservation of the IFC hierarchy seems to rest on the interrelationship between scale and agglomeration economies. Scale of operation provides a depth and breadth of market that permits specialisation with cost advantages, that promotes innovation through knowledge spillovers and the increase in the probability of proﬁtable interactions. Scale provides a skilled labour force that can be matched to the specialised jobs demanded by new products and vehicles which, in turn, attracts skilled professionals from other markets. And the capital accumulated in those cities helps to create and maintain a favourable business environment in terms of infrastructure and urban amenities, which helps to offset the congestion costs of large cities. Central to that business environment is a critical mass of ofﬁce space, as an operational base for the ﬁnancial ﬁrms – and as an immobile store of investment value for ﬁnancial ﬁrms as investors. The next part of the book, then, turns to examine the operation of ofﬁce markets in IFCs.

Part II Inside the Ofﬁce Market

Introduction In this part of the book, the dynamics of ofﬁce markets and, in particular, of the prime ofﬁce markets of international ﬁnancial centres (IFCs) are examined. The functioning of an ofﬁce market is ‘dynamic’ because the different elements of that market interact through feedback processes. Ofﬁce rents not only depend on the beneﬁts of a particular location for business activity but also on the actual supply of space relative to demand. Supply of space depends on the cost of building relative to the price of the space created and on the availability of capital. The price of the ofﬁce space depends on the rental value and on the returns required by investors – which depend, in part at least, on expectations of future rental growth and, hence, on the demand for space. Thus, the space market (where ﬁnancial and business servicebased activity is translated into demand for ofﬁce space), the supply side (where developers create new space and redevelop existing space) and the investment market (where completed developments exchange and where capital ﬂows into and out of real estate) are linked together, changes in one part of the market rippling through to others. Chapter 4 examines the link between demand for space and rent in ofﬁce markets. At its simplest, demand for ofﬁce space is a function of the number of workers employed in ofﬁce activities and the average amount of space required for each worker. In this sense, the demand for property is a derived demand, driven by service sector activity. In IFCs, employment is dominated by ﬁnancial and business services and, more than that, by globally oriented ﬁnancial activity. By implication, then, demand for ofﬁce space in such cities is tied into international ﬁnancial conditions and will ﬂuctuate as they do. The nature of ofﬁce markets, however, means that space consumption and rents do not adjust readily to changes in economic activity. Rents are ‘sticky’ – lease contracts hamper rental and space adjustment. A shift in the market rent – the rent for a new lease contract – does not necessarily affect the rents paid by existing occupiers. A ﬁnancial ﬁrm that needs more or less space cannot simply move to a larger or smaller ofﬁce building, but must wait until its lease expires or negotiate an exit with the landlord. Furthermore, with ﬂuctuating employment, a ﬁrm can alter the space per worker ratio until it is certain that it really does require more space, reducing the space per worker and over-occupying as staff numbers increase, under-occupying and creating ‘grey space’ or hidden vacancies when they have shed staff. Nonetheless, models of ofﬁce rents show that demand shifts are the key driver of rental change in cities. Chapter 5 turns to the supply of space. Ofﬁce development is perhaps the most visible aspect of the growth of IFCs, with new ofﬁce complexes

Introduction 89

transforming the urban landscape. Developers should create new space to meet growth in demand, mindful of the cost of building and the value of the completed space. However, they face uncertainty, caused in no small measure by the length of time between starting to build and completing the ofﬁce space. With ﬂuctuating demand and the sluggishness of the supply response, rental values will rise and fall above their equilibrium level. Do developers misread the rental signals? The evidence of major supply cycles suggests that they might – in many urban markets there appear to be periods of overbuilding creating a glut of space leading to falling rents and capital values. What explains this development cycle? Behavioural explanations demand individual irrational responses from developers and from the lenders providing them with capital. Economic models applying game theory approaches, however, suggest that cycles can occur even with rational responses by developers. The factors that encourage the development of cycles – volatility of demand for space, long lags between start and completion, entry barriers for developers – are all features found in the ofﬁce markets of global ﬁnancial centres. Since development of ofﬁce space involves a transformation of the built form of cities, the actions of planners and city governments will inﬂuence the development process, affecting the amount, the location and the timing of building activity. Chapter 6 turns to investment in real estate. Ofﬁce space is a store of value. Much of the ofﬁce space in global cities is occupied by tenants and owned by investors for the returns that it delivers through rental income and capital appreciation. Even where ofﬁce space is owner-occupied, it represents a ﬁxed asset for the occupier and can be used as security for borrowing. Investors acquire real estate as part of their mixed asset portfolios since it appears to offer good risk-adjusted returns and also diversiﬁcation beneﬁts – that is, real estate returns have low correlations with other ﬁnancial assets. Direct investment in real estate – actually owning and managing buildings – is problematic. Investors face entry barriers due to the high cost of buildings, face high search, transaction and management costs, are exposed to risk since it is hard to diversify the property portfolio and must confront real estate’s inherent illiquidity. This excludes many smaller investors who must seek alternative investment routes if they are to gain exposure to real estate as an asset class. Professional investors, by contrast, hold substantial portfolios of real estate assets – and a signiﬁcant proportion of that investment is in the ofﬁce markets of major cities. As those investors move away from purely domestic portfolios and pursue global investment strategies, they seek assets in markets that have good market research, that have strong tenants, where economies of scale in management can be achieved and where there appears to be reasonable liquidity. This encourages investment in global cities and in the core ofﬁce markets of

90

Towers of Capital

ﬁnancial centres. Whether the returns achieved in those markets justify that investment and whether such a strategy achieves the diversiﬁcation gains sought with a global strategy are critical questions. In IFCs, demand, supply and investment are embedded within the global ﬁnancial system. As cities become more specialised, demand for ofﬁce space – and hence rental change – is dominated by ﬁnancial ﬁrms that are externally oriented and from business and professional service ﬁrms that depend on their activities. Supply of space reacts to rental signals and to availability of capital which, because of the scale of development in IFC ofﬁce markets, must come from major ﬁnancial ﬁrms or from the capital markets and is tied into global credit cycles. Developer success depends on the achieved prices of completed ofﬁce space, which will be acquired by ﬁnancial, business or producer service ﬁrms for owner-occupation, or by large professional investors for their mixed asset portfolios – on an increasingly global basis. Capital ﬂows affect investment yields; investment yields affect capital values – along with rents. The value of the real estate in professional investors’ portfolios inﬂuences their own ﬁnancial performance. This locking together of the global ﬁnancial system and the ofﬁce markets of global cities has major implications for stability and risk.

4 Occupational Demand, Ofﬁce Space and Rents

4.1

Introduction

Although real estate researchers protest, demand for space is essentially a derived demand. Commercial real estate exists because business users require physical space in which to operate. Ofﬁce space exists because business service ﬁrms, retail services ﬁrms and the administrative units of other ﬁrms require physical space to accommodate their staff, equipment and documents. While this is self-evident, it sometimes gets lost in real estate research, with its twin focus on development and on investment activity. It can seem to be forgotten by investment and development professionals, too, as forecasts of rental growth seem divorced from any consideration of demand drivers. Ultimately it is demand for space that drives ofﬁce rents. In historical context, ofﬁce space is a comparatively recent arrival. Visitors to Pompeii can stand in the shells of Roman shops, familiar in layout and structure. There are buildings where goods were manufactured, warehouses for storage, buildings dedicated to leisure. But, outside the political fora and the halls of rulers, there are no administrative buildings, nothing that is recognisable as a specialist ofﬁce. There are ofﬁce functions: as Harris (2005) notes, the scribes of Ancient Egypt recorded transactions, inventories and accounts and had space dedicated to that work. Formal permanent market places existed, of course, the wool and corn exchanges, for example. The rise of the merchant class and business techniques (book-keeping and accounting) in medieval and renaissance Europe saw buildings with recognisable administrative functions in Italy and elsewhere – the growth of international trade increased the need for banking, insurance and foreign exchange. But even in seventeenth- and eighteenth-century London, ofﬁce work was largely

92

Towers of Capital

carried out in what were recognisably private houses.1 However, specialist ofﬁce space became more prevalent across the eighteenth and nineteenth centuries as the need to administer and manage production and exchange of goods became more pronounced in the aftermath of the industrial revolution and as ﬁrms became larger and more complex. In London, purpose-built ofﬁce blocks begin to appear in the early nineteenth century, some built speculatively (Scott, 1996; Baum & Lizieri, 1999). The rise of ofﬁce space as commercial real estate mirrors the development of the service sector of the economy and, in particular, the growth of the business and ﬁnancial services sector. Daniels (1975) stated that the number of commercial clerks in England had risen from less than 47 000 in 1841 to some 333 000 in 1901, a growth rate of around 3.5% per annum. Even then, this represented just 4% of the employed population. By 1981, 11% of UK employment was in ﬁnancial and business services. By 2001, Financial and Business Services share had risen to 19%, growing at 3.3% per annum while the overall workforce grew at an annual rate of less than 1%. Comparative international ﬁgures are harder to obtain, but the shift to service employment is evident in developing economies. In the United States, service industry’s share of civilian employment rose from 58% in 1960 to over 72% in 2000. Over that same period, the share of ﬁnancial, business and professional services in total non-farm labour had grown from 11% to 18%. In France, Germany and Japan, over that same 40-year period to 2000, overall service employment grew in excess of 2% per annum, while civil employment as a whole grew by less than 1% per annum, taking the share of service employment to 64% in France and Japan and 73% in Germany.2 These ﬁgures mask changes in the share of wealth. For example, in the United States, in 1964 the average weekly earnings of workers in the ﬁnancial activities sector were 13% below the average non-farm wage; after annual growth of 5.25% per annum, by 2000, ﬁnancial sector workers were earning 12% above the non-farm average. These aggregate differences again conceal substantial pay differentials between those providing routine processing and retail customer contact and those in ‘high-level’ business-to-business activity, with the latter concentrated in major ﬁnancial services. That growth reﬂects a growing functional specialisation and a concentration of activity within ﬁrms. Functions that were performed within manufacturing or retail companies are now provided by specialist service providers. Sole traders – lawyers, accountants for example – operating from 1

2

Exceptions in London include the Royal Exchange, originally built in 1570 and rebuilt after the Great Fire; Token House, built in 1620, and the Bank of England, built in the early 1730s. I am grateful to Steven Devaney for pointing these out to me. Comparative statistics are taken from the US Bureau of Labor Statistics website.

Occupational Demand, Ofﬁce Space and Rents

93

300 250 ToC, US$ ft2

200 150 100 50

en

rC

ne In

M tral um P N ba ew ari s i C Yo e rk nt re M T id Sh oro -to an nt wn gh o C ai B Pu D do n W g ar sa w Ly W on el lin gt Ba on ng Ph ko ila k de Li lph m a, ia As Pe cu C in ru nc c io in n n Pa ati ra gu ay

To

ky

o

C

ity

of

Lo

nd

on

0

Figure 4.1 Total office occupancy costs, May 2007. Source: CBRE (2007).

informal space merge to form partnerships and companies, those partnerships and companies merge or are taken over, the larger ﬁrms, in turn, developing internal specialist divisions and departments with speciﬁc space requirements. This changing pattern of occupational demand has a speciﬁc spatial character at a variety of geographical scales. Internationally, economies are at different stages of development3 and perform speciﬁc roles in the global economy with, in particular, international ﬁnancial services heavily concentrated nationally. At urban level, cities within the urban hierarchy perform different service functions; at the same level in the urban hierarchy, certain cities specialise in higher order ofﬁce functions. This pattern of spatial agglomeration of administrative, business and ﬁnancial services creates uneven demand for space and, hence sharp differences in rental levels. Even focusing on the major cities in the international urban hierarchy, there are signiﬁcant variations in ofﬁce rents and total occupancy costs, variations which are, at least in part, driven by the presence of global corporations and, in particular, by global ﬁnancial services ﬁrms (see Figure 4.1). This chapter begins with a discussion of the link between land values, economic activity and location, emphasising both the importance of 3

This is a convenient shorthand and should not be taken as accepting a Rostow-style stage model of economic development, as should be evident from earlier chapters.

94

Towers of Capital

agglomeration economies and historical inertia in rental levels. It continues with an analysis of research on rental adjustment processes in ofﬁce markets. Current research links together supply of space, demand for space, vacancy rates, real interest rates and real rents in an interlocked system with feedback mechanisms that seek to correct imbalances in supply and demand. However, much of the empirical research on ofﬁce rents emphasises that the key driver, particularly in the short term, is demand – either ofﬁce-based employment or service industry output. Supply side adjustment is sluggish and complex. Thus, an understanding of rents in international ﬁnancial service centres involves understanding the nature of demand for ﬂoorspace. Competitive pressures and information technology have transformed the way that service industries conduct their business. New working practices result in more intensive use of space and a fragmentation of the ﬁrm both functionally and spatially. The literature on new business practices is reviewed and then applied to the types of activities found in international ﬁnancial centres (IFCs). This permits a mapping of the linkage between developments in the international ﬁnancial markets and rents and vacancy levels in the ofﬁce markets of IFCs.

4.2

Land value, land rent and ﬁnancial services

No attempt is made here to review the literature on land rent, land-use and transport costs.4 This section brieﬂy considers the implications of urban land-use models for the rent payable by ﬁnancial services ﬁrms before attention turns to rental adjustment processes in cities and the impact of changing working practices on the demand for space. Classical location theory implies that the value of land is a residual, a derived demand from the amount that individuals are prepared to pay to gain access to a location which has advantages relative to other locations as a site for consumption or production. For any particular land-use, the costs of inputs (labour, raw material, energy), costs of transporting products to market and the rate of return on capital employed (buildings, plant and machinery) must be paid, with the residual available to pay for use of the land. Different land-uses compete for plots of land (and landowners compete to offer plots of land) with, in a theoretically efﬁcient market, the land being allocated to the land-use able to pay the highest rent, the ‘highest and best use’. Spatially and in an urban context, assumption of a site of maximum locational advantage for a land-use with transport costs increasing 4

For reviews in a neo-classical tradition, see Ball et al. (1998), DiPasquale and Wheaton (1996), Geltner et al. (2007) and for a discussion of rent theory in general, Evans (1983).

Occupational Demand, Ofﬁce Space and Rents

95

with distance from that site creates a bid-rent curve. Firms or individuals will bid higher rents to occupy the most favourable site with rent offered falling with distance, the gradient dependent on transport costs. Given the assumption of a mono-centric city as in the Alonso model, land-use would typically be arranged concentrically around the city centre with the land-use with the highest and best use and the steepest bid-rent curve occupying the central business district. However, the bid-rent model can explain clustering without the assumption of a single point of maximum accessibility in the presence of agglomeration economies that are sector speciﬁc as opposed to more general (Jacobian) urbanisation economies. For ﬁnancial service ﬁrms, the locational imperatives include access to the largest skilled labour market, access to clients, partners, service providers and ancillary services, to market presence and visibility and, more generally, to the agglomeration economies of scale and scope that relate to interactions between ﬁrms and knowledge spillovers. If, and only if, these are localised (with an emphasis on face-to-face contacts), then location becomes important and bid-rent curves will be steep (and probably non-linear). The knowledge spillovers will be greatest for ﬁnancial services that rely on nonroutine activities, business-to-business processes and on innovation: ﬁrms (and divisions of ﬁrms) participating in such activity will face higher production costs away from the point of maximum advantage and will bid more than ﬁrms with more routine, batch and mass production activity. Further, larger ﬁrms and ﬁrms engaged in international ﬁnancial services activity will gain more from clustering and will typically be able to outbid purely domestic or regional ﬁrms. For international ﬁnancial services ﬁrms, the decisions on location are global in nature and are not conﬁned by city or national boundaries, as discussed in the prior chapters. Prime ofﬁce rents and occupation costs in the leading ﬁnancial centres reﬂect the advantages of locating there in a global context and are driven by global ﬁnancial forces as much as by domestic rent gradients. For ﬁnancial service companies, space costs are but a small proportion of labour costs. KPMG (1998) analysed costs for a representative medium-sized mutual fund over a range of North American cities. In those cities, rents as a proportion of total labour costs (salaries and employer-paid beneﬁts) ranged between 8% and 15% – that is, labour costs were between 6 and 13 times more signiﬁcant as rent. Total occupancy costs, accounting for service charges and property taxes would be somewhat higher but it is clear that location decisions are going to be more sensitive to wages and salaries. However, for some activities within ﬁnancial and business services – those routine administrative functions that do not require highly skilled high-cost staff, that can be automated and that do not beneﬁt strongly from agglomeration beneﬁts, space costs may be sufﬁciently signiﬁcant relative to

96

Towers of Capital

turnover and proﬁtability to drive decentralisation and the spatial separation of functions within an organisation. The standard Alonso-style bid-rent models of urban land-use describe an ideal equilibrium outcome based on transport costs. However, the nature of the built environment is that it creates inertia. Once a site has been allocated to a particular land-use, particularly a land-use that requires capital invested into buildings and ﬁxed plant and machinery, then all the costs of change of use (including clearance of existing property) constrain the activities on that site. Further, the existing occupier may be protected by owning the site or having a lease with some form of security of tenure. This does not preclude change of use, particularly over longer timescales, but does hamper adjustment. Where there are clearly ‘higher and better’ uses, the additional value may facilitate the change. These might arise from major economic shifts, advances in production technologies, changes in transport costs, removal of barriers to trading activity or other constraints. As an example, the development of a signiﬁcant mass of high-quality ofﬁce space at Canary Wharf in London’s Docklands created sufﬁcient agglomeration economies (and sufﬁciently changed market perceptions of the area and, consequently rents and property values) to justify the demolition of recently constructed low-density warehousing and light industrial space to be replaced by further prime ofﬁce development and luxury housing. Nonetheless, the existing urban form has an impact on the future trajectory of land-use – an impact that is often magniﬁed by planning policies and development controls. One implication of this is that there is a path dependency in land values and urban rents that arises from the historic development of major cities and, in particular, of IFCs. The existing built form – the pre-existence of a critical mass of high-quality ofﬁce space and secondary space for ancillary support activities – helps to ﬁx agglomeration economies and to create the mass of ﬁnancial activity that generates the critical face-to-face meetings and knowledge spillovers. Although there will be congestion costs, high land costs and problems of site assembly to be surmounted, it is easier to add to and upgrade the existing stock of ofﬁces and to grow a ﬁnancial centre organically than it is to create a new signiﬁcant ﬁnancial centre. The examples of Dubai and Shanghai show that it is possible to create (or recreate) major ofﬁce markets in a short space of time, but these are more exceptions than the norm. Thus, the built form contributes to and reinforces the pre-existing network of IFCs. The essence of the argument here is that ofﬁce market rents in IFCs result in large measure from the beneﬁts of agglomeration and clustering for ﬁnancial services ﬁrms in a global context and cannot simply be explained as a result of the national urban hierarchy or the size of city alone. The existing built form, the existing ofﬁce market contributes to the clustering and agglomeration beneﬁts. This links the trajectory of rents in IFCs

Occupational Demand, Ofﬁce Space and Rents

97

to developments in international ﬁnancial markets. Before developing this concept, it is important to look in more detail at the micro-processes that determine rental adjustment in ofﬁce markets and the models that seek to explain those processes.

4.3

Rent models and rental adjustment

Ofﬁce market rent is the price of occupying space in (or outside) a city, the price the occupier is prepared to pay for the beneﬁts of location in a particular building. That rents differ between cities indicates that occupiers, ﬁrms are able to, are prepared to pay more for a particular building in a particular place and, as a result, they act as a benchmark for the strengths of cities. Clearly this is an oversimpliﬁcation: price results from the interplay of supply and demand, and the supply is determined by a wide range of factors: from physical, topographical constraints, through capital availability to planning and building code restrictions. Nonetheless, the level of rents in a city ofﬁce market and the growth of rents over time tell us something about the qualities of that market. There have been many attempts to model ofﬁce rents and to determine the variables that drive rental change. This next section seeks to identify the key models that have been identiﬁed and to separate out supply and demand factors in determining rental adjustment. Any quantitative model of ofﬁce rents must confront a number of deﬁnitional issues that relate both to the market under consideration and the institutional structure of landlord–tenant relationships. These must be resolved before analysis may commence. However, analysis is further constrained by data availability. Even in the largest and most-researched ofﬁce markets – the prime ofﬁce markets of IFCs – there are data inadequacies that force researchers into compromises and hamper the development of robust, sophisticated models. Differences in data availability also hamper comparative work and have led to the emergence of different modelling ‘traditions’, particularly between US and European markets, with the former emphasising rental adjustment processes driven by adjustments to vacancy rates, the latter driven more by shifts in employment level in the market being analysed. In relation to the market, there are geographical, boundary issues – where does the market begin and end? Ofﬁcial public data are typically available for administrative, political areas – but these boundaries rarely coincide with functional markets. Rent and space availability data from private suppliers – the commercial real estate agents and consultants – are more likely to coincide with market areas, but there is rarely consensus on the boundaries. In measuring employment (as an indicator of demand), a distinction needs to be drawn between place of employment and residence of workforce – since a

98

Towers of Capital

functional urban system will include the commuting zone around the central ofﬁce market – which can be vast for an IFC. City markets may contain sub-markets that are functionally specialised and that may have separate rent trajectories linked to the growth and output of the specialist sector (although competitive pressures should prevent excessive divergence). In addition, there may be spatial interaction effects. The ofﬁce rents of two neighbouring competitor cities (or sub-markets within a large city) offering relatively comparable accessibility (to markets, clients, competitors and labour markets) and factor costs must be linked together in the long run. In principle, if the rents (and costs in general) in one city became excessive, ﬁrms would relocate to the other, decreasing demand in the ﬁrst and increasing demand in the second, producing rental convergence. This would not apply at different levels of the urban hierarchy, for cities of markedly different rank in the international ﬁnancial system. The beneﬁts of urban agglomeration, critical mass, depth of markets and access to knowledge spillovers offset higher costs. Nonetheless, there should be linkages and forces that impose convergence. Later in the chapter, possible linkages across IFCs, transcending national boundaries, are considered. The City of London provides a good example of geographical issues. Formally, the City consists of the ‘Square Mile’ – the area around the Bank of England and, administratively, it consists of the area under the control of the Corporation of London, which is a separate (and highly unusual) local authority set within Greater London. As Pryke (1991) has shown, the Bank of England used formal and informal powers to ensure that major banks remained within easy reach of the Bank. However, ﬁnancial deregulation loosened those ties and allowed the spread of the core ﬁnancial areas beyond the formal boundaries of the City. The creation of a major ofﬁce cluster in the London Docklands as part of an urban regeneration strategy has created a new node spatially distinct from the City but functionally linked and integrated. Docklands – and speciﬁcally Canary Wharf – initially functioned as a back ofﬁce location but now is home to the European headquarters (HQs) of major investment banks and ﬁnancial institutions. Within the City there are distinct sub-markets (for example for insurance and reinsurance) and, outside the City, there are specialist professional service ofﬁce markets that rely on and interact with the ﬁnancial activity of the City – notably legal services clustered in the Mid-Town/Holborn market to the west of the City (Blake et al. (2000) provide empirical analysis that conﬁrms that rents and yields are shaped by these sub-markets). Modelling rents and rental adjustment processes, then, must confront these geographical issues. The dependent rent and the independent supply and demand variables used in analysis need to be, as far as is possible, for a consistent spatial area. Not all models follow this, however. For example,

Occupational Demand, Ofﬁce Space and Rents

99

Wheaton et al. (1997) attempt to model the ‘Greater London’ ofﬁce market using variables that range in scale from central London (an area that includes the ﬁnancial market of the City of London, but also the professional services and media clusters of the Holborn market and the distinct West End ofﬁce market with non-ﬁnancial HQs and service functions), through Greater London (which includes many suburban ofﬁce clusters) to the whole of the South East region surrounding London (which is in its commuter shadow but includes free-standing towns and cities). With such a range of geographical scales, it is almost impossible to place much credence on the interpretation of the results of the analysis. There are signiﬁcant deﬁnitional issues with regard to the rent variable itself. Is the rent to be measured the average ofﬁce rent (per square metre) for the ofﬁce market under consideration or should it reﬂect top rents in the market? Rents may vary markedly between ofﬁces of different quality and speciﬁcation. A common distinction is made between Class A or prime ofﬁce space and lower class or secondary space. Class A represents the best available space for the market, typically the newest, largest, best speciﬁed space. Development of new high-quality space in a market may result in a sudden jump in rental levels (although the development itself reﬂects the anticipation of that higher potential rent from ‘highest and best use’). Generally, the rent for secondary space will be related to the Class A rents although the gap between the two may well vary at different points in the property cycle. When space is scarce, tenants will be prepared to pay more to occupy poorer quality space; with high vacancy rates, prospective tenants will be unwilling to pay for poor space when better space is available. There is surprisingly little research on this topic, possibly as a result of data constraints. Having resolved that issue (or been driven by data availability) there remain measurement issues that relate to lease contracts. The rent paid by an occupier in any one time period will relate to the date of lease origination, the length of the lease, the tenant’s right to remain in occupation at lease expiry and the arrangements for variation of rent. The rent will, ﬁrst, reﬂect prevailing market rents at the time of the letting. Arrangements for rental variation and lease lengths vary considerably across national markets. Where lease lengths are short, the rent may be ﬁxed for the life of the lease. For longer leases, rents may be indexed to a consumer price or construction cost index (as with many mainland European leases), or they may adjust periodically to market rental levels (as with the British ﬁve-year ‘upwardonly’ rent review). The implication here is that the rent paid by an occupier may differ markedly from the rent that would be payable by a new tenant in the market place. In terms of analysis, the latter is probably the more valuable variable, but the differences need to be considered. Once again, the signiﬁcant differences between markets hamper comparative analysis.

100

Towers of Capital

Lease terms also determine what is paid by the landlord and what is paid by the tenant. Who is responsible for property taxes, for insurance, for repairs and maintenance of the physical structure of the building? Is there a service charge payable by the tenant and, if so, on what basis is it calculated? There thus may be some justiﬁcation for analysing total occupancy costs rather than rent alone, if such data are available. A further complication arises in relation to longer leases where the landlord grants the tenant incentives to occupy – typically in the form of a rentfree period at the start of a lease. This may be a standard market practice (allowing a new tenant a period with no or low rent while they face moving and ﬁtting out costs) or it may reﬂect periods of high vacancy where landlords are competing for tenants and are, de facto, reducing true rents while maintaining ‘headline’ or ‘face’ rents. The rent-free period thus varies according to the letting cycle and it is necessary to convert headline (face) rents to effective rents. To do this, researchers should capitalise the tenants’ rent then turn that present value into an annual equivalent to adjust the face rent based on the prevailing lease. For a standard quarterly in advance UK lease contract, this would imply  1 − (1 + i)− P  REff − RHead  −1 / 4   4[1 − (1 + i) ] 

 1 − (1 + i)− N   4[1 − (1 + i)−1/ 4 ]   

(4.1)

where REff is effective rent, RHead is headline rent, P is the rent-free period in years, N is the lease length in years, and i is the appropriate discount rate. Having determined the target variable, ofﬁce rent models now seek to determine the adjustment process by which rents change in response to market circumstances. Although distinct modelling approaches have developed, in essence, this will involve specifying supply and demand variables and modelling the interaction between the two. Demand for space in ofﬁce markets must in some sense relate to ofﬁce-based employment (since the total demand for space is employment multiplied by ﬂoorspace per worker (FSW)), although rents will also be inﬂuenced by the output and proﬁtability of the ofﬁce-based activity performed in the market. Supply of space relates to the total ofﬁce stock in the market (and to changes in that stock brought about by new construction) and to the availability of space at a particular point in time: the vacancy rate, or the supply of new and second-hand space available for occupation. Some models propose an explicit link to the capital markets by including interest rate variables: here, the rent is considered as the risk-adjusted return required by the landlord for investing capital in the built form. It is possible to identify three broad modelling approaches relating rents to ofﬁce market supply and demand. The ﬁrst is a single equation, reduced-form

Occupational Demand, Ofﬁce Space and Rents

101

Expected real effective rents Interest rates Planning policy

Expected capital gains

Starts

Depreciation/ withdrawals

Completions

Stock/space

City employment Market share economic growth

Real effective rents Rent-free periods

General price level Headline rents

Figure 4.2 A City of London office market model. Source: Derived from Blake et al. (2000).

model, based on multiple regression, which has change in rent as the dependent variable and a set of supply and demand variables on the righthand side of the equation: rental change is driven by changes in the supply and demand variables, which may be lagged.5 This provides a basic framework for understanding the rent adjustment process, but may be weak on the dynamics of ofﬁce markets. A second approach speciﬁes a system of equations – typically equations explaining rent, equations explaining new supply, and equations explaining absorption of space. These interact: changes to rent (actual and anticipated) inﬂuence developers’ decisions to create new space; absorption depends on occupier demand for space which is, in turn, inﬂuenced by rental levels.6 Figure 4.2 provides an example of a system of equations approach, derived from Blake et al. (2000). A third, 5 6

Gardiner and Henneberry (1988, 1991) and D’Arcy et al. (1997). Hendershott (1996), Wheaton et al. (1997), Hendershott et al. (1999) and Barras (2005).

102

Towers of Capital

more recent, approach is based on error correction modelling. A long-run equilibrium relationship linking rent, employment and stock is identiﬁed. Shocks to stock and employment cause rental change: the error correction mechanism speciﬁes the adjustment back to long-run real equilibrium.7 Whichever approach is adopted, a starting point is generally occupational demand for space. The demand proxy is generally taken as employment. Initially simplifying, ﬁrms in the city employ a certain number of (ofﬁcebased) workers. Each worker (on average) occupies a certain amount of ﬂoorspace (FSW). At any point in time, total workforce times FSW represents the total demand for ofﬁce space; the vacancy rate at time t represents the difference between the total demand and the total stock; an increase in employment results in increased demand for space which, if immediately realised and with stock held constant, reduces the vacancy rate. Dt = Et · FSW

(4.2)

Vt = St − Dt

(4.3)

However, the relationship between demand and occupied stock is more complex than this implies. A ﬁrm may wish to employ a certain number of ofﬁce-based workers. However, if they already occupy space in the city – either as an owner-occupier or as a tenant with a lease contract, a change in employment is not always accompanied by a change in occupation. A ﬁrm that reduces its workforce may be reluctant to release space (for example by sub-leasing part of their ofﬁces), ﬁrst because they may be unsure that the change is permanent; second because they may be concerned about business conﬁdentiality and privacy; and third because the terms of the lease may preclude or constrain lease surrender and/or sub-letting. As a result, the ﬁrm may chose to under-occupy the building, in effect increasing the FSW and creating hidden vacancy or ‘grey space’ that does not appear in published statistics. The same principle holds for a ﬁrm increasing its workforce. It may initially accommodate the new workforce within the existing occupied space, decreasing the FSW, while seeking a more permanent solution. The speed of adjustment will reﬂect market conditions (the extent to which space is available or in demand) and the institutional structure of the market and the ﬂexibility of lease forms. In addition there is a feedback loop between rent, demand and occupation. When rents are high (above their real equilibrium level) then occupiers are more likely to wait and to increase the density of occupation of their existing buildings than to take on new space; yet increases in demand are positively related to rental growth. Assuming 7

Hendershott et al. (2002), Farelly and Sanderson (2005) and Mouzakis and Richards (2007).

Occupational Demand, Ofﬁce Space and Rents

103

D*t is the amount of space that would be occupied with no constraints and Dt is the actual space occupied, then Dt = Dt − 1 + f(D*t − Dt − 1)

(4.4)

where f measures the speed of adjustment to the desired level of occupancy (as in a standard stock adjustment model). This adjustment coefﬁcient may have permanent components relating to the institutional structure of the market and time-varying components that relate to real rent levels and conditions in the supply side of the market. Furthermore, the demand for space is also a function of the (real) rent level in the market: D*t = f (Et, RRt−k)

(4.5)

∆D*t = f (∆Et, RRt−k)

(4.6)

Demand for space increases positively with increasing employment and negatively with increases in real rent levels. The relationship between rent and desired level of occupancy (which is unobservable directly) is determined by the elasticity of demand for space with respect to the price of that space. Again, with lags in the adjustment process, past changes in employment and rent will inﬂuence the demand for space and the observed absorption of space in the market. US-based ofﬁce rent models often focus on the relationship between rent and the vacancy rate. When vacancy rates are low and approaching zero, then tenants compete for the small amounts of available space and landlords are able to charge higher real rents. Drawing from labour economics and the idea of an equilibrium rate of unemployment – the non-accelerating inﬂation rate of unemployment, NAIRU – it is suggested that there exists a ‘natural’ or equilibrium vacancy rate for a market (that is that, in equilibrium, there will be space available to permit turnover in the market and refurbishment of the stock and to allow landlords ﬂexibility to wait for the optimum tenant and tenants to ﬁnd the optimum property). Real rental change is driven by the gap between the actual and natural vacancy rate, the vacancy gap: δRRt = b(V* − Vt)

(4.7)

When the vacancy rate at time t is above the natural vacancy rate, then tenants are able to negotiate real rates downward; when the observed vacancy rate is below the natural rate, then landlords can drive real rents higher. The b coefﬁcient in the rental adjustment equation suggests that adjustment is not instantaneous, but phased over time. This formulation does

104

Towers of Capital

assume that rental change is linear, increasing at the same rate when vacancies are above or below the natural vacancy rate, and invariant to the size of the gap.8 A number of models extend the natural vacancy rate idea to include the idea of an equilibrium real rental level which will equate the rental return with a user cost of capital (reﬂecting the risk-free rate of return, a risk premium for property, and adjustments for rental depreciation over time and for management costs to provide a ‘normal’ return on property investment and development). Rents should tend to that equilibrium value, although the adjustment lags may be considerable as they operate through supply of space in the market. This provides a link to the capital markets (and to the investment market for real estate) that is absent from simple supply and demand models or is more crudely proxied by inclusion of an interest rate variable. Rent also responds to supply of space – with different models estimating a direct relationship between change in stock and rent or mediating the impact through the vacancy rate adjustment mechanism. Chapter 5 considers the supply of space in more detail. Development starts should be a function of anticipated capital value growth, which in turn relates to anticipated rental growth and change in capitalisation rates (yields). However, it is more common to see starts modelled using lagged and contemporaneous rent and interest rate variables which implies developer myopia. Completions are a lagged function of starts (due to the time taken to complete ofﬁce buildings and variations in that construction period). Finally, change to the total ofﬁce stock needs to take into account the withdrawal of space due to depreciation and obsolescence. St = St−1(1−δ) + Ct

(4.8)

Ct = g(Bt−k)

(4.9)

Bt = h(RRt, RRt−k, it)

(4.10)

where δ is the depreciation rate, Ct and Bt are new construction starts and completions at time t and it is the interest rate (cost of capital) facing developers at time t. As will be explored in Chapter 5, the anticipated rent and capitalisation rate combine to produce an expected future capital value which will be related to the cost of construction to inﬂuence the construction

8

McDonald (2002) is critical of this assumption and suggests that rental adjustment is a function of both vacancy level and the amount of tenant mobility in the market. Recent research (e.g. Farelly & Sanderson, 2005; Hendershott et al., 2008; Englund et al., 2008) has tested for non-linear responses to supply and demand shocks.

Occupational Demand, Ofﬁce Space and Rents

105

decision. The relationship will be non-linear, since negative starts are not possible.9 With real rents driving new construction starts, the presence of dynamic feedback effects in the rental adjustment process is evident. These can be explicitly modelled in system of equation and error correction models (ECMs), but are harder to accommodate in single equation models (which, as a result, tend to include lagged values of the dependent rent variable). Lagged rental growth is also linked to persistence and momentum effects which relate to behavioural aspects of the occupier market. Given this overall structure for understanding rental change in ofﬁce markets, what is the available empirical evidence?

4.4

Empirical evidence of rental adjustment processes in major ofﬁce markets

Of all the major world ofﬁce markets, London has been the subject of most research. Wheaton et al.’s (1997) model has been noted above. Hendershott et al. (1999) model the City of London market using a system of equations approach (developing Hendershott’s (1996) analysis of the Sydney ofﬁce market). Rental adjustment is driven by the difference between the actual and natural vacancy rate and between actual and ‘equilibrium’ rental levels (the latter estimated from the user cost of capital – for which the main variable is the interest rate – applied to an assumed period of equilibrium). The natural vacancy rate is estimated to be around 7%. Ofﬁce completions are largely explained as a function of the (lagged) gap between actual and equilibrium rent with the rent gap having a strong positive impact on completions. Since completions have to imply earlier construction starts, the lag structure in the model suggests that developers are responding to current and immediate past rental changes with only very limited forward forecasts. Finally, occupation of stock is driven by a two-step ECM, with absorption strongly positively related to changes in employment and negatively (and less strongly) related to rent level. The error correction term in the short-run model is strongly negative implying mean reversion in the City ofﬁce market. The model generates strongly cyclical patterns, with amplitudes falling sharply after the shock. Hendershott et al. (2002) re-estimate the Hendershott et al. (1999) model using an error correction mechanism approach which takes as its starting point a long-run relationship between real rents, demand (proxied by employment) and supply (stock-adjusted for natural vacancy rate). Shocks disturb 9

Other than in the restricted sense of withdrawing ofﬁce stock, either for change of use or abandonment.

106

Towers of Capital

the long-run relationship, but the short-run error correction mechanism acts to return the system to equilibrium. In the long-run model, employment has a strongly positive impact and supply a strongly negative impact on rental levels. The model suggests that demand for space is strongly positively inﬂuenced by changes in employment (with an elasticity close to one) but more weakly negatively inﬂuenced by the real rent level. In the short-run model, the coefﬁcient on the long-run error term is signiﬁcantly negative implying an adjustment back to the long-run relationship. Positive employment shocks increase real rents, high vacancy levels depress rents, but changes to stock have a weak, non-signiﬁcant effect other than through the impact on vacancy. Farelly and Sanderson (2005) examine the City London ofﬁce market for the period 1982–2002 and test for non-linearity in rental response. They argue that it is likely that responses are more extreme in peaks and troughs than in more stable market conditions, a hypothesis tested using a smooth transition regression technique that implies different regimes of market behaviour with a gradual change in behaviour between the market environments. Their baseline results are similar to other London models, with supply variables having limited impact on rent levels, with the elasticity of demand for space with respect to price fairly low and with rents responding more markedly to changes in demand. The strongly signiﬁcant error correction term indicates that rental values are pulled back to equilibrium. However, in more extreme market conditions – that is, when there are demand shocks or vacancy rates well above the average – they ﬁnd that rents move rapidly away from equilibrium values and take longer to adjust back. Such a response would characterise a market with strong cyclical behaviour and persistence effects around peaks and troughs. Barras (2005) sets up a complex adjustment model which includes a ‘natural rate of starts’ which is the new building required to maintain the ofﬁce stock in the face of depreciation and long-run growth in output (his estimates for London suggest a rate of around 4%; the average ratio of starts to stock for the period 1977–2006 is around 3.3%). Using GDP as a demand indicator, his model suggests cyclical behaviour with a relatively slow adjustment process working through the relationship between rent, vacancy and new supply, easily destabilised by fresh exogenous shocks. As with other models, rent level only has a weak impact on the take up of space. For other markets, D’Arcy et al. (1999) adopt a single equation approach to modelling ofﬁce rents in Dublin over the 1970–1997 period. The model seeks to explain log differences in rent in terms of lagged changes to real GDP, service sector employment and the level of new construction. The employment variables do not seem to inﬂuence rents; the best ﬁt model has a one period lag on GDP and a three period lag on new construction which is explained by ‘sluggish absorption of stock produced in three

Occupational Demand, Ofﬁce Space and Rents

107

development cycles’. The underlying theory behind the model is not clear and the model does not explain the spike in rents in 1988–1989. The rapid economic growth experienced by the Irish economy (and the Dublin area in particular) towards the end of their time series might suggest a structural shift in the operation of the ofﬁce market, making modelling problematic. Karakozova (2004) compares three models of ofﬁce returns for the Helsinki market: a conventional single equation multiple regression, an ECM and an ARIMA model. Although her focus is on total returns, her capital values are developed from a rental value series. In all three models, the major inﬂuence on returns is the demand side (measured either as business and ﬁnancial services employment, GDP growth or service sector output. Her supply side variables enter weakly, if at all: where new ofﬁce construction is included it generally has the correct (negative) sign, but is rarely statistically signiﬁcant at conventional levels. In the ECM, a strong negative coefﬁcient suggests that the market contains adjustment processes that pull returns back to their trend or equilibrium values. Another approach adopted to modelling rents has been to use a panel regression model to examine adjustment processes in a number of separate cities or sub-markets. In the United States, Sivitanidou (2002) modelled rent differentials among 18 ofﬁce markets between 1986 and 1995. She suggested that there was evidence of considerable divergence from equilibrium and slow adjustment to shocks. Adjustment processes are hampered by the effect of long lease contracts, the landlord–tenant search process and construction lags. The time taken to match tenants and landlords depends on the turnover rate, on the vacancy rate and the level of demand – and on the size of the market with tenants being more likely to ﬁnd property in larger markets which should, thus, exhibit faster adjustment. Empirically, her results showed strong employment inﬂuences on rents, both from growth and size. Construction cost has a positive effect on rents (high costs suppress development and hence reduce vacancy). Rents are shown to be extremely sticky, the strong autocorrelation pointing to sluggish adjustment. D’Arcy et al. (1997) model rents for 22 European cities, with contemporaneous GDP and lagged real interest rates. They ﬁnd no market size effects but this may result from speciﬁcation problems. Mouzakis and Richards (2007) use a panel ECM approach to examine 12 European cities using a similar model to the single equation ECM approach tested in Hendershott et al. (2002). They argue that output is a better measure of demand than employment, since technical progress may mean that productivity gains reduce the ratio of employment to output, but that proﬁtability feeds into rents. While this may be the case, a substitution of capital for labour might reduce the demand for space (or dampen the long-run growth in demand for space) with lower absorption having a negative effect on rent. Nonetheless, their panel model shows a strong positive rental response to

108

Towers of Capital

their output measure (service industry gross value added), a signiﬁcant error correction term suggesting adjustment towards equilibrium and a shortterm (but not long term) supply impact. However, the supply impact is very variable across markets and is wrongly signed for Berlin, Milano and Paris. As with the Sivitanidou US model, rents exhibit persistence. Overall, then, the models of ofﬁce market adjustment processes indicate that real rental change in major city centres is driven largely by changes in occupational demand – measured either in terms of employment or output. Demand itself is affected by real rent levels but not strongly. Price elasticities are, typically, low. High real rent levels do deter demand, but this is not a strong effect. For international ﬁnancial service ﬁrms, costs are dominated by salaries and bonuses of the labour force. Rental costs and business infrastructure costs (telecommunications and other ofﬁce facilities) are substantially lower as a proportion of turnover and high-quality accommodation may help ﬁrms to attract and retain scarce quality professional staff. This would not apply to high-volume, routine ﬁnancial activities. For example, rent and ﬁtting out costs would represent a signiﬁcant component of the costs of a call centre dealing with retail insurance or account queries, while such an operation would beneﬁt less from the agglomeration economies and spillover effects found in major international city central business districts (CBDs). As a result, such operations are typically found in satellite cities, in more peripheral areas of the national economy or even in low-cost, non-domestic locations, beneﬁting both from lower property and lower labour costs. Perhaps more surprising is the apparent relatively weak impact of supply changes on rental adjustment. This ﬁnding goes against market perceptions: the idea that supply surges (as a result of ‘overbuilding’ in development booms) drive up vacancy rates and drive down rents is deeply embedded in intuitive models of the operation of city ofﬁce markets. Changes to vacancy rates (and vacancy levels that are above natural or average levels) do undoubtedly affect rents, but vacancy is a function of demand as well as supply. Further, development seems to respond to changes in rental levels: it thus operates as part of the adjustment process. New development renews the ofﬁce stock; it may be that one impact of a development boom is to create a ﬁltering process whereby poorer quality space becomes difﬁcult or impossible to let on conventional lease terms. The rent levels on such buildings are probably not reﬂected in the measures of rent used in ofﬁce market buildings and the buildings themselves may be converted to non-ofﬁce use or form the sites for the next round of (re)development. Finally, new development needs to be seen in the context of the scale of the existing stock. In any one year, the amount of new buildings completed may be relatively small as a percentage of the existing ofﬁce space – although this may not be the case in a major development boom which might induce non-linear responses.

Occupational Demand, Ofﬁce Space and Rents

109

Between 1989 and 1992, the ofﬁce space in the City of London increased by 29%, for example, while headline rents halved. However, that period was also marked by a 21% fall in ﬁnancial and business service employment in the City. The critical mass of the existing stock acts as an inertial force. It may well be that a city that does not have a well-developed historical ofﬁce market (for example, the new CBDs being created in the United Arab Emirates) might exhibit more extreme supply side response. The implication for ofﬁce markets in IFCs is that rental change is likely to be dominated by changes in the employment and output of the ﬁnancial and business service ﬁrms based there: and that those ﬁrms are locked into the global system of capital markets and ﬁnancial ﬂows. The more global the city is, the more international its focus, then the more its rental market will be driven not by domestic regional and national factors but by the global capital market environment. As will be explored below, this has signiﬁcant consequences for investment in ofﬁce markets and for the stability of those markets. It also means that the level of rents will be determined by global demand and that, hence, there is no necessary link between rent levels in IFCs and in the surrounding cities in the national urban hierarchy. That, in turn, has implications for the type of activity found in IFC ofﬁce markets and for the way that ofﬁce space is used. This, in turn, is linked to changing ofﬁce practices and ways of working within (and outside) buildings.

4.5

Changing business practices and the demand for ofﬁce space

The rental adjustment models described above are driven by occupational demand with employment driving absorption. The lags that are inherent in the adjustment process reﬂect the ability of existing ﬁrms to respond to short-term changes in employment by adjusting the density of occupation within their leased or owned buildings. That adjustment does occur, however, implies that there is a baseline FSW that is constant or changes only slowly over long periods of time. Given changes in working practices and the technology of ofﬁce work this seems an unlikely assumption. How do new business practices and IT innovation alter the use of ofﬁces? What are the implications for the demand for ofﬁce space – and, in particular, for ofﬁce space in ﬁnancial centres? There has been considerable recent research on the impact of changing business practices on corporate real estate requirements10 with a strong emphasis on the ofﬁce sector. This work draws (explicitly or implicitly) on 10

For reviews and discussions, see Manning and Roulac (2001), Gibson and Lizieri (1999, 2001) and Lizieri (2003). This section draws extensively on Lizieri (2003).

110

Towers of Capital

a diverse range of social science literature concerning ﬂexible specialisation and the ‘new economic geography’, post-Fordist production systems, postmodernism and business reorganisation to suggest that a set of interrelated business factors have changed the ways in which corporations organise their activity and, hence, their requirement for business space. Forces typically identiﬁed include globalisation, innovation and convergence in information and communications technology (ICT), reorganisation of the workplace and the drive for ﬂexibility in the production of goods and services. While these trends are not new (see, for example, Daniels, 1985), it is suggested that they have increased in intensity in the last few decades. In the management science and business organisation literature, these forces are said to drive new ways of conducting business activity. In a more competitive and uncertain environment, ﬁrms must be more ﬂexible and responsive, making greater use of information and human capital. To create that ﬂexibility, a variety of business solutions have been proposed, including an emphasis on quality management, a ﬂattening of workplace hierarchical structures, focus on core business with outsourcing of non-core activities, and increasing use of technology to alter working practices. This, in turn, must alter the pattern of demand for real estate, change the locational requirements of ﬁrms and the way in which they manage space. Information technology, as discussed in earlier chapters, creates new locational freedoms. It also facilitates new ways of working away from a core ofﬁce. Internet access via, successively, dial-up, broadband and wireless and the growing ubiquity of home computing and laptops allows home and off-site working. This, in turn, breaks the link between workers and workstations. It permits ofﬁce intensiﬁcation strategies such as ‘hot desking’ where ofﬁce workers are allocated desks on an as-need basis. IT also permits the separation of functions within a service ﬁrm, with back ofﬁces separated from front ofﬁce and relocated to lower cost locations. This, along with downsizing, outsourcing and focus on core business affects the aggregate level of demand for ofﬁce space. Moreover, outsourcing and the creation of a ‘ﬂexible’ non-core workforce implies a division of corporate space into core space, required on a permanent or long-term basis, and peripheral space which is taken on when business needs demand, but is shed when surplus to requirements (Gibson, 1998). Table 4.1 sets out some of the business and management changes proposed and points to potential impacts on the demand for space. Empirical evidence for the extent of new business practices and property effects of these changes in practice is more elusive and largely conﬁned to the ofﬁce sector. Much of the published literature relies on the views of senior executives, on anecdote or on individual case study work with limited survey-based research. There is thus little concrete evidence that new working practices have reduced the demand for space. Within the United Kingdom, the Royal Institution of Chartered Surveyors (RICS)’s Right Space

• • • •

Investing in new technology in the workplace

Source: Lizieri (2003), adapted from Gibson and Lizieri (1999).

Voice mail E-mail Internet, Intranet Video conferencing

• Core/peripheral workforce (see IPM, 1986)

Identifying key labour inputs

• Enable the above management practices to be changed • Change the location choices available for firms’ activities

• Office activities undertaken in a range of space – home, hotels, satellite offices, clients’ sites • Intensification of use of main office site, decrease in FSW

• Greater variety in labour • Long-term space for core workforce arrangements in relation to length • Shorter term for peripheral workforce and terms of employment • Need to focus on exit strategies for • Investment concentrated on key peripheral space (core) staff • Employment risk shifted from firm to peripheral workforce

• Flatter organisational structures • Less space in aggregate • Smaller organisations • Smaller units of space • Stronger relationships with suppliers • Location influenced by supplier – client and customers relationship

• Core competencies (Prahalad & Hamel, 1990) • Delayering • Downsizing • Outsourcing

Potential impact on space requirements

Concentration on core business/activities of the firm

Impact on organisation

• Change business processes within • Increased teamwork changing internal organisations configuration • Improve/alter the communication • Need for flexibility to relocate or flows reconfigure space as organisation • Greater emphasis on teamwork and evolves flexible problem solving • Less space as one worker – one desk replaced by non-territorial solutions

Management terms used

Improve quality of output and • Total quality management (TQM) productivity of whole firm’s (Deming, 1990) operations • Business process re-engineering (BPR) (Hammer, 1990; Champy, 1995) • The learning organisation (Alder, 1990; Senge, 1990) • New working practices

Key objective driving change

Table 4.1 Management change and the likely impact on space requirements.

Occupational Demand, Ofﬁce Space and Rents 111

112

Towers of Capital

Right Price research project surveyed 45 large corporations to assess their adoption of new working practices and the impact of those practices on space needs (Gibson, 1998; Gibson & Lizieri, 1999). The survey found that new practices were commonplace (for example 56% claimed to have undergone business process re-engineering, 87% had outsourced various business functions, 53% said they employed hot desking, 62% used home working schemes). These ﬁgures, however, may mislead: the proportion of workers affected by these changes was, typically, very low. For example, in a clear majority of ﬁrms, less than 15% of workers were involved in hot desking and ﬂexible team working. As a result, the reported impacts on ofﬁce space requirements were modest and largely conﬁned to internal reconﬁguration of existing space. Downsizing did lead to reduction in space needs (although realisation of that reduction was not straightforward), but outsourcing had surprisingly little impact – partly due to the nature of functions outsourced, many of which continued to be provided in the workplace. Furthermore, outsourcing of services often creates a need for more meeting rooms which may negate much of the ofﬁce space served. An RICS-commissioned study of UK ofﬁce space usage (Gerald Eve, 1997) suggested that the link between staff numbers and space required was by no means direct. Thirty-one per cent of their respondents expected to shed staff over the next three years; however, only 15% anticipated a decrease in space. With ofﬁce intensiﬁcation strategies (hot desking, home working, non-territorial space) expected to reduce ofﬁce space per worker ratios, this ﬁnding might seem surprising. However, the survey showed that, in 90% of ﬁrms less than a quarter of staff shared workstations, conﬁrming earlier survey work by the Harris Research Centre (Richard Ellis, 1994) where for two-thirds of ﬁrms less than 10% of staff shared desks. These ﬁgures are consistent with broader social evidence. Lupton and Haynes (2000) note that the proportion of people working from home based on the 1997 British Labour Force survey was around 4%, barely increased from the 2.4% ﬁgure for non-manual workers reported in the 1991 population census. In the United States, Canter and Gordon (1999) report that, when deﬁned as employees working at home at least three days per week during normal business hours, telecommuting is already used by 32% of companies with more than 1000 employees. Facility managers expected this proportion to double in three years. Other US survey evidences suggest less impact. For example, Lambert et al. (1995) found that only 44% of respondents thought workspace design contributed to improved efﬁciency and just 28% thought it reduced costs. Deloitte and Touche (1995) found that only 4% of actual change in HQ building space was attributed to hotelling, telecommuting and similar new working practices, with only 7% of future changes in requirements attributed to such factors, implying an, at best, marginal impact. More

Occupational Demand, Ofﬁce Space and Rents

113

alarmingly, the US Department of Commerce (1998) suggested that those businesses making the highest investment in ICT on average experienced a fall in productivity. However, that report noted the increased incidence of, and opportunities for, outsourcing as a result of IT developments. Gibler et al. (2002) provide comparative research on common real-estaterelated business policies, functions and activities for Australia, Hong Kong, the United Kingdom and the United States. They ﬁnd that few companies are undertaking desk sharing, hotelling or other property-related space intensiﬁcation activities. Technology, however, was ﬁrmly embedded in working practices, with e-mail standard practice, around two-thirds of respondents employing some form of teleworking and e-business and e-procurement becoming common. There are some differences across nations: UK ﬁrms were more likely to have employed innovative workplace policies, while US ﬁrms were more likely to emphasise workplace ﬂexibility and design solutions. American and Australian corporate real estate managers placed more emphasis on technological solutions than US and Hong Kong executives. Other research has examined the role of corporate real estate managers. The general ﬁndings here are that many managers do not have sufﬁcient power or authority to inﬂuence corporate strategy, that property and other business infrastructure elements (e.g. IT provision) are rarely integrated and that strategic management of the resource is hampered by lack of effective information systems that permit full cost allocation to areas of business activity. Although there are variations, these common themes emerge in the work of Schaefers (1999) in Germany, Carn et al. (1999) in the United States, Gibler et al. (2002), Gibson (2000) and Gibson and Louargand (2001). More general reviews of corporate real estate management on either side of the Atlantic can be found in Joroff et al. (1993), Nourse and Roulac (1993), O’Mara (1999b), Piershke (2001), Schulte and Schaefers (1998) and Weatherhead (1997). By implication, this may mean that changes in working practices do not result in changes in space procurement. Thus empirical research presents a rather different, and more muted, picture to that painted by the advocates and champions of new workplace designs. That global forces of change have inﬂuenced the way that work is carried out seems evident. However, the impact is less than often claimed and change is gradual, evolutionary rather than revolutionary. Nonetheless, new working practices, new labour relations and new locational drivers, allied to shorter product cycles have changed the nature of occupier requirements and placed greater stress on the need for ﬂexibility in provision to allow corporate real estate strategists to manage business risk (Gibson & Louargand, 2001). To what extent does this inﬂuence the space needs of the ﬁnancial services ﬁrms that are the core occupiers of the ofﬁce markets in IFCs?

114

Towers of Capital

New working practice and the demand for ofﬁce space in ﬁnancial markets Gordon et al. (2005), in a report for the Corporation of London, provide a valuable discussion of decentralisation, outsourcing and ‘offshoring’ in the ﬁnancial services industry which reviews much of the recent literature and survey-based research. The context for the study is that increasing competition in international ﬁnancial services following deregulation and the removal of many of the institutional barriers to capital ﬂows places a great emphasis on cost reduction. In ﬁnancial services, salaries and wages form the majority of costs (typically reckoned to be two-thirds of the costs incurred by ﬁnancial ﬁrms), with IT infrastructure, telecommunications and real estate costs also being signiﬁcant. Given that international ﬁnancial service centres like London have high labour and occupation costs, a central city location can only really be justiﬁed for activities that require rapid access to markets, rely on agglomeration economies and frequent face-to-face contacts, or where the products are complex, have high risk exposure or rely on innovation. This makes cities like London vulnerable to relocation of ﬁrms or divisions of ﬁrms. However, they conclude that the shift in the nature of employment in the City towards more specialised, high added value, business-to-business activities suggests that the impact will be relatively muted. The types of ﬁnancial service activity that are most susceptible to outsourcing and business process re-engineering are those that are relatively routine – ‘plain vanilla’ – that do not require complex interactions within the ﬁrm or with clients and customers: batch processes, high-volume, low added value activities, processes that do not require close senior supervision. These types of activity are less time-sensitive, less complex and do not require proximity to markets. Elements of them may be automated. The proﬁts from such activities are highly cost-sensitive and, hence, the high labour and property costs found in IFCs will damage the returns received by shareholders and stakeholders. However, to assess the potential impact this might have on demand for space in IFCs requires consideration of the structure of ﬁnancial service activity as it has evolved across the end of the twentieth and the start of the twenty-ﬁrst century. A ﬁrst key distinction to be made is between retail and wholesale transaction activity. Retail activity involves sales to households and individuals. These are widely distributed, making concentration unnecessary for market access. Processes are less time-sensitive and, with lower transaction values, there is less need for involvement of senior, skilled staff. There are no obvious agglomeration beneﬁts from a central city location and much of the day-to-day activity is routine. As a result, much retail activity is subject to decentralisation away from major ﬁnancial centres, ‘near-shoring’

Occupational Demand, Ofﬁce Space and Rents

115

(relocation to regional cities with lower labour and property costs) and ‘offshoring’ (relocation out of the nation state to low-cost locations). In many instances, large elements of the processing of these retail transactions may be automated, which reduces the pressure for offshoring but also reduces the required workforce – individual staff can deal with much greater volumes of activity – and hence reduces the demand for ofﬁce space. One obvious example of the spatial separation of retail ﬁnancial functions is the development of call centres well away from ﬁnancial HQs. Initially, call centres were established in lower cost peripheral regions of the home nation of ﬁnancial ﬁrms. Call centres then became more international in nature (Ireland, for example, capitalising on investment in education to become a key locale within Europe) and then global with a growing awareness of the pool of skilled graduate labour available in developing economies, notably in India. The establishment of global call centres has been enabled by developments in telecommunications that not only allow seamless intercontinental calling but also permit real-time monitoring and supervision. While call centres are the most visible (or perhaps, most audible) instance of near- and offshoring, other areas that have been separated away from head ofﬁce functions include IT functions and business support (accountancy, payroll, maintenance of electronic databases and record systems). The potential for growth in offshoring is considerable – Gordon et al. cite estimates from Forrester Research which forecast 1.2 million jobs outsourced in Western Europe, with the majority of that coming from the United Kingdom and with a third of the outsourcing being from the ﬁnancial sector. The rapid growth of ﬁnancial activity in developing economies also raises the possibility of these developing as autonomous ﬁnancial centres in their own right and becoming competitors with the existing IFCs. There are already examples of ﬁrms offering specialist ﬁnancial services to a variety of clients (rather than the ﬁnancial clients establishing remote branch ofﬁces) and establishing their own branch ofﬁces (for marketing and client contact) in traditional ﬁnancial centres. More recently, concerns have emerged which spring from the rapid success of offshoring and the growth of satellite ﬁnancial centres in the developing economies: rising labour and property costs, increasing staff turnover, issues of quality control and quality management and, above all, worries about data integrity and security. However, retail ﬁnancial operations – retail banking, household insurance, for example – still generate considerable pools of capital, assets that must be managed and require sophisticated corporate level risk management and control strategies. These are market-sensitive and do beneﬁt from agglomeration economies, knowledge spillovers and access to specialist suppliers of products and information. This implies a spatial separation of functions within retail ﬁnancial activity with the corporate level asset and

116

Towers of Capital

risk management functions centrally located. The staff employed by those central functions will be of higher status, demand better working conditions (and greater FSW) and will (at least, it is assumed) generate higher proﬁts, justifying the payment of high rents. Gordon et al. (2005) note that this process is magniﬁed by the competitive pressures that have reduced margins and commissions and led to a new focus on proprietary own-account trading to maintain proﬁtability and augment retail client income. Proprietary trading is volatile and dependent on market conditions. Wholesale ﬁnancial activity – market-sensitive trading with companies and private investors involving high-value, complex deals – can be found in investment and commercial banking, asset management, insurance and in the own-account trading of retail ﬁnancial ﬁrms. For much of this activity, face-to-face access to clients is critical and time-sensitive, creating strong agglomerative forces. Examples might include issuance of debt and equity in the primary securities markets, broking and dealing in secondary markets, mergers and acquisitions and portfolio and asset management. At a global level, as earlier chapters demonstrated, such activity is increasingly concentrated in a small number of major ﬁnancial centres: those centres’ ability to capture market share helps to ensure strong demand for ofﬁce space but, at the same time, locks turnover and the generation of proﬁts (which can be captured as property rents) to the ebb and ﬂow of global capital markets. Broking and securities trading can be split into a number of functions which, simplifying somewhat, can be split between front ofﬁce, middle ofﬁce and back ofﬁce activities. Front ofﬁce activities involve actual dealing and trading which, in turn, involve direct contact between client and trader. Some of this activity can be automated (pricing models, automated batch trading systems for index tracking funds), which may permit greater activity per broker. Much of the work, though, demands client contact and information gathering and, in turn, much of that involves developing relationships through face-to-face contact. This, allied to labour market competition, favours concentration in the major ﬁnancial centres. Middle ofﬁce functions involve trader support, reconciliation and account and risk management. Proximity is important here although not critical. There is some evidence that international ﬁnancial ﬁrms place their middle ofﬁces near their primary trading team (almost certainly in a major IFC), but that middle ofﬁce also services remote locations. Time zone differences require some local support, however. Back ofﬁce functions deal with clearance and settlement that are more routine activities with a far less pressing case for proximity, although concerns about risk management and the need for monitoring act as a consolidating pressure. Back ofﬁce areas are prone to business process design and automation, with programme trading and straight through processing systems reducing stafﬁng requirements. Even here, there is a necessity to have skilled staff to deal with exceptions, problems

Occupational Demand, Ofﬁce Space and Rents

117

and non-routine trades: it may be harder to recruit and retain such staff in non-central, non-IFC locations. The critical importance of information technology in ﬁnancial services has given rise to a new real estate demand for specialist data centres. These might be dedicated to a single company or provide a specialist environment where ﬁrms can lease space in a highly controlled environment, operational and support equipment and access to high-speed internet connections and bandwidth. As McAllister and Loizou (2007) note, this creates a locational imperative: access to secure power sources and to ﬁbre-optic networks: centres ‘must be connected to multiple backbone ﬁbre and local loops to allow connectivity with other hosting equipment and redundancy in case of disruption’. For IT connected with trading, risk management and complex clearance and settlement, global ﬁnancial ﬁrms run back up systems in parallel with their real-time trading processes: ‘synchronous replication’. Since these back up sites must replicate real-time working and with signal attenuation, these cannot be too far from the main trading operation. However, greater awareness of global terrorism threats in the wake of September 11th have also driven the need for remote data recovery centres: for example, in the United States, the SEC proposed that all those ﬁnancial institutions with ‘signiﬁcant’ roles in ‘critical’ ﬁnancial markets must have fully operational recovery sites located at least 200–300 miles away from their primary data centre site. Similarly, disaster recovery and business continuity centres need to be away from the main ﬁnancial centres but, nonetheless, accessible to the key skilled staff, hence favouring suburban locations near main cities.

4.6

Employment, volatility and the use of space

Overall, then, telematics developments permit the separation of ﬁnancial service functions and facilitate remote working. This has signiﬁcant implications for retail ﬁnancial services: the routine nature of ﬁnancial transactions, dispersal of the client base, the high-volume low added value nature of trades and competition drive cost reduction strategies that look to automation, redesign of business processes, relocation of operations to low-cost environments and outsourcing. However, much of wholesale, business-to-business ﬁnancial activity still beneﬁts from agglomeration economies and market access. The same telematics forces that favour decentralisation in retail ﬁnancial services favour concentration in wholesale ﬁnance and in the own-account asset management of retail ﬁnancial ﬁrms, as the beneﬁts of market access, product innovation, information spillover and competitive labour markets drive concentration in the largest IFCs. It may be possible to use IT to increase the productivity of traders and

118

Towers of Capital

300 000

250 000

200 000

150 000

19 71 19 73 19 75 19 77 19 79 19 81 19 83 19 85 19 87 19 89 19 91 19 93 19 95 19 97 19 99 20 01 20 03 20 05

100 000

Figure 4.3 Financial services employment in the City of London. Source: Adapted from Greater London Authority.

to separate out and relocate back ofﬁce and routine support functions but much of the key front and middle ofﬁce activity requires a central location. Furthermore, the shift towards higher value, complex products and activities, to more technologically driven business practices and to high-status, high-salary staff alters the demand for space: while ofﬁce space may be used more intensively, the space demands per worker in IFCs may increase. Redevelopment of IFC ofﬁce buildings to provide that high-speciﬁcation space, in turn, removes secondary, cheaper space from the market, driving away more cost-sensitive activities and increasing the functional specialisation of IFC occupation. Is this evident empirically in IFCs? As Figure 4.3 demonstrates, ﬁnancial and business service employment in the City of London has increased substantially from the 1970s to the present, the upward trend partially masked by a strong employment cycle, with sharp downturns after 1974, 1988 and 2001 (employment adjustments – or the reporting of those adjustments – seem to lag ﬁnancial crisis), with long employment growth upswings – a 59% increase between 1977 and 1988, a 49% increase between 1994 and 2001. Trend growth over the period is around 1% per annum. This growth does not necessarily translate into increased ofﬁce demand. Financial and business service employment may be displacing other ofﬁce-based activity – for example, the HQs functions of corporations, government and other public sector administrative functions. However, within the City of London, there seems no indication of a decline in employment and demand for ofﬁce space. By contrast, statistics for employment in ﬁnancial activities for New York City see a fall in employment between 1990 and 2007 – the trend growth rate over that period being −0.6% per annum (Figure 4.4). Clearly

Occupational Demand, Ofﬁce Space and Rents

119

550

Thousands

525 500 475 450 425

19

90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07

400

Figure 4.4 Financial employment in New York City. Source: Data taken from Bureau for Labor Statistics, www.bls.gov

employment was affected by the events of September 2001 although employment was already falling with the end of the dot.com and technology boom. The recovery between 2003 and 2007 may be curtailed by the sub-prime mortgage ‘crisis’ and credit crunch with some early evidence of falling employment in the early months of 2008. The aggregate fall, however, conceals shifts in the nature of ﬁnancial activity within New York. In 1990, the largest sub-sector was banking- and credit-related activities – with a 31% share of ﬁnancial employment and 163 000 employees, just ahead of securities and commodities trading (30% and 159 000 employed). By 2007, employment in credit and banking had fallen to 96 000, a fall of 3% per annum; by contrast, securities and commodities trading had grown at 1% per annum to 187 000. The respective shares of the two sectors stood at 20% and 40%. While part of that change may reﬂect reclassiﬁcation of functions and enterprises, it is likely that the more routine banking activities have moved out of New York and more market-anchored and complex activities have grown over the period. This is supported by the parallel fall in insurance-related employment, falling at over 2% per annum from 83 000 to 57 000. These seem relatively large swings in micro-structure over a short period. US ﬁnancial activity employment statistics include real estate and leasing, which account for a ﬁfth of all ﬁnance, insurance and real estate employment in New York. Figure 4.5 shows FSW trends in New York from 1993 to 2005. Despite the fall in employment, these show a decline from 1993 to 2001, followed by an upturn. However, this upturn is largely a function of increasing vacancy rates in the city (which rise from 7% in 2000 to 13.5% in 2003). FSW in the occupied stock drifts down from 270 square foot per worker to 244 square foot in 2001, then stabilises at around 250 square feet thereafter. The New York ﬁgures are for the whole of the city and are not conﬁned

120

Towers of Capital

340 FSW Floorspace per worker, ft2

320

OccFSW

300 280 260 240 220

05 20

04 20

ch ar M

03 M

ar

ch

20

02 M

ar

ch

20

01 ch

20 ar M

00 ch

20 ar M

99 ch

19 ar M

98 M

ar

ch

19

97 ar

ch

19 M

96 M

ar

ch

19

95 ch

19 ar M

94 M

ar

ch

19 ch

ar M

M

ar

ch

19

93

200

Figure 4.5 Floorspace per worker, New York City, 1993–2005. Source: Estimated from PPR data. OccFSW, Occupied floorspace per worker. 24 23

21 20 19

2

m per employee

22

18 17 16 15 1977–1982

1983–1986

1988–1991

1996–2000

Figure 4.6 Floorspace per worker, City of London. Source: Adapted from Corporation of London (2001).

to ﬁnancial services or the downtown area although the shock impact of September 11th is evident. There is a statistically signiﬁcant negative correlation between rental growth and change in FSW, pointing to price sensitivity, although the causal relationships between the variables are likely to be complex.

Occupational Demand, Ofﬁce Space and Rents

121

27

m2 per employee

25

23

21

19

17

15 UK banks

Foreign banks

Insurance

Other Accountancy Legal financial services

City of London

Figure 4.7 Space usage by employment sector, 1996–2000. Source: Data taken from Corporation of London (2001).

Returning to London, the Corporation of London (2001) publish results from their rolling land-use survey that show, in contrast to New York, an upward drift in FSW in the City (Figure 4.6). In the 1977–1982 survey, the occupied space per worker was 20.9 square metres (225 square feet); by the 1996–2000 survey, this had risen to 23.3 square metres (251 square feet). These ﬁgures should be treated with some caution, as they do not account for vacant space11 – but since reported vacancy was higher in the 1996–2001 survey than in 1977–1982, the difference would be magniﬁed. Given that ﬁnancial and business service employment in the City increased from 193 000 in 1980 to 247 000 in 1998, this implies a need for an additional 1.7 million square metres (18.5 million square feet) of ofﬁce space for ﬁnancial services occupation in the City. In practice, the ofﬁce stock of the City increased by about 1.3 million square metres over that period, pointing to a growing functional specialisation in ﬁnancial services in the City over this period: a topic that will be considered in depth in later chapters. Figure 4.7 shows FSW in the City of London broken down by ﬁnancial sector from the 1996–2000 survey. Foreign banks have the highest FSW – at nearly 26 square metres per person: and foreign banks increased their share of the ofﬁce market over the 1980s and 1990s. Most of the incoming foreign banks were investment banks arriving after Big Bang and focusing on high-level securities, fund management and mergers and acquisition 11

They are compatible with the occupied FSW for New York shown in Figure 4.5. The London and New York series report very similar space occupancy statistics.

122

Towers of Capital

10% 9% 35%

>10 000 5000–9999 2500–4999 1000–2499 500–999

15%

0–499

14%

Figure 4.8

17%

Unit size of office space in the City of London, 1996–2000.

activity. UK banks and insurance company space usage were both over 23 square metres per person, as were those law ﬁrms in the City – law employment in the City increased by nearly 9% per annum between the 1977–1982 and 1996–2000 surveys, with much of the growth resulting from the expansion of international law ﬁrms with a signiﬁcant focus on ﬁnancial business. Accountancy ﬁrms had an FSW of 19 square metres. Accounting ﬁrms have utilised desk-sharing policies and, with much work taking place in client ofﬁces, there is scope for ofﬁce intensiﬁcation. However, FSW in accounting increased over the four surveys. It may be that less high-value activities have been decentralised or squeezed out of the City of London market. In sum, there seems no evidence of a fall in FSW in the City of London: any business process re-engineering and ofﬁce intensiﬁcation strategies being offset by a shift in the nature of ﬁnancial activity towards higher valueadded work with greater space demands. Two other trends that emerge from the City of London land-use surveys relate to the age and size of buildings. As Figure 4.8 shows, the City has considerable diversity in ofﬁce buildings. Just over a third of ofﬁce units (deﬁned as space occupied by a single occupier) are larger than 10 000 square metres (108 000 square feet), but nearly a ﬁfth of units are smaller than 1000 square metres. However, the newer space is typically larger. The average unit size in the 1977–1982 survey was 926 square metres, but had risen to 1280 square metres by the 1988–1991 survey. Many of the older buildings are smaller and many are vacant pending redevelopment. Pre-1946 buildings and, particularly, buildings constructed between 1946 and 1969 are shown as exhibiting high vacancy rates in the 1996–2000 survey. New development is predominantly targeted towards the higher value-added occupiers, whether for owner-occupation or as tenants.

Occupational Demand, Ofﬁce Space and Rents

123

Conclusions: demand and rent in IFCs Demand for ofﬁce space results from the activities of business, ﬁnance and professional service ﬁrms and the administrative and management functions of ﬁrms in other sectors as a derived demand. Shifts in the patterns of employment and modes of working within those sectors of the economy that use ofﬁces will, consequently, drive ofﬁce rents – as the price of occupying ofﬁce buildings. In turn, those rents are capitalised into building values and the relationship between building value and the costs of land acquisition and construction will determine the supply of space. Demand for space, supply of space and rents are linked together in dynamic systems with complex feedback mechanisms with the vacancy rate playing a key role in the adjustment process. Empirical evidence from research on rental adjustment processes reveals the importance of the demand driver but also considerable inertia, adjustment lags and the ‘stickiness’ of rents in cities. Competition, globalisation, cost pressures and developments in ICT have led to major changes in the patterns of demand for ofﬁce space. ICT enables the spatial separation of functions within service ﬁrms; routine, standardised functions that do not require market proximity or close supervision can be moved away from core HQ buildings, to lower cost locations (possibly in peripheral regions, in different countries and even continents). Lower costs here are generally deﬁned in terms of wages and salaries but with real estate costs inﬂuencing location choices. ICT also enables staff to work from home, from satellite ofﬁces or from clients’ buildings. This permits ofﬁce intensiﬁcation strategies based on having less desks, less workstations than employees: hot desking, ofﬁce hotelling and similar arrangements are intended to reduce FSW ratios. Empirical evidence to date suggests that the gains are less than some of the ‘gurus’ of business process reorganisation imply, but impacts are discernible. Outsourcing reduces the size of the core ﬁrm and creates demand for space from specialised service providers and start-up ﬁrms spawned from the parent company. In turn, this creates a demand for a wide range of space, best provided in large urban ofﬁce markets. What impact have these changes in the pattern of demand had on IFCs? As discussed earlier, globalisation, deregulation and advances in information and communications technologies permit both greater concentration and dispersion, but it is the former that seems to dominate. The leading IFCs have captured greater market share of complex, high added value ﬁnancial activities. This results in greater functional specialisation in those markets – with retail ﬁnancial functions, non-ﬁnancial activities and routine administrative functions squeezed out by competition from wholesale ﬁnancial sector and the specialist business and professional services ﬁrms serving that sector and by rising labour and property costs. Competitive pressures

124

Towers of Capital

to reduce costs have also led to some decentralisation of back ofﬁce and support functions, although risk management and monitoring requirements may produce offsetting cohesive forces. This growing functional specialisation seems to be accompanied by a shift towards employment of higher status, higher wage, higher value-added workers – whose higher demand for space (a function of status, IT needs and the requirement for private meeting space to exploit the agglomeration economies and knowledge spillovers of face-to-face meetings) counterbalance ofﬁce intensiﬁcation strategies. These shifts in demand and in the use of space also require a reconﬁguration of the ofﬁce space in cities in general and in IFCs in particular. The deregulation of the 1980s and consequent spatial shift in international ﬁnancial market activity, for example, triggered a wave of construction in many of the major cities in the global urban hierarchy – a supply boom that, in turn, led to coordinated crises as cyclical falls in demand led to high vacancy levels and falling rents. The late 1980s and early 1990s development and rent cycles ﬁnd echoes in earlier and in subsequent cycles. Yet, the rental adjustment models described above emphasise adjustment processes and a return to equilibrium in ofﬁce markets. Yet it seems as if developers and investors reacted to short-run shifts in demand and above trend real rental levels in an irrational way. How can the pronounced property development cycles in major urban ofﬁce markets be reconciled with ideas of market efﬁciency and equilibrium? What are the consequences of the actions of property developers on the ﬁnancial and economic stability of IFCs?

5 The Supply of Space

5.1

Developers, development cycles and research

Standard textbook presentations of real estate economics portray property as a derived demand. The shifts in employment in global cities and international ﬁnancial centres described in Chapter 4 change the requirement for space – both quantitatively and qualitatively. Developers should respond to these changing tides of demand by producing appropriate space for the ﬁnancial and business ﬁrms, making returns appropriate to the risks of that supply. Major ofﬁce buildings, though, are not mass-produced manufactured goods. The length of time between deciding to build and completing the building creates uncertainty for the developer. Once built, a building cannot be relocated nor readily be removed; it has a long life and occupies a site in what is probably an already heavily developed city. The decision to develop cannot be taken by the developer alone: there are planning regulations which vary from city to city and even within cities,1 developers typically need to negotiate to obtain debt or equity capital, there are other interest groups in the city who are drawn into the development process. These and other factors make the development process complex and difﬁcult to model. Figure 5.1 shows new ofﬁce construction starts in the City of London between 1977 and 2006, with, superimposed upon the starts, the change in real prime ofﬁce rents. Two features are immediately obvious. First, there is a pronounced clustering of starts producing a cyclical effect, with a major 1

For example, in Central London, the West End – mostly under the jurisdiction of Westminster City Council – has both a high proportion of historic ofﬁce buildings and a very restrictive planning policy; the City of London, the main ﬁnancial core, has veered between tight and permissive planning, while to the East, the newly emerged ﬁnancial cluster in Docklands evolved from urban regeneration policies with deliberately minimal planning restrictions.

Towers of Capital

800 000

50 Starts

700 000

40

Real rent % 600 000

30 20

500 000

10 400 000 0 300 000

Per cent

126

–10

200 000

–20

05

03

20

01

20

99

20

19

97

95

19

93

19

91

19

89

19

87

19

85

19

19

19

19

19

19

83

–40 81

0 79

–30

77

100 000

Figure 5.1 City of London office starts and change in real rent. Sources: CBRE, Corporation of London, Office for National Statistics, Author.

construction boom between 1987 and 1989, a collapse in development activity in the early 1990s, then further clusters of development in the late 1990s and the second half of the 2000s. Second, while there is a relationship between construction starts and changes in real rents, there appears to be a lag. Contemporaneous change in rent has a 0.4 correlation with starts as a percentage of stock; the correlation rises to 0.67 if rental changes are lagged one year. The 1987 construction peak occurs with real rental growth near zero; in 1988 starts remain above trend even with real declines in rent levels. This trailing relationship becomes much more signiﬁcant when the lag between starts and completions is considered. With typical lags between starts on site and completions of between two and three years, it seems as if London developers are developing into falling markets. Rents are, of course, a function of supply: however, starts also seem to lag changes in employment by some two years. Average start levels of around 3.5% of stock per annum are not inconsistent with the growth in ﬁnancial services employment in the City, given reasonable assumptions on depreciation. However, the pattern of activity over time seems hard to square with a model that assumes rational economic behaviour. This pattern, while highly pronounced in London, is by no means unique. Figure 5.2 shows development completions as a percentage of total ofﬁce stock for Frankfurt and Milan. In both markets, there are waves of ofﬁce building followed by periods of low activity. In the United States, Chicago and New York both show pronounced clustering of ofﬁce development in the late 1980s and, to a lesser extent, in the early 2000s (Figure 5.3).

The Supply of Space

7

127

Frankfurt Milan

6

Per cent

5 4 3 2 1

19

8 19 4 8 19 5 8 19 6 8 19 7 8 19 8 8 19 9 9 19 0 9 19 1 9 19 2 9 19 3 9 19 4 9 19 5 9 19 6 9 19 7 9 19 8 9 20 9 0 20 0 0 20 1 0 20 2 0 20 3 0 20 4 0 20 5 06

0

Figure 5.2 Office development as % of stock, Frankfurt and Milan. 0.05

New space as % of stock

0.045

Chicago

0.04 0.035

NY Mid-town

0.03 0.025 0.02 0.015 0.01 0.005 19 88 19 89 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07

0

Figure 5.3 Change in office space Chicago Downtown, New York Mid-town. Source: From data supplied by La Salle Investment/Torto Wheaton Research.

In Japan, Oizumi (1993) suggests that between 1972 and 1990, ofﬁce space in Tokyo increased by 250% – a compound growth rate of over 7% per annum. Much of that growth was concentrated in the second half of the 1980s. Even in the already heavily developed central wards, total ﬂoorspace increased by 74% (over 3% per annum), a growth accompanied by an increase of ofﬁce’s share of total space from 44% to 55%. Machimura (1998) and Dehesh and Pugh (1999) both argue that this was a function of the growing incorporation of Japan into the international ﬁnancial system. Machimura, in particular, emphasises that the restructuring of central Tokyo in response to

128

Towers of Capital

250 000 225 000 200 000

Yen m

175 000 150 000 125 000 100 000 75 000 50 000 25 000 0 77

19

79

19

81

19

83

19

85

19

87

19

89

19

19

91

93

19

95

19

97

19

99

19

01

20

03

20

05

20

07

20

Figure 5.4 Private non-housing construction investment in Japan, 1977–2007. Source: Data taken from Government of Japan, Ministry of Land, Infrastructure, Transport and Tourism.

the demands of global ﬁnancial service ﬁrms (albeit mostly Japanese ﬁrms operating globally) represented a major cultural watershed. Whatever the cause, the consequence of the supply surge and the subsequent bursting of the Japanese bubble economy was a fall in land and property values from their early 1990s peaks variously estimated at between 45% and 85%. With over-investment in commercial property in the bubble phase, the downturn had signiﬁcant long-term effects on the Japanese economy (Figure 5.4). Such cycles are, of course, no new phenomenon. Hoyt (1933) in his classic One Hundred Years of Land Values in Chicago, set out a mechanism that has inﬂuenced thinking about property cycles: in an economic upswing, excess demand drives up rents and escalates selling prices, which in turn stimulates increasing renovation and new building: excess supply drives down rents. Hoyt chronicles a series of cycles with an average length of around 18 years. He also suggested that there were ‘exceptional’ cycles where links between capital markets and the property cycle produce a ‘mania for building’ which continues past the demand peak and is followed by a sharp bust phase with foreclosures, bank failures and ﬁnancial distress. His analysis period covers the growth and development of Chicago as a major metropolis, ﬁrst as an industrial hub but increasingly as a service and ﬁnancial centre. Creating a major ofﬁce complex in a city changes the skyline and the spatial form of the city. We recognise cities by their proﬁle; buildings, old and new, become iconic representations of the character of the city. The very visibility of real estate development – and ofﬁce development in world cities in particular – makes it an obvious research topic for a wide range of academic disciplines: a partial list would include architecture and design, urban geography and sociology and other urban social sciences, planning and regeneration, urban economics and real estate. In turn, the varied

The Supply of Space

129

epistemological, methodological and political stances found within those different subject areas have created a battleground of research methods. It would be impossible to do justice to the competing arguments about how one should seek to approach an explanation of the development process, still less to resolve those arguments. This chapter focuses on two broad approaches: economic modelling that seeks to explain aggregate development in terms of supply and demand and the adjustment processes that balance those two forces; and institutional or agency processes that put far greater emphasis on the behaviour of developers and other actors in the creation of the built environment. The cultural turn in urban social science has increasingly emphasised the symbolic meaning of buildings and the use of those meanings: this is outwith the scope of this book.2 The chapter, then, explores our understanding of the production of ofﬁce space in IFCs.

5.2

Development, building costs and property value

In standard representations of the efﬁcient operation of the real estate market, there should be no overbuilding. For example, in the much-cited DiPasquale and Wheaton (1992) or Fisher (1992) four-quadrant model, a long-run equilibrium is established for rents, asset prices, the available stock and the rate of supply of new construction. Rent is established in the space market, with the equilibrium rent matching the demand for space with the available supply. That rent is capitalised into an asset price in the investment market, the capitalisation rate providing an appropriate return for investors. Prices in the asset market combine with construction costs to determine the amount of building activity that takes place; as asset prices rise, so more space is constructed, due to lower cost of capital and with more costly sites becoming proﬁtable. The models typically assume a relatively low supply elasticity and increasing long-run marginal costs in the supply of space. Finally, new construction adds to the overall amount of space in the market which, when netted for depreciation and withdrawal of buildings, produces the total stock. Such models represent a steady-state longrun equilibrium position. They require that there are rational supply and demand responses to prices and rents and that supply adjusts to increasing asset levels.

2

Which is not to say that I do not have strong views on the topic! While I see much merit in the excellent work of, for example, Zukin, Haila or even the later writings of Jane Jacobs, much of the cultural turn seems to me to have produced work that is deliberately and unnecessarily obscure, impossible to test or verify, relativist in the worst sense and of no value for policy makers (of whatever political persuasion).

130

Towers of Capital

The four-quadrant model emphasises long-run equilibrium. The nature of real estate as an asset, however, means that adjustment can be a lengthy process. In particular, there are relatively long lags between initiating construction and supply becoming available for occupation. Rents in the occupier market can (but do not necessarily) adjust quickly to shifts in demand for ﬂoorspace. Asset prices in the investment market can (but do not necessarily) adjust quickly to a change in investment demand or to changing perceptions about the relative risk of the asset class or of future long-run real growth. Supply, however, cannot respond quickly to a price shock or to new information. The lags between taking the decision to build and making the building available are long. They also vary by type of real estate and by location. Large complex ofﬁce buildings have long lags. Locations where the supply of sites is constrained (for example by the existing built form and/or by planning constraints) will exhibit longer lags relative to other markets. Some have argued that planning policies can exacerbate cycles, particularly if they seem to operate ‘counter-cyclically’. If planning policy relaxation occurs after a sustained burst of demand pressure, this may trigger a development surge at the ‘wrong’ time, leading to over-supply.3 That over-supply must be absorbed in subsequent periods. The problem confronting the developer is that they must incur costs now to develop over a (partially uncertain) period and then sell (or reﬁnance) the building, obtaining an unknown and highly uncertain price or capital facility. A smooth and rapid transition to the equilibrium position relies on the developers’ ability to estimate future trading prices accurately. If developers in aggregate are myopic, making use of, and extrapolating forward, current market trends, then they are likely to overreact to pricing signals. An upward shock in demand, for example, cannot be instantly matched with an increase in supply. The inelastic supply curve, then, implies upward pressure on real rents. Other things equal, this might be expected to raise asset prices which, in turn, trigger development. If the shock is transitory, the new construction, when completed, will be in excess of that needed to meet demand, driving rents down. If the demand shock is permanent, but developer behaviour myopic, then the result is likely to be a classic cobweb or corn–hog cycle adjustment process, generating rent and development cycles. The ofﬁce developer, then, must assess the future asset value of the building – the present value of future net operating income, rents minus costs, accounting for the costs of rental depreciation and any tax beneﬁts of depreciation allowances. That asset value has to be set against the discounted costs of developing: land, design and construction costs, professional fees and marketing costs. With free entry of developers, the return each makes 3

I should acknowledge Phil Allmendinger for helpful comments on this point.

The Supply of Space

131

should only be normal proﬁts and hence the long-run expectation should be that value is equal to replacement cost – that is, Tobin’s Q, the ratio of market value to asset replacement cost would equal one. If Q is less that one, then assets are selling below their replacement cost – so investors should be attracted into the market. If Q is more than one, then developers can produce space at less than sale price and, hence, generate excess proﬁts. Of course, the construction lag means that the developer should not assume that the sale price will be equivalent to the current price – however, that is precisely the assumption that is embedded in the much conventional development appraisal models and software. Hendershott and Kane (1992) use the Tobin’s Q argument in their analysis of the late 1980s ofﬁce construction boom in the United States. They argue that the Value to Replacement Cost ratio should be one less the ratio of the present value of expected below equilibrium rents (BERI) to replacement costs (adjusted for any depreciation tax shields): BERI 1 V = 1− RC RC (1 − t t zt )

(5.1)

where t is the tax rate and z is the value of depreciation allowances. If BERI is greater than zero, then the V/RC ratio is less than one and development should be curtailed. If (as in Hendershott’s underlying models), rents are expected to revert to their equilibrium values, rents above equilibrium today should indicate that BERI will be positive. Yet Hendershott and Kane show that US developers increased ofﬁce construction in the second half of the 1980s, as rental values peaked, creating overbuilding and a consequent over-supply of ofﬁce space in the market. Wheaton (1999) also examines the impact of developer forecasts of future values in their decision-making. If developers extrapolate current values, then a cobweb cycle can arise as developers overreact to above equilibrium price signals, over-supplying the market and driving rents below equilibrium (choking off supply in the next period). However, even if developers are rational, they cannot time future demand shocks and oscillations in the market can develop under certain conditions. Factors that lead to a cyclical tendency include the lag between start and completion (the longer the lag, the greater the amplitude of swings) and the rates of demand growth and depreciation (the faster the growth, the more likely it is that there will be cyclical behaviour). Wheaton suggests that, for the United States, ofﬁce development has long lags, has seen strong growth in demand and has relatively inelastic demand relative to supply: all of which suggests that cyclical behaviour is likely to be a characteristic of major ofﬁce markets. While noting that UK supply elasticities with respect to price are lowered by restrictions placed by the planning system and the structure of the

132

Towers of Capital

construction, Nanthakumaran et al. (2000) ﬁnd strong value impacts from new supply. They note that ‘volatile markets will be characterised by large changes in new supply in response to changes in prices and, in turn, large falls in price levels in response to increased supply’. Formal econometric modelling of development has typically adopted an adaptive expectations framework, with development responding to signals about demand from the market place. Following Ball et al. (1998), property development cycles have ﬁve broad phases: 1 Business Upturn: an upturn in the business cycle, particularly if accompanied by credit market liquidity and low real interest rates, generates strong economic activity and demand for space. Available space is absorbed, vacancy rates fall and rents rise. Rising rents (and, possibly, falling yields – see Chapter 6) increase capital values, making development appear proﬁtable. Developers are able to obtain capital and build: but the lag between starts and completions maintains rental growth. 2 Business Downturn: the business cycle turns down, perhaps accompanied by higher interest rates (either to counter inﬂationary pressures or due simply to excess demand for loans). New demand for space slackens, while development projects completed in the latter part of the upturn are completed, leading to the start of a rise in vacancy levels. With slackening growth prospects and higher interest rates, capital values fall (although this may lag the business downturn due to appraisal inertia) and developers may ﬁnd difﬁculty in selling completed speculative developments. The rate of new construction starts falls, but projects that are on site continue. 3 The Adjustment Phase: rising vacancy levels lead to rental value falls. Developers are unable to sell their completed developments and are unable to reﬁnance loans. Even where ﬁrms can service debt, falling capital values mean that loan to value covenants are breached and developers with negative equity have incentives to default. Property developers enter bankruptcy or are forced into major capital restructurings. Nonperforming loans in the banking sector lead to policy changes that eliminate property lending or tighten lending terms considerably. 4 Slump: occupier demand is low and vacancy levels high, forcing rents below their equilibrium level which makes development appear unproﬁtable. In any case, developers are unable to obtain capital to initiate new projects. Poor real estate market performance deters investors, who seek other assets, pushing down real estate capital values. 5 Recovery and the Next Cycle: with returning occupational demand and depreciation reducing the stock of available space, vacancy levels fall and rents recover. The lack of construction starts in the previous phase(s) means that any upswing in demand cannot be readily absorbed within the existing stock, triggering rental growth and a new round of development.

The Supply of Space

133

Ball et al. also note that the impacts of a development boom–slump cycle may extend into the next business cycle if the over-supply was sufﬁciently large that the vacant space created cannot be absorbed by increased occupier demand. Equally, one could argue that the memory of the last downturn was sufﬁciently fresh to make developers, lenders and investors cautious – but that memory fades and caution evaporates with the passage of time. There is one major problem with this standard characterisation: if everyone knows this is how property cycles operate, why do developers overbuild? Why do lenders provide capital for those developments? Why do investors pay high prices for ofﬁces at the top of the cycle? In particular, since development is visible – many agents provide ‘crane surveys’ of city ofﬁce markets and publish development pipeline statistics – why are the later developments started? This may require a behavioural rather than an econometric explanation. Most econometric models start from the basis that development is driven by occupier demand which, in turn, is proxied using some output measure. For the ofﬁce sector this might be an economic aggregate (typically GDP, if possible broken down to provide service sector output) or an employment measure such as the numbers employed in business services or in ﬁnance, insurance and real estate (FIRE). The economic underpinnings of the model come from the accelerator principle that ﬁrms have a desired level of capital stock (including real estate) that is related to output. Investment in new capital stock is related to the depreciation of the existing stock and any growth in output. This would imply that capital investment – for our purposes ofﬁce development – would be linked to changes in output. Reaction would be asymmetric, since development cannot be negative, with increases in economic activity being accompanied by demand for new space. However, ﬁrms experiencing growth may delay demand for new space, to avoid disruption costs, because of the friction of existing lease contracts and because of uncertainty about the new activity levels. They may increase density of occupation of existing space or seek alternative procurement strategies before signalling their demand for additional ofﬁce space. Similarly, developers may not respond instantly to the new demand signals. As a result, lags and phasing occur. Adapting Ball et al.’s notation K*t = vYt

(5.2)

I*t = v(Yt−Yt−1)

(5.3)

It = z(K*t−Kt−1)

(5.4)

where K*t is the desired capital stock at time t for output Y at time t, with v the desired capital/output ratio; I*t is the investment required to reach the desired capital stock K* given a change in output and It is the actual

134

Towers of Capital

investment, which is a proportion of the difference between the desired and the existing capital shock, with that proportion, z, assumed to be less than one. This accelerator principle is the underlying basis, for example, of Richard Barras’s work on UK ofﬁce development (Barras, 1983, 1994; Barras & Ferguson, 1987a,b). Barras’s model suggests that lags between demand signals and supply (from the lag between starts and completions) along with the phasing of stock adjustment process can produce cyclical variation independent of any demand cycles. Development starts lead to completions with a lag; the existing stock depreciates at a (constant) depreciation rate – hence ofﬁce space in any one period is a result of the space in the previous period less depreciation plus new ofﬁce space started in a previous period. Analysis shows that with depreciation set to values between 1% and 4% (consistent with other estimates of ofﬁce depreciation rates: my own work suggests a value of between 2% and 3% for London ofﬁces) and with the value of z, the proportion of the required new stock resulting in new development, greater than 0.25, cycles will occur. With plausible lags, the model suggests that a business cycle of four to ﬁve years produces an ofﬁce development cycle of eight to ten years (this is consistent with Wheaton (1987) and Leitner (1994) both of whom see ten-year cycles in US ofﬁce development). As business and development cycles coincide (every second cycle), they reinforce each other, producing a more pronounced boom and slump. The Barras and Ferguson model adds further explanatory variables: in particular, it speciﬁes a link to investors and asset markets by considering property prices, returns for other asset classes4 and includes building and ﬁnancing costs. Investment activity seems to reinforce short cycles caused by demand ﬂuctuations while variations in development costs seem to have a weak effect on outcomes. As noted above though, an inherent problem with such models is that they must implicitly assume myopic behaviour on the part of developers and the suppliers of capital for development – and, indeed, of investors, since anticipation of below-equilibrium rents should affect capitalisation rates and hence property values before rental peaks are reached. Of course, this is only a ‘problem’ if developer behaviour is expected to conform to economic rationality, but that is an assumption of the economic theory underlying the models.5 4

5

This might suggest the impact of some sort of capital switching model (see Lizieri & Satchell, 1997a, for a discussion), but more one that operates across asset markets than the ‘circuits of capital’ model advanced by Harvey, for which there seems little empirical support. Antwi and Henneberry (1995) using a residual valuation based approach present evidence that UK ofﬁce development starts are inconsistent with perfect foresight or rational expectations, but are well explained by current price taking or by extrapolation of recent rental and value trends, which they label ‘habit persistence’.

The Supply of Space

135

Although there are severe data issues,6 long-run exploratory analyses of commercial real estate development and investment in buildings (at least in developed economies) show a correlation with economic growth. With the noise of short-run cycles and spikes removed, commercial property development accounts for a relatively low and relatively stable share of national income (see, for example, Ball & Wood, 1996; Ball & Grilli, 1997). Cycles and building booms are evident, but there do appear to be adjustment processes that pull supply back to its trend level. If that is the case, do cycles matter in aggregate? That depends on the extent to which over-investment in ofﬁces, excess lending to developers and over-investment in completed buildings have real effects on output in the adjustment phase that follows a building boom. Econometric models that focus on ofﬁce development as the dependent variable can also be criticised for failing to consider fully the dynamics of property markets. For example, most models are driven by a demand variable or set of variables: service sector employment or output. Change in demand translates into a requirement for new ofﬁce space. However, a ﬁrm’s demand for ofﬁce space is also a function of the rental costs per square metre, which is a function of the supply of space in the market. Thus effective demand for space is partly endogenous. This is recognised in some of the multi-equation models of ofﬁce markets discussed in Chapter 4 – for example in Wheaton et al.’s (1997) model of London ofﬁces. The development, occupier and investment markets are locked together, so consideration of one part in isolation can miss critical feedback loops. As we will see later, the interlocking of space, asset and supply side in the ofﬁce markets of IFCs is far greater than would be the case in, for example, regional industrial markets or suburban retail development. The models here suggest that major ofﬁce markets may be more prone to cyclical volatility: the interlocking of markets makes them more vulnerable to major shocks, to booms and to busts.

5.3 Economic models and developer behaviour That quantitative models of the development market appear to be driven by current and historic prices and demand indicators is something of a problem if the underlying assumptions are that developers, investors and providers of ﬁnance are operating in a rational economic manner. In principle, developers should be (as accurately as the current information set permits) forecasting future demand for product and the impact of future supply and demand on rents and prices. They should be aware of cyclical tendencies both in demand 6

See Ball and Tsolacos (2002) for a discussion of problems in estimating construction output.

136

Towers of Capital

for space and in real estate markets and take account of those ﬂuctuations in their decisions. They should factor in the behaviour of other developers in their decision-making. If the evidence suggests that this is not the case, then a complete explanation of the development process needs to model the actions of individual developers in different market conditions. Henneberry and Rowley (2002) insist that econometric modelling approaches are deﬁcient because they do not capture developer behaviour. By modelling, in aggregate, construction responses to shifts in demand, interest rates and values, the models are simply treating each developer as a price-taker. While this characterisation of econometric modelling research is something of a straw man, it allows them to argue that a complete explanation of developer behaviour must encompass the operational and economic environment in which construction takes place and the way that developers perceive and respond to opportunities and constraints. Outcomes in the development market result from those individual responses (and the actions of other individuals within the development process). Individual responses may be systematic or they may be haphazard and intuitive and the actions of individual actors in the development process can have signiﬁcant effects on outcomes in different markets facing similar economic conditions. This points to the second major cluster of ofﬁce development research, which focuses more on decision-making and market context. Guy and Henneberry (2002) suggest three such approaches: event-sequence or systems approaches that examine the development process as a set of interconnected stages within which events occur, the stages linked via feedback loops; agency or behavioural models, where the roles, behaviour and decisions of different actors and their interrelationships determine outcomes; and institutional models, which deﬁne the actors, organisations, networks and practices of the development process and seek to place them in a broader socio-economic context.

Institutional approaches A major strand of research that starts from such a premise has been labelled the institutional approach.7 Institutional approaches draw (fairly loosely) from institutional economics. At the risk of caricaturing a complex literature, institutional approaches start by identifying an overall institutional environment: the political, social, economic and legal structures and institutions that organise a particular society. The term institution, in this context, does not map onto a speciﬁc organisation – they are – to cite North (1990) – ‘the rules of the game’. This immediately focuses on many of the implicit assumptions of standard economic models and suggests that 7

For reviews of models that place development in an institutional framework, see Ball (1998), Ball et al. (1998) or Guy and Henneberry (2002).

The Supply of Space

137

responses observed are conditional on the institutional structure. A second level examines the institutional structure of the property market itself, which provides the context for the development process: property rights, the legal structure, the land-use planning, development control and building regulations that apply in particular markets, ownership structures and other institutional arrangements. Within this institutional framework, development occurs (or does not occur) as a result of the interaction of individuals and organisations; the players and actors in the development process who act within the constraints of the institutional context. Market outcomes, thus, are seen as contingent both on the overarching socio-economic framework and on individual behaviour. There are many variants of this model: see for example D’Arcy and Keogh (1999, 2002). One much-cited variant of the institutional approach is found in the work of Patsy Healey (Healey & Barrett, 1990; Healey, 1992): the ‘structure and agency’ approach has been widely adopted by urban social scientists researching development and urban regeneration and appears implicitly in much of the world city literature cited in earlier chapters. Structures are the ‘material resources, institutional rules and organising ideas which agencies acknowledge’ (Healey, cited in Ball et al., 1998) – material resources are the wealth, land and property rights, labour, ﬁnance, information and expertise; rules govern the ways in which these resources are used, established by organisations or the political process. Ideas ‘inform the interests and strategies of actors as they deﬁne projects’ (Healey, cited in Ball et al., 1998). These structures, then, shape and constrain the actions of agents in the development role. Ball (1998) argues that the Healey and co-workers approach sees agents as individuals, working within organisations. The activities and actions of agents, interacting with other agents, produce outcomes, albeit outcomes that are framed and inﬂuenced by the institutional structure which, in the theoretical formulation of the model (if not always in the subsequent empirical work), is given causal importance. The signiﬁcance given to individual decision-making can be seen in Guy’s (Guy & Henneberry, 2002) work. As with Michael Ball’s work, he examines different types of actor in the development process, architects, developers, occupiers, investors and agents. The views of individuals within each group are shaped by the development of that particular niche, by practice and culture: this, in turn, shapes decisions. He suggests that development practice is based ﬁrmly on entrepreneurial initiative and on ‘gut instinct’, allied to localised market knowledge, backed at best by basic ﬁnancial residual calculations, rather than on any formal modelling or forecasting process. However, developers must interact with investors and suppliers of capital; particularly in downturns, this can stiﬂe innovation and activity as potential purchasers and ﬁnance providers act with a caution lacking in the upward phase of cycles.

138

Towers of Capital

Empirical research on development from a structure and agency perspective generally starts with an identiﬁcation of the stages in the development process and identiﬁcation of which individuals and organisations are involved in that process. The decisions, strategies and interactions of those actors are then analysed and related to their structural position and access to or control of resources; from this, it can be seen that the interactions of actors, the agency role, is key in determining local outcomes. However, although the theoretical explanations stress the importance of the structural forces, it is not clear from case study research as to how the structure plays a causal role, both in determining the positions of actors and agents and in shaping their strategies and determining their success or failure In much of the research, there seems too much agency and too little structure. As an example of the beneﬁts of a more detailed analysis of the development process at city level, Fainstein’s (1993) The City Builders presents a carefully argued investigation of urban redevelopment in New York and London which does acknowledge the wider social and economic forces and constraints that shape and constrain the strategies of developers and other actors in the development process. Her focus is on the major redevelopment and regeneration schemes that took place (or were proposed) in London and New York in the 1970s and 1980s, including the London Docklands and Battery Park City schemes. She ably shows how such major developments both spring out of global economic forces and result from the interactions between developers, ﬁnanciers and investors, politicians and community groups. The global economic forces are those outlined in the earlier discussion of the evolution of world cities: the growth, in scale and velocity, of international ﬁnancial ﬂows; the growing concentration of command and control functions in global cities and the massive technical infrastructure needed by multinational ﬁnancial and business service ﬁrms. These are the driving forces that set the context for urban land-use conﬂicts in the two cities. Fainstein’s case studies of major regeneration schemes chronicle the battles that community groups fought to modify development schemes or to wrestle out local gains and beneﬁts from the projects, some successful, some failures. Importantly, she is able to show that development is a negotiated process: what is ﬁnally built is often far from the initial vision, whether that view comes from local politicians and community (whose aspirations are eroded in contact with market forces) or from developers (whose aspirations are compromised as concessions are made to win permissions and to avoid delay). That negotiation is also affected by social and economic forces: pressures for cities to compete with other urban areas tilt policy towards growth strategies and give developers and occupiers genuine threats

The Supply of Space

139

in debates with planners and politicians.8 And yet, and yet: the two cities have hugely different planning and political systems and structures (albeit with a convergence in urban redevelopment strategies in the 1980s), different spatial settings and competitive pressures, different cultural attitudes to planning and public intervention reﬂected in institutional structures, different geographical settings. But the ofﬁce development outcomes are strikingly similar, for all the nuances of the development process in individual cases. The ﬁrst edition of The City Builders was published in the aftermath of the major property downturns in London and New York and those failures cast a long shadow over the book. In particular, much attention is paid to the failure of Olympia and York (O&Y), developer of Canary Wharf in London’s Docklands and the World Financial Center in New York’s Battery Park City. Fainstein notes the connection between development, space and ﬁnance that I will focus on in later chapters: she cites the 1991 merger of Chemical Bank and Manufacturers Hannover, motivated by the two banks’ difﬁculties from exposure to under-performing real estate loans, which then released over two million square feet of ofﬁce space into an already weak New York ofﬁce market. She notes the problems that developers face in predicting demand for product in the future and suggests that the real estate industry is inherently cyclical – she draws analogies with agriculture and notes that development is a highly competitive industry with few barriers to entry and with individual developer outcomes determined by the actions of other developers. Yet that is insufﬁcient to explain the ‘overbuilding’ that occurred in both cities in the second half of the 1980s. In two of the most ﬁnancially sophisticated cities in the world system, in markets with international suppliers of capital, developers and advisors, if most impartial commentators could see over-supply on the horizon, why did developers keep building? It is at this point that Fainstein’s explanation tends to the individual, to the behavioural aspects of the development process. While arguing that capital ﬂows drive development, she argues that individuals in the development industry are by nature entrepreneurial and risk-taking. Developers want to build and will build if they are provided with capital. Listed property companies must grow, or face criticism from analysts and shareholders. Developers believe that their projects will succeed (even if there appears to be an economic logic against it) and seek to persuade others of that view – particularly the banks – who face their own competitive pressures to gain 8

Massey (2007) provides further evidence from London of the pressures on politicians, planners and city managers to accept development to maintain ‘global competitiveness’. She questions who such policies truly beneﬁt.

140

Towers of Capital

market share and grow their loan books. The lenders, the advisors, the potential investors can be convinced: a herd instinct can occur which creates an environment where short-term trends are pushed forward, where due diligence is relaxed and where capital ﬂows into development. She notes that few property companies had their own research units with decisions often based more on intuition, on ‘gut feel’ than on ﬁnancial analysis. Her discussion of the fall of O&Y, while acknowledging the unravelling of their complex ﬁnancing arrangements and the state of international commercial property markets (the simultaneous falls in values meaning that their international portfolio offered few diversiﬁcation gains), focuses on the role of the Reichmann brothers. Decision-making in O&Y is portrayed as being driven by the characteristics and attitudes of its founders; personal decisions with no strong management structure or formal analysis, a tolerance of risk and an enthusiasm for the deal, personal ambitions that were as much about scale of achievement as about accumulation of money. In the preface to the second edition of The City Builders published in 2001, Fainstein wrote ‘I did not expect the upturn to come nearly as early as it did, nor that it would be so particularly powerful in New York and London. At that time, before the Internet revolution and the 1990s stock market boom fuelled by high tech, the prospects for strong economic recovery within the two cities looked rather bleak. Now, in retrospect, I can see that these two cities, with the repositories of creative talent and their global centrality were well poised to exploit the trends of the century’s ﬁnal decade’. She goes on to claim that, despite this, no building boom was triggered: ‘Financial institutions and developers, traumatised by their losses during the recession of the early ‘90s simply did not allow themselves the speculative ﬂing on which they had embarked ten years before’. Certainly there was no equivalent building boom in New York in the late 1990s, although Figure 5.1 suggests that there was signiﬁcant development in the City of London. But had developers and ﬁnancial institutions really learnt from the early 1990s? The build up of ofﬁce starts in the mid-2000s in advance of the 2007 sub-prime mortgage crisis casts doubt on that. The dampening effect of the need to absorb excess space from the previous building boom, along with memories of difﬁcult market conditions may inhibit building booms in the ﬁrst demand cycle after a property slump – but those memories fade and it seems that demand growth and credit availability can trigger new building booms. Fainstein is sceptical of much of the critical social theory literature on urban development, for example dismissing Harvey’s ‘secondary circuits of capital’ theory as a misunderstanding of the lags inherent in real estate investment. She denies that development proﬁts are purely parasitic, pointing out the value created in development, facilitating agglomeration, promoting business efﬁciency, rebuilding disused and deteriorating areas of cities. She

The Supply of Space

141

questions the idea (in Sassen, for example) that urban redevelopment has created cities that are more divided, fractured and fragmented and characterised by social exclusion – in history, were cities really heterogeneous, mixed Jane Jacobs neighbourhoods? She does not fully accept the post-structuralist critique that development imprints the built environment with attributes of ‘privatism, competition and commodiﬁcation’, while acknowledging that the developments resulted in ‘islands of glittering structures’ amid decaying public facilities and deteriorating social and economic conditions for poor and excluded groups. Her policy conclusions do not resist development but argue for more mixed use schemes with greater social diversity. Another example of research that examines real estate from a critical stance without ignoring economic forces that shape cities can be found in Anne Haila’s work (e.g. Haila, 1997). Haila notes (as does Zukin) that development in global cities is increasingly global in nature itself. Major landmark ofﬁce developments in world ﬁnancial centres are designed by ‘signature architects’ and are (less often) developed by international construction ﬁrms. Zukin (1992) noted this in relation to London and New York, but it is a widespread tendency – seen perhaps most clearly in the emerging ﬁnancial centres, in Shanghai, in Kuala Lumpur, quintessentially in Dubai. Haila also suggests that, just as construction methods have become uniform across cities, so too have ﬁnancing methods. She also notes that these signature buildings are increasingly in international ownership – a theme that will be developed in more detail later. Finally, she returns to the idea of buildings as signs: both for the city and for the developers and investors and as a statement about the focus of a city’s growth path. While still stressing the role of dynamic individuals and ﬁrms in this development process (‘with energy and sheer force of personality, Lipton set a fantastic pace at Broadgate’), her emphasis on the common features of urban development in world cities surely emphasise structural forces over the role of individual agency. The competitive nature of ‘signature architect’ ofﬁce construction and design reﬂects the inﬂuence of individual decision-makers, even if the outcomes have a common ﬂavour – the desire to outdo rivals in terms of size or striking design produces competition within and across cities (which city will have the tallest ofﬁce building and for how long), with the process mediated by planning procedures (the tallest buildings are increasingly found in emerging global cities, in the Gulf or in Asia). Such competition does not necessarily create economic efﬁciency. As early as 1989, Vandell and Lane (1989) showed that there was at best a weak relationship between ‘good design’ and investment return, with the extra costs of design not necessarily outweighed by higher achieved rents or lower vacancy rates. They note: ‘good design may not, in fact, be more proﬁtable on average but, as with a lottery, may provide a small probability of a high return for the developer’. Of course, state-funded (and, indeed, private) mega-ofﬁces are

142

Towers of Capital

perhaps more about symbolic power than economic return. It is important to stress, though, that while such buildings are striking symbols, they are generally only a relatively small proportion of overall ofﬁce space in a global city. Institutional analyses of ofﬁce development, then, offer much greater depth and nuance in examining the development process, focusing on the interactions of the various actors in that process, on the behavioural basis of the decisions made, on the conﬂicts that are inherent in land-use change in cities. They are important in emphasising the role of planning and politics that, in much econometric work is relegated to the magnitude of an adjustment coefﬁcient or the lag on an explanatory variable and in stressing the social implications of the development process. The best institutional analysis remains conscious of the importance of structure, of the wider economic and technological forces that shape cities and do not simply emphasise the individual in their case studies. At issue, really, is the causal power of individual decision-making and behaviour. An ofﬁce development project is the result of individual entrepreneurial endeavour, a process of negotiation with bankers, potential occupiers, property consultants and ﬁnal investors; and a process of negotiation with planners, politicians and other affected groups and stakeholders. The change in the stock of ofﬁce space in a city, the ﬂow of capital into and out of real estate, the impact of completed development on rents and property values, while the aggregation of those individual negotiated processes, is qualitatively different and, surely, more determined by the wider structural forces than by individual behaviour and the whims of developers. Overbuild and prices will fall, developers will fail. The developments, though, exist and form part of the overall stock, to be absorbed, over time, by the businesses in the city as the adjustment processes take effect.

Heroes and villains It is a relatively easy step from an institutional or structure and agency approach with a focus on actors to an approach that lionises or demonises the developer. The developer as villain is a recurrent motif of critical texts on urban development, shaping the city in the interests of capitalist proﬁt and with no concern for the social impacts of that development.9 The developer as hero battles against bureaucratic obstacles and constraints to realise a vision of the completed development. The focus on the individual is also, understandably, found in histories of particular cycles or phases

9

For example, in Ambrose and Colenutt’s (1975) The Property Machine, or Fitch’s (1993) The Assassination of New York.

The Supply of Space

143

5000 4500

IPD

4000

A B

Index value

3500 3000

C D

2500 2000 1500 1000 500 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

Time

Figure 5.5 Simulated performance of four development companies.

of a city’s development.10 Traditionally, development is an archetypically entrepreneurial activity, demanding individual risk-taking and attracting ‘larger than life’ characters. But does this attention on the individual developer really help to understand the evolution of cities? Can the acts of individual developers shape a city’s fate, in the face of structural forces affecting occupier, investment and ﬁnancial markets? As a teaching tool, I often use a variant of Figure 5.5 which shows the performance of a set of developer-investors, benchmarked against Investment Property Databank (IPD) over a 30-year period. The student groups are asked to discuss the performance of Developer B, who has outperformed the market by a considerable margin. What qualities or attributes does the developer have? Typically, factors cited include entrepreneurial vision, stock selection or timing skill, optimal use of ﬁnance, being in the ‘right’ market, superior market research or market intuition, ‘feeling’. What is rarely cited is luck. This is interesting since the data series are actually generated by simply applying a random number multiplier to the IPD returns each year. Initiating the recalculation function on the spreadsheet produces 10

From the United Kingdom, examples would be Marriott’s (1967) The Property Boom, Ross-Goobey’s (1992) Bricks and Mortals, or, in a more formal economic history context, Scott’s (1996) The Property Masters.

144

Towers of Capital

a completely changed ordering of ‘developers’. Market performance is the aggregation of the outcome of individual decisions, some good, some bad, some producing above average returns and some below average. Those developers whose timing or location is unfortunate disappear from the stage, leaving the successful to take star billing. The developers who survive downturns may believe that they have done so because of superior entrepreneurial skills. But past success is no guarantee of sustained success: the lions of one cycle become gurus for the market but are as likely to fail in the next cycle.11 A focus on the individual developer may be informative on the psychology and social context underlying decisions made, but it is not at all clear or convincing that it explains the aggregate behaviour that shapes the urban landscape.

5.4

Real options and developer strategies

As is often the case, the image of economic analysis of the development process as coldly deterministic and rigid is very much a straw man. First, by their very nature, econometric models are probabilistic in nature and much attention is paid to the range of possible outcomes and to error terms. More importantly, as time series analyses have evolved, models such as vector error correction models seek to identify general trends and links between variables (for example, some measure of economic demand and new construction completions), allow for deviations from that general trend or relationship but also specify a process – the error correction mechanism – by which the long-run relationship or steady state is restored. The activities of individual developers – or time-speciﬁc behaviour by groups of developers or other actors in the development process – can pull the market away from its steady state, but adjustment processes will kick in. ‘Excess’ development in a market will, other things equal, lower rents and capital values, making development less proﬁtable and curtailing the supply – irrespective of individual behaviour. This is not to deny, of course, that, for any particular market, this ‘stable state’ is socially constructed – the adjustment processes reﬂect institutional arrangements, planning rules and legal structures, for example – nor that economic, social or technology driven structural shifts can occur that alter the underlying relationships between supply and its driver variables.

11

More remarkably, some gurus seem to be able to maintain that status despite multiple failures, reassembling companies from past failures, imparting their wisdom at professional conferences and seminars, then using their market feel and intuition to crash again.

The Supply of Space

145

More signiﬁcant, perhaps, is the developing body of work that seeks to combine real options models and game theory to explain the behaviour of economic agents in the development process. These models focus on the decisions of individual actors – generally the developer – when faced with future uncertainty and volatility. The aggregation of those individual decisions and the strategies that developers adopt produce a model for how markets behave which can be mapped onto observed empirical outcomes. Again, necessarily, such models take the overall economic and political environment as a given but provide a much more nuanced and complex explanatory framework for developer behaviour. The essential element of a real option is that it is irreversible. A developer owns a site in a city; s/he can develop (or redevelop) now – but s/he has the option to wait. The developer faces uncertainty – as to the length of the construction period and the costs of construction but, most critically, about the rental and capital values that will be achievable on completion and on the length of time taken to ﬁnd a buyer. That uncertainty gives the option to wait a positive value. Once the decision to develop is taken, it cannot be reversed.12 For an individual (monopolistic) ﬁrm, this means that the present value of the completed development must exceed the costs by a factor equivalent to the option value of waiting. For a competitive market with free entry, uncertainty and irreversibility increases the costs of investing relative to not investing so, once again, developers demand higher current price to long-run average cost ratios to justify development, and must account for competitor behaviour. A number of implications follow from this. The ﬁrst is that development starts are likely to be bunched – developers do not develop until the threshold prices that overcome the cost of building and the value of waiting are reached, but then rush to develop in the face of competition and new entrants. Second, adjustment processes are likely to be asymmetric – for example, negative demand shocks will decrease rents and prices (since there is no supply side adjustment), but positive demand shocks shift the supply curve, albeit with a lag.13 Third, these tendencies will be more pronounced in volatile markets. Sivitanidou and Sivitanides (2000) provide some empirical support for these ideas for US ofﬁce markets, although

12

13

In practice, since development is a lengthy process split into stages, a developer has a series of options. In recessions, a development may be mothballed, with construction costs saved, awaiting a more favourable economic environment. The developer may have used debt, has maintenance and security expenditure and faces property taxes, and has to remain solvent while awaiting an upturn, so this is no costless option. In the longer term, stock can be withdrawn, through change of use, abandonment or demolition.

146

Towers of Capital

they note that economic fundamentals dominate the option effects in the majority of cities. Perhaps the most complete formulation of a game theoretic approach is set out in Grenadier’s (1996) Journal of Finance article, The Exercise of Strategic Options. Grenadier sets out to explain why development appears to be clustered and why developers appear to develop into a recession, in the face of declining occupational demand and building values. His model seeks to understand developer behaviour when faced with competition from other developers and lags in delivering space to the market given the time taken to construct. He cites data from Wheaton (1987) that suggest that average time to build for ofﬁce development is around 2.5 years, with large ofﬁces in complex, regulated markets often taking ﬁve or more years. His model extends and augments Williams’s (1993) real option model, allowing for construction lags and for the lumpiness of development and his own earlier modelling work (Grenadier, 1995). The modelling framework is drawn from the ﬁnance literature on the strategic exercise of warrants and convertible securities. A developer may own an existing building that is generating rent; s/he has the option to redevelop the building incurring costs and loss of rental income but generating potentially higher rents on completion. Other developers are in a similar position. Developers can observe signals about occupational demand changes, rents, prices and costs. The action of one developer affects the returns of the other. If one commences development and a second waits, then, on completion, the leader captures the higher rents, the follower suffering a loss of rental income or tenants. However, the leader has lost rental income and incurred costs during the development phase. What strategy should the developers follow? Grenadier shows that this depends on the starting conditions in the market, on the volatility of demand and on the time to construct. Development cascades – with developers rushing to develop simultaneously – occur in particular where the volatility of demand is high. The volatility increases the option value of waiting – but once the conditions favour development, all developers race to build. Developers also are prone to build ‘defensively’ when demand signals falter, for fear of being shut out of the market if competitors do build. This helps to explain continued building after rental growth peaks have been reached. Grenadier argues that the more volatile the demand for space and the longer the lag between starts and completions, the more a market is prone to development cascades in falling markets. Finally, in markets with free competitive entry of developers, there is less risk of overbuilding, with real value staying closer to replacement cost. Grenadier’s model is theoretical in nature but nonetheless captures many of the features observed in commercial real estate development

The Supply of Space

147

in developed capitalist economies. In particular, it helps to identify the types of market which will be most prone to development cascades and ‘overbuilding’ in response to demand signals. Since the model involves exercise of real options, volatility is a key variable – here, volatility of occupational demand over the long-run trend growth. Grenadier suggests that demand volatility will be most pronounced in specialised urban markets, where the fundamental demand drivers are common to a high proportion of occupiers. By contrast, diversiﬁed markets will be less prone to volatility-induced development cascades, since differences in business cycles, product life cycles and underlying economic fundamentals will smooth out demand shocks. He cites Houston and Denver as examples of markets with an undiversiﬁed economic structure: in both cities over 50% of ofﬁce completions between 1960 and 1990 occurred in the boom of 1982–1985. In addition, a defensive cascade (building into a downturn) is more likely to occur in a market where the lag between construction start and completion is long and when there are constraints to market entry. From this, it follows that the central business district ofﬁce markets of most established IFCs are likely to be prone to development cascades, to periods of intense building amid more general low activity levels. First, there is strong specialisation of activity in the core market. Ofﬁce space is occupied by global ﬁnancial ﬁrms and professional business service ﬁrms linked to ﬁnancial activity. Demand for space and ability to pay high rents are, hence, locked into the performance of global ﬁnancial markets which are both volatile and prone to shocks. Second, the complexities of building in a heavily developed market – problems of site assembly, pre-existing transport and utilities infrastructure, planning issues – are layered onto the demand for large, technologically sophisticated buildings to create long lags between project initiation and completion of development.14 Finally, high land values, linked to the size of buildings and cost of construction, allied with the need to access debt and equity ﬁnance create signiﬁcant entry barriers for new and smaller ﬁrms. Grenadier’s model, then, points to the likelihood of pronounced development cycles in IFCs – with no need for any assumption of irrational behaviour on the part of developers. That is not to say that developers

14

Wang and Zhou (2000) suggest that markets with a large existing inventory should be less prone to overbuilding shocks (as new supply may be a smaller percentage of current stock). However, their model (a two-stage model, with ﬁrst an individual development decision and, second, a collusive rent ﬁxing game that can preserve high vacancy rates) again predicts that large ofﬁces will be more prone to cyclical overbuilding, even with rational strategic decisions by developers at stage one.

148

Towers of Capital

do not behave irrationally,15 nor that the market system produces optimal long-run outcomes: just that irrationality is not a necessary condition for the existence of development cycles and that individual strategic behaviour, in aggregate, may result in cycles and building booms. Furthermore, as we will see below, with developer, investor and occupier behaviour intertwined in ﬁnancial centre ofﬁce markets, potential risk and volatility may become magniﬁed. Before exploring these interactions, though, we need to consider capital ﬂows and investor behaviour.

15

Or some developers may behave irrationally some of the time!

6 Investment, Capital Flows and the Ofﬁce Market

6.1

Introduction

In this chapter, the forces that drive and shape real estate capital ﬂows are analysed. Investors – speciﬁcally, informed professional and institutional investors – structure their investment portfolios in an attempt to ﬁnd an optimal mix of assets; a mix that minimises risk for a given level of return. Formal portfolio optimisation models increasingly drive professional investment strategies. The models were developed for ﬁnancial assets, for bonds and equities, but now include real estate and other alternative assets and have been extended beyond national boundaries. These models, run uncritically, typically suggest that real estate should have a high weighting in mixed-asset portfolios, far higher than is observed in institutional portfolios. In part, this results from data inadequacies, in part from the problematic investment characteristics of commercial property – in particular, its illiquidity, high transaction costs and ‘lumpiness’. As new property investment vehicles have appeared that appear to solve some of those difﬁculties, institutional investment in real estate has increased, both within countries and internationally. Flows into real estate rose markedly from the late 1990s through to the late 2000s. Capital ﬂows resulting from portfolio investment inﬂuence capital values and, hence, the feasibility of investment. In theory, the capital value of an ofﬁce building should simply represent the discounted value of the estimated income stream in the future. The discount rate applied reﬂects underlying interest rates, the perceived risk of the asset class (which, in turn, reﬂects volatility, liquidity, depreciation and management costs and overall sensitivity to economic shocks), along with consideration of anticipated long-run income growth. These property-speciﬁc elements of the discount rate will shift as a market undergoes structural change: an obvious example would

150

Towers of Capital

be the fall in required returns for Eastern European ofﬁces as property rights become more established and as market transparency and market information improve. For more established city ofﬁce markets, in international ﬁnancial centres (IFCs), these shifts should be more subtle and long term, associated with long-run competitiveness and market share. However, it is clear that yields and capitalisation rates (cap rates) do change markedly, in part linked to rent cycles, in part in response to capital ﬂows. These patterns are explored below.

6.2

Real estate in the portfolio

Most professional investors (and many individual investors) build portfolios that contain a mix of asset types: cash, government securities, equities, bonds and alternative investment classes, which include real estate, infrastructure, commodities and derivatives. The intuitive idea behind a mixed-asset portfolio lies in risk diversiﬁcation. The varied characteristics of the asset types and differences in the timing and pattern of cashﬂows and returns reduce period-by-period risk since, when one asset is performing poorly, another may do well to provide compensation. This idea is embedded in folklore and investment practices that go back at least to medieval times. Quantitative modelling of the risk–return beneﬁts of diversiﬁcation has developed over 50 or more years of analysis, resulting in formal strategies that shape investment ﬂows in domestic and global markets, particularly for institutional investors, for pension funds and insurance companies. In this literature, real estate is largely absent: poor market performance data preventing effective formal modelling of real estate returns alongside the ﬁnancial asset classes. More recently, improvements in property market benchmark indices have seen property analysed alongside bonds, equities and other investment products. The portfolio allocation process directs capital ﬂows; in turn, the capital that ﬂows to real estate shapes cities, in determining property prices and the feasibility or proﬁtability of development. Changes in capital ﬂows, shifts in portfolio allocations, contribute to the apparently inherent cyclical behaviour in real estate markets.

Basic principles Theoretical models of investment strategy from ﬁnance that start from asset risk and return are predicated on rationality and market efﬁciency. Investors assemble portfolios of assets that maximise their utility. Those portfolios maximise investor return for a given level of risk (or, equivalently, minimise risk for a given level of return): they are efﬁcient portfolios in terms of risk-adjusted returns. Investors demand higher returns to compensate them

Investment, Capital Flows and the Ofﬁce Market

151

for risks and uncertainties in future cashﬂows. Building a portfolio of less than perfectly positively correlated assets results in risk diversiﬁcation. Fully efﬁcient portfolios eliminate all speciﬁc risk (that is, risk that is unique to a particular asset) leaving only systematic risk. Since a rational investor can, and therefore should, diversify away speciﬁc risk, then, in an efﬁcient market, there is no reason for that investor to receive a reward for speciﬁc risk. This implies that an asset’s returns (or a portfolio’s returns) will relate to only that asset’s sensitivity to systematic, market risk factors. The selection of assets in a portfolio will reﬂect the investor’s attitude towards risk, with highly risk-averse investors holding defensive, low volatility portfolios and more risk-tolerant investors seeking higher returns and accepting the volatility implied. These fundamental portfolio ideas – from Markowitz’s modern portfolio theory, from the Capital Asset Pricing Model and its variants, and from Arbitrage Pricing Theory – rest on a set of exacting assumptions about market structure and investor behaviour. The validity of many of the assumptions is questionable in general and particularly problematic in real estate markets. Assets are not all marketable; there are barriers to investment, investors face capital constraints, information is not freely available to all investors; there are transaction costs, and those costs differ by asset class. Furthermore, insights from behavioural ﬁnance cast doubt on assumptions about investor rationality and about the nature of investor utility functions. Nonetheless, ideas from portfolio theory have been very inﬂuential in the market and help shape formal investment strategies for institutional and professional investors. In large measure, professional investors can be thought of as behaving as if the principles of portfolio theory hold. There are clearly other inﬂuences on portfolio strategy, and strategies will vary by type of investor or fund. Funds may target a particular asset class, sector and/or geographical area or be ﬁnancially engineered to generate high expected returns (with commensurate risk): such funds may form part of an investment portfolio or be directed to return-seeking individual investors. Investors may be particularly concerned with being able to meet outgoings and hence use liabilities matching techniques to seek investment portfolios whose returns rise and fall in line with shifts in liabilities. The need to meet periodic liabilities also brings a focus on liquidity. Individuals may be seeking a safe haven for capital and be more concerned with political and economic stability than with optimal risk-adjusted returns. Corporate governance, social responsibility and ethical investment strategies may lead to a ﬁltering and exclusion of certain types of assets. In aggregate, however, it is not unreasonable to assume that investors broadly follow a mean– variance driven investment strategy. In a mean–variance strategy, the risk–return characteristics of a portfolio are a function of the return and risk of the individual assets and the covariance

152

Towers of Capital

between the assets (or the covariance of the assets and a benchmark market portfolio or set of risk factors). Pairs of assets that have low correlations will result in diversiﬁcation of risk and improvements in risk-adjusted return. Theoretically, each asset that could be included – each individual share, bond, building, derivative contract – should be considered individually. For practical purposes, though, top–down investment strategies are commonly pursued, where initial fund allocation is organised at the asset class level, then a further allocation exercise is pursued within each asset class. Hence the investment allocation committee of a large life insurance fund might decide to allocate 50% of its investment funds to equities; within the equity area, it will decide on a mix of industrial sectors and between growth and value stocks; and then, within each of those sectors, pursue stock selection. How does real estate ﬁt into this process? And how do ofﬁce returns in IFCs ﬁt within the real estate asset class?

Real estate as an asset class Direct investment in private real estate (that is, actual ownership of buildings for their rental income stream and capital appreciation) presents problems that distinguish the asset class from other ﬁnancial assets. Real estate is lumpy and expensive; problems in splitting ownership and rights to cashﬂow into smaller units create a measure of indivisibility that makes it difﬁcult to assemble fully diversiﬁed portfolios of commercial real estate. The value of the average building in the US National Council of Real Estate Fiduciaries (NCREIF) database as in December 2007 was $54 million. To assemble a portfolio of 30 typical properties would thus cost some $1.6 billion: well beyond the capital allocations of all but the largest investment funds. By implication, this means that most funds are not fully diversiﬁed, their investors exposed to speciﬁc asset level risk. Furthermore, private real estate is illiquid, often taking many months to buy or sell, hampering optimal portfolio rebalancing and adding risk and uncertainty (Bond et al., 2007); allied to this, there are high transaction costs which deter trading and force longer holding periods (Collett et al., 2003). The consequent thin trading, allied to the lack of a central market place for transactions, increases search costs and leads to information asymmetry and uncertainty as to underlying real estate market performance. For all these disadvantages, research at asset class level suggests that there are considerable beneﬁts to be gained by including real estate alongside other asset types in a mixed-asset portfolio. The beneﬁts appear to result from a combination of favourable risk-adjusted return and a low correlation with the main investment asset classes, equities and bonds. Early analyses based on an uncritical application of Markowitz portfolio optimiser

Investment, Capital Flows and the Ofﬁce Market

153

techniques produced infeasibly high allocations to real estate (which were considerably out of line with actual institutional holdings in real estate). However, more sophisticated approaches correcting for data inadequacies, illiquidity and other factors still produce high weightings, albeit ones that are more appropriate to the overall size of the real estate investment market. Much of this research is reviewed in Hoesli and Lizieri’s (2007) report for the Norwegian Government Pension Fund – Global. This massive investment fund made the decision to allocate up to 10% of its capital to real estate investment in 2008, having been previously totally invested in equities and bonds. As an example of the investment characteristics of real estate, Hoesli and Lizieri produced an analysis examining the inﬂation-adjusted investment performance of US, UK and Australian real estate compared to bonds and equities. The techniques used are standard: the results for the United States and the United Kingdom are discussed here. One important qualiﬁcation is necessary in analysing index number series of directly owned, private, commercial real estate. The indices have characteristics that distinguish them from measures of the other, ﬁnancial asset classes. First, they are ungeared asset returns, while the returns for equity markets and listed property are inﬂuenced by leverage. Second, they only represent a sub-set of investment quality real estate in the respective countries. Third, and perhaps most signiﬁcant, the returns are based on appraisals of the real estate in each database, and not on actual transactions prices, unlike, for example, equity indices which are compiled based on closing prices observed in the market place. The appraisal basis of property indices has important implications for the interpretation of performance statistics. In particular, appraisal-based indices are assumed to be ‘smoothed’ – both due to temporal aggregation effects (the appraisals are spread around the reporting date) and due to appraiser behaviour in updating prior information. This creates a moving average process that reduces the reported risk measures, creates lags in responsiveness to information shocks and may distort correlation with other asset classes. Evidence for this can be seen in the presence of serial correlation in the returns series (apparently violating market efﬁciency assumptions), particularly for higher frequency data. Not all professional investors accept the concept of smoothing. From an investment manager perspective, reported fund returns are on a valuation basis and that provides input for performance measurement and risk appraisal: hence desmoothing is often viewed as an arcane academic exercise rather than a valuable business tool. Increasingly, though, professional investors are seeking less distorted measures of underlying real estate market performance. US real estate analyses – which are typically based on NCREIF quarterly returns – are further distorted by a ‘stale appraisal’ problem: many buildings

154

Towers of Capital

are appraised only annually but are left in the quarterly analysis even where no appraisal has occurred. This is not a problem for the UK and other indices that conform to the methods established by Investment Property Databank, which only consider buildings that have been valued in the relevant time period. Nonetheless, there is likely to be an information effect where many more buildings are valued at year end or at the end of the relevant tax year. To counter the effects of smoothing, Hoesli and Lizieri desmoothed the real estate series using standard desmoothing techniques. These assume that reported valuations are a blend of the ‘true’ market price and the previous valuation; by removing serial correlation from the time series, the underlying ‘true’ return series is revealed – although results are dependent on the choice of the parameter used to desmooth the series.1 They report results for both the smoothed and desmoothed series. For the United States, they also analysed the MIT Transactions Based Index (TBI), an index estimated using a repeat sales method from the NCREIF database. Although there are some questions about the reliability of the methods used to correct for changes in the quality of buildings between transactions (e.g. as a result of capital improvements), the TBI provides valuable evidence about underlying risk in the private real estate market. For the United States, Hoesli and Lizieri compared the NCREIF2 and MIT Transactions Based indices to the National Association of Real Estate Investment Trusts (NAREIT) index of REIT returns, the MCSI USA index as a proxy for stock market returns and a benchmark index of US Government’s 10-year Treasury bonds, all with consumer price inﬂation removed. Figure 6.1 shows real quarterly return indices for selected US asset classes between 1985 and 2006. As can readily be seen, stocks have produced much higher returns than the other asset classes and sectors, despite the post-2000 correction, while the listed REIT sector produces stronger performance from the late 1990s. The three indices of private real estate markets track each other closely – although the MIT TBI and the desmoothed NCREIF series exhibit much greater periodic volatility, as expected. The mean quarterly real return for equities over the 1985–2005 period was 2.6% (an annualised real return of 10.9%). REITs produced an annualised 1

2

Let Pt be the true price and Vt be the valuation at time t. The reported valuation Vt = αVt−1 + (1−α)Pt where α is the desmoothing parameter. By rearrangement, Pt = [Vt − αVt−1]/(1−α). The value of α can be set to reduce the serial correlation of the valuation series to zero or to some other ‘acceptable’ value, or to maximise correlation to some other instrumental variable. The higher the α, the greater the degree of desmoothing and the more the impact on the volatility of the recovered series. The NCREIF index consists of properties acquired by tax-exempt institutions. At the end of 2006, the database consisted of 5333 buildings with an appraised value of $247 billion.

Investment, Capital Flows and the Ofﬁce Market

155

1000 900 800 700

MIT TBI Equities Bonds NAREIT(R)

600 500 400 300 200 100

19 8 19 4 Q 8 4 19 5 Q 8 3 19 6 Q 8 2 19 7 Q 8 1 19 7 Q 8 4 19 8 Q 8 3 19 9 Q 9 2 19 0 Q 9 1 19 0 Q 9 4 19 1 Q 9 3 19 2 Q 9 2 19 3 Q 9 1 19 3 Q 9 4 19 4 Q 9 3 19 5 Q 9 2 19 6 Q 9 1 19 6 Q 9 4 19 7 Q 9 3 19 8 Q 9 2 19 9 Q 9 1 20 9 Q 0 4 20 0 Q 0 3 20 1 Q 0 2 20 2 Q 0 1 20 2 Q 0 4 20 3 Q 0 3 20 4 Q 0 2 20 5 Q 0 1 20 5 Q 06 4 Q 3

0

Figure 6.1 US real asset return indices, 1985–2006. Source: Adapted from Hoesli and Lizieri (2007). Table 6.1 Descriptive statistics, quarterly real returns, US Markets, 1985–2005.

Arithmetic mean return (%) Compound growth (%) Standard deviation (%) Skewness Serial correlation

Stocks

Bonds

REITs

NCREIF

2.63 2.29 8.17 −0.5602 −0.034

1.33 1.25 4.12 0.3737 −0.2022

1.98 1.74 6.97 0.2107 0.4011

1.23 1.22 1.71 −1.278 0.705

Desmoothed

TBI

1.39 1.29 4.36 −1.579 0.359

1.54 1.47 3.80 0.445 0.066

Source: Adapted from Hoesli and Lizieri (2007).

return of around 8%, direct real estate as measured by NCREIF, 5%, slightly lower than bonds over the period. Given the expected risk–return trade-off, it would be expected that a ranking of the standard deviation of returns should match that of the returns themselves. Stocks do have the highest standard deviation (an annualised 16%), followed by REITs (14%) and Bonds (8%). However, the annualised reported risk for the NCREIF index, at just 3.4%, seems far too low. Evidence that the series is smoothed can be seen in the ﬁrst-order serial correlation coefﬁcient – which implies that almost half of the variation in return in any one quarter is explained by the return in the previous period. Both the desmoothed series and the MIT TBI show levels of risk that are more than double that of the NCREIF index, ﬁgures which seem closer to intuition of the risk of commercial real estate as an asset class (Table 6.1).

156

Towers of Capital

Table 6.2 Descriptive statistics, quarterly real returns, UK markets, 1988–2005.

Arithmetic mean (%) Compound growth (%) Standard deviation (%) Skewness Serial correlation

Stocks

Bonds

Property companies

IPD

Desmoothed

2.01 1.71 7.73 −0.564 −0.091

1.39 1.35 2.70 −0.124 0.430

1.90 1.40 10.07 −0.346 0.161

1.78 1.75 2.45 −0.468 0.816

1.79 1.70 4.37 0.365 0.418

Source: Adapted from Hoesli and Lizieri (2007).

For the United Kingdom, Hoesli and Lizieri estimated quarterly returns for the UK direct market from the Investment Property Databank (IPD) monthly index. The monthly index – a sub-set of the main IPD database – does not perfectly proxy the IPD annual index, as there are some compositional differences, but provides the best sub-annual indicator of the performance of UK investment-quality real estate.3 Inﬂation-corrected IPD returns were compared to the Financial Times (FT) All Share index, the FT Property Company index (which measures the performance of the approximately 40 ﬁrms classiﬁed as real estate investment holding or development companies on the London exchange) and an index of medium-dated government bonds from the WM Company for 1987–2006 (Table 6.2). Real return indices are plotted in Figure 6.2. The stocks series shows the effect of the dot.com boom and bust. In the boom period, property company returns lagged (prompting a number of prominent property ﬁrms to be taken private in the late 1990s and early 2000s), to subsequently recover sharply (aided by the declared intention to introduce a UK REIT, implemented in January 2007). The recent real estate market boom contributed to the strong relative performance of private real estate. In annualised terms, stocks produce a return of 8.3%, outperforming property company shares (7.8%), IPD (7.3%) and bonds (5.7%) – although compound growth is higher for IPD given the absence of the extreme negative returns experienced in the stock market. As with the US data, the reported risk of private real estate as measured by IPD appears to be too low. Returns are between those of stocks and bonds (which is consistent with intuition of the blended bond and equity characteristics of real estate returns), but the standard deviation is below that of bonds. Desmoothing the series produces a standard deviation that is more plausible in risk–return 3

At December 2006, the IPD monthly database contained 3820 buildings with a capital value of £50.5 billion. The IPD database as a whole covered 12 137 buildings with a capital value of £192 billion – some 49% of the real estate assets of institutional investors and listed property companies.

Investment, Capital Flows and the Ofﬁce Market

157

400 350

Equities Bonds

300

Property companies IPD desmoothed

250 200 150 100 50

D

D

ec em ec be e r D mb 19 ec e 8 e r 7 D mb 198 ec e em r 1 8 b 98 D ec er 9 em 19 9 D ec ber 0 em 19 D ec be 91 e r D mb 199 ec e e r1 2 D mb 99 ec er 3 e 1 D mb 994 ec er e 1 D mb 995 ec er e 1 D mb 99 ec e 6 em r 1 99 D b ec e 7 e r D mb 199 ec e 8 e r1 D mb 99 ec er 9 em 2 b 00 D ec er 0 2 e D mb 001 ec er e 2 D mb 00 ec er 2 e 2 D mb 00 ec e 3 r em 2 be 004 r2 00 5

0

Figure 6.2 UK real asset returns, 1988–2005. Source: Adapted from Hoesli and Lizieri (2007).

terms, although the desmoothing process reduced rather than eliminated ﬁrst-order serial correlation in the return series. After correcting for smoothing, then, real estate in the United States and the United Kingdom appears to exhibit the expected blend of equity and bond return characteristics. Property has bond-like characteristics in that the rental income provides a high and relatively stable cashﬂow; the possibility of rental value and capital growth produce equity-like characteristics. Alone, this would not justify a place in the mixed-asset portfolio: investors can replicate the cashﬂow/growth characteristics of real estate using blended paper investments in bonds and equities. It is the combination of risk–return and correlation that lead to high theoretical allocations. Tables 6.3 and 6.4 show contemporaneous quarterly correlations for US and UK asset markets, again from the Hoesli and Lizieri (2007) study and illustrate the theoretical diversiﬁcation gains to be made. For both US and UK markets, the smoothed, valuation-based returns show almost zero correlation with equity and bonds. Inclusion of real estate alongside the ﬁnancial assets should result in diversiﬁcation beneﬁts and improved risk-adjusted returns. Correlations rise slightly when the valuation smoothing issues are addressed, but the apparent diversiﬁcation

158

Towers of Capital

Table 6.3 Contemporaneous correlation, US markets, 1985–2005.

Stocks Bonds REITs NCREIF Desmoothed TBI

Stocks

Bonds

REITs

NCREIF

Desmoothed

TBI

1.000 −0.054 0.428 0.042 −0.016 0.184

1.000 0.158 −0.017 −0.097 0.010

1.000 0.041 0.136 0.181

1.000 0.767 0.488

1.000 0.479

1.000

Source: Adapted from Hoesli and Lizieri (2007).

Table 6.4 Contemporaneous correlation, UK markets, 1988–2005.

Stocks Bonds Property companies IPD Desmoothed

Stocks

Bonds

Property companies

IPD

Desmoothed

1.000 0.113 0.641 0.116 0.159

1.000 0.119 −0.014 0.209

1.000 0.277 0.391

1.000 0.772

1.000

Source: Adapted from Hoesli and Lizieri (2007).

beneﬁts remain. One result worth noting is the low correlation between private real estate returns and REITs or property companies. The public market real estate vehicles appear to have a higher correlation with the stock market than with the underlying property market: in part this reﬂects the impact of the structure of the investment vehicle, in part the tendency of returns in the listed real estate sector to lead those in the private market, with equity market prices seeming to anticipate shifts in values in the underlying market. Thus, in the 2007 UK market, REIT share prices fell and discounts to Net Asset Value widened considerably in the ﬁrst half of the year, well before observed falls in capital values in the private market. Using the risk–return results in a mean–variance optimiser framework, it is possible to obtain a theoretical allocation for real estate, equity and bonds. Desmoothing the real estate data and applying an ‘illiquidity premium’ to reﬂect the difﬁculties of buying and selling buildings, Hoesli and Lizieri demonstrate signiﬁcant weightings to real estate. Their results are paralleled in other studies. Several US authors have suggested an optimal weight for real estate in mixed-asset portfolios in the 10–20% range (Fogler, 1984; Ennis & Burik, 1991; Ziobrowski & Ziobrowski, 1997). Similar weights as those for the United States are reported in studies of non-US real estate (see, for example, Brown and Schuck (1996) for the United Kingdom and Hoesli and Hamelink (1997) for Switzerland). A study by Hoesli et al. (2004) provides evidence for seven countries on three continents, with maximum

Investment, Capital Flows and the Ofﬁce Market

159

real estate weightings on the efﬁcient frontier ranging from around 7% in Switzerland and the Netherlands, to 15% or more in Australia, France and the United States. These suggested allocation ranges are consistent with estimates of the total size of the real estate market relative to bonds and equities: although there is little consensus as to the value of investment real estate, most estimates fall between 7% and 17% of the total investment universe (see, e.g. Sirmans & Worzala, 2003; Hoesli & Lizieri, 2007). In principle, therefore, these results suggest that investors should hold substantial portfolios of real estate assets – although it should be stressed that these studies use ex post, historic data, while investment strategy should be based on expectations. Evidence on actual allocations is harder to obtain: it seems that most institutional investors hold less real estate than the portfolio models suggest. Data from Sirmans and Worzala (2003) point to a real estate weighting of a little less than 6% in institutional investor mixed-asset portfolios. PREA (2006) provide comparative ﬁgures for pension funds in a sample of countries: real estate holdings range from 3% in the United States to 12% in Germany, with an average of just over 8%. Where more detailed information is provided, it seems that smaller funds often hold no real estate at all; this may relate to the problems relating to achieving asset class diversiﬁcation noted above. Regulatory constraints may also prevent or limit funds investing in real estate. For example, in the United Kingdom, the pension fund minimum funding requirement referenced fund assets and liabilities to a benchmark that was based entirely on equities and bonds. This created matching risks for funds with portfolios that contained substantial amounts of property. The equity market downturn of the early 2000s, as the technology and dot.com growth ‘bubble’ ended, however, produced a resurgence of institutional interest in real estate as a haven for capital and as a potential diversiﬁer.

Sectors, regions and buildings Given an allocation to real estate as an asset class, investors must structure their property portfolios, both to guide stock selection and as a framework for measuring the performance of fund managers. Conventionally, this has involved sub-dividing property along two dimensions: geography and property type or sector. Sectors include ofﬁce, retail, industrial, residential and hotel property – the subdivisions varying by country to reﬂect the institutional structure of the local market. Given such a breakdown, investors may follow a variety of strategies. They may focus on a particular sector and/or area to create a specialised fund; they might seek a balanced investment strategy with property type and geography weightings matched to the size of those segments in the wider market (or, more usually, in the market benchmark); or they might create a diversiﬁed portfolio but tilted towards one or

160

Towers of Capital

more of the sectors or areas. Specialist funds beneﬁt from a concentration of management expertise (and internal economies of scale in management costs) but expose stakeholders to speciﬁc risk. Tilting a fund may indicate that the fund’s managers believe that they can identify segments of the market that will give superior (or inferior) risk-adjusted returns, or that there are other considerations – for example, particular concerns about illiquidity. Researchers have sought to identify the ‘best’ way to diversify real estate portfolios, while typically retaining the ‘sector–region’ structure (for reviews, see Hamelink et al., 2000; Nelson & Nelson, 2003; Devaney & Lizieri, 2005). US research traditionally stressed geographical boundaries with discussions concerning whether regional boundaries should be administrative or based on economic activity (the latter generally shown to be superior). More recently, however, it has been argued that ‘the free ﬂow of institutional real estate capital across the US over the past 30 years tends to homogenize transient differences in investment characteristics across geographical regions for properties of the same type’ (Young, 2008, p. 235). UK studies indicate a stronger role for property sector (with retail real estate distinct from ofﬁce and industrial) and a broad geographical factor that identiﬁes London in general, and City of London ofﬁces in particular, as being distinct (Hoesli et al. 1997; Hamelink et al. 2000). Where a sector–region segment exhibits distinct risk–return behaviour, by implication there are different returngenerating factors in operation. If City of London ofﬁces behave differently from other commercial real estate segments in the United Kingdom, is this because supply and demand are driven by international ﬁnancial markets rather than the domestic economy? Recent UK research at individual property level has cast doubt on the effectiveness of standard sector–region classiﬁcations as a basis for risk diversiﬁcation of the property portfolio. Devaney and Lizieri (2005) use multivariate techniques to test whether the individual building returns behave consistently within sector–region segments. They do not: less than one-third of buildings are correctly classiﬁed using a three-by-three sector–region grouping or the ten category IPD Portfolio Analysis Service segments that are used by institutional investors for performance measurement and attribution analysis. They suggest that yield (capitalisation rate) and tenancy structure may prove more sensitive in classifying risk–return behaviour. Callender et al. (2007) also analyse data at individual property level. If sector–region segments are valuable as ways of diversifying risk, then properties within segments should behave in similar fashion, while those properties should behave differently to properties in other segments. This does not seem to be the case. Correlations between property returns and aggregated segment returns are low – rarely over 0.4. Ofﬁce buildings have particularly low within-segment correlations. City of London ofﬁces have

Investment, Capital Flows and the Ofﬁce Market

161

an average correlation coefﬁcient of just 0.11. To some extent, this low correlation must reﬂect differences in the time buildings have been let: major differences in income return can result from the letting date, given the UK’s standard upward only rent review clause, that prevents sitting tenants beneﬁting from falls in market rent, while leaving them liable for increases. They calculate that an investor would need a portfolio of at least 40 City ofﬁces (with a value of around £887 million at 2005 prices) to achieve 90% diversiﬁcation of speciﬁc risk.4

6.3

Investment patterns and investor behaviour

That there exists a body of research seeking optimum portfolio strategies for real estate does not mean that investors follow such strategies. Actual patterns of investment may be driven by non-economic factors. The growth of quantitative models of real estate market behaviour, particularly in Englishspeaking markets, might suggest a rational process – at least for professional investors. Nonetheless, there is anecdotal evidence that in practice, the patterns of real estate investment result from intuitive decisions which may be based on sentiment as much as formal modelling processes. Beliefs about risk, return and liquidity that are not empirically founded shape portfolios. It has been suggested that a preference for investment in prime ofﬁce space in the central business districts (CBDs) of major cities is an example of an ‘irrational’ investment strategy. For the United Kingdom, Henneberry has argued that a London and South of England bias in investment exists that has led to over-development in London and underdevelopment in the peripheral regions of the United Kingdom (Henneberry, 1999; Henneberry & Rowley, 2000). This argument – which is part of a wider ‘London-bias’ body of literature, comes from an institutional stance. As set out in Guy and Henneberry (2002), investment decisions are in part driven by macro-economic forces and rational processes, but they are also a result of actual decisions made by actors, agents. UK real estate investment, development and funding decisions are dominated by ﬁnancial services, property companies and advisory ﬁrms based predominantly in London. That spatial prism distorts decision-making processes and can generate herd behaviour. As evidence, timing differences between rental cycles (which are seen as a function of occupational demand) and yield cycles (which are seen as an outcome of investment behaviour) are cited. There are lags in rental cycles but yield shifts transmit quickly from London, which, it is argued, stiﬂes development away from London. Various

4

Deﬁned as the R2 between the portfolio and the segment return.

162

Towers of Capital

service sector activity variables are compared to development activity to demonstrate regional bias. The argument is not wholly convincing and is strongly criticised by Ball (2002). If there were a development shortfall in peripheral regions, then this should be reﬂected in upward rental value pressure, with consequent impacts on capital values. Rising capital values with static construction costs should permit development, as outlined in Chapter 5. Further – as Henneberry and Rowley concede – there may be strategic advantages in London investment. The existence of large, high-value buildings permits concentration on a small number of assets, reducing management costs and creating scale economies, while the international nature of the London ofﬁce market may create greater liquidity. Nonetheless, the return performance of the London ofﬁce market does lead to questions about rational investment and pricing. Between 1980 and 2006, the City of London ofﬁces on the Investment Property Databank recorded an average annual return of 10.1% – the second lowest return of all of IPD’s portfolio analysis segments after South Eastern ofﬁces. Yet the standard deviation of returns was the second highest, at 11.9% – with only West End of London ofﬁces showing greater volatility. The three worst performing segments in terms of return per unit of risk were, in order, City ofﬁces, West End ofﬁces and South East ofﬁces. Over that period, in real terms, both rental values and capital values fell. For all this disappointing performance, IPD record positive net investment of £3982 million in City ofﬁces. While this might seem to support arguments of a London-bias, that net investment has to be set in the context of a signiﬁcant inﬂow of capital to real estate. City ofﬁces’ share of the overall capital value of the portfolio fell sharply from a peak of 18% in 1987 to around 8% in 2006, a fall that cannot be explained solely by the collapse in capital values in the early 1990s or the rise of new investment categories such as retail warehousing. Based on IPD results, then, UK institutional investors do seem to have reacted to the relative under-performance of City ofﬁces by reducing their weight in the portfolio. What is perhaps harder to explain is the persistent low capitalisation rates for those ofﬁces despite the growing evidence of low or negative real value growth. Unless there are hard-to-measure but tangible investment beneﬁts of City investment – of which, perhaps the most likely candidate is greater liquidity as a result of the global character of the market as a by-product of London’s status as an IFC – the persistent low yields and high values per square metre might appear anomalous. Investors may consider that London has better long-run economic growth prospects than other UK regions: but this assumes that such growth is translated into land and property values: the historic evidence offers at best limited support for such an assumption.

Investment, Capital Flows and the Ofﬁce Market

6.4

163

Real estate vehicles and the portfolio

Much of the research reported above focuses upon direct investment in private real estate – the ownership of buildings for their rental cashﬂow and capital appreciation. For many types of smaller investors, this is not a realistic option (other than piecemeal ownership of individual and lower capital value buildings). There do exist investment vehicles that provide investors with exposure to the real estate market while overcoming some of the obstacles and constraints faced in direct investment. Historically, the two most important forms have been public-listed property companies and listed or unlisted unit trust, managed fund or commingled real estate fund formats. From the mid-1990s, there has been a wave of innovation in real estate vehicles, creating many new ways to access the property market. These will be examined in more depth later. One key common feature of real estate vehicles is that the design and structure of the vehicle alters the risk and return characteristics of the investment – and also the diversiﬁcation beneﬁts that it offers. This is perhaps most clearly seen in relation to public-listed property companies. Owning shares in a public property company confers on the equity holder the right to receive dividends and the right to the residual assets of the company. Dividends are paid from rental income and the asset base of the company is its real estate so the investor is exposed to the risk– return characteristics of the real estate market(s) in which the company is active. Yet the cashﬂow returns experienced by the shareholder will be very different to those reported by private property market performance indices. There are many reasons for this: I will focus on four. First, the share price is subject to equity market volatility. Tables 6.1 and 6.2 show that REIT and property company volatility was far greater than underlying real estate market volatility in the United States and the United Kingdom, even once the private market indices are corrected for smoothing. Second, the shareholder’s exposure to capital market volatility is greater to the extent that the property company’s managers may retain and reinvest proﬁts rather than distributing income, reducing the bond-like characteristics of the real estate market. Third, equity markets anticipate changes in the underlying market and react to information shocks. Hence, the shares of a property company or REIT may trade at a discount or at a premium to the Net Asset Value of the company (crudely the value of the properties owned less the costs of redeeming the debt in the company) as the equity market forecasts rises or falls in property capital values. Finally, the equity returns are inﬂuenced by the capital structure of the ﬁrm. The leverage or gearing of a property company – the balance between equity and debt – alters the pattern of cashﬂow and the volatility of returns.

164

Towers of Capital

150

100

Per cent

50

0

–50 All equity

–100

50% Debt 80% Debt

–150 87

19

89

19

91

19

93

19

95

19

97 19

99

19

01

20

03

20

05

20

Figure 6.3 Delivered returns for different gearing levels. Underlying data sources: IPD, British Bankers Association.

Figure 6.3, above, shows hypothetical delivered annual returns for different levels of gearing based on UK IPD returns and the London Interbank Offer Rate. As gearing increases, returns become more volatile, with peaks and troughs exaggerated. Over the 20 years period, underlying asset returns average just less than 12% with a standard deviation of 8%. With a debt to equity ratio of one (50% loan to value ratio), average return rises to 15%, but the standard deviation increases to 18%. At an 80% loan to value ratio (by no means uncommon in opportunity funds), average returns reach 25% but the standard deviation increases to over 48% and the investor would confront negative equity in the 1990 trough. While making simplifying assumptions, the results indicate how higher expected returns are accompanied by greatly magniﬁed volatility and an increase in the risk of bankruptcy and loan default. For tax liable property companies (if not for tax-exempt REITs), the interest payments on debt act as a tax shield, reducing the Government call on proﬁts. The impact of this gearing effect in real estate varies over time as debt capital availability veers between glut and famine. In a capital famine (or credit crunch), lenders – principally banks – seek to reduce their exposure to real estate markets and refuse to lend to positive net present value projects or set onerous conditions and demand high interest rates to ration lending. In capital gluts, competition for business drives a relaxation of lending terms and can erode due diligence standards. In turn, that can contribute to

Investment, Capital Flows and the Ofﬁce Market

165

the supply cycle, with real estate construction driven by capital availability. The standard description of lending’s role in the property cycle now needs to be augmented with the development of real estate debt securitisation and the evolution of mortgage-backed securities markets. These provide banks with new sources of capital, an exit strategy and liquidity, which changes lending behaviour: they also permit disintermediation: property investors may by-pass conventional lenders and raise debt directly in the capital markets. The dynamics of this process and their implication for the stability of the ofﬁce markets of IFCs are considered in later chapters.

6.5

International real estate investment

A striking feature of the commercial real estate market of the late 1990s and the ﬁrst decade of the twentieth century has been the rise of global commercial real estate investment. From relatively modest levels, cross-border portfolio investment has become a signiﬁcant part of the real estate investment market. JLL (2006) reported that in 2005, 35% of the US$475 billion direct investment in commercial property, $164 billion – 35% – was cross-border, an increase from the 29% share of 2004. Nearly a third of those international transactions were inter-regional, with European investors and global investment funds in particular investing outside their home continent. The following year, JLL (2007) reported that 63% of the €244 billion direct investment in European real estate was cross-border – with 52% of that international investment going into ofﬁce buildings. At individual nation or city level, international ﬂows can dominate local ﬂows – and not just in emerging or transitional economies. Nappi-Choulet (2006) claims that around two-thirds of French real estate investment in the 2000–2003 period was non-domestic, with Paris the most signiﬁcant recipient of inﬂows. What has driven the creation of a global real estate investment and what are the implications for property markets?

Global diversiﬁcation Although diversiﬁcation is only one motivation for global investment strategies, the dismantling of barriers to capital ﬂows across the last quarter of the twentieth century permitted active pursuit of international diversiﬁcation as a way of generating superior risk-adjusted returns. That, in turn, shaped the new international ﬁnancial system and enhanced the role of IFCs in coordinating the investment and managing the global portfolios. Other factors promoting international investment included return-chasing strategies (investment in new markets, in the developing and transitional economies), safe-haven investing (moving capital from unstable domestic regimes

166

Towers of Capital

to locations with stability, transparency and strongly enforced property rights), lack of opportunity in the domestic market (important for the growing sovereign and natural resource based funds) and strategic reasons. The principle underlying global diversiﬁcation is that domestic assets are subject to a set of systematic risk factors – movements in the domestic market, the local business cycle, interest rate and inﬂation shocks and institutional change – which are, at least in part, localised. A domestic portfolio, fully diversiﬁed, has a sensitivity to those domestic risk factors; efﬁcient portfolios cannot generate further reduction in risk for a given level of return. Investment across borders can generate further diversiﬁcation provided that the systematic risk factors in the target country are less than perfectly correlated with those of the domestic economy. This will generally be the case; although there are global inﬂation, interest rate and growth cycles, there are sufﬁcient regional and local differences to generate lower average covariance between assets and hence greater risk reduction. These theoretical advantages, proposed by economists such as Solnik in the 1960s and 1970s, were rapidly transformed into practice once cross-border capital controls were removed – at least for paper securities, creating truly international bond, derivative and equity markets. With rapid growth in international penetration of domestic securities markets, one consequence was an increase in the correlation between markets – global shocks were transmitted rapidly around the ﬁnancial system, differences between real expected returns and interest rates between countries were reduced as capital ﬂows sought to beneﬁt from arbitrage opportunities – and hence increasingly eliminated them. As will be shown later, correlation between national equity market indices increased rapidly, leading to arguments that industrial sector was a more important dimension of diversiﬁcation than national location.

Global diversiﬁcation and real estate Real estate is essentially local in nature and, as a result, may be less correlated across countries than traded paper securities, bonds and equities. Analysis has attempted to quantify the beneﬁts of global real estate investment strategies, generally in a standard mean–variance framework. The results in general do indicate that there are beneﬁts to be gained from including global real estate in a mixed-asset portfolio and, within a specialist real estate portfolio, holding a global rather than a purely domestic mix of property assets. However, there are severe practical difﬁculties in implementing an international investment strategy that are not fully addressed in research. These concerns have not prevented the rapid growth of an international real estate investment industry, with global investment funds, international advisory services and benchmark indices.

Investment, Capital Flows and the Ofﬁce Market

167

Due to data restrictions limited evidence exists concerning the beneﬁts of international direct real estate diversiﬁcation (see Sirmans and Worzala, 2003, for a review). Analysis is limited to those countries which have robust time series data for commercial real estate returns (which are usually short time series, appraisal based and subject to smoothing) or must use hypothetical returns derived from rental value and cap rate estimates provided by real estate consultants. In general, results suggest that there are beneﬁts to be gained from a global real estate strategy, both in a mixed-asset portfolio and in a specialist real estate portfolio context. For example, Case et al. (1997) analyse returns for 21 markets and argue that international real estate diversiﬁcation within the three major sectors of commercial property would have been beneﬁcial for a US-based investor. In follow-up research (Case et al., 1999) they examine correlation structures in relation to GDP: much of the common movement of real estate returns can be linked to the global economy; national GDP explains more of the local variation but there remains an unexplained ‘local’ real estate factor. Quan and Titman (1997) ﬁnd that US real estate returns are less highly correlated with real estate returns in other countries, than is the case of US stocks with international stocks, suggesting signiﬁcant beneﬁts from international real estate portfolio strategies. Similar results are reported by Newell and Webb (1996) for the United States, Canada, the United Kingdom, Australia and New Zealand. Exchange rate movements have an impact on the risk, return and covariance patterns of international investments. A real estate investor faces exchange rate risk that is more difﬁcult – and more costly – to manage than would be the case for equity or bond investment. The risk relates to the longer holding periods for real estate (as a result of higher transaction costs) and the illiquidity of private real estate markets which makes the timing of exit uncertain. The somewhat limited evidence in the area of currency hedging suggests that currency ﬂuctuations can be – at least partially – hedged, and that swaps constitute the best hedging instrument (Worzala et al., 1997; Ziobrowski et al., 1997). When hedging costs are taken into account, however, the beneﬁts of international diversiﬁcation appear more limited. However, Hoesli et al. (2004) after allowing for the costs of hedging and for currency risk ﬁnd that for many investors, adding international real estate to a portfolio improves risk-adjusted returns (with allocations of 15–25% in real estate on the efﬁcient frontier). These results do not hold for UK and US investors, however, which may explain why other studies are more pessimistic. These studies, however, have only investigated the case of real estate in one foreign country. If a full-blown diversiﬁcation strategy is carried out, the risk reduction beneﬁts should be substantially greater (as illustrated in the studies by Case et al., 1997; Quan & Titman, 1997), and some of the currency effects will diversify away. In such a context, it can be argued that hedging

168

Towers of Capital

strategies may not be needed when several currencies are considered; in any case, the cost of hedging should be weighed against the risk it eliminates. The results in Hoesli et al. (2004) make it possible to examine the impact of including both domestic and international direct real estate in a mixedasset portfolio. Generally speaking, the allocation to domestic real estate diminishes as international real estate is added to the investment universe. The only exception to this is the United States where the allocation remains exclusively in domestic real estate. This result reﬂects the fact that in small countries the size of the real estate market will not allow for a large allocation to a diversiﬁed real estate portfolio and that investment opportunities have to be sought after in other countries. Data are more readily available for indirect real estate investments, which have observable, public market prices. The general ﬁnding is that publiclisted real estate equity indices show lower correlations across national boundaries than do general equity indices. Eichholtz (1996), for example, presents evidence that international real estate share returns are correlated less strongly than international common stock and bond returns. As a result, international indirect real estate portfolio diversiﬁcation is found to work better than for stocks and bonds. Gordon et al. (1998) construct international mixed-asset frontiers and ﬁnd both that international stock diversiﬁcation is more effective than purely domestic inter-asset diversiﬁcation and that international real estate stock diversiﬁcation is more effective than international common stock diversiﬁcation. Similar conclusions are reported by Liu and Mei (1998), Stevenson (2007) and Conover et al. (2002). Nonetheless, there is some evidence of global real estate factors and convergence of real estate returns both internationally and within regions. Global and regional factors are found by Bond et al. (2003) and Ling and Naranjo (2002); Bond et al. and Eichholtz (1997) ﬁnd strong regional patterns too. Lizieri et al. (2003) and McAllister and Lizieri (2006) examine the impact of the introduction of the Euro on the return performance of listed real estate in Europe. An increase in cross-national correlation is observed, for both Eurozone and non-Eurozone countries. There is no clear evidence that this is related to monetary integration (as opposed to broader economic integration). There is evidence that national returns are inﬂuenced both by a global and a European real estate factor. Although there is evidence of convergence, the cross-market linkages still appear to be lower than for common equities and small capitalisation stocks. These results do suggest that listed property company returns are driven by local factors, national factors, regional factors and global factors. Investors seem to gain from a global public real estate strategy, through greater diversiﬁcation and, hence, improved risk-adjusted returns. As a note of caution, though, the lower correlations might simply result from choice of an individual equity market sector rather than the whole market or it might be linked to size of

Investment, Capital Flows and the Ofﬁce Market

169

ﬁrm: small cap stocks tend to exhibit lower cross-national correlations than the equivalent national all share indices.

Barriers, issues and problems Investors seeking to diversify their equity or bond portfolio internationally face few technical obstacles. Provided that there are no constraints on crossnational investment, they can acquire shares from a country that will closely track the underlying national market index and they can choose an exposure level that ﬁts their portfolio objectives. It is also relatively straightforward (and not excessively costly) to rebalance or restructure their portfolio over time. The paper securities are publicly traded, pricing information is transparent and search and monitoring costs are low. Even ignoring exchange rate risk, the problems confronting the direct real estate investor are much more difﬁcult. Much of the research underplays the very real practical issues in front of an investor in implementing international direct investment real estate strategies: global investment strategies are much more complex than much of the body of research literature implies (Lizieri & Finlay, 1995; Baum, 2002; Sirmans & Worzala, 2003). Theoretical real estate allocation models typically utilise national indices (or hypothetical city returns) and implicitly assume that an investor will be able to track such indices. However, one key feature of real estate – its large lot size – makes it difﬁcult to diversify away speciﬁc risk, since most investors will only hold a handful of properties. Callender et al.’s (2007) study cited above argues that it would be necessary to assemble a 150 property, £2 billion portfolio of UK property to diversify away 90% of speciﬁc property risk in that country. Repeated internationally, it is evident that only the very largest investment funds – a sovereign wealth fund, perhaps – could reasonably be expected to create global property portfolios that were not exposed to high levels of building level risk or that might be expected to match the performance of the underlying international real estate markets. It is possible that lower average covariance between buildings in different countries would result in more rapid falls in speciﬁc risk, but even then, the size of portfolio required would be beyond the reach of most investors. Furthermore, the illiquidity inherent in property markets prevents investors from rebalancing portfolios. The rise of global real estate funds that permit investors to pool capital and gain exposure to diversiﬁed property portfolios has created a way around the lot size problem and (in principle) reduces liquidity constraints, although, in turn, fund investment structures create new issues and problems (explored in Chapter 7). Private real estate markets are characterised by information asymmetry and local (often informal) knowledge can be very important: a complete understanding of the operation of the land-use planning system, of political

170

Towers of Capital

interventions in the local economy, on the true nature of supply and demand processes. Non-domestic investors are at a disadvantage compared to local players and face higher search and monitoring costs. International investors may also face currency risk. To achieve satisfactory returns in non-domestic markets, they must possess an advantage over local players: superior access to capital, better analytic or forecasting tools, for example. Even here, local players can free ride on the analytic and strategic skills of the outside investors by replicating their strategies, while exploiting local knowledge advantages. As an alternative, the external investors can buy in local expertise either by working with a local joint venture partner (which raises questions of alignment and management control) or by paying for local research and consultancy (which adds further costs and erodes returns). To some extent, this suggests that outside investors are most likely to beneﬁt from a global strategy either by investing in emerging markets (for example, the transitional Eastern European economies or in Asian or Latin American markets) or by seeking out distressed markets where local players may be unable to exploit investment opportunities. In developed markets, global players face disadvantages: Eichholtz et al. (1998) ﬁnd, for example, that domestic property companies outperform internationally diversiﬁed ﬁrms in absolute and in risk-adjusted returns. Furthermore, in order to seek economies of scale and minimise search, information and monitoring costs, many investors concentrate investment in large buildings in a small range of property markets. These are typically in global cities with more mature, transparent, markets, greater liquidity and for which market research is available. Such a strategy – on the surface rational – may remove the anticipated diversiﬁcation gains due to induced correlation risk (Lizieri & Finlay, 1995). Returns in these large world cities are more likely to be driven by common global demand factors: booms and slumps may well be coincident. This will be a signiﬁcant issue in the ofﬁce markets of IFCs – as will be explored in detail in subsequent chapters.

6.6

Real estate capital ﬂows

Capital ﬂows result from the investment decisions of investors. As Chapter 3 showed, while it is often suggested that global capital ﬂows are a new phenomenon, international investment has a long tradition and has helped to create the interlocking global ﬁnancial system and the hierarchy of world cities and ﬁnancial centres. Nonetheless, the more recent phase of global capital ﬂows that emerged from the removal of barriers to both outward and inward investment initially focused either on paper securities (for portfolio building) or on foreign direct investment as part of globalisation of industrial production and distribution. With the creation of new property

Investment, Capital Flows and the Ofﬁce Market

171

investment vehicles, many of the barriers to direct real estate investment outlined above have been eroded. Prices in property investment markets are inﬂuenced both by domestic and international capital: those prices, in turn, inﬂuence development activity and help determine the quantity and quality of space in those markets. Capital ﬂows and capital availability inﬂuence asset prices both generally and within an asset class. Where there is ‘excess’ capital in the global economy seeking investment opportunities, interest rates will be forced downward with the lower discount rate increasing the value of future expected cashﬂows and raising asset prices. Capital will also ﬂow to assets that are believed to offer superior investment opportunities – higher returns or diversiﬁcation beneﬁts. Those favoured assets might be in particular geographical markets or might represent a particular class of asset. Common belief systems or herd behaviour may cause excess capital to ﬂow into an asset class or market, raising prices above what might be considered the underlying fundamental asset value. Can the strong growth in capital values in the commercial real estate markets of the developed economies in the decade up to 2007 be seen in this light? What drove the capitalisation rates, the yields of commercial property markets down? Were the movements rational or a function of herd instinct?

Capital ﬂows, required returns and yields In an equilibrium, efﬁcient market world, the return for investing in real estate would be just sufﬁcient to reward the investor for their exposure to systematic risk factors as a result of acquiring the building or property investment vehicle. With all investors discounting future cashﬂows at the appropriate target rate of return, then all property would be fairly priced. Tenants occupying space will be paying optimal rents. The relationship between the rent and the capital value of the property is the capitalisation rate or yield. kt = Rt/CVt

(6.1)

Traditional appraisal or valuation practice uses the capitalisation rate as a valuation tool for freehold, rack rented property, dividing the rent by the yield to estimate the market value of the property. Yields are observed in the market place from transactions of comparable properties, then adjusted for the subject property. This adjustment at property level is not consistent with the idea that investor return relates to systematic risk, since micro-adjustments to yield imply a required return adjusting to reward speciﬁc risk. The capitalisation rate or yield contains, implicit within it, a number of separate elements. First, the rental income produced by the building may rise (at least in nominal terms). Assuming annual rental adjustment and in

172

Towers of Capital

keeping with the Gordon dividend growth model, then, the cap rate is the required return less the anticipated nominal income growth. kt = rt − gn

(6.2)

In turn, the required return consists of the risk-free rate and a risk premium to reﬂect the systematic risk of the asset in question. kt = RFRt + RPt − gn

(6.3)

The risk-free rate itself consists of two components – a compensation for anticipated inﬂation and the real risk-free rate, the minimum return that an investor requires to compensate for investing rather than consuming capital or wealth. In turn, the nominal income growth consists of a real growth element and anticipated inﬂation. Ignoring interaction effects, these two inﬂation components will cancel. kt = RRFRt + Et(F) + RPt – (gr + Et(F))

(6.4)

kt = RRFRt + RPt – gr

(6.5)

Finally, since property is a real asset, investors must be concerned about depreciation (decline in rental income due to obsolescence, physical and functional), the required capital input to check depreciation and maintain the income-producing qualities of the building, and the proportion of rental income lost in management, operating and collection costs. These may be embedded in the income (that is, Rt is the net operating income, rent less collection costs), the risk premium or the real growth components, but, setting them out explicitly: kt = RRFRt + RPt – gr + δ + MOC

(6.6)

Given this theoretical structure of yields, what causes yields to vary from ofﬁce market to ofﬁce market and what causes them to change over time? The real risk-free rate should be constant across markets, at least within a national economy and arguably more widely given the erosion of barriers to global capital ﬂows. While the real risk-free rate cannot be observed directly, the redemption yield on a government inﬂation-linked bond could act as a proxy, as could an inﬂation-adjusted zero coupon bond yield. Figure 6.4 shows real zero coupon bond yields, as estimated by the Bank of England, from 1985 to 2006. These appear to decline sharply from the late 1990s – a fall mirrored in declining index linked bond yields. In part, the fall reﬂects changing perceptions about inﬂation: not only is there a decline in the expected average level of inﬂation, there is also a decline in the expected volatility of inﬂation in the macro-economy. But bond yields are also a product of supply and demand. Large quantities of capitalseeking investment opportunities (often attributed to ‘excessive’ Asian

Investment, Capital Flows and the Ofﬁce Market

173

5 4.5 4 3.5 3 2.5 2 1.5 1 0.5

M

ar c M h1 ar 9 c 8 M h1 5 ar 9 c 8 M h1 6 ar 9 c 8 M h1 7 ar 98 c M h1 8 ar 9 c 8 M h1 9 ar 99 c M h1 0 ar 9 c 9 M h1 1 ar 99 c M h1 2 ar 9 c 9 M h1 3 ar 99 c M h1 4 ar 9 c 9 M h1 5 ar 9 c 9 M h1 6 ar 99 c M h1 7 ar 9 c 9 M h1 8 ar 9 c 9 M h2 9 ar 00 c M h2 0 ar 0 c 0 M h2 1 ar 0 c 0 M h2 2 ar 0 c 0 M h2 3 ar 0 c 0 M h2 4 ar 00 ch 5 20 06

0

Figure 6.4 UK real zero coupon bond yields, 1985–2006. Source: Estimated from Bank of England data.

savings) drive asset prices up and yields down. This demand push, allied to concerns about the solvency of pension schemes and about poor equity returns in the aftermath of the post dot.com stock market adjustment, coincides with relatively low Government bond issuance, forcing down redemption yields. This supply–demand imbalance does not necessarily imply a change in the underlying risk-free rate but would have an impact on asset prices and required returns. The risk premium rewards investors for exposure to systematic risk factors, which would include sensitivity to shocks in macro-variables such as interest rates, inﬂation and the economic cycle and to changes in term structure and risk preference. Some assets will be more affected by shocks than others: for example, highly geared indirect real estate investment vehicles are more vulnerable to interest rate shocks than direct, ungeared, equity investment in real estate. Illiquidity was identiﬁed above as a key investment characteristic of real estate. There may be an illiquidity premium embedded in the real estate risk premium to reﬂect the direct costs and time taken to trade facing a property investor. Illiquidity premia may vary by property type, by geography and over time. Historically, very high value assets – major regional shopping malls or large prime ofﬁce developments in the CBDs of global markets had restricted markets: few investors could raise the capital to acquire them, or would be prepared to accept exposure to an asset that would have a very large weighting in their portfolio. This might suggest that investors would demand higher returns, since their ability to exit was constrained. However, globalisation of investment and innovation in the design of indirect real

174

Towers of Capital

estate investment vehicles permitting co-investment in single assets have eroded that disadvantage. The growth of a wide range of property investment funds in public and private markets, allowing both retail and professional investors ways to increase and decrease their real estate exposure with less constraints and costs than in the private direct investment market can be seen as a structural shift in liquidity – although, as the events of 2007/8 revealed, some of the supposed liquidity of, for example, open-ended funds, proved to be illusory. Yield shifts may thus represent true structural shifts in risk but, possibly, just market perceptions of structural shifts. In terms of geographical variations in liquidity, there may be markets where trading is more active and hence entry in and exit from the market are subject to less friction and risk. It is often argued that this is the case for the prime ofﬁce markets of major ﬁnancial centres: lower delivered returns are justiﬁed by the security of knowing that there will be a buyer, should the investor wish to exit. In a similar fashion, in the largest urban markets, loss of a tenant may be less of a risk factor if there exists a wide, diverse pool of demand, and it is likely that landlord–tenant search costs will be lower as a beneﬁt of agglomeration economies, reducing income risk. Again, it is not clear as to how much this is perception rather than reality. For example, while it is undoubtedly true that there are more investment transactions in the City of London than in a small regional city in the periphery of the United Kingdom, this does not mean that the transaction rate (total sales over total assets) is higher: the higher volume might simply be a function of the greater scale of the market. Key (2004) analysed transaction evidence from IPD for 2000–2002 and found that turnover rates in Central London ofﬁces were no higher than for other regions as a proportion of value and rather lower in terms of the numbers of buildings changing hands (Table 6.5). He does, however, suggest that central London transaction rates are less variable from peak to trough than other markets. Markets will vary in terms of their transparency and institutional structure. There will be cross-national variations. The ofﬁce markets of cities in the transitional economies of Eastern Europe and the emerging Chinese Table 6.5

Office turnover rates: selected cities.

Location

% Sold by number

% Sold by value

16.7 28.2 23.5 27.6 25.6 23.0 32.6

22.4 31.2 17.1 10.7 22.6 28.9 25.7

City of London Birmingham Bristol Edinburgh Leeds Oxford Manchester Source: From data in Key (2004).

Investment, Capital Flows and the Ofﬁce Market

175

markets have faced uncertainties with regard to title and property rights and an absence of any historic data on the operation of a market-led system of land-use development. There will also be variations across cities within a country, as the level of research and robust data differs signiﬁcantly. Given that the underlying private real estate is characterised by the strong information asymmetry that is characteristic of second-hand markets (the existing owner knows more about the property than prospective buyers can, and acquisition of information is costly and difﬁcult), lack of transparency creates an ex ante risk that might call for a higher risk premium. The ofﬁce markets of IFCs are typically well-researched and have comparatively good supply, demand and price data. This both reduces risk and enables investment by funds that are driven by quantitative portfolio strategies that demand market data. Of other components of the cap rate, the growth rate might seem to be an obvious contender for spatial or temporal variation. It needs to be appreciated, however, that the gr is the long-run real growth rate. Short-term rises or falls in rents in a market ought not to have a signiﬁcant impact on the yield, particularly if rents are seen as mean- or trend-reverting. For two markets to have signiﬁcantly different real growth trajectories has to imply large structural differences in economic prospects, sufﬁcient to overcome convergence forces (as the increasing rental levels of the market with higher growth damage the proﬁtability of ﬁrms locating there) or substantial and sustained differences in supply responsiveness due to land constraints or planning policy.5 A city faced with economic decline and lack of competitiveness will be unable to retain existing business and attract new business; reduced demand is likely to drive real rents downwards. An emerging city ofﬁce market, where growing agglomeration economies and the establishment of a critical mass of activity feed back into occupational demand, may experience high levels of real rental growth. What is not clear is whether higher levels of real growth can be sustained in established major cities. The limited empirical evidence from large cities suggests that expectations of continued real income and capital value growth may be misplaced. Figure 6.5 (based on ﬁgures kindly provided by Steven Devaney) shows real rents in the City of London from 1905 to 1959. There are evidently major cycles linked to geo-political events; over the whole period, however, real rental growth has been very close to zero. Figure 6.6 shows real ofﬁce rents in the City of London from 1977 to 2006. Once again, over the whole period, there is no real growth: rents rise rapidly in the late 1980s with ﬁnancial restructuring but fall back sharply in the early 1990s, never recovering to their nominal peak values. 5

Even here, if there are long-run supply constraints, then high real rents may choke off growth.

176

Towers of Capital

150

Real rent 1971 = 100

125

100

75

50

20 19 23 19 26 19 29 19 32 19 35 19 38 19 41 19 44 19 47 19 50 19 53 19 56 19 59

17

19

14

19

11

19

08

19

19

19

05

25

Figure 6.5 Real City of London office rents, 1905–1959. Source: Data taken from Devaney (forthcoming, 2009).

200 175

Real rent 1971 = 100

150 125 100 75

50

19 77 19 79 19 81 19 83 19 85 19 87 19 89 19 91 19 93 19 95 19 97 19 99 20 01 20 03 20 05

25

Figure 6.6 Real City of London office rents, 1977–2006. Source: CBRE, ONS, Hendershott et al. (2008).

Investment, Capital Flows and the Ofﬁce Market

177

Depreciation Depreciation reﬂects the impact of the ageing of a building on rent chargeable or, equivalently, the requirement to invest capital in a building to maintain its income-generating potential. In aggregate, depreciation rates reﬂect the need to replace the stock in an urban ofﬁce market. Headline rent ﬁgures published by real estate consultants often reﬂect the ‘best rent’ achievable in a market. Thus, headline growth rates overstate the income growth received by an investor (average rent ﬁgures are less affected by this problem). A new building may attract the highest rent in a city but its rent relative to the best rent will fall as it ages. Evidence suggests that this fall is non-linear: initially high relative rents are maintained; they then decline sharply before stabilising. Depreciation is linked both to physical and functional obsolescence. Changing working technologies can render an ofﬁce stock obsolescent – as noted above, the restructuring of international ﬁnancial services brought a demand for large ﬂoorplates, under ﬂoor cabling and suspended ceilings for heating, ventilation and air conditioning that rendered many traditional ofﬁce buildings inappropriate in global ﬁnancial centres, promoting a wave of new building and redevelopment. Do depreciation rates vary systematically across geographical markets? Baum and Turner (2004) argue that rental depreciation varies as a function of lease structure. Where there are very long leases (as, for example, in the United Kingdom), landlords have less incentive to reinvest income in their buildings to maintain their value. They found that the owners of ofﬁces let to a single tenant in London retained just 0.22% of annual income, compared to over 1% in Frankfurt, Paris and Stockholm, differences that were statistically signiﬁcant. These retention rates translate directly into rental depreciation: as London’s single-let buildings age, their rents relative to top rents in the market falls at a much faster rate than rents in Amsterdam, Frankfurt and Paris – and in particular, faster than in Stockholm, where short lets, multiple-tenanted ofﬁces and more active landlord involvement are the norm. This suggests that one might expect spatial variations in depreciation rates, and hence yields, across national boundaries or even across cities and regions within a country if typical lease arrangements vary. Depreciation rates might also be higher in a market where functional obsolescence was more prevalent or where innovations in operational and production processes were more frequent. In principle, that could apply to IFCs, which might imply higher rather than lower depreciation rates. These might appear as step changes. For example, the shift towards electronic trading ﬂoors created a demand for large open ﬂoorplate ofﬁces with sufﬁcient ﬂoor-to-ceiling height to accommodate cabling and heating and air conditioning in raised ﬂoors and suspended ceilings – suddenly rendering older buildings apparently obsolete and in need of replacement. Yet the rise

178

Towers of Capital

of wireless and low heat emission technologies and a move back to more personal, compartmentalised workspace might make the older space more acceptable and present problems for the adaptation of new space. Changes in regulation (for example on health and safety or environmental impact) can affect the demand for buildings and hence their ability to generate rent and, again, these may vary by nation and by city (for example, German workspace legislation about the maximum allowable distance to natural light affects the design and conﬁguration of ofﬁce space). For all these potential differences, change in required depreciation premium is (usually) likely to be gradual and result in only minor shifts to yields. While a theoretical decomposition of yields might suggest that capitalisation rates should be relatively stable, it is evident that they vary considerably both across space and time. Figure 6.7 shows initial and equivalent yields for City of London ofﬁces between 1981 and 2006. The more stable equivalent yield series (which corrects for reversionary income potential caused by the ﬁve-yearly rent review) shows a sharp upward movement in the early 1990s, with a consequent correction, then a steady fall from the early 2000s. This pattern is only partially explained by changing real bond yields and reﬂects investment supply and demand in the market. While a rational market model suggests that the price of an asset should reﬂect the present value of the expected future cashﬂow given current information, excess investment demand will drive prices up and excess supply relative to demand will drive prices down. Are the cap rate changes observed, then, based on a rational

12 Equivalent yield 10

Initial yield

8

6

4

2

19

8 19 1 8 19 2 8 19 3 8 19 4 8 19 5 8 19 6 8 19 7 8 19 8 8 19 9 9 19 0 9 19 1 9 19 2 9 19 3 9 19 4 9 19 5 9 19 6 9 19 7 9 19 8 9 20 9 0 20 0 0 20 1 0 20 2 0 20 3 0 20 4 0 20 5 06

0

Figure 6.7 IPD Data.

Office capitalisation rates in the City of London. Source: Estimated from

Investment, Capital Flows and the Ofﬁce Market

179

assessment of future prospects and required risk premia, or do they result from sentiment that is not based on economic fundamentals? Hendershott (2000) examines capital values and cap rates in the Sydney ofﬁce market over the 1985–1996 boom and bust. He argues that the value of a property in an overbuilt market should be equal to its replacement cost less the present value of the below equilibrium rents; similarly, in an under-supplied market, the value should be replacement cost plus the present value of over-equilibrium rent. This implies, ﬁrst, that real rents are expected to mean revert; second, that investors should anticipate the mean reversion caused, respectively, by rising vacancy rates or by new development triggered by the above equilibrium values. Empirically, however, he ﬁnds that cap rates do not follow such a model; they are consistent with investors extrapolating current rental growth forward at rental peaks and cutting risk premia. This, it is suggested, contributed to an asset price bubble that, in turn, led to the over-supply and reversal of rents and values. Bubbles and cycles are considered further in Chapter 8. Sivitanides et al. (2001) ﬁnd a similar investor myopia in US cap rates: the ratio of current real rent to average rent is negatively correlated with cap rates. Similarly, Chen, Hudson-Wilson and Nordby ﬁnd cap rates fall as net operating income rises, which is consistent with the contemporaneous effect observed by Sivitanides et al. and Hendershott. They note that cap rates fell in the early 2000s even though rental market fundamentals were deteriorating, but attribute this to a fall in bond yields, suggesting that the risk premium rose. Their analysis is couched in nominal terms and assumes single period adjustments rather than long-run real expectations and their argument that real estate pricing is ‘rational’ does not really convince. Bjorklund and Soderburg (1999) also point to optimistic overvaluations at the rental peak in Stockholm that suggest extrapolation of short-run growth trends. By contrast, Hendershott and MacGregor (2005a) analyse UK cap rates and make a stronger argument that British ofﬁce yields are consistent with a forward-looking view of rental change. They allow both for mean reversion in rents (by calculating the deviation of real rent to average or trend) and serial correlation (to allow for stickiness in rental adjustment) and show that both are signiﬁcant in explaining shifts in cap rate. They also ﬁnd a link to the capital markets via equity dividend growth. Although the model is forward looking and does suggest investors to consider mean reversion, there is an assumption that rental adjustment will be slow and that rental growth generates momentum effects. In addition, they include an ‘investment weight’ variable (which measures property’s share of institutional investment portfolios) which has a negative impact on cap rates that is statistically and economically signiﬁcant: that is, weight of money has an impact on yields that is separate from any consideration of rental fundamentals.

180

Towers of Capital

Intuitively, UK – and speciﬁcally London – ofﬁce yields seem to be too low at the top of rent cycles, leading to sharp capital market adjustments at the end of property booms.

6.7

Investment, capital ﬂows and IFC ofﬁce markets

Improvements in commercial real estate performance data have allowed professional investors to assess the potential beneﬁts of property as part of mixed-asset portfolios. Asset allocation models typically suggest a relatively high weighting to real estate: allied to the development of a range of public and private market real estate investment vehicles that overcome some of the problematic characteristics of property as an asset class, this has led to an increase in capital ﬂows and an increase in property’s share of the portfolio. Real estate investment is increasingly international in nature, driven by removal of barriers to acquisition, globalisation of property advisory and research services and by the appearance of multinational investment funds. Those ﬂows appear to affect capitalisation rates and, hence, real estate values. Some of the observed yield shifts can be explained rationally in terms of falls in the risk premium through improved information and the impact of new vehicles on liquidity. Other shifts – both over time and across markets – are harder to explain and seem to be based on sentiment or on myopic reactions to short-term price signals. Within the real estate portfolio, diversiﬁcation should be based on risk– return characteristics and on the correlation between property types or markets. Research on diversiﬁcation does not produce consensus. In general, sector-based diversiﬁcation (between, say, ofﬁces, retail and industrial property) seems to offer greater beneﬁts than geographical diversiﬁcation, but this depends on the spatial frame of reference: there appear to be beneﬁts from international property diversiﬁcation, although a global real estate investment strategy faces many practical difﬁculties. Analysis, however, is typically at an aggregate level. When individual buildings are considered, it seems that the pattern of property returns do not conform to standard sector–region groupings, suggesting that most portfolios are exposed to considerable asset-speciﬁc risk and are far from fully diversiﬁed. Individual real estate investment portfolios may be specialised (to capture specialist management or stock selection skills) or diversiﬁed. In aggregate, however, professional real estate holdings should be consistent with expectations about risk and return. Based on historical returns, it is by no means clear this is the case. High portfolio weightings are found in sectors and markets that appear to offer returns that are low relative to risk. This appears true for the ofﬁce sector in general and for the ofﬁce markets of major cities in particular. The proportion of property of a particular type in

Investment, Capital Flows and the Ofﬁce Market

181

the portfolio reﬂects anticipated, future returns: but historical evidence of low or negative real growth in rents and capital values and highly volatile, cyclical returns sit uneasily with observed holdings. Based on historical returns data and conventional portfolio allocation models, then, the share of real estate investment going to the ofﬁce markets of world cities in general, and IFCs in particular, seems hard to justify. What might explain the discrepancies observed? Is it possible that investors hold beliefs about rental growth that are not empirically grounded? The attractions of IFCs as a business location, the agglomeration economies and knowledge spillovers, encourage ﬁrms to locate there and generate higher proﬁts that can be captured in rent. This creates higher rental levels than those found in markets unable to offer the same economies of scale or scope. But, once this relative gap is established, that does not mean that real rents will rise at a faster rate. A new or emerging IFC should exhibit strong initial rental growth, but for an established centre, adjustment processes should limit rises. First, increasing rents translate into increasing capital values, encouraging new development and a supply side adjustment. Second, rising rents affect the proﬁtability of the city as a business location and drive out marginal occupiers, producing a demand adjustment. These processes may be masked by the impacts of inﬂation and by myopia in capitalisation rate responses to the rent cycle. Another factor might be the existence (or more importantly, the belief in the existence) of greater liquidity in IFC markets. Since IFC markets are large, there is generally a high volume of sales creating an impression of liquidity; since they are international, then there are often international buyers that might seem to offer a protection against domestic business cycles and an exit route in a market downturn. This is further reinforced by the availability of market data and research that can be supported by large, high-value markets. Greater transparency, allied to historic time series to input into allocation and performance models, encourages investment. Data quality is also is an important factor in obtaining ﬁnance and funding for real estate investment. Further, since prime ofﬁces in IFCs tend to be large and of high value, international investors can concentrate their investment and gain scale economies in management and monitoring costs. These characteristics, then, create a rationale for investing in IFC ofﬁce markets. That rationale may not, of course, be ﬁrmly based on rigorous empirical analysis; the importance is that it is believed. Layered on these rational economic arguments are more behavioural factors linked to presence in global markets, ownership of ﬂagship or trophy buildings and the beneﬁts of location of assets in cosmopolitan cities with all their services. Investment demand, occupational demand, rental levels, supply of space and the price of space are interlinked and that linkage is particularly strong in the ofﬁce markets of IFCs and world cities. If, as suggested above, yields

182

Towers of Capital

are both affected by investment demand and myopic in relation to rental growth, it is likely that there will be capital value cycles that are locked together with the development and rent cycles, creating potentially greater volatility, larger peaks and troughs in IFC ofﬁce markets. In the downturns, the precise nature of liquidity becomes critical. Investors – whether domestic or international, institutional or individual – may ﬁnd that their supposed ease of exit is illusory. Since real estate is traded in private markets and since it is ﬁxed in location, adjustment in down markets can come as much from a slow down in trading activity as a drop in prices. The fall of transactions results from a reluctance of investors to crystallise losses and a lack of effective investment demand. By contrast, in a booming market, investment demand is likely to outstrip the supply of available assets – existing buildings – and the lag inherent in the development process constrains the supply of new assets. This drives yields down and asset prices up. To develop this argument further, two strands need to be developed. First, the nature of ownership structures and vehicles needs more detailed consideration: how is the investment process structured? Second, an analysis of the nature of cycles and the possibility of bubbles in asset markets is required. Part III addresses these two critical issues.

Part III International Finance, Global Ofﬁce Markets and Systemic Risk

Introduction Part I of the book examined the evolution of the global ﬁnancial system and the key role played by a small set of global cities that play a key role in coordinating capital ﬂows and which increasingly capture a major share of global ﬁnancial activity – the international ﬁnancial centres (IFCs). IFCs, as a focus for global ﬁnancial activity, must have major core ofﬁce market to create working space for ﬁnancial ﬁrms and their supporting business and professional services. Part II examined the dynamics of ofﬁce markets in those IFCs, linking demand for space, ofﬁce development and investment in real estate. In this third and ﬁnal part, the two themes are brought together to explore the interaction between global ﬁnancial markets and real estate markets, the impact of that integration in the ofﬁce markets of IFCs and the implications for risk, stability and urban policy. Chapter 7 considers the behaviour of international ﬁnancial markets and the existence of asset market bubbles and ﬁnancial crises. Financial crises often spread out beyond a single national market, with effects spilling over into surrounding regional and global markets. As Chapter 3 revealed, this is no new phenomenon: there are examples of regional and global ﬁnancial crises from the eighteenth and nineteenth centuries with a shock in one country spreading rapidly to other ﬁnancial centres around Europe and beyond. In the modern era of global capitalism, developments in information and communications technologies have altered the speed and the extent of the transmission of shocks. Equity market, currency market or interest rate shocks can be transmitted near instantaneously around global markets. The speed of adjustment to shocks has also accelerated. Real estate markets in IFCs are directly affected by ﬁnancial crises. Financial ﬁrms laying off staff affect demand and hence rents; interest rate changes and the availability of capital affect prices. But real estate can also have causal effects – the real estate cycle interacting with credit cycles and ﬁnancial booms and busts and contributing to systemic risk as banks and other ﬁnancial ﬁrms are affected by their exposure to property markets. In a similar fashion, real asset market bubbles may be linked to global credit cycles and appear simultaneously in cities across the world. Chapter 8 highlights a set of interrelated changes in commercial real estate markets that have had a major impact on risk and return in IFC ofﬁce markets. The last two decades have seen waves of innovation in debt and equity vehicles in real estate. Two features appear particularly signiﬁcant: the rapid growth of private real estate funds that pool together capital from many investors to acquire properties and the development of securitised debt products that simultaneously provide a new source of real estate capital and distribute real estate exposure more widely across ﬁnancial markets.

Introduction

185

The development of private real estate funds, along with the spread of tax-efﬁcient listed Real Estate Investment Trusts, has enabled a globalisation of real estate investment as professional investors pursue international strategies. Much of that global activity is concentrated in IFCs – both directly (through the acquisition of ofﬁce buildings in the CBDs of global cities) and indirectly through fund management activities and provision of capital. In turn, the new investment vehicles have contributed to the fragmentation of ofﬁce ownership in major cities, with implications for the management of the built environment and the planning of cities. Finally, Chapter 9 analyses the implications of all of these changes. Real estate ﬁnance, investment and development are closely intertwined in the ofﬁce markets of global cities. In IFCs, demand for space comes from an externally oriented ﬁnancial services sector; thus ﬂuctuations in global ﬁnancial market conditions directly affect space requirements and, hence rents. That same ﬁnancial sector provides capital for the development of space and commissions new building. The ﬁnancial sector acquires completed developments – for owner occupation or, more commonly, as investors and thus is exposed to the performance of ofﬁce markets – which is, in turn, driven by the ﬁnancial market environment. The way that ofﬁce markets inﬂuence and are inﬂuenced by capital markets is critical to understanding risk in global cities. Moreover, the factors that drive real estate market performance are increasingly global rather than local – which links together the fortunes of ﬁnancial centre ofﬁces across global cities. For example, a major investment bank will be represented in many, if not all, key ﬁnancial centres around the world. In a ﬁnancial crisis, staff will be lost in all those cities and, hence, there will be reductions in demand and downward pressure on rents. The precise timing and magnitude of the impacts may well vary, inﬂuenced by property market structures, by labour laws and by the extent to which ﬁnance dominates demand – but the fundamental demand drivers will coincide across markets. These interactions between ofﬁce markets and ﬁnancial markets, allied to cyclical tendencies in both markets and the inherent illiquidity of real estate create a potential systemic risk, where shocks in one part of the system can spread rapidly from one city to another and from one sector of a city’s economy to another. The 2007 credit crunch provides a good example, as the impacts of the US sub-prime mortgage crisis rippled across the world, as a result of global investment in securitised debt products; poor performance affected banks and other ﬁnancial ﬁrms across many global cities, affecting demand; the linked liquidity crisis affected interest rates, credit availability and asset prices; falling asset prices affected ﬁnancial sector performance, deepening problems. The chapter concludes by exploring the policy implications of this interlocking set of markets and the inherent risk and volatility that they create.

7 Booms, Bubbles, Crises and Contagion

7.1

Introduction

Any discussion of the globalisation of economies and ﬁnancial markets has to take into account a great variety of approaches to the topic. Within a varied literature, terms are used in different ways and are often loosely deﬁned, the type of fuzzy concepts criticised by Markusen (1999): convergence, integration, globalisation all have multiple layers of meaning.1 Much of the critical writing2 on globalisation sees globalisation as a product of capitalist expansion, eroding national governments’ ability to make independent policy decisions. The move to a globally integrated market is seen as a recent phenomenon, the product of geo-political events and technological innovation, which have served to lock markets together. Amin and Thrift (1992), for example, argue that a transition occurred, starting in the 1970s, towards a globalised economy. As major features they cite: an increasing mobility of international capital; the emergence of a knowledge or information society not conﬁned to the nation state; developments in communications technology that transcend national boundaries; technological innovation and rapid technological innovation; the growing importance of multinational enterprises whose operations have been facilitated both by deregulation and by information technology; the growth of multinational institutions whose purpose is to remove barriers to global business; the decline in importance of the nation state and a cultural homogenisation driven by new international media. From this perspective, the world economy is more than a set of interlinked nation states. The observation of globally coordinated ﬁnancial market crises, common patterns of boom and bust, is explained by an assumption of international division of labour and dispersion of economic activity. 1 2

For an extensive review, see McAllister (2008). In particular that emerging from post-Marxian, structural and critical perspectives.

188

Towers of Capital

The rapid rise in equity markets and the subsequent inevitabe market crashes – 1987’s Black Monday, the collapse of the dot.com boom in 2000 – are seen as a product of a globalised ‘casino capitalism’; in a similar fashion, real estate booms and busts can be explained in terms of capital switching between primary and secondary circuits of capital in an extension of David Harvey’s model These explanations merge into the global cities literature with urban development in cities explained by the logic of global capitalism and analysed in terms of conﬂict between the interests of ﬁnance capital and those of a marginalised urban population. Yet, while there are valuable insights here, the explanations are rather unsatisfactory. First, they lack a historical perspective. As earlier chapters have demonstrated, global trading activity focused on a small number of cities playing roles out of proportion to their size is no new phenomenon. Second, the mechanisms that link economies and markets together and that transmit shocks around the world tend to be weakly speciﬁed. For example, technology enables 24-hour trading and allows near-instantaneous observation of market behaviour around the world – but it does not force investors and market makers in, say, Sydney, to behave in the same way as those in Toronto. Third, if two assets trading in different national markets have the same risk characteristics and drivers of returns, it is not clear why they would be expected to behave in a different manner – that is, economic and asset market integration has simply removed pricing inefﬁciencies and frictions. Finally, it is clear that ‘global crises’ are not universally coordinated with effects across all market. Some have impacts that are largely conﬁned to one region or that have lagged and complex impacts on external markets (the East Asian currency crisis for example). Other shocks remain conﬁned to national markets. Thus co-movement in asset markets and ﬁnancial market integration processes are more complex than often portrayed. Nonetheless, there are co-movements. Equity market prices rise and fall across the world: a sharp jump or fall in a major market is echoed in other markets on a near-instantaneous basis. Commodity prices are set globally and have impacts for global inﬂation rates and, hence, nominal interest rate shocks. There are parallel movements in bond yields, reﬂecting some common attitudes to risk and future volatility. And real estate cycles appear to be increasingly coordinated, at least in developed economies. This chapter explores those common movements and examines the linkage between ﬁnancial markets and real estate. First, evidence on co-movement in global asset markets and on contagion effects is considered. Then asset market bubbles, slumps and crises are examined in more detail, exploring the link between banking crises, currency crises and the real estate cycle. Real estate is strongly affected by the credit and investment cycles – but also plays a signiﬁcant role in exacerbating downturns. The emphasis here is on understanding the mechanisms by which asset market bubbles can form

Booms, Bubbles, Crises and Contagion

189

and burst – without resorting to some deus ex machina explanation based on untestable meta-theories.

7.2

Financial markets, co-movement and crises

The growth of international ownership of shares, allied to technological changes in trading mechanisms make equity markets prone to common patterns of movement and higher correlation. For example, in a European context, Freimann (1998) and Rouwenhorst (1999) found increases in average correlation between national markets from approximately 0.2 in the 1980s to 0.6 in the 1990s, Baele and Vennet (2001) found signiﬁcantly positive contemporaneous correlations between local excess returns and pan-European returns ranging from 0.57 in Belgium to 0.88 in the United Kingdom and Adjaouté and Danthine (2002) identiﬁed clear evidence of convergence in equity returns up to 1992. For lower frequency data, Fraser and Oyefeso (2005) reported correlation coefﬁcients of monthly real returns for the United States and major European markets typically of 0.9 and above. For the major equity markets, there is evidence of reducing country effects relative to sector effects (see Baca et al., 2000; Cavaglia et al., 2000) implying that cross-national economic fundamentals are driving asset prices. At individual sector level, real estate securities appear less correlated across national boundaries. Property company and Real Estate Investment Trust (REIT) returns are correlated with national equity indices, but correlations with real estate in other countries are lower than for general equities. This would be consistent both with the idea of strong sectoral effects and with the localised nature of real estate investment. Lizieri et al. (2003) also found that national real estate markets were much more segmented than common stocks in that correlations were substantially lower and that loadings on a common factor were less. This implies that there are greater diversiﬁcation beneﬁts to be gained from a global property investment strategy than from construction of a general international equity market portfolio. Even here, though, correlations have been increasing over time (see, for example, McAllister & Lizieri, 2006), particularly in those markets with a mature, sizable listed real estate sector. Focusing implicitly on global integration, there have been a number of studies examining the relative contribution of global systematic relative to country-speciﬁc factors in explaining the behaviour of national real estate security markets and individual real estate company returns. Ling and Naranjo (2002), using a simple two-factor model to measure the relative contribution of a worldwide factor and country factors in company real estate returns found evidence of a strong worldwide factor but that the country-speciﬁc factor was signiﬁcant in over 90% of cases. Bond et al. (2003)

190

Towers of Capital

built on the work of Ling and Naranjo by incorporating country value factors. They found that the coefﬁcient of the global market factor was positive and signiﬁcantly different from zero for all countries except Germany. However, they also reported that the portion of the variation in individual country excess returns explained by the global market was low. Hamelink and Hoesli (2004) similarly ﬁnd a strong common factor in listed real estate returns but also signiﬁcant individual national factors. Eichholtz et al. (1998) tested for the existence of ‘continental’ (or regional) factors in real estate securities. They found evidence of a strong European effect with a signiﬁcant continental factor which appeared to increase in strength from the early 1990s with the completion of the Single European Market and the moves toward Monetary Union. By contrast, they found little evidence of a signiﬁcant Asian continental factor. The extent of co-movement in global equity markets can be seen in Figure 7.1, which shows rolling three-year correlations, based on monthly stock market returns, for the major global stockmarkets: the United States, Canada, United Kingdom, Germany, France, Switzerland, Japan, Hong Kong, Singapore and Australia, all based in major international ﬁnancial centres (IFCs). Correlation varies across time but is generally high: over 1985–2007 the mean correlation of national stock indices with the Morgan Stanley Capital International World stock market index is 0.714 – that is, around 51% of the variation in national market equity returns is explained by global market ﬂuctuations. The degree of co-movement with the global equity index varies: the United States has the highest correlation over the whole period (at 0.886) – this is perhaps unsurprising given the signiﬁcance of US stocks in the world index, followed by the United Kingdom (0.794) and Canada (0.774). Asian markets have lower correlations to the market: Hong Kong (0.556) and Singapore (0.609) trailing behind the 0.850 0.800 0.750 0.700 0.650 0.600 0.550

Ju

ne Ju 19 ne 86 Ju 19 ne 87 Ju 19 ne 88 Ju 198 ne 9 Ju 19 ne 90 Ju 19 ne 91 Ju 19 ne 92 Ju 199 ne 3 Ju 19 ne 94 Ju 199 ne 5 Ju 199 ne 6 Ju 19 ne 97 Ju 19 ne 98 Ju 199 ne 9 Ju 200 ne 0 Ju 20 ne 01 Ju 20 ne 02 Ju 20 ne 03 Ju 200 ne 4 Ju 20 ne 05 20 06

0.500

Figure 7.1 Major equity markets – average correlation. Source: Author’s calculations, data sourced from DataStream.

Booms, Bubbles, Crises and Contagion

191

North American and European markets. Nonetheless, these are still statistically signiﬁcant positive correlations, showing a strong common world equity market factor. Running a principal components analysis on the individual market returns, the ﬁrst component explains 63% of total variance, with all ten markets having factor loadings of 0.69 or more.3 This common movement has implications for investment strategies. One of the assumed beneﬁts of a global investment strategy comes from superior risk-adjusted returns from the diversiﬁcation effects of investing across different national markets. Domestic stocks returns are affected by common national factors: if these national factors are less than perfectly correlated across countries, then global portfolios should dominate purely domestic portfolios. However, such diversiﬁcation beneﬁts may be dissipated if there are strong common global factors, leaving international investors facing higher portfolio management, monitoring and information costs with less performance gains to compensate them. Of course, the markets analysed here are developed markets and it may well be that emerging and smaller markets exhibit less common movement. While much research has focused on the linkages between global equity markets, this is only one aspect of international ﬁnancial integration. Other dimensions include capital ﬂows between countries including foreign direct investment and borrowing, interest rate and bond spreads, international insurance contracts and cross-national derivative trading, with all such activities triggering off active currency trading. Elton et al. (2007) cite an average correlation between long-dated bond indices (converted into US$) for seven developed economies of 0.54 for the period 1990–2000, noting that this was a higher ﬁgure than studies analysing the previous decade. Not only are IFCs linked together through this global activity, they also play a key role in coordinating cross-border trades between countries without fully developed ﬁnancial centres. As noted earlier, much social science literature (embodied, for example, in the idea of a ‘new international ﬁnancial system’) suggests that there has been a signiﬁcant increase in global ﬁnancial integration from the late 1970s, linked both to economic liberalism and to developments in information and communications technology. However, more recent historical analyses call into doubt this assumed break with the past. Obstfeld and Taylor (2004) argue that global capital mobility follows a u-shaped trajectory (Figure 7.2). There was a surge in capital mobility in the period 1870–1914 (linked to the widespread acceptance of a currency system based on the gold standard). 3

However, a Varimax rotation of the ﬁrst two components divides the equity markets into two groups, with the European and North American markets loading strongly on factor one, the Asian markets loading strongly on factor two, with Australian equity returns loading on both factors.

192

Towers of Capital

Foreign assets as % GDP

120 100 80 60 40 20 0 1870

1900

1914

1930

1938

1945

1960

1980

1985

1990

1995

2000

Figure 7.2 Foreign owned assets in the global economy. Source: From data tabulated in Obstfeld and Taylor (2004). Note that growth in financial products means that total asset values exceed value of world GDP in the modern era. Countries in sample increase over time.

Chapter 3 set out how this global activity helped establish the hierarchy of IFCs that – in large measure – persists to today. However, the two world wars and the 1920s recession brought international capital ﬂows to a virtual standstill between 1914 and 1945. The post-war period brought a slow recovery in activity, followed by another surge in global capital market activity from 1974 with the ending of the Bretton Woods system and the growing dominance of economic liberalisation. Obstfeld and Taylor suggest that the evidence of an earlier globally linked set of capital markets implies that international ﬁnancial market integration is, by implication, not a function of technological change but, rather, conditional by political economy factors.4 Global ﬁnancial integration, from a conventional economics perspective, brings a number of theoretical beneﬁts (drawing, inter alia, on Obstfeld and Taylor):

• Risk Diversiﬁcation. Investors and residents of different countries can pool risks, producing more effective insurance than can be obtained through purely domestic arrangements. Portfolio risk pooling can occur through 4

Obstfeld and Taylor do, however, distinguish the nature of capital ﬂows in the modern and earlier era. Much of the modern era ﬂow of global capital is for diversiﬁcation, between rich countries of the ‘North’. In the earlier era, more was development ﬁnance ﬂowing from core to periphery – but, in particular, to low-density ‘settler’ countries – Argentina, Australia, Canada, the United States – most of which are now rich. Where capital ﬂowed to more densely populated, less temperate climate, it tended to be associated with colonial, extractive institutional structures that neither promoted development nor protected individual property rights.

Booms, Bubbles, Crises and Contagion

193

equity, bond and derivative investment, insurance and reinsurance, FDI and, more generally, through contracts subject to exchange rate risk. Following Solnik (1991), if the performance of assets in one country is less than perfectly correlated with the performance in another country, then the average covariance of a global portfolio will be lower than for an equivalent domestic portfolio, reducing investor exposure to speciﬁc risk. • Economic Imbalances. In an internationally integrated ﬁnancial system, countries facing temporary economic recessions or recovering from natural or political disasters can borrow abroad to provide ﬁnance for activity. Developing countries with low domestic capital formation can borrow to ﬁnance productive investment and promote economic growth. • Promoting Growth. A system with no barriers to mobility allows international capital markets to channel savings to their most productive use, enhancing growth in world economic growth. Furthermore, the existence of the portfolio diversiﬁcation and insurance gains already noted may encourage investors to pursue investment strategies that focus on higher expected returns (that is, to have a greater tolerance of asset and project level risk). • Policy Discipline. In an open economy, domestic policy makers must be aware of the capital market implications of their actions. Thus ‘excessive’ borrowing may result in a surge of capital outﬂows, with implications for domestic interest rates and the local currency. From other political perspectives, these may not seem unalloyed nor universal beneﬁts. Countries with weak institutional structures may ﬁnd themselves vulnerable to volatility and capital ﬂight if pursuing a policy of ﬁnancial openness, while any discussion of foreign trade and capital ﬂows between developed and developing nations needs to be mindful of the termsof-trade debate (which is outwith the scope of this book). National governments may ﬁnd their policies limited in a ﬁnancially integrated world. Obstfeld and Taylor describe this as the policy ‘trilemma’. A government may have three goals: to peg the exchange rate, to keep capital markets open and to conduct an active monetary policy. However, it can only simultaneously pursue two of those three aims, reducing national autonomy. This provides a parallel to critical social science assertions that the international ﬁnancial system erodes national autonomy. The measures often demanded in the terms of World Bank or International Monetary Fund rescue arrangements give some credence to such criticisms. More signiﬁcantly for our purposes, does integration into the global ﬁnancial system lead to a higher risk of ﬁnancial or economic crisis? Are crises in one country transmitted to other countries through a contagion process even where economic fundamentals differ across countries? Casual

194

Towers of Capital

observation would suggest that this is the case: a sharp stock market fall in one country seems to coincide with market falls in other countries and correlation between markets seems to increase in the presence of shocks – particularly negative shocks. Early empirical work (for example, Kasa, 1992) seemed to conﬁrm this. There appear to be periods of turmoil in capital markets where many, if not all, markets appear to behave in similar ways. This was the case for equities with sharp falls worldwide in October 1987 (Black Monday) with similar global shocks in 2000 with the bursting of the dot.com bubble; the East Asian ﬁnancial turbulence following the Thai devaluation of 1997 had widespread effects which, while most evident in the Asian markets, have been linked to the Russian debt default and the failure of hedge fund Long-Term Capital Management (LTCM) in 1998; similar regional and global impacts followed the Mexican Peso devaluation in 1994. This evidence of interdependence, with shocks in one market rippling out across asset markets, countries and continents and spilling over into economic activity, would imply that investors may not get diversiﬁcation when they need it most – when ﬁnancial asset prices in a market are falling; that is, ﬁnancial market integration can create systemic risks and greater instability. Fragility in one market sector can trigger panics in other sectors, magnifying effects – for example, a currency crisis may lead to a run on domestic banks as both domestic and foreign investors withdraw capital and seek foreign-denominated assets, deepening the currency problems and regional instability (see Bleaney et al., 2008, for a theoretical model of this process). Furthermore, technological developments mean that shocks can be transmitted around the world’s ﬁnancial markets almost instantaneously, marking a qualitative shift in the risks of contagion. It is, of course, important to deﬁne contagion carefully. A very broad deﬁnition, adapted from the World Bank, might be that contagion ‘is the crosscountry transmission of shocks or the general cross-country spillover effects’. However, most research on international contagion attempts to distinguish between fundamental linkages and contagion effects. Fundamental linkages reﬂect trade links, common economic or ﬁnancial characteristics, crossnational assets and liabilities and similar factors that will drive co-movement of asset prices and/or exchange rates. Contagion effects are common movements that cannot be linked to those fundamentals, or, adopting a much more precise deﬁnition, ‘contagion occurs when cross-country correlations increase during “crisis times” relative to correlations during “tranquil times”‘. Even here there are technical, measurement issues. As Forbes and Rigobon (2002) noted, if a ﬁnancial crisis is accompanied by an increase in volatility and two countries have a common factor driving asset returns, then there will be an increase in correlation with no break or regime shift in the relationship between the countries.

Booms, Bubbles, Crises and Contagion

195

Let the returns on an asset class in country i, Ri, be inﬂuenced by the returns in country j, Rj – for simplicity assuming a single-index model.5 Ri = β0 + β1Rj + υi

(7.1)

The variance of asset returns var(Ri) = β12 var(Rj) + var(υi)

(7.2)

cov(Ri,Rj) = β1 var(Rj)

(7.3)

the covariance,

and, ﬁnally, the correlation,  var(υ i )  ρ(Ri , Rj ) =  1 + 2 β var(rj )  

(7.4)

Thus, the overall variance of asset returns in country i and the covariance and correlation of asset returns between countries i and j are all increasing with the volatility of returns in country j, with no change in the β needed. Hence, ﬁnancial turmoil in country j linked to higher volatility of prices will induce an increase in correlation. Forbes and Rigobon suggest that the increase in correlation is not ‘contagion’ but results from the common factor. They use this restrictive deﬁnition to conclude that, in the ﬁnancial crises they analyse, there was ‘no contagion, only interdependence’. Subsequent empirical research in this vein has sought to test this assertion using different techniques to separate out fundamental and spillover effects, with mixed results. For example, Corsetti et al. (2005) use a factor model of stock returns to examine the 1997 Hong Kong stock market crisis and its impact on 17 other countries. Correlation increases across 15 of the 17 markets, but much of this can be explained, as in the Forbes and Rigobon model, by the effect of increased volatility. However, they detect ‘pure’ contagion effects for ﬁve countries: Singapore and the Philippines in Asia and France, Italy and the United Kingdom in Europe. By contrast, Candelion et al. (2005) use a common cycle model to argue that there is no signiﬁcant evidence of an increase in common features in either the Hong Kong stock market crisis or the 1994 Mexican Peso crisis and that shocks are being transmitted through existing mechanisms. However, Masih and Masih (2001), using a co-integration and error correction approach to examine transmission of shocks across developed ﬁnancial markets, ﬁnd a strong single global market factor, with shocks transmitted across markets. Japanese, 5

Adapted from Corsetti et al. (2005).

196

Towers of Capital

US and UK markets seem to lead – receiving the shocks then transmitting them to other markets. Japan is more signiﬁcant in Asian markets, while the United Kingdom appears more signiﬁcant as a leading market than the United States. They attribute this to London’s more international nature and to the higher percentage of foreign companies listed on the London Stock Exchange, giving it an inﬂuence greater than its size would seem to warrant. While the results are mixed, even using the most restrictive deﬁnition of contagion, many researchers ﬁnd that shocks are transmitted across markets even where there are no direct links or common structures. Others have sought to identify the mechanisms that transmit ﬁnancial shocks, whether based on economic fundamentals or on contagion. Valuable summaries may be found in Kaminsky et al. (2003) and Dornbusch et al. (2000). Spillovers may occur from ‘normal’ interdependence among market economies from trade links and ﬁnancial linkages and from economic similarities creating the possibility of simultaneous shocks. Closely integrated ﬁnancial markets are likely to exhibit simultaneously price movements and closely integrated economies may see highly correlated movements in fundamental economic variables. It is also possible that macro-economic policy decisions may trigger reactions from other governments. For example, a devaluation by one country to capture export share may lead to competitive devaluations from regional competitors, with the non-cooperative game pushing all currencies down lower than economically necessary. Rational investors, anticipating the competitive devaluations may withdraw capital, exacerbating the situation. However, contagion effects, particularly what Kaminsky et al. describe as ‘fast and furious chain reactions’ may have behavioural origins based on herd behaviour, information asymmetry, the actions of individual agents and prior investment structures. This does not necessarily imply individual irrational behaviour; but individual rational behaviour in the absence of complete information may result in sudden aggregate shifts in ﬁnancial and economic activity. Information asymmetry may be an important mechanism promoting contagion effects. Individual investors may have limited information about individual countries but can observe the actions of more informed investors and mimic them. If there is a negative shock in country one, an informed investor may seek to sell assets in anticipation of redemption demands: however, they may be unable or unwilling to sell assets in country one (there may be liquidity constraints or concerns about selling into a falling market), so sell in country two. The uninformed investor, holding assets in country two can observe both the shock in country one and the actions of the informed investor in country two – and will replicate the sale strategy pushing asset prices down and triggering chain reactions based on information cascades. Falls in value in one country or region also alter both the market portfolio and evaluations of asset risk which may drive portfolio rebalancing

Booms, Bubbles, Crises and Contagion

197

strategies. Shleifer and Vishny (1997) further argue that international arbitrage is carried out by relatively few specialist, highly leveraged investors. When prices move out of line, such investors ﬁnd it very difﬁcult to function, meaning that arbitrage and bail out is unavailable exactly when it is needed. One implication here is that widely traded, liquid assets and open, integrated ﬁnancial markets are more vulnerable to contagion shocks. Kaminsky et al. identify three interrelated factors that can lead to major contagion episodes. First, a country or region experiences a rapid reversal of capital ﬂows. Secondly, the event is triggered by a sudden announcement, a shock. Third, transmission is driven by the presence of highly leveraged common creditors across countries. They argue that contagion often follows a surge of international capital into a region. Dornbusch et al. (2000) show that private capital ﬂows to Indonesia, Korea, Malaysia, the Philippines and Thailand surged across the mid-1990s, averaging $40 billion per annum and peaking at $70 billion: but in the second half of 1997, over $100 billion in short-term bank loans were recalled as the East Asian ﬁnancial crisis evolved. The shock factor is necessary to prevent a gradual rebalancing of positions. For example, the 1998 Russian debt default was not preceded by a worsening of credit ratings (in fact Russia’s credit rating was upgraded in June 1998) while there had been ﬁve downgrades prior to Argentina’s 2001 default and, consequently, there were far fewer spillover effects. Finally, the presence of highly leveraged common creditors can cause contagion through liquidity and portfolio effects. Thus the 1980s Latin American debt defaults were transmitted largely via US banks; European and Japanese banks had been very active in Asian and transitional economies in the 1990s as had US mutual funds; highly leveraged hedge funds were key mechanisms in the Russian debt and LTCM crises and, earlier, in the Exchange Rate Mechanism (ERM) difﬁculties. Hedge funds have two potential magnifying effects: one from signiﬁcant problems if they are exposed in the wrong position in a crisis, the other from exacerbating falls by shorting the market. A shock in one country or region or even in one ﬁrm or investment product may also have an information effect as a ‘wake up call’ – reminding investors of economic fundamentals at the end of a fad or long period of herd behaviour. It may also represent a ‘change to the rules of the game’. Thus the Russian debt default in 1998 increased concerns that other countries might follow such unilateral actions and reduced conﬁdence in the ability of international institutions such as the IMF or the World Bank to control national actions and to organise bail outs and restructuring in the face of multiple crises. This inevitably leads to a reappraisal of portfolio risk and a re-pricing of assets that will not be conﬁned to the country originating the shock. Finally, models of bank runs can be extended to national asset markets to suggest that there may be multiple equilibria. In a bank run, a depositor’s decision to withdraw capital is conditional on the actions of

198

Towers of Capital

others: at some point, it becomes rational for all investors to withdraw. The same model can be applied to contagious ﬁnancial crises. Although, at the time of writing, it is too early for an analytic review of the causes and consequences of the 2007–2008 ‘credit crunch’, many of the features described above can be seen. Capital gluts and asset price inﬂation created competitive pressures to ﬁnd returns and capture market share, with securitisation of debt – for example, through mortgage-backed securities, collateralised debt obligations and structured investment vehicles in principle spreading risk but also providing additional sources of capital and an exit strategy that allowed for more risky lending. Real estate markets, both residential and commercial, relied on increasingly higher proportions of debt in capital structure. From mid-2007, problems in the US residential sub-prime mortgage market, linked to falling asset prices and to underlying interest rates then rippled through ﬁnancial markets. Some of the ripples were based on fundamentals; banks and investors around the world had exposure to structured debt products exposed to sub-prime lending and these assets had been used as collateral to raise capital. But there seem to be behavioural factors too: information asymmetry (what is the exposure to risky structured products, what is the true market value of the asset base) contributed strongly to a banking liquidity crisis (particularly but by no means exclusively in the United Kingdom and the United States), with rapid impacts on market interest rates and on lending policies which, in turn, had impacts on prices in asset markets. Attitudes to risk were reassessed: required returns on commercial mortgage-backed securities (CMBS) increased, lowering the market value of CMBS portfolios. Falling real asset prices meant that real estate loans were in danger of falling in breach of loan to value covenants; investors seeking to reﬁnance loans found either a lack of availability or much more onerous terms. Financial ﬁrms failed or produced poor ﬁnancial performance and began shedding staff or, at the least, shelving expansion plans. As I write, the trajectory of the crisis is uncertain, but the linkages between localised shocks and widespread global outcomes across continents and across asset markets seem clear. Just as Chapter 3 showed that IFCs were by no means a recent phenomenon, so it should not be imagined that international ﬁnancial crises ﬁrst appeared in the twentieth century. Bordo and Eichengreen (1999) provide a longer perspective on ﬁnancial crises and show contagion and spillover effects in the earlier, nineteenth-century era of globalisation. In the 1820s, British capital ﬂows to ﬁnance Latin American mining and infrastructure presaged the 1825 ﬁnancial crisis, with stock exchange crashes, international debt default and banking panics affecting London and a number of other European markets. A second wave of Latin American loans in the 1860s to ﬁnance railroads led to the 1873 crisis; the 1890s Barings Crisis, precipitated by Argentinean private debt defaults had widespread impacts

Booms, Bubbles, Crises and Contagion

199

across Europe with bank solvency and liquidity issues. European harvest failures led to a drain on UK gold reserves, which raised interest rates and had implications for lending to the new world – with banking crises in New York in the absence of a lender of last resort. Even in the 1920s, with the contraction of global capital ﬂows, the collapse of US stock and commodity prices initiated the great depression. Bordo and Eichengreen analyse the economic impacts of ﬁnancial crises in the modern and the earlier globalisation eras, by examining changes in real GDP before and after crises. Their results suggest that the magnitude of negative impacts on output is approximately the same in the pre-1914 and post-1972 periods. They do suggest, however, that twin currency and banking crises are more severe in the later period. There is some suggestion that, in the earlier era, countries and regions recover faster from currency crises (perhaps due to the constraints provided by the gold standard) and slower from banking crises (in the absence of formal domestic lenders of the last resort or global ﬁnancial institutions with a brief to restore stability). As with Kaminsky et al., they see strong capital ﬂows to emerging economies and sectors as an important pre-condition of ﬁnancial crises. Bordo and Murshid (2000) suggest that ﬁnancial shocks were more global in the pre-1914 period since ﬁnancial and commercial centres were strongly concentrated in a small number of Western European cities and policy tools to restore ﬁnancial stability were less well developed. They examine interest rate spreads and ﬁnd that in both periods there are strong co-movements; in the modern era, these tend to be short run, in the earlier period, more long lasting. For currency crises, in the pre-1914 era, impacts were global, while post-1972 impacts tend to be regional and conﬁned either to advanced or to emerging economies. Regional effects have become more pronounced both for interest rates and for exchange rate risks in the modern era although their component analysis indicates a clear global factor explaining around 60% of variation both pre-1914 and post-1972. Goetzmann et al. (2005) provide a complementary review of long-run international market correlation, analysing data series going back to the mid-nineteenth century. They note that the idea of global diversiﬁcation has long been present. They cite the example of a 1774 Dutch investment fund which had Viennese, Prussian, Russian, Spanish, English and Latin American investment assets, mixing government and corporate bonds and securities and note the existence of published works advocating international equity strategies around the turn of the twentieth century. They split their analysis into a number of periods and show that higher correlations are found in both in the 1972–2000 period and in the earlier era of globalisation, from 1872 to 1913, echoing Obstfeld and Taylor’s ﬁndings on capital ﬂows. They note (as with Bordo et al., 1999) that with lower transaction costs, better institutional structures and less information asymmetry ‘integration is broader

200

Towers of Capital

and deeper than 100 years ago’. Much of the reduction in risk from global investment strategies seems to come from the increase in the number of markets and assets available and, correcting for size of markets, it appears that many of the gains come from small emerging markets which, as they mature, come to behave in a more similar manner to pre-existing major markets. They also suggest that correlation between markets increases when crises occur – conﬁrming the ﬁndings of Longin and Solnik (2001). This is particularly noticeable around the 1929 crisis, in an era where, otherwise, average stock market correlations are low. Hence investors may not obtain the beneﬁts of international diversiﬁcation when they most need it. Recent empirical ﬁnance literature has reported that equities very often exhibit some form of asymmetric dependence6 – that is, pairs of stocks or indices exhibit common behaviour patterns at particular points in time, but diverge at other times. Such research, using a battery of econometric techniques, including dynamic conditional correlation analysis, non-parametric tail dependence measures and the copula approach, examines dependence relating to extreme events, or the tails of the distributions in equity, currency and bond markets and explores contagion effects in market ‘crises’. Knight et al. (2005) and Michayluk et al. (2006) provide real estate examples showing that co-movements increase in the tails – and in particular in the negative tail – of property company return distributions, suggesting that there is a higher probability of a coincident crisis than overall correlation might suggest. Financial crises, then, do seem to be a persistent feature of a globalised system of ﬁnancial markets and capital ﬂows. This is not a new phenomenon: in the earlier age of global capital, local shocks had widespread ﬁnancial and economic impacts across nations and continents. Co-movement in asset prices, interest rates and exchange rate volatility can be related to market fundamentals – trade linkages, cross-holdings of assets and liabilities and to similarities in the characteristics of assets and markets that are priced in the global market. Alongside these fundamental effects, contagion can occur due to the behaviour of investors (both domestic and international) whose actions, while individually rational, particularly in the face of limited information, have negative aggregate impacts and can lead to a switch from one equilibrium position to another. What facilitates these international crises is the liberalisation of capital ﬂows, not technology per se. Technology, though, must affect the speed of transmission. Contagion effects can be near instantaneous. This suggests that crises may be deeper but more short lived in the modern era – aided by the greater sophistication of policy tools for promoting ﬁnancial stability. 6

See, for example, Ang and Chen (2002), Ang and Bekaert (2002), Bae et al. (2003), Campbell et al. (2002), Hartmann et al. (2004) and Longin and Solnik (2001).

Booms, Bubbles, Crises and Contagion

7.3

201

Booms and bubbles

The dramatic stock market price reversals of 1974, 1987, 2000 and 2007 and parallel real estate booms and slumps have led to a resurgence in interest in asset market bubbles. The idea of a bubble, that asset prices rise above their fundamental values for a sustained period, before collapsing back to (or below) their long-run real trend, is not new: Kindleberger’s (1978) much cited Manias, Panics and Crashes provides a history of ﬁnancial booms and crises that goes back to Dutch tulip mania in the seventeenth century and the South Sea Bubble of the early eighteenth century. The existence of periods of rapidly escalating stock market and/or property prices where the growth does not seem to be justiﬁed by the performance of the underlying economy seems to cast doubt on the assumption of the rational and efﬁcient operation of asset markets. Accordingly, many explanations of bubbles focus on the actions and perceptions of individual investors and their ‘irrational exuberance’. Shiller (2005) discusses the causes underlying the rise in asset prices – in particular equity and real estate prices – in the US and many other developed economies across the late 1990s. He suggests that between 1994 and 2000, the Dow Jones index more than trebled in value, from 3600 to 11 723 – a nominal compound growth rate of nearly 22% in the context of low consumer inﬂation. Analysis of broader US equity market indices over the same period suggests similar strong growth, with the S&P 500, DataStream and Morgan Stanley indices all exhibiting returns of 16–18% and with inﬂation running at less than 2%. It is hard to explain such growth in relation to the standard economic variables that are considered to drive returns. While US ﬁrms did experience real earnings growth and real interest rates did decline (itself in part a function of supply and demand in capital markets), this alone is not sufﬁcient to explain the rise in price/earnings ratios. Shiller lists a set of ‘fundamental factors’ driving US equity growth which are, in reality, behavioural rather than economic in nature. First, he suggests that the period coincides with a ‘capitalist explosion’ and an outward spread of ownership. This, combined with increasing concerns about long-term job security, led to many more households seeking investment in stocks and real estate for future security. The growth of deﬁned contribution pension schemes (and the decline of deﬁned beneﬁt pension schemes) also boosted investment, as did the growth of mutual funds that could be placed in tax-efﬁcient wrappers. This does not necessarily increase the savings ratio. In fact, as asset prices rise, feelings of growing wealth (and the ability to borrow against that increased wealth) can fuel consumer spending which, in turn, can boost company earnings, which then feeds back into rising equity prices.

202

Towers of Capital

Shiller goes on to suggest that there are cultural and political factors favouring growth – a supportive monetary policy, a belief that the Federal Reserve Board would intervene if the market dropped, tax reductions and anticipated further falls in capital gains tax (which encouraged equity owners to hold onto assets). He also argues that the baby boom generation investing in stocks had a greater appetite for risk, since they had no direct experience of major recessions or economic disasters. As supporting evidence he cites the increase in gambling activity over this period. The low inﬂation environment (and low nominal interest rates) set against historic equity market returns also creates a ‘money illusion’ effect where the public fail to account for historic inﬂation in returns. He claims that there is an unprecedented growth in media reporting of market performance creating a perception of success, widespread awareness of markets and knowledge of investment opportunities. This allows optimistic equity analyst forecasts and market commentaries to be widely disseminated. The standard explanation of the growth in equity prices up to 2000 relates to dot.com ﬁrms and the internet explosion. As Shiller observes, this cannot be a sufﬁcient explanation. The Dow Jones index does not (did not) contain internet or high technology ﬁrms and yet rose faster than broader stock market indices.7 However, given that the growth of the internet and the performance of technology stocks coincided with equity market success, the public (and indeed professional analysts) could conclude that the growth in prices and earnings were explained by IT, that there was a fundamental technological shift or a ‘new paradigm’. The expansion of public access to the internet and growth in information and communications technology also created the phenomenon of day trading – individuals trading in stocks from home or ofﬁce – which, Shiller claims, increased the volume of trading activity and tended to push prices upwards. The internet also acts as a communications medium, spreading information, views and opinions widely. Shiller continues to set out a series of ‘ampliﬁcation mechanisms’ that tended to push share prices (and also house prices) upwards. These include growing investor expectations of, and conﬁdence in, asset price growth; a fear of missing out on high returns and wealth, fuelled by articles and media coverage emphasising the potential beneﬁts to be had (and downplaying the risks) from equity market returns; public attention focusing on the markets (a phenomenon also observed in the late 1920s); and, in general, a series of feedback mechanisms that relate to adaptive expectations, a belief in future price rises conditioned by recent price rises and also to wealth effects. 7

However, over that period, some information technology and internet-based stocks grew dramatically in value and were increasingly represented in the leading equity market indices such as the FTSE 100 – for example Freeserve, Misys, Baltimore Technologies, Energis or Colt Telecom.

Booms, Bubbles, Crises and Contagion

203

He argues that there are also feedback effects between the equity and housing markets: rising house prices created a wealth effect that encouraged equity investing; rising equity portfolio values encouraged investors to trade up their housing. Finally, when the equity bubble burst, investors switched into real estate as a ‘safer’ investment. The equity market did reverse. The S&P 500 fell 43% from the second quarter of 2000 to the end of 2004 and conﬁdence in the stock market fell with it. Even here, though, perceptions play a part. That 43% fall over a two and a half year period can be matched by the 69% increase over the preceding two and a half years. For many of the investing public, that demonstrates that stocks over the whole of the ﬁve-year period must have increased in value. In practice, of course, they have not: $1000 invested in the period would have been worth around $963, assuming all the growth is capital gain and ignoring the effects of inﬂation. Higgins and Osler (1999) demonstrate a similar effect for commercial real estate prices for 12 OECD countries around the late 1980s and early 1990s cycle. The average growth in capital values in the ﬁve years before the peak (generally around 1989–1990) was around 97%; the fall from peak to trough around 55%. By the same process of compounding, this implies real estate prices have fallen by 11–12%. While Shiller’s behavioural explanations of the equity market boom of the late 1990s (and other asset price booms) seem intuitively reasonable, it is not entirely analytically satisfactory. First, it places great importance on individual investors and the retail market. While this is plausible for the residential real estate market, equity and commercial real estate market transactions are largely conducted by professional, informed, investors, who should be less inﬂuenced by media, by money illusion and more inﬂuenced by underlying factors driving returns. Second, it relies on systematic and sustained irrational behaviour with no investors taking contrary positions and dampening down ‘excess’ growth. As with the discussions of property cycles and developer behaviour in earlier chapters, it would be preferable to seek an explanation which relied less on illogical actions and more on the impact of market structures, economic context and information on individual behaviour and aggregate outcomes. Rising asset prices might result from sharply changing fundamentals and rational expectations. Given available information, investors may be rational in expecting growth in earnings, sustained periods of low volatility, benign inﬂation or low real interest rates and reﬂect those expectations in prices. Garber (2000) argues that even the Dutch tulip mania and the South Sea and Mississippi bubbles can be explained in terms of optimistic but plausible expectations of high returns. Similarly Hall (2001) suggests that the rise in equity prices at the end of the 1990s can be explained by the rise in intangible values in ﬁnancial and business service ﬁrms which would be linked to increases in expected earnings and the present value of

204

Towers of Capital

those earnings; furthermore, dividend growth rates implied in the market do correlate with expected earnings.8 Thus, it is unwise to assume that a sharp rise in asset prices necessarily implies the existence of a bubble. Even where asset prices cannot be explained by underlying economic variables, bubbles are not necessarily irrational or speculative in nature. In a ‘normal’ market, prices should be relatively stable at close to fundamental value, since buyers will generally only seek to purchase if they perceive prices to be below fundamental value and sellers will only seek to dispose of assets where they feel the price is above fundamental value. The action of buyers and sellers serves to bring prices back to their equilibrium level. However, such a benign model relies on a number of critical conditions. First, buyers and sellers need to have estimates of fundamental value that are relatively close together. Second, there needs to be a balance between buyers and sellers, and between available assets and capital. Third, if one set of investors bids prices up (down) to excessive levels then another set should trade against them, to arbitrage away differences. Finally, information should be evenly distributed and available in the market. With those assumptions removed, then prices can move above or below their fair value. Hendershott et al. discuss a market which contains pessimists and optimists with very divergent views about the value of an asset. If most investors are pessimists, then average prices will be low. Optimists enter the market and push prices up – until they exceed the value of the majority of pessimists who sell and exit. With no pessimists left, prices can leap to the average value of the optimists and not reﬂect the full aggregation of beliefs and expectations. Pessimists will not buy at the higher level, allowing the excessive prices to be sustained, until and unless events prove the optimists wrong. In that event, prices fall below the expectation of the optimists, to be replaced by pessimists, causing a rapid deﬂation of the bubble. Such a model does require there to be more investors than assets in that asset class. The conditions for a rational bubble, then, are that there exists wide uncertainty and disagreement about asset values; initial good performance in the asset class; an inability to create new product and additional substitutes in the short term; use of debt in investment strategies and high transaction costs. The uncertainty is critical, allowing discrepancy in valuations. Good asset performance generates optimism and allows the optimists to leverage up their gains. Supply side inelasticity prevents new product exploiting and damping the increasing asset prices. Use of debt provides the optimists with more capital to invest (and also, as we shall see, shifts risk away from investors). That leverage, though, exacerbates problems when the bubble deﬂates. Finally transaction costs discourage selling, creating short-run 8

See Hendershott et al. (2003) for an extended discussion.

Booms, Bubbles, Crises and Contagion

205

supply constraints. The late 1990s internet stock bubble, for example, can be explained in these terms. There was considerable uncertainty over the future proﬁtability of dot.coms and early ﬁrms were very successful.9 While implied growth rates were unsustainable, company valuations required only relatively short periods of very high growth followed by a plausible slowing in growth. Widespread capital availability allowed optimists to dominate. However, increasing numbers of internet ﬁrms coming into the market through initial public offerings helped to prick the bubble. In the decline phase, rapid capital losses and the need to liquidate stock to repay debt led to forced sales and a decline in prices to the levels thought reasonable by the pessimists. These conditions conducive to the development of rational bubbles exist in commercial real estate markets. Since the property market is a private market with high transaction costs, trading is infrequent and there are few public sources of information. Those that do exist typically rely on appraisals of value rather than on sales data. As a result, there is considerable uncertainty as to fair or fundamental value, exacerbated by heterogeneity in product. The length of the development process creates lags between price signals and the availability of new supply. Finally, much of the investment in commercial real estate makes substantial use of debt capital, magnifying upside and downside returns and creating a tendency for development to continue even with faltering market conditions. Given these characteristics, it must be possible for rational – and speculative – pricing bubbles to form. Allen and Gale (1999, 2000) make a similar argument regarding the development of asset bubbles in relation to the East Asian equity and currency crises of 1997, but with an international perspective. They suggest that credit expansion is critical in the build up of a bubble. In particular, investors in equity or real estate make considerable use of external funds. If the fund providers are unable to observe the characteristics of the investments being made, they misprice loans, resulting in the risk shifting from investors to lenders. The investors retain upside potential (if the assets gain rapidly in value, the lenders do not participate in that growth) but the lender will suffer loss from sharp falls in value. When a signiﬁcant proportion of investors possess such advantages, asset prices can rise above their fundamental values. This combination of credit availability and uncertainty works directly in private real estate markets but more subtly in equity markets. They also suggest that the credit expansion is linked to ﬁnancial liberalisation and deregulation from the mid-1980s, making the era more crisis prone. Liberalisation moves may include a relaxation of central bank 9

More accurately, some early ﬁrms were very successful – those that survived. The failures tend to be quickly forgotten. Survivorship bias can distort ex post analysis.

206

Towers of Capital

controls on lending or interest rate reductions; the opening up of economies and capital markets to external, international investment; and an expansion of ﬁnancial institutions involved in capital funding, competing for market share. These occur within individual countries but also are coordinated effects across regions and globally. Bubble bursting may be as a result of a real shock (a sudden leap in commodity prices, for example) or a ﬁnancial shock (for example, a tightening of credit policy). Extensive use of debt can then create systemic or contagion effects. Default becomes widespread (which may force central bank intervention or cause a currency crisis, as the government must balance lowering interest rates to stabilising the banking sector or raising interest rates to protect the exchange rate). Banks have a debt overhang and the suppliers of capital to the banking sector, aware of the risk of bank default and liquidation may be unwilling to supply new capital, creating liquidity problems. Banks could sell some of their loan portfolio – but the most likely buyers would be other ﬁnancial institutions facing their own liquidity problems. This, in turn, may create contagion problems as banks seek to liquidate assets held in other banks, in other countries and in other regions in order to release operating capital, exacerbating ﬁnancial fragility.

7.4

Real estate in booms and bubbles

In a ﬁnancial crisis, real estate will be affected. Most obviously, as ﬁrms get into trouble, they shed staff, which reduces the demand for space. This is most stark for bankrupt ﬁrms – landlords face immediate vacancy and loss of rental income (although legal process may create delays in obtaining possession of the buildings). For survivor ﬁrms, existing lease terms may mean that it is not possible to vacate excess space, resulting in under-occupation or hidden vacancy10 and creating lags between ﬁnancial crisis and property market reaction. The overall fall in occupational demand will, however, create downward pressure on rents. It has been estimated that ﬁnancial ﬁrms in New York shed 17% of their staff in the aftermath of the bursting of the dot.com bubble; in the City of London, ﬁnancial and business service employment fell by 55 000 from the 1988 peak to the 1993 trough, and by 26 000 between 2001 and 2004. The converse is, of course, true in growing markets, as the supply side lags new occupational demand, pushing rental values higher. These processes will be most pronounced in ofﬁce markets, since demand for business and ﬁnancial service sector space is most directly affected. There will, nonetheless, be impacts on other property sectors – housing 10

See Englund et al. (2008) for a discussion.

Booms, Bubbles, Crises and Contagion

207

and retail sectors being affected by falling disposable income and increasing uncertainty about employment prospects, for example, and business demand in general slackening as all ﬁrms face worsening credit conditions. Loss of business and consumer conﬁdence may lead to delays in taking up space for expansion, again reducing demand. But are these impacts in property markets secondary symptoms or does real estate contribute to the boom and bust cycle? In documenting recent ﬁnancial crises and shocks around the world, many commentators have highlighted the important role played by real estate both in creating the boom and contributing to the bust phases of sharp cyclical reversals. A valuable starting point is the work of Davis and Zhu (2004), who provide an extensive survey for the Bank for International Settlements. Their argument is, in essence, that bank ﬁnance for real estate interacts with the property cycle, with high levels of lending in the up phase leading to banking problems as markets turn down. This, then, has spillover effects in other markets (both sectoral and geographical). They note the widespread coincidence of property cycles and banking crises in both developed and emerging economies. Their model of property cycles is drawn from Wheaton, Grenadier and others, where underlying economic and business ﬂuctuations generate cyclical oscillations as a result of the inherent characteristics of real estate. Speciﬁc characteristics inﬂuencing cyclical behaviour include supply side lags and stickiness in rental values which hamper adjustment processes; an inability to short the market, which allows momentum effects to exacerbate cycles; market structures that lead to information asymmetry and pricing uncertainty and a reliance on external ﬁnance, which links the real estate market to credit cycles, creating feedback effects. It is this last aspect that points to combined banking and real estate crises. They explore the mechanisms linking the two markets to provide some empirical evidence. Davis and Zhu argue that property prices affect the volume of bank lending. There are direct and indirect effects. Direct effects come from banks’ own ownership of real estate as a capital asset and from lending to the real estate market for development and investment. Indirect effects result from the use of real estate as collateral for lending (by both business and private individuals) and from the impact of rising property prices on company balance sheets. With rising property prices and falling loan to value ratios for the existing loan portfolio, lenders may underestimate default risk and also be less concerned about moral hazard and adverse selection since – in that phase – borrowers have much more to lose from default. Bank lending then feeds back into property prices via liquidity effects: credit availability and lender attitudes affect the demand for real estate and construction investment. Thus ‘ﬂoods of capital seeking investment opportunities and

208

Towers of Capital

the industrial competition amongst ﬁnancial institutions after ﬁnancial deregulation helped to stimulate the building frenzy phenomenon in a number of countries in the 1980s and 1990s’. In time, bank ﬁnance of new development will, other things equal, lead to a fall in property prices – but this is over the long run. There are also common factors: the credit cycle is linked to economic conditions and prospects which lead to shifts in economic activity that translate into real estate demand. This linkage means that shocks in either credit or real estate markets spill over into the other. Whether these shocks produce sustained oscillations or there is rapid convergence to a steady state depends on market structure and the relative elasticities of supply and demand. The model requires lenders to be strongly inﬂuenced by the current price of real estate, to form either myopic or adaptive expectations about prices. They argue that investors do tend to extrapolate existing price trends forward, partly as a result of weak information in real estate. Second, lending capacity is automatically linked to property prices through the value of bank assets (property loan portfolios and real estate asset portfolios increase in value with rising prices). Third, they cite Bernanke and Gertler (1989) on inherent credit market imperfections. They apply their model to 17 countries using an error correction framework and ﬁnd a clear positive relationship between credit growth and real estate prices. Credit impacts are positive in the short run (capital availability drives up prices) but negative in the long run (presumably due to a supply effect operating in a lagged fashion). The model holds for different sub-groups of countries: credit growth seems to have a much stronger impact in the 1985–1995 period, particularly in those countries that experienced both a collapse of real estate prices and a banking crisis. They ﬁnd evidence of property prices driving11 bank lending in Australia, Denmark, Finland, France, Ireland, Italy, the Netherlands, Sweden and the United Kingdom – over half the 17 countries in the sample. Higgins and Osler (1999) provide evidence of the coincidence of equity and commercial real estate cycles for OECD countries across 1984–1993 (Table 7.1, Figure 7.3). They link the boom and bust phases to the underlying economy (pointing to above trend GDP growth across the 1980s and the tightening of monetary policy at the end of the decade as central banks responded to accelerating inﬂation), but also suggest that the rise in asset prices cannot be explained in terms of market fundamentals pointing to the existence of some form of bubble phenomenon. Their explanation of asset price bubbles relates to rational investors discounting the risk of large losses. They are aware of the risk, but consider it to be of such low probability that it can be ignored. Combined with investors who believe that 11

That is, real estate prices Granger causing general bank lending volumes.

Booms, Bubbles, Crises and Contagion

209

Table 7.1 Commercial real estate cycles in developed economies. Equity peak

Real estate peak

Rise to peak value (%)

Fall to trough (%)

Net price impact

1987 1989 1989 1989 1990 1987 1989 1990 1989 1989 1987 1989

1988 1989 1989 1990 1991 1990 1989 1987 1990 1988 1988 1987

76.4 28.2 51.9 78.0 136.2 111.0 174.0 62.3 209.3 150.5 65.6 24.5

66.2 41.6 46.3 47.0 45.5 52.5 50.9 56.8 71.9 66.1 67.6 44.6

0.60 0.75 0.82 0.94 1.29 1.00 1.35 0.70 0.87 0.85 0.54 0.69

Australia Canada Finland France Germany Italy Japan Norway Spain Sweden UK USA

Source: Estimated from data in Higgins and Osler (1999).

250 Growth to peak

200

Fall to trough

Per cent

150 100 50 0 –50

SA U

K U

Ja pa n N or w ay Sp ai n Sw ed en

y

ly Ita

an m

G er

an ad a Fi nl an d Fr an ce

C

Au

st ra

lia

–100

Figure 7.3 Real estate boom and bust, OECD countries, 1984–1993. Source: Based on data in Higgins and Osler (1999). Recall that a 100% rise is cancelled by a 50% fall in value: in the majority of countries here, asset values are lower at the end of the period than at the start.

the pricing is correct, this can sustain bubble conditions. They go on to argue that the fall in asset prices has real impacts on output, with anaemic economic growth in investment over the ﬁrst half of the 1990s as the direct consequence of asset price falls either intensifying recessions or acting as a drag on recovery.

210

Towers of Capital

Herring and Wachter (2002) discuss the frequency at which real estate and banking bubbles and crises coincide and make similar arguments to those advanced by Hendershott et al. and Davis and Zhu, in focusing on bank behaviour in exacerbating ‘excessive’ real estate price rises and on the impact of falling real estate prices in the aftermath of a bubble on bank activity. Their model of price rises rests on the assumption of short-term supply inelasticity and no ability to short sell in real estate markets. In such circumstances, optimists are likely to push prices above fundamental levels – since if prices are below fundamental value, knowledgeable investors will buy to capture any arbitrage proﬁts but, if prices are above fundamental values, knowledgeable investors do not have a mechanism to exploit the imbalance. So long as prices trend upwards, optimists can survive and proﬁt, and can borrow to fund investment provided that banks use current market values to set loan to value ratios. Market prices increase as fundamental values rise, as the number of investors (and the resources available to them) rises and as there is price uncertainty. Thus the rises in prices seen in the 1980s are attributed to ﬁnancial deregulation (which opens domestic markets to international investors and new players) allied to credit availability (in part also conditioned by ﬁnancial deregulation). Once again, this raises questions about bank lending behaviour: why should banks take current prices as the basis for lending decisions when prices are above their trend level for no sound fundamental reason? Bank decisions are driven by expectation of returns, asset values and cost of funds. Real estate price rises coincide with increases in the economic value of banks’ assets allowing them to lend more which, in turn, boosts property values in the short run. Herring and Wachter also cite ‘disaster myopia’ – the discounting of signiﬁcant default risk where this has a low probability of occurrence – and moral hazard, where individual bank agents make decisions based on short-term proﬁts and bonuses.12 As with other models, competition between banks and other ﬁnancial institutions for market share erodes due diligence and drives banks to take on higher risk projects or accept higher leverage ratios to maintain proﬁtability. In the downturn, depositors, aware of both the banks’ exposure to real estate and the impact of leverage effects, have incentives to run. This creates a liquidity problem for banks, promoting ﬁre sales, the additional supply putting further downward pressure on prices. Government and central bank deposit insurance, while protecting smaller retail investors, provides incentives for banks to take greater risks and insulates them from the consequences of adverse portfolio decisions and encourages herd behaviour (if the majority of banks have similar portfolios, the regulatory authorities may be forced to intervene, while an isolated bank may be allowed to fail). 12

See also Pavlov and Wachter (2002).

Booms, Bubbles, Crises and Contagion

211

In a parallel work, Malpezzi and Wachter (2005) suggest that property downturns often act as the trigger for ﬁnancial crises, citing the Asian crises of the late 1990s as examples. Collapsing property prices weaken the banking system and create vulnerability to bank runs and currency crises. Renaud (2003) advances a similar argument to explain the 1997 Asian ﬁnancial crisis. While accepting multiple causes for the crises that followed the Thai currency crisis, he notes the high exposure to real estate lending in those countries most affected by the crises: 30–40% of bank loan books in Thailand and Malaysia were real estate loans, in Indonesia, property’s share of the portfolio was 25–30% and in Korea, around 15–25%. Renaud suggests that real estate impacts were greater because the worst affected Asian markets were more shallow (such that new development represented a signiﬁcant increase in stock levels) and institutional structures and information weaker than in developed economies. This means the real estate markets were more prone to shocks – but also that recovery could be speedy. Renaud cites the example of the Bangkok ofﬁce market. In 1991, the total stock was around 1.5 million square feet; by 1997 this had risen to around 6 million square feet and by the end of the crisis, around 40% of this was vacant (considerably more than the original stock level). Why was there so much built? First, land owners sought to realise rent on otherwise undeveloped land; second, for many ﬁrms with land, real estate was the main source of collateral demanded by banks; third there was easy access to loan capital; fourth investing in real estate had tax advantages both through preferential treatment and by deferring taxes on proﬁts and ﬁnally there were behavioural components – a ‘trophy building’ mentality and ‘euphoria-based price expectations’. As with the other authors here, bank behaviour is seen as contributing to boom and bust, with disaster myopia, current price taking and an over-optimistic view of future demand overlaid on an absence of capital market signals for real estate debt and weak real estate information, reliant on uncertain appraisals of value. While real estate, banking and currency crises do not necessarily coincide (for example, Hong Kong experiences a real estate downturn but no banking crisis), Renaud suggests that there is a strong association between the three. Renaud (1999) also examines the real estate cycles experienced in the OECD and many middle income developing countries in the 1985–1993 period, the same cycle analysed by Higgins and Osler. His explanation is based on global deregulation but, in contrast to other authors, ascribes a signiﬁcant role to Japanese investment activity. The lowering of Japanese interest rates after the 1985 Plaza Accord led to a rapid increase in investment in land and buildings, with wealth effects that encouraged further lending and saw investors using their real estate asset base as collateral for overseas investment activity. Japanese external investment reverses radically

212

Towers of Capital

in 1990 with the ending of the bubble, as a combination of domestic and international shocks (increases in the Bank of Japan interest rates and the invasion of Kuwait and ﬁrst Gulf war, for example) hit asset prices and bank liquidity. This cannot be a complete explanation (the rapid rise in values in the main cities in Spain was not accompanied by a surge in Japanese inward investment) but offers useful insights into global contagion effects. Renaud describes common features: a rapid expansion of bank lending in general and also in the share of bank loan books assigned to property; increase inter-bank competition for market share following deregulation; poor lending practices in smaller and niche ﬁnancial institutions; real estate asset prices accelerating rapidly then reversing, creating liquidity issues, the need to rebalance loan portfolios and the need to bring banks back into conformance with local and Basel capital adequacy and solvency regulatory requirements and the development of a liquidity crisis and credit crunch. He also suggests that the restructuring of ﬁnancial ﬁrms, non-performing loan portfolios and real estate holdings created new product markets and ﬁrms of specialist asset managers. We return to this topic in the next chapter. As a ﬁnal example of the links between real estate and ﬁnancial crises, Peter Englund (1999) provides a detailed analysis of the 1989–1990 Swedish banking crisis. As before, it is argued that the newly deregulated credit markets and expansion of lending drove up asset prices from the mid-1980s. Real estate borrowers beneﬁted from favourable terms due to bank competition and also from the lending activities of ﬁnance companies (who themselves obtained capital by borrowing from banks or who had banks guaranteeing their capital raising vehicles such as investment certiﬁcates). The ﬁnance companies had more appetite for risk. Lending increased ﬁrst in the household sector and then in the corporate sector, with tax-motivated savings investment driving up both stock and real estate prices – notably in the Stockholm ofﬁce market. Although Swedish real estate had been appreciating before the mid-1980s mainly due to deregulation in rental markets, the increase in asset prices was argued to be far greater than fundamentals could support (capitalisation rates fell by 6% while real interest rates were about the same in 1990 as in 1980, such a shift would require either a complete change in risk premium or a huge increase in anticipated growth in rental income). The crisis phase came suddenly, with no real signs beyond general concerns about overheating and the currency in the ﬁrst half of 1989. In August 1989, the Swedish stock market index was 42% higher than its start of year level. However, concerns emerged in the commercial property sector: listed real estate and construction ﬁrms’ share prices fell by 25% (with an overall fall of 11% in the stock market). Bank and ﬁnancial stocks did not initially suffer. However, sharp increases in interest rates following German uniﬁcation and changing domestic macro-economic and tax policies began

Booms, Bubbles, Crises and Contagion

213

300 Rental value

250

Capital value 1985 = 100

200 150 100 50 0 1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

Figure 7.4 Stockholm office market values over the Swedish banking crisis. Source: Author, various data sources.

to expose systemic weakness. The September failure of ﬁnance company Nyckeln, heavily exposed to real estate and unable to roll over its maturing investment certiﬁcates, spread across other ﬁnance companies within days and created a liquidity crisis that led to company failure.13 Banks, while competing with the ﬁnance companies directly, were also ﬁnancing them and holding their loans and certiﬁcates as assets – and had incomplete information about their credit portfolios. Direct bank exposure to real estate lending was around 10–15% but real estate losses accounted for up to 40% of the non-performing loans and, combined with exposure to the failing ﬁnance companies, this was sufﬁcient to trigger a major banking crisis. Five of the six largest banks needed central support, one going bankrupt and passing into government ownership. Meanwhile the real estate market plunged downwards, notably in central Stockholm (which implies rapid falls in the value of ofﬁce property occupied by business and ﬁnancial services ﬁrms) – see Figure 7.4. Finally, the damage to the Swedish banking sector greatly exacerbated the Krona crisis when the Swedish currency was left to ﬂoat following the collapse of the European ERM in 1992, with an effective devaluation of 20% taking place by year end. Real estate, then, is affected by ﬁnancial crises and, through the interaction of the property cycle and debt ﬁnance, can contribute to banking crises, exacerbating slumps and downturns. The bursting of asset price bubbles has spillover effects in ﬁnancial markets and, from there, into the real economy. With bubbles and crashes coincident across nations, it 13

There are strong similarities to the UK ‘secondary banking crisis’ of the early 1970s.

214

Towers of Capital

seems likely that there will be coordinated property market downturns at an international scale.

7.5

Bubbles, crashes and IFC ofﬁce markets

There is substantial evidence that ﬁnancial markets move together across the world. Equity markets and, to a lesser extent, bond and real estate markets exhibit common patterns of movement. There are local and regional effects – particularly in real estate – but strong common factors. By implication, shocks are transmitted rapidly around the ﬁnancial system – particularly when those shocks are negative. Crises and crashes appear to be inherent features of a global capital system. There is evidence of contagion effects and coincident ﬁnancial crises in the early era of global capital, in the nineteenth and early twentieth century. However, the modern era seems to be characterised by more intense crises with more rapid transmission of shocks through the linkage of markets via advanced communications technologies. In market crises, it seems that the correlation between national markets increases – as does the correlation between different asset markets. This asymmetric relationship between assets and markets reduces some of the beneﬁts of risk diversiﬁcation approaches, whether these involve construction of mixed asset portfolios or global investment strategies. Interpenetration of markets – foreign ownership of assets, ﬁrms listing on more than one market or raising debt in different countries to their home listing, creation of investment funds with a global asset base, foreign direct investment – creates a set of fundamental linkages between national asset markets that drives co-movement and ensures that negative shocks and downturns appear near simultaneously in different markets across a region or across the world. In addition to these fundamental links, there are contagion effects where investors, faced with price uncertainty and volatility, take individually rational investment decisions that have an aggregate impact on the market – for example, provoking a bank run, a liquidity crisis or a sharp sell off of equities – that is not clearly justiﬁed by underlying market fundamentals. Contagion effects may be geographical (with impacts across nations within a region or across the whole world) or cross-market: a banking crisis provoking a currency crisis, for example, or a credit crisis triggering a property crash. International asset markets also appear to be prone to coincident periods of sustained rising prices. These do not necessarily constitute bubbles: price rises might result from structural changes in the economic drivers of value. For example, a reduction in the real rate of interest should be accompanied by rising asset prices (as the present value of future cashﬂows rises with falling discount rate) – and factors such as real interest rates may be globally

Booms, Bubbles, Crises and Contagion

215

determined. Asset market bubbles, sustained periods where prices are above their fundamental values, also seem to occur and, with globalisation of ownership, they too appear to coincide internationally – in equity markets, in residential property markets and in commercial real estate, for example. The bursting of a bubble, then, can trigger off a ﬁnancial crisis which will, once again, be coordinated across national markets. Real estate is affected directly by ﬁnancial crises. There are impacts on the demand for space (as employment falls), on rental and capital values and on the ability of investors and developers to obtain debt ﬁnance for their activities. Supply side lags, the constraints imposed by lease contracts and the stickiness of property values hamper adjustment processes. However, real estate also contributes to ﬁnancial crises – particularly banking and liquidity crises. Property is a signiﬁcant source of collateral supporting business debt. Falling real estate values reduce the borrowing capacity of industry with impacts on investment and subsequent economic growth. Falling real estate values reduce the value of loans secured on real estate and reduce the value of banks’ capital assets that determine their level of lending activity. Uncertainty as to the true value of real estate assets and liabilities creates information asymmetries that can provoke investor decisions that exacerbate market difﬁculties. Hence real estate markets have a causal impact in crises periods as well as suffering directly from the effects of downturns in other asset and capital markets: there are complex feedback loops. The impacts of coincident international real estate and ﬁnancial market booms and slumps are likely to be most pronounced in IFCs. IFCs concentrate employment in ﬁnancial services and, in particular, act as the focus for complex global ﬁnancial activity. Shocks in lending markets, in equity markets and other areas of the ﬁnancial sector directly affect the performance of ﬁrms located in IFCs which, in turn, directly affects the demand for ofﬁce space and the ability to pay rent for that space. In a boom or bubble, demand for space increases, rents rise and the consequent rise in capital values (possibly exaggerated by credit availability) triggers off investment and development. In a downturn, failing ﬁnancial ﬁrms and ﬁrm consolidations through mergers and acquisitions lead to rising vacancy rates in world city ofﬁce markets, while survivor ﬁnancial ﬁrms shed staff, depressing rental demands. Falling real estate values affect banks’ capital base and, hence, their lending activity which, in turn, creates problems for investors and developers with high levels of gearing, potentially creating a vicious circle. Further, as will be shown, a signiﬁcant proportion of international ownership of real estate has been concentrated in IFC ofﬁce markets, linking these cities together in boom and in slump. Chapter 9 will examine the mechanisms that lock city ofﬁce markets together in more detail. Before that, however, it is important to consider recent changes in commercial property markets that are signiﬁcant in

216

Towers of Capital

transmitting shocks – positive and negative – across real estate markets in different cities and countries. First, real estate investment markets, once largely domestic in nature, have become increasingly global in nature with signiﬁcant inter-regional capital ﬂows. This increases cross-holdings in real estate markets across nations and cities. Second, facilitating this globalisation has been considerable innovation in real estate investment vehicles, which has created new ways of gaining exposure to property market returns. This has allowed a much greater range of investors to build international property portfolios and to hold stakes in high value, prime, real estate assets such as world city ofﬁce buildings. However, this wave of innovation has more complex effects on the relationship between real estate and other asset market returns and on the volatility of real estate returns. Third, globalisation and product innovation are linked to a transformation in the nature of real estate ownership. The new property investment vehicles fragment ownership which has beneﬁts in terms of the spread of risk, but downsides in terms of policy coordination and transparency. These themes are explored in depth in the next chapter.

8 Globalisation, Ownership and Innovation in Real Estate Markets

8.1

Introduction: innovation in real estate investment and ownership

In Chapter 7, we saw how ﬁnancial and banking crises affect real estate markets directly and indirectly – directly through impacts on demand for space and indirectly through constraints on the supply of capital, raising the cost of debt and through increases in risk premia, pulling down asset values. Real estate bubbles may also contribute to banking and credit crises, with escalating real asset prices contributing to a surge of lending and then exacerbating problems when the bubble bursts. Property markets have traditionally been considered as segmented – both from other asset classes and from other geographical markets. As a result, local market factors should retain a greater importance relative to regional and global factors than might be the case with purely ﬁnancial assets like equities or bonds. However, in recent decades innovations in commercial real estate investment may have resulted in an increase in the signiﬁcance of international factors. Here, three interlinked changes are examined: the growth of global property investment strategies, the development of new collective real estate investment vehicles and the transformation of ownership patterns. Global real estate investment is hardly a new phenomenon. Foreign direct investment has long been accompanied by ownership of real estate assets; much of the direct capital investment in the earlier phase of global capitalism was real estate or infrastructure related, and much of the indirect ﬁnancial investment through securities and debt instruments was secured on real estate collateral. In developed markets, however, non-domestic investors have faced a number of barriers – for example, possible ownership restrictions, lack of local market knowledge, higher search and monitoring costs. Further, the costs of assembling a portfolio of global real estate assets

218

Towers of Capital

that is fully diversiﬁed within each country are prohibitive for all but the very largest of investment funds, leaving investors exposed to local speciﬁc risk. As a result, there has been a distinct home bias in real estate investment. More recently, though, it has become increasingly possible to build international property portfolios and overcome information frictions. As a result, global capital ﬂows in real estate have increased markedly. In part, that globalisation has been enabled by innovation in both real estate ﬁnance and real estate investment vehicles. The US real estate investment trust model – a tax-efﬁcient listed property vehicle readily accessible to retail and professional investors – has spread widely and almost all developed economies (and many emerging ones) now have similar vehicles. The rapid spread of tax-efﬁcient listed property companies has been accompanied by an upsurge in private, unlisted, real estate vehicles which enable investors to pool equity and debt. There are many different forms of private real estate investment vehicle, with a variety of portfolio structures and investment strategies, allowing investors to select different levels of risk and expected return and to ﬁne tune their portfolios. It is really the growth in private vehicles that has enabled the current wave of global real estate capital ﬂows. Innovation in debt structures has accompanied the new equity vehicles. These are discussed below: a critical change has been the development of securitised forms of debt – in particular, mortgage-backed securitisation which has both transformed the nature of property lending and enabled still more investors to gain exposure to the real estate market – sometimes without a clear understanding of the extent of that exposure. Globalisation and new investment vehicles have also had the effect of transforming the nature of ownership and exposure to real estate. The traditional model of investment ownership was typically characterised by a single entity – an institutional investor or a property company for example – controlling a building in entirety and letting it to one or more tenants to beneﬁt from the rental cashﬂow. Debt was raised at corporate level but secured on the real estate asset, with much of the debt being bank lending on a relationship basis. Financial innovation has contributed to a fracturing of ownership: not only might there be many ﬁrms and organisations with a stake in the building, but also the rights to the cashﬂow are also fragmented, split between debt holders and equity holders and with different risk characteristics dependent on the investment vehicle utilised. This fragmentation has implications for investment strategy and for urban planning and development policies. These three elements of change are clearly interrelated. Innovation in property investment vehicles has clearly enabled more investors to pursue regional or global investment strategies. Those same innovations in investment, allied to transformation in debt structures have contributed to the shifting patterns of ownership observed in both developed and emerging

Globalisation, Ownership and Innovation

219

markets. Equally, the growth of global property vehicles has been driven by the demands of international investors seeking a means to place capital into real estate in non-domestic markets; and that same global demand has shaped the structure of the new investment vehicles as investors seek higher returns through more aggressive portfolio strategies or use of debt. The changing pattern of ownership has been accompanied by a changing time horizon, with more aggressive return-seeking behaviour again reﬂected in the structures of the investment vehicles used. We will start by examining the innovations in investment vehicles and debt ﬁnance – which has driven the change in ownership and globalisation, then look in more detail at the development of global investment strategies. Finally, the implications of the changing nature of ownership and exposure to real estate will be considered.

8.2

Innovation in property investment vehicles

How do investors gain exposure to real estate as an asset class? The obvious route might seem to be to acquire buildings and to let them to beneﬁt from the rental income stream and any future capital growth. However, as suggested in Chapter 6, an investor seeking to assemble a commercial real estate portfolio faces a number of major obstacles. The critical barrier is the ‘lumpiness’ of real estate assets. The average value of a building on the UK IPD databank in December 2007 was £15 million, the average ofﬁce building £19 million. To assemble a portfolio that begins to diversify away speciﬁc property reach would require a level of capital investment that is beyond the resources of all but the largest professional investors. Furthermore, a directly owned real estate portfolio requires specialised management, either through an in-house team or employment of an external manager. An alternative strategy might be to invest in the shares of property companies: this should provide considerable ﬂexibility in the size of real estate exposure and, because of the greater ﬂexibility, allow an investor to change the proportion of their portfolio allocated to property. This, though, presents its own problems: property company returns appear to be highly correlated with the equity market (negating diversiﬁcation beneﬁts) and exhibit higher volatility than direct real estate market indices – which is only partially explained by the measurement issues resulting from the appraisal basis of conventional property indices. There have been attempts to create investment products that seek to solve these problems. One way would be to create a collective investment vehicle that is not publicly traded (and hence less subject to public market volatility). This would allow groups of investors to pool their capital and acquire portfolios of buildings, obtaining the diversiﬁcation they require

220

Towers of Capital

and generating economies of scale in management. In the United Kingdom, for example, unauthorised property unit trusts have allowed smaller tax exempt institutional investors to gain equity exposure to real estate through open-ended pooled investment, with the units being priced on a net asset value (NAV) basis – that is the appraised value of the real estate less the value of any liabilities (principally debt capital). The NAV basis means that the returns generated by property unit trusts are similar in characteristics to those reported in the real estate market indices. For other investors – retail investors and non-exempt organisations – similar structures have existed, but have been difﬁcult to structure in a tax-efﬁcient manner. Many pooled investment vehicles suffered from double taxation: proﬁts generated in the vehicle were taxed and then the investors were taxed again. From the mid-1990s, there has been a wave of product innovation that has attempted to solve some of the problems faced by potential real estate investors. In public markets, the tax-efﬁcient Real Estate Investment Trust (REIT) model, long established in the United States and, more recently, in Australia, has spread rapidly to most developed and many emerging markets. Perhaps more signiﬁcantly, there has been a surge in the development of private real estate equity vehicles mainly, but not exclusively, targeted at professional investors. These new vehicles have transformed the real estate investment landscape. More recently, attempts have been made to create synthetic real estate investment products based on ﬁnancial derivatives that allow investors to gain exposure to the market, to hedge their existing exposure and, critically, to short the market. Finally, the growth in securitised real estate debt products creates a different form of real estate investment. We will examine each in turn, starting with, and focusing on, the growth of the private real estate equity vehicle market.

Private equity vehicles Since the late 1990s, investors wishing to build a commercial real estate portfolio have been able to place capital in a wide range of private real estate vehicles or real estate funds as a substitute for directly purchasing buildings themselves. They allow different investors to pool equity capital to acquire assets, adding debt capital to increase the size of the fund and to access specialist investment and asset management services. Most of these vehicles are structured to be tax efﬁcient or tax neutral – the vehicles themselves do not pay tax on proﬁts but the rental cashﬂow and capital gains are passed directly to the investor whose tax position is as near identical to that of an investor owning the building(s) directly. This is vital for tax-exempt institutional investors such as pension funds, charities or endowments who would face a tax drain if the real estate fund were itself taxed and they were unable to reclaim that ‘loss’. Real estate funds and vehicles permit smaller

Globalisation, Ownership and Innovation

221

investors to acquire stakes in high-value property assets (such as shopping malls or prime ofﬁces) in a wide range of locations. Baum (2008) estimates that, at the end of 2006, private real estate funds owned around 10% of global commercial real estate by value; 16% was owned by listed real estate vehicles, the remaining 74% represents private direct investment ownership and owner-occupation of commercial space. There are marked regional differences in this distribution of ownership. In emerging markets, property ownership by listed (41%) and unlisted (37%) vehicles is much more signiﬁcant; in Asian markets, listed property companies own around 26% of real estate; at 12% the unlisted sector has a larger share in Europe than in North America (8%). Such ﬁgures are, necessarily, estimates as there are no ofﬁcial statistics, but give an indication of the importance of private real estate funds in the ownership of real estate. If owner-occupied property were excluded, the share of investment ownership would, of course, be far higher. The signiﬁcance of the private real estate fund market is a comparatively recent phenomenon. In mid-June 2008, Property Funds Research’s (PFR’s) database recorded 646 real estate funds with a gross asset value in excess of €317 billion. INREV (the European Association for Investors in Non-listed Real Estate Vehicles) recorded a smaller number of funds but a higher gross asset value. Figure 8.1 shows PFR’s estimate of the growth of the market for European private real estate vehicles, with slow stable growth across the 1980s and ﬁrst half of the 1990s replaced by explosive growth in the decade from 1997. The PFR ﬁgures imply that, over the decade to 2007, the value of real estate under management by private funds has increased by around 16% per annum, double the rate from 1987 to 1997. A number of factors fuelled this growth: the relatively poor performance of listed 350 300

1000 900 GAV €bn

Number

800

250

700

200

600 500

150 100

400 300 200

50 0

100 0

Figure 8.1 Growth in the gross asset value private real estate funds in Europe. Source: Data supplied by PFR, constant prices.

222

Towers of Capital

property during the technology and dot.com bull market of the late 1990s encouraged investors to take public market property companies private to beneﬁt from large discounts to NAV; listed property companies themselves seeking new business opportunities by offering fund management services; institutional investors seeking to increase their exposure to real estate in the aftermath of the bursting of the dot.com bubble and poor equity market performance in the early 2000s; and the general excess supply of capital (both debt and equity) seeking product that drove up asset prices in the mid-2000s. Whatever the underlying reasons, the net result was the creation of a large asset sub-market that, for many investors, became the prime mechanism for gaining exposure to real estate. There are many different types of private real estate fund, distinguished by their structure and their investment strategy. Structural dimensions include legal and regulatory status, whether the fund is open or closed, and whether or not it is ﬁnite life. Legal structures include unit trust structures, limited partnership structures, corporate vehicles and contractual agreements. Many have complex management structures and organisations and, increasingly, many are based in offshore tax havens. The structure affects liquidity and regulatory control, which, in turn, inﬂuence the type of investor who can access the fund. In general, only funds that are subject to control by an appropriate ﬁnancial regulator can be (directly) marketed to smaller retail investors, while unregulated funds are only available to professional investors. This, too, has been subject to innovation with regulated feeder funds investing in unregulated products, providing a channel for retail investment.1 For the purposes of this book, the precise structure is less important than the fact that investor capital is pooled and ownership of the rights to receive real estate cashﬂows is shared between those investors. It is worth noting, though, the difference between open-ended and closed-ended funds. With an open-ended fund (for example, a property unit trust), the fund manager can receive capital at any point and the investors can request redemption at any point. For unit trusts, the units are valued on an NAV basis (subject to a bid–ask spread). If the investors have the right to invest and redeem, this can cause problems for fund managers – particularly if the investors are inﬂuenced by current returns. If so, in a rising market, capital ﬂows into the fund, requiring the manager to buy at high prices; more seriously, if there is a strong demand for redemptions in a downturn, managers are forced to sell into a falling market: the achieved sale prices may be less than the quoted NAV and the act of selling may further depress the market, leading 1

A key issue here is the extent to which retail customers’ ﬁnancial advisors understand the nature of the products they are recommending. Anecdotal evidence suggests that many do not have a clear awareness of the risks implicit in particular products.

Globalisation, Ownership and Innovation

223

to further demands for redemption and creating the equivalent of a bank run on the fund.2 This has been a problem for German open-ended funds, a signiﬁcant segment of the European real estate market. Open-ended funds, in theory provide liquidity for investors: in practice, there may be restrictions on redemption or delays. Closed-end funds make calls for capital which close on an agreed date. Thereafter, the investors must collectively agree to any new call for capital and exit involves sale of the stake on a secondary market. Linked to open- and closed-end structures is the issue of the life of the fund. Funds may be inﬁnite or ﬁnite life. For funds with a deﬁned lifespan (typically six to ten years), a key issue is what happens at the end of the investment period. In principle, the fund should liquidate, selling assets, repaying debt and returning residual capital to investors. However, many have some provision for continuation, if market conditions make liquidation inappropriate. In addition to the structure of the fund, investment style is critical in determining the investment characteristic of a real estate vehicle. Funds are typically classiﬁed as core, core plus, value-added or opportunity funds, although the deﬁnition of, and boundaries between, these characteristics is imprecise (see Figures 8.2 and 8.3). In essence, the difference reﬂects expected return and risk. Simplifying, core funds invest in good quality, fully let real

Return

Value-added Repositioning, redevelopment, releasing

Opportunistic Distressed sellers, speculative development, financial engineering, new markets

Core plus Stable lease roll, more active management income enhancement strategies

Return: 18%+ Return:20%+ Leverage: 70%+ Leverage:70%+ Return: Return:15–18% 1619% Leverage: Leverage:60–65% 6065%

Core Fully leased prime property Return: 10–14% Return:1114% Leverage: 40–60% Leverage:30 -50% Return: 6–10% Return:8-10% Leverage: 0–50% Leverage:0 -30% Risk

Figure 8.2 Private vehicle investment styles.3 2 3

See Baum (2008). One issue with collaborative work is that ownership and origin of presentation slides becomes obscured. I am pleased to acknowledge the contribution of Andrew Baum, PFR and CBRE Investors in the evolution of this slide.

224

Towers of Capital

Opportunity 16%

Core 51%

Value added 33%

Figure 8.3 Fund styles, European private real estate funds, mid-2008. Source: Based on PFR data.

estate in mature markets: as a result, they offer relatively modest returns with low risk; opportunity funds offer very high expected returns with high risk and seek out distressed or emerging markets, undertake speculative development and use ﬁnancial engineering to generate returns; core plus and value-added funds occupy an intermediate position. In order to generate returns, the different fund styles employ different levels of debt, with core funds being typically modestly geared and opportunity funds being highly geared, often in excess of 85%. The expected returns for each fund style should compensate investors both for the riskiness of the business activity and the risks induced by the capital structure. The general trend in real estate funds has been for an increase in gearing levels and a tendency to move from core to core plus or value-added styles (Figure 8.4). Examining funds in the PFR database, average permitted gearing stood at around 50% and actual debt to gross asset value at around 33%. Although there is some volatility on an annual basis, both permitted and actual gearing increase with each ﬁve-year period from 1990 until the very end of the analysis period. Funds launched before 1995 had average debt to value ratios of 12.5% and permitted gearing of 33%; for funds launched between 2005 and 2008, average debt was close to 40% and permitted gearing 50%. INREV ﬁgures are very similar. Funds launched for the 2005– 2008 period have weighted average gearing of 49%; funds launched during 2000–2004, 33%; funds launched during 1995–1999 just 24%. In part, this reﬂected low interest rates and easy credit availability over the ﬁrst half of the 2000s. High gearing levels and rising prices contributed to strong global real estate fund returns which fuelled both increased capital inﬂows and an

Globalisation, Ownership and Innovation

225

Debt as % GAV

50 40

Gearing

30

Trend

20 10 0

e-

19

90 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08

–10

Pr

Launch date

Figure 8.4 Private real estate fund average gearing levels. Source: Estimated from data provided by PFR. Table 8.1

Leveraged returns, assuming 12% underlying property return.

Gearing level D/V Equity stake Debt level Initial property value Final property value Loan repayment Final equity One year return

30%

60%

85%

€25m €10.7m €35.71m €40.00m €11.57m €28.43m 13.7%

€25m €37.50m €62.50m €70.00m €40.50m €29.50m 18.0%

€25m €141.67m €166.67m €186.67m €153.00m €33.67m 34.7%

appetite for risk and gearing. The 2007–2008 credit crunch and falling capital values may reverse these trends (not least in that funds may be unable to obtain debt ﬁnance, particularly at high loan-to-value ratios, and where they are able to borrow, may ﬁnd lending terms more onerous and less ﬂexible). Highly leveraged funds can generate very high returns in rising markets, but investors face loss of all their capital in a falling market – although a substantial amount of the downside risk is shifted onto the lenders in such structures. As a simple illustration, consider €25 million invested in one of three funds which have a strict one-year ﬁnite life: Fund A is 30% geared, Fund B 60% geared and Fund C 85% geared. The loan facility is provided by a bank and borrowing rates are 8%. Table 8.1 shows the returns for equity holders assuming a 12% growth in the value of the underlying real estate; Table 8.2 the equivalent returns if growth is only 4%. In strong markets (where asset growth is greater than the loan rate), gearing enhances returns, but in weaker markets, gearing reduces the return for the equity holder. For the 85% debt-to-value ratio case, if real estate values fall by 10%, then the ﬁnal value is less than the required repayment: assuming the fund goes

226

Towers of Capital

Table 8.2 Leveraged returns, assuming 4% underlying property return. Gearing level D/V Equity stake Debt level Initial property value Final property value Loan repayment Final equity One year return

30%

60%

85%

€25.00m €10.71m €35.71m €37.14m €11.57m €25.57m 2.3%

€25.00m €37.50m €62.50m €65.00m €40.50m €24.50m −2.0%

€25.00m €141.67m €166.67m €173.33m €153.00m €20.33m −18.7%

into bankruptcy, the investors lose all their equity, the bank fails to make its required return and must either sell the property into a falling market or take the property onto its books. Thus, real estate lending is not an arms length ﬁnancial transaction and the bank is exposed to the volatility of the underlying property market. Funds also differ by target market. Finance theory would suggest that funds should be specialised – targeting particular sectors and geographical areas. This would allow individual investors to select their own mixture of funds to ﬁne tune their portfolio and to beneﬁt from specialist management and market knowledge. The practice has been somewhat different. INREV’s European real estate vehicle database shows that, excluding the German open-ended funds, 54% of funds were diversiﬁed by sector (with just 13% specialist ofﬁce market funds) and just over half held assets in more than one country. For many investors, a diversiﬁed fund offers a ‘one-stop shop’ – they can obtain a diversiﬁed real estate exposure with a single investment, without having to acquire detailed market knowledge of the asset class, sector or geographical markets and beneﬁt from scale economies. For smaller investors, since many funds have a substantial minimum capital investment requirement, available resources may be insufﬁcient to permit a more sophisticated specialist fund investment strategy. The idea of a one-stop shop is given more credence by the development of ‘fund-of-fund’ structures – investment funds which pool capital and then invest in other individual real estate funds. The layering of management fees and loss of control may be offset by additional diversiﬁcation of risk and the fund of fund manager’s skill or knowledge in selecting vehicles for investment. The ﬁrst fund of real estate fund products was offered in 2005 and by early 2008 over 80 products were available (Baum, 2008). The principal advantages of private real estate investment vehicles, then, are that, ﬁrst, by pooling equity from different investors, they allow smaller investors to build diversiﬁed real estate portfolios and gain access to types of real estate previously out of their reach; second that they allow smaller investors to achieve scale economies in specialist asset management and,

Globalisation, Ownership and Innovation

227

third, that they produce real estate returns that more closely track aggregate real estate return indices than do listed property vehicles – although it is not clear that that is the case for highly geared opportunity funds. These beneﬁts, allied to tax efﬁciency, have led to the dramatic growth in the number of vehicles and the value of commercial real estate managed through private real estate funds. What impact, then, does the growth of the private real estate funds market have on the underlying asset market? One clear impact is that an active private market facilitates globalisation of real estate investment. Investors can, at realistic cost, gain an exposure to property markets in a variety of markets without being exposed to high levels of property-speciﬁc risk or damaging their returns through high search, management and monitoring costs; they can even make single investments into international vehicles or fund of funds products, provided that they are prepared to accept the portfolio structure decisions of those fund-of-fund managers. Second, creating new investment avenues may serve to increase the ﬂow of capital into real estate, although a considerable amount of the capital ﬂowing into funds has been from institutional investors switching from directly owned and managed property portfolios to fund-based investments. In markets where there is already excess capital, additional buying pressures could force up prices, creating or sustaining asset market bubbles. Rising real asset prices could have supply side impacts, both by triggering new consumption or by inducing sales of owner-occupied property by corporate occupiers, to release capital for investment in their core businesses. A third impact relates to transparency in real estate markets. Private funds are under no obligation to provide public information on their performance. As a result, it is difﬁcult to establish effective benchmarks for real estate fund performance, despite the efforts of AREF, INREV, IPD or PFR to build databases and publish indices. There are signiﬁcant measurement issues which relate, inter alia, to the valuation of fractional interests in a property,4 to the impact of realisation-based management fees, to complex capital structures, to the market value of the fund’s debt and to the timing of capital calls.5 These difﬁculties, allied to lack of public information, reduce transparency. For professional investors with in-house research capacity, these may not 4

5

If ten investors own equal shares of an ofﬁce complex valued at €200 million, is each share worth €20 million? It could be worth less, if loss of management control and ﬂexibility of decision-making has a price; it could be worth more, since there are many more potential investors for assets priced at around €20 million than at €200 million. See Kutsch (forthcoming, 2009) for an extended discussion. An investor might commit $50 million to a fund on 1 January of an accounting year – but the fund may not draw it down until much later in the year, when they seek to acquire an investment. When does the real estate investment take place? The capital is allocated to real estate yet, until called down, cannot earn real estate returns. Typically, it would be placed on short-term deposit, earning lower cash returns.

228

Towers of Capital

be signiﬁcant drawbacks, but they represent a potential problem for smaller retail investors who may not be fully aware of the nature of the investment they are making. This appears to be the case for vehicles which have been advertised on the basis of real estate’s risk diversifying characteristics which, on investigation, turn out to be highly geared single property vehicles which rely on strong rental growth to generate the promised returns and are, hence, highly reliant on the point in the market cycle. It was reported that investors to one such vehicle – which acquired a single new London headquarters ofﬁce building – had lost all their equity when that building’s proposed tenant got into ﬁnancial difﬁculty and failed to take up occupancy. It seemed that many of the investors and their advisors had been unaware of the risks faced, even in a relatively strong property market. A ﬁnal impact relates to ownership itself, which appears to be more fragmented as a result of the growth of private real estate funds. Fund structures allow multiple ownership of buildings. Equity stakeholders are more distant from the building and its management than they would be if they had individual ownership and may be more interested in short-run ﬁnancial games than the long-run economic efﬁciency of the building. This suggests that the growth of real estate funds may promote more short-termism in commercial real estate. The counter-argument is that funds beneﬁt from specialist management (which can reap scale economies) and that the fund managers do have an interest in maintaining the functional qualities of the building since this will affect the exit value. However, for a ﬁnite life fund with high promised returns, decisions are likely to reﬂect the need to generate profits than any sense of long-term stewardship of the asset. Fragmentation of ownership also creates planning and urban regeneration issues. With multiple stakeholders, it becomes a much more complex coordination process to obtain collective agreement on, for example, transport infrastructure or environmental improvements that might require capital contributions or loss of control on the part of owners.

Listed real estate vehicles Investors in private real estate funds must confront problems of illiquidity – particularly in falling markets. Even for an open-ended unit trust structure, managers may struggle to redeem units in falling markets, while investors in closed-ended formats may be locked in or have to negotiate a sale both with potential buyers and other stakeholders. For investors who require greater liquidity, a public-listed vehicle offers an alternative. Listed property companies are real estate focused businesses; shareholders receive dividend payments from the proﬁts generated by real estate activities and the share price is underpinned by the value of the assets owned by the company. Research traditionally distinguishes between property trader-developers and property investment companies. Trader-developers acquire sites to construct

Globalisation, Ownership and Innovation

229

buildings then sell them on upon completion for proﬁt, or acquire existing buildings aiming to refurbish, upgrade, reposition or redevelop the properties for sale. Property investment companies acquire or develop tenanted properties, intending to retain them and beneﬁt from their rental income stream. As might be expected, developer-traders tend to use higher proportions of debt and have more volatile share price performance. Many larger listed companies have both development and investment divisions and such ﬁrms dominate activity relating to the largest property developments in major cities – including large ofﬁce developments in ﬁnancial centres.6 There are also exchange traded funds and real estate mutual funds that perform the same function in the listed market as fund-of-fund structures perform in the private market. Another signiﬁcant distinction between property companies relates to tax status. The traditional property company has been a taxable entity. This causes a potential tax loss for investors: real estate proﬁts are taxed within the company and then dividends, paid out of post-tax proﬁts, are taxed for the investor. The tax position is, thus, not equivalent to actually owning the real estate. Property companies are likely to use higher levels of debt to obtain tax shield effects, with the gearing bringing higher volatility of returns for shareholders. However, an increasing number of countries have followed the example of the US Real Estate Investment Trust or the Australian Listed Property Trust (LPT) in creating a tax-transparent listed property vehicle. The precise format of such vehicles (for simplicity we will use REIT as a shorthand throughout) varies by country, but a number of common features emerge. REITs are subject to a number of qualifying conditions: they must be focused on real estate activity with the vast majority of their income, proﬁts and assets property-based; generally they must be investment rather than trading vehicles; there are restrictions on ownership and control to ensure a spread of ownership; there may be rules relating to gearing levels and trading activity. Critically, REITs must distribute a very high proportion of their proﬁts directly to shareholders – typically 90–95% after adjustment for depreciation. Subject to meeting those qualifying conditions, REITs are typically tax transparent: the REIT is not taxed as an entity on its property activities. The distributions to shareholders are taxable, making the theoretical cashﬂow equivalent to owning the property. In practice, individual shareholders cannot control the timing of the sale of REIT assets and, hence, the great part of their capital return comes from any appreciation of the share price. Figure 8.5 illustrates the growth in REIT market capitalisation in the United States, accelerating as the result of a series of tax reforms and 6

Increasingly, major city centre developments are joint ventures between property companies, institutional investors and private funds, usually set up as a limited partnership or equivalent. The property company provides the development expertise and will generally retain a stake in the completed development.

230

Towers of Capital

500 000

$million

450 000

Hybrid

400 000

Mortgage

350 000

Equity

300 000 250 000 200 000 150 000 100 000 50 000

07

05

20

03

20

01

20

99

20

97

19

95

19

93

19

91

19

89

19

87

19

85

19

83

19

81

19

79

19

77

19

75

19

73

19

19

19

71

0

Figure 8.5 US Real Estate Investment Trusts: equity market capitalisation.

liberalisation measures in the 1990s with the creation of the Umbrella Partnership (UPREIT) structure critical since it allowed developers and private owners to transfer real estate assets into public ownership without triggering a capital gains tax sale. This measure, along with more favourable treatment of institutional investors, brought a rapid increase in market capitalisation and, with it, in the signiﬁcance of REITs as market players. The fall in market capitalisation in 2007 in part reﬂects concerns following the sub-prime mortgage crisis and credit crunch, but also results from a number of REITs being taken private – most notably Equity Ofﬁce Property Trust, which was acquired by private equity fund Blackstone for a reported $38 billion. Equity Ofﬁce owned around 50 million square feet of ofﬁce space, much of it in major cities, including Manhattan, Boston, San Francisco and Los Angeles. A graph of the growth of Australian LPTs would show a similar pattern with slow initial growth being followed by rapidly rising market capitalisation. LPTs were so successful in attracting capital from both retail and professional investors that they came to dominate prime Australian real estate markets and were pushed to adopt global investment strategies over the late 1990s and ﬁrst half of the 2000s to ﬁnd investment opportunities. More recently launched tax-efﬁcient real estate vehicles in developed economies in European and Asian markets have, after initial conversion of qualifying property companies, shown less explosive growth patterns, although this may be a function of relative immaturity of REIT markets in those countries, market timing and competition from other forms of capital targeting real estate assets. Figure 8.6 gives some indication of the global reach of tax-efﬁcient public-listed real estate vehicles, although changes happen

Figure 8.6

Distribution of tax-efficient public real estate vehicles.

Globalisation, Ownership and Innovation 231

232

Towers of Capital

so rapidly that any map is overtaken by events and many of the unshaded countries have developed mechanisms for tax-efﬁcient public market investment. The majority of developed economies have some form of REIT-like vehicle. However, the countries of the ‘South’ and the emerging economies are, as yet, less likely to have listed tax-transparent property companies: of the BRIC states, only Brazil has adopted the REIT model thus far. From a real estate investment perspective, a key question is the extent to which buying shares in a listed property company produces real estate returns. Applying the ‘duck test’ to property companies,7 they must represent a real estate investment. Their asset base consists of land and buildings; their cashﬂow is generated from rental income and the sale of property. This will be particularly true for REIT vehicles, whose distribution rules ensure that a very high proportion of proﬁts are passed directly to investors and not re-invested in the business. Empirically, however, the evidence calls this assumption into doubt. Listed real estate return indices tend to exhibit high correlation with equity indices and low correlation with direct property market indices – even where these have been corrected for serial correlation and lagging effects. Hoesli and Lizieri (2007) show that quarterly US REIT returns have a 0.43 correlation with the equity market but only a 0.18 correlation with the MIT Transaction Based Index of US commercial real estate returns; similarly Australian LPT returns have a 0.46 correlation with the Australian Stock Exchange but a weakly negative correlation with the unsmoothed PCA-IPD index. They also exhibit far higher volatility than that observed in the underlying property market indices, again even when corrected for smoothing. Econometric analysis tends to reveal long-run relationships between REIT and unlisted direct market performance, but those long-run links are masked by considerable short-term noise. Explanations for these divergent results vary. Some have argued that there is very little information content in appraisal-based property indices and that the REIT returns represent the true behaviour of property markets. Others have noted that there are gearing effects in REIT returns since, even with some restrictions on lending and no tax shield beneﬁts, most listed property vehicles have substantial amounts of debt in their capital structures. This increases volatility and increases the exposure of REIT returns to interest rate shocks. Equity markets (in principle) anticipate changes in market environment: it may be that real estate equity prices reﬂect future rises and falls in underlying asset values above and below trend, which may not be fully reﬂected in appraisals in the property market (this would amount to pricing inefﬁciency in the direct market or, more accurately, in the measurement of the market). Finally, the higher variance in REIT returns might 7

If it waddles and quacks, it is probably a duck …

Globalisation, Ownership and Innovation

233

reﬂect more general equity market volatility. For example, REIT stocks that are held in index tracker funds will tend to rise and fall with the market, while the presence of ‘noise traders’ – less-well-informed investors whose buy–sell decisions may not be based on pricing fundamentals –induce higher volatility and leave investors demanding higher returns which may in part explain discounts to NAV (Barkham & Ward, 1999). The existence of a growing REIT market provides opportunities for smaller investors to gain exposure to the real estate market and to build global portfolios of real estate securities. This is no recent phenomenon, of course: securities based on international real estate assets were traded on European Stock Exchanges in the nineteenth century. The correlations between listed property company returns and general equity indices, allied to increasing co-movement between international stocks do suggest that global stock market bubbles and crises will also be experienced in REIT and property company markets and there is some evidence that correlations are higher in the tails of return distributions and, in particular, when prices are falling sharply. To what extent does this affect the underlying property market, though? There are a number of possible mechanisms. First, strong property company equity performance may encourage ﬁrms to come to market through IPOs or to raise new equity capital via rights issues and seasoned equity offerings; it may also make it easier to raise new debt. This provides additional capital for investment in new acquisitions and projects which, in turn, may drive markets higher. In a downturn, poor performance may make it very difﬁcult to raise new capital, constraining activity. This would tend to exacerbate cyclical tendencies (particularly where property companies make up a signiﬁcant share of the investment real estate market, for example in Paciﬁc Rim markets) – which would also tend to be coordinated within regions and globally. Second, falling real estate share prices imply growing NAV discounts: these might act as a signal to the direct market and to property lenders, affecting their behaviour. Third, property companies and REITs serve as an exit route for property developers and construction companies, who sell completed developments, taking proﬁts to reinvest in future activity. If listed property company activity is curtailed through capital shortages (as might be the case with falling share prices), then developers face fewer buyers which could drive prices down and, in sharp downturns, leave the developers unable to ﬁnance their activities and at risk of bankruptcy as development loans fall due for repayment.

Real estate debt markets Real estate investors have traditionally blended debt and equity capital in acquiring property assets, while much development activity is debt funded. From the mid-1980s, with ﬁnancial market deregulation, the nature of the

234

Towers of Capital

debt market changed from a traditional model based largely on relationship lending by banks or on conventional bond issuance to a more complex market with securitised debt products, complex ﬁnancial engineering and niche lenders assuming an increasing importance. Financial deregulation also broke down national barriers and international banks became key sources of capital in both developed and emerging markets. Debt is not just a source of capital for real estate developers and investors: it is also an investment in its own right. The evolution of securitised debt markets in real estate creates a new route for gaining exposure to real estate returns – and risks. Conventional bank lending still plays a very important role in the funding and ﬁnancing of real estate activity and remains the main source of debt capital for smaller investors. In principle, banks assess the risks associated with each loan – the probability of delinquency or default, the scale of loss in the event of default, the risk of loan prepayment – and the relationship between the risks of that loan and those of other loans in their portfolio to assess their overall position. The price of the loan – the interest rate margin – should be based on that risk and the loan contract should contain a series of terms that deal with risk – loan to value ratios, debtservice coverage ratios, covenants and guarantees. International capital adequacy rules and, in particular, the requirements of Basel II require that banks formalise this process through risk assessment models that quantify default probabilities and loss given default and that are used to deﬁne the amount of capital the bank is required to hold to ensure solvency. In principle, then, loans should be correctly priced to reﬂect risk and to deliver appropriate returns to the bank’s shareholders. Writing in the aftermath of the 2007–2008 credit crunch, such assurances seem somewhat dubious. Banks operate in a competitive environment. In order to maintain market share and to lend, they must be mindful of the actions of their competitors: other banks and, increasingly, other capital market players. In conditions of plentiful capital – as in the so-called ‘capital glut’ of the ﬁrst half of the 2000s – capital is readily available for investors and ﬁrms and they can play one lender off against others to obtain better terms, lower margins, less onerous covenants. With capital availability also compressing yields and driving real asset values upwards, banks may be drawn to lending more to real estate on more generous terms. This capital availability fuelled the growth of highly geared private real estate funds described above and it permitted listed property vehicles to increase the proportion of debt in capital structures. Smaller private investors were also able to access debt funding for investment activity. The growth of lending can be readily illustrated by examining Bank of England data on loans outstanding to commercial real estate in the United Kingdom (Figure 8.7). Commercial real estate lending fell from its 1991 peak of around £38 billion in 1991 to a plateau of around £30 billion, before

Globalisation, Ownership and Innovation

235

250 000

£m outstanding

200 000

150 000

100 000

50 000

De

ce

m

be r1 De Jun 98 6 e ce m 198 be 8 r1 9 De Jun 89 ce e 1 m 9 be 91 r1 J 9 De une 92 19 ce 94 m be r1 J 9 De une 95 19 ce m 97 be Ju r 19 ne 98 De 20 ce 00 m be r Ju 20 ne 01 De 20 ce 03 m be r2 J 0 De une 04 20 ce m 06 be r2 00 7

0

Figure 8.7 Bank lending on UK commercial property. Source: Author, based on Bank of England data.

beginning to rise in the late 1990s, passing £50 billion in 2000, £75 billion in 2002 and £100 billion in 2004. Despite warnings about possible market heating, net lending continued to rise and passed £200 billion at the start of 2008, around 12% of UK regulated banks’ loan portfolios. Between December 1997 and December 2007, the total debt outstanding rose at 19% per annum. Over the same period, IPD ﬁgures suggest that UK commercial real estate values rose at around 4.4%. Even allowing for strong development and sale and leaseback and disposal by the corporate sector, these ﬁgures suggest that UK real estate experienced a substantial increase in gearing. The Bank of England data may even understate the level of real estate lending – since borrowers are classiﬁed by main business activity, real estate lending to corporates may not be identiﬁed. Moreover, many niche ﬁnance companies providing high risk, high interest rate mezzanine ﬁnance to real estate borrowers are not required to report lending ﬁgures to the Bank – many are formally established offshore for tax reasons.8 Other developed economies experienced similar growth in lending. 8

Some of the niche lenders are in effect subsidiaries of investment banks and other regulated ﬁnancial bodies. While they may be set up to be formally bankruptcy remote, failures in this segment of the market are likely to have contagion effects elsewhere in ﬁnancial services, as was the case in 1973.

236

Towers of Capital

Securitised debt and the capital markets This rise in bank lending needs to be set alongside the growth of securitised commercial real estate debt. Commercial mortgage-backed securities (CMBS) markets grew in the United States on the back of the development of residential mortgage-backed securitisation in the 1980s and evolved rapidly to become a standard feature of real estate debt markets. The essential structure of a CMBS issue is set out in Figure 8.8. An originator – for example a bank – has issued loans to property owners who are repaying those loans based on the rental income stream of their tenants (or, if the borrower is a corporate occupier, making repayments from their business income). The originator pools a number of loans and sells them to a special-purpose vehicle or SPV (usually a bankruptcy-remote company set up to securitise the loans). The SPV then issues bonds or commercial paper into the capital markets to provide the capital to pay the originator for the mortgages. The loan repayments go to the SPV and are used to pay the CMBS investors their coupon (interest) payments and to repay the capital and redeem the bonds. Bonds are rated by the rating agencies, helping to determine the coupon rate offered and are then fully tradable in secondary debt markets. The rating reﬂects the agencies’ assessment of default risk and prepayment risk in the mortgage pool and the impact that default and prepayment will have on the returns delivered to investors. There are many complexities to CMBS structures, determining how loan repayments are passed through to investors, what insurance and guarantee procedures are in place, trustee arrangements and much more. One critical feature of CMBS is the deﬁnition of different classes of CMBS – tranches – which have different rights to the cashﬂow coming into the SPV. Senior

Tenants

Liquidity

Insurance

Rent Mortgage collateral

Loan Owner

Rating Issuer Servicing

Repayment and security

Originator (lender)

Issue proceeds Payments

Investors

Figure 8.8

Basic CMBS structure.

Sale proceeds

Globalisation, Ownership and Innovation

237

tranches must be paid before junior tranches so, in the event of some default in the mortgage pool, losses will tend to be concentrated in the more junior classes of bonds. Senior classes are, thus, considered to be low risk and tend to attract very high bond ratings – much of a CMBS issue is likely to be rated AAA or AA. The tranche structure allows investors to select the risk– return combination that suits their portfolio needs. Risk-averse investors can choose the low-risk lower return AAA senior tranches, while investors seeking high returns can acquire lower-rated junior tranches or even the unrated (non-investment grade) bond and paper which have rights to any residual interest and capital payments once the rated bonds have been paid and redeemed. High-risk bonds and paper may be repackaged and mixed with other assets into other products such as collateralised debt obligations (CDOs) and structured investment vehicles (SIVs). The development of European MBS markets lagged behind North America – partly due to competitive debt products such as mortgage bonds and pfandbriefe, partly due to lack of scale and critical mass caused by national boundaries (and differences in national default and possession procedures); much of the early development was largely conﬁned to the United Kingdom. The market began to grow rapidly in the late 1990s (see Figure 8.9) and achieved sufﬁcient mass and volume of activity to create scale economies and reduced issuance costs in the ﬁrst half of the 2000s. European markets differ from US markets in a number of ways. In particular, European markets have evolved routes where ﬁrms can come directly to market to raise debt, bypassing conventional bank lending in a process of disintermediation. An early example was UK property company British Land’s 1999 securitisation which raised £1.5 billion based on 13 ofﬁces from the

90

75

Issuance Transactions

70 60

60 €billion

80

50 45

40 30

30

20

Number of transactions

90

15 10 0

0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

Figure 8.9 Growth of European CMBS markets. Source: From data supplied by Barclays Capital, Bloomberg.

238

Towers of Capital

Broadgate development on the edge of the City of London. The mortgages which backed the securitisation were in turn backed by the rents of the high-quality tenants (many of whom were international ﬁnancial service ﬁrms) in the Broadgate ofﬁces and the value of the ofﬁce space itself. As a result, much of the issue was AAA rated and the ﬁrm was able to obtain a lower overall interest rate and higher loan to value ratio than had they gone direct to a bank or issued corporate bonds. European corporate occupiers have also used the securitisation route to raise capital through structured sale and leaseback deals. Disintermediation and the ability of borrowers to go straight to market created competition for traditional real estate lenders and placed downward pressure on interest rates and loan terms. However, it also provided an exit strategy for traditional and new lenders. In the absence of a CMBS market, a bank lending on a major real estate project carries the full risk of an extremely large loan. Commercial loan books are lumpy, ‘granular’. Risk could be spread through syndication, although this added costs. Now the bank has an exit strategy: large mortgages can be securitised individually or in combination with other loans and exposure to a particular sector managed by full or partial sale. For many banks – both traditional lenders and investment banks – commercial real estate lending origination was strongly inﬂuenced by the securitisation process. The size and terms of the loan were increasingly being determined by the exit strategy and the ability to create loan book liquidity.9 More generally, asset-backed securitisation came to be an important source of working capital for lenders – with short-term assetbacked commercial paper issues critical for short-run liquidity. As the failure of Northern Rock showed, reliance on short-term capital market routes for working capital made ﬁnancial institutions vulnerable to sudden market shifts and to bank runs. What determined the rapid growth of CMBS markets? The growth requires there to be beneﬁts both for the originating bank and the investor purchasing the CMBS in capital markets. The cited advantages of securitising a pool of commercial property loans include, from the originating lender’s perspective, the following:

• a reduction of the bank’s exposure to speciﬁc risk (in the form of large single loans) or to the real estate sector concerned;

• loans taken off balance sheet, thus improving capital adequacy and solvency ratios;

• tapping of a new source of capital from a wider pool of investors – particularly valuable if conventional sources from deposits are constrained;

9

See Lizieri et al. (2001) for a more detailed discussion.

Globalisation, Ownership and Innovation

239

• proﬁts from arrangement and service fees and from the spread between the loan interest and that payable on the bonds or commercial paper; • the ﬂexibility to structure the securitisation to match liabilities, to adjust duration and, in general, to manage assets. These advantages have to be set against costs: the arrangement and underwriting costs, rating costs, credit enhancement and insurance guarantees and any requirement to over-collateralise or ring-fence proceeds as additional security. For investors, claimed advantages include the following:

• the ability to gain exposure to a market that would otherwise be subject • • • • •

to entry barriers (primarily due to high entry costs); the ability to acquire real estate exposure through debt assets that are marketable and liquid; improved capacity to tailor and actively manage investment portfolios both through the tranche structure and through exposure to real estate as a sector; the ability to invest in returns from a diversiﬁed pool of real estate loans, avoiding speciﬁc risk problems; low transactions, management and monitoring costs (particularly for rated securities); an enhanced yield when compared to similarly risky corporate bonds.

These beneﬁts result from transformations achieved in the securitisation process, with the CMBS originator performing a broking role. The bank provides a wholesale function, collecting together loans and then repackaging them for investors who beneﬁt from cost economies and reduced search and information costs. Securitisation also involves risk transformation. At its most basic, a CMBS pools together loans to create diversiﬁcation gains for smaller investors who could not achieve such beneﬁts individually. The tranche structure also disaggregates risk which can then, in principle, be distributed to appropriate investors. These transformations mean that the average return required by CMBS investors should be lower than the loan rate charged by banks to individual borrowers – and it is this margin that allows CMBS markets to operate and grow. Since CMBS are traded on secondary markets, mortgage-backed securitisation results in greater transparency in real estate lending. There are market signals that, again in principle, provide information that banks can use in shaping their lending policies. If CMBS spreads increase sharply

240

Towers of Capital

in the secondary market, that should be a warning to banks about their lending practices. In addition, it has been suggested that the securitisation process helps integrate the commercial property lending market with the wider capital markets. This reduction in segmentation was claimed to overcome periodic problems of capital ﬂow – as in the late 1980s and early 1990s ‘capital crunch’ in the United States – improve capital market monitoring of the real estate market and possibly result in lower interest rate spreads and high loan-to-value ratios for property borrowers (in essence, lenders can adopt more liberal policies knowing that the loans can be securitised and risk spread post-arrangement). The claim that problems of uneven capital ﬂow are solved by securitisation might seem somewhat misplaced in the light of the 2007 credit crunch which, for many commentators, was triggered by problems in the mortgage-backed securities (MBS) markets – albeit residential mortgage-backed securities (RMBS) and allied products such as CDOs and SIVs and credit default swaps that were tainted by exposure to sub-prime US residential mortgages. The role of MBS, CDOs and other structured products in the development of the credit crunch was to a large extent misrepresented in media coverage of the sub-prime crisis, where the focus was largely on lending practices. The liquidity problems that emerged in the second half of 2007 came more from banks using securitised debt as a way of raising operational capital and from their purchase of structured debt products as investment assets. Problems with sub-prime mortgage debt not only reduced the market value of RMBS and other products that had an exposure to RMBS. There were contagion effects in the CMBS market, partly due to loss of conﬁdence in the ratings agencies, partly due to information asymmetry. As a result, banks could no longer securitise parts of their loan books to raise capital; nor could they offer securitised debt from their investment portfolios as collateral for their short-term inter-bank lending. These both drove up short-term interest rates and reduced the supply of operational capital. This, in turn, raised lending rates and led to severe curtailment of lending and a tightening of lending terms – both as a rationing device and as an attempt to produce more saleable loans as assets. With conduit lending also curtailed, the reduced amount of debt ﬁnance available in real estate markets and the higher cost of that debt contributed to falling asset prices. Falling asset prices, in turn, led to ﬁrms and funds breaching loan to value covenants, creating a higher volume of loans in technical default and leading risk models to point to higher default probabilities. Increased default risk also highlighted bank exposure to credit default swaps with MBS as the underlying asset. It was thus the use of securitised debt as an exit strategy by banks, as a source of new capital for banks and as an investment asset in bank portfolios that locked the performance of banks to the performance of the real estate market.

Globalisation, Ownership and Innovation

241

The impact of change in debt markets The net result of these changes has been greater fragmentation and diversity of debt products. The market has evolved rapidly over the last quarter century from a situation where relationship banking and conventional bond ﬁnancing dominated to one where the decision to lend can result in a whole series of products with different maturities, returns and risk proﬁles with a variety of issuers, investors and associated service providers. Disintermediation means that borrowers can tap directly into capital markets; niche providers meet niche needs and can expand operations when conventional providers of debt capital switch their attention elsewhere, which – in principle, if not always in practice – should reduce periodic ‘capital crunch’ shortages of lending for viable projects. The evolution of CMBS markets also creates a new type of investment product which provides both a debt-based cashﬂow and an exposure to the commercial real estate market. This evolution has obvious advantages: with the separation of loans into different tranches of risk acquired by investors, there are more sensitive price signals and a greater ﬂow of information – facilitated by the closer integration of property debt into the capital markets. Securitisation and allied developments have improved the liquidity of the debt market and, in normal markets, provided an exit strategy for conventional lenders. However, the greater variety of products and providers can, perversely, make debt more widely available as competition for borrowers and market share can drive margins down and lead to relaxation of underwriting standards in certain market environments. This is not necessarily to argue that the scale of real estate debt is an indicator of a high level of systemic risk. However, by their nature, high-risk, high-return products are vulnerable to cyclical downturns in the underlying asset markets. At issue would be the contagion effects that might follow from any downside shock. That the 2007 credit crunch was triggered by problems in the securitised debt markets and by bank investment in securitised debt products illustrates both the complexities of the new structures and the risks that are inherent in the new real estate lending market.

Property derivatives For completeness, the evolving property derivatives market should be considered. Commercial real estate was arguably, until recently, the only major asset class without a well-developed derivatives market. Early attempts to establish such markets were unable to achieve critical mass or trading volume. In the United Kingdom, for example, the failure of the property forwards markets created on London FOX in 1991 blighted attitudes of

242

Towers of Capital

investors and regulators alike. Allied to regulatory constraints and unfavourable tax treatment which, in large measure, excluded institutional investors from participating in such markets, this meant that there was little activity during the 1990s, despite the work of energetic advocates from the industry, beyond the creation of a modest property index based Eurobond market in Property Index Certiﬁcates, more recently rebranded as Property Linked Notes. The UK property industry continued to seek a derivative vehicle that would facilitate strategic and tactical portfolio management and enable investors to alter their exposure to real estate quickly and without incurring high transaction costs or being exposed to public market price volatility. However, attempts were hamstrung by complications concerning regulatory requirements (in particular whether they were admissible investment assets for institutional investors), by the tax treatment of property derivative cashﬂows and by concerns about the nature of the underlying property market indices. Industry groups continued to lobby Government for a relaxation of the constraints on investment in, and trading of, commercial real estate derivatives, making progress at the turn of the century. By 2005, most of the regulatory restrictions on the development of a property derivatives market had been lifted. Commercial real estate index total return swaps have been traded in the United Kingdom since early 2005, once regulatory issues for institutional investors were resolved. During this period, a degree of standardisation of commercial terms has occurred. It is the standard form of total return swap commonly traded in the inter-bank market which is the focus of analysis in this paper. The UK market grew very rapidly, with outstanding notional principal passing the £1 billion mark by the end of 2005, £7 billion by the end of 2006 and £9 billion by the end of 2007 (see Figure 8.10). By the ﬁrst quarter of 2008, the notional value of trades had surpassed £15 billion, with over 1100 trades executed. Developments in the United Kingdom form part of a wider global move towards the creation of property derivatives. In the US market, the development of derivative products in the real estate market is also quite recent, although the ﬁrst real estate linked swap was launched in 1991. Fisher (2005) describes a total return swap announced and to be offered by Credit Suisse First Boston (CSFB) in the USA. The credit for each contract would be backed by CSFB which will play a role as counterparty in an exchange of property returns – based on the NCREIF Property Index, either total return or capital growth – and priced at a spread which could be either positive or negative, depending upon market conditions. In 2006, Real Capital Analytics and MIT announced a set of indices tracking US commercial property prices which were designed to be the basis for derivative trading. CMBS swaps have existed for some time, based on the Lehman or Bank of America MBS indices. In 2006, the S&P CME housing futures and options contracts were

Globalisation, Ownership and Innovation

243

10 9

Trades

8

Notional outstanding

Notional, £bn

7 6 5 4 3 2 1

Figure 8.10

20 Q1 05 20 Q2 05 20 Q3 05 20 Q4 06 20 Q1 06 20 Q2 06 20 Q3 06 20 Q4 07 20 Q1 07 20 Q2 07 20 Q3 07 20 Q4 08 Q 1

05

20

Pr

e-

20

05

0

Growth of the UK total property return swap market.

launched in Chicago, based on the Shiller–Weiss indices. There is an active Listed Property Trust futures contract on the Australian Stock Exchange, there are exchange traded products based on the EPRA indices, residential property derivatives have been traded in Switzerland and Sweden and commercial real estate total return swaps based on IPD indices have been traded in Australia, Canada, France, Germany and Japan. Real estate derivatives, in principle, allow investors to hedge their exposure to real estate. An investor owning a portfolio of buildings but anticipating poor property returns for a period faces a problem. If they sell all or part of their portfolio, then they face sales and marketing costs: if they then wish to rebuild their portfolio, they face substantial acquisition costs and have no guarantee that they will be able to repurchase similar or equivalent buildings to their original portfolio. Using a derivative, they could retain the portfolio but swap out property returns for an interest rate product. An active derivatives market also allows investors to short the market, to sell synthetic property returns, if they believe that the market is overpriced, creating a mechanism for arbitrage that is absent from the underlying real estate market. The balance of trading in the derivatives market and changes to swap margins should also provide additional pricesensitive information for market participants. However, at this stage, the property derivatives market is still evolving and lacks the critical mass and liquidity to have a signiﬁcant material effect on underlying real estate market behaviour.

244

Towers of Capital

Innovation in vehicles and structures: summary The most striking feature of the changing real estate environment is the rapidly growing variety and diversity of vehicles and products available in debt and equity markets. The commercial real estate market has evolved rapidly from one that could be characterised relatively simply in terms of ownership and sources of capital to a more complex, varied market with the boundaries between equity and debt increasingly blurred and more closely integrated to other capital and ﬁnancial asset markets. New forms appear and become rapidly signiﬁcant: recent years have seen sovereign wealth funds (notably from the Middle East) and Sharia-compliant investment vehicles become important. The range of private and public vehicles enables investors to select levels of risk and return appropriate to their preferences, and to structure their portfolios in a more innovative fashion. Lenders have access to new sources of capital and can use the capital markets to provide exit strategies; structuring of securitisation issues allows investors to select different levels of risk and return; and niche providers meet the gap between conventional, investment-grade debt and equity input, and can repackage such debts for risk-taking investors. These changes provide new opportunities for ﬁrms to manage their corporate real estate. Legislative, regulatory and ﬁscal changes have as much driven as constrained this developing diversity – not always intentionally. For example, changes to the tax regime for property in the United Kingdom encouraged the development of private vehicles and the ﬂow of capital offshore, creating a model that applies in other markets. This, in turn, encouraged ﬁnancial regulatory authorities to relax their stance to taxefﬁcient real estate investment vehicles that were listed and fell in their regulatory ambit. Similarly, the implementation of Basel II capital adequacy rules, while possibly accelerating the trend towards shorter maturity debt, is likely to lead to the development of specialist funds and vehicles to provide and recycle risky and non-standard loans that would lead to capital penalties in the conventional banking sector. Similarly, although there is great concern in European markets surrounding potential legislation on commercial leases and tenant security, it is likely that the market would develop new forms of tenancy arrangements with marketable investment characteristics, as seen by the ﬁxed or indexed leases used in structured sale and leaseback deals. As Lizieri and Ward (2004) noted, there are many beneﬁts from this more fragmented market in terms of efﬁcient allocation of capital. The range of private vehicles and investment opportunities available – particularly as these have been extended to permit access for private individual capital – have removed many of the entry barriers facing real estate investors. Previously, smaller investors, whether professional or private, were largely

Globalisation, Ownership and Innovation

245

excluded from participating in the performance of large, prime and trophy buildings – which, in turn, has price impacts on those assets. In principle, investors now have greater ability to diversify the property components of their portfolios and to structure and restructure them according to changing expectations, preferences and liabilities. In practice, the current prevalence of diversiﬁed, rather than specialised, vehicles hampers this process although the trend appears to be toward more focused vehicles, particularly in the listed, public real estate market. Further, many of the available private vehicles have limited – or untested – liquidity, which might hamper capital switching, and there is less diversity in the range of gearing available. Liquidity issues quickly emerged in the open-ended fund market in 2007 as many funds blocked redemptions when faced with the prospect of selling into a falling market. The integration of real estate with capital markets should bring beneﬁts in terms of pricing and transparency, particularly in debt markets. In deciding the margins on property lending, banks will be mindful of prevailing interest rate and rating shifts – particularly those that recycle debt rather than act as term lenders. However, there is a danger that rating and interest rate margins will be subject to herding behaviour. Again, the contagion effects of problems in the sub-prime and alt-A mortgage markets on commercial real estate debt products in the 2007–2008 credit crunch provide an illustration. Market concerns about the true value of securitised real estate debt essentially blocked conduit lending and conventional bank lending predicated on a capital market exit, leading to a rapid fall in credit availability and tightening of lending terms. These beneﬁts of transparency are less manifest in relation to real estate equity investment. While REIT prices are in the public domain, there is limited evidence of transaction prices in the private vehicle arena and it is difﬁcult to interpret the price signals of those vehicles that are listed and traded (not least due to tax leakage and gearing effects). As a result, there is little evidence that traded share or unit prices are inﬂuencing valuers directly. The market believes, against theory, that property derivative margins provide forward information on market returns and it is undoubtedly true that UK property swap margins turned negative well in advance of falls in reported property values. It is too early to tell whether there is market information in swap prices or whether this is simply a phenomenon that reﬂects the early development of the market and the lack of critical mass in trading activity. There also appears to be a lack of transparency in the private vehicle market. It is this lack of transparency that is perhaps the most concerning aspect of the property investment market as more private retail investors are drawn in. To what extent are such investors and their advisors fully aware of the nature of the products in which they are investing? Is there

246

Towers of Capital

sufﬁcient understanding of the risks involved and of the impact of gearing and issues surrounding illiquidity and lock in? How are investors (actual and potential) to benchmark the performance of vehicles and managers? These become critical issues as moves are made to extend the sources of property investment beyond professional and qualiﬁed investors to general retail customers.

8.3

Global real estate investment

Over the last decade, a change in the investment strategy of pension funds and other professional investors has generated an increased investor appetite for global real estate investment. The world’s top investors are going global and real estate managers have facilitated this through the creation of innovative indirect real estate investment solutions. The most popular of these indirect solutions are called real estate funds. (Baum, 2008) This extract from Andrew Baum makes clear the link between indirect investment and global property. Once barriers to international investment were removed, many professional investors sought to build global investment portfolios. However, while this was comparatively straightforward for equity and bond portfolios, the characteristics of real estate as an asset class made it extremely difﬁcult for all but the very largest of investment funds to assemble an international property portfolio that was reasonably diversiﬁed within each of the target countries. Given the heterogeneity of real estate and the high cost of individual buildings, any international portfolio that included a broad range of countries would inevitably be carrying speciﬁc risk. Investors unprepared to take that risk or without the capital to build a directly owned portfolio could buy real estate securities. Yet, while these offered diversiﬁcation beneﬁts over a general equity portfolio, investors faced higher volatility, higher correlations between national portfolios and a pattern of cashﬂow that did not mirror that expected of commercial property and that often created adverse tax positions. Thus, the development of the private equity vehicle market enabled a growth in global investment. Another factor that has contributed to the growth of international real estate investment has been the establishment of international networks of property consultants and advisors through a process of global consolidation, mergers, acquisitions and strategic alliances.10 As an example, CBRE state that they employ over 29 000 people in 400 ofﬁces spread across some 60 countries; Jones Lang LaSalle (JLL) claim a presence in over 40 countries. Such networks help to break down information barriers, increase market 10

For discussions, see de Magalhães (2001), D’Arcy and Keogh (1998).

Globalisation, Ownership and Innovation

247

transparency and contribute to the provision of more robust and consistent market data – although poor data quality remains the Achilles’ heel of commercial real estate. Improved market performance indices such as those provided by Investment Property Databank (who now cover around 20 countries) help investors ﬁne tune global strategies and provide the initial steps towards benchmarking international property market performance. While the development of new ways of investing in real estate internationally and improvements in market transparency facilitate global real estate capital ﬂows, the impact on the spatial concentration of those capital ﬂows is less clear cut. At one level, the market developments should encourage a greater spread of investment. An investor making direct acquisitions of buildings faces formidable information, search and monitoring costs and many investors sought economies of scale by acquiring large prime developments in or near major global cities. The private real estate fund route reduces this constraint. An investor could, for example, invest in a fund that specialised in central and Eastern European logistics warehouses and another in Balkan apartment buildings as part of a global strategy, relying on the local knowledge and specialist management expertise of the fund managers.11 However, the investor still requires sufﬁcient knowledge and understanding of those markets to include them in the portfolio and must perform due diligence on the managers themselves. There is thus likely to still be a tendency for capital to be drawn to those cities and asset types that are subject to greater market research and analysis and where the density of advice and consultancy is greatest. As a result investment is still concentrated both in the major world cities and in the major property sectors. It is not the intention here to provide a detailed analysis of global real estate capital ﬂows. As a snapshot indicator of the level of global activity, Real Capital Analytics record major individual commercial real estate transactions around the world. In 2007, at least 128 (26%) of the 500 largest sales involved cross-border acquisitions – 31% by value of deals.12 By value, 37% of those top 500 acquisitions were in international ﬁnancial centres (IFCs) ranked in the top 20 by Z/Yen (all bar three being ofﬁce buildings) and 12% involved top 20 IFC cross-border deals. RCA ﬁgures suggest that 114 cities had major sales of over $1 billion in 2007 – mainly the major cities of North America, Western Europe and developed Asia, but also 13 cities in mainland China, two in Brazil and two in India. Also for 2007, PFR estimate that 13% of private real estate funds are ‘global’ – that is their target markets include

11

12

Éamonn D’Arcy made this point in discussion at the European Real Estate Society conference in Krakow in 2008. These ﬁgures are an underestimate since the domicile of ownership is not always known and is sometimes masked behind a nominee local purchaser.

248

Towers of Capital

300 000 250 000

Cross-border Domestic

€million

200 000 150 000 100 000 50 000 0 2000

2001

2002

2003

2004

2005

2006

2007

Figure 8.11 European direct market real estate investment. Source data: JLL.

two or more regions – and those 278 global vehicles manage 27% of the real estate by value in the PFR database. JLL (2007) provide direct real estate investment statistics for European real estate markets that illustrate both the rise in the level of investment across the ﬁrst half of the 2000s and the increase in cross-border activity (Figure 8.11). Between 2000 and 2007 (and noting the fall in investment in 2007 in the aftermath of the sub-prime and credit crises) overall investment rose by around 18% per annum; over the same period, cross-border investment increased by 25% per annum, increasing its share of overall investment from around a third to over 60%. The introduction of the single currency clearly facilitates cross-border investment within the Euro zone by removing currency risk. Innovation in real estate investment vehicles has also played its part: JLL report that in 2007, 36% of acquisitions by value were by unlisted funds, with a further 18% attributed to private property companies. Similar ﬁgures for global real estate capital ﬂows show a 41% per annum increase in cross-border property investment between 2003 and 2007, and a 44% annual increase in inter-regional transactions, in the context of an overall growth rate of 21% per annum. Cross-border deals’ share of total investment increased from a quarter to 47% over those ﬁve years (Figure 8.12). At individual city level, the research conducted as part of the Who Owns the City project at the University of Reading provides detailed information on global ownership in the City of London (Lizieri et al., 2000a; Lizieri & Kutsch, 2006). Figure 8.13 shows the proportion of City ofﬁces on the Reading database that are owned by non-UK entities. From a relatively steady proportion of around 10% from the early 1970s to the mid-1980s, the international share increases markedly to around 20% in the early 1990s, then jumps to 47% in 1999, falling away slightly then returning to 46% by

Globalisation, Ownership and Innovation

249

800 700

US$billion

600

Inter-regional Intra-regional Domestic

500 400 300 200 100 0 2003

Figure 8.12

2004

2005

2006

2007

Global real estate capital flows. Source data: JLL.

50 45 40 Per cent

35

Other International Middle East W Europe Japan USA Germany

30 25 20 15 10 0

1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

5

Figure 8.13 International ownership of offices in the City of London. Source: Adapted from Lizieri and Kutsch (2006).

2005. Within that overall upward trend, it is possible to discern different waves of foreign ownership: the Japanese in the late 1980s and early 1990s, German funds in the late 1990s and ﬁrst half of the 2000s, the return of US investors in the 2000s. Also apparent is the increasing importance of ‘international’ ownership – buildings owned by different nationalities in partnership or by an offshore fund structure bringing together global investors. Those ownership patterns reﬂect capital ﬂows into the City’s real estate market. For a shorter time period, ﬁgures for CBRE showing quarterly investment in central London ofﬁces further reveal the importance of foreign investors in London’s real estate markets. In the ﬁve years between 2003 and 2007, non-UK investors invested £30 billion in central London

250

Towers of Capital

(a)

£billion

3.5 3.0

Non-UK

2.5

UK

2.0 1.5 1.0 0.5

(b)

1 Q

3 20

08

Q

1 20

07

Q

3 20

07

Q

1 20

06

Q

3 20

06

Q

1 20

05

Q

3 20

05

Q

1 20

04

Q

3 04

Q 20

03 20

20

03

Q

1

0.0

7.0 6.0

Non-UK

£billion

5.0

UK

4.0 3.0 2.0 1.0

1 Q

3 20

08

Q

1 20

07

Q

3 07 20

06

Q

1 20

06

Q

3 20

05

Q

1 20

05

Q

3 20

04

Q

1 20

04

Q

3 Q 20

03 20

20

03

Q

1

0.0

Figure 8.14 Office investment in (a) City of London, (b) Central London, 2003 Q1– 2008 Q1. Source: Calculated from CBRE Central London Office Market Data.

ofﬁces, 48% of total investment in that market. In the City of London, foreign investors acquired interests in ofﬁces valued at £16 billion – 53% of total investment. As Figure 8.14 shows, overseas investors have been critical in maintaining liquidity in the central London ofﬁce markets. It is this liquidity – combined with a perception of long-run demand for space – that may underpin continued investment in London ofﬁces despite their apparently poor investment performance relative to other sectors and geographical markets. JLL (2007) also report that a signiﬁcant proportion of net foreign investment in UK real estate goes to central London and, speciﬁcally, to City of London ofﬁce buildings: 35% of net investment in 2006 and 60% of net investment in the ﬁrst half of 2007 involved acquisition of London ofﬁces. In aggregate, commercial real estate investment and development still remain a largely domestic, localised activity. Barriers to the development of a truly global investment strategy are still real and considerable and

Globalisation, Ownership and Innovation

251

local market participants possess considerable beneﬁts in terms of market knowledge and understanding of the institutional structure of the local property market – including all the informal, uncodiﬁed ‘rules of the game’. However, those barriers have been increasingly eroded, ﬁrst by market deregulation, second by the creation of new ways of placing capital in real estate and, third, by improvements in the advisory, research, and consultancy services that support the investment process. As a result, from the 1990s there has been dramatic growth in global real estate capital ﬂows. The new investment vehicles should permit acquisition of a wide range of strategies and property types. In practice, however, investment still seems heavily concentrated in large, high-value ‘traditional’ assets and in the world cities identiﬁed in earlier chapters. Development activity may be somewhat less concentrated in the major cities of the developed regions, given the wave of development in mainland China, in the transitional economies and in the Middle East, although there are many examples of major international ofﬁce developments and urban regeneration schemes in world cities and ﬁnancial centres. We return to this topic in the ﬁnal chapter: however, there is one more theme to explore here – the impact of new investment vehicles and globalisation on the nature and structure of real estate ownership.

8.4

Ownership patterns

The innovations in equity and debt vehicles described earlier have signiﬁcant implications for the ownership of commercial real estate. Above all, they bring a fragmentation of ownership, with many stakeholders having a claim on a property’s cashﬂow and a say in the usage of that building. This is not to suggest that traditional forms of ownership such as owner-occupation or long-term ownership by a single landlord have disappeared – they still dominate – but rather to note the growing signiﬁcance of fractured ownership and fractional interests. The new forms will be most prevalent in larger developed markets with high-value properties, such as the CBDs of major cities, where real estate funds can gain economies of scale and where the new investment vehicles are most effective at breaking down investment barriers. For the owner-occupier, a property is held for operational purposes. At the same time, ownership provides occupational security and stability and adds to the value of the ﬁrm’s ﬁxed assets. The building can be used as security for lending (and estimates suggest a very substantial proportion of business lending is secured on property assets), with much lending on a bank–client relationship basis. Owner-occupation, though, ties capital up in real estate (which may produce expected returns lower than those from the

252

Towers of Capital

ﬁrm’s central area of business,13 and the ﬁrm must manage the property, which is not their core business activity. Moreover, property values may not be fully reﬂected in share prices and the absence of a rental market discipline makes it hard to assess the proﬁtability of operating units. These factors have contributed to the trend towards corporate real estate outsourcing (Gibson & Lizieri, 2001) and the growth of structured sale and leaseback arrangements. In these, the ﬁrm sells off some or all of its corporate real estate assets to a specialist property manager, then leases the property back on agreed terms. The manager then organises an asset-backed securitisation to raise capital, with the rental payments under the leaseback and the capital value of the property backing the securities. Managers are often private equity vehicles or specialist arms of listed property companies; thus there has been a shift away from a single owner with a long-term interest in the functioning of the property to a fragmented set of interests, many of which are short-term ﬁnancial claims removed from consideration about the building itself. While high-proﬁle structured sale and leaseback deals have been in the retail or telecoms sectors, there are many examples of ﬁnancial and business service ofﬁce deals. Similar shifts can be seen in the move away from the traditional landlord role. Traditional owners include institutional investors (life insurers, pension funds), endowments and charities, landed estates and established property companies. It could be argued that such owners have a longterm interest in the building and, hence, in preservation of the property’s value and its income-generating capacity. By contrast, a real estate fund, particularly if ﬁnite life, is focused more on short-run returns. The same is increasingly true for listed real estate ﬁrms as shareholders (and analysts) demand growth in dividend and asset base. The impacts of these shifts are debatable. Some argue that traditional landlords neglected to invest in their properties adopting an arm’s length, distant approach (this will be particularly true where long leases prevail), while the active management required to generate higher short-run returns will be accompanied by an upgrading on stock and a focus on ﬁnding the most proﬁtable tenant mix and usage of space, promoting efﬁciency gains. On the other hand, the stakeholders in real estate funds have little interest in the long-run earning capacity of the buildings in their portfolio, except insofar as this has an impact on exit values at fund liquidation.

13

It is sometimes argued that ownership of property diversiﬁes risk for a shareholder investing in a company operating in a volatile sector. However, shareholders can diversify their own risk (for example, holding a portfolio of shares split between the risky ﬁrm and property company shares). It is not the job of the ﬁrm’s managers to diversify, it is the responsibility of the shareholders.

Globalisation, Ownership and Innovation

253

100 90 80 Per cent

70

Traditional Institutions Other Specialist real estate Finance

60 50 40 30 20 10 05

03

20

01

20

99

20

97

19

95

19

93

19

91

19

89

19

87

19

85

19

83

19

81

19

79

19

77

19

75

19

19

19

73

0

Figure 8.15 Office ownership by type of owner, City of London. Source: Adapted from data in Lizieri and Kutsch (2006).

As an example of shifting ownership patterns, the City of London ofﬁce market once again provides a useful case study. The Who Owns the City studies (Lizieri et al., 2000a; Lizieri & Kutsch, 2006) trace shifts in ofﬁce ownership in the City from 1972 to 2005. Figure 8.15 shows how ownership by traditional landlords (the City livery companies, endowments, traditional estates and the public sector) declined from 37% in 1972 to less than 7% in 2005 while ﬁnancial and specialist real estate ownership rose from 40% to 65%. These bald ﬁgures mask the extent of the change – ﬁrst, in that they do not account for the change in the nationality of the owners outlined in the previous section; and second, in that the ‘specialist real estate’ segment has shifted in character from a one dominated by relatively passive traditional property companies, to one where the dominant players are real estate funds actively managing ofﬁces and employing complex ﬁnancial engineering techniques to generate the promised returns for stakeholders. The continued and accelerating growth of new real estate investment vehicles makes the deﬁnition of ownership more complex and slippery. This can be seen in relation to Canary Wharf (see Box 8.1). In terms of domicile, with equity market investment, the convention would be to deﬁne the nationality of a company in terms of its place of listing and/or its effective headquarters – even though the shares may be held, substantially, by foreign investors (a third of UK shares are owned outside the country). However, a company with a low free ﬂoat and a substantial majority of shares held by a ﬁrm outside the jurisdiction of listing might be considered a non-domestic subsidiary. Presumably, similar principles apply to unitised forms of real estate investment, although the situation is complicated where the formal head ofﬁce is in a tax haven while the de facto management is UK-based. However, co-ownership investment clubs such as joint venture limited

254

Towers of Capital

Box 8.1 Who owns Canary Wharf? Canary Wharf provides an interesting illustration of the complexities of defining ownership. The easy answer to the question would be ‘Songbird owns Canary Wharf’: the Morgan Stanley-led consortium having won the battle for Canary Wharf Group in 2004 and taken it private. US Investment Bank Morgan Stanley was part of the original consortium of banks that found itself effectively the owner of the Docklands development having foreclosed on Olympia and York, its Canadian-origin owners. The development was then sold to a consortium that included Paul Reichmann (founder of O&Y), Saudi Prince Al-Waleed, Simon Glick and US fund management interests, who (partially) floated the company in 1999. Songbird outbid another consortium led by Canadian developer Brascan – which was, in turn, supported by Paul Reichmann. But who or what is Songbird? In Songbird’s last annual results the equity interests in Songbird includes 29% held by Morgan Stanley private equity funds (we were unable to find information on the equity holders in those funds). 30% by Glick Entities (investment vehicles for original CWG consortium partner Simon Glick); 16% by British Land joint ventures; 6% by Kingdom Trust, a Saudi investment vehicle; and 5% by Vidacas Nominees which is linked to AXA (the nationality of which is equally open to question given its historical antecedents). There are four classes of shares: the B shares, representing just over 12% of the total are listed on AIM. Songbird’s consolidated balance sheet implies a debt to equity ratio of 0.781 so around 44% of the value of the enterprise is linked to debtholders (which, given Canary Wharf’s history seems relevant when considering ownership). A number of the buildings on Canary Wharf were used as the basis for two major asset backed securitisations. Thus there are rights to those properties that are held in trust for the bond and commercial paper-holders (Songbird was forced to repay some of the securitisations, with penalties, to give it the right to dispose of some of the assets). However, Songbird does not fully control the original Canary Wharf Group company. In mid-2005, after Paul Reichmann’s investment vehicle IPC Advisors had exercised warrants, the proportion of shares in CWG held by Songbird had fallen to 62% from 66%. In addition to Reichmann, Brascan holds around 20% of the share capital. A consortium of Reichmann, Brascan and Barclays Capital combined to buy one of the Canary Wharf buildings (20 Canada Square) from Songbird as part of its disposal strategy. As can be seen then, the ownership structure is highly complex and beneficial interests in the development are widely spread. Songbird has a controlling interest in Canary Wharf, is UK-registered and based and is a specialist property vehicle. But to attribute ownership as ‘UK property company’ is to simplify a highly complex situation. The EGi London Office Database run for the Docklands office market (which is wider than Canary Wharf, but dominated by it) shows 79% of ownership as being UK (92% of known, disclosed ownership) and 79% of ownership being attributed to property investment companies. This may well be correct legally but is only part of the story. Source: From Lizieri and Kutsch (2006, p. 4).

Globalisation, Ownership and Innovation

255

partnerships with small numbers of participants are more complex. The vehicle is not a taxable entity and the partners clearly hold a beneﬁcial stake in the property and exercise some form of management control or inﬂuence. The situation is further complicated by the creation of feeder fund and fundof-fund products and by the impact of asset-backed securitisations with the layered ownership and restrictions on management actions that these bring. The partners and investors themselves may be very different entities, differentiated by nationality, by type of organisation and many other dimensions including investment motivation. There are further implications of the changing nature of ownership for the quality of stock and for planning of the built environment. Shorter holding periods and a focus on maximising cashﬂow over short time horizons potentially leads to underinvestment in the built form, increasing the impact of depreciation and obsolescence. Fragmented ownership patterns (particularly where the true owners are masked behind nominee companies or tax-driven offshore structures) make negotiations on major planning initiatives and on the overall asset management of the real estate markets of major global cities much more complicated. It would be wrong to overstate this problem: in many instances there will be a local fund or asset manager with decisionmaking powers. However, it is evident that the fragmentation of ownership and the widening of the distance between the planning authorities and the end investor make achievement of optimal decisions – particularly where these require delicate negotiations and concessions – much more difﬁcult than under traditional single ownership modes.

8.5

Conclusions: innovation, ownership and real estate risk in IFCs

The traditional segmented model of real estate with buildings owned by individual property investors using a combination of their own equity and bank debt no longer represents the norm for developed real estate markets. The growth of private real estate funds as a means of pooling equity from different investors and acquiring portfolios of buildings, along with the spread of the tax-efﬁcient REIT model around the world has transformed real estate investment, removing many of the barriers to commercial property ownership for retail, private professional and institutional investors. Smaller investors can now gain exposure to large prime real estate assets such as shopping malls or CBD ofﬁce complexes; larger investors can pursue realistic global property investment strategies. Allied to increasing deregulation of real estate markets and improvements in the quality and coverage of property market research, this has led to a rapid growth in cross-border real estate capital ﬂows.

256

Towers of Capital

Real estate debt markets have also seen rapid innovation and evolution. One key change has been the development of asset-backed securitisation and, in particular, CMBS markets, ﬁrst in the United States and more recently in Europe and other developed economies. CMBS markets provide a new source of capital for property lending, either as a means for traditional lenders to sell existing loans and generate new funds for lending or, through disintermediation, by borrowers going straight to the capital markets to raise debt. New lenders have become increasingly active, including ﬁnance ﬁrms lending on more risky projects or providing mezzanine funding to increase loan-to-value ratios on real estate investments. These additional sources of capital, allied to low nominal interest rates and competition led to an overall increase in the gearing of real estate investment from the late 1990s. As outlined above, there are beneﬁts and costs to such a market transformation. By integrating the commercial real estate market more fully with other capital markets, entry and exit barriers are removed, capital can theoretically ﬂow more efﬁciently to appropriate assets and there may be gains in transparency where there are clearer market pricing signals. Fragmenting the claims to real estate cashﬂow can create diversiﬁcation beneﬁts, reducing the exposure of individual investors or lenders to particular buildings and allowing more investors to include domestic and global property assets in their portfolio. Integrating the real estate market with other ﬁnancial markets, however, may induce greater volatility and increase co-movements between property and other ﬁnancial asset classes reducing its diversiﬁcation beneﬁts; international real estate strategies and the growing importance of cross-border capital ﬂows might similarly increase the signiﬁcance of any global real estate factor relative to local return drivers. Fragmentation of ownership diversiﬁes risk but also makes coordinated local urban planning strategies more difﬁcult, while the nature of funds can lead to a shift to a more short-term investment horizon with implications for the longer term maintenance of the quality of the built environment. Fragmentation of ownership and diversity of investment forms can damage market transparency – to an extent offset by demands for more market research and performance analysis. In principle, the evolving global commercial real estate market provides opportunities for investors that are neither constrained by geographical location nor by property type. Evidence, though, suggests that global ﬂows remain heavily concentrated in a relatively small number of countries – the major developed economies in Western Europe, North America and the Paciﬁc Rim – and, within those in the major world cities of the developed world. There are many counter-examples: the massive real estate developments in mainland China and in the Gulf, the growth of commercial real estate markets in the emerging Eastern European economies. But, by volume

Globalisation, Ownership and Innovation

257

of activity, it is the existing mature real estate markets of the world cities that dominate global capital ﬂows, both for equity and debt. Given that IFCs are central to these capital ﬂows both as sources and as recipients of capital, what does this mean for risk and return in IFC ofﬁce markets? Do the evolution of the real estate market and the integration of commercial real estate and ﬁnancial markets on a global scale change the drivers of risk and return there? With ﬁnancial ﬁrms heavily exposed to the performance of IFC ofﬁce markets, are there systemic ﬁnancial risks of contagion from real estate downturns? These questions frame the next and ﬁnal chapter of this book.

9 All Fall Down: Global Financial Centres, Real Estate and Risk

9.1

Introduction

In this ﬁnal chapter, the threads from the earlier chapters are drawn together. Part I of the book outlined the evolution of the world urban system and the network of international ﬁnancial centres (IFCs) that is at the heart of it. While that network represents an evolution, not a revolution, with strong historic continuity in the domination of a small number of cities, there has been a growing integration of ﬁnancial markets and concentration of activity in the post-Bretton Woods era. Market interconnectedness has been enhanced by developments in information and communication technologies. Part II examined the operation of real estate markets in general and ofﬁce markets in particular, examining demand for space; supply and development; and investment in property. Key features identiﬁed included the importance of demand from employment in determining rental values and the inherent cyclicality in property development – with cyclical tendencies more pronounced in global city ofﬁce markets. Chapter 7 examined asset market bubbles, global ﬁnancial crises and contagion effects; Chapter 8 looked at the evolution of the commercial real estate market, with the development of new investment vehicles, new debt products, fragmentation of ownership and the growth of an active global real estate market. How do these elements interact in international ﬁnancial markets? The basic argument advanced here is simple, but has signiﬁcant implications. First, international ﬁnancial activity remains strongly concentrated in a relatively small number of IFCs. That concentration of ﬁnancial activity requires a critical mass of ofﬁce occupation and creates demand for highspeciﬁcation, high-cost space. The demand for that space is increasingly linked to the fortunes of global capital markets – which then determine ofﬁce asset values. Real estate market performance feeds back into ﬁnancial

Global Financial Centres, Real Estate and Risk

259

markets through a number of channels linked to investment portfolios, credit markets and liquidity. Those linkages and feedback mechanisms have been accentuated by developments in real estate markets, notably the development of global real estate investment, innovation in property investment vehicles and the growth of debt securitisation. The resultant interlinking of occupier, asset, debt and development markets within and across global ﬁnancial centres and the feedback between ﬁnancial and real estate markets are a source of potential volatility and systemic risk, with major policy implications. Before expanding on this thesis, it is useful to retrace the arguments made in the earlier chapters, pulling out key points that support the case for enhanced and linked volatility and risk. Next, the thesis will be set out in more detail, examining the transmission mechanisms and linkages. Some supporting empirical evidence is provided – although it is not the intention here to provide any detailed econometric or statistical analysis (an effort that would, in any case, be frustrated by lack of robust data). Finally, some general policy implications are explored.

9.2

Retracing the threads

Systems of cities and IFCs The starting point is an understanding of the idea of an urban hierarchy, that certain cities grow by performing high-level, specialised services for an extended hinterland, capturing market share from smaller cities in their shadow. As cities grow, they beneﬁt from agglomeration economies, either from specialising in certain activities or from spillover gains from a range of activities. There exists a global hierarchy of cities, at the top of which there are a set of world cities whose importance, catchment and trading areas are not bounded by the nation state. These cities play a key coordinating role in international trade, in managing a global division of labour and activity and act as political, cultural and social hubs. The importance of such cities is emphasised rather than diminished by technological developments, by improvements in communications and IT, since these allow the separation of headquarters functions from production and other operational activities, allowing the former to be spatially concentrated and the latter to be geographically dispersed. A key role played by world cities is the management of capital ﬂows. Within world cities, a small sub-set of cities plays a major and signiﬁcant role in the global system of ﬁnance – the IFCs. IFCs collect capital from savers and investors and redirect that capital to seek out returns through direct or indirect investment, through debt or equity investment. They facilitate

260

Towers of Capital

international trade and business activity by organising currency markets and guaranteeing payments. They act as centres for asset management and for product innovation. Scale economies and agglomeration economies favour concentration of high-order ﬁnancial services and, as a result, a small number of cities dominate ﬁnancial activity in equity trading, bond trading, foreign exchange activity, derivatives trading and wealth management. Although this system of IFCs and global ﬁnancial ﬂows has been seen as a recent phenomenon and dubbed the ‘new international ﬁnancial system’, there are strong historical antecedents, particularly in the late nineteenth century and the years running up to 1914, when global capital ﬂows were strong. Many of the current major IFCs played key roles in the earlier era of global capitalism. The more recent era has been enabled by ﬁnancial deregulation and liberalisation of markets and been given additional impetus by the opening up of the Eastern European and Chinese markets. Developments in infomatics – information and communications technologies – have undoubtedly facilitated a globalisation of ﬁnance, but may not have the causal role sometimes suggested. What high-speed global communication technologies do permit, however, is the rapid transmission of information and shocks, positive and negative, around the ﬁnancial system.

Real estate in global cities The global cities literature has observed the waves of property development that have transformed the built environment of IFCs. This is frequently treated as a consequence of the demands for space from international ﬁnancial service ﬁrms and the professional and technical service providers that cluster around them. However, real estate plays a more complex role in such cities. First, the availability of adequate and appropriate space for ﬁrms is a key competitive attribute of a city, not simply a derived response to demand. Further, ‘adequate and appropriate’ does not simply mean provision of large, new, technologically sophisticated ofﬁce complexes. Outsourcing of activities and the need for niche service providers (and for service ﬁrms meeting the needs of those service providers) creates a demand for a wide range of space, varying in size and quality. This favours cities with large ofﬁce markets, it favours cities with ofﬁce markets with a history and track record and it favours cities that are ﬂexible in the provision of space and are able to ﬁnance new and refurbished space. The real estate market is, hence, an important part of the agglomeration economies that promote clustering of ﬁnancial activity. Second, real estate is more than simply a space in which business activities take place. It is an investment asset, a store of value. The construction of new space requires capital, debt and equity. On completion, owner-occupiers commit business capital to the property, and raise debt against the value of

Global Financial Centres, Real Estate and Risk

261

the building; tenanted space is owned by landlords as an investment asset, that ownership funded through a mixture of equity and debt. Real estate forms a signiﬁcant part of professional investors’ mixed-asset portfolios and global city ofﬁces form a signiﬁcant part of the investment real estate portfolio. Lending to commercial real estate forms a signiﬁcant part of banks’ ﬁnancial activity and a relatively high proportion of banks’ commercial real estate loan books are secured on global city ofﬁces. Since real estate is, by deﬁnition, a ﬁxed asset – spatially ﬁxed – then the existence of major ofﬁce markets in IFCs locks capital in those cities.

Demand, supply and investment in ofﬁce markets Econometric models of ofﬁce rents are strongly driven by demand variables. That demand comes from the activities of business, ﬁnancial and professional service ﬁrms (and the administrative functions of other business sectors). Shifts in the patterns of employment and new ways of working alter occupational demand; changes in occupational demand move rental values; those rents are then capitalised into ofﬁce values. Rental levels are not only demand driven, there are supply effects, but supply is also linked to rent levels through a complex dynamic process with feedbacks. Rental models are typically dominated by demand proxies such as ofﬁce employment, but rent levels are also sticky – there is a market inertia that hampers adjustment processes. Rents are also business costs and ofﬁce rents are generally higher in the central business districts (CBDs) of global cities and IFCs – as would be expected as a result of their size and economic importance. Given progress in information and communication technology, why would ﬁrms locate in such expensive areas if they can conduct their business via telephone or the internet, perhaps retaining a virtual presence in the city (for example, by renting meeting rooms in a serviced ofﬁce or business centre)? Firms will only remain in high-cost markets if the agglomeration economies, knowledge spillovers and competitive advantage of locating there outweigh the space and other congestion costs. As a result, domestic-focused ﬁrms are increasingly driven out of the central ofﬁce markets of IFCs, as are smaller ﬁrms and those that cannot exploit the beneﬁts of clustering. Even within ﬁrms, the ability to separate functions spatially leads to decentralisation of low value-added, high-volume and retail functions along with basic data processing and back ofﬁce functions – although these may be retained strategically close to the core central activities. The net result of this process of cost-driven decentralisation is a growing functional specialisation in IFC ofﬁce markets: towards an employment structure that is tilted towards high status, higher wage, higher added value jobs; to activities that rely on formal and informal networks, innovation,

262

Towers of Capital

knowledge transfer and face-to-face meetings; and on business-to-business, ﬁnancial activity. The supporting business and professional service ﬁrms that underpin ﬁnancial activities in IFCs must become similarly specialised and focused. These shifts produce a demand for more space per worker (for status reasons, to create environments that attract and retain scarce labour, because of IT needs and because of the need for private meeting space). The shifts in ﬁnancial employment that resulted from ﬁnancial deregulation from the mid-1980s produced a substantial reconﬁguration of urban ofﬁce markets in the late 1980s and early 1990s in the majority of IFC cities around the world; with echoes of that development cycle reappearing thereafter. Real estate developers should respond to demand signals, but the lags between a shift in demand, rental movements and the delivery of completed new ofﬁces cause problems in balancing demand for and supply of space and contribute to the existence of development cycles. Most IFCs have experienced pronounced development cycles over the last 25–30 years, but property cycles are no new phenomenon. The problem confronting developers is that they must incur costs and raise capital now to construct but sell some time in the future at an uncertain price. If developers are myopic (and, indeed, if those who supply capital to developers are myopic) then they may overreact to short-run upward rental and price trends, overbuilding and triggering off over-supply, that will, in turn, create downward pressure on real rents and increase vacancy rates in cities. Recent research on developer strategies suggests that it is not necessary to assume myopia to generate building cascades and to initiate cyclical oscillations in property markets. Individually rational developers, faced with pricing uncertainty and competitor ﬁrms, can initiate and contribute to development cascades and even continue building into a falling market. There are a number of market features that make such development cascades more likely. These include high price and demand volatility which can be exacerbated by a functionally specialised employment and business structure; long lags between construction starts and completions and entry barriers for developers. All of these apply in IFC ofﬁce markets. Investment banks and other global ﬁnance ﬁrms switch between strong recruitment and labour shedding phases; IFCs are increasingly functionally specialised; since many IFCs are historic cities, site assembly is difﬁcult and the start to completion time for a major ofﬁce complex is long; and the capital requirements of such developments are very high, excluding many smaller development companies from participating. It is thus unsurprising that such cities are prone to ofﬁce booms and slumps. Ofﬁce market activity implies investment activity – development ﬁnance and investment in the completed buildings. For all the difﬁcult characteristics of the asset class, real estate forms a signiﬁcant component of professional investors’ mixed-asset portfolios. Property has (or appears to have) beneﬁts

Global Financial Centres, Real Estate and Risk

263

in diversifying away risk, due to its low correlation with ﬁnancial assets, appears to have good risk-adjusted returns, and is generally believed to act as a hedge against inﬂation. To some extent, these beneﬁts may be overstated: there are measurement issues with appraisal-based market performance indices; investors must be prepared to accept illiquidity and the high cost of commercial real estate assets makes it difﬁcult to diversify away speciﬁc risk fully. For all these caveats, there is a strong case for inclusion of real estate in the mixed-asset portfolio. Within the real estate portfolio, investors should seek to diversify their property exposure, usually using geography and property sector as dimensions. Observed investor behaviour reveals a high weighting to ofﬁce properties in large cities, despite evidence of poor relative risk–return performance. There are a number of possible explanations of CBD ofﬁces’ share of investment. It may result from institutional biases and behaviour; investors may be able to make scale economies in management and monitoring expenses due to the high value of individual investments; there may be a belief in greater liquidity and a lower risk of long-run vacancy in such markets. The availability of market research and performance benchmarks may give greater conﬁdence to investors.1 Whatever the reason, ofﬁce markets in IFCs continue to be the target for real estate investment, both domestically and globally. Global real estate investors face many barriers: in particular lack of local knowledge and lack of market transparency and higher search, monitoring and management costs than domestic players. It may be that such barriers are minimised in densely researched and actively traded IFC ofﬁce markets. Yields or capitalisation rates provide the link between occupational rent and investment activity. The yield, crudely the ratio of the rent to the capital value of the property, should reﬂect investors’ required returns and anticipated rental growth. The required return, in turn, reﬂects underlying interest rates, a risk premium relating to systematic real estate risk and a compensation for depreciation and additional management costs. Yield shifts should reﬂect changes in these fundamental components. A market that becomes more transparent, where property rights improve, should be less risky and, hence, yields should fall (that is prices increase relative to rents). This might apply, for example, to the ofﬁce markets of the transitional Eastern European economies. A city that loses market share in a key 1

It was suggested to me at the European Real Estate Society meeting at Krakow in 2008 that the reported higher volatility of ofﬁce markets in global cities might actually be a function of the greater levels of research. In smaller and more peripheral markets, thin trading and more casual research could mask true market volatility. There may be some truth in that, but infrequent transactions could equally cause sharp movements in prices, as is the case with small cap equities with a low free ﬂoat.

264

Towers of Capital

business activity as a result of some structural shift is likely to experience lower long-run rental growth and, hence, yields should rise. But many yield shifts appear to result as much from ﬂuctuations in investment demand as from structural change or core interest rate movements. The price should reﬂect the building’s fundamental value, but it seems that periodic ﬂuctuations in capital ﬂows do affect asset prices. An excess supply of capital, allied to positive sentiment towards real estate may trigger property price rises and initiate a pronounced real estate cycle. The price rises may trigger an excess supply response and, from that, drive subsequent rents lower. Ofﬁce markets appear to adjust back to equilibrium pricing – but at a slow rate, with considerable inertia and with cyclical behaviour.

Financial market bubbles, shocks and crises Real estate investment activity needs to be placed in the wider context of the behaviour of asset markets. In the last decades, such markets appear to have been prone to rapid price escalations – bubbles – and sudden shock falls. Moreover, these pronounced rises and falls in asset values appear to be increasingly coordinated across countries, regions and internationally. This is perhaps most apparent in relation to equity market indices, where correlations are increasing over time, where sector characteristics are increasingly more important than nationality in explaining returns and shocks in one index appear to be transmitted around world markets nearly instantaneously. Bond prices exhibit similar trends. The behaviour of real estate, as might be expected from the importance of location, has tended to have a more local component, but more recent studies have detected common global and regional property factors. Furthermore, sharp real estate downturns seem clustered in time. Such co-movements in asset markets have been attributed to technology and, again, to the new international ﬁnancial system which has served to integrate markets. However, the ‘casino capitalism’ explanation is weak on causal mechanisms and weak on history. There is evidence of similar contagion effects across ﬁnancial markets in the nineteenth and early twentieth century early era of global capital. Major banking crises triggered by global lending and investment were rarely conﬁned to a single country or ﬁnancial centre. Technology does facilitate very rapid transmission of shocks and it may be that recent contagion crises are deeper but shorter lived than their precursors; technology allows near-instantaneous trading and rapid spread of information; it also, through electronic media, makes shocks much more visible. In analysing common movements of asset prices, it is important to distinguish between fundamental drivers (common risk factors, common ownership across countries) and contagion or spillover effects. For the latter

Global Financial Centres, Real Estate and Risk

265

to occur, impacts are secondary, linked to changes in investor behaviour, reaction to information (or lack of it) and capital ﬂow impacts. With contagion, return correlations are higher in crisis phases than in more stable, tranquil periods. Factors contributing to contagion and spillover effects include: information asymmetry, which can cause herd behaviour as information cascades from one market to another; a sudden information shock – a real economy or ﬁnancial event or unexpected price and growth sensitive announcements; inward surges of foreign capital followed by sudden reversals; and the existence of highly leveraged investors common to a number of countries. These factors may be sufﬁcient to trigger rapid asset price falls, bank runs, currency instability and liquidity crises. The impact of these effects can spread rapidly across sectors of the economy and across countries. Asset markets also seem to be prone to rapid price escalation that cannot be explained by underlying fundamental economic factors: asset market bubbles. As with crises, it is possible to identify factors that can contribute to the development of a bubble. They include widespread uncertainty as to the true fundamental price (that is substantial differences of opinion between optimists and pessimists); initial good asset market performance (to encourage optimists); supply side inelasticity (so that rising prices do not trigger new assets coming to market) and problems in shorting the market (so that arbitrageurs cannot easily take contrary positions). Further, ready availability of credit, allows optimists to expand their investment portfolios, driving out the pessimists. Credit-based explanations suggest that lenders, with information uncertainty, misprice loan risk and relax risk controls in the face of competitive pressures. Such models still require a behavioural component: belief in a ‘new paradigm’ or that a structural change has occurred; disaster myopia, a failure to account for the low probability of large losses; money illusion. Real or ﬁnancial shocks can burst a bubble with systemic effects through default, curtailment of credit and a rapid price adjustment to the consensus view of the pessimists. Commercial real estate (and real estate in general) has many characteristics that would suggest it should be prone to bubbles. The thinly traded private property market, the lack of robust public performance data, the presence of information asymmetry, few opportunities to short the market2 and a heavy reliance of debt in capital structures are all pre-conditions for the development of asset bubbles. So too is the institutional structure of the private market with the role of agents and brokers advising both buyers and sellers as to the appropriate price, creating the possibility of common shared errors. Against this, the lumpiness of private real estate and its high cost creates signiﬁcant entry barriers for smaller investors – which might 2

At least until the recent development of commercial property derivatives contracts.

266

Towers of Capital

serve to exclude noise traders who can be inﬂuential in equity markets and small depositors who can trigger bank runs. High transaction costs and illiquidity – which contribute to thin trading and lack of transparency – also act as an inertial force. Nonetheless, there is evidence of periods of rapid real estate price escalation that appear to be bubbles. In ﬁnancial market crises, real estate suffers – and, in particular, ofﬁce markets suffer. Firms shed staff, depressing occupational demand, ﬁrms fail increasing vacancy rates; which places downward pressure on rents (since supply inelasticity means that buildings cannot readily be withdrawn from the market). Falling rents depress values and damage investment performance; curtailment of credit reduces investment demand, pushing prices down further. These difﬁculties feed back into the ﬁnancial system: investors with real estate in their portfolio perform poorly; falling capital values affect the security for bank lending not only from real estate investors but also from the corporate sector. As banks face increasing numbers of loans in breach of covenant, higher default risk and non-performing loans, they are forced to reduce lending further, driving interest rates up creating a vicious spiral. There is increasing evidence that real estate can play a more causal role in ﬁnancial and banking crises, with the interlocking of credit, business and property cycles creating systemic risk. A sudden real estate shock can thus trigger banking crises which then spread to other areas of the ﬁnancial and real economy. Examples cited include the 1973 UK banking crisis, the late 1990s East Asian currency and banking crises, the Japanese bubble economy, the 1989–1990 Swedish banking crisis and, albeit from the residential sector, the 2007 sub-prime crisis.

Innovation in commercial real estate markets The two decades preceding the 2007 credit crunch saw considerable transformation of the commercial real estate market: with rapid product innovation, increased use of debt instruments, a wave of globalisation and shifts in the nature of ownership creating a markedly different business environment. These interrelated changes have many implications for the linkage between real estate and the ﬁnancial markets, property’s exposure to ﬁnancial market crises and to the distribution of real estate risk and return amongst investors. There are clear efﬁciency gains from the innovations as barriers to trading and investing in real estate are removed; but there may also be a greater vulnerability to periodic shifts and cyclical behaviour. Innovations in the structure of real estate vehicles were largely aimed at remedying some of the disadvantages of real estate as an asset class: illiquidity and high transaction costs; lumpiness and high costs leading to entry barriers and problems in constructing diversiﬁed property portfolios; and the need for costly specialised management. A wide range of fund and

Global Financial Centres, Real Estate and Risk

267

vehicle products have emerged in both public and private markets, notably the spread of the tax-efﬁcient public market Real Estate Investment Trust (REIT) model across the majority of developed economies and, crucially, the development of specialist private real estate funds. Private real estate funds come in a wide range of formats, open- and closed-ended, ﬁnite life and continuing, onshore or offshore, and with a wide range of investment strategies and risk/return characteristics. Neither format solves all of direct real estate investment’s problems. REIT-like vehicles provide liquidity and transparent pricing but bring exposure to equity market volatility. Real estate funds give tax-efﬁcient real estate returns (albeit geared returns), access to diversiﬁed portfolios and specialist management but raise issues regarding transparency, performance measurement and illiquidity. The growth of fund-of-fund, mutual fund and exchange traded fund products provide smaller investors with a diversiﬁed exposure to REITs or private funds. Finally, the beginnings of a property derivatives market opens up routes to obtain a synthetic real estate exposure and creates the opportunity to short the market. Developments on the debt side of the real estate market should not be neglected. There has been an overall increase in gearing in the property sector with real estate funds and many listed companies taking more aggressive, leveraged positions. That credit was supplied by traditional lenders, through use of ﬁnancial engineering techniques, through niche lenders providing mezzanine funding and higher risk loans and, above all, through the growth of debt securitisation. Commercial mortgage-backed securitisation provided lenders with an exit strategy and more liquidity for their loan portfolios and tapped into new sources of capital, increasing the supply of credit. New players in the market used the mortgage-backed securitisation route to capture business share and real estate ﬁrms could go straight to the market to raise debt. This process of disintermediation was used not only by property companies but also by corporate owners of real estate raising capital through structured sale and leaseback and outsourcing deals. CMBS and other forms of property debt also need to be placed in an investment context. They increasingly formed part of the asset portfolio of pension funds and other institutional investors, investment banks and deposit banks, hedge funds and retail investment vehicles. Real estate debt products also increasing served as underlying security, as collateral for bank operational borrowing underpinning asset-backed commercial paper programmes and as the underlying asset for derivative products such as credit default swaps. The new vehicles and market structures removed many of the entry barriers to commercial real estate, spread ownership and interests in property more widely and distributed risk. Risk distribution is a two-edged sword: it reduces individual investor exposure to large speciﬁc risk but it also links many more investors and organisations to real estate risk factors

268

Towers of Capital

and the property cycle. The risk–return characteristics of real estate have increasingly been separated and repackaged, sold to investors with different appetites for risk. There should be efﬁciency gains from this process compared to the bundled risk–return in the traditional real estate process, but it creates a high-risk segment of the asset class which is particularly vulnerable to shocks and the downward phase of the cycle – with possible contagion effects then spreading to less risky parts of the market. The plethora of products adds layers of complexity to a market that is already less than transparent. It is possible that investors in the new market environment were less aware of the fundamentals of the underlying asset market than was the case in a market dominated by direct acquisition and ownership of buildings. The efﬁciency gains and the additional risks must be balanced against one another. The new market environment also supported and encouraged the growth of global real estate investment strategies. International real estate investment is no new phenomenon but, from the mid-1990s, there has been a rapid increase in global capital ﬂows, largely enabled by capital availability and facilitated by the growth of private real estate funds. Private funds remove many of the obstacles to foreign property investment – the lack of local knowledge and specialist management expertise, the problems and costs of diversifying portfolios within the target countries, the high search, information, management and monitoring costs and the difﬁculties of obtaining scale economies in investment. Global investment was also aided (and even driven) by the growth of international property advisors and consultants and by improvements in market information and performance benchmarks – although these last remain a key weakness in real estate markets. International exposure also came from investment in listed REIT securities and in real estate debt products and from direct lending by the subsidiaries of global banks.3 The impact on foreign ownership of property assets is clear – but very unevenly distributed. While the existence of highly specialised funds means that, in principle, an investor can gain exposure to Balkan retail property or Eastern European logistics, there remains a clear focus on prime property, on large retail outlets and on major ofﬁce buildings, particularly those in the CBDs of global cities. It is in those markets – partly due to the concentration and mass of prime ofﬁce space – that foreign ownership is most marked. Thus, in the City of London in 2005, some 45% of the equity interests in ofﬁce space were owned by non-UK investors, who play a key role in maintaining liquidity and supplying capital to the market.

3

It is perhaps worth raising the issue of the nationality of such banks – is UBS really Swiss, is Deutsche Bank really German?

Global Financial Centres, Real Estate and Risk

269

Innovation in investment vehicles and in real estate ﬁnance and funding also changes the nature of property ownership. Interests in buildings, claims to property cashﬂows are increasingly fragmented and dispersed. Corporate owner-occupiers outsource property provision or sell buildings into structured sale and leaseback vehicles which distribute claims to the cashﬂow widely. Real estate funds pool equity and debt capital to assemble property portfolios with many investors having rights to receive rental income and capital growth from individual buildings. Limited partnerships allow groups of private investors to acquire buildings that would be beyond their individual capacities. There are potential impacts on the maintenance of the physical stock of space. Buildings come under the control of specialist asset managers with expertise and resources to work the stock. This must be set against pressures to generate short-run returns from investors in ﬁnite life private funds and from the shareholders of listed property companies and from the distance created between actual property and individual investor that may make a belief in long-term stewardship and sustainability of the building less prevalent. Fragmentation of ownership may also have implications for planning and urban regeneration, making coordination of activity more complex.

Real estate, ﬁnancial markets and IFCs Innovations in commercial real estate markets have created stronger links between property and other ﬁnancial asset markets. The links were always there – real estate investment requires capital and hence is tied into credit cycles, occupational demand rests on the growth or contraction of ﬁrms and hence is locked into business cycles. Traditional models of property ownership, ﬁnance and funding, though, contributed to a separation of real estate from the ﬁnancial asset classes, while the lags between demand signals and supply side responses meant that property, business and ﬁnancial cycles typically did not coincide. The cycles were connected, the links between ﬁnancial markets and real estate were there. The linkages in the new market environment, though, are stronger and more immediate and, with globalisation of real estate investment and funding, are much less spatially constrained. Commercial real estate markets are less segmented and less local than they have been historically. This process of integration is perhaps most apparent in ofﬁce markets. Occupational demand for ofﬁce space comes from ﬁnancial, professional and business service ﬁrms whose performance, at least in developed economies, those economies that are strongly service-oriented, is closely tied into the rise and fall of ﬁnancial markets. Ofﬁce markets, too, are more spatially concentrated than the industrial and retail sectors and, with the urban hierarchy, those ﬁrms that are most strongly inﬂuenced by ﬁnancial market and business cycles are typically based in the CBDs of the largest cities

270

Towers of Capital

and occupy large, complex, high-value ofﬁce buildings. This process is most pronounced in global cities and in IFCs, where that concentration of activity is greatest, where there is most functional specialisation, where ﬁrms’ orientation is international, not domestic. And it is in IFCs that the processes of innovation in real estate markets are at their most intense, where the fragmentation of ownership and the dispersion of claims on real estate cashﬂows are most evident and which are the focus for much of global real estate investment activity. What, then, are the wider implications of the transformation of IFC ofﬁce markets?

9.3

Market integration, globalisation and systemic risk

The development and growth of IFCs has driven the creation of large ofﬁce markets that are increasingly dominated by ﬁnancial ﬁrms – both as investors and as users of that space. That combined investment and occupation (exchange and use) function creates risk – risk that is magniﬁed both by developments in commercial real estate markets and by the nature of the global ﬁnancial system. The essence of the argument here is an extension of that found in Lizieri et al. (2000a) who suggest that the distinction between the funding of real estate development, ownership of real estate as an investment and occupational of property has become blurred. They describe an integration of property and ﬁnancial markets (Figure 9.1 extends their characterisation of that integration) and argue that this integration can create systemic risk as shocks in one area of the property market are transmitted throughout the system. A downturn in demand for space – perhaps from a downward shock in ﬁnancial markets – may lower rents and,

Financial sector firms Funding vehicles

Occupation

Ownership

Debt Letting/owner-occupation Developer property companies

Construction

Office

Sales

Private fund

Debt exposure Equity ownership

Figure 9.1 Integrated property–finance market relationships. Source: Adapted from Lizieri et al. (2000a).

Global Financial Centres, Real Estate and Risk

271

as a result, depress capital values. Falls in the value of real estate damage the performance of property held as an investment and increase the risks of property loan default and, hence, lower the value of debts held as assets. This damages the performance of ﬁnancial ﬁrms; retrenchment and contraction lowers occupational demand. A similar story can be told in most major ﬁnancial services markets – and not just in the largest global centres. For example, I examined the concentration of ﬁnancial services in the core ﬁnancial district of Toronto (Lizieri, 1995), where the same pronounced ofﬁce construction boom and slump and the adverse effects of property developer failure and non-performing real estate debt occurred in the early 1990s. Tracing the ownership and debt structures of buildings around Bay Street revealed tight interdependence between real estate and ﬁnancial services. Toronto provides a good example of a regional ﬁnancial service centre, dominating Canada and performing key niche services globally. Toronto has the third largest downtown ofﬁce market in North America, after New York and Chicago. Despite the early 1990s ofﬁce market downturn, strong growth in ﬁnance, insurance and real estate (FIRE), business and professional service employment from 1993 to 1999 meant that 62% of downtown ofﬁce employment (and 41% of total employment) was in ﬁnancial and business services, over a third being FIRE employment. As an indicator of domestic concentration and functional specialisation, 58% of Canadian mutual fund management and 78% of mutual fund professional services were located in Toronto – compared to just 25% of mutual fund retail distribution (GHK International Ltd, 2000). These effects will be at their most pronounced in international ﬁnancial service centre ofﬁce markets. First, the size and complexity of developments demand complex ﬁnance and funding arrangements provided by the major banks, ﬁnance houses and institutional investors. Only the largest banks will be willing to advance sufﬁcient capital to construct a major ofﬁce complex. For the largest projects, syndication will be needed – best arranged in IFCs, which cluster together banks and ﬁnance houses willing to lend. For capital market lending, the investment banks and specialist conduits with the expertise to arrange bond issuance or securitised debt programmes are based in IFCs as are the debt markets and investors seeking such products. Secondary trading of these debt products happens in IFCs, with banks playing a key role in broking deals and analysing the risk–return characteristics of the bonds and securities. Debt capital is also critical in real estate investment: both the listed property companies and private real estate funds active in the ofﬁce markets of global cities are leveraged investors – with many private property vehicles increasing their gearing ratios substantially from the mid1990s. That long-term funding again comes largely from the major banks, the capital markets and from niche specialist providers of real estate capital who are heavily concentrated in the major ﬁnancial centres of the world.

272

Towers of Capital

Table 9.1 Office space in the City of London: type of occupier.

Finance insurance and real estate Business and professional services Public and charitable Other uses

Core city (%)

City fringe (%)

All City (%)

57.2 27.5 4.1 11.2

25.2 36.8 14.4 23.6

46.7 31.4 7.3 14.6

Source: Adapted from Lizieri and Kutsch (2006).

Now, those same ﬁnancial ﬁrms that are providing capital for the supply of space are the occupiers of space in IFC ofﬁce markets, either as owners or, more generally, as tenants. Thus rents and capital values are linked to the fortunes of international ﬁnancial ﬁrms and their demand for space. International ﬁnancial service ﬁrm employment is volatile. The need to compete and capture market share brings international ﬁrms into IFCs seeking space. Firms grow aggressively in boom periods and then shed staff in downturns. Firm failure and takeovers of weakened companies create vacancies and lead to rationalisation of space requirements. Fainstein (1993) cites the implications on demand for space in the New York ofﬁce market of the 1991 merger of Chemical Bank and Manufacturers Hannover, driven by their exposure to under-performing real estate returns. Seventeen years later, the fall of Bear Stearns, in part through exposure to mortgage-backed securities and structured real estate debt products, may have similar implications and there is emerging evidence of ﬁnancial ﬁrms shedding space or putting moves on hold in many ﬁnancial capitals in the wake of the 2008 credit crunch. In IFCs, this volatility of demand is intensiﬁed because of the functional specialisation that is so much of a feature of the ofﬁce markets of global ﬁnancial centres. Table 9.1 shows the proportion of ofﬁce space in the City of London occupied by ﬁnancial, business and professional service ﬁrms. Many of those service ﬁrms are located in and around the City to service the needs of the ﬁnance ﬁrms who occupy over half the space in the core areas of the City. Similarly, Lizieri et al. (2000b) estimate that 69% of ‘prestige’ ofﬁces in central Frankfurt were occupied by banks or associated ﬁnancial ﬁrms. In high ranking ﬁnancial services markets, functional specialisation is more intense: there are beneﬁts of a core location for activity involving complex products and services where there is a premium on the management of risk, a need for face-to-face contact for knowledge transfer or where product innovation is critical.4 High added value business involving relationship banking, investment and corporate banking activities, asset management, primary market business, product innovation, front ofﬁce functions gain 4

Gordon et al. (2005).

Global Financial Centres, Real Estate and Risk

273

most from the agglomeration economies and spillovers of a central location. The demand for space for such activities will tend to drive out wholesale functions, back ofﬁce and routine administrative functions, retail and domestic focused ﬁnancial services and lower added value tasks. This links occupational demand more tightly to the state of global ﬁnancial markets and further reduces diversity in employment structures. Completing the integration of the elements of real estate markets in IFCs, it is those same ﬁnancial ﬁrms that invest in the ofﬁce buildings in IFCs – directly by acquisition for their investment portfolios, indirectly through investment in funds acquiring buildings, by holding shares of the major property companies owning the buildings or by investing in the securitised debt products whose underlying cashﬂow and security is based on the ofﬁce buildings. Many of the property companies and private real estate funds also have their headquarters in major ﬁnancial centres – or are managed from ﬁnancial centres, where the notional head ofﬁce is in a tax haven. Those real estate investments are signiﬁcant parts of the asset base of the ﬁnancial ﬁrms and act as collateral for their operational activities including property lending. A fall in ofﬁce values damages asset returns and the performance of ﬁnancial ﬁrms. A fall in ofﬁce values increases the default risk of debt securities, which drives up the required return and hence reduces the present value, the market value of the debts as asset. This curtails the ability of banks and other ﬁnance ﬁrms to use the debt as collateral for their operational borrowings – which affects liquidity and constrains lending. Less lending puts upward pressure on interest rates, which feeds through into discount rates and further depresses property values. Thus the occupier, supply and investment markets are locked together. The innovations in real estate markets described in Chapter 8 interact with the integration processes set out above. The growth of global real estate funds has permitted an expansion of international property investment amongst professional investors. With a signiﬁcant proportion of funds invested in ofﬁces, and in IFC markets, this has served to increase exposure to such markets. At one level, risk has been dispersed, not concentrated. With entry barriers to investment imposed by the high value of IFC ofﬁce stock, a direct building ownership model excluded many small investors and, with a domestic focus, exposure to speciﬁc markets was highly concentrated. With a fund model, exposure is much more dispersed. However, the great majority of investors in private real estate funds are professional investors (who are, typically, located in ﬁnancial centres) or high net wealth individuals (many of whom are resident in global cities). Thus exposure has been geographically dispersed, but across different global cities and ﬁnancial centres. Diversiﬁcation beneﬁts thus depend on the performance of the ofﬁce markets in different IFCs being uncorrelated. Moreover, the strategic management of private real estate funds remains strongly concentrated.

274

Towers of Capital

Table 9.2 Major property deals in 2007: location and type of activity. % of deals in top 20 IFCs Top 250 Top 500 Top 1000

47.39 50.45 45.67

% of deals % of deals % of office deals % of cross-border office cross-border in top 20 IFCs deals in top 20 IFCs 43.89 50.64 49.82

42.28 47.65 43.47

74.18 71.17 62.77

47.39 50.45 45.67

Note: Figures represent percentage of $ value in each category. Source: Estimated from data supplied by Real Capital Analytics.@2007 Real Capital Analytics, Inc. All rights reserved.

To conﬁrm the importance of IFCs in global real estate investment, it is worth expanding on the Real Capital Analytics’ statistics cited in Chapter 7. RCA record details of the top 1000 property transactions in 2007 (deals ranging from $134 million to $5.4 billion with a total value of $319 billion). As Table 9.2 shows, 46% of the top 1000 deals took place in IFCs ranked by Z/Yen in the top 20: 50% of the top 1000 deals were ofﬁce investments, and over 60% of those ofﬁce deals were in the top 20 IFCs. Almost threequarters of the ofﬁce transactions that were in the top 250 deals by value took place in leading IFCs. In 2007, RCA report $32 billion of deals involving cross-border, ofﬁce transactions in top ranked IFCs.5 The ofﬁce ﬁgures exclude deals designated as development sites, although since a number of these were in IFCs, they could well represent ofﬁce projects. RCA also published statistics ranking ﬁrms and funds in terms of their real estate trading activity in 2007 (based on all property and portfolio sales of $10 million or more). It is possible to trace the headquarters of those ﬁrms and funds. Of the leading 50 ﬁrms, 31 were in Z/Yen’s top 20 ﬁnancial centres, those ﬁrms responsible for 75% of the transactions; 60% of funds in the top 100 were based in the top 20 ﬁnancial centres and were responsible for over 70% of deals by value. As one moves down the list, more ﬁrms based away from ﬁnancial centres – and even major cities – appear, but even so, half of funds in the top 500 are based in the top 20 IFCs and those funds were responsible for over 62% of deals by value. This represents a considerable concentration of activity – a very signiﬁcant proportion of the real estate ﬁrms and funds active in the global property market are themselves based in IFCs. This spatial clustering reﬂects, in part, the high proportion of activity 5

These proportions are conservative. The Z/Yen top 20 includes Jersey and Guernsey – which will have few large property transactions – and excludes a number of large cities that would feature on many lists of leading ﬁnancial centres. For example, inclusion of Shanghai would result in a substantial increase in the proportion of deals occurring in IFCs. The Z/Yen ranking is used for consistency. Similarly, trades have been classiﬁed as cross-border only where the headquarter of the acquirer is clearly in another nation. It was not always possible to determine the location of the buyer and some ‘domestic’ buyers may be subsidiaries of foreign investors.

Global Financial Centres, Real Estate and Risk

275

that takes place in those cities but, above all, reﬂects the need to obtain capital for trades – with both equity and debt providers heavily concentrated in IFCs or drawn to them to ﬁnd investment opportunities. The globalisation of ownership needs to be set alongside globalisation of occupation. IFCs are characterised by their global orientation and the presence of foreign banks and ﬁnancial ﬁrms – a critical indicator in many of the indices of ﬁnancial competitiveness. The GaWC measures discussed in relation to the world cities use the presence of branch ofﬁces in cities as an indicator of interconnectedness; similar measures feed into Z/Yen’s global ﬁnancial competitiveness index. Lizieri and Kutsch (2006) estimate that 40% of core ofﬁce space in the City of London is occupied by non-UK ﬁrms. An earlier survey on Frankfurt suggested that over a quarter of prime ofﬁce buildings in the City had non-German occupiers. Indeed, there are deﬁnitional issues, particularly given mergers and acquisitions activity in ﬁnancial services – how is the nationality of a global investment bank determined? Presumably, it is related to the location of the formal head ofﬁce of that bank, but this does not necessarily reﬂect the weight of activity. The key point here is that many of the ofﬁce buildings in the core areas of IFCs are occupied by the same ﬁrms. By implication, some of the factors that determine demand for space will affect many IFC markets simultaneously. Shifts in demand affect rents; rents affect supply of space and capital values: those capital values affect the performance of investment portfolios of ﬁnancial ﬁrms based in IFCs locally and internationally – the same ﬁrms that occupy the space. Hence, not only are investment, supply, ﬁnance and funding of ofﬁce markets locked together in IFCs: they are locked together across IFCs. This process of lockstep is important in the context of the volatility of global capital markets and to the existence of contagion and spillover effects. As we have seen, ﬁnancial markets are closely integrated and correlated. Those correlations increase when market performance is above trend and, in particular, when it is below trend, in downturns and in response to negative shocks. Given the link between the performance of global ﬁnancial markets and the demand for ofﬁce space by ﬁnancial ﬁrms and given the functional specialisation that occurs in the core CBD markets of IFCs, then demand shocks are likely to occur in a coordinated fashion across the major global ﬁnancial capitals. With ﬁnancial market crises, falling proﬁts and ﬁrm failures lead to job losses and rising vacancy rates, depressing rents and, since the global ﬁnancial ﬁrms concerned have multiple locations, the effects are widely dispersed across cities. The converse applies to positive shocks driven by booming capital markets. Larger ﬁrms based in ﬁnancial centres capture greater market share and seek to expand, placing upward pressure on rents; rising property prices enhance asset values and encourage lending and development, with the additional activity enhancing short-term proﬁtability.

276

Towers of Capital

The integration of occupational, asset and development markets means that demand shocks (positive or negative) are reinforced and affect the investment performance of real estate assets and the supply of space. In ﬁnancial crises, poor real estate performance feeds back into poor ﬁnancial performance through declining portfolio values and write downs, non-performing loans and debt default and, with declining collateral values, liquidity problems and a credit crunch can quickly emerge. Conversely, where real estate values rise above their fundamental values – in particular, where there is an asset market bubble – this feeds back into ﬁnancial market performance and can create excess credit. Credit availability was one factor encouraging the development of bubbles; rapid increases in capital inﬂows, followed by sudden reversals one of the factors underlying the existing of ﬁnancial crises and contagion effects. This implicit volatility in ﬁnancial centre ofﬁce markets and the way that volatility feeds back into ﬁnancial performance is further reinforced by the cyclicality of property and, in particular, by the development cycle. Game theoretic and behavioural explanations of the existence of development cycles, building cascades and periodic over-supply specify a set of conditions that encourage cyclical behaviour: as set out above, these include price volatility, an undiversiﬁed economic structure, entry barriers for developers and long lags between start of construction and completion of space. By implication, development cycles are likely to be a persistent feature of IFC ofﬁce markets. With supply triggered by demand signals and credit conditions that are likely to be coincident across ﬁnancial centres, the probability of common supply side peaks and troughs increases. Where periodic ofﬁce market over-supply and ﬁnancial market downturns occur simultaneously, the conditions for a self-reinforcing downward spiral exist. From the arguments above, this suggests that the amplitude of cyclical ﬂuctuations in the ofﬁce markets of IFCs will be high and peaks and troughs will be coincident across ﬁnancial markets. To be clear, this is in no way suggesting that real estate and ﬁnancial markets will always coincide or that all ﬁnancial centres will simultaneously experience booms and slumps. First, there will always be local factors operating in real estate markets (for example, the impact of German reuniﬁcation on Frankfurt and Berlin; or, more starkly, the 2001 attack on the World Trade Centre in New York). Second, institutional structures vary across markets which can create lead and lag relationships. For example, supply side response is constrained by planning policy and, to an extent, by the attitudes to risk and behaviour of domestic ﬁnancial institutions; similarly, local labour laws alter the ability of ﬁrms to shed staff quickly or to grow quickly. The leads and lags might mean that one city experiences coincident over-supply and ﬁnancial crisis, while another, lagging in the property cycle, is relatively unaffected. IFCs vary in the extent to which they are diversiﬁed, globalised and externally

Global Financial Centres, Real Estate and Risk

277

oriented and the effects of global crises will, thus, differ. Finally, not all ﬁnancial crises are ‘global crises’: they may be conﬁned to particular regions or particular types of countries, markets and cities. For all these caveats, it seems clear that the performance of ofﬁce markets in global ﬁnancial centres is likely to be volatile and to be subject to pronounced peaks and troughs – and that there is a strong probability of interacting real estate and ﬁnancial crises occurring at the same time in many of those global ﬁnancial cities. Change and innovation in ﬁnancial and in real estate have served to increase the integration between property and ﬁnancial markets, to increase the linkage between global markets and to magnify the feedback effects that can occur in rising and in falling markets. However, it will be difﬁcult to identify such impacts quantitatively and disentangle them from overall market volatility. What evidence is there to support this thesis of enhanced volatility and coincident extreme events?

9.4

Some preliminary evidence

The aim of Towers of Capital is to provide a theoretical framework and set of ideas as a springboard for future research on ﬁnance and ofﬁce markets in major cities. It was never the intention to undertake a detailed statistical analysis of the relationships between rent and ﬁnancial market performance in IFCs. This section sets out some of the issues confronting empirical research and provides some preliminary evidence and stylised facts that support the overall concepts set out in the book. Before providing some analysis and results, however, it is necessary to consider the problems to be faced in researching international real estate markets.

Confronting data issues Any real estate researcher faces signiﬁcant data problems in attempting to undertake international real estate research. Since the direct investment market is a private market, it is difﬁcult to obtain robust, comparable data series – particularly series disaggregated to city or sub-market level. Even where such data series exist for individual cities, deﬁnitions employed differ considerably. Consider, for example, rent levels. Rental levels are typically measured on a square metre or square foot basis, which might seem standard: but will vary depending on whether ﬂoor area is measured on a net or gross area, includes or excludes common areas and service space and so on. What buildings are included in calculating rents? An average rent measure may look at all lettings in an area, not correcting for quality; properties may be split into prime (class A) and secondary space and rents averaged for prime space; or the best achievable rents may be quoted. Practices differ

278

Towers of Capital

across cities, countries and regions – and even vary within individual data providers: … this is inﬂuenced by the methodology used in the United States, where costs represent an average for Grade A space in the market, whereas elsewhere the measure reﬂects typical costs for the best buildings and locations in the market. (CBRE, 2007) The CBRE quotation highlights other issues. First, the rental ﬁgures given are very often ‘hypothetical’ – best estimates by agents operating in the local market of the rents that might be achieved – even where these are anchored on actual transactions in the period preceding the valuation date. Second, what does the ‘rent’ ﬁgure include? CBRE quote ‘total occupancy costs’ – which include management fees, service charges and property taxes payable by the tenant. In many markets it is common to quote ‘net rents’ – the lease rent excluding such variable costs. Varying lease structures place different responsibilities on landlord and tenant which, to some extent, need to be reﬂected in any comparative work. Additionally, the issue of ‘effective rent’ – raised in Chapter 4 – needs to be considered. How are tenant incentives such as rent-free periods at the start of a lease or ﬁtting out allowances, or tenant premia, such as ‘key money’ or deposits reﬂected in the rent? Even where such data issues can be resolved, there remain difﬁculties. Data are published relatively infrequently (usually at best quarterly but often only available annually) and, for all but a handful of markets, time series are short. Where markets are covered by more than one agent or consultant, the separate time series may differ both in magnitude and direction (Kennedy et al., 2007) – a problem compounded by change and consolidation amongst real estate service providers, where one ﬁrm’s coverage may cease, creating a need to splice possibly inconsistent series. In addition, there are (as with all international research) currency conversion issues. If a common benchmark is used – the US dollar or the Euro, for example – rather than local currency, it becomes important to distinguish between property market movements and exchange rate impacts. Again citing CBRE (2007): ‘the other factor that has inﬂuenced movement in dollar-denominated occupancy costs is changes in currency exchange rates. The weakening of the US dollar in recent years has had a signiﬁcant impact of the rankings [of the world’s most expensive ofﬁce markets]’. Not to labour the point, there are similar problems to be confronted with other real estate market variables: yields (capitalisation rates) are not measured consistently across markets; stock estimates and measures of vacancy and space absorption, where available, are rarely robust. There are also important issues of scale – what geographical boundaries are employed? Should analysis be conﬁned to a tightly deﬁned CBD, or should fringe, peripheral

Global Financial Centres, Real Estate and Risk

279

or suburban markets be considered. This is of particular importance in IFCs where, for example, data centres and back ofﬁces may be moved away from the core central ofﬁce location but must be accessible to front ofﬁce and key strategic staff. Hence the size and quality of space in the fringe, peripheral markets are critical to the effective functioning of the businesses located in the core. It is also rare to ﬁnd robust data on ofﬁce investment returns at small area level other than in the very largest of markets. While commercial real estate data are particularly problematic, collecting robust economic, demographic and other contextual data is no easy task. Again, scale issues are important. Even where, for example, detailed employment data are available broken down by type of activity, the geographical coverage often does not coincide with the real estate markets under consideration. Some employment series are collected at household rather than enterprise level, which means that researchers need to consider travel to work patterns and catchment areas, which can be particularly problematic in densely populated regions. Proxies for GDP and other output measures at city level are difﬁcult to obtain. Even basic demographic data present problems. Do ﬁgures relate to administrative areas or to the urban agglomeration? If the latter, is urban agglomeration measured consistently across countries? Confronted with such problems, it is tempting to abandon any attempt at empirical research. Academic pressures to publish exacerbate this problem – in economics and ﬁnance journals, lack of robust data is seized upon by referees as a reason to reject submissions, while the cultural turn in urban social science has led many researchers to abandon quantitative empirical approaches. I would argue, though, that basic but formal empirical work is still vital in shedding light on interrelationships between markets and providing preliminary tests of concepts, hypotheses and theories. Any such analyses must, of course, be mindful of the data issues, must rely on exploratory techniques and needs not to overstate the robustness of the results. Subsequent researchers can build on the preliminary work, reﬁne datasets and contribute to a fuller understanding of international real estate market processes.

Rental levels and IFCs As a ﬁrst step, is there any evidence that IFC status affects rental levels? CBRE publish a biannual survey of ofﬁce rents and total occupancy costs in a range of cities across the world. As noted above, there are issues relating to compatibility, deﬁnitions employed and robustness, but they do provide a benchmark with which to compare cities. The rents from the May 2007 survey are compared to the third Z/Yen Global Financial Centres Index (GFCI). There are 40 cities that are ranked in the top 50 by Z/Yen which are also

280

Towers of Capital

2000 Occupancy costs, $/metre

1800 1600 1400 1200 1000 800 600 400 200 0 400

450

500

550

600 650 GFCI

700

750

800

850

Figure 9.2 Office occupancy costs and IFCs. Source: From data in CBRE (2007), Z/Yen (2008).

covered by the CBRE survey.6 One might expect that the higher the ﬁnancial centre score, the more desirable the city would be as a business location and, therefore, rents would be high due to increased demand for space. Figure 9.2 provides weak conﬁrmation for this: there is a 0.39 correlation between GFCI score and total occupancy costs (and a 0.35 non-parametric correlation between city rankings on GFCI and ofﬁces costs), 30 of the 50 most expensive ofﬁce locations are high ranking IFCs. However, IFC ranking alone is clearly insufﬁcient to explain variations in rent. Although many IFCs are highly functionally specialised, most perform other urban functions. Population forms a general proxy for aggregate demand – and has a 0.45 correlation with ofﬁce costs using the same sample of 40 IFCs. Further, urban economics suggests that as population rises and the spatial extent of the city increases, so land rents at the centre should increase. US cities appear to have lower rents than those found in other nations’ cities which may reﬂect data collection factors (US numbers represent average prime rent, while most others are on a ‘best rent’ basis). There may also be currency factors involved – with the rise in strength of the Euro relative to the dollar. Finally, UK cities appear to have high ofﬁce costs relative to their size and ﬁnancial signiﬁcance7 – possibly as a result of planning constraints and their impact on supply side response to occupier demand. These factors are combined together into a standard regression model in an attempt to explain rental levels (Table 9.3). 6

7

The majority of the Z/Yen high-ranked IFCs not covered by CBRE’s report are small tax havens. Seven of the twenty non-IFC cities in CBRE’s ranking of the 50 most expensive ofﬁce locations are UK cities.

Global Financial Centres, Real Estate and Risk

281

Table 9.3 Explaining office occupancy costs (dependent variable: total office occupancy per square metre). Variable Constant Population (millions) GFCI score US dummy Euro (ex UK) dummy UK dummy

Coefficient

t-Statistics (significance)

−558.05 31.55 1.66 −399.12 +223.58 +503.59

−1.77 (0.085) +5.57 (0.000) +3.06 (0.004) −3.35 (0.002) +2.61 (0.014) +3.29 (0.002)

Adjusted R 2: 0.621, F-statistics: 13.766***, RMSE: 215.98.

This simple model (which appears robust to functional format), explains around 62% of the variation in ofﬁce rents, with all coefﬁcients correctly signed and signiﬁcant. The higher ranked the IFC, the higher the ofﬁce occupancy costs, other factors equal. Thus, the model suggests that London’s rents should be $307 per square metre higher than Toronto’s as a result of the UK city’s ﬁrst-ranked GFCI score of 795 compared to the Canadian city’s ﬁfteenth-ranked 610. London’s costs should also be higher due to its UK location and larger population size. The model predicts a difference in total ofﬁce occupancy costs of $1019 per square metre – the observed difference is a little larger at $1220 per square metre. The largest anomalies are Mumbai, Dubai and Dublin (where predicted rents are substantially higher than observed rents) and Shanghai, Johannesburg and New York, where observed rents are sharply lower than the estimated costs from the model. There are local stories for all six of these anomalies, emphasising the importance of immediate market and economic context in explaining rental values.

Development cycles in cities As Chapter 5 indicated, there are many local, institutional factors that affect the relationship between business demand for ofﬁce space, rental adjustment and the development of new ofﬁce space. Factors that inﬂuence supply response include the physical conﬁguration of the city, the extent to which it is already densely developed, transport infrastructure, the capacity of the construction industry and, critically, planning and conservation constraints. Add these to local speciﬁc factors affecting the market (such as the 2001 terrorist attack on New York) and it is, therefore unlikely that it will be possible to discern strong common statistical patterns in development between cities. Analysis is complicated by the absence of robust and publicly available data on completions that are compatible across cities. Examining a small group of eight cities for which basic completion or stock change ﬁgures were available for 1988–2006 it is hard to discern

282

Towers of Capital

non-regional patterns. Change in ofﬁce stock in downtown Chicago has a 0.75 correlation with mid-town New York (downtown New York’s series is distorted by the shock decline in stock as a result of September 11) and a 0.75 correlation with the City of London. London and New York midtown have a 0.58 correlation. Weaker positive correlations can be observed between Copenhagen and Chicago (and, to a lesser extent, between Copenhagen and London) and for Frankfurt with Paris and Milan. Other correlations, though, are close to zero. Applying a principal components analysis (PCA), the ﬁrst component explains 32% of variation in development, but only London, Chicago and New York have strong loadings on it, with Copenhagen also linked. A second component, explaining 20% of the variation, identiﬁes Frankfurt, Paris and (to a lesser extent) Milan, suggesting some mainland European factor in operation. There is some weak evidence of a lag between development in the City of London and in Frankfurt of around one to two years.

Rental relationships over time The CBRE data provide a snapshot of global ofﬁce occupancy costs. The theoretical model outlined above, though, suggests that there should be coincident patterns of movement in rents across IFCs, as a result of common supply and demand factors and capital ﬂows between the cities. This requires an analysis of changing rents over time. Using annual data provided by JLL and LaSalle Investment Management, change in prime ofﬁce rental values were calculated for a sample of 32 cities for the period 1990–2007. Rental changes were calculated in domestic currency to avoid the results being inﬂuenced by the currency movement effects described in the previous sections. For all 32 cities in the sample, the average correlation in the sample was 0.26. For the 11 cities that were ranked in the top 15 ﬁnancial centres by Z/Yen, the average correlation was 0.46. This difference is statistically signiﬁcant at the 0.05 level. Of the 55 possible correlations between the 11 IFC markets, even given the small sample size, 24 are signiﬁcant at the 5% level and beyond, with 33 signiﬁcant at 10% or beyond. New York has the highest number of signiﬁcant correlation coefﬁcients with other IFCs: of the non-US cities, the three with the highest average correlation are the ‘world alpha cities’ of London, Paris and Tokyo. There are distinct regional factors at play, particularly in North America. The average correlation between all US cities is 0.51; between European cities 0.35. Despite this regional dimension, there does appear to be a stronger link between IFC ofﬁce markets than between those found in world cities in general. Exploratory data analysis using PCA was conducted on 28 of the 32 cities (those with no missing data). The PCA approach seeks common patterns of

Global Financial Centres, Real Estate and Risk

283

movement between objects; here city ofﬁce markets. The ﬁrst component explains as much common variation as possible; the second component, orthogonal, uncorrelated with the ﬁrst, explains as much of the residual variation as possible, and so on until the ﬁnal component. The factor loading of a city on a component gives some indication of its importance in explaining rental movement. PCA will produce as many components as there are cities. If the cities behaved distinctly and individually, then each component would explain much of the variation in one city and little of the variation in others. If there are common factors driving rental movements then the early components will explain much of the variation with many cities exhibiting high loadings: while late components will explain very little variation. The results largely conﬁrmed the Goetzmann and Wachter (1995) ﬁnding of the existence of a global ofﬁce market factor (Figure 9.3). The ﬁrst component extracted explained 38% of the variation in the dataset, with 20 of the 28 cities having loadings of 0.5 or higher, including 9 of the 11 high-ranked IFCs – the exceptions being Hong Kong and Sydney. In general, Asian cities have lower factor loadings than European and North American cities. A varimax rotation of the retained components again conﬁrms a regional pattern, separating the US and the Asian cities more clearly. Focusing just on the 11 high ranking IFCs in the JLL dataset, a further focused PCA was run. The ﬁrst retained factor explains 53% of return variation, with all IFCs loading at 0.62 or higher with the exception, once again, of Hong Kong (0.495) and

Component

% of variance

1 2 3 4

52.7 13.2 10.1 9.2

All other components 6% or less Loadings on Factor 1: Boston Chicago Frankfurt Hong Kong London New York Paris San Francisco Singapore Sydney Tokyo

0.795 0.850 0.622 0.495 0.721 0.921 0.798 0.817 0.676 0.375 0.736

Figure 9.3 Principal components analysis: 11 IFCs, rental change.

284

Risk Risk Risk

Towers of Capital

0.10 + 2.26 Mean (7.25**) (6.50**) = 0.12 + 2.26 Mean − 0.07 US (8.36**) (7.83**) (−3.70**) = 0.10 + 2.00 Mean − 0.07 US + (7.25**) (6.50**) (−3.51**)

Adjusted R2 = 0.61

=

Adjusted R2 = 0.68 0.05 IFC (2.00*)

Adjusted R2 = 0.73

All estimates OLSQ, Newey–West heteroscedasticity corrected coefficient estimates. ** significant at 0.01 and beyond; * significant at 0.05 level. Data: annual rental change 1990–2007, 32 cities.

Figure 9.4

Explaining rental volatility in global cities.

Sydney (0.375). An hierarchical cluster analysis8 of the 11 cities emphasises that, over this short time period, Hong Kong and Sydney behave differently from the other IFCs, with Singapore somewhat separate, too. The other ﬁnancial centres merge quickly: there is initial regional clustering (particularly for the US cities), but most cities swiftly form a single group. Overall, then, even with a very short dataset, there is evidence of a strong common factor in rental change in major global cities, and in the ofﬁce markets of IFCs in particular, but also that there are regional factors at play within ﬁnancial centres. The model also suggests that IFCs will be more volatile. The average standard deviation of rental change for the 11 leading IFCs was 21.95%; for the remaining cities, the average standard deviation was 14.86%. Using an unequal sample t-test, the difference between the two means is statistically signiﬁcant at the 5% level. The higher rental volatility, consistent with the model outlined above, is compensated by a higher mean rental growth (4.8–2.2%), a difference which is, as with the risk measure, statistically signiﬁcant. Simple regression analysis with the standard deviation as the dependent variable shows that 61% of variation in risk is associated with the average growth rate; adding a dummy variable that identiﬁes US cities increases the level explanation to 68%; a dummy variable identifying IFCs further increases the R2 to 73%, with the coefﬁcient on the dummy statistically signiﬁcant at the 5% level (Figure 9.4). Given the short time series, it is not possible to conduct robustness tests so it is important not to overstress the signiﬁcance of the results, but it does provide some indication of higher rental volatility in international ﬁnancial services ﬁrms. A similar dataset from Property & Portfolio Research (PPR) covering ofﬁce markets in 29 cities for 1991–2007 produced very similar results when 8

The cluster analysis used a hierarchical procedure using squared Euclidean distance metric and Ward’s method of clustering. See Hoesli et al. (1997) for a discussion of the method.

Global Financial Centres, Real Estate and Risk

285

rental change was examined. The overall average correlation of all 29 cities in the sample was 0.368: for the 14 cities ranked in the top 20 on Z/Yen IFC index, the average correlation increases slightly to 0.417: but as with the CBRE data, the Asian IFCs – in particular Hong Kong – appear to have relatively low correlations. Excluding the Asian cities (but retaining Sydney), the remaining 11 IFCs have an average correlation of 0.65 suggesting strong common movement. The PPR dataset also includes a yield or capitalisation rate, which allows a hypothetical capital growth series to be estimated for each city. These series do not seem to be very robust, but there does seem to be evidence of common capital growth between the non-Asian IFCs, with average correlation of around 0.5. Capital growth is considerably more volatile in IFC than in non-IFC cities even if the extreme volatility of the Hong Kong market is excluded. Capital values grow strongly in most IFCs from 2003 to 2007, averaging 12.3%. London and the volatile Hong Kong market experienced declines in the year to Q4 2007, while others showed early indications of slower growth. Comparing rental volatility across national real estate markets is made problematic by differences in individual market structures and arrangements – for example by differences in lease contracts, in supply side responsiveness, in the role of brokers, agents and appraisers and in the quality of market information. Within national boundaries, the model suggests that rental volatility should be higher in and around IFCs than in other markets. The United Kingdom provides some supporting data for this. IPD’s Annual Digest allows examination of rental value change for different 15 ofﬁce market segments for the period 1980–2007. The segments with the highest standard deviation are the four inner London segments – the City of London (14.5%), Mid-Town (15.1%), West End (14.7%) and Inner London (14.5%). The average standard deviation of the remaining 11 segments is just 7.8%. The central London volatility is not compensated by higher rental value growth. City ofﬁces have below average growth and the worst risk/ return ratio of any of the 15 segments. This result is not simply a function of the dramatic rental falls in 1991–1993; conﬁning analysis to the 1995–2007 period, the four central London ofﬁce market segments have the four highest standard deviations, all more than double the average standard deviation. Of the four central London markets, the City has the lowest rental growth over this period. Figure 9.5 compares annual ofﬁce rental value change in the City of London with change in the South East of England – the region surrounding London, and the location for much of the decentralisation of retail ﬁnancial service and back ofﬁce activity. The higher volatility of the core ﬁnancial market is evident. Given the lack of robustness of the private real estate data, a more productive line of enquiry might consider the co-movements of economic driver variables. Unfortunately, it is not easy to obtain comparable city level data

286

Towers of Capital

40 30 20

Per cent

10 0 –10 –20 –30

City

07

05

20

20

03

01

20

20

99 19

97 19

95 19

93 19

91 19

89 19

87 19

85 19

83 19

19

81

–40

South East

Figure 9.5 Office rental value change, City of London, South East. Source: Estimated from IPD Annual Digest data.

that measure ﬁnancial market activity linked to occupational demand. Co-movement of equity markets is a well-known phenomenon. Examining low-frequency stock market returns for 1990–2006, the correlation between 18 leading markets was 0.50; for markets located in leading IFCs, this rose to 0.60. However, as noted above, for ofﬁce demand, what is relevant is where trading activity takes place, making equity performance at best a fuzzy indicator. Examining Bank for International Settlements (BIS) data for change in the volume of debt issuance over the period 1994–2007, the mean correlation between all markets surveyed by BIS is 0.19; this rises to 0.37 in markets which contain a leading IFC; to 0.48 if Hong Kong is excluded; and to 0.56 just considering debt issuance in the United States, United Kingdom, Japan and Germany. Finally, preliminary analysis of growth in ﬁnancial and business service employment shows a correlation of 0.68 between London and New York in the period 1990–2007, despite opposite growth trends over that period, with the two series apparently co-integrated allowing for the trend. Change in FIRE employment in New York and Chicago employment shifts shows a strong annual correlation – with, at a ﬁner, monthly, grain, New York leading by around three to six months. However, there is little evidence that Frankfurt’s ﬁnancial services employment follows the London and New York trend. All three markets show a decline in ﬁnancial services employment in the aftermath of the bursting of the dot.com and technology

Global Financial Centres, Real Estate and Risk

287

bubble in the early 2000s – although the falls in London and New York precede Frankfurt by one year, perhaps reﬂecting institutional inﬂexibility in German labour markets.9 These results clearly do not amount to deﬁnitive evidence in support of the idea of ofﬁce market linkage advanced: but they are, at the least, supportive of the predictions of the model. It is, perhaps, too early to observe the impact of some of the changes to the commercial real estate environment – many date back only to the late 1990s so, while the effects of the boom and market adjustment phases can be observed, robust quantitative analysis is not possible. What evidence is available does, though, point to higher volatility and to patterns of common movement, particularly at the turning points of cycles. Combined with individual observation of the linkages between real estate and ﬁnance and the feedback mechanisms as they play out in the 2007 credit crunch and the earlier 1990s downturn, it does seem that the integration of capital, ﬁnancial asset and real estate markets and the evolution of the ofﬁce markets of IFCs create the possibility of systemic risk. What are the policy implications of this developing feature of ﬁnancial markets?

9.5

Global ﬁnancial centres and ofﬁce market risk: some policy implications

This ﬁnal section addresses a number of crucial policy issues. First, the integration of ﬁnancial markets and commercial property markets produces systemic risk. How should regulators and central banks address that risk? Moreover, these spillover and contagion risks are spatially focused on global ﬁnancial centres and their ofﬁce markets: what are the implications for city planners and city governments? The problems facing city planners are compounded by shifts in ownership patterns, with fragmentation and globalisation. What does this mean for the maintenance and improvement of the built environment – both at individual property and at urban level? What beneﬁts are there from the development and integration of real estate and ﬁnancial markets? Finally, what needs to be done to improve our understanding of the ofﬁce markets of major world cities and the linkage between those markets and global ﬁnance?

9

The time series is too short for robust results, but New York ﬁnancial service employment changes Granger cause Frankfurt changes at the 0.05 signiﬁcance level over the period 1994–2006. New York changes Granger cause Chicago changes at the 0.05 signiﬁcance level using monthly data of 1990–2007.

288

Towers of Capital

Financial markets, real estate markets and systemic risk The growing integration of ﬁnancial markets, the dominance of ﬁnancial markets by a relatively small number of cities and rapid ﬁnancial product innovation creates the possibility of systemic risk effects. Systemic risk exists where failure in one part of the (ﬁnancial) system has spillover effects in other parts, threatening the stability of the system as a whole. Systemic risk has been a subject of growing concern since the ability of regulatory authorities to control and manage crises has diminished. With power and assets concentrated in fewer institutions and markets, the threat of a cascade through the system is real. Heinman and Alexander (1997, p. 82) suggest that ‘rapid changes occurring in the international ﬁnancial system have resulted in new sources of, and transmission mechanisms for, systemic shocks’. Similarly, Phillips (1993, p. 6) notes that ‘the expansion of market linkages which cut across national boundaries and embrace a wide range of ﬁnancial and non-ﬁnancial ﬁrms raises concerns about the ability of central banks to contain systemic difﬁculties, should they emerge’. The international ﬁnancial system has proved relatively robust in dealing with major failures in recent years – as the ﬁnancial system’s ability to cope with the collapse of major hedge funds LTCM in 1998 and Amaranth in 2006 and the international response to the fallout from the 2007 sub-prime crisis show. Information and communications technology developments and their integration in trading systems permit rapid transmission of ﬁnancial shocks – but they also permit rapid adjustment processes and response to coordinated policy interventions by central banks. However, the localised effects of internationally induced crises are considerable and regulators and central banks have struggled to manage domestic impacts of global crises, particularly where there are regional contagion effects. The linkage of real estate to ﬁnancial markets has been a cause for concern amongst regulatory bodies for over a decade. The coincident real estate downturns of the 1990s led to a number of warnings of potential risk (see, for example, Bank for International Settlements, 1993; Goldstein & FolkertsLandau, 1993; Folkerts-Landau & Ito, 1995; Renaud, 1997). In that cycle, the expansion of lending into real estate, fuelled by rapid asset price inﬂation and deregulation (giving access to new sources of funds) created problems for Japanese, Nordic, UK and US banks in the aftermath of the world recession. Cyclical decline was associated with major falls in property values. The losses incurred by banks were sufﬁciently large to force curtailment of portfolio growth to meet regulatory capital requirements such as the capital adequacy and solvency requirements of the Basle Accord. This created the early 1990s ‘credit crunch’ where reduced (and more costly) lending hampered business expansion and prolonged the recession. Ghosh et al. (1994),

Global Financial Centres, Real Estate and Risk

289

for example, demonstrate the adverse impact of the collapse of Olympia and York’s Canary Wharf project on bank proﬁtability around the world. Unfortunately, many of the warnings sounded in the mid-1990s went unheeded; the lack of signiﬁcant global real estate market impacts from the ﬁnancial crises of the late 1990s and early 2000s created a more bullish, less risk-averse view which, consequently, failed to constrain the expansion of real estate credit and the escalation of real asset prices in the mid-2000s. For example, Zhu (2002) writing for the BIS, argues that there was a ‘missing’ commercial real estate cycle with no above trend escalation of prices and subsequent sharp adjustment: arguably partly attributable to the rapid growth of real estate securitisation in the past decade. First, the emergence of new ﬁnancing methods provided a substitute for traditional banking ﬁnance and may have helped even out the ﬂow of capital into the commercial property sector. Second, the development of public markets improved information transparency and may have strengthened market discipline. And ﬁnally, the development of public real estate equity and debt markets made it possible for commercial property risk to be spread through capital markets to a wider array of investors. Zhu went on to note that closer integration of commercial real estate markets with capital markets suggested that commercial real estate was still prone to shocks and, indeed, might even be subject to new sources of market volatility, presciently adding that ‘capital markets are also vulnerable. The liquidity appeal of securitisation to investors is a double-edged sword. Just as capital users can obtain rapid access to funds on a broad basis, so capital suppliers can quickly move their funds out of the markets’. However, the idea that debt and equity securitisation and the new (global) investment vehicles represented a structural change and a new business environment became received wisdom. As ever, claims of structural change and ‘a new paradigm’ presage a reassertion of economic fundamentals, sharp falls in asset values and company failures. The idea that real estate securitisation helps prevent cycles might seem ironic writing in 2008, but the risk diversiﬁcation and efﬁciency gains still exist, even in the aftermath of the 2007 credit crunch. The growing linkage of real estate and ﬁnancial markets, then, creates the potential for greater volatility – or extreme movements – where cyclical peaks and troughs coincide. Given that the economic forces which drive the cyclical ﬂuctuations are international in nature, then extreme events are likely to be coordinated around the world – and at their most extreme in global cities, in international ﬁnancial service centres where integration of occupation, investment, supply and ﬁnance are tightest and where ofﬁce

290

Towers of Capital

occupation is dominated by internationally oriented ﬁnance ﬁrms. Again, this is not a new idea: I was only one of a number of authors highlighting the issue after the 1990s real estate downturn. The basic principles were outlined in Lizieri and Finlay (1995), providing evidence that real estate ﬁnancial crises tend to be coordinated in world ﬁnancial capitals, frustrating global property risk diversiﬁcation strategies (see also Goetzmann & Wachter, 1995; Ball et al., 1996). However, the developments of the last decade have increased the risk of coordinated ﬁnancial city behaviour, with further concentration of international ﬁnancial activity, with growing functional specialisation of the core ofﬁce markets in IFCs and with the changing real estate investment and ﬁnance environment. Functional specialisation of global ﬁnancial cities is thus also a ‘two-edged sword’. The concentration of international ﬁnancial and business activity in IFCs like London, New York or Frankfurt contributes to the depth of markets, permits specialisation and creates the agglomeration economies that are critical to the cities’ competitive position. Without the scale and concentration, it is hard to achieve information spillovers and to create and market the product innovations that are required in an evolving ﬁnancial market. Without the scale and concentration, it is hard to maintain a labour market that provides mobility and that attracts international talent. At the same time, an undiversiﬁed economic structure and interlocking markets create dangers of downward spirals and negative feedback effects whenever the international ﬁnancial system is subject to shocks and, in particular, when ﬁnancial and real estate shocks coincide. Developments in the institutional organisation of global city ofﬁce markets, with their tightly linked ownership, occupational and ﬁnancial structures, may contribute to that danger. This creates a policy dilemma for city planners. In order to maintain ﬁnancial competitiveness and to capture international market share, clustering needs to be encouraged – which also means permitting the real estate development that is necessary to upgrade the stock to ‘global standards’ and to accommodate growth.10 This can be seen explicitly in the policies of many of the major IFCs – in the City of London’s prodevelopment stance, in Frankfurt’s ‘Frankfurt 2000’ ofﬁce development plan, in New York (Fainstein, 1993). Even in the more diversiﬁed Tokyo market, Machimura (1992) reports planned concentration on global ﬁnancial control functions in certain central districts with concomitant construction activity. However, permitting the creation of a critical mass of 10

Massey (2007) questions whether cities should compete for global ﬁnancial market share – and asks who beneﬁts from the pursuit of global city status. This links back to the global cities literature on income and wealth inequality. Whether a loss of market share would have positive distributive effects must be moot.

Global Financial Centres, Real Estate and Risk

291

ﬁnancial ofﬁce space and encouraging clustering and functional specialisation creates an undiversiﬁed economy, vulnerable to ﬂuctuations in global markets. And that undiversiﬁed employment structure is one of the factors that encourages cyclicality in real estate development.

Ownership, time horizon, the built form and planning The continued and accelerating growth of new property investment vehicles makes the deﬁnition of real estate ownership more complex and slippery. With equity market investment, the convention would be to deﬁne the nationality of a company in terms of its place of listing and/or its effective headquarters – even though the shares may be held, substantially, by foreign investors. Similar principles apply to real estate funds, although the situation is complicated where the formal head ofﬁce is in a tax haven while the de facto management is based in a global city. However, co-investment through, for example joint venture limited partnerships with small numbers of participants are more complex. The vehicle is not a taxable entity and the partners clearly hold a beneﬁcial stake in the property and exercise some form of management control or inﬂuence. The situation is further complicated by the creation of feeder fund and fund-of-fund products and by the impact of asset-backed securitisations with the layered ownership and restrictions on management actions that these bring. The changing nature and form of property ownership in ﬁnancial cities is, potentially, accompanied by a change in attitude towards investment in real estate. The use of partnership and corporate forms of ownership provides (again, in principle) an enhanced liquidity; this, along with the globalisation of ownership, is associated with much more of a focus on short-run risk and return performance and active portfolio management. Fund managers (including REIT managers) who are best able to work their portfolios, to ‘add alpha’, will, over time (and in rising markets?) be more successful in attracting capital. In many IFC ofﬁce markets, there is a clear increase in transactions activity over time, implying much shorter holding periods. Buy-and-hold strategies may become less common. At one level, this brings beneﬁts in the form of greater market-based information and transparency and a reduction in any risk premium in return to reﬂect illiquidity and the need to hold property for longer than may be optimal. On the other hand, it also creates the potential for greater volatility in the market, if substantial numbers of owners seek to exit the market at the same point. Such a risk may be increased by the growing proportion of ﬁnite life funds. If fund creation is clustered over time (linked to the credit cycle and to the relative performance of real estate compared to other asset classes) and ﬁnite life fund length becomes standardised, then this may

292

Towers of Capital

create clusters or swarms of funds reaching the end of their life (and, even if those funds vote to continue, their ﬁnance and funding arrangements may require rearrangement). If a swarm of fund renewals coincides with a real estate market downturn, then funds may seek to – may be forced to – sell into a falling market, exacerbating problems in the market and pushing asset values down further. At 2007, 48% of European real estate funds by value on the INREV database are classed as ﬁnite life funds – with a gross asset value of €134 billion: 28% of those funds – with property under management valued at an excess of €37 billion – are due to close in 2010 or 2011. There are further implications of the changing nature of ownership for the quality of stock and for planning of the built environment. Shorter holding periods and a focus on maximising cashﬂow over short time horizons potentially leads to underinvestment in the built form, increasing the impact of depreciation and obsolescence. Fragmented ownership patterns (particularly where the true owners are masked behind nominee companies or tax-driven offshore structures) makes negotiations on major planning initiatives and on the overall asset management of the major global city ofﬁce markets much more complicated. It would be wrong to overstate this problem: in many instances there will be a local fund or asset manager with decisionmaking powers. However, it is evident that the fragmentation of ownership and the widening of the distance between the planning authority and the end investor make achievement of optimal decisions – particularly where these require delicate negotiations and concessions – much more difﬁcult than under traditional single ownership modes.

Beneﬁts and gains in the new real estate markets Lest this appear too pessimistic, it is important to stress that there are many beneﬁts from the restructuring of commercial real estate markets and from the spread of global investment that have both changed the nature of real estate investment and the patterns of ofﬁce ownership in ﬁnancial centres. First, the new investment vehicles have removed entry barriers and permitted a far wider set of investors to participate in major urban ofﬁce markets. While this is hardly a democratisation of ownership – most investors in real estate funds are professional investors and high net wealth investors – it has the effect of spreading risk and does allow smaller pension funds, charities and endowments to gain exposure to such markets and to diversify their portfolios. The new vehicles permit smaller investors to pursue realistic global property investment strategies, removing many of the information and cost constraints which acted as a barrier. There are efﬁciency gains from this, since capital can ﬂow to markets offering the best returns and since investors can achieve better risk-adjusted returns by holding an

Global Financial Centres, Real Estate and Risk

293

internationally diversiﬁed portfolio. That there are common global factors that can lead to coordinated booms and slumps across cities does not mean there are no diversiﬁcation gains from local factors and leads and lags in cycles. Next, there are information gains. Securitisation brings public market pricing. REIT prices are observable and change quickly with shifts in market sentiment. Real estate bonds and mortgage-backed security prices are observable; redemption yields reﬂecting investor attitudes to risk. Secondary market trading in debt and equity securities provides regular, high-frequency information. Shifts in property derivative spreads will, increasingly, provide information on changing market sentiment. These signals are, of course, by no means perfect – particularly where markets are new and evolving. CMBS spreads and ratings in the ﬁrst half of 2007 probably did not truly reﬂect the risks of investing in such products and it is quite possible that yields in the ﬁrst half of 2008 do not truly reﬂect that risk either. Private real estate funds offer less transparency, but there are attempts to create secondary markets in property units and to provide at least indicative pricing. Additionally, globalisation of investment activity brings demands for greater market information, greater transparency and more stable property market structures, through the spread of real estate research, performance indices, benchmarking services and parallel developments. This again breaks down barriers and helps remove misunderstandings about the characteristics of local markets. There is an argument that this amounts to a form of cultural imperialism, that it exports an Anglo-American model of real estate investment,11 but while this has some validity, the process has generated more robust market data and analysis for a wide range of markets. The suggestion that the new ownership structures bring short-termism in investor attitudes to real estate management can, in part at least, be offset by arguing that active asset management implies an upgrading of the stock – higher returns are achieved by repositioning, refurbishment and redevelopment. A long-term investor may adopt a laissez-faire attitude to the property – particularly where there are long leases – relying on tenants to make improvements. The ﬁnite life fund manager or the REIT manager under pressure to deliver dividend growth might be more concerned with maintaining the building’s capacity to deliver rental income and to retain value on sale. Moreover, many of the major ﬁnancial ﬁrms based in IFCs are owner-occupiers rather than tenants. Why would such ﬁrms own rather

11

It has been suggested that there are distinct Asian and Middle Eastern processes which ﬁt uneasily within the dominant Anglo-American investment paradigm. Certainly, providing executive education in the Middle East presents challenges that reﬂect such cultural differences.

294

Towers of Capital

than rent space? A number of possible reasons may be advanced. These include the following:

• that space is seen simultaneously as an operational base and as a ﬁnancial asset;

• that ownership of space brings control of space, allowing the ﬁrms to adapt the building to their changing operational needs; • that ownership of space signals a commitment to local clients, an indication that the ﬁrms are forging a long-term relationship with the city and the region; The ﬁrst two reasons would suggest that overseas ownership may signal commitment to IFCs as a location implying some form of stability. In the event of a real estate market downturn, an owner-occupier is less likely to take ﬂight, is less concerned with short-run return ﬂuctuations. The ownership of space also acts to retain a global presence in IFCs through international ﬁnancial market crises, which helps retain integration of ﬁnancial markets and facilitates adjustment processes and ﬂows of capital. Where ﬁnancial ﬁrms are tenants, they will also typically be investors in prime ofﬁce space. They thus have ofﬁce space as an asset (in their investment portfolios) and as a liability (their lease commitments), forming a hedge. Provided that rental commitments fall along with capital values, ﬁnancial ﬁrms are partially hedged against property market downturns. Lease contracts do, however, create a stickiness in rents that hampers adjustments.In the City of London, with its long leases and upward only rent reviews, many major ﬁnancial ﬁrms who rented space in the late 1980s were committed to paying rents that were well above market levels across the whole of the 1990s and into the 2000s, unable to beneﬁt from the sharp rental falls in the early 1990s.

Summing up Real estate investment and real estate development does not take place globally or in nations, it takes place locally. Major ofﬁce development and investment happens in cities. The development of global real estate investment and real estate ﬁnance markets gives that local focus an international ﬂavour – and this is most clearly observed in the core ofﬁce markets of global cities and, in particular, in IFCs. In the major IFCs, the activities of ofﬁce occupiers are functionally specialised – not simply focused on ﬁnancial services but on externally focused, international ﬁnancial services. This means, inevitably, that the demand for ofﬁce space – and, from this, the determination of rental values – is strongly inﬂuenced, even determined, by the performance of global ﬁnancial markets. As a result, IFC ofﬁce markets are exposed to volatility from ﬁnancial markets alongside

Global Financial Centres, Real Estate and Risk

295

the inherent cyclicality that results from the property development cycle. As the major global cities of ﬁnance capture a greater market share of highlevel ﬁnancial activity, asset and wealth management, and as ﬁnancial product innovation becomes more concentrated, this link between international ﬁnancial markets and ofﬁce demand becomes more pronounced. Developments in commercial real estate markets have tended to bring a closer integration of real estate and capital markets. This is evident in the use of debt products and, in particular, of debt securitisation; it is evident in the growth of listed real estate securities; it is evident in the growth of private real estate funds that can collect capital from a wide range of investors and channel it into real estate markets around the world – with prime ofﬁce real estate capturing a signiﬁcant share of that investment. These developments have tended to bind together the development of ofﬁce space, the funding and ﬁnancing of that space, real estate investment and ofﬁce occupation more closely together. Again, this is most evident in the core ofﬁce markets of IFCs, where the occupiers of space are global ﬁnancial ﬁrms, it is banks and other ﬁnance ﬁrms that provide the debt capital that ﬁnances development and funds investment, institutional investors and other ﬁnancial investors that acquire prime ofﬁce space for their portfolios, and where the major real estate developers, investors, fund managers and service providers are located. This locking together of the different elements of the real estate market coincides with a locking together of ﬁnancial and real estate markets. It also occurs on a global scale, with ﬁnancial ﬁrms exposed to IFC ofﬁce real estate in many different cities. International ﬁnancial markets are prone to booms and, in particular, to shocks. Periodic crises spill over from one sector of capital markets to others and from one country to others; this inherent volatility is embedded in the ofﬁce markets of global ﬁnancial capitals. But real estate markets are not simply a passive recipient of ﬁnancial market shocks. They have a material, an instrumental impact on ﬁnancial market performance and stability. Credit expansion brings ﬂows of capital into real assets and places upward pressure on property prices. The higher real estate values then form the collateral for further borrowing – by ﬁrms to fund their business activity and by banks to expand their lending. Rising values bring higher real estate investment returns and investment values. Thus the link between ﬁnance and real estate can fuel a rising cycle of prices and lending. Conversely, problems in the real estate market can create a downward spiral as falling asset values damage ﬁnancial performance, lead to non-performing loans and default and can create liquidity crises, credit crunches. This is exacerbated in IFCs since the occupiers of space are those ﬁnancial ﬁrms struggling to deal with the ﬁnancial crisis, with implications for space demand, for rents and, from that, for ofﬁce values. This inherent volatility and the existence of systemic risk have policy implications for ﬁnancial regulators, central banks – and for city planners.

296

Towers of Capital

The dilemma faced by policy makers is that concentration and clustering of ﬁnancial markets, functional specialisation of activity and the transformation of urban real estate markets to meet the needs of global ﬁnancial ﬁrms bring efﬁciency gains, enhance competitiveness and help maintain and capture market share: but, at the same time, they expose cities to sharp cyclical movements and risk from shocks. For city planners, this risk is compounded by shifts in the pattern of ownership – as fragmentation of ownership and the dispersal of stakeholders in city ofﬁces makes coordination of urban strategy more difﬁcult and complex. There are policy implications, too, for investors. There are beneﬁts from holding global ﬁnancial centre ofﬁces in portfolios – liquidity, reduced risk of tenant default, economies of scale in management and monitoring, high levels of transparency and market research. The investor must decide whether those beneﬁts offset the potential volatility and risk that the exposure brings – in particular the downside risk that can occur when ﬁnancial market and ofﬁce market downturns coincide – and whether the returns justify those risks. Moreover, since rents, values, returns and the supply cycle in such markets are, at least in part, driven by the same fundamental global factors, the investor must consider whether there is any true diversiﬁcation gain from holding a portfolio of ofﬁces that are located in different IFCs. Local and regional factors will provide some diversiﬁcation gains, but those might disappear when needed most, in the face of extreme negative returns and adverse capital market conditions. What is needed to address these policy questions? Most of all, it is a proper and full understanding of the processes that operate in ﬁnancial markets and in ofﬁce markets and the way that those processes interact. Academic research is increasingly compartmentalised. There are honourable exceptions: but much of ﬁnance research has typically either neglected real estate as an asset class or analysed listed real estate securities without really addressing the link between securities returns and the behaviour of the underlying property assets. Real estate research from an urban social science tradition has typically declined to address the mechanics of ﬁnancial markets and relied on high-level – even superﬁcial – interpretations of their operation. Much urban and ﬁnance research seems blind to historical antecedents and resonances. Working practice seems similarly segmented. The writing of Towers of Capital was an attempt to bridge those divides, to pull together strands from ﬁnancial markets, real estate and urban theory to help understand how ofﬁce markets in big cities function. The next critical task is careful empirical research – both through detailed case studies and more aggregate quantitative analysis, although the latter will always be hampered by data problems, real estate research’s Achilles’ heel. My hope is that Towers of Capital will form a framework for, and inform, that programme of empirical work.

Epilogue As I ﬁnished the writing Towers of Capital, I did not experience any real sense of completion. At each stage and in each chapter, there were areas of discussion that I wished I could expand or analyses I wished I could run. In writing the historical chapters, I wanted to go deeper, to ﬁnd out more about nineteenth-century global property investment schemes and the spillover effects of banking crises on real estate. In writing about ofﬁce ownership in global cities, I wanted more evidence about cross-holdings and the nature of ownership in different cities and regions. In examining demand for ofﬁce space and its impact on rents, I wanted to conduct more analysis on the links between ﬁnancial employment in different cities and how local property market structures and labour market regulations affected response to the global ﬁnancial environment in individual cities. The book, then, is a staging post, a base camp for a continuing programme of research on real estate in global ﬁnancial cities. As an afterword, then, this brief section points to where that programme might lead.

Financial centres and regional difference One unresolved issue is the extent to which the models and trends described here apply to cities in all regions of the world. There is some danger of a form of cultural imperialism in assuming that models developed and tested in one market have a universal applicability. Much of the evidence used to support the interlocking model is based on London, on US cities and on Western European cities. There are, thus, political and cultural factors that are common to those cities. They are also, for the most part, historic cities, many of which have played a key role in international ﬁnance and global capital ﬂows for centuries. As historic cities, they have a characteristic urban form and the legacy of a downtown commercial form. How well does this model ﬁt Asian cities? Are there differences in the organisation of property markets, in the ways that business is conducted, in the legal and institutional structures that inﬂuence spatial outcomes and the interaction between ﬁnance and real estate? Similar considerations apply to emerging cities in the developing and transitional economies. For example, in Dubai or Abu Dhabi, an understanding of the evolution of the city has to be an awareness of the way that physical growth is tied to a development-driven model, and to the strong role of the state both in directing development and providing capital. Given the weaker role of occupational demand in determining the nature of space provided,

298

Epilogue

the parallel use of Western and Sharia-compliant ﬁnance and the absence of transparency in the establishment of rents and prices, it is hard to assert that standard rental adjustment and investment appraisal models apply. Fundamentally, supply and demand must interact. The underlying value of the property created rests on the demand for that property, which, in turn is based on its utility for occupiers. Yet that cannot explain the current trajectory of development and the evolution of those cities based on a ‘build it and they will come’ philosophy. A different dynamic applies in transitional economies, where the legacy effects both of historic cities and of land-use patterns from the era of planned economies create a path dependency that must affect outcomes in ofﬁce markets. There may also be impacts from the speciﬁc role played by a city in the global ﬁnancial system. New York’s ﬁnancial markets grew initially from its role as a recipient of capital, for investment in the growing United States economy. Its location on the eastern seaboard, facing Europe, its own industry, trading activities and role as a port allowed it to develop critical mass and, with the economic and political growth of the United States, assume a global role. Shanghai performed a similar historic role for China and may have resumed that function as the Chinese economy opens up; as the staging post for capital investment, and as the focus for capital accumulation and, ultimately, export. If Shanghai is to become a major global ﬁnancial centre, does this point to the importance of the national economy in the evolution and growth of international ﬁnancial centres (IFCs) and erode the idea of such cities as somehow independent of their geographical setting? The growing presence of Mumbai and Sao Paulo in rankings of global ﬁnancial cities gives some credence to this idea. One implication of this is that the City of London is, to an extent, unique in maintaining its pivotal role despite relative national decline economically and politically.1

Empirical tasks The model advanced here points to patterns and interrelationships that are, in principle, testable. For example, rental adjustment models developed for individual cities can be adapted using panel methods to analyse rental response across cities. This would enable us to test formally for the existence of common factors – in particular, whether there is a common global ﬁnancial sector employment or output factor. Such an analysis would complement and augment the early attempts to explore co-integration in rents and 1

The relative decline hypothesis itself sometimes seems overstated: UK GDP remains more than double that of India, Russia or Brazil, despite population, land mass and physical resource disparities.

Epilogue

299

capital markets. Similar analyses exploring linkages between movements in capital values, construction starts and investment capital ﬂows would greatly enhance our understanding of the dynamics of global city ofﬁce markets. Such analyses need to be linked to research on the integration and convergence of ﬁnancial markets (but with a careful consideration of the wealth impacts of price movements), on contagion and ﬁnancial crisis and on the relationships between labour markets in ﬁnancial cities. A whole programme of empirical research beckons. That programme, though, requires a massive effort in data collection. The analyses that have been conducted rely on readily available series that are not necessarily fully ﬁt for the purpose. At the most basic level, ofﬁce rent series for cities are needed. How robust are the existing series? How compatible are they? Some series report best rents, others average rents; some are adjusted for lease terms and tenant incentives, some are headline ﬁgures alone; some series reﬂect total occupation costs, others just the core, base rent. Not all series are spatially ﬁne grained – do they apply just to the core CBD ofﬁce market or to a wider metropolitan area? Now, to some extent, this may not matter. If we are mainly interested in the direction and magnitude of change, then precise deﬁnitions may not be crucial. Yet differences in measurement and deﬁnition may mask movements. For example, if tenant incentives are ignored, then falls in effective rent levels in property downturns will be understated. A €200 per square metre rent with no concessions is not the same as a €200 per square metre rent with an 18 months rent-free period at lease commencement. Furthermore, how robust are the time series in smaller, less densely researched markets? Jumps and discontinuities need to be investigated, to ensure that they represent actual market changes rather than shifts in data collection methods or approach. This task is made much more complex given the private nature of much commercial real estate data. Weak public sources of data allow private providers to create lucrative businesses collecting and selling on information. They will be reluctant to turn these data into a public domain resource. The same careful identiﬁcation, collection and cleaning operation is required for other variables. It is not just the real estate variables that are problematic. For example, there are signiﬁcant problems in collecting local employment statistics that match the spatial extent of the core ofﬁce markets and that are broken down by sector and sub-sector to allow identiﬁcation of ﬁnancial, business and professional services users of space. Some data are collected at enterprise, workspace level, others at residential level (which raises questions about travel to work and catchment areas), some are only available at city-wide or regional level. There are differences in national industrial classiﬁcations and, once again, deﬁnitions change over time. A casual examination of ﬁnancial services employment in Frankfurt, for example, would detect a discontinuity between 1998 and

300

Epilogue

1999 that reﬂected a change in deﬁnitions. Failing to correct for this change would distort results of any analysis using employment data as an ofﬁce demand proxy. There is also a case for much more detailed work at individual city level, case study based work that traces the ownership patterns, the nature of occupation and capital ﬂows at building level. The Who Owns the City projects, which analyse ofﬁce ownership and occupation in the City of London demonstrate the beneﬁts of such an approach but have also revealed what effort is required to complete the task. Once again there are complex deﬁnitional issues to be confronted – on what constitutes beneﬁcial ownership and on the domicile of particular stakeholders, for example. Case-based research can also help to explore the interrelationship between urban planning and ofﬁce market outcomes, where local differences in policy can alter the physical form, quantum and location of buildings, in turn affecting rental adjustment and the investment characteristics of the market. Data assembly, then, is a daunting task – but not an impossible one. The problem confronting researchers, though, is that most projects are short and resource constrained. Researchers working in isolation collect the data for their speciﬁc purpose, use it and move on. The next team to address similar problems must repeat the exercise. That is why I have emphasised that what is really needed is a programme of empirical research – which will, over time, build a repository of data where these deﬁnitional and robustness issues have been addressed and where effort can go into analysis and model building rather than primary data collection. To succeed, then, requires a collective effort and, importantly, international cooperation. This will be difﬁcult to achieve: it requires funding, it requires acceptance of a common purpose, it must wrestle with issues of data conﬁdentiality, data access and free rider problems. However, the creation of a more robust, accessible database of international city level real estate data offers huge potential in enabling us to understand how global ofﬁce markets and capital markets function. As I wrote Towers of Capital this data issue loomed large and, it seems to me, it is the next, critical step in advancing our knowledge of this topic.

Cycles, timing and perspective I completed the writing of this book in 2007 and 2008, to the backdrop of the evolving sub-prime crisis and credit crunch. It would have been easy to let those events inﬂuence my interpretation of the relationship between ofﬁce markets and ﬁnance. I hope that this has not happened – it is far too early to know the fundamental impacts of the events triggered off in the summer

Epilogue

301

of 2007 and the extent to which those impacts have spread across cities and regions. I should stress that the basic model outlined here has not been shaped by the credit crunch – I had set out the structure of the book well in advance of the downturn and the basic ideas can be found in earlier published work. Indeed, I had been sounding warnings about the level of asset prices and high gearing levels in the UK commercial real estate market for some time,2 arguing that the property market innovation could exacerbate the impact of shocks. A more difﬁcult question of perspective is the extent to which the innovations in real estate described in Chapter 8 genuinely represent a fundamental shift in the organisation of commercial property. It is easy to believe that one is experiencing revolutionary times and commentators are quick to pronounce a ‘new paradigm’ – usually just before the old paradigm returns to reassert its potency. Debt securitisation, the creation of private real estate funds with a global reach, the spread of tax-efﬁcient public REITs, and synthetic property investment vehicles all seem to point to signiﬁcant changes in the nature of real estate investment with risk and return unbundled and distributed and entry barriers lowered, if not removed. And yet, there remain historic resonances. The 1990s property crash was preceded by a ﬂurry of ﬁnancial engineering activity, innovative off balance sheet structures and attempts to create new investment vehicles, with the impact of the downturn magniﬁed by debt structures and by global contagion effects; the 1973 crash was fuelled and exacerbated by niche property market lenders raising credit in the capital markets to lend on in complex layered ways. Even the Barings crisis in the late nineteenth century has many of the characteristics of the sub-prime crisis and credit crunch of the ﬁrst decade of the twenty-ﬁrst century and there are many examples of collective international real estate investment schemes in the earlier era of global capital. It may be that the nature of commercial real estate – its ﬁxed location, its illiquidity, and the scale of investment required – drives attempts to create liquidity and to combine capital sources. Innovation such as mortgage-backed securitisation ﬁts within that model – while having real and signiﬁcant impacts both in terms of the functioning of real estate debt markets and the spread of exposure to property risk and return. Even if the changes to the structure of property markets described in Towers of Capital are simply the latest chapter of a long story, I do not think this affects the fundamental proposition. The ofﬁce markets of IFCs are fundamental to the operation and competitiveness of those cities. They are not simply a response to the demand for space by globally oriented ﬁnancial services ﬁrms, they play a more important, causal role, as the physical manifestation of the agglomeration economies and spillover 2

To cite the old joke, I have correctly predicted six of the last two property downturns.

302

Epilogue

effects that determine the need for a core location, as a store of capital and as an investment asset. As the core areas of IFCs become more and more functionally specialised, the fortunes of the ofﬁce markets in those cities become more and more tied into global capital markets, with the performance of ﬁnancial ﬁrms in those cities, in turn, linked to real estate market behaviour through direct and indirect investment in the built form. The new debt and equity vehicles allow that exposure to be parcelled up and distributed – but the spatial concentration of ﬁnancial activity in the IFCs ensures that much of that exposure remains in global ﬁnancial cities, further linking them together. By implication, then, global ﬁnancial ofﬁce markets are likely to exhibit high volatility and vulnerability to shocks – which will be felt in many markets across the world, with real economic effects.

Bibliography Abdel-Rahman, H. (1996) When Do Cities Specialise in Production? Regional Science and Urban Economics, 26, 1–22. Addison, T., Le Billon, P. and Murshed, M. (2001) Finance in Conﬂict and Reconstruction, WIDER Discussion Paper No. 2001/44, United Nations University. Adjaouté, K. and Danthine, J.-P. (2002) European Financial Integration and Equity Returns: A Theory-Based Assessment, Paper presented at Second ECB Central Banking Conference, October 2002. Adler, P. (1990) Shared Learning, Management Science, 36, 938–957. Agnes, P. (2000) The ‘End of Geography’ in Financial Services? Local Embeddedness and Territorialization in the Interest Rate Swaps Industry, Economic Geography, 76(4), 347–366. Alderson, A. and Beckﬁeld, J. (2004) Power and Position in the World City System, American Journal of Sociology, 109, 811–851. Allen, F. and Gale, D. (1999) Bubbles, Crisis and Policy, Oxford Review of Economic Policy, 15(3), 9–18. Allen, F. and Gale, D. (2000) Financial Contagion, Journal of Political Economy, 108(1), 1–33. Ambrose, P. and Colenutt, B. (1975) The Property Machine, Harmondswoth: Penguin. Amel, D., Barnes, C., Panetta, F. and Salleo, C. (2004) Consolidation and Efﬁciency in the Financial Sector: A Review of the International Evidence, Journal of Banking and Finance, 28, 2493–2519. Amin, A. and Thrift, N. (1992) Neo-Marshallian Nodes in Global Networks, International Journal of Urban and Regional Research, 16, 571–587. Ang, A. and Bekaert, G. (2002) International Asset Allocation with Regime Shifts, Review of Financial Studies, 15, 1137–1187. Ang, A. and Chen, J. (2002) Asymmetric Correlations of Equity Portfolios, Journal of Financial Economics, 63, 443–494. Antwi, A. and Henneberry, J. (1995) Developers, Non-Linearity and Asymmetry in the Development Cycle, Journal of Property Research, 12, 217–239. Auerbach, F. (1913) Das Gesetz der Bevolkerungskoncentration, Petermanns Geographische Mitteilungen, 59, 74–76. Baca, S., Garb, B. and Weiss, R. (2000) The Rise of Sector Effects in Major Equity Markets, Financial Analysts Journal, September–October, 34–40. Bae, K.-H., Karolyi, A. and Stulz, R. (2003) A New Approach to Measuring Financial Contagion, Review of Financial Studies, 16, 717–763. Baele, L. and Vennet, R. (2001) European Stock Market Integration and EMU, Mimeo, Department of Financial Economics, Ghent University. Ball, M. (1998) Institutions in British Property Research: A Review, Urban Studies, 35, 1501–1517. Ball, M. (2002) Cultural Explanation of Regional Property Markets: A Critique, Urban Studies, 39(8), 1453–1469. Ball, M. and Grilli, M. (1997) UK Commercial Property Investment: Timeseries Characteristics and Modelling Strategies, Journal of Property Research, 14, 279–296.

304

Bibliography

Ball, M., Lizieri, C. and MacGregor, B.D. (1998) The Economics of Commercial Property Markets, London: Routledge. Ball, M., Morrison, T. and Wood, A. (1996) Structures Investment and Economic Growth: A Long-Term International Comparison, Urban Studies, 33, 1687–1706. Ball, M. and Tsolacos, S. (2002) UK Commercial Property Forecasting: The Devil is in the Data, Journal of Property Research, 19, 13–38. Ball, M. and Wood, A. (1996) Does Building Investment Affect Economic Growth, Journal of Property Research, 13, 99–114. Bank for International Settlements (BIS) (1993) 63rd Annual Report, Basel: Bank for International Settlements. Bank for International Settlements (BIS) (2007) Triennial Central Bank Survey, December 2007, Basel: Bank for International Settlements. Bardhan, A., Edelstein, R. and Tsang, D. (2008) Global Financial Integration and Real Estate Security Returns, Real Estate Economics, 36, 285–311. Barkham, R. and Ward, C. (1999) Investor Sentiment and Noise Traders: A Study of the Discount to Net Asset Value in Listed Property Companies in the UK, Journal of Real Estate Research, 18(2), 291–312. Barras, R. (1983) A Simple Theoretical Model of the Ofﬁce Development Cycle, Environment and Planning A, 15, 1381–1394. Barras, R. (1994) Property and the Economic Cycle: Building Cycles Revisited, Journal of Property Research, 11, 183–197. Barras, R. (2005) A Building Cycle Model for an Imperfect World, Journal of Property Research, 22(3), 63–96. Barras, R. and Ferguson, D. (1987a) Dynamic Modelling of the Building Cycle: 1. Theoretical Framework, Environment and Planning A, 19, 353–367. Barras, R. and Ferguson, D. (1987b) Dynamic Modelling of the Building Cycle: 2: Empirical Results, Environment and Planning A, 19, 493–520. Baum, A. (2002) Commercial Real Estate Investment, London: Estates Gazette. Baum, A (2008) The Emergence of Real Estate Funds, in Peterson, A. (ed.), Real Estate Finance: Law, Regulation and Practice, London: LexisNexis. Baum, A. and Lizieri, C. (1999) Who Owns the City? Ofﬁce Ownership and Overseas Investment in the City of London, Real Estate Finance, 16(1), 87–100. Baum, A. and Turner, N. (2004) Retention rates, Reinvestment and Depreciation in European Ofﬁce Markets, Journal of Property Investment & Finance, 22(3), 214–235. Beaverstock, J. (2004) Managing Across Borders: Knowledge Management and Expatriation in Professional Service Law Firms, Journal of Economic Geography, 4(2), 157–179. Beaverstock, J., Smith, R. and Taylor, P. (2000) World-City Network: A New Metageography?, Annals of the American Association of Geographers, 90(1), 123–134. Beaverstock, J., Taylor, P. and Smith, R. (1999) A Roster of World Cities, Cities, 16(6), 445–458. Beckers, S., Connor, G. and Curds, R. (1996) National versus Global Inﬂuences on Equity Returns, Financial Analysts Journal, 52(2), 31–39. Bell, A., Brooks, C. and Dryburgh, P. (2007) Interest Rates and Efﬁciency in Medieval Wool Foward Contracts, Journal of Banking and Finance, 31(2), 361–380. Berger, A., Demsetz, R. and Strahan, P. (1999) The Consolidation of the Financial Services Industry: Causes, Consequences, and Implications for the Future, Journal of Banking & Finance, 23, 135–194. Bernanke, B. and Gertler, M. (1989) Agency Costs, Net Worth and Business Fluctuations, American Economic Review, 79, 14–31.

Bibliography

305

Björklund, K. and Söderberg, B. (1999) Property Cycles, Speculative Bubbles and the Gross Income Multiplier, Journal of Real Estate Research, 18, 151–174. Blake, D., Lizieri, C. and Matysiak, G. (2000) Forecasting Ofﬁce Supply and Demand, London: RICS Research Foundation, p. 36. Bleaney, M., Bougheas, S. and Skamnelos, I. (2008) A Model of the Interactions Between Banking Crises and Currency Crises, Journal of International Money and Finance, 27, 695–706. Bond, S., Crosby, N., Hwang, S., Key, T., Lizieri, C., Matysiak, G., McAllister, P. and Ward, C. (2004) Liquidity in Commercial Property Markets, London: Investment Property Forum, p. 116. Bond, S., Karolyi, G. and Sanders, A. (2003) International Real Estate Returns, a Multifactor, Multi country approach, Real Estate Economics, 31(3), 481–500. Bond, S.A., Hwang, S., Lin, Z. and Vandell, K.D. (2007) Marketing Period Risk in a Portfolio Context: Theory and Empirical Estimates from the UK Commercial Real Estate Market, Journal of Real Estate Finance and Economics, 34(4), 447–461. Bordo, M. and Eichengreen, B. (1999) Is Our Current Economic Environment Unusually Crisis Prone?, Mimeo, Sydney: Reserve Bank of Australia, 57pp. Bordo, M. and Murshid, R. (2000) Are Financial Crises Becoming Increasingly More Contagious? What is the Historical Evidence on Contagion? NBER Working Paper W7900, National Bureau of Economic Research. Bordo, M. and Murshid, R. (2006) Globalization and Changing Patterns in the International Transmission of Shocks in Financial Markets, Journal of International Money and Finance, 25, 655–674. Bordo, R., Eichengreen, B. and Irwin, D. (1999) Is Globalization Today Really Different from Globalisation a Hundred Years Ago?, NBER Working Paper WP7195, Cambridge, MA. Brenner, N. and Keil, R. (2006) The Global Cities Reader, London: Routledge. Brown, G.R. and Schuck, E.J. (1996) Optimal Portfolio Allocations to Real Estate, Journal of Real Estate Portfolio Management, 2(1), 63–73. Bruner, R. and Carr, S. (2007) The Panic of 2007, New York: Wiley. Buch, C. (2005) Distance and International Banking, Review of International Economics, 13(4), 787–804. Buckley, P. and Ghauri, P. (2004) Globalisation, Economic Geography and the Strategy of Multinational Enterprises, Journal of International Business Studies, 35, 81–98. Callender, M., Devaney, S., Sheahan, A. and Key, T. (2007) Risk Reduction and Diversiﬁcation in UK Commercial Property Portfolios, Journal of Property Research, 24, 355–375. Campbell, R., Koedijk, K. and Kofman, P. (2002) Increasing Correlation in Bear Markets: A Downside Risk Perspective, Financial Analysts Journal, 58, 87–94. Candelion, B., Hecq, A. and Veischoor, W. (2005) Measuring Common Cycle Features During Financial Turmoil, Journal of International Money and Finance, 24, 1317–1334. Canter, T. and Gordon, J. (1999) Alternative Workplace Strategies, Wharton Real Estate Review, Spring, 21. Carlos, A. and Nicholas, S. (1988) Giants of an Earlier Capitalism, The Business History Review, 62(3), 398–419. Carn, N., Black, R. and Rabianski, J. (1999) Operational and Organization Issues Facing Corporate Real Estate Executives and Managers, Journal of Real Estate Research, 17, 281–299.

306

Bibliography

Case, B., Goetzmann, W.N. and Rouwenhorst, K.G. (1999) Global Real Estate Markets – Cycles and Fundamentals, Yale International Center for Finance, unpublished Working Paper. Case, B., Goetzmann, W.N. and Wachter, S. (1997) The Global Commercial Property Market Cycles: A Comparison Across Property Types, Paper presented at the AREUEA 6th International Real Estate Conference, University of California at Berkeley. Cassis, Y. (2006) Capitals of Capital, Cambridge: Cambridge University Press. Castells, M. (1989) The Informational City: Information Technology, Economic Restructuring, and the Urban-Regional Process, Oxford: Blackwell. Castells, M. (1999) The Informational City, Cambridge, MA: Blackwell. Cavaglia, S., Brightman, C. and Akev, C. (2000) The Increasing Importance of Industry Factors, Financial Analysts Journal, September–October, 41–53. CBRE (2007) Global Market Rents, May 2007, London: CB Richard Ellis Global Research. CBRE (various) Central London Ofﬁce Market Report, Quarterly, London: CB Richard Ellis Global Research. Cetorelli, N., Hortle, B., Morgan, D., Peristianio, S. and Santos, P. (2007) Trends in Financial Market Concentration and Their Implications for Market Stability, FRBNY Economic Policy Review, March 2007, Federal Reserve Bank of New York. Champy, J. (1995) Re-Engineering Management: The Mandate for New Leadership, London: Harper Collins. Charlot, S. and Duranton, G. (2006) Cities and Workplace Communication: Some Quantitative French Evidence, Urban Studies, 43(8), 1365–1394. Cheshire, P. (2006) Resurgent Cities, Urban Myths and Policy Hubris: What We Need to Know, Urban Studies, 43, 1231–1246. Chinitz, B. (1961) Contrast in Agglomeration: New York and Pittsburgh, American Economic Review, 51, 279–289. Choi, S., Park, D. and Tschoegl, A. (2003) Banks and the World’s Major Banking Centers, 2000, Review of World Economics, 139(3), 550–568. Christaller, W. (1933) Die Zentralen Orte in Süddeutschland, Trans. Baskin, C., Englewood Cliffs, NJ: Prentice-Hall. Collett, D., Lizieri, C. and Ward, C. (2003) Timing and the Holding Periods of Institutional Real Estate, Real Estate Economics, 31(2), 205–222. Collyns, C. and Senhadji, A. (2002) Lending Booms, Real Estate Bubbles and the Asian Crisis, International Monetary Fund WP 02/20, New York. Combes, P., Duranton, G. and Overman, H. (2005) Agglomeration and the Adjustment of the Spatial Economy, Papers in Regional Science, 84(3), 311–349. Conover, C., Friday, H. and Sirmans, G. (2002) Diversiﬁcation Beneﬁts from Foreign Real Estate Investments, Journal of Real Estate Portfolio Management, 8(1), 17–25. Corbridge, S., Martin, R. and Thrift, N. (eds) (1994) Money, Power and Space, Oxford: Blackwell. Corporation of London (2001) Ofﬁce Stock in the City of London, London: Department of Planning and Transportation, Corporation of London. Corsetti, G., Pericoli, M. and Sbracia, M. (2005) ‘Some Contagion, Some Interdependence’: More Pitfalls in Tests of Financial Contagion, Journal of International Money and Finance, 24, 1177–1199. Coval, J. and Moskowitz, T. (1999) Home Bias at Home: Local Equity Preference in Domestic Portfolios, Journal of Finance, 54, 2045–2074.

Bibliography

307

D’Arcy, E. and Keogh, G. (1997) Towards a Property Market Paradigm of Urban Change, Environment and Planning A, 29, 685–706. Daniels, P. (1975) Ofﬁce Location: An Urban and Regional Study, London: G Bell and Sons. Daniels, P. (1983) Business Service Ofﬁces in British Provincial Cities: Location and Control, Environment and Planning A, 5, 1101–1120. Daniels, P. (1985) Service Industries: A Geographical Appraisal, London: Routledge. D’Arcy, E. and Keogh, G. (1998) Territorial Competition and Property Market Process: An Exploratory Analysis, Urban Studies, 35, 1215–1230. D’Arcy, E. and Keogh, G. (1999) The Property Market and Urban Competitiveness: A Review, Urban Studies, 36, 917–928. D’Arcy, E. and Keogh, G. (2002) The Market Context of Property Development Activity, in Guy, S. and Henneberry, J. (eds), Development and Developers, Oxford: Blackwell. D’Arcy, É., McGough, T. and Tsolacos, S. (1999) An Econometric Analysis and Forecasts of the Ofﬁce Rental Cycle in the Dublin Area, Journal of Property Research, 16, 309–321. D’Arcy, É., Tsolacos, S. and McGough, T. (1997) An Empirical Investigation of Retail Rents in Five European Cities, Journal of Property Valuation and Investment, 15, 308–322. Davis, C. (2007) Banking Concentration, Financial Stability and Public Policy, Paper presented to the Reserve Bank of Australia Conference, August 2007. Davis, E. (1995) Debt, Financial Fragility and Systemic Risk, Oxford: Oxford University Press. Davis, E. and Zhu, H. (2004) Bank Lending and Commercial Property Cycles, BIS Working Paper 150, Bank for International Settlements, Basel. Dehesh, A. and Pugh, C. (1999) The Internationalization of Post-1980 Property Cycles and the Japanese ‘Bubble’ Economy, 1986–96, International Journal of Urban and Regional Research, 23, 147–164. De Magalhães, C. (2001) International Property Consultants and the Transformation of Local Markets, Journal of Property Research, 18(2), 99–121. Deloitte and Touche LLP (1995) Space Needs of Corporate Headquarters in America, Chicago, IL: Deloitte and Touche. Deming, W. (1990) Out of the Crisis, Cambridge: Cambridge University Press. Department of Commerce (1998) The Emerging Digital Economy, Washington, DC: Department of Commerce. Derudder, B. and Taylor, P. (2005) The Cliqueishness of World Cities, Global Networks, 5(1), 71–91. Derudder, B. and Witlox, F. (2005) An Appraisal of the Use of Airline Data in Assessing the World City Network, Urban Studies, 42(13), 2371–2388. Derudder, B., Taylor, P., Witlox, F. and Catalano, G. (2003) Hierarchical Tendencies and Regional Patterns in the World City Network, Regional Studies, 37, 875–886. Deutsche Grundbesitz (1998) Real Estate Markets in European Financial Centres, Frankfurt, Deutsche Grundbesitz/Deutsche Bank Research. Devaney, S. (forthcoming, 2009) Long Run Relationships between Ofﬁce Rents and the Economy in the UK, Doctoral Thesis, Henley Business School, University of Reading. Devaney, S. and Lizieri, C. (2005) Individual Assets, Market Structure and the Drivers of Return, Journal of Property Research, 22, 287–307.

308

Bibliography

DiPasquale, D. and Wheaton, W. (1992) The Markets for Real Estate Assets and Space: A Conceptual Framework, Journal of the American Real Estate and Urban Economics Association, 20, 181–197. DiPasquale, D. and Wheaton, W. (1996) Urban Economics and Real Estate Markets, Englewood Cliffs, NJ: Prentice-Hall. Dornbusch, R., Park, Y. and Claessens, S. (2000) Contagion: Understanding How It Spreads, World Bank Research Observer, 15(2), 177–197. Dunning, J. (ed.) (2000) Regions, Globalization and the Knowledge Based Economy, Oxford: Oxford University Press. Dunning, J. and Morgan, V. (eds) (1971) An Economic Study of the City of London, London: Allen & Unwin. Duranton, G. and Puga, D. (2000) Diversity and Specialisation in Cities: Why, Where and When Does it Matter?, Urban Studies, 37(3), 533–555. Duranton, G. and Puga, D. (2005) From Sectoral to Functional Urban Specialisation, Journal of Urban Economics, 57, 343–370. Eaton, J. and Eckstein, Z. (1997) Cities and Growth: Theory and Evidence from France and Japan, Regional Science and Urban Exonomics, 27, 443–474. Economist, The (1998) Capitals of Capital, a Survey of Financial Centres, The Economist, May 9th, 1998. Eichholtz, P. (1996) Does International Diversiﬁcation Work Better for Real Estate than for Stocks and Bonds?, Financial Analysts Journal, 52(1), 56–62. Eichholtz, P., Huisman, R., Koedijk, K. and Schuin, L. (1998) Continental Factors in International Real Estate Returns, Real Estate Economics, 26(3), 493–509. Elton, E., Gruber, M., Brown, S. and Goetzmann, W. (2007) Modern Portfolio Theory and Investment Analysis, Hoboken, NJ: Wiley. Englund, P. (1999) The Swedish Banking Crisis: Roots and Consequences, Oxford Review of Economic Policy, 15(3), 80–97. Englund, P., Gunnelin, A., Hendershott, P. and Soderberg, B. (2008) Adjustment in Commercial Property Space Markets: Taking Long-Term Leases and Transaction Costs Seriously, Real Estate Economics, 36, 81–109. Ennis, R. and Burik, P. (1991) Pension Fund Real Estate Investment under a Simple Equilibrium Pricing Model, Financial Analysts Journal, 47(3), 20–30. Evans, A. (1983) The Determination of the Price of Land, Urban Studies, 20, 119–129. Fainstein, S. (1993) The City Builders: Property Politics and Planning in New York and London, Oxford: Blackwell. Fainstein, S. (2001a) Inequality in Global City Regions, in Scott, A. (ed.), Global City Regions, Trends, Theory, Policy, Oxford: Oxford University Press. Fainstein, S. (2001b) The City Builders: Property Politics and Planning in New York and London, 2nd edition, Oxford: Blackwell. Farrelly, K. and Sanderson, B. (2005) Modelling Regime Shifts in the City of London Ofﬁce Rental Cycle, Journal of Property Research, 22, 325–344. Fisher, J. (1992) Integrating Research on Markets for Space and Capital, Real Estate Economics, 20, 161–180. Fisher, J., Gatzlaff, D., Geltner, D. and Haurin, D. (2003) Controlling for the Impact of Variable Liquidity in Commercial Real Estate Price Indices, Real Estate Economics, 31(2), 269–303. Fisher, J.D. (2005) New Strategies for Commercial Real Estate Investment and Risk Management, Journal of Portfolio Management, September, 154–161. Fitch, R. (1993) The Assassination of New York. New York: Verso.

Bibliography

309

Fogler, H. (1984) 20% in Real Estate: Can Theory Justify It?, Journal of Portfolio Management, 10(2), 6–13. Folkerts-Landau, D. and Ito, T. (1995) International Capital Markets: Developments, Prospects and Policy Issues, Washington: International Monetary Fund. Forbes, K. and Rigobon, R. (2002) No Contagion, Only Interdependence: Measuring Stock Market Co-Movements, Journal of Finance, 57(5), 2223–2261. Fraser, P. and Oyefeso, O. (2005) US, UK and European Stock Market Integration, Journal of Business Finance and Accounting, 32(1), 306–686. Freimann, E. (1998) European Integration and Country Allocation in Europe, Financial Analysts Journal, September–October, 32–41. Friedmann, J. (1986) The World City Hypothesis, Growth and Change, 17, 69–83. Friedmann, J. and Wolff, G. (1982) World City Formation, an Agenda for Research and Action, International Journal of Urban and Regional Research, 6(3), 309–344. Frobel, F., Heinrichs, J. and Kreye, O. (1980) The New International Division of Labour, Cambridge, MA: Cambridge University Press. Fujita, M., Krugman, P. and Mori, T. (1999a) On the Evolution of Hierarchical Urban Systems, European Economic Review, 43, 209–251. Fujita, M., Krugman, P. and Venables, A. (1999b) The Spatial Economy, Cambridge, MA: MIT. Gallimore, P. and McAllister, P. (2005) Expert Judgement in the Processes of Commercial Property Market Forecasting, Journal of Property Research, 21(4), 337–360. Ganesan, S., Malter, A. and Rindﬂeisch, A. (2005) Does Distance Still Matter? Geographic Proximity and New Product Development, Journal of Marketing, 69(4), 44–60. Garber, P. (2000) Famous First Bubbles: The Fundamentals of Early Manias, Cambridge, MA: MIT. Gardiner, C. and Henneberry, J. (1988) The Development of a Simple Regional Ofﬁce Rent Prediction Model, Journal of Valuation, 7, 36–52. Gardiner, C. and Henneberry, J. (1991) Predicting Regional Ofﬁce Rents Using Habit Persistence Theories, Journal of Property Valuation and Investment, 9, 215–226. Garreau, J. (1991) Edge City: Life on the New Frontier, New York: Doubleday. Gehrig, T. (1998) Cities and the Geography of Financial Centres, Centre for Economic Policy Research, London, Discussion Paper No. 1894. Geltner, D., Miller, N., Clayton, J. and Eichholtz, P. (2007) Commercial Real Estate Analysis and Investments, Cincinnati, OH: South-Western Educational Publishing. Gerald Eve (1997) Overcrowded Under-Utilised or Just Right?, Report of a Study of Ofﬁce Occupational Densities. London: Gerald Eve in association with the RICS. GHK International Ltd (2000) The Ofﬁce Market in Downtown Toronto, The Future of Downtown Toronto, Background Studies, #2, Toronto: City of Toronto. Ghosh, C., Guttery, R. and Sirmans, C. (1994) The Olympia & York Crisis: Effects on the Financial Performance of US & Foreign Banks, Journal of Property Finance, 5, 5–46. Gibler, K., Black, R. and Moon, K. (2002) Time, Place, Technology and Corporate Real Estate Strategy, Journal of Real Estate Research, 24, 235–262. Gibson, V. (1998) Changing Business Practice and its Impact on Occupational Property Portfolios, Research Paper, Right Space Right Price, London: RICS Books. Gibson, V. (2000) Property Portfolio Dynamics: The Flexible Management of Inﬂexible Assets, Facilities, 18, 150–154.

310

Bibliography

Gibson, V. and Lizieri, C. (1999) New Business Practices and the Corporate Property Portfolio: How Responsive is the UK Property Market?, Journal of Property Research, 16(3), 201–218. Gibson, V. and Lizieri, C. (2000) Space or Service? The Role of Executive Suites in Corporate Real Estate Portfolios and Ofﬁce Markets, Real Estate Finance Journal, Spring 2000, 21–29. Gibson, V. and Lizieri, C. (2001) Friction and Inertia: Corporate Change, Real Estate Portfolios and the UK Ofﬁce Market, Journal of Real Estate Research, 21(1/2), 59–79. Gibson, V. and Louargand, M. (2001) The Workplace Portfolio as Contractual Arrangements, in Joroff, M. and Bell, M. (eds), The Agile Workplace: Supporting People and Their Work, Boston: Gartner and Massachusetts Institute of Technology, pp. 37–55. Gibson, V., Hedley, C., Proctor, A. and Fennell, B. (2000) Evaluating Ofﬁce Space Needs and Choices, A Research Report for MWB, Reading: University of Reading, http:// www.reading.ac.uk/LM/LM/ofﬁcespace.pdf Gibson, V., Lizieri, C. and Oughton, M. (2001) The Impact of Flexible Working on Central Government Core Ofﬁce Space, London: Property Advisers to the Civil Estate. Glaeser, E., Sheinkman, J. and Sheifer, A. (1995) Economic Growth in a Cross-section of Cities, Journal of Monetary Economics, 36, 117–143. Goddard, J. (1973) Ofﬁce Linkages and Location, Oxford: Pergamon. Goddard, J. (1975) Ofﬁce Location in Urban and Regional Development, Oxford: Oxford University Press. Goddard, J. and Pye, R. (1977) Telecommunications and ofﬁce location, Regional Studies, II, 19–30. Goetzmann, W. and Wachter S. (1995) The Global Real Estate Crash: Evidence from an International Database, Working Paper, the Yale School of Management, July. Goetzmann, W., Li, L. and Rouwenhorst, G. (2005) Long Term Market Correlation, Journal of Business, 78(1), 1–38. Goetzmann, W., Ukhov, A. and Zhu, N. (2007) China and the World Financial Markets 1870–1939: Modern Lessons from Historical Globalization, The Economic History Review, 60(2), 267–312. Goldstein, M. and Folkerts-Landau, D. (1993) International Capital Markets Part II: Systemic Issues in International Finance, Washington: IMF. Gordon, I. (1999) Internationalisation and Urban Competition, Urban Studies, 36(5/6), 1001–1016. Gordon, I. and McCann, P. (2000) Industrial Clusters: Complexes, Agglomeration and/or Social Networks?, Urban Studies, 37(3), 513–532. Gordon, I., Haslam, C., McCann, P. and Scott-Quinn, B. (2005) Off-Shoring and the City of London, London: Corporation of London. Gordon, J.N., Canter, T.A. and Webb, J.R. (1998) The Effect of International Real Estate Securities on Portfolio Diversiﬁcation, Journal of Real Estate Portfolio Management, 4(2), 83–92. Grenadier, S. (1996) The Strategic Exercise of Options: Development Cascades and Overbuilding in Real Estate Markets, Journal of Finance, 51, 1653–1679. Graham, S. (1999) Global Grids of Glass: On Global Cities, Telecommunications and Planetary Urban Networks, Urban Studies, 36(5–6), 929–949. Graham, S. and Marvin, S. (1996) Telecommunications and the City, London: Routledge. Granovetter, M. (1985) Economic Action and Social Structure: The Problem of Embeddedness, American Journal of Sociology, 91, 481–510.

Bibliography

311

Grenadier, S.R. (1995) The Persistence of Real Estate Cycles, Journal of Real Estate Finance and Economics, 10, 95–119. Guy, S. and Henneberry, J. (2002) Bridging the Divide? Complementary Perspectives, in Guy, S. and Henneberry, J. (eds), Development and Developers, Oxford: Blackwell. Haila, A. (1997) The Neglected Builder of Global Cities, reprinted in Brenner, N. and Keil, R. (2006) The Global Cities Reader, London: Routledge. Hakfoort, J. and Lie, R. (1996) Ofﬁce Space per Worker: Evidence from Four European Markets, Journal of Real Estate Research, 11, 183–196. Hall, P. (1966) The World Cities, New York: McGraw-Hill. Hall, P. (1984) The World Cities, London: Orion. Hall, R. (2001) Struggling to Understand the Stock Market, The American Economic Review, 91, 1–11. Hamel, G. and Prahalad, C. (1989) Strategic Intent, Harvard Business Review, May–June, 63–76. Hamelink, F. and Hoesli, M. (2004) What Factors Determine International Real Estate Security Returns, Real Estate Economics, 32, 437–462. Hamelink, F., Hoesli, M., Lizieri, C. and MacGregor, B. (2000) Homogenous Commercial Property Market Groupings and Portfolio Construction in the UK, Environment and Planning A, 32(2), 323–344. Hammer, M. (1990) Re-Engineering Work: Don’t Automate, Obliterate!, Harvard Business Review, July–August, 104–112. Hamnett, C. (1994) Social Polarisation in Global Cities: Theory and Evidence, Urban Studies, 31, 401–424. Harris, R. (2005) Property and the Ofﬁce Economy, London: Estates Gazette Books. Hartmann, P., Straetmans, S. and de Vries, C. (2004) Asset Market Linkages in Crisis Periods, The Review of Economics and Statistics, 86, 313–336. Hau, H. (2001) Location Matters: An Examination of Trading Proﬁts, Journal of Finance, 56(5), 1959–1983. Healey, P. (1992) An Institutional Model of the Development Process, Journal of Property Research, 9, 33–44. Healey, P. and Barrett, S. (1990) Structure and Agency in Land and Property Development Processes: Some Ideas for Research, Urban Studies, 27, 89–104. Heinman, J. and Lord Alexander (1997) Global Institutions, National Supervision and Systemic Risk, Bank of England Financial Stability Review, 3, 82–91. Hendershott, P. (1995) Real Effective Rent Determination: Evidence from the Sydney Ofﬁce Market, Journal of Property Research, 12(2), 127–135. Hendershott, P. (1996) Rental Adjustment and Valuation in Overbuilt Markets: Evidence from the Sydney Ofﬁce Market, Journal of Urban Economics, 39(1), 51–67. Hendershott, P. (2000) Property Asset Bubbles: Evidence from the Sydney Ofﬁce Market, Journal of Real Estate Finance and Economics, 20(1), 67–81. Hendershott, P. and Kane, E. (1992) Causes and Consequences of the 1980s Commercial Construction Boom, Journal of Applied Corporate Finance, 5, 61–70. Hendershott, P. and MacGregor, B. (2005a) Investor Rationality: Evidence from UK Property Capitalization Rates, Real Estate Economics, 33(2), 299–322. Hendershott, P. and MacGregor, B. (2005b) Investor Rationality: An Analysis of NCREIF Commercial Property Data, Journal of Real Estate Research, 27(4), 445–475. Hendershott, P., Hendershott, R. and Ward, C. (2003) Corporate Equity and Commercial Property Market ‘Bubbles’, Urban Studies, 40(5/6), 993–1009.

312

Bibliography

Hendershott, P., Lizieri, C. and MacGregor, B. (2008) Asymmetric Space Market Adjustment in the London Ofﬁce Market, Papers in Real Estate & Planning 07/08, The University of Reading, p. 34. Hendershott, P., Lizieri, C. and Matysiak, G. (1999) The Workings of The London Ofﬁce Market, Real Estate Economics, 27(2), 365–387. Hendershott, P., MacGregor, B. and Tse, R.Y.C. (2002) Estimation of the Rental Adjustment Process, Real Estate Economics, 30(2), 165–183. Henderson, J. (1974) The Sizes and Types of Cities, American Economic Review, 64, 640–656. Henneberry, J. (1999) Convergence and Difference in Regional Ofﬁce Development Cycles, Urban Studies, 36(9), 1439–1465. Henneberry, J. and Rowley, S. (2000) Property Market Processes and Development Outcomes in Cities and Regions, RICS Research Foundation Research Papers, 3(9), 1–59. Henneberry, J. and Rowley, S. (2002) Developers’ Decisions and Property Market Behaviour, in Guy, S. and Henneberry, J. (eds), Development and Developers, Oxford: Blackwell. Herring, R. and Wachter, S. (1999) Real Estate Booms and Banking Busts: An International Perspective, School Working Paper 99–27, University of Pennsylvania, Wharton. Herring, R. and Wachter, S. (2002) Bubbles in Real Estate Markets, Zell/Lurie Real Estate Center Working Paper 402, University of Pennsylvania. Higgins, M. and Osler, C. (1999) Asset Market Hangovers and Economic Growth: The OECD 1984–1993, Oxford Review of Economic Policy, 13(3), 110–134. Hill, R. and Kim, J.W. (2000) Global City and Developmental States: New York, Tokyo and Seoul, Urban Studies, 37(12), 2167–2195. Hobsbawm, E. (1995) The Age of Extremes: A History of the World, 1914–1991, Harmondsworth: Penguin. Hoesli, M. and Hamelink, F. (1997) An Examination of the Role of Geneva and Zurich Housing in Swiss Institutional Portfolios, Journal of Property Valuation and Investment, 15(4), 354–371. Hoesli, M. and Lizieri, C. (2007) Real Estate in the Investment Portfolio, A report for the Investment Strategy Council of the Ministry of Finance, Norway, p. 95. Hoesli, M., Lekander, J. and Witkiewicz, W. (2003) Real Estate in the Institutional Portfolio: A Comparison of Suggested and Actual Weights, Journal of Alternative Investments, 7(3), 53–59. Hoesli, M., Lekander, J. and Witkiewicz, W. (2004) International Evidence on Real Estate as a Portfolio Diversiﬁer, Journal of Real Estate Research, 26(2), 161–206. Hoesli, M., Lizieri, C. and MacGregor, B. (1997) The Spatial Dimensions of the Investment Performance of UK Commercial Property, Urban Studies, 34(9), 1475–1494. Hoesli, M., Lizieri, C. and MacGregor, B. (2008) The Inﬂation Hedging Characteristics of US and UK Investments: A Multi-Factor Error Correction Approach, Journal of Real Estate Finance and Economics, 36(2), 183–206. Holloway, S. and Wheeler, J. (1991) Corporate Headquarters Relocation and Changes in Metropolitan Corporate Dominance, Economic Geography, 67, 55–72. Hoyt, H. (1933) 100 Years of Land Values in Chicago, Chicago: Chicago University Press. Hymer, S. (1979) The Multinational Corporation: A Radical Approach, Cambridge: Cambridge University Press. Ingham, G. (1984) Capitalism Divided? The City and Industry in British Social Development, London: MacMillan.

Bibliography

313

IPM (1986) Flexible Patterns of Work (ed. Curson, C.), London: Institute of Personnel Management. Jacobs, J. (1960) The Economy of Cities, New York: Random House. Jaffe, A.B., Trajtenberg, M. and Henderson, R. (1993) Geographic Localization of Knowledge Spillovers as Evidenced by Patent Citations, Quarterly Journal of Economics, 108, 577–598. Jensen, M. and Meckling, W. (1976) Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure, Journal of Financial Economics, 3, 305–360. JLL (2006) Overseas Investment into UK, London, Jones Lang LaSalle, July 2006. JLL (2007) Overseas Investment into UK, London, Jones Lang LaSalle, August 2007. Johnson, R., Lizieri, C., Soenen, L. and Worzala, E. (2006) Simulating Currency Risk on Private Investments in Real Estate, Journal of Real Estate Portfolio Management, 12(2), 91–102. Joroff, M., Louargand, M., Lambert, S. and Becker, F. (1993) Strategic Management of the Fifth Resource: Corporate Real Estate, Corporate Real Estate 2000 Report #49, Norcross, GA: IDRC. Kaminsky, G, Reinhart, C. and Végh, C. (2003) The Unholy Trinity of Financial Contagion, Journal of Economic Perspectives, 17(4), 51–74. Karakozova, O. (2004) Modelling and Forecasting Ofﬁce Returns in the Helsinki Area, Journal of Property Research, 21, 51–73. Kasa, L. (1992) Common Stochastic Trends in International Stock Markets, Journal of Monetary Economics, 29, 95–124. Kaufman, G. (2001) Emerging Economies and International Financial Centers, Review of Paciﬁc Basin Financial Markets and Policies, 4(4), 365–377. Keil, R. (2006) The Global Cities Reader, London: Routledge. Kennedy, P., Lee, S. and McAllister, P. (2007) Data Dilemmas in Forecasting European Ofﬁce Market Rents, University of Reading: Working Papers in Real Estate & Planning 10/07. Key, T. (2004) The Analysis of Transactions Evidence, Working Paper 3, Liquidity in Commercial Property Markets, London, Investment Property Forum. Kim, S. (1995) Expansion of Markets and the Geographic Distribution of Economic Activities, Quarterly Journal of Economics, 110(4), 881–908. Kimber, L. and Rutherford, R. (1993) Corporate Real Estate Outsourcing: A Survey of Issues, Journal of Real Estate Research, 8, 525–540. Kindelberger, C. (1974) The Formation of Financial Centers: A Study of Comparative Economic History, Princeton Studies in International Finance 36, Princeton University. Kindleberger, C. (1978) Manias, Panics and Crashes: A History of Financial Crises, New York: Wiley. King, A. (1990) Global Cities: Post-Imperialism and the Internationalism of London, London: Routledge. King, A. (2006) World Cities: Global? Postcolonial? Post-imperial? Or Just the Result of Happenstance?, in Brenner, N. and Keil, R. (eds), The Global Cities Reader, London: Routledge. Knight, J., Lizieri, C. and Satchell, S. (2005) Diversiﬁcation When it Hurts? The Joint Distributions of Real Estate and Equity Markets, Journal of Property Research, 22, 309–323. KPMG (1998) Business Cost Comparison: Downtown Core of Major North American Cities, Toronto, KPMG/Ontario Investment Service/City of Toronto.

314

Bibliography

Kutsch, N. (forthcoming, 2009) The Fractional Valuation of Private Real Estate Funds, Unpublished PhD Thesis, University of Reading. Lambert S., Poteete, J. and Waltch, A. (1995) Generating High Performance Corporate Real Estate Service, Corporate Real Estate 2000 Series Report #52, Norcross, GA: IDRC. Laurenceson, J. and Tang, K. (2005) Shanghai’s Development as an International Financial Center, Review of Paciﬁc Basin Financial Markets and Policies, 8(1), 147–166. Leitner, H. (1994) Capital Markets, the Development Industry and Urban Ofﬁce Market Dynamics: Rethinking Building Cycles, Environment and Planning A, 26, 779–802. Ling, D. and Naranjo, A. (2002) Commercial Real Estate Return Performance: A CrossCountry Analysis, Journal of Real Estate Finance and Economics, 24(1/2), 119–142. Liu, C.H. and Mei, J. (1998) The Predictability of International Real Estate Markets, Exchange Rate Risks and Diversiﬁcation Consequences, Real Estate Economics, 26(1), 3–39. Lizieri, C. (1991) The Property Market in a Changing World Economy, Journal of Property Valuation and Investment, 9, 201–214. Lizieri, C. (1992) City Ofﬁce Markets, Global Finance and Local Constraints, Discussion Papers in Property Research No. 2, Property Investment Research Centre, City University, London. Lizieri, C. (1995) Real Estate Ownership, Finance & Risk: Evidence from The Toronto Ofﬁce Market, Proceedings, RICS Research Conference The Cutting Edge, Vol. 1, pp. 13–28. Lizieri, C. (2003) Occupier Requirements in Commercial Real Estate Markets, Urban Studies, 40(5/6), 1151–1169. Lizieri, C. and Finlay, L. (1995) International Portfolio Strategies: Problems and Opportunities, Journal of Property Valuation and Investment, 13(1), 6–21. Lizieri, C. and Kutsch, N. (2006) Who Owns the City 2006: Ofﬁce Ownership in the City of London, Reading: University of Reading Business School and Development Securities plc, pp. 27 + iii. Lizieri, C. and Satchell, S. (1997a) Interactions Between Property and Equity Markets: An Investigation of Linkages in the United Kingdom 1972–1992, Journal of Real Estate Finance and Economics, 15(1), 11–26. Lizieri, C. and Satchell, S. (1997b) Property Company Performance and Real Interest Rates: A Regime-switching Approach, Journal of Property Research, 14(2), 85–97. Lizieri, C. and Ward, C. (2004) Financial Innovation in UK Property Markets: A Review of Trends and Prospects, London: Corporation of London, pp. 116. Lizieri, C., Baum, A. and Scott, A. (2000a) Ownership, Occupation and Risk: A View of the City of London Ofﬁce Market, Urban Studies, 37(7), 1109–1129. Lizieri, C., Baum, A., Williams, N. and Oughton, M. (2000b) Space Race: The Contribution of Property Markets to the Competitiveness of London and Frankfurt, Reading: University of Reading and Development Securities plc. Lizieri, C., Crosby, N., Gibson, V., Murdoch, S. and Ward, C. (1997) Right Space, Right Price? A Study of the Impact of Changing Business Patterns on the Property Market, London: RICS Books. Lizieri, C. McAllister, P. and Ward, C. (2003) Continental Shift: An Analysis of Convergence Trends in European Real Estate, Journal of Real Estate Research, 25(1), 1–22. Lizieri, C., Satchell, S., Worzala, E. and Dacco’, R. (1998) Real Interest Rates and Real Estate Performance: A Comparison of UK and US Markets, Journal of Real Estate Research, 16(3), 339–356.

Bibliography

315

Lizieri, C., Ward, C. and Lee, S. (2001) Financial Innovation in Property Markets: Implications for the City of London, Final Report, London: Corporation of London. London Planning Advisory Council (1991) London: World City Moving into the 21st Century, London: HMSO. Longin, F. and Solnik, B. (2001) Extreme Correlation of International Equity Markets, Journal of Finance, 56(2), 649–675. Lösch, A. (1940) The Economics of Location, trans. Fischer, New Haven, CT: Yale University Press. Lupton, P. and Haynes, B. (2000) Teleworking – The Perception–Reality Gap, Facilities, 16, 92–197. Machimura, T. (1992) The Urban Restructuring Process in Tokyo in the 1980s: Transforming Tokyo into a World City, International Journal of Urban and Regional Research, 16(1), 114–128. Machimura, T. (1998) Symbolic Use of Globalization in Urban Politics in Tokyo, International Journal of Urban and Regional Research, 22, 183–194. Malpezzi, S. and Wachter, S. (2005) The Role of Speculation in Real Estate Cycles, Journal of Real Estate Literature, 13(2), 143–164. Malpezzi, S., Seah, K.-Y. and Shilling, J. (2004) Is It What We Do or How We Do It? New Evidence on Agglomeration Economies and Metropolitan Growth, Real Estate Economics, 32, 265–295. Manning, C. and Roulac, S. (1999) Corporate Real Estate Research within the Academy, Journal of Real Estate Research, 17, 265–280. Manning, C. and Roulac, S. (2001) Lessons from the Past and Future Directions for Corporate Real Estate Research, Journal of Real Estate Research, 22, 7–58. Markham, J. (2002) A Financial History of the United States, New York: ME Sharpe. Markusen, A. (1999) Fuzzy Concepts, Scanty Evidence, Policy Distance, Regional Studies, 33, 869–884. Markusen, A. and Schrock, G. (2006) The Distinctive City: Divergent Patterns in Growth, Hierarchy and Specialisation, Urban Studies, 43(8), 1301–1324. Markusen, J. (1995) The Boundaries of Multinational Enterprises and the Theory of International Trade, Journal of Economic Perspectives, 9(2), 169–175. Marriott, O. (1967) The Property Boom, London: Pan Books. Marshall, A. (1883) Principles of Economics, 1920 edition, London: MacMillan. Martin, R. (1994) Stateless Money, Global Financial Integration and National Economic Autonomy: the End of Geography, in Corbridge, S., Martin, R. and Thrift, N. (eds), Money, Power and Space, Oxford: Blackwell. Masih, R. and Masih, A. (2001) Long and Short Term Dynamic Causal Transmission amongst International Stock Markets, Journal of International Money and Finance, 20, 563–587. Massey, D. (2007) World City, Cambridge: Polity. Massey, D.B. and Catalano, A. (1978) Capital and Land: Landownership by Capital in Great Britain, London: Edward Arnold. MasterCard Worldwide (2007) The Dynamics of Global Cities and Global Commerce, Insights, 2Q 2007. MasterCard Worldwide (2008) Worldwide Centers of Commerce Index 2008, www. mastercardworldwide.com/Insights. McAllister, P. (2008) European Monetary Integration and Convergence in European Real Estate Markets, Unpublished PhD Thesis, University of Reading.

316

Bibliography

McAllister, P. and Lizieri, C. (2006) Monetary Integration and Real Estate Markets: The Impact of the Euro on European Real Estate Equities, Journal of Property Research, 23(4), 281–304. McAllister, P. and Loizou, P. (2007) The Appraisal of Data Centres: Deconstructing the Cash Flow, Working Papers in Real Estate & Planning 04/07, Reading: University of Reading. McDonald, J.F. (2002) A Survey of Econometric Models of Ofﬁce Markets, Journal of Real Estate Literature, 10(2), 223–242. Michayluk, D., Willson, P. and Zurbruegg, R. (2006) Asymmetric Volatility, Correlation and Returns Dynamics between the US and UK Securitized Real Estate Markets, Real Estate Economics, 34(1), 109–131. Mitchener, K. and Weidenmier, M. (2007) The Baring Crisis and the Great Latin American Meltdown of the 1890s, NBER Working Papers 13403, National Bureau of Economic Research. Mouzakis, F. and Richards, D. (2007) Panel Data Modelling of Prime Ofﬁce Rents: A Study of 12 Major European Markets, Journal of Property Research, 24, 31–53. Nanthakumaran, N., Watkins, C. and Orr, A. (2000) Understanding Property Market Dynamics: Insights from Modelling the Supply-side Adjustment Mechanism, Environment and Planning A, 32, 655–671. Nappi-Choulet, I. (2006) The Role and Behaviour of Commercial Property Investors and Developers in French Urban Regeneration: The Experience of the Paris Region, Urban Studies, 43(9), 1511–1535. Nasr, S. (1978) Backdrop to Civil War: The Crisis of Lebanese Capitalism, MERIP Reports 73 (December 1978). Nelson, T. and Nelson, S. (2003) Regional Models for Portfolio Diversiﬁcation, Journal of Real Estate Portfolio Management, 9, 71–88. Newell, G. and Webb, J. (1996) Assessing Risk for International Real Estate Investments, Journal of Real Estate Research, 11(2), 103–115. North, D. (1990) Institutions, Institutional Change and Economic Performance, Cambridge: Cambridge University Press. Nourse, H. and Roulac, S. (1993) Linking Real Estate Decisions to Corporate Strategy, Journal of Real Estate Research, 8, 475–494. Noyelle, T. and Stanback, T. (1984) The Economic Transformation of American Cities, Totawa, NJ: Rowman and Allanheld. O’Brien, R. (1992) Global Financial Integration: The End of Geography, London: RIIA. O’Mara, M. (1999a) Strategic Drivers of Location Decisions for Information-Age Companies, Journal of Real Estate Research, 17, 365–386. O’Mara, M. (1999b) Strategy and Place: Managing Corporate Real Estate and Facilities for Competitive Advantage, New York: The Free Press. Obstfeld, M. and Taylor, A. (2003) Globalization and capital markets, in Bordo, M., Taylor, A. and Williamson, J. (eds), Globalization in Historical Perspective, Chicago: University of Chicago Press. Obstfeld, M. and Taylor, A. (2004) Global Capital Markets: Integration, Crisis and Growth, Cambridge: Cambridge University Press. Ofﬁce of Real Property (1999) The Integrated Workplace, Washington, DC: Government Ofﬁce of Government Wide Policy. Oizumi, E. (1993) Property Finance in Japan: Expansion and Collapse of the Bubble Economy, Discussion Papers in Urban and Regional Science, 82, University of Reading on Property, Urban Studies, 39(8), 1471–1478.

Bibliography

317

Olds, K. and Yeung, H. (2004) Pathways to Global City Formation: A View from the Developmental City State of Singapore, Review of International Political Economy, 11, 489–521. Pavlov, A. and Wachter, S.M. (2002) The Option Value of Non-Recourse Lending and Inﬂated Asset Prices, Wharton Real Estate Working Paper Series, February 27. Peek, J. and Rosengren, E. (1997) Collateral Damage: Effects of the Japanese Real Estate Collapse on Credit Availability and Real Activity in the United States, Working Paper 97–5, Federal Reserve Bank, Boston. Peet, J. (1992) Financial Centres: Rise and Fall, The Economist, June 27 1992. Pericoli, M. and Sbracia, M. (2003) A Primer on Financial Contagion, Journal of Economic Surveys, 17(4), 571–608. Phillips, S. (1993) Derivative Markets: Challenges and Regulatory Issues, World of Banking, 12(2), 4–7. Piershke, B. (2001) Die Organisatorische Gestaltung des Betrieblichen Immobilienmanagements, Köln: Immobilien Informationsverl. Piore, M. and Sabel, C. (1984) The Second Industrial Divide: Possibilities for Prosperity, New York: Basic Books. Porteous, D. (1995) The Geography of Finance, Aldershot: Avebury. Porter, M. (1990) The Competitive Advantage of Nations, London: MacMillan. Porter, M. (1998a) Clusters and the New Economy of Competition, Harvard Business Review, 76, 77–90. Porter, M. (1998b) Location, Clusters and the ‘New’ Micro Economics of Competition, Business Economics, 33, 7–13. Portes, R. and Rey, H. (2005) The Determinants of Cross-Border Equity Flows, Journal of International Economics, 65, 269–296. Prahalad, C. and Hamel, G. (1990) The Core Competence of the Corporation, Harvard Business Review, May–June, 79–91. PREA (2006) An International Comparative Study of the Pension Plan Community and Real Estate Investments, with contributions by Worzala, E., Sirmans, C.F., Lizieri, C., Schulte, K.-W., Ooi, J., Hordijk, A. and Newell, G., Hartford, CT: Pensions Real Estate Association. PricewaterhouseCoopers (2007) UK Economic Outlook, March 2007, London: PWC. Pryke, M. (1991) An International City Going ‘Global’: Spatial Change in the City of London, Environment and Planning D: Society and Space, 9, 197–222. Pryke, M. (1994) Looking Back on the Space of a Boom: (Re)Developing Spatial Matrices in the City of London, Environment and Planning A, 26, 235–264. Quan, D. and Titman, S. (1997) Commercial Real Estate Prices and Stock Market Returns: An International Analysis, Financial Analysts Journal, 53(3), 21–34. Quan, D. and Titman, S. (1999) Do Real Estate Prices and Stock Prices Move Together? An International Analysis, Real Estate Economics, 27, 183–207. Reed, H. (1980) The Ascent of Tokyo as an International Financial Centre, Journal of International Business Studies, 11(3), 19–36. Renaud, B. (1997) The 1985 to 1994 Global Real Estate Cycle: An Overview, Journal of Real Estate Literature, 5(1), 13–44. Renaud, B. (2003) Speculative Behavior in Immature Real Estate Markets: Lessons of the 1997 Asian Financial Crisis, Urban Policy and Research, 21(2), 151–173. Richard Ellis (1994) Tomorrow’s Workplace, A Survey for Richard Ellis by the Harris Research Group, October, London: Richard Ellis.

318

Bibliography

Rodriguez, N. and Feagin, J. (1986) Urban Specialization in the World-System, Urban Affairs Quarterly, 22, 187–219. Ross-Goobey, A. (1992) Bricks and Mortals, London: Century Business. Rouwenhorst, K. (1999) European Equity Markets and the EMU, Financial Analysts Journal, May/June, 57–64. Rutherford, J. (2005) Networks in Cities, Cities in Networks: Territory and Globalisation Intertwined in Telecommunications Infrastructure Development in Europe, Urban Studies, 42(13), 2389–2406. Sassen, S. (1991) The Global City, Princeton: Princeton University Press. Sassen, S. (1999) Global Financial Centres, Foreign Affairs, 78(1), 75–87. Sassen, S. (2000) Cities in the World Economy, Thousand Oaks, CA: Pine Forge. Sassen, S. (2002) Locating Cities on Global Circuits, Environment and Urbanization, 14, 13–20. Saxenian, A. (1994) Regional Advantage, Culture and Competition in Silicon Valley and Route 192, Cambridge, MA: Harvard University Press. Schaefers, W. (1999) Corporate Real Estate Management: Evidence from German Companies, Journal of Real Estate Research, 17(3), 301–320. Schenk, C. (2002) Banks and the Emergence of Hong Kong as an International Financial Centre, Journal of International Financial Markets, Institutions and Money, 12, 321–340. Schulte, K.-W. and Schaefers, W. (1998) Handbuch Corporate Real Estate Management, Köln: Immobilien Informationsverl. Scott, P. (1996) The Property Masters: A History of the British Commercial Property Sector, London: E & FN Spon. Senge, P. (1990) The Fifth Discipline, Australia: Random House. Shiller, R. (2005) Irrational Exuberance, 2nd edition, Princeton: Princeton University Press. Shilton, L. and Stanley, C. (1999) Spatial Patterns of Headquarters, Journal of Real Estate Research, 17, 341–364. Shleifer, A. and Vishny, R. (1997) The Limits of Arbitrage, Journal of Finance, 52(1), 35–55. Short, J., Kim, Y., Kuss, M. and Wells, H. (1996) The Dirty Little Secret of World Cities Research, International Journal of Urban and Regional Research, 20, 697–717. Sirmans, C.F. and Worzala, E. (2003) International Direct Real Estate Investment: A Review of the Literature, Urban Studies 40, 1081–1114. Sivitanides, P., Southard, J., Torto, R. and Wheaton, W. (2001) The Determinants of Appraisal-Based Capitalization Rates, Real Estate Finance, 18, 27–37. Sivitanidou, R. (2002) Ofﬁce Rent Processes: The Case of US Metropolitan Markets, Real Estate Economics, 30, 317–344. Sivitanidou, R. and Sivitanides, P. (2000) Does the Theory of Irreversible Investments Help Explain Movements in Ofﬁce–Commercial Construction? Real Estate Economics, 28(4), 623–661. Smith, D. and Timberlake, M. (2001) World City Networks and Hierarchies 1979–1999: An empirical analysis of global air travel links, American Behavioral Scientist, 44, 1656–1677. Solnik, B. (1991) International Investments, 2nd edition, Reading, MA: AddisonWelsley. Soo, K. (2005) Zipf’s Law for Cities: A Cross-Country Investigation, Regional Science and Urban Economics, 35(3), 239–263.

Bibliography

319

Stevenson, S. (2007) International Real Estate Diversiﬁcation: Empirical Tests using Hedged Indices, Journal of Real Estate Research, 19, 105–131. Tamaki, N. (1995) Japanese Banking, a History 1859–1959, Cambridge: Cambridge University Press. Taylor, P. and Aranya, R. (2008) A Global ‘Urban Roller Coaster’? Connectivity Changes in the World City Network, 2000–2004, Regional Studies, 42, 1–16. Taylor, P. and Walker, D. (2001) World Cities: A First Multivariate Analysis of Their Service Complexes, Urban Studies, 38(1), 23–47. Taylor, P., Catalano, G. and Walker, D. (2002) Measurement of the World City Network, Urban Studies, 39(13), 2367–2376. Terrabit Consulting (2007) The Worldwide Submarine Cable Market, Cambridge, MA: Terrabit. Tsolacos, S. and McGough, T. (1999) Rational Expectations, Uncertainty and Cyclical Activity in the British Ofﬁce Market, Urban Studies, 36(7), 1137–1149. Tsolacos, S., Keogh, G. and McGough, T. (1998) Modelling Use, Investment and Development in the British Ofﬁce Market, Environment and Planning A, 30, 1409–1427. United Nations (2004) World Urbanisation Prospects: The 2003 Revision, New York: UN Department of Economic and Social Affairs. Van Rijckeghem, C. and Weder, B. (2003) Spillovers Through Banking: A Panel Data Analysis of Bank Flows, Journal of International Money and Finance, 22(4), 483–509. Vandell, L. and Lane, J. (1989) The Economics of Architecture and Urban Design: Some Preliminary Findings, Journal of the American Real Estate and Urban Economics Association, 17(2), 235–260. van Winden, W., van den Berg, L. and Pol, P. (1997) European Cities in the Knowledge Economy: Towards a Typology, Urban Studies, 44, 525–549. Venables, A. (1995) Economic Integration and the Location of Firms, American Economic Review, 85, 296–300. Waley, P. (2002) ‘Moving the margins of Tokyo’, Urban Studies, 39, 1533–1550. Waley, P. (2007) Tokyo as a World City: Reassessing the Role of Capital and the State in Urban Restructuring, Urban Studies, 44(8), 1465–1490. Wallerstein, I. (1974) The Modern World System, New York: Academic. Wang, K. and Zhou, Y. (2000) Overbuilding: A Game Theoretic Approach, Real Estate Economics, 28(3), 493–522. Weatherhead, M. (1997) Real Estate in Corporate Strategy, London: MacMillan. Webber, M. (1963) Order in Diversity, Community without Propinquity, in Wingo L. (ed.), Cities and Space, Baltimore: Johns Hopkins. Webber, M. (1964) The Urban Place and the Non-urban Realm, in Webber M. et al. (eds), Explorations in Urban Structure, Philadelphia: University of Pennsylvania Press. Wharf, B. (1994) Vicious Circle: Financial Markets and Commercial Real Estate in the US, in Corbridge, S., Martin, R. and Thrift, N. (eds), Money, Power and Space, Oxford: Blackwell. Wharf, B. (1995) Telecommunications and the Changing Geographies of Knowledge Transformation, Urban Studies, 32, 361–378. Wheaton, W. (1996) A Perspective on Telecommunication Technology and Real Estate Ofﬁce, Industrial and Retail Markets, Real Estate Finance, 13, 13–17. Wheaton, W. (1999) Real Estate ‘Cycles’: Some Fundamentals, Real Estate Economics, 27(2), 209–230.

320

Bibliography

Wheaton, W. and Torto, R. (1988) Vacancy Rates and the Future of Ofﬁce Rents, Journal of the American Real Estate and Urban Economics Association, 16(4), 430–436. Wheaton, W., Torto, R. and Evans, P. (1997) The Cyclic Behaviour of the Greater London Ofﬁce Market, Journal of Real Estate Finance and Economics, 15(1), 77–92. Wheaton, W.C. (1987) The cyclic behaviour of the national ofﬁce market, Journal of the American Real Estate and Urban Economics Association, 15(4), 281–299. White, J. (1998) Old Wine, Cracked Bottle? Tokyo, Paris and the Global City Hypothesis, Urban Affairs Review, 33(4), 451–477. Williams, S. (1992) The Coming of the Groundscrapers, in Budd, L. and Whimster, S. (eds), Global Finance and Urban Living, London: Routledge. Williams, J. (1993) Equilibrium and Options on Real Assets, Review of Financial Studies, 6, 825–850. Wójcik, D. (2007) The International Finance Index and its Derivatives, Mimeo, Oxford University Centre for the Environment. World Federation of Stock Exchanges (2006) Annual Report and Statistics, Paris: WFSE. Worzala, E. and Sirmans, C.F. (2003) Investing in International Real Estate Stocks: A Review of the Literature, Urban Studies, 40, 1115–1149. Worzala, E., Johnson, R. and Lizieri, C. (1997) Currency Swaps as a Hedging Technique for an International Real Estate Investment, Journal of Property Finance, 8(2), 134–151. Wu, W. (1999) City Proﬁle: Shanghai, Cities, 16(3), 207–216. Young, M. (2008) Revisiting Non-normal Real Estate Return Distributions by Property Type in the US, Journal of Real Estate Finance and Economics, 36(2), 233–248. Z/Yen (2007) The Global Financial Centres Index – 2, London: Corporation of London, September 2007. Z/Yen (2008) The Global Financial Centres Index – 3, London: Corporation of London, February 2008. Zaheer, S. and Manrakhan, S. (2001) Concentration and Dispersion in Global Industries: Remote Electronic Access and the Location of Economic Activities, Journal of International Business Studies, 32(4), 667–686. Zhu, H. (2002) The Case of the Missing Real Estate Cycle, BIS Quarterly Review, September 2002, 56–66. Ziobrowski, A. and Ziobrowski, B. (1997) Higher Real Estate Risk and Mixed-Asset Portfolio Performance, Journal of Real Estate Portfolio Management, 3, 107–115. Ziobrowski, A.J., Ziobrowski, B.J. and Rosenberg, S. (1997) Currency Swaps and International Real Estate Investment, Real Estate Economics, 25(2), 223–251. Zipf, G. (1949) Human Behaviour and the Principle of Least Effort, Reading, MA: AddisonWesley. Zukin, S. (1992) The City as a Landscape of Power: London and New York as Global Financial Capitals, in Budd, L. and Whimster, S. (eds), Global Finance and Urban Living, London: Routledge.

Index Abu Dhabi, 80 accessibility, 58 agglomeration economies, 9 inﬂuence of, 52–60 Alcatel, 46 Alonso-style bid-rent models, of urban land-use, 95–6 alpha world cities, 24 amount of aggregate space, 61–2 Amsterdam, 67–9 Amsterdam Stock Exchange, 68 ArcaEx, 45 Asian and Middle Eastern markets, 76–81 Asian IFCs, 76–81 asset market bubbles, 215 asset prices, 129–30 Bahrain, 80 Bank for International Settlements (BIS) data analysis, 37–8 bank M&A across developed economies, 41–2 Bank of America, 79 Bank of England, 68, 72–3 Bank of Japan, 77 bank penetration, 56 banks by assets, 42 Banque de France, 70 Banque de l’Indo Chine, 72 Barings, 68, 78 ﬁnancial crisis, 72 Basel II capital adequacy and solvency regulatory requirements, 212 Battery Park City schemes, 138–9 Beirut, 80 below equilibrium rents (BERI), 131 Berlin, 68, 72 beta world cities, 24 Black Monday, 1987, 188 Bretton Woods Agreement, 65–6 Bretton Woods exchange rate system, 65, 76, 79 broking and securities trading, 116 Brussels, 69

business downturn, 132 business upturn, 132 call centres, 115 capital value, of an ofﬁce building, 149 Castells’s ‘network society’, 22 catchment area, 8 CBOT, 45 central business districts (CBDs), 161 central place theory, 9 Chemical Bank, 139 Christaller’s central place system, 7–9 cities agglomeration economies, 9 command and control, 12–13 congestion and crowding costs, 9, 14 distribution in size, 6 diversiﬁed, 15 edge, 9 ﬁrm growth and ﬁrm trurnover, 12 high-order, 9, 30 impact of shifts in nature of activities, 14–15 and Jacobian economies, 10, 12 lower order, 9 market place, 8 North American and UK, 12–14 outside states, 15–18 proportions of headquarters (HQs) buildings, 12–14 and provision of telecommunications networks, 47 ranked by GDP, 17 settlement hierarchy, 8 spatial distribution of, 6 specialised, 10–12 subordinate, 12 transport costs, 10–11 city, idea of, 5 City of London, 57, 68, 70–71, 73, 84–5, 114, 118, 162, 206 classical location theory, 94 CME, 45 cobweb or corn–hog cycle adjustment process, 130

322

Index

commodity prices, 188 co-movement in asset markets and ﬁnancial market, 188 in global equity markets, 190 competitive international service centre, 62 competitiveness, 31, 63 rating of, 35 consolidation of ﬁnancial services, inﬂuencing factors, 43 construction lag, 131 Cooper Brothers, 71 Crédit Lyonnais, 72 cross-continental strategic alliances, 45 Daiichi, 77 DataStream index, 201 decentralisation, of ﬁnancial activity, 48–50 Deloitte, 71 Deutsche-Asiatische Bank, 78 Deutsche Bank, 48, 70, 72 Deutsche Börse, 45 development boom–slump cycle, 133 development cycles in IFCs, 147–8, 281–2 DiPasquale and Wheaton model, for rent, 129 Disconto-Gesellschaft, 70 distance and information on equity trading, combined effect of, 56–7 diversiﬁcation, 10 diversiﬁed cities, 15 dot.com crisis, 188 see also ﬁnancial markets Dow Jones index, 201–2 Dubai, 80 Dutch East India Company, 28, 68 Dutch tulip mania, 201, 203 ease of doing business and status as a business centre, 36–7 Eastern European ofﬁces, 150 economic growth, in a region, 60 edge cities, 9 effects of distance on equity ﬂows between countries, 55–6 electronic trading platforms, 44–7

employment, in ﬁnancial activities City of London, 118 New York, 118–19 employment cycle in IFCs, 117–22 equity market prices, 188 Eurex, 45 Eurobond markets, 65 Eurocredits, 65 Eurodollar markets, 65 EuroLAN, 47 Euronext, 44–5 European centres, 69 European international ﬁnancial centre (IFC), 69 European overseas banks, 78 European real estate, 165, 168 European Union urban agglomeration, 7 face-to-face communications, 50–52 factors determining choice of cities in attracting ﬁnancial ﬁrms, 57 Falcon, 46 FEA, 46 Federal Reserve System, 73 ﬁnance, insurance and real estate (FIRE), 133 ﬁnancial markets asset market bubbles, 201–6, 265 attributes and functions of, 33 average correlation between national markets, 189 Barings Crisis, 198 Black Monday, 194 bubble bursting, 206 capital ﬂows between countries, 191–2 coefﬁcient of global market factor, 189–90 commercial mortgage-backed securities (CMBS) portfolios, 198 commercial real estate, 265 co-movement and crises, 189–200, 264–6 co-movement in global equity markets, 190 competition between banks and other ﬁnancial institutions, 210 conditions for a rational bubble, 204–5

Index

contagion and spillover effects, 196–200 correlation coefﬁcients of monthly real returns for US and major European markets, 189 credit cycle, 208 development of asset bubbles in relation to the East Asian equity and currency crises of 1997, 205 Dutch investment fund correlation analysis, 199–200 Dutch tulip mania, 201, 203 East Asian ﬁnancial crisis, 197 Economist views, 33–4 employment status, 206 equity and commercial real estate cycles for OECD countries, 208–9 Exchange Rate Mechanism (ERM) difﬁculties, 197 feedback effects between the equity and housing markets, 203 fundamentals of, 200 Hong Kong stock market crisis, 195 impact of rise in intangible values, 203–4 implications for investment strategies, 191 information asymmetry, 196–8 integration of property and, 270–77 Latin American loans, crisis, 198 Long-Term Capital Management (LTCM), 194 Mexican Peso crisis, 195 Morgan Stanley Capital International World stock market index, 190 occupational demand and rents, 206 optimists vs pessimists, 204 private capital ﬂows, 197 Property company and Real Estate Investment Trust (REIT) returns, 189 and property downturns, 211 property prices and bank lending, 207–8 real estate booms, 201–6 real estate crisis, 206–14, 264–6 real estate cycles in OECD countries, 211–12 risks of global ﬁnancial integration, 193–6

323

role played by real estate in ﬁnancial crisis, 207 Russian debt default, 197 Shiller’s behavioural explanations of equity market boom, 203 and shocks in lending markets, 215 South Sea Bubble, 201 Swedish banking crisis, 212–13 theoretical beneﬁts of global ﬁnancial integration, 192–3 US residential sub-prime mortgage crisis, 198 variance of asset returns in a country, 195 Financial Modernisation Act (1999), 66 ﬁrms’ decision-making, on location, 60–61 Fisher four-quadrant model, for rent, 129–30 FLAG (Fibre-Optic Link Around the Globe), 46 FNAL, 46 foreign assets held by countries, 71–2 foreign exchange trades, 37–8 Frankfurt, 69, 71, 76 French overseas banks, 72 French real estate investment, 165 Fujitsu, 46 gamma world cities, 24 Geneva, 69 geo-political factors, 58 Glass–Steagall Act, 40, 66, 75 global and market risk, 287–94 global capital ﬂows, 170 global cities analysis of air travels, 25–7 Anne Haila’s work, 141 Asian cities, 25, 27, 29–30 basis of intra-national regional trips, 27 on the basis of linkages, 20 built space and ownership, 31 capital and reinvestment, 29–30 central business districts (CBDs), 261 by corporate service provision, 24 cosmopolitan nature, 28 Eurocentric/Atlantacist bias, 29 Friedman’s and Sassen’s ideas, 18–23, 30

324

Index

global cities (Contd.) functional specialisation, 290 GaWC database analysis, 24–7 growth of transnational ﬁrms, 19 high factor loadings, 25 impact of national and local institutions, 30 importance of London and New York, 25 multinational enterprise (MNE) generated system of, 22 occupational space, 31 real estate, 260–61 religion, 30 social impacts of urban restructuring in, 22 social structure, 30 taxonomic approach, 24 Tokyo, 29 Global Crossing, 46 Global Financial Centres Index, 34–5 Globalisation and World Cities (GaWC) project, of world cities, 23–7 global real estate investment, 217, 246–51 Gold convertibility, 73 great depression era (1930s), 74 Hamburg, 68–9, 76 Harvey’s ‘secondary circuits of capital’ theory, 140 Henderson’s model, of agglomeration economies, 9–10 hexagonal market areas, 8 high-order cities, 9, 30 history, 67–74 ‘home bias’ effect, 55 Hong Kong, 79–80 Hope & Co, 68 HSBC, 78 impact of technology on ﬁnancial services, 48–50, 117, 188 indicators and attributes, 34 Industrial Bank of Japan, 77 inﬂation and oil prices, 66 information technology and globalisation, 47–50, 81–3, 117, 188

changing ofﬁce space requirements, 110 control and maintenance of linkages, 51–2 decision-making process, 50 and face-to-face communications, 50–52 impact of technology, 48–50 intensity of communication, 52 portfolios of sites, 51 innovations in commercial real estate markets, 269 insurance companies, 71–2 integrated property–ﬁnance market relationships, 270–77 interbank non-domestic claims in developing and transitional economies, 55 internal geography, 5 international arbitrage, 197 International asset markets, 214 international debt issuance ﬁgures by country, 38–9 International Financial Index (IFI), 37 international ﬁnancial services, 32 international real estate investments, 165–70 Investment Property Databank (IPD), 143 Islamic ﬁnance, 83 Jacobian economies, 10, 12 Japanese banking system, 77 Japanese investment banks, 77 JP Morgan, 73, 75 JS Morgan, 72–3 K = 3 structure, 8 KDDI, 46 keiretsu networks, 29 Kleinwort, 72 knowledge base and infrastructure, 58 Lebanon, 80 lending role, in the property cycle, 164–5 liberalisation policies and ﬁnancial market, 66–7 locational ﬁxity, 31 location of business activity, 60–61 London Docklands, 138–9 London Interbank Offer Rate, 164

Index

London International Financial Futures Exchange, 45, 57 London Stock Exchange, 68, 71, 196 lower order cities, 9 major law and accounting ﬁrms, 71 Manufacturers Hannover, 139 market place, 8 Marshallian externalities, 9 Marshall’s ideas, of localisation economies, 53 MasterCard Worldwide Index, 36–7 mergers and acquisitions (M&A), 40–41 Middle Eastern IFCs, 76–81 Middle East ﬁnancial centres, 36 Mitsubishi, 77 Mitsui, 77 Morgan Harjes et Cie, 73 Morgan Stanley index, 201 multinational enterprise (MNE), 22, 28 Muscovy Company, 28 Myrdal and Latin American dependency theory, 27 NASDAQ, 45 national boundaries, 15 national economies, 6 National Stock Exchange of India, 45 NEC, 46 negotiated process, 138 New International Financial System, 67, 83 New York, 71, 73, 118–19 niche centres, 36 Nomura, 77 Nyckeln, 213 ofﬁce developer, 130 ofﬁce space agency approach to development process, 142–4 Alonso-style bid-rent models of urban land-use, 95–6 approaches measuring development process, 136 building costs and property value, 129–35 Chicago, property development, 128 City of London, property development, 125–7

325

City of London, rental adjustment, 105–6, 109 cyclical behaviour of markets, 131 demand for, 109–17, 261–4 developers and development cycles, 125–9 development cascades of, 146–7 Dublin, rental adjustment, 106 econometric modelling research, 135–6 estimated development of the service sector, 92 European markets, rental adjustment, 107 ﬂoorspace per worker, City of London, 120–22 ﬂoorspace per worker, New York, 118–20 Helsinki, rental adjustment, 107 impact of changing business practices and, 109–17 impact of developer forecasts of future values, 131 impact of management changes, 110–11 and information technology, 110 innovative workplace policies, 113 institutional framework of development process, 136–42 and interlocking of markets, 135 international city central business districts (CBDs), 108–9 for international ﬁnancial services, 114–17, 123–4 land rent, land-use and transport costs, 94–7 links between capital markets and the property cycle, 128 options and developer strategies, 144–8 quantitative models of the development market, 135–44 rental lags and phasing, 133–4 rent models and rental adjustments, 97–105 role of corporate real estate managers, 113 specialist, 92–3 supply and investment, 261–4 Tokyo, 127

326

Index

ofﬁce space (Contd.) UK banks and insurance company space usage, 121–2 UK ofﬁce space usage, 112, 113 United States, rental adjustment, 107 United States ofﬁce space usage, 112–13 volatile demand for space, 146–7 offshore, 36 Olympia and York (O&Y), 139–40 OMX, 45 OPEC, 66 OTC derivatives trading, 37–8 outsourcing, 43–4 ownership, changing pattern of, 219, 251–5 Paciﬁc Coast Exchange, 45 Paris, 68, 70, 72 Paris Borse’s liberalisation measures, 66, 72 People’s Republic of China (PRC), 78–9, 86 period of crises, 74–6 Plaza Accord (1985), 211 Porter’s Diamond Model of competitive advantage, 58–9 portfolio allocations, in real estate markets barriers and issues, 169–70 basic principles, 150–52 capital ﬂows, 170–80 diversiﬁcation, 160 domestic real estate, 167–8 equity returns, 163–4 global diversiﬁcation, 165–9 illiquidity premium, 158 and international ﬁnancial centres (IFC), 180–82 international investments, 165–70 investment in private real estate, 152–9 investment patterns and investor behaviour, 161–2 IPD Portfolio Analysis Service segments, 160 lending terms, 164–5 London and South of England bias in investment, 161

market volatility in US and UK, 163 non-domestic investors, 170 payment of dividends, 163 private real estate markets, 169 real estate vehicles and, 163–5 return-chasing strategies, 165–6 sector–region segments, 159–61 strategic advantages in London investment, 162 theoretical real estate allocations, 169 UK real estate investment, 161 Price Waterhouse, 71 Property company and Real Estate Investment Trust (REIT) returns, 189 property development cycles, 132 property investment vehicles, 266–8 beneﬁts of transparency, 245 closed-ended funds, 223 diversifying characteristics, 228 ‘fund-of-fund’ structures, 226 fund styles, 223–4 German open-ended funds, 223, 226 impact of change in debt markets, 241 inﬁnite or ﬁnite life funds, 223 leveraged funds, 225–6 listed real estate vehicles, 228–33 NAV basis, 220, 222 open-ended funds, 223 pension funds, 220 private equity, 220–28 property derivatives, 241–3 real estate debt markets, 233–5 real estate funds, 220–22 Real Estate Investment Trust (REIT) model, 220 securitised debt and the capital markets, 236–40 and transparency, 227 purchasing power parity, of top cities, 17 rank size rule, 6 rating of competitiveness, 35 real estate capital markets and capital availability, 171 capital ﬂows between countries, 191–2 capitalisation rate or yield, 171

Index

capital values and cap rates in the Sydney ofﬁce market, 179 City of London ofﬁce rents, 175–6 depreciation, 177–80 equilibrium state, 171 global, 170 individual real estate investment portfolios, 180 long-run real growth rate, 175–6 and rental growth, 181 required returns and yields, 171–6 risk-free rate and risk premium, 172–3 structural shifts in risk, 174 theoretical decomposition of yields, 178 transparency and institutional structure, 174–5 UK cap rates, 179 US cap rates, 179 zero coupon bond yields, 172 real estate investment and ownership beneﬁts and gains, 292–4 changing pattern of ownership, 219, 251–5, 291 foreign direct investment, 217 global, 217, 246–51 innovations in property investment vehicles see property investment vehicles traditional model of investment ownership, 218 US real estate investment trust model, 218 Real Estate Investment Trust (REIT) returns, 189 redevelopment, 118 regional differences, 297–8 Regulation Q, 65–6, 76 relationships between ﬁrms in IFCs, 54 relationships between rent and ﬁnancial market performance, 277–81 rental adjustment models for individual cities, 298–300 rental relationship overtime, 282–7 retail and wholesale transaction activity, 114–16 risk of cable disruption, 46, 51–2

327

role of city governments, 59–60 role of government, 58 Rothschild, 68, 72 Schröders, 68–9, 72 SEA-ME-WE-3/4 cables, 46 settlement hierarchy, emergence of, 8 Shanghai Share Brokers Association, 78 signature architect ofﬁce construction, 141 signiﬁcance of land and property markets, 61–4 Slovenian Stock Exchange, 45 slump, 132 South Sea Bubble, 201 spatial behaviour of ﬁrms, 53–4 spatial model, of urban system, 7–9 specialised city, 10–12 ST Telemedia, 46 subordinate cities, 12 Sumitomo, 77 Swedish banking crisis, 212–13 Swedish real estate, 212 Swiss Stock Exchanges, 45 Teleglobe, 46 telematics, 50–52, 81–2 Telstar, 83 Tobin’s Q, 131 Tokyo, 76–7 Toronto Stock Exchange, 45 Transatlantic radio communication, 83 transnational corporations (TNCs), 19, 28, 32 transoceanic ﬁbre-optic cables, establishment of, 46 Treaty Ports, 77 Tyco, 46 UK real estate investment, 161 United States, 73 urban agglomerations PricewaterhouseCoopers estimates, 17 UN estimates, 16–17 urbanised nation, 5 urban land-use conﬂicts, 138 urban population, 16 urban redevelopment, 22

328

Index

urban redevelopment, in New York and London, 138–9 US cities, 12–14 US ﬁnancial activity employment statistics, 118–19 US urban hierarchy, 13

world cities see global cities World Federation of Stock Exchange (WFSE) data for market capitalisation, 39–40 World Financial Center, 139 World Wars, 74, 76

value to replacement cost ratio, 131 Vienna, 69 VSNL, 46

Xetra trading platform, 56

Wall Street Crash (1929), 75 Wójcik’s index, 37

Zipf’s law, 6 Zurich, 69, 76