Innovation U.S. Services Sector (Routledge Studies in Innovation, Organizations and Technology)

Innovation in the US Service Sector

The US service sector is the largest sector in the US economy and accounts for an increasingly significant share of US gross domestic product, currently 68 percent. Both in the United States, as well as in other industrialized nations, the service sector is a dynamic component of economic activity and growth. As pervasive and economically important as the service sector is, innovative activity in service-sector firms remains somewhat of an enigma; it is not well understood and not well defined because it differs dramatically from the traditional model of innovation in manufacturing. Innovation in the US Service Sector attempts to begin to fill this void with an emphasis on the United States, but the implications of the book have global relevance. This book contributes toward a better theoretical understanding of innovation in the US service sector by focusing specifically on the disparate nature and role of R&D in the service sector compared to the manufacturing sector. Based on this understanding of the nature and scope of R&D in the service sector—and taking as given that R&D is a critical input to the innovation process and innovation drives economic growth—the book illustrates a theoretical model of service-sector innovation and compares and contrasts it to the traditional theoretical model of manufacturing innovation. Michael P. Gallaher is Director of the RTI International’s Technology, Energy and Environment Program, and his research focuses on modeling the economic impact of new technologies. Albert N. Link is Professor of Economics at the University of North Carolina at Greensboro, USA. His research has generally been within the areas of innovation policy, university technology transfer strategy, and the economics of R&D. Jeffrey E. Petrusa is an Associate Economist at RTI International in the Technology, Energy and Environment Program, where he specializes in technology adoption, and industrial innovation.

Routledge Studies in Innovation, Organization and Technology

1

Innovation in the US Service Sector Michael P. Gallaher, Albert N. Link, and Jeffrey E. Petrusa

Innovation in the US Service Sector

Michael P. Gallaher, Albert N. Link, and Jeffrey E. Petrusa

First published 2006 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Simultaneously published in the USA and Canada by Routledge 270 Madison Ave, New York, NY 10016 Routledge is an imprint of the Taylor & Francis Group, an informa business

This edition published in the Taylor & Francis e-Library, 2007. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.” © 2006 Michael P. Gallaher, Albert N. Link, and Jeffrey E. Petrusa All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data A catalog record for this book has been requested

ISBN 0–203–96631–7 Master e-book ISBN

ISBN10: 0–415–39068–0 (hbk) ISBN10: 0–203–96631–7 (ebk) ISBN13: 978–0–415–39068–2 (hbk) ISBN13: 978–0–203–96631–0 (ebk)

Contents

List of figures List of tables Acknowledgments

vi vii ix

1

Introduction

1

2

Innovation in the service sector

6

3

Telecommunications industry

36

4

Financial services industry

51

5

Systems integration services industry

69

6

Research, development, and testing service industry

81

7

Dimensions of innovation and productivity growth

97

8

Public policies to enhance innovation

111

Notes Bibliography Index

120 124 131

Figures

1.1 US GDP and value added by industry, 1999–2004 ($ billions) 2.1 Total and nonmanufacturing R&D expenditures, 1986–2001 ($1996 billions) 2.2 Model of innovation relevant to the manufacturing sector 2.3 Model of innovation relevant to the service sector 3.1 Development supply chain for Wi-Fi services 3.2 Development activities performed by small-sized network operators 3.3 Development activities performed by large-sized network operators 4.1 Iso-technology curve 5.1 Illustration of the information flows for systems integration activities 5.2 Production spectrum of systems integration services firms 6.1 Drug development supply chain 6.2 The innovation process in biopharmaceuticals over several phases of R&D 6.3 The stages in drug development 8.1 Spillover gap between social and private rates of return to R&D

2 15 19 21 43 46 49 60 72 76 85 86 88 113

Tables

1.1 Industrial and institutional composition of US sectors 1.2 Traditional comparison of manufacturing and services sectors’ innovation-related characteristics 2.1 Systems integration in selected US service firms 2.2 Alternative definitions of R&D 2.3 Public good characteristics of the elements of innovation in the manufacturing sector 3.1 Summary of the telecommunications industry (2003) 3.2 Top 10 telecommunications service providers, by R&D expenditures (2003) 4.1 Summary of financial services industry (2003) 4.2 Firms interviewed related to Web services 4.3 Differences in investment service and retail banking development strategies 4A.1 Ten largest firms in the financial services industry (2003) 4A.2 Firms with significant R&D expenditures (2003) 4A.3 Select sample of firms reporting R&D expenditure in the financial services industry 5.1 R&D investments of representative systems integrators (1998) 5.2 Systems integration services firms interviewed 5.3 R&D investments of representative systems integrators, by class of systems (1998) 6.1 Largest RD&T firms, by R&D expenditures 6.2 Largest biotechnology firms, by R&D expenditures 6.3 RD&T biotechnology services firms interviewed 7.1 Descriptive R&D statistics for manufacturing and nonmanufacturing firms

2 3 8 16 20 37 38 52 53 55 64 65 66 70 72 74 82 83 90 98

viii Tables 7.2 Regression results on the elasticity of R&D 7.3 Membership in RJVs 7.4 Regression results on sectoral differences in induced technology 7.5 R&D and productivity growth 7A.1 Disaggregated R&D-to-productivity growth studies

99 100 104 104 108

Acknowledgments

Many individuals contributed to the research and preparation of this book. First, we are pleased to acknowledge the support of the National Science Foundation and the National Institute of Standards and Technology; these two organizations commissioned the study from which this book is derived. Second, we appreciate the administrative support provided by both the Research Triangle Institute and the University of North Carolina at Greensboro in the preparation of the final book manuscript. Third, the book has benefited from the comments and suggestions of our colleagues and the referees on earlier versions of our manuscript. And finally, our families and friends have been appreciatively supportive throughout the entire process.

1

Introduction

It is well known that productivity improvements fueled by technological change have contributed to a level of economic well-being that is higher today than at any time in our nation’s history. Further, sector-level increases in productivity growth coupled with consumer preferences have substantially changed the economic landscape of the United States over the last century. In 1900, for example, farming (agriculture) was the largest sector of the US economy. During the twentieth century, the number of farm workers decreased dramatically, while farm output grew fast enough to feed a growing population and have a surplus remaining for exports. In the second half of the twentieth century, manufacturing has followed a similar course—output grew continually while employment decreased. In both of these sectors, R&D led to innovations that significantly increased productivity and freed up labor for alternative uses (services). Today, the US private service-producing industries, hereafter referred to as the US service sector, is the largest sector in the US economy and accounts for an increasingly significant share of US gross domestic product (GDP).1 Figure 1.1 shows US GDP in billions of current dollars for the years 1999 through 2004, and it shows value added by industry for the government (federal, state, and local), the private-goods producing industries, and the private service-producing industries. In 2004, the service sector accounted for about 68 percent of GDP. If government services are included with serviceproducing industries, this contribution by the more broadly defined service sector exceeds 80 percent. Clearly, the service sector makes the greatest contribution to the US economy.2 Despite the contribution the service sector makes to the economy, people cannot agree on its definition.3 From a national data collection perspective, the Bureau of Economic Analysis within the US Department of Commerce defines service-producing industries to be the nonagriculture, forestry, fishing, and hunting; nonmining; nonconstruction; and nonmanufacturing industries. Table 1.1 provides the industrial and institutional composition of

$14,000

$12,000

$10,000

$8,000

$6,000

$4,000

$2,000

$0 1999

2000

2001

2002

GDP Goods-producing industries

2003

2004

Government Service-producing industries

Figure 1.1 US GDP and value added by industry, 1999–2004 ($ billions).

Table 1.1 Industrial and institutional composition of US sectors Private service-producing industries or service sector

Private goods-producing industries

Government

Arts, entertainment, recreation, accommodation, and food services Educational services, health care, and social assistance Finance, insurance, real estate, rental, and leasing Information Other, except government Professional and business services Retail trade Transportation and warehousing Utilities Wholesale trade

Agriculture, forestry, fishing, and hunting Construction

Federal

Note Industries listed alphabetically.

Manufacturing Mining

State and local

Introduction

3

Table 1.2 Traditional comparison of manufacturing and services sectors’ innovationrelated characteristics Characteristics

Service sector

Manufacturing sector

Intellectual property rights Technology orientation

Weak, copyright Technology pull, consumer/client-led (co-terminality) Outsourced—embodied in purchased inputs Long cycles (except for computer services) Intangible, difficult to store Regional, national

Strong, patents Technology push, science and technology led In-house

Research/innovation Innovation cycle times Product characteristics Spatial scale of system

Short cycles Tangible, easy to store National, global

Source: Adapted from Howells (2000a) and NIST (2005).

US sectors. From the North American Industry Classification System (NAICS) data classification perspective, the definition of services is still evolving and the industries included change, albeit only a small amount, with each revision (Mohr, 1999). In addition to being the largest, the service sector is also the fastestgrowing sector. In 2004, according to Bureau of Economic Analysis data, the percentage change in value added by the service sector was 4.9 percent, compared to 4.2 percent for GDP in total, 3.9 percent for the private goods-producing sector, and 1.0 percent for the government. Ironically, however, the service sector has historically been viewed as having little or no productivity growth and as void of an ability to innovate (Tether et al., 2001; Miles and Dachs, 2005). The sector has also been characterized as having low-paying jobs, low levels of technological dependence, and a relatively undeveloped level of institutional organization. In contrast, the manufacturing sector, which collectively dominates all other private goods-producing industries, was seen as the source of most innovations and thus the engine of economic growth. Table 1.2 provides a partial comparison of innovation-related characteristics between the service sector and the manufacturing sector. The contrast between these two sectors is dramatic: ●

●

Intellectual property rights are weak in services and strong in manufacturing. Service-sector firms pull new technologies in-house, whereas manufacturingsector firms develop new technologies and push them into the market.

4 ●

● ●

●

Introduction Traditionally defined R&D occurs in-house in manufacturing firms. Service-sector firms outsource their research and innovation activity. Innovation cycles are short in manufacturing and long in services. Products produced by service-sector firms are relatively intangible compared to the tangible nature of manufacturing products. Many manufacturing-sector firms operate on a national or global scale, whereas service-sector firms are more regional, or perhaps national, in spatial scale.

These innovation-characteristic differences are due in large part to the nature of R&D that occurs in services, more so than to the level of R&D conducted in services. In fact, service-sector firms are, on average, active in all aspects of R&D, contrary to the stereotypical description of that sector’s innovative activities. In 2001, according to National Science Foundation (NSF) data, nearly 45 percent of all industrial basic research was performed in the service sector, just over 36 percent of all applied research, and nearly 41 percent of all development research. In recent years, the service sector has come to be viewed, both in the United States as well as in other industrialized nations, as a dynamic component of economic activity and growth. The observable growth in Internet and Web-based services and high-technology environmental services has brought attention to the service sector—knowledge-intensive services, in particular— as a significant contributor to economic growth (Howells, 2001). As pervasive and economically important as the service sector is, innovative activity in service-sector firms remains somewhat of an enigma; it is not well understood and not well defined because it differs dramatically from the traditional model of innovation in manufacturing.4 This book attempts to begin to fill this void. The purpose of this book is to contribute toward a better theoretical understanding of innovation in the service sector by focusing specifically on the disparate nature and role of R&D in the service sector compared to the manufacturing sector. Based on this understanding of the nature and scope of R&D in the service sector—and taking as given that R&D is a critical input to the innovation process and innovation drives economic growth—we illustrate a theoretical model of service-sector innovation and compare and contrast it to the traditional theoretical model of manufacturing innovation. Our model is capable of offering new insights into the innovation process, including not only an understanding of the role of R&D as an input into innovation but also the general knowledge flows that influence the level and scope of innovation within service-sector firms. In addition we investigate the role of entrepreneurial activity as it specifically relates to the service sector.

Introduction

5

In this way our model has the potential to provide a broad framework on which innovation and technology policies, as well as future empirical research, could be based. As an overview, we set forth in Chapter 2 our model of service-sector innovation, and we contrast it with a traditional model of innovation in the manufacturing sector. Our model evolved from four extensive case studies of service-producing industries: the telecommunications industry, discussed in Chapter 3; the financial services industry, discussed in Chapter 4; the systems integration services industry, discussed in Chapter 5; and the research, development, and testing service industry, discussed in Chapter 6. Our model of innovation in Chapter 2 is not derived specifically from any specific industry; rather it is a composite of aspects of the innovation process gleaned from these four industry studies.5 In Chapter 7, we quantify firm-specific aspects of innovative activity in the service sector and compare them to the manufacturing sector using a representative sample of US public R&D-active firms. Finally, Chapter 8 concludes the book with a discussion of policy prescriptions related to the public sector enhancing innovation in the service sector.

2

Innovation in the service sector

Introduction Much of the service sector’s growth since the early 1980s has been based on new services with significant knowledge content. A large share of the knowledge content in services is built on advances in hardware and software that were imported or purchased from the manufacturing sector. However, the service sector adds value in a different way than the manufacturing sector—by integrating purchased physical technology into systems. To add this value, service-sector firms have made considerable investments in capabilities such as systems-level integration. The impact of these investments has been substantial. R&D investments in service-sector information technology (IT) have generated an estimated rate of return of 196 percent, while noninformation technology investments in general have only created an 11 percent return (TASC, 1998). For the US economy to continue to experience historical rates of productivity growth, the future performance of the service sector will be critical. Productivity growth is associated with applying resources to inventive and innovative initiatives as measured by spending on R&D. These expenditures, which currently represent only a few percentage points of GDP, support both basic research and applied research, which includes research into new specialized products for sale to industries and into the development of processes and process improvements internal to the innovating organization. Traditionally, analysis of productivity growth in the service sector and the development of more accurate R&D estimates for that sector have suffered relative to other sectors of the economy. This is because economic output and productivity measures were originally developed in an era when services were a smaller share of the economy and the absence of more accurate information was not critical to policy making. In addition, because improvements to service outputs tend to be related to changes in quality rather than in quantity, productivity improvements are very difficult to measure.

Innovation in the service sector 7 Today that view has changed. Services are a large part of the economy and information technologies have, in many service industries, revolutionized the way business is conducted. However, use of the manufacturing R&D model to measure service-sector innovation activities is not completely appropriate and is likely to lead to biased estimates of innovative investments by the service sector. Because the service-sector innovation process is a relatively new phenomenon and less structured compared to the manufacturing sector’s process, the nature and magnitude of service-sector innovative investments are less understood. Different taxonomies, objectives, and research processes create confusion over classification and data collection. This confusion makes NSF’s job of determining how to quantify total investments in service-sector R&D and describing innovation outputs difficult. The lack of an adequate R&D taxonomy and a consensus framework for analyzing service-sector R&D and innovative activity hampers the development and evaluation of public policy.

Taxonomies of service-sector innovations Innovation and technological change in the service sector are increasingly dependent on internal service-sector R&D, in addition to technology acquired from the manufacturing sector (Pilat, 2001). However, there is concern that current NSF R&D statistics do not fully capture the level of innovative activity being performed within the service sector or, indirectly, the rate of change of innovative activity. NSF’s industrial R&D survey reports that manufacturing performed 62 percent and nonmanufacturing performed 38 percent of total industrial R&D in 2001. This distribution could be interpreted to imply that manufacturing may be performing a disproportionate share of national R&D relative to the sector’s contribution to economic growth and may be doing a preponderance of the innovative activity in the economy. Recent studies have found that innovations in high-technology equipment have been increasingly the product of end-user R&D activities and less from vendor R&D. These end users, which are frequently service-sector firms, have firsthand understanding of what the technology needs are within an industry and how to innovate the existing technology to ultimately improve competitive advantage (von Hippel, 1988). For example, systems integration is an integral component of most service sectors’ innovative activities. As shown in Table 2.1, this process involves customizing components for specific applications. The service-sector firms (or specialized consultants) are essential in the integration process because of their detailed knowledge of the specific applications.

8 Innovation in the service sector Table 2.1 Systems integration in selected US service firms Name

Application description

Aetna

● ●

●

Amazon.com

●

●

●

Citicorp Merrill Lynch

● ● ● ●

●

Sabre Group

● ●

Insurance classification plan loss control system Expert system for providing interactive assistance in solving problems such as health care management Three-level distributed control for networking input/output devices System and method for providing multimedia bookmarks for hypertext markup language Secure method for communicating credit card data when placing an order on a nonsecure network Secure method and system for communicating a list of credit card numbers over a nonsecure network System and method for delivering financial services Distributed network agents Check alteration detection system and method Integrated system for controlling master account and nested subaccount(s) Securities trading workstation Information aggregation and synthesization system System to predict optimum computer platform

Source: Compiled from information contained in the US Patent and Trademark Office’s Web patent database. Available online at http://www.uspto.gov. Taken directly from CSTB (2000, p. 84).

The term “innovation” has historically been used to encompass a wide range of processes that include both R&D- and non-R&D-related activities. In a broad sense, innovation may be new products, new processes, or new organizational methods that are novel and add value to economic activity. These developments or changes are shaped by interactions between a firm and various other organizations, including suppliers, collaborators, competitors, customers, technological infrastructures, and professional networks and environments. A firm’s innovation pattern depends on changes in the behavior of these organizations and the expectations of other organizations’ behaviors (Gallouj and Weinstein, 1997; Hauknes, 1998). In general, innovation has also been described as any change in the characteristics of new products in terms of service, competence, and/or technical knowledge, brought about by evolution, variation, disappearance, appearance, association, or disassociation (Gallouj and Weinstein, 1997). In this light, innovation captures product and service modifications that may or may not be derived from R&D. To date, most researchers believe that an accurate model of service innovation is still absent from the literature (Howells, 2000b). With a few

Innovation in the service sector 9 notable exceptions, such as, for example, Barras’s reverse product cycle (Barras, 1986), the extant literature related to service-sector innovation focuses on differentiating service activities from manufacturing innovation paradigms, as opposed to building on the specific nature of service-sector products and process. This state of the literatures suggests a need to integrate the unique traits of service-sector innovation into the existing taxonomies and innovation paradigms. The goal is to capture more diverse types of innovation in conceptual models of innovation, including what was once two distinct segments of the economy but has over time become increasingly less disparate (Gallouj and Weinstein, 1997; Amable and Palombarini, 1998). Although existing taxonomies have begun to address how innovation occurs in service firms, much more work remains to be done in terms of modeling an innovation system that incorporates services. The model we posit in this chapter is only one step toward that end. Recent efforts to incorporate the service sector into models of innovation include, for example, Pavitt’s (1984) taxonomies for classifying sectoral patterns of technological change. By dividing a national economy into three sectors—supplier-dominated, production-intensive, and science-based, Pavitt outlined a dynamic relationship between technology and service industries. Professional, financial, and commercial services are captured in the supplier-dominated category. However, the firms associated with this sector were primarily described as firms that expend few resources developing processes and products, usually having weak in-house R&D capabilities where most innovations come from the supplier of equipment and materials.1 Soete and Miozzo (1989) have taken the Pavitt model a step further by expanding on the supplier-dominated sector, offering two new classifications: a production-oriented category and an innovative-specialty category. The production category includes those service firms performing large-scale processing and administrative activities and developing physical or information networks. The specialized technology suppliers’ category includes firms performing science-based activities to develop proprietary technology through innovation. In addition to understanding the distinctions between manufacturing and service-sector innovative activities, a second line of research has focused on the differences between products (e.g. new services) and process (e.g. new organizational and delivery processes) within the service industry. Gallouj and Weinstein (1997) make this distinction by dividing innovation for the service firm into two classifications: technical characteristics (front-office tangible technologies for the part of a firm in direct contact with clients) and

10

Innovation in the service sector

process characteristics (both tangible and intangible back-office technologies such as methods, working tools, and organizational theory). Evidence from Sirilli and Evangelista (1998) suggests that service-sector firms are able to distinguish between service and process innovations. These authors identify engineering, technical consultancy, and computing services as the most innovative service sectors. Sundbo (1997) offers a strategic innovation paradigm where the firm’s strategy is the core innovation determinant. He offers empirical evidence through a study of Danish service firms where he breaks services into three categories: top strategic organizations, characterized by large to medium service-sector firms that are mass producing services; network organizations, described as a loosely tied association of small firms that innovate little on their own; and professional organizations, defined as collective action groups with shared interest in technology interoperability. Evidence of the service sector’s technological maturity has been established throughout the literature (Amable and Palombarini, 1998; Pilat, 2001). Service-sector firms are taking on more proactive roles in innovation activities and in some cases are leading the innovation process. Given the distributive and dynamic nature of the innovation process, both manufacturing and service-sector firms are beginning to collaborate through bilateral and multilateral networks. This collaboration of firms is referred to as a distributed innovation process (DIP) and suggests that service-sector firms are beginning to take more of a leading role in innovation. The largest contribution to growth in the service sector is from a small subset of all services known as knowledge-intensive business services (KIBSs) (e.g. telecommunications, IT, networking, and organizational consultancy). In these areas, services are playing an increasingly important role in technological change and productivity growth by promoting standards and systems integration. Researchers predict that the developed economies of the world will soon enter a postindustrial period in which services drive economic growth (OECD, 2001a). As summary, over the past two decades, the literature has attempted to develop a conceptual framework aimed at understanding how service firms innovate. In recent years, the discussion has turned from the sizeable differences between the two sectors to their convergence. However, a growing number of authors point to such firms as IBM and/or Siemens (both large mass-production service firms) as examples of traditional manufacturers who now have a dominant share of their business activities associated with the sale of services (Howells, 2000b; OECD, 2001b). Over time the once sizeable distinction between manufacturing industries and service-producing industries is narrowing (Amable and Palombarini, 1998;

Innovation in the service sector 11 OECD, 2001a). The literature suggests that distinguishing between the primary (agriculture and raw material extraction), secondary (manufacturing), and tertiary (services) sectors of the economy is obsolete, as more firms in both manufacturing and services adopt the practice of encapsulating physical products with services (Howells, 2000b). Xerox was one of the first manufacturers to bundle its products with full-scale service agreements. Today, full-service leasing directly from manufacturers is common, and the trend is expected to grow in the future. This broadening of scope is referred to as servicisation, or the trend in manufacturing to encapsulate the physical product in a shell of services (i.e. finance, monitoring, maintenance, and repurchase). Servicisation is also a growing trend in the automobile and aerospace industries, and, as it becomes more pervasive and better understood, researchers will begin to rethink how they characterize the service sector and its innovation activities.

Measurement of service-sector innovation Recent empirical evidence demonstrates that service-producing industries innovate and have been doing so increasingly over time (Sirilli and Evangelista, 1998). As national institutions began measuring innovation activities in the service sector, the original models and data collection instruments were based on an understanding of the technology innovation process as it applied to manufacturing firms. However, because of the intangible nature of the service sector’s output, measuring the productivity of R&D performed using the historical measures of innovation, such as new products or patents (Gallouj and Weinstein, 1997), is difficult. The manufacturing sector adds value to inputs—R&D is an input in the innovation process—by continuously improving the materials and design of their products, whereas the service sector applies accumulated knowledge to build organizational models or systems, a more abstract output than in manufacturing (Jankowski, 2001). In the absence of such appropriate metrics, any resulting measure of innovative activity would by definition be limited. Service innovations draw less directly on scientific breakthroughs and are often small or incremental in nature, this means that small changes can lead to new applications or reorganization of an existing technology or system (Pilat, 2001). Some large service-sector firms actually have an innovation department that promotes and collects ideas. However, this organizational design may be the exception rather than the rule. Service-sector innovations are typically based on both market-wide and consumer-specific needs. Innovation in service-sector firms is generally not a systematic process and often consists of spontaneous ideas developed internally to meet the real-time

12

Innovation in the service sector

needs of a specific client. In contrast, innovation in manufacturing firms is typically highly structured, with a systematic approach to developing new products following the product life cycle. Although there is an attempt to systematically organize and account for innovation across all sectors, measuring innovation in the service sector is extremely subjective because of the intangible nature of its products (Pilat, 2001). Measuring innovation is further complicated in the service sector because it occurs throughout the organizational process. Patenting in the service sector is at times more difficult because of high visibility or the inexcludability of the product or process. If a manufacturing firm develops an innovative process or product, it can keep the process or product a secret by not allowing anyone outside the firm to see it. Servicesector firms offering intangible products have much higher visibility, which makes it hard to contain trade secrets and easier for other firms to imitate the product or process.

R&D in the service sector R&D is an integral component of the innovation process. R&D is distinguished from the broader category of innovative activities in that R&D is defined to include activities that systematically use research findings and expand the frontier of knowledge. Based on this definition, many activities frequently included as innovation are excluded from R&D, such as market research and technology adoption/or imitation. The classical definitions of R&D are grounded in creating an artifact or physical product (Tether et al., 2001) and, therefore, are harder to apply to intangible outputs of services, such as methods or organizational theory. For example, Sundbo and Gallouji (1999) identify four major categories of service innovation—product, process, organizational, and market—not all of which are considered R&D. However, the difficult distinction in services between product and process, also referred to as coterminality, often makes it difficult to interpret what is R&D and what is non-R&D (Sirilli and Evangelista, 1998; Evangelista, 1999). Innovation surveys have found that R&D accounts for a much smaller share of activity and expertise as related to service-sector innovation compared to innovation in the manufacturing sector. For example, R&D typically accounts for about half of manufacturing innovation expenditures, whereas, on average, R&D accounts for only about one-fourth of service-sector innovation expenditures. The exceptions are telecommunications, computer services, and engineering services where R&D accounts for over three-fourths of expenditures related to innovative activity.

Innovation in the service sector 13 Innovation inputs (elements) that are not classified as R&D include market research, training in innovation, adoption and adaptation of new technology, start-up activities, organizational changes, and incremental impacts. Such inputs characterize much related activity in service-sector firms. Because service activities are generally labor intensive, investment in human capital costs often represents a larger share of total innovation expenditures. For example, staff training for evolving computer systems or new product offerings is a continual process in service industries. Customization is also a gray area for distinguishing between R&D and non-R&D innovation activities. Service-producing industries commonly take products developed in the manufacturing sector and add value to them by assembling customized systems or networks. Frequently, a system is specifically tailored to an individual client and its assembly represents a unique product. However, it may be unclear if this is the development of a new and improved product, hence to be included as R&D, or the reapplication of existing methods and processes in providing a service. Possibly new knowledge and refined techniques are developed as part of most custom systems integration services that could be classified as R&D. Acquisition and integration of technology may also be an important component of innovation that is frequently not included as R&D. These activities are of particular importance to service-sector firms because much of the R&D associated with services is embodied in products acquired from outside of the service sector. The issue becomes what share of acquisition and integration activities is R&D and what share is simply technology adoption or imitation that is not classified as R&D. Business interactions and joint product development between the serviceproducing industries and manufacturing industries represent an increasing trend in product and service innovation. For example, American Airlines (AA) played a significant role in developing the design specifications for Boeing’s 777 series. However, much of AA’s activities were likely conducted by staff in market research divisions, and it is unclear what share of this work was or should have been counted as R&D expenditures. Similarly, telecommunications and financial service providers are integrally involved in specifying components that go into their systems. The NSF has traditionally measured R&D activity in the nonmanufacturing sector from responses to its annual survey of private businesses using its Form RD-1. Form RD-1 is an instrument that was originally developed for measuring R&D activity in manufacturing firms (Link, 1996); thus, by its design its applicability to service-sector firms is reasonably questionable. For reporting purposes, R&D is defined slightly differently across different US and international agencies. However, most US agencies and many foreign

14

Innovation in the service sector

agencies follow NSF’s definition of R&D. Research is defined as the systematic study directed toward fuller knowledge or understanding of the subject studied. Development is defined as the systematic use of knowledge or understanding gained from research, directed toward the production of useful materials, devices, systems, or methods, including the design or development of prototypes or processes. Research is further classified as either basic or applied, depending on the objectives of the investigator. Basic research is research directed toward increases in the knowledge or understanding of fundamental aspects of phenomena and of observable facts without specific application toward processes or products. This type of research is typically limited to the federal, university, and nonprofit sectors. However, NSF does include basic research as a reporting option on Form RD-1. Expenditures for basic research include the cost of projects that adhere to the aforementioned definition, but the projects could originate in fields that are of potential business interest in the future. Applied research is research directed toward gaining knowledge that will meet a specific need. This includes research for specific commercial objectives. Based on NSF’s definition, an activity is considered R&D if it is related to one or more of the following goals: pursue a planned search for new knowledge, whether or not the search has reference to a specific application; apply existing knowledge to problems involved in creating a new product or process, including work required to evaluate possible uses; and/or apply existing knowledge to problems involved in improving a present product or process. Inflation-adjusted R&D performed in industry over the years 1986 through 2001 in total and for nonmanufacturing (defined as total less manufacturing R&D) is illustrated in Figure 2.1.2 In 2001, the nonmanufacturing share of total industry R&D was nearly 40 percent. That percentage had increased steadily over time. Nonmanufacturing R&D as a percentage of total R&D has increased from about 8 percent in 1986 to its current level. This NSF-reported dramatic increase in reported nonmanufacturing R&D is partly due to improved sampling procedures to collect nonmanufacturing sector data, and to increased innovative behavior in nonmanufacturing firms. The NSF has collected and published industrial R&D statistics since 1953. Historically, a sample of firms was selected every 4 to 6 years to complete Form RD-1. In the intervening years, a subsample of only the largest firms was surveyed. In the early years of the survey, R&D was performed only in a small number of industries, and R&D in those industries was reported along with a catch-all category simply called nonmanufacturing. For example, in 1987 a sample of about 14,000 firms was selected to receive Form RD-1.

Innovation in the service sector 15 $180 $160 $140 $120 $100 $80 $60 $40 $20 $0 1986

1988

1990

1992

Real total

1994

1996

1998

2000

Real nonmanufacturing

Figure 2.1 Total and nonmanufacturing R&D expenditures, 1986–2001 ($1996 billions).

From 1988 through 1991, about 1,700 of these firms were resurveyed, and about 300 of the 1,700 were nonmanufacturing firms. In the early 1970s, recognition by NSF that more detailed information on nonmanufacturing R&D was needed began to spread. It was not until 1987, however, that NSF’s annual R&D reports included R&D estimates separated into three broad nonmanufacturing groupings: (1) communication, utility, engineering, architectural, research, development, testing, computer programming, and data-processing service industries; (2) hospitals; and (3) medical laboratories. By 1992, decisions were made at NSF to draw new samples annually with broader coverage, to increase the sample size from 14,000 firms to 23,000 firms, and to add 25 new nonmanufacturing industries to the sampling panel. Included in these 25 new industries were finance, computer and other business services, and engineering and management services. In 1992, the number of manufacturing firms nearly equaled the number of nonmanufacturing firms—11,818 compared to 11,558. However, in 1993 and 1994, the nonmanufacturing firms sampled fell below the number of manufacturing firms sampled. In 1995, the sample of nonmanufacturers was expanded by 60 percent. In that year, R&D expenditures for the following industries were, for the first time, reported in NSF’s annual publication, Research and Development in Industry: transportation and utilities, including communications; trade; finance, insurance, and real estate; services, including computer-based

Table 2.2 Alternative definitions of R&D US NSF

EU Frascati

US R&E tax credit bill

The following are included in the definition of R&D: Basic research: Pursue new Basic research is work Research that is undertaken knowledge whether or done to acquire new to discover information not the search has knowledge, without any technical in nature and reference to a specific particular application or holds applications useful application use in view in developing a new or Limited to federal, improved business university, and nonprofit component of the organizations taxpayer Applied research: Apply Applied research is original Research that seeks a new existing knowledge to investigation to acquire or improved function problems involved in new knowledge directed performance, reliability, creating a new product toward a practical or quality or process objective or a single product, operation, method, or system Development: Apply Experimental developNot specified existing knowledge to ment is systematic work, problems involved in drawing on existing improving an existing knowledge aimed at product or process producing new, or to improving substantially, existing products The following are not included in the definition of R&D: Not specified Not specified

R&D from acquired firms prior to acquisition Amortization above actual cost of property and equipment related to firm R&D Test and evaluation once a prototype becomes a production model Routine product testing

Not specified

Adaptation of existing business components to fit a particular customer’s requirements Not specified

Not specified

Not specified

Not specified

Consumer, market, and opinion R&D; advertising new products or processes

General purpose data collection

Research after commercial production of the business component Market research, testing, or development (including advertising or promotions) Routine data collection

Routine product testing

Innovation in the service sector 17 Table 2.2 Continued US NSF

EU Frascati

US R&E tax credit bill

Geological and geophysical exploration activities Quality and quantity control

Analysis of soils and atmosphere Not specified

Not specified

Troubleshooting for breakdowns in production Social sciences, etc.: any research in the social sciences, arts, or humanities Not specified

Scientific and technical information assistance

Management and organization R&D

Social sciences, etc.: any research in the social sciences, arts, or humanities Feasibility studies (e.g. a study of the viability of a petrochemical complex in a certain region) Administration and other supporting activities

Routine or ordinary testing or inspection for quality control Scientific and technical information assistance Social sciences, etc.: any research in the social sciences, arts, or humanities Efficiency survey

Activity relating to management function or technique

Source: NIST (2005).

business services, health services, and engineering and management services; and other. In 1998, 19,973 nonmanufacturing firms were surveyed compared to only 4,836 manufacturing firms. Although NSF responded to underrepresentation of the nonmanufacturing sector in national R&D statistics, it has done so using Form RD-1, a survey instrument that was originally developed on the basis of an understanding of the innovation process in manufacturing (Link, 1996). By so doing, NSF was implicitly assuming that the innovation process underlying the expenditure of R&D is generally similar in manufacturing as in nonmanufacturing. And such an assumption underlies any attendant policies based on NSF’s R&D data. That may or may not be the case. Three different definitions of R&D, from three alternative sources, are presented in Table 2.2 to illustrate similarities and differences in that activity. NSF identifies basic research, applied research, and development as the three distinct types of activities classified for reporting purposes under the rubric of R&D.3 All three activities require either the creation of new knowledge or a novel application of existing knowledge. Once a production process is established, activities associated with any further development of that process are not considered as R&D. Furthermore, NSF omits all social science research from its definition of R&D activities, as also shown in Table 2.2.

18

Innovation in the service sector

The Frascati Manual, which guides data collection throughout the European Union (EU) classifies R&D activities into three similar categories: basic research, applied research, and experimental development, and all activities classified as R&D must be in the pursuit of new knowledge or the discovery of new applications for existing knowledge, product, or process. The Frascati definition lists relatively fewer disqualifying criteria for R&D activities. See Table 2.2. Finally, the US 1981 R&E tax credit definition uses the term “qualified research” to identify activities aimed at creating new information or products and new applications of existing knowledge as applied to existing products. The Act also addresses the issue of modification or adaptation, omitted by the other two institutional definitions. The Act states that modification or adaptation of existing products to meet a client’s needs is not considered R&D. See Table 2.2.

Models of innovation in the manufacturing and service sectors Based on insights from four case studies of service-sector industries, summarized in Chapters 3 through 6, we illustrate a theoretical model of innovation in the service sector. First, for comparative purposes, we describe schematically innovation in manufacturing firms and then, building off that schematic model, we set forth our model of innovation relevant to service-sector firms. Innovation in manufacturing sector firms One well-established model of innovation activity relevant to technologybased firms in the manufacturing sector comes from Tassey (1997, 2005).4 Figure 2.2 builds on the Tassey model; the figure illustrates different technology elements within the overall model of innovation. Each technology element has a slightly different degree of public good attributes. These distinctions in terms of public good attributes make the model especially relevant for policy analysis, but also useful as a benchmark for manufacturing and thus a point of departure for modeling service-sector innovation. At the root of the model is the science base, referring to the accumulation of scientific and technological knowledge. The science base resides in the public domain. Investments in the science base come from basic research, primarily funded by the government and primarily performed globally in universities, federal laboratories and some large companies. Consider a representative manufacturing firm. Technology development, in the form of basic or applied research, generally begins within the firm’s

Innovation in the service sector 19 R&D laboratory. Technology development involves the application of scientific knowledge from the science base toward the proof of concept of a new technology. Such fundamental research, if successful, yields a generic technology. If the generic technology has potential commercial value, followon applied research takes place toward development, and if successful, a proprietary technology results. Basic, applied, and developmental research occur within a firm as a result of the firm’s overall strategic planning. Strategic planning defines the environment for entrepreneurial activities. And entrepreneurial activity influences production process development. Entrepreneurial activity related to innovation implies different things to different individuals, depending on their academic or professional backgrounds. Therefore, we have reviewed in an appendix to this chapter the intellectual history of the concept of entrepreneurship as it relates to innovation. In terms of Figure 2.2, we think of entrepreneurial activity in the Schumpeterian sense. Technology development corresponds to Schumpeter’s concept of innovation being, among other things, the creation of a new good or new quality of good or the creation of a new method of production. In this sense, entrepreneurial activity involves “the carrying out of new combinations” (Schumpeter, 1934, p. 74).5

Strategic planning

Production process development

Entrepreneurial activity

Technology development Proprietary technologies

Generic technologies

Market development

New value-added product

Risk reduction

Infrastructure technologies

Science base

Figure 2.2 Model of innovation relevant to the manufacturing sector. Source: NIST (2005).

20

Innovation in the service sector

Infrastructure technologies, or infratechnologies, support the processes that lead to both generic and proprietary technologies, and hence technology development. Infrastructure technologies are a diverse set of technical tools that are necessary to conduct all phases of R&D efficiently. Following Link et al. (1983b) and Tassey (1997, 2005), examples of infrastructure technologies include measurement and test methods, process and quality control techniques, evaluated scientific and engineering data, and the technical basis for product interfaces. Infrastructure technologies thus influence the science bases and are so influenced by it. The managerial skills necessary for a firm to move its proprietary technologies to a value-added product or process are also shown along the top horizontal area of the schematic in Figure 2.2. After production, market development takes place. Markets do not always accept new technology for a number of reasons, including transaction costs associated with verifying the new technology’s attributes and interoperability of the new technology with existing technologies. Infrastructure technologies can reduce such market risks and thus speedup market development. And to conclude, if market development is successful, value added will result. Of course, all of the private-sector decision nodes in Figure 2.2 are influenced by the overall strategic planning of the manufacturing firm. As shown in Table 2.3, the ten elements in the model in Figure 2.2 can be categorized in terms of their public good characteristics. This categorization is important because it highlights those aspects of manufacturing innovation that can be leveraged by public-sector innovation or technology policy. The knowledge that resides in the science base is a pure public good, as in knowledge per se. Generic technologies and infrastructure technologies are quasi public goods, and the results of risk reduction that Table 2.3 Public good characteristics of the elements of innovation in the manufacturing sector Elements of the innovation model

Public good characteristics

Science base Generic technologies Proprietary technologies Entrepreneurial activity Infrastructure technologies Risk reduction Strategic planning Production process development Market development New value-added product

Pure public good Quasi public good No No Quasi public good Quasi public good No No No No

Innovation in the service sector 21 stem from the use of infrastructure technologies also have a public good nature since the innovating manufacturing firm cannot appropriate these results fully. Innovation in service-sector firms Many of the fundamental characteristics of the innovation process differ between the model relevant to manufacturing in Figure 2.2 and the model relevant to a service-sector firm in Figure 2.3. Both manufacturing and service-sector innovation processes are driven by strategic planning. However, service-sector activities are more influenced by competitive planning because the technology-based service firm is more likely to innovate on the basis of customer input and on the basis of competitors that continually seek to challenge the firm for its customers. Whereas manufacturing firms strategically formulate road maps for developing new emerging technologies, in the service sector, firms strategically formulate road maps for deploying modifications of existing products. Because manufacturing firms are more likely to target discrete technology jumps, creating new technologies that make their competition obsolete, their strategic plans are long term and are linked less to current competitive planning. In contrast, service-sector firms’ strategies are typically focused on retaining or gaining market shares and involve more continual or incremental transition strategies that are or will be integrated with competitive planning. Both strategic and competitive planning drive the firm’s entrepreneurial activity. Whereas entrepreneurial activity in a manufacturing firm was Strategic planning

Product or service enhancement and systems integration

Market development

Competitive planning

Purchased technologies

Entrepreneurial activity

Risk reduction

Purchased technical services

Infrastructure technologies Science base

Figure 2.3 Model of innovation relevant to the service sector. Source: NIST (2005).

Enhanced value-added service

22

Innovation in the service sector

related to innovation creating a state of disequilibrium, in a service-sector firm the entrepreneur is more of a manager or perhaps an arbitrageur. In the sense of Baudeau (1910), the service-sector entrepreneur collects and processes knowledge and information. The identification and use of others’ technologies are at the heart of entrepreneurial activity. R&D activity (subsumed under entrepreneurial activity in Figure 2.3) fulfills an adaptive role to ensure that purchased technologies and technical services are used efficiently. One could generalize about the service-sector entrepreneur as being an individual who allocates resources with an equilibrium state and to maintain an equilibrium state. Whereas entrepreneurial activity, and its motivating perception of opportunity, drive the manufacturing firm toward producing new products and processes, the entrepreneurial activity of the service-sector firm drives redesigned or reconfigured enhancements of its existing products. At the root of entrepreneurial activity are others’ intellectual capital and technologies that are licensed or purchased to meet the firm’s road maps for deploying modifications of its existing products. Still, perception of opportunity is the defining entrepreneurial characteristic. This product and service enhancement often involves systems integration where systems integration facilitates the intersection of hardware, software, and the synthesis of application domains such as finance, manufacturing, transportation, and retail. A key distinction between manufacturing and service-sector firms’ R&D is that manufacturing firms conduct a larger share of their R&D in-house, and the output of that internal activity is more likely to be a proprietary technology. In the service sector, little research (R) occurs in house, and the development (D) activity that occurs is primarily related to enhancing, redesigning, or reconfiguring others’ proprietary technologies. Whereas manufacturing firms license or purchase others’ technologies in the form of intellectual capital or equipment to be used to produce proprietary technology, service-sector firms purchase others’ technology in the form of equipment to be modified and integrated into their operational system to deliver modifications to existing products. In addition, manufacturing firms strategically, through their research, introduce new technologically advanced products and processes to anticipate new consumer wants; whereas servicesector firms strategically, through information gathering, modify existing products to meet existing consumer needs. These differences underscore the differences between entrepreneurial activities in the two sectors. Both manufacturing and service innovation are built on the scientific base of knowledge. The manufacturing sector is more likely to build on the science base directly or in collaboration with universities. In contrast, the service sector purchases products and services as inputs that incorporate others’ research, which of course draws on the science base.

Innovation in the service sector 23 The role of infrastructure technologies is also different between the two sectors. Whereas infrastructure technologies reduce the market risk associated with the market introduction of a new product or process to the manufacturing firm, infrastructure technologies ensure that purchased technologies interface or integrate with the service-sector firm’s existing systems. Such infrastructure technologies emanate from the science base, and it is the science base that is at the root of the production of purchased technologies. An important component of the innovative process in both the manufacturing and service sectors is risk reduction. However, the focus of the activities differs. In the manufacturing sector, innovation is likely to be less integrated with marketing. Once a new product has been designed and tested, technical risk may be relatively low, but market risk may be significant because the product needs to be accepted and integrated into existing systems. In contrast, service-sector innovation is more likely to involve enhancements to products in existing markets, lowering market risk. However, limitations and the cost of testing increase technical risk, making risk reduction a key objective of the product enhancement phase of service innovation. As part of their risk reduction strategy, it is not uncommon for servicesector firms to outsource key components of product development or systems integration. However, many service-sector firms provide research as their primary service; thus, a component of their service is to assume the risk other firms are looking to outsource. These firms provide a key input into entrepreneurial activities similar to purchasing technology embedded in products or licensing technology.6 In summary, Figure 2.3 illustrates the process by which service-sector firms access and integrate technology with the goal of developing and providing enhanced services to their customers. These firms lead the strategic planning and entrepreneurial activities, as well as market development. They are likely to be heavily involved in the final stages of developmental research but may outsource a large share of the applied research and earlystage developmental research. Their role is often an integrator of existing technologies; however, they may also outsource significant systems integration activities.7

Appendix The entrepreneur as innovator Throughout intellectual history as we know it, the entrepreneur has worn many faces and played many roles.8 Neither economic theory nor economic

24

Innovation in the service sector

history has fully defined his visage. At least 12 distinct themes reside within the economics literature: ●

● ● ● ● ● ● ● ● ● ● ●

The entrepreneur is the person who assumes the risk associated with uncertainty. The entrepreneur is the person who supplies financial capital. The entrepreneur is an innovator. The entrepreneur is a decision maker. The entrepreneur is an industrial leader. The entrepreneur is a manager or superintendent. The entrepreneur is an organizer and coordinator of economic resources. The entrepreneur is the owner of an enterprise. The entrepreneur is an employer of factors of production. The entrepreneur is a contractor. The entrepreneur is an arbitrageur. The entrepreneur is an allocator of resources among alternative uses.

In this appendix we trace briefly the intellectual history of the entrepreneur as innovator because innovation is the role that we have defined to be subsumed in our models within the activity that we have named as entrepreneurial activity. Richard Cantillon (1680–1734) The term “entrepreneur” does not appear often in the prehistory of economics. It is a word of French origin that first made an obtrusive appearance in the writing of Cantillon, an eighteenth-century businessman and financier. Cantillon is significant as an introductory figure in this intellectual history not merely because he is often credited with first using the term entrepreneur but because he infused it with precise economic content and gave it analytic prominence. The fact that common, though imprecise, usage of the term existed prior to Cantillon is corroborated by an entry in Savary’s 1723 Dictionnaire Universel de Commerce in which entrepreneur is defined as one who undertakes a project; a manufacturer; a master builder. An earlier form of the word, entrepredeur, appears as early as the fourteenth century (Hoselitz, 1960). Throughout the sixteenth and seventeenth centuries, the most frequent usage of the term connoted a government contractor, usually of military fortifications or public works. Nicholas Baudeau (1730–92) One subsequent writer who developed a theory of entrepreneurship that foreshadowed future developments was the clergyman, the Abbe Nicolas

Innovation in the service sector 25 Baudeau. Baudeau treated the agricultural entrepreneur as a risk bearer, a concept that Cantillon discussed, but he added a distinctly modern twist. He made the entrepreneur an innovator as well, one who invents and applies new techniques or ideas to reduce his costs and thereby raise his profit. These new aspects of entrepreneurship, invention, and innovation, represent an important advance over Cantillon’s writings because they anticipate the most prominent twentieth-century formulation of entrepreneurship, Joseph Schumpeter’s theory of creative destruction. Consider the nature of risk faced by the agricultural entrepreneur. The rent he pays to the landlord is the surplus of farm revenue over necessary costs of production, including some payment for his own services. For the tenant farmer, rent is a cost determined in advance of production. The Physiocrats favored stabilizing these costs as much as possible through long-term leases, while wage rates were usually fixed at or near subsistence levels.9 Thus, the farmer operating with a long-term lease faced certain fixed costs, but uncertain harvests and hence uncertain sales prices. Baudeau emphasized and analyzed the significance of ability. Baudeau underscored the importance of intelligence, the entrepreneur’s ability to collect and process knowledge and information. Intelligence—knowledge and the ability to act—also gives the entrepreneur a measure of control, so that he is not a mere pawn to the capitalist. Baudeau paints the entrepreneur as an active agent (1910, p. 46): Such is the goal of the grand productive enterprises; first to increase the harvest by two, three, four, ten times if possible; secondly to reduce the amount of labor employed and so reduce costs by a half, a third, a fourth, or a tenth, whatever possible. Baudeau’s theory of entrepreneurship presupposes that economic events fall into two categories, those that are subject to human control and those that are not. To the extent that the entrepreneur confronts events under his control, his success depends on knowledge and ability. To the extent that he confronts events beyond his control, he places himself at risk. In this sense, Baudeau’s theory of entrepreneurship is more general than Cantillon’s, which concentrated on the effects of uncertainty without reference to administrative control. Jeremy Bentham (1748–1832) One British writer paid considerable attention to the entrepreneur. That writer was Bentham, whose ties with France and its intellectual tradition

26 Innovation in the service sector were much stronger than those of his contemporaries, especially Adam Smith (1723–90). The locus classicus of economic analysis in the eighteenth century was Adam Smith’s Inquiry Into the Nature and Causes of the Wealth of Nations (1776). Although a certified classic in the history of economics, it is notably deficient in one important sense. Smith failed to separate the entrepreneurial decision maker from among the various kinds of industrious people in the economy. Like Smith, Bentham understood that the regime most favorable to the development of inventive faculties was one of absolute liberalism. But, unlike Smith, Bentham (1952) defended usurers and projectors as useful sets of men. Both helped to advance the cause of inventive genius, each in his own way. It is something of a puzzle that Smith would, on the one hand, recognize innovation as a professional activity, while, on the other hand, ignore its importance in another context. In his denunciation of usury, Smith failed to see the importance of the innovator. Bentham aptly pointed this out in his Defence of Usury (1787), the first publication that brought him public recognition as an economist. There Bentham detailed how laws against usury limit the overall quantity of capital lent and borrowed and how such laws keep away foreign money from domestic capital markets. Both of these effects tend to throttle the activities of successful entrepreneurs. Although Bentham used the customary term projector, he was quite precise in his definition of this term as any person who, in the pursuit of wealth, strikes out into any new channel, especially into any channel of invention. He argued that interest rate ceilings tend to discriminate against entrepreneurs of new projects, because, by their novelty, such projects are more risky than those already proven profitable by experience. Moreover, legal restrictions of this sort are powerless to distinguish bad projects from good ones. In pleading the cause of the projectors, Bentham, the inventor of the Panopticon, was to some extent pleading his own case. Panopticon was the name Bentham gave to his idea of a model prison. The concept involved both an architectural and an institutional innovation. Bentham’s ideal prison was circular. All the cells were arranged concentrically round a central pavillion, which contained an inspector, or at most a small number of inspectors. From his central position the inspector could see at a glance everything that was going on, yet he was rendered invisible by a system of blinds. In this way, too, outside visitors could inspect the prisoners, as well as the prison’s administration, without being seen. According to Bentham, this constant scrutiny of the prisoners would deprive them of the power, and even the will, to do evil. The site that Bentham proposed for his model prison is now occupied by the Tate Gallery in London. Bentham was never able to attract enough backers to make his model prison a reality. The architectural idea behind the

Innovation in the service sector 27 Panopticon was first applied in Russia by Bentham’s brother Samuel, who in fact, deserves priority for the idea. Bentham’s unique contribution was an administrative innovation that is more to the point of our subject than the general problem of prison reform. Bentham completed the architectural innovation of the Panopticon by introducing an administrative arrangement that involved management by contract. What is especially interesting about this arrangement is the critical way that its success depends on the dynamic activities of the entrepreneur and the proper structuring of economic incentives. To Bentham, true reform would obtain in prisons only if the administrative plan simultaneously protected convicts against the harshness of their warders and society against the wastefulness of administrators. The choice, as he saw it, was between contract management and trust management. The differences are as follows (Halvy, 1955, p. 84): Contract-management is management by a man who treats with the government, and takes charge of the convicts at so much a head and applies their time and industry to his personal profit, as does a master with his apprentices. Trust-management is management by a single individual or by a committee, who keep up the establishment at the public expense, and pay into the treasury the products of the convicts’ work. In Bentham’s mind, the latter arrangement did not provide the proper junction of interest and duty on the part of the entrepreneur. Its success therefore depends on public interest as a motivating factor. Bentham, like his proclaimed mentor, Smith, had much more confidence in individual selfinterest as the spur to human action. The beauty of contract management was that it brought about an artificial identity of interests between the public on the one hand and the entrepreneur on the other. The entrepreneur in this case was an independent contractor who purchased, through competitive bid, the right to run the prison, thereby also acquiring title to whatever profits might be earned by the application of convict labor. Such an entrepreneurmanager could maximize his long-term gains by preserving the health and productivity of his worker-convicts. In this manner public interest became entwined with private interest. In 1787, Bentham completed the idea of contract management by a new administrative arrangement: he thought that life insurance offered an excellent means of joining the interest of one man to the preservation of a number of men. He therefore proposed that after consulting the appropriate mortality tables, the entrepreneur (prison manager) should be given a fixed sum of money for each convict due to die that year in prison, on condition that at the end of the year he must pay back the same sum for each convict

28

Innovation in the service sector

who had actually died in prison. The difference would be profit for the entrepreneur, who would thereby have an economic incentive to lower the average mortality rate in his prison (Bentham, 1962, p. 53). Aside from the fact that Bentham was virtually alone among British classical economists in his repeated emphasis on the entrepreneur as an agent of economic progress, it is noteworthy that his administrative arrangement of contract management recast the entrepreneur in the position of government contractor, that is, a franchisee who undertakes financial risk to obtain an uncertain profit. Bentham also explicitly tied his notion of entrepreneurcontractor to the act of invention. He defended contract management as the proper form of prison administration on the ground that it is a progressive innovation and should therefore be rewarded accordingly, no less than an inventor is rewarded for his invention (1962, p. 47). J.H. Von Thünen (1785–1850) Thünen is best known in the history of economics for his contributions to location theory, but in the second volume of The Isolated State (1850) he set forth an explanation of profit that clearly distinguished the return of the entrepreneur from that of the capitalist. What Thünen labeled entrepreneurial gain is profit minus (1) interest on invested capital, (2) insurance against business losses, and (3) the wages of management. This residual represents, for Thünen, a return to entrepreneurial risk. This last item Thünen identified as uninsurable risk, insofar as “there exists no insurance company that will cover all and every risk connected with a business. A part of the risk must always be accepted by the entrepreneur” (1960, p. 246). As Kanbur (1980) has argued, opportunity costs provide the basis for measuring this element of risk. Thünen seems to have had the same argument in mind when he wrote: He who has enough means to pay to get some knowledge and education for public service has a choice to become either a civil servant or, if equally suited for both kinds of jobs, to become an industrial entrepreneur. If he takes the first job, he is guaranteed subsistence for life; if he chooses the latter, an unfortunate economic situation may take all his property, and then his fate becomes that of a worker for daily wages. Under such unequal expectations for the future what could motivate him to become an entrepreneur if the probability of gain were not much greater than that of loss? (1960, p. 247) Thünen clearly appreciated the difference between management and entrepreneurship. He maintained that the effort of an entrepreneur working

Innovation in the service sector 29 on his own account was different from that of a paid substitute (manager), even if they have the same knowledge and ability. The entrepreneur is open to the anxiety and agitation that accompanies his business gamble; he spends many sleepless nights preoccupied with the single thought of how to avoid catastrophe, whereas the paid substitute, if he has worked well during the day and finds himself tired in the evening, can sleep soundly, secure in the knowledge of having performed his duty. Anyone who has nursed along a new enterprise knows precisely of what Thünen speaks. What is especially interesting about Thünen’s treatment is how he turns the discussion from the trials of the entrepreneur into a kind of crucible theory of the development of entrepreneurial talent. The sleepless nights of the entrepreneur are not unproductive; it is then that the entrepreneur makes his plans and arrives at solutions for avoiding business failure. Adversity in the business world thereby becomes a training ground for the entrepreneur. As Thünen put it: Necessity is the mother of invention; and so the entrepreneur through his troubles will become an inventor and explorer in his field. So, as the invention of a new and useful machine rightly gets the surplus which its application provides in comparison with an older machine, and this surplus is the compensation for his invention, in the same way what the entrepreneur brings about by greater mental effort in comparison with the paid manager is compensation for his industry, diligence, and ingenuity. (1960, p. 248) What makes this a significant step forward in the theory of entrepreneurship is the fact that Thünen successfully married the separate strands of entrepreneurial theory that, on the one hand, characterized the entrepreneur as risk bearer (Cantillon), and, on the other hand, portrayed him as innovator (Baudeau, Bentham). Economic analysis having come this far by 1850, we may well question whether Schumpeter took a step backward in the next century by excluding risk bearing from the nature of entrepreneurship, confining its meaning instead solely to innovative activity. Thünen was quite explicit about the fact that there are two elements in entrepreneurial income: a return to entrepreneurial risk and a return to ingenuity. Labeling the sum of these two as business profit, Thünen drew a succinct and precise distinction between entrepreneurship and capital use: Capital will give results, and is in the strict sense of the term capital, only if used productively; on the degree of this usefulness depends the rate of interest at which we lend capital. Productive use presupposes an

30

Innovation in the service sector industrial enterprise and an entrepreneur. The enterprise gives the entrepreneur a net yield after compensating for all expenses and costs. This net yield has two parts, business profits and capital use. (1960, p. 249)

Gustav Schmoller (1838–1917) The development of economic thought in the late nineteenth and early twentieth centuries progressed differently in Germany than it did in England or the rest of the Continent. This was partly because of the influence on economic method of the German Historical School. The historicists believed that in order to understand man’s economic behavior and the institutions that constrain it, economics must describe human motives and behavioral tendencies in psychologically realistic terms. The second generation of historicists is represented best by Schmoller. He amassed mountains of historical data to analyze actual economic behavior. From his examination of these data he discovered a unique central factor in all economic activity—the enterprising spirit, the Unternehmer, or entrepreneur. Schmoller’s entrepreneur was a creative organizer and manager whose role was innovation and the initiation of new projects (Zrinyi, 1962). He combined factors of production to yield either new products or new methods of production. Schmoller’s entrepreneur possessed imagination and daring. Werner Sombart (1863–1941) and Max Weber (1864–1920) Schmoller’s ideas were extended by third-generation historicists, Sombart and Weber (1864–1920). Sombart introduced a new leader who animates the entire economic system by creative innovations. This entrepreneur combined the powers of organization described by Schmoller with a personality and ability to elicit maximum productivity from individuals engaged in the productive process. Sombart painted the entrepreneur as a profit maximizer, whether he be a financier, manufacturer, or trader. The German historicists characterized the entrepreneurial process as a breaking away from the old methods of production and the creation of new ones. This disequilibrating process was particularly emphasized by Weber. He sought to explain how a social system, as compared to an individual enterprise, could evolve from one stable form (perhaps under an authoritarian structure) to another type of system. Historically, he identified such changes with a charismatic leader, or entrepreneur-like person (Carlin, 1956).

Innovation in the service sector 31 Like so many writers, Weber began his analysis of change with a stationary state construct: We may . . . visualize an economic process which merely reproduces itself at constant rates; a given population, not changing in either numbers or age distribution. . . . The tasks (wants) of households are given and do not change. The ways of production and usances of commerce are optimal from the standpoint of the firm’s interest and with respect to existing horizons and possibilities, hence do not change either, unless some datum changes or some chance event intrudes upon this world. (1930, p. 67) In such a stationary society there is nothing that requires the activity traditionally associated with the entrepreneur. “No other than ordinary routine work has to be done in this stationary society,” declared Weber, “either by workmen or managers” (1930, p. 67). Yet, inevitably, change occurs. Weber described a likely scenario: Now at some time this leisureliness was suddenly destroyed, and often entirely without any essential change in form of organization . . . What happened was, on the contrary, often no more than this: Some young men from one of the putting-out families went out into the country, carefully chose weavers from his employ, greatly increased the rigor of his supervision of their work, and thus turned them from peasants into laborers . . . he would begin to change his marketing methods . . . he began to introduce the principle of low prices and large turnover. There was repeated what everywhere and always is the result of such a process of rationalization: those who would not follow suit had to go out of business. The idyllic state collapsed under the pressure of a bitter competitive struggle. (1930, p. 68) From this context, the alteration described is driven by an entrepreneur type. The critical characteristics of Weber’s successful entrepreneur are his religious imperatives, which make up what is called the Protestant ethic. In the final analysis, therefore, Weber’s theory of social and economic change is as much sociology as economics. Joseph Schumpeter (1883–1950) Schumpeter revealed his concept of the entrepreneur within the broader scope of a theory of economic development. To Schumpeter, development is a dynamic process, a disturbing of the economic status quo. He looked on

32 Innovation in the service sector economic development not as a mere adjunct to the central body of orthodox economic theory, but as the basis for reinterpreting a vital process that had been crowded out of mainstream economic analysis by the static, general equilibrium approach. The entrepreneur is a key figure for Schumpeter because, quite simply, he is the persona causa of economic development. Schumpeter combined ideas from many earlier writers, but rather than slavishly imitate the work of others, he melded these elements into something uniquely his own. To Schumpeter, competition involved mainly the dynamic innovations of the entrepreneur. This view is most clearly and completely set forth in his Theory of Economic Development, first published in 1911, and echoed in later works of 1939 and 1950. Schumpeter used the concept of equilibrium—a theoretical construct and a point of departure. He coined a phrase to describe this equilibrium state—the circular flow of economic life. Its chief characteristic is that economic life proceeds routinely on the basis of past experience; there are no forces evident for any change of the status quo. Schumpeter outlined the nature of production and distribution in the circular flow in the following passage: [I]n every period only products which were produced in the previous period are consumed, and . . . only products which will be consumed in the following period are produced. Therefore workers and landlords always exchange their productive services for present consumption goods only, whether the former are employed directly or only indirectly in the production of consumption goods. There is no necessity for them to exchange their services of labor and land for future goods or for promises of future consumption goods or to apply for any “advances” of present consumption goods. It is simply a matter of exchange, and not of credit transactions. The element of time plays no part. All products are only products and nothing more. For the individual firm it is a matter of complete indifference whether it produces means of production or consumption goods. In both cases the product is paid for immediately and at its full value. (1934, pp. 42–43) Within this system, the production function is invariant, although factor substitution is possible within the limits of known technological horizons. The only real function that must be performed in this state is “that of combining the two original factors of production, and this function is performed in every period mechanically as it were, of its own accord, without requiring a personal element distinguishable from superintendence and similar things” (1934, p. 45). In this artificial situation, the entrepreneur is

Innovation in the service sector 33 a nonentity. “If we choose to call the manager or owner of a business ‘entrepreneur’,” wrote Schumpeter (1934, pp. 45–46), then he would be an entrepreneur of the kind without special function and without income of a special kind. For Schumpeter, the circular flow is a mere foil. The really relevant problem, he wrote in Capitalism, Socialism and Democracy (1950), is not how capitalism administers existing structures, but how it creates and destroys them. This process—what Schumpeter called creative destruction—is the essence of economic development. In other words, development is a disturbance of the circular flow. It occurs in industrial and commercial life, not in consumption. It is a process defined by the carrying out of new combinations in production. It is accomplished by the entrepreneur. Schumpeter reduced his theory to three elemental and corresponding pairs of opposites: (1) the circular flow (i.e. tendency toward equilibrium) on the one hand versus a change in economic routine or data on the other, (2) statics versus dynamics, and (3) entrepreneurship versus management. The first pair consists of two real processes; the second, two theoretical apparatuses; the third, two distinct types of conduct. The theory maintained that the essential function of the entrepreneur is distinct from that of capitalist, land-owner, laborer, or inventor. According to Schumpeter, the entrepreneur may be any and all of these things, but if he is, it is by coincidence rather than by nature of function. Nor is the entrepreneurial function, in principle, connected with the possession of wealth, even though “the accidental fact of the possession of wealth constitutes a practical advantage” (1934, p. 101). Moreover, entrepreneurs do not form a social class, in the technical sense, although they come to be esteemed for their ability in a capitalist society. Schumpeter admitted that the essential function of the entrepreneur is almost always mingled with other functions, such as management. But management, he asserted, does not elicit the truly distinctive role of the entrepreneur. “The function of superintendence in itself, constitutes no essential economic distinction,” he declared (1934, p. 20). The function of making decisions is another matter, however. In Schumpeter’s theory, the dynamic entrepreneur is the person who innovates, who makes new combinations in production. Schumpeter described innovation in several ways. Initially he spelled out the kinds of new combinations that underlie economic development. They encompass the following: (1) creation of a new good or new quality of good, (2) creation of a new method of production, (3) the opening of a new market, (4) the capture of a new source of supply, and (5) a new organization of industry (e.g. creation or destruction of a monopoly). Over time, of course, the force of these new combinations dissipates, as the new becomes part of the

34

Innovation in the service sector

old (circular flow). But this does not change the essence of the entrepreneurial function. According to Schumpeter, “everyone is an entrepreneur only when he actually ‘carries out new combinations’, and loses that character as soon as he has built up his business, when he settles down to running it as other people run their businesses” (1934, p. 78). Alternatively, Schumpeter defined innovation by means of the production function. The production function, he said, “describes the way in which quantity of product varies if quantities of factors vary. If, instead of quantities of factors, we vary the form of the function, we have an innovation” (1939, p. 62). Mere cost-reducing adaptations of knowledge lead only to new supply schedules of existing goods, however, so this kind of innovation must involve a new commodity, or one of higher quality. However, the knowledge undergirding the innovation need not be new, as Schumpeter recognized. On the contrary, it may be existing knowledge that has not been used before. According to Schumpeter: [T]here never has been anytime when the store of scientific knowledge has yielded all it could in the way of industrial improvement, and, on the other hand, it is not the knowledge that matters, but the successful solution of the task sui generis of putting an untried method into practice—there may be, and often is, no scientific novelty involved at all, and even if it be involved, this does not make any difference to the nature of the process. (1928, p. 378) In Schumpeter’s theory, successful innovation requires an act of will, not of intellect. It depends, therefore, on leadership, not intelligence, and it should not be confused with invention. On this last point, Schumpeter was explicit: To carry any improvement into effect is a task entirely different from the inventing of it, and a task, moreover, requiring entirely different kinds of aptitudes. Although entrepreneurs of course may be inventors just as they may be capitalists, they are inventors not by nature of their function but by coincidence and vice versa. Besides, the innovations which it is the function of entrepreneurs to carry out need not necessarily be any inventions at all. (1934, pp. 88–89) The leadership that constitutes innovation in the Schumpeterian system is disparate, not homogeneous. An aptitude for leadership stems in part from the use of knowledge, and knowledge has aspects of a public good. People of

Innovation in the service sector 35 action who perceive and react to knowledge do so in various ways; each internalizes the public good in potentially a different way. The leader distances himself from the manager by virtue of his aptitude. According to Schumpeter, different aptitudes for the routine work of “static” management result merely in differential success at what all managers do, whereas different leadership aptitudes mean that “some are able to under-take uncertainties incident to what has not been done before; [indeed] . . . to overcome these difficulties incident to change of practice is the function of the entrepreneur” (1928, p. 380). Schumpeter’s influence on the theory of economic development has been enormous, in part because it derives from its simplicity and its power. This simplicity and power is summed up in the Schumpeterian phrase, “The carrying out of new combinations we call ‘enterprise’; the individual whose function it is to carry them out we call ‘entrepreneurs’ ” (1934, p. 74).

3

Telecommunications industry

Introduction Telecommunications has facilitated the death of distance, enabling service providers and users to conduct business without regard to geographical boundaries. Information and communication technology has become an inseparable component of any business in the economy of the twenty-first century. The telecommunications industry’s innovations have lowered transaction costs to domestic businesses and individuals through continued efforts toward developing and improving standards, technological infrastructure, reliability, and security. R&D in the telecommunications industry ranges from basic and applied research for electronic devices to development for the integration of networks and systems. In contrast with other sectors, almost all of the R&D expenditures reported for the industry—NAICS 517— appear to be directly related to the industry (i.e. providing communications services).1 This chapter begins with a profile of the telecommunications industry and its ongoing R&D expenditures and activities, followed by a case study focused on the development of wireless communications (specifically, Wi-Fi—wireless fidelity—applications) to investigate the roles different stakeholders play in the technology development and deployment process.

Industry profile and R&D statistics The telecommunications industry consists of a broad range of firms operating, maintaining, and/or providing access to facilities for the transmission of voice, data, text, and full motion picture video between network termination points and telecommunications reselling. Local, long-distance, and international voice telephony and data transmission services are the three main sources of service provider revenue. In 2003, total sales for this sector exceeded $388 billion and employment exceeded 1 million. See Table 3.1.

15 8 13 23 172

5173 5174 5175 5179 517

Note NA: not available from COMPUSTAT.

Source: COMPUSTAT (2003).

72 41

5171 5172

Wired telecommunications carriers Wireless telecommunications carriers (except satellite) Telecommunications resellers Satellite telecommunications Cable and other program distribution Other telecommunications Telecommunications

Number of firms

NAICS

Name

Table 3.1 Summary of the telecommunications industry (2003)

3,316 7,625 22,159 3,635 388,599

271,927 79,937

Sales ($ millions)

10 33 56 74 503

314 16

R&D ($ millions)

12 3 74 21 1,098

811 176

Employment (thousands)

887 9,635 748 3,488 NA

388 91

R&D intensity ($/employee)

Note NA: not available from COMPUSTAT.

Source: COMPUSTAT (2003).

92 381 289.18

141 5,732 178 103

517510 517410 517110 517212

517410 517910 517110

34,529 314 5,739

517110 517910 517510

AT&T Corporation Corvis Corporation Echostar Communication Corporation TIVO Inc. Echostar DBS Corporation IBASIS Inc. Boston Communications Group XM Satellite Radio Inc. PTEK Holdings Inc. ZTEL Technologies Inc.

Sales ($ millions)

NAICS

Name

0.4 2.0 1.19

0.3 NA 0.2 0.4

61.6 1.2 15.0

Employees (thousands)

12 9 6

22 20 13 13

277 59 32

R&D ($ millions)

Table 3.2 Top 10 telecommunications service providers, by R&D expenditures (2003)

Cable and other program distribution Wired telecommunications carriers Wired telecommunications services Cellular and other wireless telecommunications Satellite telecommunications Other telecommunications Wired telecommunications carriers

Multimedia delivery service Networking equipment design Satellite network services

Service description

Telecommunications industry 39 In recent years, the telecommunications industry has been integral in the development of the information economy. Telecommunications is a dynamic industry that has undergone tremendous change. Today’s telecommunications market is the product of deregulation, consolidation, and technological innovations. During the 1980s, partial deregulation focused on long-distance operators at a modular point of intersection between local and long-haul voice networks. Deregulation intensified in the 1990s, resulting in the 1996 US Telecommunications Reform Act. This Act prompted investment for innovations in areas outside the traditional telephone market of the telecommunications industry. The local area network (LAN), a simple, flexible technology enabled by Internet protocol (IP) and Ethernet standards, spawned innovations such as the multiprotocol router that disrupted a series of highly reliable but rigid incumbent technologies and accelerated the rise of a new paradigm. Improvements in Voice over IP (VoIP) technology are now enabling the multiplexing of voice, video, and data onto a single data network (Christensen et al., 2001). Once closely linked, a division has emerged between transport and services since the emergence of IP networking. IP networking allows the decoupling of network services from their reliance on transmission media. Another important distinction is that of wireless versus wired connections. Although still dependent on the transmission network infrastructure for origination and completion, wireless technologies have untethered devices from the Public Switched Telephone Network (PSTN) and enterprise data networks, creating new opportunities associated with mobility. Regarding the size and technology areas of telecommunications firms, the top 10 publicly traded telecommunications firms, ranked by R&D expenditures, are in Table 3.2.2 AT&T, the largest firm in terms of R&D expenditures, allocated $277 million in 2003. It provides a variety of services covering the spectrum of media delivery systems, from television and telephone to Internet and satellite. AT&T operates its own R&D research laboratories dedicated to seeking innovations in communications based on network operations, IP applications, and systems automation, in addition to other more long-term interests in communications technology.

R&D activities in wireless communications services Moving into the twenty-first century, wireless communications services have expanded rapidly. While maintaining wireless telephone service, service providers have begun to offer business applications such as wireless LANs

40

Telecommunications industry

(W-LAN), pagers, laptops with wireless modems, personal digital assistants (PDAs) with wireless connectivity, and cellular phone service. In general, wireless communications services can be grouped as follows: ●

●

●

●

Hybrid approach: Personal communications services (PCS), which incorporate both wire and wireless technology. Hybrid technology allows carriers to bypass the traditional stationary wireless local loop technology. Dispatch services: Intercom communication for individuals who need to communicate frequently each day. Wireless data and Internet access: Combination of W-LAN and broadband to create fixed wireless broadband services like local multipoint distribution service (LMDS) and multichannel multipoint distribution service, which support simultaneous voice and data transfers and eventually video. Wi-Fi (or W-LAN): Transmittance of data over high radio frequencies (2.4 GHz) designed for short distances such as a hotel, office building, or college campus.

A CIO Magazine survey reported that enterprise wireless solutions (a.k.a., wireless broadband) are becoming increasingly important in business operations for both manufacturing and nonmanufacturing firms. Firms are increasing consumption of the wireless infrastructure motivated by productivity increases, streamlining, and customer satisfaction. Today’s wireless service providers are working to improve security, accessibility, reliability, and speed through advances in the wireless technology. Security As businesses have become more dependent on wireless networks for their operations, the demand for security has increased dramatically. Security was the primary topic of a recent meeting of the world’s largest communications firms, as the industry organization Alliance for Telecommunications Industry Solutions (ATIS) hosted a security summit entitled “Security of Service Provider Infrastructure in the Era of Convergence.” The wireless communications industry is increasingly interested in eliminating unauthorized access to operations support systems. The 3 G Partnership Project, a consortium of wireless industry global leaders formed to develop and standardize 3 G technology, is addressing the security issues related to 3 G wireless technology. Their R&D activities include general packet radio service (GPRS) ciphering algorithms, immediate service termination (IST) and access security for IP-based services, network domain security (NDS), requirements for security architecture and interfaces,

Telecommunications industry 41 interoperability of multitechnology networks and encryption, and user authentication of information as it travels from server to user. Improved accessibility R&D to support accessibility includes developing and standardizing wireless markup languages. Wireless providers are working to further the wireless application protocol (WAP) programming languages. WAP is a grouping of standards related to accessing the Internet, using e-mail, receiving faxes, and conducting monetary transactions via digital wireless products such as mobile phones, pagers, and PDAs. Other R&D activities include Bluetooth technology research to offer short-range wireless connections between mobile phones and headsets, keyboard or mouse with personal computer, and mobile phone and PDA. Ongoing Wi-Fi and W-LAN research is being conducted to support the transmittance of data over radio frequencies designed for short distances such as at hotels, office buildings, or college campuses. Wi-Fi technology is preferable to its wired counterpart LAN, mostly because of the increase in mobility of workers across a workspace. Reliability Wireless networks quickly pass large amounts of information. In early wireless applications, reliability was always an issue. Mobile phone users could suddenly lose their connection or intercept other information. The Institute of Electrical and Electronics Engineers (IEEE) developed a set of standards known as 802.11 that were designed to manage data packets as they are transmitted across the wireless network to avoid collision and disruption.

Case study of the Wi-Fi technology supply chain Interviews were conducted with organizations throughout the supply chain for Wi-Fi products and services. Wi-Fi has become a catch phrase to describe W-LANs, electronic equipment, and services enabled by one of several 802.11 standards from the IEEE. Wi-Fi functionality includes the ability to share resources such as Internet content, e-mail, voice communications, and video and pictures. Description of Wi-Fi services and technologies Wi-Fi technology allows the exchange of information, such as Internet content, e-mail, and digital images, at speeds ranging from 2 to 100 Mbps

42

Telecommunications industry

(megabytes per second) from a remote location over a W-LAN. The market for Wi-Fi products, such as notebooks and other portable devices, has been growing for several years, and the number of individual Wi-Fi subscribers is projected to increase from 12 million to over 700 million in the next 5 years. There are a number of variations to the 802.11 standard. The current version relating to access speed is 802.11 g. Ratified by the IEEE standards committee on June 12, 2003, this standard allows information transmission rates of 20
Mbps in the 2.4 GHz radio frequency spectrum. Wi-Fi network equipment currently available to consumers is built on 802.11 g or one of two older standards known as 802.11 a and 802.11 b. Devices enabled with the most recent standard are compatible with the 802.11 b standard equipment and devices. The equipment market for Wi-Fi includes network infrastructure, end-user devices, back-end software applications, authentication software, and subscriber usage tracking software. Network infrastructure includes access points or hotspots that transmit and receive Wi-Fi signals from individual user’s devices. Hotspots consist of a radio frequency base station and a wired high-speed network connection. Depending on the location, hotspots also use amplifiers, antennae, cellularto-Wi-Fi switching devices, and network management platforms to manage and extend the reach of a network connection. Furthermore, networking equipment designed for commercial use is integrated with back-end network management software applications, allowing the network operator to monitor, manage, and track subscriber usage and to allocate additional bandwidth during peak usage. A network operator provides high-speed connection services to subscribers using this equipment. Wi-Fi service is defined as the aggregation of the equipment outlined above to allow an end user or service subscriber, such as a business traveler, to connect to the Internet via one of the network’s hotspots. Wireless Internet service provider (WISP) is the term used to describe firms that provide wireless Internet connectivity. Similar to cellular phone carriers, WISPs monitor their networks, track usage, and expand their coverage to remain competitive. Activities performed by the service providers include developing customized access portals to ensure security and privacy to subscribers, monitoring subscriber usage, adjusting broadband allocations to individual hotspots, and expanding the reach of their existing network infrastructure. Software development of access portals is performed in house or outsourced depending on the level of complexity and required security encryption. Billing system software applications integrated with subscriber usage tracking software, which represents component technology, can be purchased or developed through in-house R&D. Expansion of the network is another activity service providers perform. However, networks suffer from the limited

Telecommunications industry 43 reach of access points. As a result, WISPs, much like cellular phone providers, enter into service roaming agreements with other regional-specific networks as a way of expanding the number of access points a network operator can offer to its subscribers. Wi-Fi equipment manufacturers and service industry description The supply chain for developing and deploying Wi-Fi services and products begins with original equipment manufacturers (OEMs) and software programmers, many of which participated in developing the 802.11 standard within the IEEE standards organization. Equipment and software applications developers then assist network operators by developing network connectivity equipment and security and network management software. Also supporting network operators are system integration firms and business support software developers. Network operators then provide services either directly to end users (e.g. individual consumer, institution, or enterprise) or through aggregators. Figure 3.1 illustrates the development supply chain for Wi-Fi services. Original equipment manufacturers The IEEE develops network standards using input from manufacturers, vendors, and network operators from around the world. OEMs have traditionally dominated participation in wireless networking standards development.

Supply chain participants

Subscribers end users

Aggregators

Products or services

Access to network

OEM

Equipment

Integration services

RD&T firms

Standards orgs.

Systems integration

Figure 3.1 Development supply chain for Wi-Fi services. Source: NIST (2005).

Commercial Subscribe Network deployment

Network operator

Accounting services Network/ settlement agency

44

Telecommunications industry

However, the number of service providers participating in these standards committees has been growing rapidly (Meyers, 2004). Telephony Online, a communications industry information source, points out the trend in increased participation from service providers in the standards development of wireless broadband access technologies. Two major reasons for the increased participation reported in the TelephonyOnline.com article for involving service providers were to educate them on the technology’s capabilities and to gain information on the functional requirements that need to be included in the standard. The equipment market is built on a number of strategic partnerships between OEMs and software development firms. OEMs spend considerable time identifying the operational needs of their Wi-Fi service provider customers. Operation requirements articulated by service firms include interoperability with existing data services, security, and accounting systems. OEMs, in cooperation with specialized back-end business process software development firms, then create both a Wi-Fi networking technology business platform and the equipment required to operate a network. OEMs often contract with other upstream suppliers to build and install the network access points or hotspots. Network operators Wi-Fi service providers’ (network operators’) overall strategy for market growth is to target people in transit. This targeting includes business travelers, public service employees, and metropolitan public networks. The Wi-Fi service market, or hotspot market, revolves around expanding the overall reach or scope of a service provider’s network. The goal is to maximize the number of locations that the service provider can offer to subscribers as access points. Taking the purchased technology platform developed by the OEM, the network operator builds out a physical network of access points in a certain geographical location. Most likely this location is in an urban setting, where there is potential for a large number of mobile users. Another strategy has been to build networks in more remote locations to act as a last mile option for rural communities, where the cost-effective concerns have limited the development of wired network infrastructure. The network operator then develops and manages access and authentication, security, and roaming subscriber authentication through an operation support system (OSS) that enables the operator to manage a large hotspot market. Aggregators Aggregators specialize in developing roaming agreements that expand the reach of any single network.3 These firms do not operate networks but instead

Telecommunications industry 45 develop a subscriber base under a brand name (e.g. Boingo Wireless, iPass, and GRIC) and then contract with existing Wi-Fi network operators to gain access for their subscribers to an assortment of networks, thereby expanding the reach of a given network for the end user. Conversely, the aggregator can work as a contract vehicle, providing a network operator with extended network roaming capabilities for its subscriber base via the network roaming agreement that the aggregators have already established. Additional supply chain participants The origin of the Wi-Fi service market lies in specifying the new 802.11 standard for wireless networking from IEEE. Following, and even during, ratification of 802.11 g by the IEEE in 2003, electronic and networking equipment manufacturers began to integrate this new wireless networking technology into their existing products, such as laptops, and develop a network management platform for Wi-Fi network operators. The development and deployment process is described for two types of service firms: network operators without traditional R&D activities and network operators with traditional R&D activities. Wireless service providers use technology products manufactured by OEMs to allow subscribers to wirelessly connect to the service provider’s network. The extent to which a service-sector firm plays a role in developing new products and services is determined largely by the overall size of the firm, either in revenue or market share. Related to the development process for Wi-Fi services, a typical firm might be MHO Networks, which offers regional service in Denver, Colorado. This firm started out as a retail custom computer reseller and has evolved to a full-blown wireless network operator offering wireless network access in select cities and communities throughout the state of Colorado. Figure 3.2 depicts a small network operator’s involvement in developing its wireless service capabilities (technologies) and its interactions within the supply chain. The figure illustrates that small network operators are involved in product development primarily through providing feedback into OEM road-mapping activities, providing feedback into product specifications, participating in deployment tests, coordinating system integration, and leading service support software development. The OEMs coordinate the product development process, building on basic and applied research performed through a combination of in-house, institutional, and commercial partnerships. The OEM typically identifies those industries or markets that would derive the greatest benefit from wireless technology. Then in conjunction with the current technology capabilities and trends in future standards development, the OEM articulates a technology

Installation and maintenance of access points

Process improvement

Systems integration

Source: NIST (2005).

Integration services

Deployment tests

Small network operator

Figure 3.2 Development activities performed by small-sized network operators.

Standards specifications

Contract research

Trouble shooting

Product specification

Conformance testing and error Beta version correction

Market assessment

Product specification customization

Standards orgs.

Platform

Product conceptualization

RD&T firms

Roadmap Basic research Technology development Strategic trajectory Target markets

OEM

Products or services

Flows of information

Activities

Supply chain participants

Network deployment

Accounting software development

Network/settlement agency

Accounting services

Authentication software versions

Billing recognition software development

Network monitoring software development

Standards configuring

Commercial

Subscribers end users

Access to network

Aggregators

Subscribe

Telecommunications industry 47 road map. The road map aligns the firm’s internal strategic trajectory with the current and future trends of wireless technology. The manufacturing firm leverages input from the industry leaders, such as existing network operators, to fine tune a technology road map, aligning its development strategy with market trends reported by the perspective consumers. OEMs spend considerable time iterating with network operators to ensure that the product or equipment being developed under the 802.11 standard addresses the operator’s business process needs. Equipped with knowledge of the network operator industry’s market trends, the OEM engages in product conceptualization and begins to build the wireless networking platform aimed at addressing the business needs of the targeted industry of wireless network operators. The OEM partners with software application development firms that specialize in back-end business software for the wireless network operators to build the initial structure of the wireless networking platform. Following the development of the platform, the OEM again seeks network operators’ feedback on product specifications, and for larger service firms this may include some customization. The network operators articulate the specific capabilities and operational requirements that the networking platform must address to meet their business needs. Next, the platform enters a conformance testing phase, also referred to as an alpha testing phase, where network operators take part in developing test scenarios that will allow the product designers to evaluate how well the product will perform in a real business environment. The service firm provides feedback on bugs and glitches in the system. The OEM then sends the initial product to wireless network operators to use in a real-world environment. This portion of the process is known as a deployment testing phase, also referred to as beta testing phase. At this point, the network operator takes a role in the development process through testing and offering feedback to the OEM, creating a virtually integrated supply chain for R&D activities. Once all issues and bugs have been sufficiently addressed, the OEM can roll out its new wireless networking platform. The network operator purchases the platform from the OEM and then begins to build out its wireless network. The operator integrates the new platform with any existing data services it operates. The operator then develops customized software for authentication and access applications and ensures security and a methodology to ensure that the network is configured to meet all current standards in use. In addition, network operators work to develop methods to more effectively track subscriber usage and link usage to billing software applications. Development of the physical infrastructure, such as access points, is typically outsourced to network construction service firms.

48

Telecommunications industry

Once their local network is in place, network operators use aggregators to maximize geographical coverage. This optimization enables the subscribers to have the greatest mobility and variety in access points from a particular network operator. Roaming agreements allow networks to expand without large capital investments. Network operators with larger development budgets have more traditional research labs. Product and service development occurs in much the same way as described above, but the large network providers internalize more of the development process within technology divisions. These larger firms tend to lead, as opposed to participate in as do smaller firms, development activities related to technology road mapping, service conceptualization, compliance testing, integration of new systems with existing legacy systems, deployment testing, standards configuring, and accounting and billing software development. Technology divisions for larger network operators serve to maintain an awareness of the existing and emerging equipment and software on the market for various networking standards and technologies, including Wi-Fi, virtual private network (VPN), and WiMax. This technology monitoring is used to inform their strategic planning, and these service providers are likely to lead the development of their own technology road mapping activities (whereas in the case of small network providers, the firm simply provides input on emerging market industry trends). Large network operators have knowledge of available equipment and systems and their capabilities. The network operator’s in-house resources allow it to conduct compliance and deployment testing and network modification internally. However, the carrier still offers feedback to manufacturers during the development of the platform and networking equipment. Figure 3.3 suggests an optional scenario in which these larger firms conduct more development activities in house. In Figure 3.3, the network operator is still providing market assessment input to the OEM’s product development process and some product specification. However, compliance and deployment testing, as well as required modification, are typically conducted internally. Larger network operators also invest significant resources in standards development organizations, as noted earlier in this discussion. Their involvement in developing standards serves three purposes. First, they are exposed to the array of competing vendors for various technologies, which allows the network operator to make a better informed purchase of networking equipment that will meet the specific business needs of individual network operators. Second, by offering knowledge of business operations

Telecommunications industry 49 Supply chain participants Aggregators

Activities

Subscribe

Flows of information Products and services Market assessment

Network access Large network operator

OEM R&D

Equipment Software applications

Roadmapping Target markets Strategic trajectory Service conceptualization

Deployment testing

Standards configuration Accounting and billing software modification

Subscribe

Standards organization Standards specification

Busines operations needs

Compliance tests

Network equipment installation

Modification to address legacy issues

Figure 3.3 Development activities performed by large-sized network operators. Source: NIST (2005).

requirements, network operators shorten the time required to bring new or improved services to subscribers. Finally, these firms are enhancing their ability to absorb future technology change through establishing a continued presence in the standards development process.

Activity categories and research taxonomies Network operators are involved in the product development process throughout the technology supply chain. Both large and small network operators contribute to similar development tasks to roll out new services. However, in the case of a small network operator, many R&D activities take place outside the firm, with suppliers receiving input from the network operator, while the larger-sized firms are more likely to internally lead many of the activities, such as compliance and deployment testing activities and customization. Potential R&D activities conducted by in-house network operators or by supporting suppliers include road mapping activities; participation in standards organizations; product conceptualization, specification, and customization; compliance testing; deployment testing; and development and integration of network monitoring and business support software.

50

Telecommunications industry

Conclusions This case study represents an example of how one technology-intensive service industry can develop new services by acquiring new or emerging technologies and adapting them to fit into an existing service infrastructure. In the case of a small network provider, technology is adapted and then managed; feedback loops are established with OEMs to improve the networking products being manufactured. In the case of the larger network operators, innovation rests in the service firm articulating the development and deployment of the new service, bundling it with existing services. In relation to the service-sector model of innovation presented in Chapter 2 (Figure 2.3), the service provider—in this case, the network operator— innovation is taking place as the supporting infrastructural technologies acquired from OEMs are brought together to create a system that provides a novel or enhanced service. The network operators bear the majority of the product and service enhancement risk. Market risk, which is secondary, is in large part passed on to aggregators.

4

Financial services industry

Introduction Over the past decade, changes in the US economy and the increased use of IT in financial services have significantly influenced innovative activities within the financial services sector. Increased competition through deregulation, consolidation, and disintermediation is forcing financial services firms to innovate to maintain profitability. At the same time, the demand for financial services has increased as the baby boom generation approaches retirement. The simultaneous increase in competition and demand is driving the industry to increase R&D to develop both new services and low-cost delivery devices for new and existing services. The introduction of e-money, smart cards, e-checks, e-funds transfer, and improved encryption are some of the innovations developed in financial services in the twenty-first century. An additional trend influencing innovation is the disintermediation of financial services as manufacturers begin to encapsulate their products with services and deal directly with customers. This trend is forcing service-sector firms to reduce costs by developing more efficient technologies and move into new product and service areas to maintain profitability. This chapter profiles the financial services industry and its ongoing R&D expenditures and innovative activities. Findings from industry interviews are presented to provide an overview of the technology development and deployment process, with a focus on financial Web-based services.

Industry profile and R&D statistics Firms in the financial services industry, generally NAICS 52 and 53, consist of depository institutions; nondepository institutions; security and commodity

52

Financial services industry

brokers; insurance carriers; insurance agents, brokers, and services; real estate; and holding and other investment institutions. Depository institutions include firms engaged in deposit banking and fiduciary activities. Nondepository institutions include firms engaged in extending credit in the form of loans but not engaged in deposit banking. Security and commodity brokers, dealers, exchanges, and services include firms engaged in the underwriting, purchase, sale, or brokerage of securities and other financial contracts on their own account or for the account of others, and exchanges, exchange clearinghouses, and other services allied with the exchange of securities and commodities. Insurance carriers include carriers of insurance of all types, including reinsurance. Insurance agents, brokers, and services include agents and brokers dealing in insurance and organizations offering services to insurance firms and policyholders. Real estate includes real estate operators and owners and lessors of real property as well as buyers, sellers, developers, agents, and brokers. Holding and other investment offices include investment trusts, investment firms, holding companies, and miscellaneous investment offices. In 2003, total revenue for this sector exceeded $1.9 trillion with employment over 4.3 million. See Table 4.1 for aggregate and subindustry information. The financial services industry is too large and too diverse to examine in its entirety using interview-based tools. And much of the R&D performed by firms classified under NAICS 52 and 53 is not directly related to the provision of financial services. Please refer to this chapter’s appendix for a discussion of the diversity of firms classified in the financial services sector. The remainder of this chapter focuses on financial firms classified under NAICS 523 as securities, commodities, contracts, and other financial investments and related activities. Table 4.1 Summary of financial services industry (2003) NAICS Name 5222 5222 5231 5241 5242 5311 5511 5253

Number of Employment Sales firms (thousands) ($ millions)

Depository institutions 672 Nondepository institutions 99 Security, commodity broker 98 Insurance carriers 200 Insurance agents, brokers, services 43 Real estate 92 Holding and other investment offices 875 Finance, insurance, and real estate 2,079

Source: COMPUSTAT (2003).

2,228 359 322 1,007 165 79 168 4,329

715,938 217,974 188,342 694,385 30,021 14,533 66,600 1,927,792

Financial services industry 53

Case study of web services technology Web services technology was identified for this project as a case study to demonstrate how innovation occurs within financial services firms. The firms interviewed are listed in Table 4.2. However, in a number of our interviews there was a shift in the discussion from Web services to more general concepts about R&D, how those activities are measured, and how they relate to the innovation process in general. The case example of Web services is provided in the appendix to this chapter. The following discussion focuses on the financial services industry, in general, and R&D and innovation in the industry, in particular, in an effort to complement the information about the industry and relate to our model of innovation in Chapter 2.1 Defining financial services The firms interviewed defined financial services as an industry that provides services in, but not limited to, retail banking, debt and asset management, and private and institutional investment. Financial services firms provide service to a variety of clients, ranging from individuals to large institutions. Institutional client services include debt management, capital financing, public and private offerings of debt (stock and bonds), and equity, as well as other securities. The provision of these services is based on a skills set of underlying knowledge and experience to ensure high returns on investment for the client (consumer) and the financial services firm (producer). Financial services change over time in response to consumer preferences. Firms are continuously working to develop higher quality, more efficient, and Table 4.2 Firms interviewed related to Web services Name Computer Service Firms Accenture

Position of the interviewee

Niteo Partners

Associate partner for financial services group Project manager

Financial service institutions Merrill Lynch JP Morgan Chase Wachovia

Chief technology architect VP of treasury security services VP of retail integration

Industry research firm Forrester research

Principal analyst

Source: NIST (2005).

54

Financial services industry

less expensive products that will allow them to gain competitive advantage in their industry or sector. To this end, technology is developed and/or acquired and used to increase the productivity of the existing service and to develop new services. The development and application of intellectual capital (IC) is the driving force behind innovation. Financial services can be either labor intensive, relying heavily on personal interactions and human capital, or capital intensive, relying on automation to lower the cost of transactions and the dissemination of information. There are two general categories of these services: investment services and retail banking. The distinction between these two segments lies in the level of technology needed to meet consumer demand for services in the respective segments. Investment services rely heavily on a number of different technologies to ensure that accurate and complete information is available to investors. A high rate of innovation is required as the leading firms compete for customers through differentiation via technology. We found that investment services firms were conducting a large share of innovation in-house. The largest investment services firms even have teams referred to as Advanced Development Groups (ADGs) that are responsible for innovation and development projects within the firm. Conversely, retail banks see themselves as competing for customers through the quality of service rather than the development of and innovation within services. Retail banks believe that they have found an optimal distribution between maintaining and enhancing the level of technology facilitating the provision of retail banking service. In recent history, the trend in the provision of services has been to move from human interaction toward total automation via the Internet and ATMs. In some cases, banks began to charge customers for using bank branches and human tellers to make transactions. However, recent literature suggests that this trend is reversing. Although retail banks are maintaining the use of the Internet and ATMs for simple banking services, such as balance inquiries, deposits, and withdrawals, these banks are relying more on face-to-face interactions when providing more complex services, such as mortgages and mutual fund investments. R&D activity The NSF’s definition of R&D activities includes the following components: planned, systematic pursuit of new knowledge or understanding toward general application (basic research); acquisition of knowledge or understanding to meet a specific, recognized need (applied research); and, application of knowledge or understanding toward the production or improvement of a product, service, process, or method (development).

Financial services industry 55 Based on our interviews, we found that these concepts of R&D do not resonate well within the financial services industry. None of the firms interviewed indicated that they conduct activities in either basic or applied research. Most of the individuals interviewed could relate to the concept of development; however, traditional research, either basic or applied, was not commonplace in their operations. We found that the ideas underlying the development phase of R&D resonated best for investment services institutions. However, only the largest retail banking firms thought that they were performing development activities, and the smaller retail banks indicated they were not doing any R&D. For large investment firms, development activities mentioned in the interviews included conceptual design, articulation of technical specifications and capabilities, integration, and deployment of new technology. The research activities of large retail banks are primarily related to initial conceptual design, with only a small role in the development and deployment of the technology. Table 4.3 identifies and compares development activities performed within investment and retail service firms. Bullet marks in Table 4.3 indicate the stages in the innovation process where the financial services firm or the external technology vendor is conducting development activities. This table highlights the fundamental difference between investment services and retail banking in terms of development activities performed in house versus those performed by vendors. Large investment services firms conduct most of their development activities inhouse, relying on external technology vendors only for existing technologies that meet capability requirements and for assistance in implementing the purchased technology. Retail banking firms perform very few development activities in-house, relying on off-the-shelf technologies developed by vendors. Table 4.3 Differences in investment service and retail banking development strategies Innovation process

Investment services In-house

Generation of new technology idea Initial development Sources of generic technology Technology infrastructure Modification and implementation Operation and maintenance Source: NIST (2005).

Vendor

Retail banking Consortiums

●

In-house Vendor ●

●

● ●

●

●

●

●

●

Consortiums

●

●

●

●

●

●

56 Financial services industry Our interview information suggested that different types of financial services require different levels of technological innovation. In addition, a firm that is considered a market leader may differentiate itself from other competing firms through the continued development of new products or services. Smaller industry participants are reluctant to compete in innovation and instead find it more cost-effective to adopt off-the-shelf innovations from vendors. The innovation process The innovation process is a term used to describe the continuum of activities associated with how firms or industries develop and deploy new ideas and technologies. This process has long been discussed in the academic and policy literatures with respect to the manufacturing and industrial sectors of the US economy (e.g. research, prototype development, and scale-up to full production). However, we found the same process does not characterize innovative activities in the financial services industry. Innovation in financial services seems to be driven by customer demand, where only the larger firms participate in innovative activity and that activity is a strategic response to compete for customers, as is the case in manufacturing firms. Smaller financial services firms do innovate but generally not in terms of enhancing state-of-the-art consumer services. Rather, smaller firms innovate by providing consistency in the level of customer service.2 Larger financial services firms are conducting activities that conceptualize and develop technological advancements. Retail banking firms innovate by adopting and somewhat modifying existing technologies from vendors. Large investment firms either develop technologies in-house or purchase technologies from vendors. Large firms then significantly modify vendors’ technologies to meet their system needs. Smaller investment firms, on the other hand, could be considered imitators. They adopt technologies developed by external vendors, similar to the approach taken by smaller commercial banks, and use that modified technology to provide a differentiated product. The innovation process for financial services firms begins in-house with the following steps: ●

●

● ●

The firm identifies a new service product to meet an actual or perceived customer need. In-house development occurs to specify the attributes of the needed innovative technology that will support the new or enhanced service. The firm contracts an IT specialist to build the new technology. The firm incorporates the purchased technology into its business process.

Financial services industry 57 Retail banks conduct very little of the development activities in-house. However, based on our interviews, investment firms spend less than 20 percent of a project’s total cost on purchased technology from external vendors. The majority of the costs are incurred through in-house activities relating to the adoption and modification stages of this innovation model. The following are examples of innovation from our interviews in retail banking: ●

●

ATMs and the Internet: Retail banks incorporate Internet capabilities into retail banking ATMs by purchasing ATMs from vendors and then incorporating Internet capabilities (in-house) to allow customers to perform online activities, such as bill paying and financial transfers to third parties. The distribution of development costs for this service was 40 percent to purchased IC and 60 percent to in-house adoption and modification. Purchased IC is embedded in ATMs from vendors and the application of existing technologies related to Internet Web services. In-house activities include the large amounts of computer programming necessary to synchronize the ATM’s Internet-based transactions so that the customer’s accounts are updated automatically. Web Services (see the appendix for detailed discussion): this area involves the adoption of Web service technology to increase efficiency in intrainstitutional banking. This innovation is taking place through a collaborative cost-sharing project at the FSTC in cooperation with NEC and Stanford researchers. The distribution of development costs was 80 percent to purchased IC and 20 percent for in-house activities. In the case of Web services, only a small amount of development was done in-house. Purchased IC is embedded in technical consulting services from Niteo Partners. In-house activities include the bank’s time spent overseeing the project and offering insight into the business practices the consultants were trying to model in their Web services applications.

The following are examples of innovation from our interviews in investment services: ●

Virtual desktop: This technology allows a financial firm’s employees to log in to an institution’s internal system via the Internet to conduct business or modify documents. The innovation is that the virtual desktop will not lose data if the user’s connection is terminated. This means that an employee working off-site can log in to a firm’s internal system, work off of shared documents that are housed on a server, terminate his or her connection, reconnect, and continue working on the same document

58 Financial services industry

●

without losing any information. The distribution of development costs for this example was 5 percent to purchased IC and 95 percent for in-house development activities. Purchased IC is embedded in the computer software platform developed by a vendor. In-house activities include performing tasks such as writing code, integrating the information format from the vendor with the in-house code, testing, and implementing the technology. Virtual private networks (VPNs): Aimed at addressing the security issues associated with connecting and conducting monetary transactions over the Internet, this technology ensures that the customer is able to access all or almost all of the firm’s internal information system from a remote location as if he were accessing the system on site or from within the firm’s network firewall. The distribution of development costs for this example was 20 percent to purchased IC and 80 percent to in-house development activities. Purchased IC is embedded in networking consultancy services specializing in network security. The consultant was responsible for developing the security code specification that ensured the security of information shared between the financial institution and the customer over the VPN. In-house activities include the articulation of specifications and capabilities that the VPN needed to meet. After the consultant built the VPN, the financial institution took the system and integrated it into the institution’s existing line of service products.

R&D metrics for financial services The NSF has traditionally reported the ratio of R&D to sales as a metric to characterize the investment innovation intensity of manufacturing firms. Academics and policy makers have similarly relied on this measure, and the measure is one that aptly characterizes manufacturing’s view. From a policy perspective, maximizing the R&D intensity of firms is viewed as a positive (and meaningful) objective to achieve growth. However, our interviews indicate that this metric may be less useful for firms in the financial services sector. First, R&D is not a generally accepted descriptor of innovative investments, and second, no individual in financial services spoke of maximizing innovative investments as being associated with growth. Growth, in the traditional paradigm, comes from in-house development of proprietary knowledge that is either cost reducing or product enhancing. In fact, whereas innovation in manufacturing is often cost reducing, that concept is orthogonal to strategic planning in services. Semantically, financial services firms do not call this R&D, but compared with activities

Financial services industry 59 that occur in manufacturing, this activity is the same in nature as what many manufacturing firms call R&D. For the financial services sector, a more relevant metric proposed was the ratio of dollars spent to maintain existing technology versus the dollars spent to enhance or purchase new technology. An industry rule among larger retail banking institutions is 70 percent maintenance and 30 percent enhancement, or 2.3 to 1. If a firm moves closer to 80 percent maintenance and 20 percent enhancement, or 4 to 1, that firm would not view itself as competing successfully for customers. Small retail banks do not follow this metric because of their strong reliance on prepackaged software and off-the-shelf solutions. The 30 percent in enhancement or innovation is then optimally (i.e. in a cost-minimizing manner) allocated between in-house development activities and the purchase of IC from vendors. This rule could be considered the optimized innovation ratio for financial services firms. In investment services, an efficiency measure in technological innovation was suggested as the best way to evaluate a firm’s level of innovativeness. This metric measures technological success through a ratio of the cost of technology to the firm’s revenues. The cost of technology represents the sum of purchased intellectual property from outside the firm and in-house development activities. Holding the quality of service constant, the firm’s goal should be to minimize the ratio of its total technology-related expenditures to services. As part of meeting this goal, the firm minimizes technology costs by optimizing the ratio of in-house development activities to the purchase of IC from vendors. To illustrate the in-house development versus IC purchase decision, consider Figure 4.1. It depicts what we call an iso-technology curve, T0. The vertical axis represents purchased IC. The horizontal axis represents in-house IC. Purchased IC generally takes the form of purchased equipment and labor including human capital, and in-house IC generally takes the form of human capital. For the financial services sector firm at point A, corresponding to a given level of purchased and in-house IC, innovation by the firm could be described as an outward shift in T0 to T1, where T1 represents a new bundle of services to meet customer needs. A firm maximizes technology output by selecting the most cost-effective combination of in-house to purchased IC. As technology demands for new products and services increase, the optimal pathway for service firms may diverge from manufacturing firms because of differences in core capabilities and business models. Although we did not collect financial information from those interviewed, it was our impression that a greater percentage of IC in service firms came from purchased technology than from in-house

60

Financial services industry

PIC2 Technology pathway (Engle curve)

Purchase PIC1 intellectual capital (PIC)

• C

PIC0

T2

• B

T1

• A

T0

IHIC0

IHIC1

IHIC2

In-house intellectual capital (IHIC)

Figure 4.1 Iso-technology curve. Source: NIST (2005).

technology. And this trend is likely to increase as systems become more complicated, resulting in the curvature of the technology pathway.

Examples of R&D activity in security and commodity broker services Mintel International Group Ltd.3 suggests that the financial services industry is in the midst of redefining its business processes by developing and adopting technologies such as data warehousing and mining, customer service and support software, and client relationship management (CRM) tools. The industry’s goal is to provide either the individual or the corporate investor with easy and reliable access to real-time market investment information. These services are being provided via networks that require higher levels of security to ensure the confidentiality of monetary transactions. In addition to improving the services for customers, financial institutions are also interested in developing technology for investment tools that will enhance a financial analyst’s ability to make the most lucrative investment decision and gain competitive advantage. Industry leaders have focused on developing databases that update information in real time, thus equipping

Financial services industry 61 investors with up-to-the-minute market information and assisting investment managers in decision making. The following examples describe the types of R&D activities performed by financial services firms. Financial markets can be extremely volatile in both the short run and the long run, and without a way of providing up-to-the-minute market information, investment managers’ firms may realize large losses. To this purpose, industry leaders in investment services have turned to innovation as a strategy to improve process quality and acquire additional market share. Morgan Stanley has developed real-time international hedge funds and equity indices informed by market data via Reuters and Bloomberg. Morgan Stanley will be able to offer emerging market, regional, country, and sector equity indices in real time. Their real-time indices are said to provide a unique insight into the intraday movements of the global equity markets and thus enable clients to evaluate their portfolios’ performance versus the benchmark Morgan Stanley index at any point in the day. JP Morgan’s Investor Services product development division announced in October 2002 a strategic alliance with Investors, a leading supplier of performance measurement services for equity research. The collaboration goal is to develop an integrated research and benchmarking tool that will help institutions maximize returns through better analysis of their supplier networks and relative contribution to performance (JP Morgan, 2002). Benchmarking tools are a way for financial institutions to demonstrate the value added from their research and evaluate comparatively the performance of research analysts in-house against other institutions. Merrill Lynch decided to enhance their services by developing a superior platform for their financial advisors. In November 2002, they partnered with Thomson Financial, a unit of the Thomson Corporation, to develop a Wealth Management Workstation (WMW). The workstation will be designed to support financial advisors through the use of robust market data, news and portfolio management tools, and CRM software. In addition to product development efforts using partnerships as mentioned above, broad research consortiums are formed to conduct and coordinate generic and infrastructure R&D. Interoperability issues can be addressed through the cooperative efforts of member organizations, with individual institutions realizing gains in competitive advantage through the development of proprietary technology. For example, the Financial Services Technology Consortium (FSTC)4 is a member organization of the leading financial firms in North America. The FSTC coordinates collaborative technology research and development through pilots, proof of concept, tests, and demonstrations to develop interoperable, open-standard technologies that answer core competency needs for industry. FSTC prototypes new

62

Financial services industry

infrastructures for financial transactions, confirms new specifications for the industry, and evaluates new technologies in lab settings. The FSTC has conducted projects since 1994 in areas such as customer authentication, branch automation, check truncation, Web services, wireless banking, and biometrics. By developing these services through cooperative efforts of the organization’s members, the technology developed ensures open architectures and interoperability. The financial services industry identified the need for technological innovation in the areas of voice authentication, Web services for corporate cash management, automated intrainstitutional exchange systems, and electronic checking. The following are some examples of the type of projects performed by the consortium’s working groups. Universal value exchange (UVX) is a set of protocols that define the internal architecture, interfaces, and gateways to existing payment systems for financial institutions. Thought to be middleware, the UVX protocols will support payment processing such as paper check processing, wire transfer services, ACH, ATM/EFT, and credit card processing. Initially, the architecture is planned for transactions between banks and patrons; however, following the adoption by two or more banks, UVX is designed to conduct intrainstitutional transactions. The technology goal is to reduce the operational costs of legacy payment systems by connecting existing systems to a modern payment infrastructure using XML, state-of-the-art security technology, and current Internet protocols. The ANS X9.85 project will test the viability and performance of a prototype check validation program by designing a system that uses the ANS X9.85 standard “Specifications for Automated Identification of Security Features.” The system will be evaluated using a metric developed by financial institutions. The goal of the project is to surmise the degree of difficulty in modifying current check-processing systems to identify security features, assess the scalability and ease of integration with existing systems, and specify technical and operational barriers to implementation.

Conclusions This chapter demonstrates the diverse and abundant examples of R&D and innovation activities conducted in the financial services sector. The financial services sector has used IT to enhance its existing services and, in some cases, to even create new services. In addition, we have seen that these firms conduct R&D activities both collectively and independently in their efforts to innovate. Furthermore our interviews demonstrate that in highly competitive service industries, increasing market share is a major driver of

Financial services industry 63 innovation. Service firms use innovation as a strategy to differentiate their services in a competitive market. In relation to the service sector innovation model presented in Chapter 2 (Figure 2.3), the financial service provider has engaged in a variety of cost-sharing strategies that use purchased IC, software, or other technologies to articulate, develop, and implement innovation in the form of new or enhanced services. Because information systems are complex and tend to be custom built, much of the systems development (and hence the risk) occurs in house, although financial services firms actively pursue risk reduction by participating in consortia where interoperability issues are addressed through prototype platforms and test beds.

Appendix Diversity in the financial services sector This appendix provides additional background related to innovation in the financial services sector. Table 4A.1 lists the 10 largest financial services firms, ranked by sales. Table 4A.2 lists the publicly traded firms for which significant R&D expenditure information is available. As shown in Table 4A.2, most of the firms with significant R&D expenditures are either holding or technology investment firms and are not necessarily research-related firms conducting R&D for the provision of financial services. For example: ●

●

SEI Investments Company is a consulting firm specializing in financial management and investment technology solutions. Although the firm is oriented around the financial sector, the R&D reported seems to be emerging from the development of software applications that enable clients to make decisions concerning their investment portfolios. Anglo American PLC-ADR, with the second largest reported R&D expenditure, is a holding company, controlling the majority of shares for some of the world’s largest diamond (45 percent of DeBeers), gold (53 percent Anglogold), and platinum (50 percent Anglo American Platinum) firms. Also Anglo American PLC-ADR is one of the world’s largest independent coal miners, with interests in ferrous and base metals, industrial minerals, and forest products. Only $4 million, or 0.17 percent, of the total $2.4 billion operating profit for Anglo American was dedicated to financial services in 2001, yet they reported a $34 million investment in R&D. Given the brief firm description, it is unlikely that the amount invested in R&D is being used to innovate in the financial sector.

64

Financial services industry

Table 4A.1 Ten largest firms in the financial services industry (2003) Name

NAICS code

Employment (thousands)

Sales ($ millions)

Citigroup Inc. Prudential PLC-ADR Bank of America Corporation ING Groep NV-ADR American International Group Fannie Mae Merrill Lynch & Company Morgan Stanley Dean Witter Chase Manhattan Corporation Allstate Corporation

5223 5241 5221 5241 5241 5222 5231 5231 5221 5241

115.0 23.0 155.9 82.7 55.0 NA 67.2 55.3 74.8 52.0

82,005 51,745 51,526 43,819 40,656 36,968 34,879 33,928 33,544 26,959

Source: COMPUSTAT (2003). Note NA: not available in COMPUSTAT.

IGEN Inc. is another example of a firm whose product is not related to the financial sector. IGEN designs and manufactures diagnostic systems that aid in the mapping of the human genome. It holds the patent rights for this cutting-edge technology and leases its use to clients such as the Human Genome Project and other molecular biologists. The firm’s Standard Industrial Classification (SIC) code, 6794, classifies the firm as a patent owner and lessor. The firm’s description demonstrates that the R&D reported to COMPUSTAT is advancing medically related service industries, not the financial sector. MIPS Technologies is a design firm, classified as a patent owner and lessor, specializing in developing the low-power 34 and 64 bit core chips that are found in most video game consoles in today’s markets. In addition to gaming system chips, they offer microprocessors and architecture design systems. Instead of manufacturing these products, they simply license their intellectual property to large manufacturers of high-tech products, such as Hewlett-Packard, NEC, and Philips Semiconductors. Rambus provides chip and system firms with interface solutions to enable high performance and system bandwidth for a range of consumer, computing, and networking applications. Rambus provides its customers with interface solutions and comprehensive engineering services to support implementation of its interfaces in customer products. Example of R&D activities in select financial services sector firms These examples provide evidence of the breadth of activities that are accounted for in the financial sector. Table 4A.3 lists firms reporting R&D

0.1 0.3 0.2 0.3 0.9 0.1 0.2

5331 5331

5331 5231

5242 5331 5331

Source: COMPUSTAT (2003).

1.5 0.2 0.3

5239 5239 5331

SEI Investments Company Anglo American PLC-ADR Gemstar-TV Guide International Inc. MIPS Technologies Inc. Interdigital Communications Corporation IGEN Inc. Investment Technology Group Inc. Health Risk Management Inc. Macrovision Corporation Rambus Inc.

Employment (thousands)

NAICS code

Name

Table 4A.2 Firms with significant R&D expenditures (2003)

162 37 43

14 232

71 70

456 11,923 241

Sales ($ millions)

9 9 8

14 10

21 20

43 34 24

R&D expenses ($ millions)

Managed health care and consulting Patent manager and software design Licenses semiconductor chip connections

Technology design and patent leasing Designer of wireless telecom technologies Biodetection platform-leasing Technology-based trading firm

Investment technologies firm Holding company Global media and technology firm

Business description

66

Financial services industry

Table 4A.3 Select sample of firms reporting R&D expenditure in the financial services industry Name

NAICS Employment Sales R&D (thousands) ($ millions) expenses ($ millions)

Investment Technology Group Inc. A B Watley Group Inc. ILIFE.com Inc. Mortgage.com Inc. LendingTree.com Inc.

5231 5231 5231 5223 5223

0.3 0.1 NA 0.5 0.1

232 21 12 43 7

9.7 6.0 3.0 2.9 1.1

Source: COMPUSTAT (2003). Note NA: not available in COMPUSTAT.

expenditures related to the financial sector. For this nonrepresentative sample of firms, information on their Web sites revealed the following innovation-related information. Investment Technology Group Inc. is a full-service trade execution firm that uses technology to increase the effectiveness and lower the cost of trading by emphasizing R&D in the products they offer, which are as follows: POSIT: an electronic stock crossing system; QuantEX: a Unix-based decisionsupport, trade management, and order routing system; SmartServers: serverbased implementation of trading strategies; Electronic Trading Desk: an agency-only trading desk offering clients the ability to efficiently access multiple sources of liquidity; ITG Platform: a PC-based order routing and trade management system; ITG ACE and TCA: a set of pre- and posttrade tools for systematically estimating and measuring transaction costs; ITG/ Opt: a computer-based equity portfolio selection system; ITG WebAccess: a browser-based order routing tool; and Research: research, development, sales, and consulting services.5 A.B. Watley Group Inc. is a New York-registered broker-dealer that operates both direct-access trading and third market institutional sales trading brokerage businesses. The firm offers a proprietary technology called Direct-Access Vertical Exchange (DAVE) to brokerage and banking industries. DAVE consists of a ticker plant, order entry and trade processing, and data delivery engines. LendingTree.com Inc. is a lending exchange that attracts customers looking for loans and processes loan requests through a number of banks. LendingTree licenses their technology platform LEND-X(SM) (which powers their Internet-based lending exchange) to other businesses to create exchanges on their own Web sites.

Financial services industry 67 Web services case study Consortia such as the FSTC serve an important role in technology development activities, and they contribute significantly to applied research being performed to support the financial services industry. The following is a case study of the proof-of-concept project performed by a group of retail banks in cooperation with NEC and Stanford University through the FSTC.6 We found little evidence that retail banks were involved in applied research other than through funding of consortium activities. Our interviews indicate that NEC performed all basic and applied research and that only a small percentage of the total project cost was spent by banks for in-house development activities. The FSTC recently completed the proof-of-concept project related to Web services. The project’s goals were to promote shared learning and develop technologies related to Web services for identification, aggregation, and composition of corporate account data and services. The Web services project was cosponsored by NEC’s Niteo Partners and three retail banking firms— Wachovia, Bank of America, and JP Morgan Chase. Niteo Partners is a wholly owned subsidiary of NEC (NEC has worked on innovation in technology for over 100 years). NEC has a large investment in technology research labs around the world. These laboratories work on issues related to Internet software, nanocomputing, quantum cryptography, and other networking-related technologies. In addition to applied technology research, NEC also invests in academic or basic research. NEC has worked closely with a multidisciplinary research faculty at Stanford University. Areas of research include knowledge representation, machine-to-machine communication and interaction, and automated computing. The NEC invested $2 to $3 million for academic research in machine-tomachine interactions and wanted to set up a project that would advance awareness of its technology. NEC was looking for market exposure in the financial services sector specifically. NEC called on its wholly owned subsidiary, Niteo Partners, to design or craft a project in conjunction with the Stanford research team that would demonstrate the machine-to-machine interaction technology in combination with existing standards related to Web services to reduce the cost of intrabank interactions. The goals of this project were to gain market exposure for NEC technology related to Web services and to demonstrate what Web services have the potential to provide in the future for the financial sector. When Niteo approached the FSTC, they discovered that the financial services sector was just beginning to understand the current capabilities and applications for Web services. Niteo spent 6 months educating industry

68

Financial services industry

participants on the state of standards related to Web service applications. Niteo then created case examples to demonstrate the market potential for Web services in the financial services sector. Once the returns on investment were visible, retail banking institutions entered into a cost-share project with Niteo through FSTC. This project allowed industry experts to share knowledge of the business and specify the special needs that the technology needed to address. Niteo used this input to develop reference materials that could be implemented in-house by the participating banks. The latter project was cofunded by NEC and the FSTC participants, including Wachovia, Bank of America, and JP Morgan Chase. These firms agreed to fund the proof of concept jointly in return for the rights to core findings and any intellectual property developed. Niteo ran the project. The knowledge and learning occurred at Niteo but was informed by the business expertise of the banking executives from the participating banks. The project has moved into a second phase concerned with implementing the findings from Phase 1. Phase 1 lasted approximately 2 years. As part of the implementation phase of the project, participating banks have the opportunity to become more involved in developing and deploying the technology. At this point JP Morgan Chase is the only one of the three banks that is proceeding with implementation. They are funding the implementation of the basic standardized codes developed as part of Phase 1 and are creating a test bed to work through the remaining technological issues. Most of these activities will be conducted in-house, with NEC working as a consultant. Insights gained in Phase 2 will be IC owned by JP Morgan Chase and will not be shared with the other banks that participated in the first phase of the project.

5

Systems integration services industry

Introduction Systems integration services facilitate the intersection of hardware, software, and pragmatic industry knowledge that provides the foundation of IT systems. The Computer Science Telecommunications Board (CSTB, 2000) defines systems integration as the wiring together, via hardware and frequently very complex software, of the often already existing islands of computer applications into a coordinated enterprise-wide distributed network system. Systems integration includes more than just physically allowing incompatible components to communicate. It is the synthesis of application domains such as finance, manufacturing, transportation, and retail and the supporting information infrastructure including databases, operating systems, architectures, networks, communications devices, and security measures (CSTB, 2000).

Systems integration industry Since 1992, systems integration has been the most rapidly growing component of the US computer industry. Total revenues for custom integrated system design and custom programming services rose from $34 billion in 1990 to $76 billion in 1997 (CSTS, 2000). Table 5.1 illustrates the diversity of companies providing systems integration services and highlights the variance in their R&D reporting. Service-only firms are generally less likely to report their systems integration activities as R&D expenditures. In contrast, diversified firms such as IBM and Hewlett Packard classify one-quarter to one-half of systems integration activities as R&D; thus, the level of reported R&D spending by diversified service firms is significantly greater than that of service-only firms. This variance in reported R&D for systems integration is likely due to several factors. Larger systems integration firms differentiate themselves

70

Systems integration services industry

Table 5.1 R&D investments of representative systems integrators (1998) Company Services-only firms Accenture Consulting American Management Systems Computer Sciences Corporation Electronic Data Services Keane Diversified firms IBM Hewlett Packard Lockheed Martin

Systems integration revenues ($ millions)

Percentage systems integration R&D (%)

8,307 1,058 7,660 16,891 1,076

0 77 0 0 3.5

28,916 6,956 5,212

25 50 36

Source: CSTB (2000).

through the IC and assets they offer in conjunction with traditional integration services. This can be in the form of detailed knowledge of the hardware and software products these firms provide or through specific industry knowledge of selected sectors’ IT needs. In these instances, companies introduce products and services with IC developed through strategic R&D initiatives. In contrast, smaller service-only systems integration firms are more likely to solve their problems in the field through on-site modification of software programs and development of work-arounds. From this perspective they have no “R&D staff,” and any “research” they might conduct is considered an integral part of the integration service being provided.

Systems integration services The market for systems integration services, although rapidly growing, is still evolving. This evolution is largely based on how data is managed and communicated internally between different areas’ business operations and externally between other organizations within a supply chain. As supply chains increasingly become virtually vertically integrated, the level of detail and the quantity of data a firm is required to manage are increasing exponentially. Today’s systems integration services help client organizations manage data and provide innovative ways to exploit the knowledge nested in the large amounts of data that exist both inside and outside of a firm’s business operations. To this end, systems integration services have been evolving over time from the simple task of wiring (a term used literally and figuratively in the industry) two or more systems together to enabling

Systems integration services industry 71 information transfer across different firms, platforms, and standards protocols. As this highly competitive industry grows, integration services firms are forced to seek out ways to differentiate their services. In the 1980s and 1990s, almost all systems integration activities involved computer programming labor services. Customizing software programs, translation algorithms, and other middleware were commonly needed to interconnect legacy information infrastructures to more modern business software applications. Today’s systems integration firms rely less on a set of software programs that were traditionally developed as a unique solution for each project and more on developing and leveraging capital and intellectual assets that can provide a competitive advantage. The latter includes developing either generic technology solutions that require relatively minor adaptations to individual customer needs or software that is sufficiently flexible to allow the customer to make adaptations. The trend is for an increasing share of systems integration services to be commoditized. For example, software vendors have begun to specialize in a specific industry or business process software. Over time, the high costs of setting up custom-built systems has motivated the software industry to identify common needs across numerous clients and collapse the commonalities into products that are closer to the concept of shrink-wrapped software. These packaged applications can be configured with relatively little additional effort to meet the needs of many firms within a single industry. As the components of systems integration services become commoditized, leaders in systems integration are differentiating their services and are no longer simply linking several systems in a business operation through computer programming. Today, the service provider differentiates its services through asset innovation, leveraging its internal assets to assist in developing solutions for a client. For example, many firms’ strategic plans include accumulating and maintaining industry-specific knowledge in business operation optimization, and that knowledge is sold to industry as a service. As a result, a distinction is emerging between traditional programmer-based service systems integration firms and more diversified firms that also have detailed scientific or industry-specific knowledge that can be bundled with their systems integration services.

Innovation in the systems integration industry As the foundation for this case study, we interviewed a number of firms involved in systems integrations, as shown in Table 5.2. Many of these firms are highly diversified, such as IBM, in both the broadly defined computer

72

Systems integration services industry

Table 5.2 Systems integration services firms interviewed Name

Position

Carolina Advanced Digital Inc. Computer Science Corporation

President and Chief Scientist Practice Manager of IT Strategy and Architecture Inside Account Manager Director of Systems Integration Chief Information Officer Founder and President Business Development Systems Integration Project Manager Vice President of Assets Innovation Project Manager

Computer Service Partners Concurrent Technologies Corporation EMC Corporation Infosystems Technology Inc. (ITI) Integrian Inc. IBM IBM PeopleSoft Source: NIST (2005).

Enabling knowledge (outside firm)

Market research Consumer preference Technology Status Regulations

·· ·

Physical dimension

R&D knowledge (in-house)

Strategic planning

standards · Industry standards · Interface and license · Patent R&D (public) · Basic embedded in · R&D components · Imitation

Product and service development

Labor to provide service

Service provision

Value added received by customer

of quality integration of · Systems · Development systems management physical components and operation ·· Customization Enhance tools and process customer · Enhanced service/billing systems Testing methodology · R&D lab work Labor (human capital) · ·

Figure 5.1 Illustration of the information flows for systems integration activities. Source: NIST (2005).

industry and broadly defined software industry. Respondents to our interviews were asked to describe the processes by which their firms develop the IC underlying their products and services as a first step toward our understanding of innovation. Figure 5.1 represents our understanding of the information flows that characterize innovation for the activities that are related to systems integration for a generic firm. We also discussed with these firms industry trends in general, and much of what we learned through this broader dialog is discussed below. When references are made to noninterviewed firms, the related information was

Systems integration services industry 73 learned either from broad discussions as part of the interview process or from Web sites. Information flows In Figure 5.1, the center horizontal row represents an average product life cycle and the different stages associated with it. The downward pointing arrows in the upper portion of the figure represent information flows into the firm’s product development from outside the firm and/or industry. External information is relevant for the firm’s strategic planning as well as for its product and service development. The upward pointing arrows in the lower portion of the figure represent knowledge flows resulting from in-house R&D. Such knowledge is fundamental to the firm’s product and service development process as well as to its delivery of services. In the highly competitive systems integration industry, induced (through in-house R&D) knowledge is considered a trade secret. As a result, each firm in the industry could potentially be required to reinvent existing technology already developed by a competing firm to maintain a competitive edge in the market. R&D activities Systems integration firms devote a significant resource to developing and enhancing the skills and knowledge required to provide their services. Many of these innovative activities occur in real time as part of software customization or troubleshooting system problems. However, it is questionable as to whether these innovation-enhancing activities meet the definition of R&D. Table 5.3 separates firms into service-only firms and diversified firms; these two classes of systems integration services firms are discussed in more detail below. Based on CSTB (2000) information, the information in the table illustrates the diversity of firms providing systems integration services in two dimensions—the variance in size, measured in revenues and in terms of the variance in the percentage of systems integration activities that the firm categorizes as R&D. Service-only firms are generally less likely to report their systems integration activities as R&D expenditures. In contrast, diversified firms attribute a much larger share of their budgets to R&D activities, in some cases upwards of 50 percent of their systems integration activities. Thus, the level of reported R&D spending by diversified service firms is significantly greater than that of service-only firms, and this fact may or may not reflect differences in innovative activity among the firms. This variance in reported R&D for systems integration is likely due to several factors. Larger systems integration firms differentiate themselves

74

Systems integration services industry

Table 5.3 R&D investments of representative systems integrators, by class of systems (1998) Name Services-only firms Accenture Consulting American Management Systems Computer Sciences Corporation Electronic Data Services Keane Diversified firms IBM Hewlett Packard Lockheed Martin

Systems integration revenues ($ millions)

Percentage systems integration R&D (%)

8,307 1,058 7,660 16,891 1,076

0 77 0 0 3.5

28,916 6,956 5,212

25 50 36

Source: CSTB (2000).

through IC and assets they offer in conjunction with traditional integration services. This can be in the form of detailed knowledge of the hardware and software products these firms provide or through specific industry knowledge of selected sectors’ IT needs. In these instances, firms introduce products and services with IC developed through strategic R&D initiatives. For example, IBM conducts research to ensure that its hardware and software will interoperate with legacy and competing systems and then promotes this research as part of a strategy to market its systems integration services. In this way it leverages its technical expertise related to its proprietary hardware and software technology. In contrast, smaller service-only systems integration firms are more likely to solve their problems in the field through on-site modification of software programs and development of work-arounds. From this perspective, they have no R&D staff, and any research they conduct is considered an integral part of the integration service being provided. In addition, it is likely that their customized integration services have less potential for reuse (although lessons learned build IC and capabilities of service-only firms). Larger firms follow a model of innovation moving from concept stage, where a potentially marketable solution is developed, to proof-of-concept stage, where risk is evaluated in terms of marketability, and finally to pilot project, where risk assessment is continued before moving into mass production. Smaller firms are more likely to implement existing systems integration technology, which involves some development/engineering of a commodity service integration approach, but minimal activities related to seeking out novel approaches to integration.

Systems integration services industry 75 Categories of R&D Systems integration firms generally group research into four categories, some of which match NSF’s definition of R&D by character of use (basic/applied/development): ●

●

●

●

Building new systems for clients: Activities for this category resemble software design and the synthesis of hardware and interface applications. Products developed in this category are one of a kind. Maintenance of code: Rewriting code is necessary to maintain the system’s integrity with respect to its operational environment. Support: (Applied research) researching better processes for server maintenance and monitoring and system tracking for troubleshooting. Break/fix maintenance: (Applied research) research consists of designing better processes and establishing generic protocols across all products.

Larger firms, such as IBM and Computer Science Corporation, reported that they were undertaking at least some basic research. These firms maintain large facilities dedicated to noncommercial research programs. Smaller firms consistently reported zero activity in basic research. Respondents explained that the smaller business model for systems integration was focused primarily in applied research with respect to the overall objective of completing the project. Respondents from the larger firms identified applied research as the area in which their division spent the largest share of time, whereas respondents from smaller firms attributed only a small portion of their activities to such research. Identifying practical applications for emerging technology was the primary activity that respondents associated with this category of research. Development activities were reported as the second largest share of a systems integrator’s time from large firms and the largest share by smaller firms.1 The types of activities respondents associated with this category of research were writing code, setting up infrastructure, wiring, and putting the system together. While building systems was a major activity for the larger firms, it was not an activity mentioned by the smaller firms. Over time, the high costs of setting up custom-built systems have motivated the computer services industry to identify common needs across numerous clients and collapse the commonalities into a shrink-wrapped product. Labor versus intellectual capital Figure 5.2 highlights the spectrum of systems integration services and the shift between labor and IC as inputs to providing systems integration

76

Systems integration services industry

100%

Service only

Diversified

Intellectual assets % of service inputs

Labor 0% 100%

R&D as a % of integration services

0%

Development

Applied research Basic research

Figure 5.2 Production spectrum of systems integration services firms. Source: NIST (2005).

services. On the left-hand side of the figure, the service-only firm relies heavily on labor (computer programmers and some managerial workers) to implement a standard set of activities that integrate multiple systems. As one moves from left to right in the figure, the share of the labor versus IC that is used in providing a systems integration service increases. IC in this case consists of input from skilled scientists (possibly involved in hardware or software development) and industry specialists, in addition to proprietary technologies such as algorithms and other patentable or licensed technology. R&D intensity increases as intellectual assets become an increasing share of providing integration services. The lower portion of the figure, R&D intensity (R&D as a percentage of integration services), increases as the firm becomes more diversified in the scope of its innovative activities. In general,

Systems integration services industry 77 systems integrators, especially service-only integrators, are most likely to engage in development, working to enhance or create new service offerings. As a firm continues across the spectrum and leveraging related technologies become more important, the firm may conduct applied research to incorporate these technologies into its service offerings. In the extreme, larger firms may conduct basic research to support systems integration, such as involvement in standards and protocol development, as well as basic research into business organization and communications.

Service-only versus diversified firms’ R&D activities As discussed earlier, there are two general types of systems integration firms: service-only and diversified service and product firms. Service-only systems integration firms primarily provide services in computer programming and reconfiguration of packaged software, leveraging experience gained over time from previous, often similar, projects. Diversified systems integration firms provide hardware and software products as well as integration services that can be bundled. In addition, diversified systems integration firms are typically much larger service firms that acquire and use a pool of technology and intellectual assets that the firms can leverage to differentiate their services from those of competitors. R&D activity in service-only integrators From our interviews, service-only firms responded that they engaged in few activities that would be considered R&D. Service-only firms commonly build on business software applications using open architecture that allows the integrator to easily configure prepackaged applications purchased from a vendor to meet the needs of the client. Using XML and other open-source networking languages, the integrator brings together various components of a client’s business operation. However, those interviewed in the service-only firms were concerned with protecting and managing their IC, which is largely the ability to engineer commodity software systems. Service-only firms interviewed reported that they were managing the IC developed from years of experience on different projects. Larger service-only firms often create executive positions known as Chief Knowledge Officers to manage and coordinate the IC developed by the firm. Larger firms reported developing formal methodologies for enterprise architecture, product life-cycle management (PLM), enterprise resource

78 Systems integration services industry management (ERM), and managing operations centers. These methodologies are documented through white papers and in-house presentations at firm research conferences. Service-only firms generally grouped their methodologies and development activities into four categories: ●

●

●

●

Building new systems for a client: Including software design and systems engineering (synthesis of hardware and interface applications). Products developed in this category are one of a kind. Systems configuration: Turning certain options within purchased software applications off or on to maintain the system’s integrity as products/ systems evolve. Support research: Researching better processes for server maintenance and monitoring and system tracking for troubleshooting (one firm classified this as applied research). Break/fix maintenance: Designing better processes and establishing generic protocols across all products.

In addition, service-only firms also perform scoping exercises to determine the capabilities of future technologies. As part of these exercises, they research the technologies that will be most influential over the next 30 years by scanning the scientific literature and technology news sources. This information is then used to conceptualize how the integrator can create new services or incorporate the capabilities into existing services. One of the industry’s leading firms interviewed reported that 24 months is a minimum required lead time that an integration firm must have on the emerging technology markets if the firm wants to remain competitive. The optimal lead time is 4 to 5 years. However, this amount of lead time is generally only possible if the integration firm is taking part in the technology development process. Service-only integrators rarely have access to emerging technology at its earliest stages of development. R&D activities in diversified integrators In contrast, diversified integrators not only have access to the IC accumulated through experience on previous integration projects, but they also have access to the technology being developed in the design and manufacturing division of their firm and to the researchers leading these activities. IBM Global Services report that they are working with their firm’s computing equipment experts to conceptualize how existing equipment or capabilities can generate new service offerings. IBM is creating new services by moving

Systems integration services industry 79 away from the labor exercise that integration has become and attempting to leverage internal and external IC to innovate its integration services. One example of leveraging IBM equipment technology to innovate integration services is Global Services’ new offering called WebFountain. This new service offers clients the ability to harness the power of a super computer to scour the Internet for all the published or online information on a given topic. The service is composed of five servers with over a petabyte of data storage programmed with unique algorithms designed to seek out Web content on specific topics. The amount of data the technology can gather allows the client to examine trends in consumer preference or identify cultural norm differences across society. In its traditional form, this type of research would be very expensive and require several years of research. WebFountain reduces the costs of research in product design and increases revenue by shortening the time required to bring a product to market. Integration of systems in a business operation setting requires understanding the role that information plays in the business. Global Services is taking this concept one step further to better understand the nature of work within different industries. IBM has anthropologists who specialize in researching the way professionals in different industries use information to maintain large business operations systems. IBM considers their research in this area as a basic research exercise. Although the study of human behavior is a social science area, the end use is to increase the effectiveness of an integration service. Global Services has also used its engineering staff, which normally specializes in designing super computers and deep computing technology, to develop a cryptographic scheme to code electronic signatures for the French land titling system. The French government wanted to modernize their real estate titling system, converting all hard copies to digital online documents. However, the French government required a guarantee that no technology would emerge that could make the system vulnerable to computer attacks for the next 25 to 30 years. IBM was hired to create a system that met the French government’s requirements. The system required a combination of hash and cryptographic schemes. As part of this project, IBM Global Services leveraged its deep computing R&D division to develop a scheme that met the government’s requirements.

Conclusions Systems integration services will become more important as industries expand their reliance on technologies and systems. The commoditization of basic integration services (i.e. the act of connecting hardware, software, and

80 Systems integration services industry communication equipment) has led to a divergence in the strategies and approaches integrators are offering. This has led to segmentation of the marketplace into applied and diversified service firms. Applied service firms do not perform R&D and innovate only in ad hoc ways to solve issues as they arise. Pure service firms also monitor trends and emerging technologies, creating robust good practice guidelines that apply existing methodologies of integration to new or evolving environments. In contrast, diversified firms have shown high levels of innovation, leveraging computer equipment and manufacturing IC, as well as actively accumulating expertise in process optimization for specific industries. Innovation is serving to differentiate the diversified firm’s service offerings. All this ultimately leads to enhanced value or the creation of new services. This characterization of integration services complements the model of innovation for service-sector firms presented in Chapter 2 (Figure 2.3). Strategy and competition drive both applied and diversified systems integration firms. These service firms rely heavily on information from suppliers and customers to assess technology trends and consumer preferences to identify potential opportunities for enhanced value of services. Entrepreneurial activities focus on adapting existing or emerging technologies and practices to meet their consumers’ needs.

6

Research, development, and testing service industry

Introduction The expansion in service-sector R&D is in large part due to an increasing dependence by large firms on outsourcing as the vehicle for accomplishing innovation in products and production (Amable and Palombarini, 1998; Howells, 1999; Jankowski, 2001; Pilat, 2001). Outsourcing is a common approach for conducting research in areas of interest that require expertise outside of a firm’s core competencies. As of 2000, engineering and scientific contract research accounted for between 5 and 12 percent of industrial R&D in most industrial economies (OECD, 2001b). There is no general agreement within the literature on a common set of factors that motivate firms to contract out scientific R&D. In general, research, development, and testing (RD&T) services commonly outsourced have traditionally been viewed as formal, routine, repetitive, and cost-based with short time horizons (Andersen et al., 2000). However, this appears to be changing as RD&T firms are establishing long-term partnerships with client industries and increasingly providing core research functions to support strategic initiatives (Howells, 2000b). This chapter begins with an overview of the broader RD&T service sector and its R&D activities. This is followed by a more in-depth analysis of R&D activities and the innovation process in the biopharmaceutical branch of the biotechnology industry, which is one of the largest and fastest growing segments of the RD&T service sector.

RD&T industry profile and R&D statistics The RD&T industry (NAICS 5417, also referred to as Scientific R&D Services) performs R&D activities in the fields of physical engineering, life sciences, and social sciences and humanities. Most firms classified as RD&T perform R&D by contract or fee for either manufacturing or nonmanufacturing industries.1

82

Research, development, and testing service industry

Table 6.1 Largest RD&T firms, by R&D expenditures Name

Employment sales

R&D Business description expenses (Thousands) ($ millions) ($ millions)

Exelixis Inc.

0.57

41

89

Curagen Corporation Maxygen Inc.

0.51

23

66

0.31

30

64

Genaissance Pharmaceuticals Deltagen Inc.

0.16

5

46

0.30

10

45

Symyx Technologies Inc. Microvision Inc.

0.20

60

39

0.23

11

33

Paradigm Genetics Inc.

0.25

24

28

Arena Pharmaceuticals Inc. Exact Sciences Corporation

0.21

18

23

0.07

0.05

14

Genomics-based pharmaceuticals firm Genomics-based pharmaceuticals firm Genomics-based pharmaceutical and agriculture Genomics-based pharmaceuticals firm Genomics-based pharmaceuticals firm Genomics-based pharmaceuticals firm Optical scanning systems Life sciences in agriculture and human health Biopharmaceutical firm Applied genomics firm

Source: COMPUSTAT (2003).

Table 6.1 lists some of the attributes associated with R&D-intensive RD&T firms. In general, these firms are of modest size with fewer than 5,000 employees. Most of these firms are engaged in some form of biotechnology research.2 As with many pharmaceutical firms that have yet to introduce a product to market, R&D expenditures exceed annual revenue for all but one (Symyx Technologies) of the firms listed. Because of the predominance of RD&T firms engaged in biotechnology research (including genetic bioscience research, drug discovery, and pharmaceutical testing, for example), the remainder of this chapter focuses on the biotechnology-based biopharmaceutical industries and their relationship with RD&T firms (outsourcing of R&D).

Biotechnology The term “biotechnology industry” is widely used by public policy makers as well as the popular press. However, no group of homogeneous firms or

Research, development, and testing service industry 83 organizations (or NAICS code) clearly defines the biotechnology industry (Toole, 2003) because of the rapidly evolving nature of the industry and the diverse set of technologies used to develop applications from bioscience research.3 The biotechnology industry crosses over many disciplines from agriculture and the environment to health care and industrial applications. However, individual biotech firms typically specialize in and develop core competencies in specific genes and treatments that increase their probability of success. In general, the biotechnology industry is not identified by a set of common products that are produced by similar techniques, but rather by its applications. Prevezer (1998) defines the biotechnology industry in terms of the following applications: therapeutics (drugs), diagnostic applications, chemicals (pesticides, insecticides, and new chemicals), agricultural (seed, plant, and animal applications), food and cosmetics, environmental, and energy (biomass). By all definitions, the biotechnology industry has experienced rapid growth in the past two decades. October 2002 marked the 20th anniversary of the first Food and Drug Administration (FDA)-approved biotechnology drug (Biotechnology Industry Organization, 2002). Today, there are 141 biotech-based medicines and vaccines on the market in an industry valued at Table 6.2 Largest biotechnology firms, by R&D expenditures Name

Estimated 2002 R&D ($ millions)

Percentage change from 2001 (%)

R&D as percentage of 2002 revenue (%)

Amgen Genentech Millenium Pharmaceuticals Biogen Chiron Immunex Genzyme Corp. Vertex Pharmaceuticals Human Genome Sciences Incyte Genomics ICOS Inhale Therapeutic Abgenix Gilead Sciences Celera Genomicsa

967.1 613.1 510.0

11.8 16.5 27.3

19.4 24.0 123.4

365.7 361.0 258.6 238.5 212.0 205.8

16.2 4.8 26.3 27.2 55.1 40.7

30.9 27.0 20.0 20.1 113.9 954.2

183.0 173.0 160.0 155.0 145.0 142.4

10.0 54.7 14.6 54.4 21.9 31.4

116.6 211.0 190.5 328.7 36.2 103.4

Source: Standards & Poor’s Biotechnology Industry Survey, May 2002. Note a Celera was a major contributor to the Human Genome Project and completed a draft of the human genome in June 2000.

84

Research, development, and testing service industry

$198 billion. The biotech industry, which includes 1,457 firms and employs 191,000 people in the United States, spent $15.7 billion on R&D activities in 2001 (Ernst and Young, 2002). The US Department of Commerce (DOC) surveyed 1,031 firms conducting R&D in biotechnology-related applications. A large majority were conducting R&D activities in the area of human health in support of the pharmaceutical and the medical device industries (DOC, 2003). A large number of these biotechnology firms are relatively young, independent firms (established after 1990), employing fewer than 100 researchers. The DOC report states that most small firms conducting human health research focused on applied research and larger firms emphasized product and process development. Table 6.2 reports the largest biotechnology firms in terms of R&D expenditure in 2002. These firms are increasingly outsourcing R&D activities to RD&T service firms.

Biopharmaceutical industry A larger, more encompassing industry description used in professional literature and policy discussions is the term “biopharmaceutical industry.” Biotechnology firms commonly enter into commercialization agreements with large pharmaceutical firms. Large pharmaceutical firms finance the R&D for biotechnology firms with potential drug targets in exchange for exclusive distributing rights and a royalty fee should the drug survive clinical trials and gain FDA approval. The term “biopharmaceutical industry” captures all major participants in the drug development supply chain, including identification and discovery, pre- and post-clinical trials, and production activities. In addition to large biotechnology pharmaceutical firms, this includes smaller RD&T firms, contract research organizations (CROs), and contract manufacturing organizations (CMOs). The pharmaceutical industry is rapidly expanding with the advent of biotechnology-related drugs and gene therapies otherwise known as biopharmaceuticals, in addition to efficiency improvements in technologies that analyze potential drugs. As the “new drug” market expands, many firms are focusing their internal resources on drug discovery. As a result, the contract research industry has experienced rapid growth in the past two decades and now is extremely diversified in the variety of services and tools offered. The tools and services vary dramatically across the spectrum of R&D activities required to discover and develop a new drug. Figure 6.1 shows the four major contributing groups in the supply chain involved in the drug development, identification and discovery, pre- and postclinical trials, and production activities. These include large biopharmaceuticals

Research, development, and testing service industry 85 Product development

Process development

Small RD&T firms

Large biopharmaceuticals

CROs

CMOs

Figure 6.1 Drug development supply chain. Source: NIST (2005).

firms, smaller RD&T firms, CROs, and CMOs. RD&T firms and CROs are mostly involved in the product development phase of the supply chain and are typically classified as service providers. CMOs support the process side of drug development. This chapter discusses the drug and therapies development supply chain, focusing on the roles the RD&T firms and CROs play in the development process.

The drug development process The RD&T biotechnology firms are integrally involved in drug development. Development of a new drug costs on average $800 million. A large share of the total R&D dollars from these industries pays for outsourcing development services and process and manufacturing services. Drug discovery and development is a well-documented process that includes several stages, starting with discovery of a potential targeted compound or therapy, followed by FDA-regulated preclinical and clinical trials, and ending with the successful marketing of a new drug or therapy. Assuming the newly discovered compound or therapy clears the FDA approval process, the total time from discovery to market can be 10 to 15 years. This process is both costly and time consuming. The innovation process for biotechnology-related drug treatment products and processes is closely aligned with that of the pharmaceutical industry’s drug development model. Figure 6.2 illustrates the innovation process in biopharmaceuticals over several phases of R&D. The model identifies phases

86

Research, development, and testing service industry Basic research Physical and biological sciences R&D

Information technology R&D

Equipment and manufacturing R&D

Tools and processes to aid discovery in biotechnology Genomics Proteomics Physisomics Molecular modeling Cell culture technology Molecular cloning Recombinant DNA Protein engineering Tissue engineering Nanobiotechnology Bioinformatics Automated analyzer Biosensor Microarry

Applied R&D activities

Biotechnology industry: pharmacogenomics

Discovery Target development Preclinical development

Process research for target (1 gram– 1 kg) Explore alternative synthetic routes

Biological tests Analytical characterization of molecules Animal screening Pharmacological and pharmokinetic studies Reproduction Mutagenicity tests Design model

Development R&D activities

Evaluate production at lab level Produce small amounts of drug for clinical trials

Clinical trials PhaseI: Safety Healthy volunteers Maximum tolerable dose Side effects Phase II: Efficacy Afflicted patients Bioavailability of different formulations and doses • Phase III: Efficacy Large-scale multisite trials Proof of safety and efficacy in long-term use Comparative studies

Pilot development

(1–100 kg) Evaluation of process in pilot plants. Kinetic studies Optimization of reaction conditions Supply materials for clinical trials Design of equipment/facility

Commercial plant transfer/startup (100 kg– metric tons) Transfer of process to commercial plant

Approval Documentation of clinical trial data Expert opinion on data Documentation and validation of process technology Final preparation and submission of NDA

Optimization of processs under commercial conditions Validation of manufacturing process

Marketing and production

Figure 6.2 The innovation process in biopharmaceuticals over several phases of R&D.

required to bring a drug to market in the United States. The arrows indicate the direction for the flow of information. Basic research originating in the information technology, manufacturing, and physical and biological sciences sectors is synthesized by the biotechnology sector and applied to identify possible drug targets. Tools and processes used in the discovery include genomics, proteomics, molecular cloning, recombinant DNA, protein engineering, bioinformatics, automated analyzer, and microarray technology. In most cases, biotechnology firms, like large pharmaceutical firms, are scanning millions of possible compounds or genetic combinations for any

Research, development, and testing service industry 87 variation that may improve human health in some measurable way. Much of this applied research is conducted by small RD&T firms and then transferred (through purchasing or licensing agreements) to larger pharmaceutical firms. Once a compound is discovered, the biopharmaceutical firm must decide where the drug has the greatest potential for market success. Firms will often contract with large CROs that have knowledge of various disease markets in the hopes of tailoring their R&D efforts toward a disease with the highest probability of success both in clinical trials and marketability. Following discovery of a new drug or treatment, the development diverges on two tracks: product and process. Product development On the product side, the drug firm is required to demonstrate a drug’s efficacy and safety to FDA, evaluated through several years of clinical trials. Stages in the product development process include preclinical testing; clinical trial Phases I, II, and III; FDA review and approval; and Phase IV postmarket testing. Each stage takes at least 1 year or more to complete. Figure 6.3 outlines the various stages in the product development process. The preclinical testing stage indicates likely safety and efficacy attributes in living organisms through animal testing and toxicology studies. These studies show biological activity of a compound against a targeted disease. Additionally, risk assessment and market viability are established at this stage as well. The type of information that comes out of these studies includes detailed descriptions of the new drug’s chemical structure, how it works in the body, any toxic effects, and how the drug will be manufactured. These sets of information are submitted to FDA in an Investigational New Drug (IND) application. The IND specifies the how, where, and by whom the clinical trials will be conducted. The remaining compounds enter into clinical trials. Over 50 percent of the product development effort is spent in clinical trials. Over the course of 6 years, the drug is tested on real human patients to evaluate the drug’s safety and dosage range. Phase I of clinical trials generally takes about 1 year to complete. These studies enroll about 20 to 80 normal, healthy volunteers to determine how the drug is absorbed, distributed, metabolized, and excreted and the duration of its effects. Phase II uses volunteers that suffer from the targeted disease to study the drug’s effectiveness. Phase II usually lasts 2 years. Finally Phase III is a nationwide study enrolling 1,000 to 3,000 real patients in clinics and hospitals around the country. Phase III requires physicians to monitor patients to determine efficacy and adverse effects. Upon successful completion of clinical trials, trial data, expert opinion data, and validation of process manufacturing data are compiled and analyzed.

20–80 healthy volunteers

Laboratory and animal studies

Assess safety and biological activity

5,000 compounds evaluated

Test Population

Purpose

Success rate

Figure 6.3 The stages in drug development.

Source: NIST (2005).

Evaluate effectiveness, look for side effects

100–300 patient volunteers

2

Clinical trials Phase II

5 compounds enter trials

Determine safety and dosage

1

3.5

Years

File IND at FDA

Phase I

Preclinical testing

Verify effectiveness, monitor adverse reactions from long-term use

1,000–3,000 patient volunteers

3

Phase III

File NDA at FDA

1 compound approved

Review process and approval

2.5

FDA

Additional postmarketing testing required by FDA

Phase IV

Research, development, and testing service industry 89 A New Drug Application (NDA) submitted to FDA includes the final analysis and all the data gathered since the discovery of the new drug compound. The NDA regulatory submissions are typically over 100,000 pages and require an average review period by FDA of 30 months. Following the approval and marketing of the new drug, the drug firm is still required to periodically report any adverse patient reactions and conduct (Phase IV) long-term effect studies on existing patients. Additionally some drugs are discovered to have additional positive effects that were not specified in the IND submission. One example is Welbutrine; this drug was originally approved as an antidepressant but has been found to mitigate the withdrawal symptoms associated with quitting smoking. To market the existing drug for a new purpose, the drug firm must go through the entire FDA approval process again. Process development Process development begins even before preclinical testing. Newly discovered compounds must be manufactured in small amounts and the process for manufacturing must be determined in the IND regulatory submission to FDA. Depending on the size of the drug firm, the firm may or may not have the resources and knowledge to efficiently evaluate how best to produce the newly discovered compound. Drug firms with limited resources, especially small biopharmaceutical firms, will outsource the drug manufacturing to a CMO that specializes in chemical or biological manufacturing. Although the CMO provides a service, these firms are generally classified under the industry classification medicinal manufacturing. The CMOs, like the traditional pharmaceutical manufacturing firm, perform research to determine the availability and affordability of the components for a drug compound. Process research is intent on developing more cost-effective avenues for the drug’s production. A team of process engineers, chemists, and/or geneticists is assigned to examine the underlying chemistry behind the new drug or therapy, in the hopes of finding ways to produce the drug on a larger scale in a commercial manufacturing plant. Following the firm’s IND with FDA, clinical trials will begin, and additional testing demands larger quantities of the drug to be available. The CMOs also offer services to develop the packaging and labeling for a new drug. This process is generally done in close cooperation with the drug firm and the CRO managing clinical trials. Controlled testing requires adherence to a strict protocol in the drug’s composition and exterior appearance and packaging. The dosages must look the same for all three phases of clinical trials. Finally the drug’s packaging must meet child safety

90

Research, development, and testing service industry

regulations while being used in the clinical trial. Blister packets are one example of packaging services that a CMO might offer. The CMOs perform process engineering, manufacturing, and package and labeling development services on a contract basis with sponsor drug firms. CMOs play a significant role in the supply chain for drug development. Although it is important to highlight the services CMOs provide, they are primarily manufacturing facilities and are classified as such and therefore are not in the scope of this study.

R&D and drug development To investigate the nature of R&D in the drug development process, we interviewed several small biotechnology RD&T firms and several larger CROs. Interviews were conducted with the firms’ research directors. The biotechnology firms were small start-up firms, conducting contract research for larger pharmaceutical firms or conducting independent research (funded by venture capitalists) with the hope of being bought out or going public once innovations were proven successful (see Table 6.3). Their activities are characterized as high-risk, high-return research, and risk sharing is one of the services they provide. Although their research eventually affects health care services, most of their research activities are in biological sciences similar to those conducted by large pharmaceutical and medicine manufacturing firms. In addition to small biotechnology firms, we interviewed large CROs (see Table 6.3). CRO is a generic term used in the pharmaceutical industry to Table 6.3 RD&T biotechnology services firms interviewed Name Small firms BioMachines Nobex Zen-Bio Large firms Covance InGenium Research Inveresk Quintiles Transnational

Research activities Designs bench-top computer analysis solutions to automate proteomic R&D Develops and markets various methods for delivering drugs to a patient (e.g. transdermal patch, time release pills) Licenses tissue engineering technology associated with human fat cells (adipocyte) Preclinical pharmacological studies, and Phases I through IV clinical trials and postmarketing studies Phases II through IV clinical trials Process engineering, data management software, assay development Preclinical pharmacological studies, and Phases I through IV clinical trials and postmarketing studies

Research, development, and testing service industry 91 label firms that perform research on a contractual basis for a sponsor pharmaceutical firm. The contract research industry has experienced rapid growth in the past 2 decades, and research performed by CROs spans the preclinical and clinical phases. Several biotech firms interviewed are developing technologies, such as assays,4 to support drug discovery and development of new vaccines. These firms either sell the assays directly to biopharmaceutical firms or use them to provide services. Few of the experts interviewed considered their firm a service provider. Most saw their business as a small player in the rapidly growing sector of pharmacogenomics (development of patient-specific therapies based on the individual’s genetic makeup). Although most of the experts did not consider their business a service, their research was primarily funded by pharmaceutical firms. They indicated that Big Pharma typically focus on their areas of core competencies and outsource to biotech firms to research unexplored gene therapies. This practice reduces risk while expanding the number of markets in which a pharmaceutical firm can compete. Of those interviewed, none reported performing basic research. However, they often partner with local universities to adopt and leverage knowledge as it emerges. In addition, many firms purchase the rights to existing research on a particular compound and continue its development. Respondents reported that basic research is often too costly to perform in-house because it generates little or no revenue. The biotechnology firms interviewed attributed the largest share of their R&D activities to performing highly specialized applied research to support the creation of proprietary technology, which in turn can be marketed as a service through licensing or contracting. Responses varied for development activities depending on the firm’s area of research and business interest. Following the completion of preclinical trials, biopharmaceutical firms either complete development in-house, through CROs, or by allowing larger pharmaceutical firms to take over development. Firms specializing in assay design reported that their largest share of R&D was spent in development. The biotechnology industry crosses over many disciplines from agriculture and the environment to health care and industrial applications. However, smaller individual biotech firms typically specialize and develop core competencies in specific genes and treatments that increase their probability of success in discovering a new marketable compound, therapy, or process. These smaller biopharmaceutical firms want to rapidly bring compounds to market using as few internal resources as possible. Outsourcing by small R&D firms to CROs is becoming increasingly more common for achieving this goal. Preclinical testing is normally kept in-house to ensure that proprietary information is kept confidential.

92

Research, development, and testing service industry

However, given very few resources to spare, smaller biopharmaceutical firms will often outsource some of the more time-consuming studies to a CRO. Examples include bioavailability, drug optimization, and toxicology. In summary, the biotechnology service firms we spoke with had little problem classifying their activities as R&D and were familiar with the distinction between basic, applied, and developmental research. However, even though many were conducting contract research for larger biotech or pharmaceutical firms, none of the firms we contacted consider themselves to be service firms. Although many research projects fail, the small biotech firms that encounter a successful outcome from their applied research are frequently purchased by a large pharmaceutical firm, transfer all or part of the intellectual property to a partner through a licensing of the proprietary technology, or create a third jointly owned entity. Such business models may be supported by the RD&T service sector just as independent biotechnology firms use these services. However, in recent years, many smaller biotech firms have become more independent because of increases in capital availability. Standard & Poor’s (S&P’s) industry profile reports that from 1999 to 2002 the biotech industry raised approximately $60 billion through a variety of public and private financing sources, whereas only $15 billion were raised from 1996 to 1998. Nevertheless, the huge total cost of bringing a new drug to market necessitates both hybrid organizational strategies and the extensive use of efficiency-enhancing RD&T service firms.

CRO R&D activities Outsourcing by both large and small pharmaceutical firms is generally done when the drug firm believes that it is overburdened by work relating to the drug development process or believes that a CRO may have expertise that allows it to conduct aspects of the process more efficiently. CROs work with their sponsor client to develop trial protocols, standardize their methods for data analysis and regulatory submissions, and create a seamless interchange of communication and data exchange. The CROs specialize in every aspect of drug development, including medicinal chemistry, preclinical testing, clinical trials, sales and marketing, and postmarket testing. Of those we interviewed, none reported performing basic research. CROs leverage technology emerging from universities and software vendors to supply their clients with the most efficient tools and processes to conduct the studies they have been contracted to perform. The CROs interviewed described the small share of their R&D activities directed

Research, development, and testing service industry 93 at applied research to be targeted at the creation of proprietary technology, which in turn can be marketed as a service through licensing or contracting. This research is centered in process engineering with a focus on assay development and testing techniques. The CROs interviewed attributed the largest share of their R&D to performing development activities. CROs adopt existing technologies from various industries and convert the technology to meet their business needs. Much of the development is aimed at streamlining the information flows between the sponsor and the CRO and improving the speed and accuracy at which the data can be acquired and analyzed. Following the completion of preclinical testing and the submission of an IND to FDA, biopharmaceutical firms will most likely contract with a leading CRO that can conduct the clinical trials and efficiently process and analyze data from the field to maximize the speed at which the new drug moves through the FDA approval process. As discussed earlier, many types of CROs specialize in one or all of the services a biopharmaceutical firm may need. However, the core mission is always the same—to provide efficient and accurate methods and tools that ensure timely completion of the clinical trials and regulatory submissions. From our interviews, we found that some of the largest CROs were conducting R&D in support of their research services. The most common areas of development were in process engineering in service delivery and software development in data acquisition, exchange, and security. Process engineering was cited as an area of constant reworking in the industry. A CRO competes on the promise that it has the fastest, most efficient, and streamlined system for moving the drug firm’s IND through clinical trials. As a result, the larger, more competitive CROs spend large sums of money on research to streamline processes in site recruitment, communication with sponsors, and the efficiency and accuracy of data delivery and analysis. The CROs interviewed also mentioned software development as another source of R&D conducted internally to enhance their services. In all cases, this generally included purchasing existing software from third-party vendors such as Oracle that specialize in clinical trial database management applications. Once purchased, CROs modify these applications to meet the needs and specifications that they believe will have the greatest impact on the quality of services they provide. In addition to modifying existing software, the CROs in cooperation with the pharmaceutical industry have recently developed an industry standard for how data are defined across the industry. The Clinical Data Interchange Standards Consortium (CDISC) is the industry organization leading this initiative. Electronic data capture (EDC) is another area of software development that CROs are working on. Massive quantities of clinical data are generated in

94 Research, development, and testing service industry each of the three phases of trials. This information was traditionally filled out on a bubble sheet questionnaire by the physician or patient volunteer and then mailed or faxed back to the CRO where the data were entered into a database. In the late 1990s, CROs began to conceive of compiling trial data over the Internet. Today, Web services allow physicians and patients around the world to enter clinical data into laptops, cell phones, or palm devices, which then send the data over the Internet directly to the CRO’s database. Security becomes an important issue as EDC technologies are developed. CROs are forced to comply with the sponsor’s strict confidentiality requirements while transferring data and must develop security protocols that ensure that the data will not be captured during transmission from the field to the CRO.

CMO R&D activities Biopharmaceutical drug development also requires some special research on the manufacturing side as well. CMOs research the most efficient process for manufacturing large quantities of the new drug. For this reason CMOs adopt many of the same technologies the biopharmaceutical firm used to discover the drug. The following tools are used to perform basic and applied research that will ultimately lead to the discovery of a novel pharmaceutical application or gene therapy, new medical diagnostic device or method, or a preventative vaccine. These technologies can also inform and facilitate the product and process development activities conducted at CMOs further down the pipeline. Other examples of R&D activities that may be conducted at biotechnology service firms include the following: ●

● ●

●

●

Bioprocessing technology: Uses single-cell microorganisms to catalyze biochemical reactions that in turn produce new products. This technology is used to create insulin, biodegradable plastics, and vaccines for various blood-borne pathogens. Cell culture: Growing living cells outside of living organisms. Recombinant DNA technology: Preferential genetic selection and expression. This technology is used in combination with cloning or protein engineering to achieve new genetic properties in existing cells or proteins. Cloning: Production of genetically identical copies of a molecule, cell, or animal allows researchers to study genetic diseases and processes and discover potential drugs and therapies. Protein engineering: Technology used to improve the chemical structure of proteins and enzymes to be used in drug development, food processing, and industrial manufacturing.

Research, development, and testing service industry 95 ●

●

●

Biosensors: Attaches biological components to transducers to measure extremely low concentrations of certain substances. This technology allows for the development of products that measure nutritional value of food or allows physicians to measure the levels of vital blood components. Nanobiotechnology: Combination of nanotechnology with microbiology to discover stable micro-structures that can be employed to create nanoprocessors to channel electronic signals. Microarrays: Used to study gene structure and function at the DNA or protein molecular level. Arrays are used to monitor and detect mutations in genes and diagnose infectious diseases. These arrays also support drug discovery and development of new vaccines.

Conclusions The RD&T sector provides highly technical services by engaging in research and experimental development in the physical, engineering, and life sciences. More specifically, biotechnology services firms focused in human medical therapeutics perform many of the required activities in the discovery, development, and production phases of drug development. Examples of these activities include assay development, high throughput screening, toxicity and metabolism studies, clinical trial management (in the case of CROs), process engineering and manufacturing (in the case of CMOs). The RD&T sector provides evidence that R&D itself can be a service. The firms that conduct contract R&D are focusing on innovations to optimize processes and reliability of results to enhance the value of their services and differentiate in a competitive market place. The nature of contract R&D suggests that clients set the trajectory for any internal R&D activities performed by the service firm. While innovations in services may be based on customer needs, there is evidence that many of the biopharmaceutical service firms have their own proprietary drug R&D programs. New or small biotechnology firms have no permanent revenue source because they have yet to develop a marketable product. Many firms have begun to market their internal core competencies to perform specific tasks in drug R&D. Marketing services provide a revenue stream that allows a firm to sustain its own internal R&D program. In reference to the service-sector model of innovation presented in Chapter 2 (Figure 2.3), these firms provide contract R&D services, using a combination of external (or purchased) IC in the form of hardware, licensed software, and infratechnologies such as measurement standards. The service firm also contributes its own internal IC through employees and proprietary technologies. Strategic and competitive planning play a large role in

96

Research, development, and testing service industry

informing the firm’s risk analysis and its decision about the optimal mix of each type of IC. Entrepreneurial activity in this case can be defined as identifying and combining purchased and internal IC in a way that substantially enhances the value of the service, providing sufficient differentiation and thereby allowing the firm to remain viable in a competitive market.

7

Dimensions of innovation and productivity growth

Introduction In this chapter we quantify firm-specific aspects of innovative activity in the service sector and compare them to the manufacturing sector. We use firm data from COMPUSTAT for 2003 for these comparative analyses. Our population of firms is certainly not inclusive of all innovative firms, but we are not constrained by that limitation and work with the COMPUSTAT sample of public firms in the two sectors.

Innovation characteristics of manufacturing and service-sector firms The results of several empirical analyses are presented in this section. They are intended to shed light on aspects of innovation characteristics of firms in the manufacturing and service-producing sectors. We note, when appropriate, the extent that our findings corroborate implications of our models in innovation from Chapter 2. Given the burgeoning nature of empirical research on innovation in the service sector, and the fact that the empirical analyses below are some of the first systematic studies of dimensions of innovation, we refrain from positing hypotheses and testing them with our samples of data. Rather, we present our descriptive findings followed by a possible interpretation of the results with reference to the models of innovation in Chapter 2. R&D Spending In an effort to characterize the R&D spending of firms in the manufacturing and service sectors, we examined the population of public firms on COMPUSTAT that reported both R&D expenditures and sales in 2003. There were 1,660 public firms in the manufacturing sector for which positive R&D

98

Dimensions of innovation and productivity growth

Table 7.1 Descriptive R&D statistics for manufacturing and nonmanufacturing firms Sector

R&D expenditures ($ millions)

R&D-to-sales ratio

Manufacturing (n  1,660) Nonmanufacturing (n  698)

Mean  146.6 Median  9.14 Mean  66.9 Median  7.31

Mean  0.13 Median  0.06 Mean  0.15 Median  0.11

Source: COMPUSTAT (2003).

expenditures and positive sales figures were reported; there were 698 public firms in the nonmanufacturing sector for which the same data were reported.1 Table 7.1 reports descriptive statistics for the sample of manufacturing and nonmanufacturing firms. As shown, both the mean and median levels of R&D expenditures are greater in the sample of manufacturing firms than in the sample of nonmanufacturing firms. For public firms doing R&D, those in the manufacturing sector do more of it. But, as shown in the models of innovation in Chapter 2, the activities to which these investments are directed are different. The mean ratio of R&D to sales tells a different story about the intensity of investments in innovation. The mean and median levels of R&D intensity are greater among nonmanufacturing firms doing R&D than among manufacturing firms. Also, we calculated the elasticity of R&D with respect to sales using firm data from each sector. Many have called such an empirical examination a test of the Schumpeterian hypothesis (e.g. Link (1981b)). Our logarithmic regression model is: ln RD  ␣
␤1 ln SALE
␤2 ln SALE * MD
␤3 MD


(7.1)

where RD represents firm R&D expenditures, SALE represents sales, MD is a binary variable equaling 1 if the firm is in the manufacturing sector and 0 other wise, and ␧ is an error term assumed to be random and normally distributed. The ordinary least squares results from equation (7.1) are reported in Table 7.2. The estimated coefficient on ln SALE is positive and significant, implying that a 10 percent increase in sales in service-sector firms (MD  0) is associated with a 6.65 percent increase in R&D expenditures. The estimated coefficient on (ln SALE * MD) is also positive and significant, implying an elasticity of R&D with respect to sales among manufacturing firms (MD  1) of 0.761 (0.665
0.096). Clearly, although the models in Chapter 2 do not address this point, the estimated elasticity of R&D with respect to sales is greater among manufacturing firms than among nonmanufacturing firms.

Dimensions of innovation and productivity growth 99 Table 7.2 Regression results on the elasticity of R&D Variable

Estimated coefficient (t-statistic)

Intercept

 1.17 ( 10.15) 0.665* (29.57) 0.096* (3.69)  0.58* ( 4.19)

ln SALE ln SALE * MD MD R2  0.6416 n  2,358

Note * Statistically significant at the .01-level or higher.

Research joint ventures2 The National Cooperative Research Act (NCRA) of 1984, Public Law 98-462, was legislated, as stated in the Preamble to the Act:3 to promote research and development, encourage innovation, stimulate trade, and make necessary and appropriate modifications in the operation of the antitrust laws. The NCRA of 1984 created a registration process, later expanded by the National Cooperative Research and Production Act (NCRPA) of 1993 and the Standards Development Organization Advancement Act of 2004 (SDOAA), under which research joint ventures (RJVs) can voluntarily disclose their research intentions to the US Department of Justice; all disclosures are made public in the Federal Register.4 The RJVs gain two significant benefits from filing with the Department of Justice, and these benefits are what could be referred to as a safe harbor for participants in the venture. One, if the venture is subjected to criminal or civil action, the charges would be evaluated under a rule of reason that analyzes whether the venture improves social welfare. And two, if the venture is found to fail a rule-of-reason analysis, it is subject to actual damages rather than treble damages. There is a vast theoretical literature, recently summarized by Hagedoorn et al. (2000) and Combs and Link (2003), that concludes that, among other things, collaborative research increases the efficiency of the R&D conducted by the collaborating members. Link and Rees (1990) provided early empirical evidence of

100

Dimensions of innovation and productivity growth

this proposition, but to date related empirical research has been sparse owing to the paucity of public domain data related to research collaborations. However, there has been conspicuously absent from the policy landscape rich public domain information about cooperative research activity. The CORE (COoperative REsearch) database was constructed under NSF’s sponsorship and is maintained under their support by Link, for the purpose of chronicling what public information there is. Its resource base is information in the RJV filings with the Department of Justice as disclosed publicly in the Federal Register. The unit of observation in the CORE database is the RJV. As shown in Table 7.3, between 1995 and 2004, nearly three times as many RJVs were filed with members from, primarily, the manufacturing sector than from the service sector. Perhaps, manufacturing firms view collaborative research to be a more strategically competitive innovation arrangement. Also, based on Table 7.3, RJVs among manufacturing firms are noticeably more likely to include a university as a research partner than are those among servicesector firms, perhaps reflecting for firms in that sector a greater importance of access to the science base in the innovation process. See Figures 2.2 and 2.3. Also, service-sector firms are noticeably more likely to be engaged in RJVs related to infrastructure technology than are manufacturing firms, perhaps reflecting the more direct relationship between that form of technology and market development. See Figures 2.2 and 2.3. Finally, of descriptive interest is the fact that RJVs among service-sector firms have a greater likelihood of having a foreign firm as a member of the joint venture than RJVs among manufacturing firms. Perhaps this reflects the fact that service-sector innovations have less of a proprietary nature than do those in manufacturing. Table 7.3 Membership in RJVs Characteristics

Manufacturing industry RJVs

Service-sector industry RJVs

Number of filings Mean number of members Percentage of filings with process technology focus Percentage of filings with university research partners Percentage of filings with infrastructure technology focus Percentage of filings with foreign members

641 13.4 50.1

219 14.6 49.3

16.1

11.9

12.1

20.1

26.5

35.1

Source: CORE database.

Dimensions of innovation and productivity growth 101

R&D and productivity growth framework5 The theoretical model Since the early 1960s, researchers have conducted empirical analyses of the impact of investment in R&D on productivity and productivity growth. The framework used by most researchers begins with a Solow-like (1957) model. A generalizable production function applicable to the ith firm, ith industry, ith sector, or the economy, can be written as: Qi  Ai F (K, L, T)i

(7.2)

where Q represents output. This model is commonly referred to in the literature (Griliches and Lichtenberg, 1984) as the R&D capital stock model. In equation (7.2), A is a neutral disembodied shift factor. The stock of physical capital and labor or human capital are K and L, respectively. The stock of technical capital available to the unit of observation, hereafter referred to as the firm for simplicity, is represented as T. T in turn can be written in terms of the alternative sources on which the firm acquires technical knowledge. Although this extended model in equation (7.3) has most often been applied in an abbreviated form primarily because of data limitations, its general representation includes four sources of technical knowledge. Following Charles River Associates (1981): Ti  G (OTi, PTi, GTi, IT)

(7.3)

where OTi is the ith firm’s own or self-financed stock of technical knowledge, PTi is the ith firm’s purchased stock of technical knowledge, GTi is the ith firm’s government-financed stock of technical knowledge, and IT is the infrastructure technology that the ith firm uses in its production of technology process. Note that there is no firm subscript on IT since it, as discussed with reference to the models in Chapter 2, is public domain technology. As is most common in the literature, OTi is assumed to be the relevant source of technical knowledge affecting the firm’s productivity, and further it is assumed to be related to the ith firm’s internal or self-financed previous R&D expenditures, RD, as: OTi 
ai,j RDi,t j

(7.4)

where the ith firm’s accumulation weights, aj, reflect the influence of a j-period distributed lag and obsolescence rate of R&D.

102

Dimensions of innovation and productivity growth

Early empirical studies of the technology-productivity growth relationship were based on a simplified version of the model in equation (7.4), where the only argument defining Ti was OTi and the production function was Cobb-Douglas: Q  Ae␭ t K ␣ L(l␣) T ␤

(7.5)

where ␭ is a disembodied rate of growth parameter and ␣ and ␤ are output elasticities. Constant returns to scale are assumed with respect to K and L, but not with respect to T. Using logarithmic transformations and differentiating the resulting version of equation (7.5) with respect to time, t, one obtains: Q/Q  {(Q/t)/Q}  ␭
␣ {(K/t)/K}
(1␣){(L/t)/L}
␤ {(T/t)/T}

(7.6)

Residually measured productivity growth is defined as: A/A  Q/Q  ␣ {(K/t)/K}  (1␣) {(L/t)/L}  ␭
␤ {(T/t)/T}

(7.7)

where A/A  {∂A/∂t)/A}. In equations (7.5) through (7.7), the parameter ␤ is the output elasticity of technical capital: ␤  (Q/T) (T/Q)

(7.8)

For ␤  0, equation (7.5) exhibits constant returns to scale in K and L and increasing returns to scale in K, L, and T. Substituting the right-hand-side of equation (7.8) into equation (7.7) and rearranging terms yields: A/A  ␭
␳ (T/Q)

(7.9)

for T  (T/t) and for ␳  (Q/T). From equation (7.9), ␳ is the marginal product of technical capital and T is the decision making unit’s net private investment in the stock of technical capital. It is generally assumed in the early empirical work, that the stock of R&Dbased technical capital does not depreciate, or if it does depreciate it does so very slowly. Thus, T is reasonably approximated by the flow of self-financed

Dimensions of innovation and productivity growth 103 R&D expenditures in a given period of time, RD, as: A/A  ␭
␳ (RD/Q)

(7.10)

Empirical estimates of ␳ from equation (7.10) have been interpreted as an estimate of the marginal private rate of return to investments in R&D. To the extent that the R&D stock of technical capital does in fact depreciate, such an estimate of the marginal private rate of return is downwardly biased, as demonstrated by Scherer (1982). Also, Schankerman (1981) discussed the impact of double-counting in the calculation of the private returns to R&D, since R&D expenditures are often already included in measures of K and L. Thus, it is common in the later variants of this literature to refer to ␳ as an excess rate of return, meaning a rate of return in excess of normal remuneration to conventional factors of production. The extant empirical literature related to estimates of the R&D to productivity growth relationship, using equation (7.10) or a variant, is reviewed in the appendix to this chapter.

Empirical estimates related to manufacturing and service-sector industries Based on survey results from a sample of 114 COMPUSTAT firms, information was collected on the percentage of existing (2003) technologies that were induced, meaning that resulted from in-house R&D activity, as opposed to purchased.6 Earlier work related to induced technology was conducted by Link (1983a,b) and Link et al. (1983a). Here, we formulated our survey instrument to parallel their data collection. Table 7.4 shows the estimated results from the following model: INDUCE  ␣
␤1 SALE
␤2 MD
␧

(7.11)

where INDUCE is the percentage of existing (2003) technologies that were induced, meaning that resulted from in-house R&D activity, as opposed to being purchased, and where SALE, MD and ␧ have been previously defined. The mean value of INDUCE for manufacturing firms is 0.62 and for servicesector firms it is 0.29.7 These differences in mean statistics correspond to the models of innovation in Chapter 2, in which in-house R&D is a driver of innovation in manufacturing, and purchased technology and technical service are drivers of innovation in the service sector. As shown in Table 7.4, larger firms induce more technology than do smaller firms, and those in the manufacturing sector induce more than those in the service sector.8 Based on our data on INDUCE, the R&D-to-productivity growth model in equation (7.10) was estimated, and the regression results are in Table 7.5.

104

Dimensions of innovation and productivity growth Table 7.4 Regression results on sectoral differences in induced technology Variable

Estimated coefficient (t-statistics)

SALE

2.87 (2.17)** 4.52 (3.78)*

MD R2  0.39 n  114

Notes * Statistically significant at the .01 level or higher. ** Statistically significant at the .05 level.

Table 7.5 R&D and productivity growth Variable

Estimated coefficient (t-statistics) (1)

(2) 0.381** (2.01) —

RDinduce/Q

0.051* (2.71) 0.052 (1.83) 0.071* (2.64) 0.813 (1.56) —

RDpurchase/Q

—

RDinduce/Q * DM

—

RDpurchase/Q * DM

—

R2 n

0.371 114

Intercept RD/Q RD/Q * DM DM

— — 0.064 (1.08) 0.211* (3.11) 0.224* (3.92) 0.105 (1.56) 0.482 114

Notes * Statistically significant at the .01 level or higher. ** Statistically significant at the .05 level.

Following Griliches (1986), Link (1980, 1981a), and Mansfield (1980), among others, A/A was calculated over the period 1999 to 2003 using COMPUSTAT data. From column (1) the estimated rate of return to R&D among service-sector firms is 5.2 percent, but the associated coefficient is

Dimensions of innovation and productivity growth 105 only marginally significant; the estimated rate of return to R&D among manufacturing firms is 12.3 percent (0.052
0.071). Total R&D expenditures were divided between those in-house activities that are devoted to inducing technology, presumably proprietary technology, as opposed to those in-house activities that are devoted to adopting and modifying purchased technology. This disaggregation was constructed by multiplying R&D expenditures for each firm by the firm’s reported value for INDUCE. When R&D expenditures are so disaggregated, the results in column (2) of Table 7.5 show that the rate of return to R&D devoted to induced technology among manufacturing firms is 28.8 percent (0.064
0.224); the rate of return to R&D devoted to purchased technology among service-sector firms is 21.1 percent. We do not interpret the results in Table 7.5 to reflect that manufacturing firms are more productive with their R&D investments than service-sector firms. As posited in Chapter 2, the nature of R&D is distinctively different between the two sectors. Rather, we interpret the results to mean that both manufacturing firms and service-sector firms are allocating their internal R&D resources in ways that generate appropriate productivity growth enhancements.

Appendix Empirical evidence on the R&D-to-productivity growth relationship The large body of literature on estimating equations similar to equation (7.10) can be summarized in various ways. Here, we review relevant studies in a chronological order, based on the sample period used by the authors in an effort to better document the evolution of this body of research. Studies focusing on the early postwar decades reported a strong positive relationship between R&D and productivity growth. However, findings reported for later years are mixed. For the most part, these studies have been based on US data. The 1950s and 1960s Minasian’s (1962) investigation focused on 85 industrial firms over the period 1947 to 1957. Using a simplified version of equation (7.10), he found that R&D spending was a statistically significant determinant of productivity growth in chemical firms. Griliches (1973) reached a similar conclusion from an analysis of two-, three-, and four-digit manufacturing industries between 1958 and 1963. In a subsequent article, Griliches (1980b) extended his analysis to examine a sample of firms during the 1965 to 1975 period. His later results were similar to earlier results in that he found that

106

Dimensions of innovation and productivity growth

the R&D-to-productivity growth relationship was positive and statistically significant. Perhaps the most extensive industry-level investigation from this early literature was by Terleckyj (1974). Based on his analysis of two- and three-digit industries, he too found ␳ in equation (7.10) to be positive and statistically significant. More important than simply adding to the body of empirical literature that R&D expenditures are an important determinant of productivity growth, he was perhaps the first to go beyond the specification in equation (7.10) to a specification that approached the conceptual representation of a production function using T as defined by equation (7.2). Specifically, he attempted to quantify the impact of interindustry spillovers of technical knowledge by adding to equation (7.10) a variable for government-financed R&D and R&D embodied in purchased inputs, intermediate goods, and capital goods, specifically. Government-financed R&D had no statistical impact on measured productivity growth but embodied R&D did. One might date the origin of the literature on diffusion or spillovers as an indicator of economic performance to this Terleckyj study. The early 1970s The US economy saw a modest decline in total factor productivity growth until about 1973, which then became more severe and lasted until the end of the decade. This post-1973 decline was especially steep for the manufacturing sector. Meanwhile, industrial R&D spending had begun to decline in both relative and absolute terms during the late 1960s. Not surprisingly, given the empirical research reviewed just above, researchers as well as policy makers focused on the declining R&D spending pattern to identify a technology-based culprit for the 1970s productivity growth slump. Total US R&D as a percentage of GNP peaked at 3 percent by 1964 and then fell to 2.3 percent by 1975. There was disagreement regarding the quantitative impact of the R&D decline on measured productivity growth. Nadiri (1980) and Nadiri and Schankerman (1981) estimated that the reduced rate of R&D stock accumulation might have accounted for as much as one-third of the post-1973 productivity decline. Denison (1979) and Griliches (1980a) averred that the R&D slowdown accounted for at best one-tenth of the decline in productivity growth. Griliches (1980a) posited that it was not the slowdown in R&D that was important but rather the collapse in the productivity of R&D itself. But, in subsequent years, based on analyses of newer microeconomic data sets, Griliches’ opinion changed. This latter group of studies included the works of Griliches and Mairesse (1984), Cuneo and Mairesse (1984) for French manufacturing firms, and Clark and Griliches (1984). Lichtenberg and Siegel (1991)

Dimensions of innovation and productivity growth 107 reported that the private, or firm level, returns to company-funded R&D in the United States remained high during the productivity slowdown of the 1970s. Post-slowdown policy It is well documented that R&D spending, within the framework that underlies equation (7.10), is positively and statistically significantly correlated with measured productivity growth in the United States and in selected OECD countries. Thus, it is not surprising that policy makers worldwide advocated policies to stimulate industrial R&D in an effort to reverse the productivity slowdown that persisted until the early 1980s. The immediate policy response, and the timing of this response varied by country, was to encourage a higher level of industrial R&D spending. R&D and R&D-related tax credits were widely introduced, in many cases because of the overwhelming empirical evidence of the relationship between productivity growth and R&D spending. Simply put, it was believed that a tax credit would lower a firm’s cost of conducting R&D and, holding marginal benefits constant, provide firms with an incentive to increase their level of spending. Feldman et al. (2002) and Link (2006) discuss these credits with reference to the United States. Based on this body of research, including an article by Leyden and Link (1993) and the Hall and van Reenan (2000) synthesis and interpretation of the literature, it is fair to conclude that the literature remains mixed about the effectiveness across countries of this policy. This suggests that such tax credits seem to be revenue neutral, that is that dollar for dollar the amount of the credit equals the amount of the increase in R&D spending. The most complete review of this literature is presented in Hall and van Reenen (2000). Disaggregate R&D studies Referring to equations (7.3) and (7.10), the state of the art has, for the most part, been to view privately financed R&D as the primary determinant of productivity growth. This finding appears to be quite robust at the firm and industry levels. This focus on R&D in total was due in large part to the paucity of publicly available data at a microeconomic level on R&D either by character of use—basic research, applied research, and development—or by source of funding—self-financed R&D versus R&D embodied in purchased technology versus government-financed R&D. As noted above, Terleckyj’s (1974) work was pioneering with regard to its attempt to disaggregate R&D using public domain data. Since then, a number of studies have extended this avenue of investigation, using primarily

108

Dimensions of innovation and productivity growth

Table 7A.1 Disaggregated R&D-to-productivity growth studies Author

Level of analysis

Disaggregation of R&D

Terleckyj (1974)

US industries

Mansfield (1980) Link (1980) Link (1981a,b)

US firms US firms US firms

Link (1982a) Terleckyj (1982) Scherer (1982, 1983b)

US firms US industries US industries

Griliches (1986)

US firms

Lichtenberg and Siegel (1991)

US firms (with productivity estimates based on plant-level data)

Government-funded R&D R&D embodied in purchased capital Private R&D by character of use Private R&D by size of firm Private R&D by character of use Government R&D by character of use Product and process R&D Product and process R&D Product and process R&D R&D embodied in purchased capital Private R&D by character of use Government R&D by character of use Private R&D by character of use Government R&D by character of use

survey data. These studies are reviewed in Table 7A.1. The most important conclusion to be drawn from these studies is that firms rely on myriad sources of technical knowledge, and each source does affect not only the firm’s productivity growth but also the efficiency with which it conducts R&D. Mansfield (1980) and Link (1981a) disaggregated self-financed R&D by character of use and found that the marginal return to basic research was the greatest. Further, Link (1981a) disaggregated government-financed R&D by character of use and found that government-financed basic research was statistically more important than government-financed applied research or development at the firm level. Using more recent and more comprehensive data, Lichtenberg and Siegel (1991) also found a productivity premium associated with basic research. Both the Mansfield and Link studies of self-financed firm R&D by character of use were verified by Griliches’ (1986) analysis using an alternative data set. Link (1982a) and Terleckyj (1982) disaggregated R&D, firm R&D, and industry-level R&D, respectively, into process-enhancing expenditures and product development expenditures and found that the former had the greater statistical relationship to productivity growth. Their findings were verified by Scherer’s (1982, 1983b) industry-level line of business analysis. More

Dimensions of innovation and productivity growth 109 importantly, Scherer’s work, followed by Link’s (1983b), emphasized the productivity growth-enhancing nature of technology embodied in purchased capital equipment, that is interindustry technology flows. The relationship between government-financed R&D and productivity growth is more complex. The complexity is because the impact of government-financed R&D is not independent of its relationship with self-financed R&D. Beginning with Blank and Stigler (1957), scholars have debated the complementarity or pump-priming versus substitution effect between the two. The latter is often referred to as the crowding out hypothesis. This literature—extended by Higgins and Link (1981), Link (1982b), Lichtenberg (1984, 1987, 1988), Holemans and Sleuwagen (1988), Antonelli (1989), Leyden et al. (1989), Leyden and Link (1991, 1992), Wallsten (1999), and reviewed by David et al. (2000)—is important because it emphasizes that alternative technical sources are interrelated and appear to affect productivity growth. Three external sources of technical knowledge have not yet been discussed with reference to the specifications in equations (7.3) and (7.10) that are considered to be important in the technological change and productivity growth literature. The first source is research partnerships. The US economy did recover from the productivity growth slowdown, in terms of measured domestic productivity growth in the early to mid-1980s, but so did productivity growth in other countries. Unfortunately for US industry, a nagging consequence of declining industrial R&D and productivity growth was a loss in global market share for many American industries. Although not documented officially until after the fact, the trend in the competitiveness of American industries in emerging technologies was not optimistic. The Technology Administration of the US Department of Commerce (1990) reported that as of 1989 the United States was losing badly—the Department of Commerce’s worst of four categories— to Japan in advanced materials, biotechnology, digital imaging technology, and superconductors. With respect to the Europeans, American firms were losing market share in digital imaging technology and flexible integrated manufacturing. The Council on Competitiveness (1991) reached a similar conclusion, noting that the United States was, at that time, experiencing declining market share in materials and associated processing technologies, engineering and production technologies, and electronic components. One response to the anticipated trends was a policy effort to increase the speed with which US firms could conduct their R&D. The R&E tax credit in 1981 was intended to affect the level of US R&D spending, but that emphasis was considered insufficient for increasing the pace of innovation at the time. The NCRA of 1984 created a process whereby firms could file RJVs with the

110

Dimensions of innovation and productivity growth

US Department of Justice with a strong likelihood that the government would not prosecute them for engaging in collusive behavior. While the overall effectiveness of this policy initiative is still somewhat unclear, Link and Bauer (1989) have shown that collaboration increases the efficiency with which firms conduct their in-house R&D. In terms of equation (7.10), they show that ␳ is statistically greater among US manufacturing firms that engage in collaborative research than among those that do not. Hence, through collaboration, firms learn to conduct their R&D more efficiently and thus enjoy greater productivity growth and, according to Link and Bauer, greater competitive advantage. Link et al. (2002) have also shown that the NCRA is facilitating its goal of creating a favorable environment, or safe harbor, for conducting research cooperatively. Infrastructure technologies are a second key source of technical knowledge. Few technologies developed through in-house or indigenous R&D can be successfully commercialized and achieve market penetration in the absence of infrastructure technologies. Following Tassey (1992) and Link and Tassey (1993), infrastructure technologies originate outside of the boundaries of the firm and are accordingly elements of the technology base on which all firms in the industry rely on. These technologies include measurement methods, test methods, scientific and engineering data bases used in conducting R&D, process controls, and various other standards. Infrastructure technologies increase the efficiency with which technology-based economic activity is conducted. They act collectively as a leveraging agent on R&D, as well as on production. Most infrastructure technologies are in the public domain, especially those technologies created through research conducted in federal laboratories, as Leyden and Link (1992) have shown. However, the private sector also invests in infrastructure technology for important internal purposes, such as to benchmark against national standards. Using aggregated US manufacturing industry data, Link and Tassey (1993) showed that total factor productivity growth was greater in those industries where firms invest a larger portion of their self-financed R&D in infrastructure technology. The science base resides in the public domain, but it is enriched primarily through basic research conducted at universities. Thus, university-based research is an important source of technical knowledge that will leverage the efficiency with which firms conduct their R&D and hence will affect productivity growth. Referring to equation (7.10), Link and Rees (1990) showed, using a sample of US manufacturing firms, which could be divided into those involved with universities in research and those not, that the private rate of return to R&D, ␳ in equation (7.10), was nearly three times as large for the group of firms with university research relationships.

8

Public policies to enhance innovation

Introduction The US government has historically had an important partnership role with the private sector in fostering innovation. This intuitive conclusion logically follows from the following: innovation leads to technology; technology is the prime driver of economic growth; in the absence of government intervention, firms in all sectors will underinvest in the innovation process (especially in R&D) due to knowledge spillovers and appropriability issues; and government has a responsibility to address this underinvestment by providing incentives for the continued conduct of, or perhaps increase in, R&D. Such sequential reasoning to justify the role of government in innovation has dominated the history of public-sector involvement in the innovation process and more recently the growth of public/private partnerships as related to innovation since the colonial period in US history, in general, and since the Patent Act of 1790, in particular. However, the economic underpinnings of government’s role in innovation are more complex than the above logic might suggest, as are public policies to enhance innovation in the service sector.1

Government’s role in innovation2 The theoretical basis for government’s role in market activity is based on the concept of market failure. Market failure is typically attributed to market power, imperfect information, externalities, and public goods. The explicit application of market failure to justify government’s role in innovation, in R&D activity in particular, is a relatively recent phenomenon within public policy. Many point in the United States to President George Bush’s 1990 U.S. Technology Policy as that nation’s first formal domestic technology policy

112 Public policies to enhance innovation statement. Albeit an important initial policy effort, it, however, failed to articulate a foundation for government’s role in innovation and technology. Rather, it implicitly assumed that government had a role and then set forth the general statement: The goal of U.S. technology policy is to make the best use of technology in achieving the national goals of improved quality of life for all Americans, continued economic growth, and national security. (1990, p. 2) President William Clinton took a major step forward from the 1990 policy statement in his 1994 Economic Report of the President by articulating first principles about why government should be involved in the technological process: The goal of technology policy is not to substitute the government’s judgment for that of private industry in deciding which potential “winners” to back. Rather, the point is to correct market failure.3 (1994, p. 191) Subsequent Executive Office policy statements have echoed this theme; Science in the National Interest (1994) and Science and Technology: Shaping the Twenty-First Century (1998) are among the examples. President Clinton’s 2000 Economic Report of the President elaborated on the concept of market failure as part of US technology policy: Rather than support technologies that have clear and immediate commercial potential (which would likely be developed by the private sector without government support), government should seek out new technologies that will create benefits with large spillovers to society at large. (2000, p. 99) Relatedly, Martin and Scott observe: Limited appropriability, financial market failure, external benefits to the production of knowledge, and other factors suggest that strict reliance on a market system will result in underinvestment in innovation, relative to the socially desirable level. This creates a prima facie case in favor of public intervention to promote innovative activity. (2000, p. 438)

Public policies to enhance innovation 113 Market failure, in particular, technological or innovation market failure, results from conditions that prevent organizations from fully realizing or appropriating the benefits created by their investments. Arrow (1962) identified three sources of market failure related to knowledge-based innovative activity—“indivisibilities, inappropriability, and uncertainty” (p. 609). To explain, consider a marketable technology to be produced through an R&D process where conditions prevent full appropriation of the benefits from technological advancement by the R&D-investing firm. Other firms in the market or in related markets will realize some of the profits from the innovation, and of course consumers will typically place a higher value on a product than the price paid for it. The R&D-investing firm will then calculate, because of such conditions, that the marginal benefits it can receive from a unit investment in such R&D will be less than could be earned in the absence of the conditions reducing the appropriated benefits of R&D below their potential, namely the full social benefits. Thus, the R&D-investing firm may underinvest in R&D, relative to what it would have chosen as its investment in the absence of the conditions. Stated alternatively, the R&D-investing firm may determine that its private rate of return is less than its private hurdle rate; therefore, it will not undertake socially valuable R&D. The basic concept can be illustrated with Figure 8.1, which follows from Tassey (1997) and Jaffe (1998). The social rate of return is measured on the vertical axis along with society’s hurdle rate on investments in R&D. The private rate of return is measured on the horizontal axis along with the private hurdle rate on R&D. A 45-degree line (dashed) is imposed on the figure under the assumption that the social rate of return from an R&D investment will at least equal the private rate of return from the same Social rate of return Private hurdle rate A

45°

B

Social hurdle rate

Private rate of return

Figure 8.1 Spillover gap between social and private rates of return to R&D.

114

Public policies to enhance innovation

investment. Two separate R&D projects are labeled as project A and project B. Each is shown, for illustrative purposes only, with the same social rate of return. For project A, the private rate of return is less than the private hurdle rate because of barriers that limit the firm’s ability to appropriate returns. As such, the private firm will not choose to invest in project A, although the social benefits from undertaking project A would be substantial. The principle of market failure illustrated in the figure can reflect differences in risk preferences and a firm’s ability to diversify risk, as well as issues related to appropriability of returns to investment. The vertical distance shown with the double arrow for project A is called the spillover gap; it results from the additional value society would receive above what the private firm would receive if project A were undertaken. What the firm would receive (along the 45-degree line) is less than its hurdle rate because the firm is unable to appropriate all of the returns that spill over to society. Project A is the type of project in which public resources should be invested to ensure that the project is undertaken. In comparison, project B yields the same social rate of return as project A, but most of that return can be appropriated by the innovator, and the private rate of return is greater than the private hurdle rate. Hence, project B is one for which the private sector has an incentive to invest on its own or, alternatively stated, there is no economic justification for public resources being allocated to support project B. For projects of type A where significant spillovers occur, government’s role has typically been to provide funding or technology infrastructure through public research institutions that lowers the marginal cost of investment so that the marginal private rate of return exceeds the private hurdle rate. Note that the private hurdle rate is greater than the social hurdle rate in the figure. This is primarily because of management’s (and employees’) risk aversion and issues related to the availability and cost of capital. These factors represent an additional source of market failure that is related to uncertainty. For example, because most private firms are risk averse (i.e. the penalty from lower than expected returns is weighted more heavily than the benefits from greater than expected returns), they require a higher hurdle rate of return compared to society as a whole that is closer to being risk neutral.

Public policies to enhance innovation With reference to Figure 8.1, technology policy in the United States has a long history of initiatives targeted to overcome the market failures related to R&D investment (Link 1999, 2006). Beginning with the introduction of the patent

Public policies to enhance innovation 115 systems and continuing through more recent tax policies, the government has realized the importance of stimulating private-sector R&D. In general, technology policies related to R&D can be bundled into several major groups: ●

●

●

market structure policies (such as patents) that affect firms’ abilities to appropriate returns; government support for basic R&D and infratechnologies that increase the efficiency of private-sector R&D; direct funding to firms to lower the private cost of R&D in the form of grants, tax credits, or other incentives.

Each of these policy initiatives provided incentives to private-sector firms to increase the level of their R&D spending. The patent system is targeted at addressing appropriability issues associated with innovation. By granting, what is in effect a temporary monopoly, patents provide innovators the opportunity to generate the revenue needed to offset their R&D investments. For service-sector R&D, however, patents frequently are less effective. Much of service-sector R&D is related to process innovation (also referred to as business methods), for which it is inherently more difficult to obtain patents as well as to enforce them. In addition, it is more difficult for process innovations to demonstrate that they are clearly distinct from all other processes (prior art). The government also supports private R&D through the funding of basic research and infratechnologies. Basic research funded through universities or research laboratories leads to the development of generic technologies for which technical risk and spillovers are greatest. For example, the Preeminence Act of 1991 created the Advanced Technology Program whose mission is to fund research to “accelerate the development of innovative technologies for broad national benefit through partnership with the private sector.” Infratechnologies, such as standards and measurement technologies, have a strong public goods nature that leads to underinvestment by the private sector.4 Infratechnologies, such as standards, are particularly important in supporting service-sector innovation. A large share of service-sector R&D is focused on integrating technologies and knowledge imported from the manufacturing sector. Systems integration and customization of information systems depend heavily on the evolution of IT standards (Gallaher and Rowe, 2006). Without well-defined standards and protocols, most of the innovations in the telecommunications, financial services, and other IT-dependent industries would not be possible. Recent decades have seen the introduction of new public/private partnerships to overcome market failure in the form of tax incentives. For example,

116

Public policies to enhance innovation

the R&E Tax Credit of 1981 and its subsequent renewals, the National Cooperative Research Act of 1984 and its subsequent amendments, and the American Technology Innovation-based policies that are designed to increase the level of R&D spending by removing relevant barriers that cause there to be an underinvestment are economically appropriate. Reflecting on the models of innovation for a manufacturing firm and a service-sector firm in Chapter 2, it is reasonable to ask if the United States, or any industrialized nation, needs differential innovation policies, some aimed at the manufacturing sector and some aimed at the services sector. We do not think so. As we have illustrated, R&D is a constant driver in both sectors, although its purposes are sometimes very different. Certainly, the innovation process is distinctively different between the two sectors. However, the categories of market failures and the policy approaches to address these failures are just as relevant for the service-sector R&D investment as for manufacturing R&D investment. Among manufacturing firms, in-house R&D is intended to develop proprietary technology that leverages productivity growth. Among servicesector firms, in-house R&D is intended to leverage the process of adopting and modifying purchased technology that also leverages productivity growth. Thus, innovation polices that stimulate R&D will enhance the productivity growth of firms in both sectors. However, in the service sector, activities that are of an R&D nature are often decentralized through different business units of a firm, making them difficult to measure. Often these activities occur in very small groups that, reasonably, cannot segment their time between traditionally defined R&D and other technical areas. In addition, the product/service development process in the service sector does not fit well with the fundamental R&D concepts of engineering design, prototype testing, and manufacturing process design leading to mass production (concepts that emerged from the manufacturing sector). The following section discusses the issues associated with identifying and measuring R&D in the service sector.

Service-sector R&D: measurement issues and implications for public policy Worldwide, services account for an increasing share of reported R&D. However, the service sector’s share of R&D varies greatly across OECD countries. For example, in Canada and Australia, the service sector accounts for approximately 35 percent of industry R&D, whereas in the United States and Great Britain, the service sector’s share is about 20 percent, and in Germany and Japan it is less than 10 percent (OECD, 2001b). In contrast,

Public policies to enhance innovation 117 the service sector’s role in terms of its share of economic activity and growth is similar across most of the OECD countries. These cross-country differences highlight the difficulties in identifying and measuring R&D expenditures in the service sector. The ability to accurately identify and measure R&D expenditures is key to the success of public policies targeted at stimulating research. Tax credits and other incentives rely on strict definitions of R&D to determine which activities are eligible. The applicability of definitions and difficulties in identifying and measuring service sector R&D has implications for the effectiveness of such policies. Identifying and measuring service-sector R&D As discussed in previous chapters, classifying innovative activities in the service sector as being driven by R&D or not being driven by R&D is difficult. For example, the service sector has developed and implemented significant innovations related to Web-based applications, information management, and data transfer. In addition, new sectors have evolved that supply system and component integration services to manufacturers and service providers. However, the historical view of R&D and product development (engineering design, prototype testing, and manufacturing process design leading to mass production) does not fit well when services are built on unique applications that are continually customized, incorporating incremental improvements (as opposed to discrete new product models or software versions). In addition to difficulties in identifying R&D activities, service-sector R&D is also inherently more difficult to measure. For example, R&D can be carried out in formal R&D departments or in an informal nature carried out in facilities where R&D is not the main activity. In theory, data collection instruments, such as NSF’s Survey of Industry Research and Development, should identify and measure all resources devoted to R&D. However, breaking out informal R&D may be difficult or costly in some businesses. Because service innovations are more likely to be customer driven, related R&D activities are typically integrated into user companies’ business units. There is less likely to be a stand-alone R&D facility or division in which R&D activities can be readily quantified by existing accounting systems. Service innovation is more likely to be carried out in a multidisciplinary business unit that combines IT system integrators, managers, and market researchers. This places more responsibility on the organization to determine what share of its innovation expenditures is to be classified as R&D when completing the survey.

118

Public policies to enhance innovation

In addition to business units composed of teams of multidisciplinary staff, it is also common for individual staff members supporting innovation in the service sector to have multiple responsibilities and functions. For example, IT managers need to maintain and improve their networks and systems. It is not uncommon for them to spend part of their time managing and overseeing the day-to-day operations of the system and also to contribute to developing the next generation system or service. Again, this makes quantifying R&D difficult, because fractions of expenditures based on existing accounting systems (primarily staff labor) need to be assigned as R&D. Multidisciplinary business units and staff with multiple responsibilities are not business structures unique to the service sector. IT managers in manufacturing companies also pose the same measurement issues. However, the problem has historically been more pervasive in service firms because the R&D supplier and user are more often the same firm, implying that the same individual can be involved in operations/delivery and R&D activities. Patenting service-sector innovations Patents provide short-term (monopoly) rights to innovators as an incentive for stimulating R&D and innovative activity. In addition, patents require public disclosure of the innovation and its enablement, which can lead to greater information dissemination and spillover benefits. However, it is not always clear that patents are an effective way for service sector firms to protect intellectual capital resulting from innovation (Hall, 2003). Historically, many parts of the service sector have engaged in limited patenting, even though a new and useful process has always been explicitly included. However, recently the now-famous State Street Bank & Trust Co. vs. Signature Financial Group Inc.5 reaffirmed that business methods are patentable subject matter. Business methods, as defined by the US Patents and Trademark Office (USPTO), include primarily data processing as related to financial, business practice, management, or cost/price determination (USPTO, 1999). This led to a significant increase in the number of business method patent applications and grants from 1999 through 2001. However, business method patents still account for only about 1 percent of all patents (Hall, 2003). Many experts have questioned the potential benefits of business method patents, stating that they tend to be of lower quality and could potentially stifle innovation. According to Hall (2003), for example, when different and incremental innovations are combined to make a useful product, it is not always obvious that the benefits of the patent system outweigh the costs.

Public policies to enhance innovation 119 In addition, legal scholars have argued that the nonobviousness test is difficult to assess in the case of Internet and business method patents, and the lack of well-established prior art makes assessing the quality of patents difficult. However, other scholars disagree, saying there is no evidence that business method patents are of lower quality than patents in general (Allison and Tiller, 2003). Based on our interviews, we find that firms in the service sector place significant value on patents, related to both scientific and business method innovations. They view patents as an important tool to protect IC. However, firms expressed concern over inconsistencies and costs related to obtaining and enforcing business method patents. Innovative processes frequently fall into a gray area where patents exclude methods of doing business but include business methods. Precedents are only recently being established, and uncertainty over intellectual property rights represents a major barrier to innovative activity in the service sector.

Conclusions The traditional theories of market failure and policy implication appear to be just as valid for service-sector R&D issues as for manufacturing. However, issues of defining and measuring service-sector R&D are different in many instances and need to be investigated. We recommend reviewing and potentially broadening the definition of R&D, for reporting purposes and for targeted innovation policy purposes. Being able to identify and measure service-sector R&D is critical, if accounting-based policies (such as tax credits) are to effectively stimulate R&D investment. In addition, there are uncertainties and inconsistencies in how process and business methods innovation fits into the definition of R&D and how it is covered under the current patent system. Potentially, NSF could devise an instrument to collect this type of R&D information, even if the information is not specifically classified by the firm as R&D or R&D personnel.

Notes

1 Introduction 1 The service sector is also often referred to as the nonmanufacturing sector. However, there is some disagreement among policy makers and economists as to a precise definition of the service sector or service-producing industries. 2 Globally, the service sector accounts for over 60 percent of GDP in industrialized nations, and it is the fastest growing sector in most Organisation for Economic Cooperation and Development (OECD) economies. 3 Service-sector firms are defined in many ways: as being within industries with a close interaction between production and consumption; as producing products with high information content, albeit some are of an intangible nature; and as having a production process with a heavy emphasis on labor relative to capital in the delivery of output (Sirilli and Evangelista, 1998). 4 This point has been emphasized by many scholars, including Lecht and Moch (1999), Jankowski (2001), and Pilat (2001). 5 Of course, to the extent that these four case studies are viewed as nonrepresentative of innovation in the service sector, our model may be limited in its generalizability to the sector as a whole. 2 Innovation in the service sector 1 Howells (2000b) suggests that this manufacturer-centrist view has permeated most attempts to reformulate how service sectors innovate. 2 Nominal R&D data came from the National Science Board (2004). The GDP deflator was used to adjust the annual data for inflation. 3 The first definition of basic research was developed by Dearborn et al. (1953) based on industrial interviews, although the term “basic research” traces to Vannevar Bush (1945), where he proffered the characterization that basic research is performed without thought of practical ends. The Dearborn et al. definition, used in the NSF 1953–54 survey, was (NSF, 1956) the following: Basic or fundamental research—projects that are not identified with specific products or processes applications, but rather have the primary objective of adding to the overall scientific knowledge of the firm. However, NSF expressed concern about the appropriateness of the reporting definition. Thus, in NSF’s 1955–56 survey, the word “fundamental” was deleted and the definition used then, and basically ever since, is the following: Basic research—research projects that represent original investigation for the advancement of scientific knowledge and that do not have specific

Notes 121 4 5 6 7 8 9

commercial objectives, although they may be in fields of present or potential interest to the reporting firm. The Tassey (1997, 2005) model has formed the conceptual basis for several US public/private partnership initiatives. See Link (1999, 2006) for an overview. And the new combinations are the way through which the entrepreneur deals with the disequilibrium brought about by technology development. These include biotechnology firms as well as systems integration firms in the information services sector. Telecommunications and financial services sector firms are examples of technology integrators. This appendix draws directly from Hébert and Link (1988, 2006). After Cantillon’s death, economic analysis in France was dominated by a group of writers who called themselves, simply, “The Economists.” As that term became more general in use, however, historians began to refer to this particular group of French writers as “The Physiocrats” (the term physiocracy means role of nature). Its leader was François Quesnay.

3 Telecommunications industry 1 NAICS codes were revised in 2002. Previously, the telecommunications industry was coded as 5133. 2 This table is delimited by the paucity of firms for which R&D expenditures are reported in COMPUSTAT. 3 Aggregators include firms such as Boingo Wireless, iPass, and GoRemote (formerly GRIC). These firms are involved in building access agreements between numerous independent wireless networks enabling nationwide access to customers without having to build out independent network infrastructure such as wireless towers and access nodes. 4 Financial services industry 1 Because the following information comes from a niche sample of firms, care should be exercised in generalizing to the industry as a whole. 2 Small manufacturing firms behave similarly in that they locate close to the downstream customer (such as an OEM) to provide quick and consistent delivery and after-sales service. 3 See www.marketresearch.com/product/display.asp?productid805899. 4 See www.fstc.org. 5 Product information was retrieved from the Securities and Exchange Commission’s (SEC’s) 10-K fillings archive. 6 The term “proof of concept” was used by the participants in this case study to refer to applied research to demonstrate real-world applications of a technology. 5 Systems integration services industry 1 National statistics indicate that development activities are typically the largest component of R&D expenditures. The emphasis on applied research provided by participants interviewed as part of this study highlights the difficulties in distinguishing between basic, applied, and developmental research in the systems integration services sector.

122

Notes

6 Research, development, and testing service industry 1 Because social science and humanities research falls outside NSF’s definition of R&D, this chapter focuses primarily on physical, engineering, and life science research firms—specifically biotechnology firms. 2 The composition of biotechnology research and definitions of the biotechnology industry are discussed below. 3 Throughout the following discussion, we present statistics and discuss research activities based on similar, but not necessarily identical, definitions of the biotechnology industry. 4 Assays are in vitro cells with characteristics similar to a human cell. They are marketed for use in preclinical tests to prove efficacy of a drug before entering costly clinical trials. 7 Dimensions of innovation and productivity growth 1 We made no effort to distinguish among firms doing $0 R&D and those simply not reporting their R&D expenditures. 2 This background information draws from Link (2005). 3 While the Act set forth these objectives, it did not place them in a historical perspective. In the early 1980s, there was growing concern that the US industrial sector was losing its competitive advantage in global markets. This was explicitly noted in the Research and Development Joint Venture Act of 1983, HR 4043. In the Joint Research and Development Act of 1984, HR 5041, the supposed benefits of joint research and development were first articulated from a policy perspective: “Joint research and development, as our foreign competitors have learned, can be procompetitive. It can reduce duplication, promote the efficient use of scarce technical personnel, and help to achieve desirable economies of scale [in R&D].” After revisions, the NCRA of 1984 was passed. 4 An RJV is a collaborative research arrangement through which firms jointly acquire technical knowledge. 5 This section and the appendix to this chapter draw directly from Link and Siegel (2003). 6 All COMPUSTAT 1999 firms with reported sales and R&D expenditures defined the population from which a sample of firms would be selected to receive a survey related to a number of different aspects of their innovative activity. From this defined population, a balanced sample—balanced in terms of sales and R&D expenditures—of 400 firms (200 manufacturing and 200 service-sector firms) was selected. Each of these 400 firms was surveyed electronically and responses were received from a total of 114. This represents a response rate of 28 percent. 7 The mean value for INDUCE in the Link et al. (1983a) sample of 275 manufacturing firms from either COMPUSTAT or the Register of Corporations is 0.46. 8 This interpretation remains the same when equation (7.11) was estimated using a Tobit specification. 8 Public policies to enhance innovation 1 Much of the introductory material in this chapter draws directly from Link and Scott (2005). See also Feldman et al. (2002) and Link (2006). 2 This section also draws from Link and Scott (2005). It was also presented in Link (2006).

Notes 123 3 The conceptual importance of identifying market failure for policy is also emphasized, although without any operational guidance, in Office of Management and Budget (1996). 4 See Tassey (2005) for a discussion of infratechnologies. 5 149 F.3d 1368 (Fed. Cir. 1998).

Bibliography

Allison, J.R. and Tiller, E.H. (2003) “The business method patent myth,” Berkeley Technology Law Journal, 18: 1–65. Amable, B. and Palombarini, S. (1998) “Technical change and incorporated R&D in the service sector,” Paris: OECD Publication Service. Andersen, B., Howells, J., Hull, R., Miles, I., and Roberts, J. (2000) Knowledge and Innovation in the New Service Economy, North Hampton, MA: Edward Elgar. Antonelli, C. (1989) “A failure-inducement model of research and development expenditure,” Journal of Economic Behavior and Organization, 12: 159–80. Arrow, K.J. (1962) “Economic welfare and the allocation of resources for invention,” in Universities-National Bureau Committee for Economic Research, The Rate and Direction of Inventive Activity, Princeton, NJ: Princeton University Press. Barras, R. (1986) “Towards a theory of innovation in services,” Research Policy, 15: 161–73. Baudeau, N. (1910, originally 1767) Premiere Introduction a la philosophie economique, edited by A. Dubois. Paris: P. Geuthner. Bentham, J. (1952) Jeremy Bentham’s Economic Writings, edited by W. Stark. London: Allen & Unwin. Bentham, J. (1962, originally 1838–43) The Works of Jeremy Bentham, edited by J. Bowring. New York: Russell & Russell. Biotechnology Industry Organization (2002) Biotechnology Industry Organization. 2004. The Editors’ and Reporters’ Guide to Biotechnology, Washington, DC: Biotechnology Industry Organization. Blank, D.M. and Stigler, G.J. (1957) The Demand and Supply of Scientific Personnel, New York: National Bureau of Economic Research. Bush, V. (1945) Science—the Endless Frontier, Washington, DC: U.S. Government Printing Office. Carlin, E.A. (1956) “Schumpeter’s constructed type—the entrepreneur,” Kyklos, 9: 27–43. Charles River Associates (1981) Productivity Impacts of Government R&D Labs: The National Bureau of Standards’ Semiconductor Technology Program—A Production Function Approach, Washington, DC: National Bureau of Standards. Christensen, C.M., Anthony, S.D., and Roth, E.A. (2001) “Innovation in the telecommunications industry: separating hype from reality,” Cambridge, MA: Harvard Business School Working Paper. Clark, K.B. and Griliches, Z. (1984) “Productivity growth and R&D at the business level: results from the PIMS data base,” in Z. Griliches (ed.), R&D, Patents, and Productivity, Chicago, IL: University of Chicago Press.

Bibliography 125 Combs, K.L. and Link, A.N. (2003) “Innovation policy in search of an economic foundation: the case of research partnerships in the United States,” Technology Analysis and Strategic Management, 15: 177–87. COMPUSTAT (2003) Standard Poor’s Compustat, division of the McGraw-Hill Companies. Computer Science and Telecommunications Board—CSTB (2000) Making It Better: Expanding Information Technology Research to Meet Society’s Needs, Washington, DC: National Academy Press. Council on Competitiveness (1991) Gaining New Ground: Technology Priorities for American’s Future, Washington, DC: Council on Competitiveness. Cuneo, P. and Mairesse, J. (1984) “Productivity and R&D at the firm level in French manufacturing,” in Z. Griliches (ed.), R&D, Patents, and Productivity, Chicago, IL: University of Chicago Press. David, P.A., Hall, B.H., and Toole, A.A. (2000) “Is public R&D a complement or a substitute for private R&D: a review of econometric evidence,” Research Policy, 29: 497–529. Dearborn, D.C., Kneznek, R.W., and Anthony, R.N. (1953) “Spending for industrial research, 1951–52,” Cambridge, MA: Harvard University Graduate School of Business report. Denison, E.F. (1979) Accounting for Slower Economic Growth: The United States in the 1970s, Washington, DC: Brookings Institution. Ernst and Young (2002) “Convergence: Ernst & Young’s biotechnology industry report, millennium edition,” New York: Electronic report. Evangelista, R. (1999) Knowledge and Investment, Northampton, MA: Edward Elgar Publishing Inc. Feldman, M.P., Link, A.N., and Siegel, D.S. (2002) The Economics of Science and Technology, Norwell, MA: Kluwer Academic Publishers. Gallaher, M.P. and Rowe, B.R. (2006) “The costs and benefits of transferring technology infrastructures underlying complex standards: the case of IPv6,” Journal of Technology Transfer, 31: 519–44. Gallouj, F. and Weinstein, O. (1997) “Innovation in services,” Research Policy, 26: 537–56. Griliches, Z. (1973) “Research in growth accounting,” in B.R. Williams (ed.), Science and Technology in Economic Growth, New York: John Wiley. Griliches, Z. (1980a) “R&D and the productivity slowdown,” American Economic Review, 70: 343–48. Griliches, Z. (1980b) “Returns to research and development expenditures in the private sector,” in J.W. Kendrick and B.N. Vaccara (eds), New Developments in Productivity Measurement and Analysis, Chicago, IL: University of Chicago Press. Griliches, Z. (1986) “Productivity growth, R&D, and basic research at the firm level in the 1970s,” American Economic Review, 76: 141–54. Griliches, Z. and Lichtenberg, F.R. (1984) “R&D and productivity growth at the industry level: is there still a relationship?” in Z. Griliches (ed.), R&D, Patents, and Productivity, Chicago, IL: University of Chicago Press. Griliches, Z. and Mairesse, J. (1984) “Productivity and R&D at the firm level,” in Z. Griliches (ed.), R&D, Patents, and Productivity, Chicago, IL: University of Chicago Press. Hagedoorn, J., Link, A.N., and Vonortas, N.S. (2000) “Research partnerships,” Research Policy, 29: 567–86. Hall, B.H. (2003) “Business method patents, innovation, and policy,” UC Berkeley Competition Policy Center Working Paper No. CPC03-39.

126

Bibliography

Hall, B.H. and van Reenan, J. (2000) “How effective are fiscal incentives for R&D? A review of the evidence,” Research Policy, 29: 449–69. Halvy, E. (1955) The Growth of Philosophic Radicalism, translated by M. Morris. Boston, MA: Beacon Press. Hauknes, J. (1998) “Dynamic innovation systems: do services have a role to play?” Oslo, Norway: STEP Group report. Hébert, R.F. and Link, A.N. (1988) The Entrepreneur: Mainstream Views and Radical Critiques, Second Edition. New York: Praeger. Hébert, R.F. and Link, A.N. (2006) “The entrepreneur as innovatoz,” Journal of Technology Transfer, 31: 589–97. Higgins, R.S. and Link, A.N. (1981) “Federal support of technological growth in industry: some evidence of crowding-out,” IEEE Transactions on Engineering Management, EM-28: 86–88. Holemans, B. and Sleuwagen, L. (1988) “Innovation expenditures and the role of government in Belgium,” Research Policy, 17: 375–79. Hoselitz, Bert F. (1960) “The early history of entrepreneurial theory,” in J. Spengler and W.R. Allen (eds), Essays in Economic Thought: Aristotle to Marshall, Chicago, IL: Rand McNally. Howells, J. (1999) “Research and technology outsourcing,” Technology Analysis and Strategic Management, 26: 17–29. Howells, J. (2000a) “Services and systems of innovation,” in B. Andersen (ed.), Knowledge and Innovation in the New Service Economy, Northampton, MA: Edward Elgar Publishing Inc. Howells, J. (2000b) “Understanding the new service economy,” in B. Andersen (ed.), Knowledge and Innovation in the New Service Economy, Northampton, MA: Edward Elgar Publishing Inc. Howells, J. (2001) “The nature of innovation in services,” OECD Proceedings Industry, Services and Trade, Paris: Organisation for Economic Co-operation and Development. Jaffe, A.B. (1998) “The importance of ‘spillovers’ in the policy mission of the ATP,” Journal of Technology Transfer, 23: 11–19. Jankowski, J.E. (2001) “Measurement and growth of R&D within the service economy,” Journal of Technology Transfer, 26: 323–36. J.P. Morgan Press Release, October 2, 2002. Kanbur, S.M. (1980) “A note on risk taking, entrepreneurship, and Schumpeter,” History of Political Economy, 12: 489–98. Lecht, G. and Moch, D. (1999) “Innovation and information technology in services.” Canadian Journal of Economics, 32: 363–83. Leyden, D.P. and Link, A.N. (1991) “Why are government and private R&D complements?” Applied Economics, 23: 1673–81. Leyden, D.P. and Link, A.N. (1992) Government’s Role in Innovation, Norwell, MA: Kluwer Academic Publishers. Leyden, D.P. and Link, A.N. (1993) “Tax policies affecting R&D: an international comparison,” Technovation, 13: 17–25. Leyden, D.P., Link, A.N., and Bozeman, B. (1989) “The effects of governmental financing on firms’ R&D activities: a theoretical and empirical investigation,” Technovation, 9: 561–75.

Bibliography 127 Lichtenberg, F.R. (1984) “The relationship between federal contract R&D and company R&D,” American Economic Review, 74: 73–78. Lichtenberg, F.R. (1987) “The effect of government funding on private industrial research and development: a re-assessment,” Journal of Industrial Economics, 36: 97–104. Lichtenberg, F.R. (1988) “The private R&D investment response to federal design and technical competitions,” American Economic Review, 78: 550–59. Lichtenberg, F.R. and Siegel, D. (1991) “The impact of R&D investment on productivity-new evidence using linked R&D-LRD data,” Economic Inquiry, 29: 203–28. Link, A.N. (1980) “Firm size and efficient entrepreneurial activity: a reformulation of the Schumpeter hypothesis,” Journal of Political Economy, 88: 771–82. Link, A.N. (1981a) “Basic research and productivity increase in manufacturing: some additional evidence,” American Economic Review, 71: 1111–12. Link, A.N. (1981b) Research and Development Activity in U.S. Manufacturing, New York: Praeger. Link, A.N. (1982a) “A disaggregated analysis of industrial R&D: product versus process innovation,” in D. Sahal (ed.), The Transfer and Utilization of Technical Knowledge, Lexington, MA: D.C. Health. Link, A.N. (1982b) “The impact of federal research and development spending on productivity,” IEEE Transactions on Engineering Management, EM-29: 166–69. Link, A.N. (1983a) “Alternative sources of technology: an analysis of induced innovations,” Managerial and Decision Economics, 4: 40–44. Link, A.N. (1983b) “Interfirm technology flows and productivity growth,” Economics Letters, 11: 179–84. Link, A.N. (1996) “On the classification of industrial R&D,” Research Policy, 25: 397–401. Link, A.N. (1999) “Public/private partnerships in the United States,” Industry and Innovation, 6: 191–217. Link, A.N. (2005) “Research joint ventures in the United States: a descriptive analysis,” in A. Link and F.M. Scherer (eds), Essays in Honor of Edwin Mansfield, New York: Springer. Link, A.N. (2006) Public/Private Partnerships: Innovation Strategies and Policy Alternatives, New York: Springer. Link, A.N. and Bauer, L.L. (1989) Cooperative Research in U.S. Manufacturing: Assessing Policy Initiatives and Corporate Strategies, Lexington, MA: D.C. Health. Link, A.N. and Rees, J. (1990) “Firm size, university based research and the returns to R&D,” Small Business Economics, 2: 25–32. Link, A.N. and Scott, J.T. (2005) Evaluating Public Research Institutions: The U.S. Advanced Technology Program’s Intramural Research Initiative, London: Routledge. Link, A.N. and Siegel, D. (2003) Technological Change and Economic Performance, London: Routledge. Link, A.N. and Tassey, G. (1993) “The technology infrastructure of firms: investments in infratechnology,” IEEE Transactions on Engineering Management, 40: 312–15. Link, A.N., Tassey, G., and Zmud, R. (1983a) “The induce versus purchase decision: an empirical analysis of industrial R&D,” Decision Sciences, 14: 46–61. Link, A.N., Tassey, G., and Zmud, R. (1983b) “The technology infrastructure of firms: investments in infratechnology,” IEEE Transactions on Engineering Management, 40: 312–15.

128

Bibliography

Link, A.N., Paton, D., and Siegel, D.S. (2002) “An analysis of policy initiatives to promote strategic research partnerships,” Research Policy, 31: 1459–66. Mansfield, E. (1980) “Basic research and productivity increase in manufacturing,” American Economic Review, 70: 863–73. Martin, S. and Scott, J.T. (2000) “The nature of innovation market failure and the design of public support for private innovation,” Research Policy, 29: 437–47. Meyers, J. (2004) “Telephony’s complete guide to WiMAX: the business case for service provider deployment.” http://telephonyonline.com/ar/ telecom_telephonys_complete_guide/ index.htm. Miles, I. and Dachs, B. (2005) “The future of R&D in services: implications for E.U. research and innovation policy,” Report for the European Commission Directorate—General Research. Minasian, J.R. (1962) “The economics of research and development,” in R.R. Nelson (ed.), The Rate and Direction of Inventive Activity: Economic and Social Factors, New York: National Bureau of Economic Research. Mohr, M.F. (1999) “Identification and classification of service projects: phase I of an initiative to create a North American product classification system,” mimeograph. Nadiri, M.I. (1980) “Sectoral productivity slowdown,” American Economic Review, 70: 349–52. Nadiri, M.I. and Schankerman, M.A. (1981) “Technical change, returns to scale, and the productivity slowdown,” American Economic Review, 71: 314–19. National Institute of Standards and Technology (NIST) (2005) Measuring ServiceSector Research and Development, Planning Report 05–1, Gaithersburg, MD: National Institute of Standards and Technology. National Science Board (2004) Science and Engineering Indicators, 2004, Arlington, VA: National Science Foundation. National Science Foundation (1956) Science and Engineering in American dynasty: Final Report on a 1953–1954 Survey, Washington, DC: National Science Foundation. Office of Management and Budget (1996) “Economic analysis of federal regulations under Executive Order 12866,” Washington, DC. Organisation for Economic Co-operation and Development (OECD) (2001a) Innovation and Productivity in Services, Paris: OECD Publication Service. Organisation for Economic Co-operation and Development (OECD) (2001b) Research and Development Expenditures in Industry 1987–1999, Paris: OECD Publication Service. Pavitt, K. (1984) “Sectoral patterns of technical change: towards a taxonomy and a theory,” Research Policy, 134: 343–73. Pilat, D. (2001) “Innovation and productivity in services: state of the art,” in Innovation and Productivity in Services, France: Organisation for Economic Co-operation and Development. Prevezer, M. (1998) “Clustering in biotechnology in the USA,” in M. Prevzer and D. Stout (eds), The Dynamics of Industry Clustering, Oxford: Oxford University Press. Schankerman, M.A. (1981) “The effects of double-counting and expensing on the measured returns to R&D,” Review of Economics and Statistics, 63: 454–58. Scherer, F.M. (1982) “Inter-industry technology flows and productivity growth,” Review of Economics and Statistics, 6: 627–34. Scherer, F.M. (1983a) “The propensity to patent,” International Journal of Industrial Organization, 1: 107–28.

Bibliography 129 Scherer, F.M. (1983b) “R&D and declining productivity growth,” American Economic Review, 73: 215–18. Schumpeter, J.A. (1928) “The instability of capitalism,” Economic Journal, 38: 361–86. Schumpeter, J.A. (1934) The Theory of Economic Development, translated by R. Opie from the 2nd German edition [1926], Cambridge: Harvard University Press. Schumpeter, J.A. (1939) Business Cycles, New York: McGraw-Hill. Schumpeter, J.A. (1950) Capitalism, Socialism and Democracy, Third Edition, New York: Harper & Row. Sirilli, G. and Evangelista, R. (1998) “Technological innovation in services and manufacturing: results from Italian surveys,” Research Policy, 27: 881–99. Soete, L. and Miozzo, M. (1989) “Trade and development in services: a technological perspective,” Maastricht: MERIT Working paper No. 89-031. Solow, R.M. (1957) “Technical change and the aggregate production function,” Review of Economics and Statistics, 39: 312–20. Standard & Poor’s (2002) “Standards & Poor’s biotechnology industry survey,” New York: mimeographed. Sundbo, J. (1997) “Management of innovation in services,” The Services Industries Journal, 17: 432–56. Sundbo, J. and Gallouji, F. (1999) “Innovation in services in seven European countries,” Report for the European Commission. TASC (1998) The Economics of a Technology-based Service Sector, NISY Planning Report 98–2, Gaithersburg, MD: National Institute of Standards and Technology. Tassey, G. (1992) Technology Infrastructure and Competitive Position, Norwell, MA: Kluwer Academic Publishers. Tassey, G. (1997) The Economics of R&D Policy. Westport, CT: Greenwood Publishing Company. Tassey, G. (2005) “Underinvestment in public good technology,” Journal of Technology Transfer, 30: 89–113. Terleckyj, N.E. (1974) Effects of R&D on the Productivity Growth of Industries: An Exploratory Study, Washington, DC: National Planning Association. Terleckyj, N.E. (1982) “R&D and the U.S. industrial productivity in the 1970s,” in D. Sahal (ed.), The Transfer and Utilization of Technical Knowledge, Lexington, MA: D.C. Health. Tether, B., Hipp, C., and Miles, I. (2001) “Standarisation and specialisation in services: evidence from Germany,” Research Policy, 30: 1115–38. Thünen, J.H. von (1960) The Isolated State in Relation to Agriculture and Political Economy, vol. 2, translated by B.W. Dempsey, in The Frontier Wage, Chicago, IL: Loyola University Press. Toole, A.A. (2003) “Understanding entrepreneurship in the U.S. biotechnology industry,” in D.H. Hart (ed.), The Emergence of Entrepreneurship Policy: Governance, Start-Ups, and Growth in the U.S. Knowledge Economy, Cambridge: Cambridge University Press. US Department of Commerce (2003) “A survey of the use of biotechnology in U.S. industry,” Washington, DC: U.S. Department of Commerce. US Department of Commerce, Technology Administration (1990) Emerging Technologies: A Survey of Technical and Economic Opportunities, Washington, DC: US Department of Commerce.

130

Bibliography

US Patents and Trademark Office (1999) “White paper on automated financial or management data processing methods (business methods),” http://www.uspto.gov/ web/menu/busmethp/index2.htm. von Hippel, E. (1988) The Sources of Innovation, New York: Oxford University Press. Wallsten, S. (1999) “The effects of government-industry R&D programs on private R&D: the case of the small business innovation research program,” Rand Journal of Economics, 31: 82–100. Weber, M. (1930) The Protestant Ethic and the Spirit of Capitalism, translated by T. Parsons. New York: Scribner’s. Zrinyi, J. (1962) “Entrepreneurial behavior in economic theory: an historical and analytical approach,” PhD dissertation, Georgetown University.

Index

A.B. Watley Group Inc. 66 Advanced Development Groups (ADGs) 54 Advanced Technology Program 115 Alliance for Telecommunications Industry Solutions (ATIS) 40 Anglo American PLC-ADR 63 ANS X9.85 project 62 Antonelli, C. 109 applied research 4, 6, 14–19, 23, 36, 45, 54–55, 67, 75–78, 84, 87, 91–94, 107–08, 121n.6 (Chap.4), 121n.1 (Chap.5) Arrow, K.J. 113 automatic teller machines (ATMs) and internet 57

Clinical Data Interchange Standards Consortium (CDISC) 93 cloning 94 Combs, K.L. 99 Computer Science Telecommunications Board (CSTB) 69, 73 contract management 27 contract manufacturing organizations (CMOs) 84–85, 89–90; R&D activities in 94–95 contract research organizations (CROs) 84–85, 89–91; R&D activities 92–94 CORE (COoperative REsearch) database 100 Covance 90 Cuneo, P. 106 customization 13

Barras’s reverse product cycle 9 basic research 4, 6, 14–18, 46, 54, 67, 75–79, 86, 91–92, 107–08, 110, 115; definition 120n.3 (Chap.2) Baudeau, N. 22, 24 Bauer, L.L. 110 Bentham, Jeremy 25–28 BioMachines 90 biopharmaceutical industry 84–85; innovation process in 86; see also drug development bioprocessing technology 94 biosensors 95 biotechnology industry 82–84 Blank, D.M. 108

David, P.A. 109 Dearborn, D.C. 120n.3 (Chap.2) Denison, E.F. 106 Department of Commerce (DOC), US 84 depository institutions 52 development: definition 14, 16; drug development process 85–92; large-sized network operators, development activities by 49; process development 89–90; product development 87–89; small-sized network operators, development activities by 46; software development 42–43; Wi-Fi services, development supply chain for 43; see also under R&D Direct-Access Vertical Exchange (DAVE) 66 disaggregate R&D studies 107–08 dispatch services 40 distributed innovation process (DIP) 10 diversified firms 69–73, 76; R&D activities in 78–79; services-only versus diversified firm’s R&D activities 77

Cantillon, Richard 24 cell culture 94 Charles River Associates 101 Clark, K.B. 106 client relationship management (CRM) 60

132

Index

drug development 85–87; process 85–90; process development 89–90; product development 87–89; R&D in 90–92; stages 88; supply chain 85; see also biopharmaceutical industry Electronic data capture (EDC) 93 entrepreneurship 19, 24–25, 28–29, 33 enterprise resource management (ERM) 78 European Union (EU) 18 Evangelista, R. 10 Feldman, M.P. 107, 122 nn.1, 2 (Chap.8) financial services industry 51–68; ATMs and internet 57; defining 53–54; diversity in 63–64; firms with significant R&D expenditures (2003) 65; industry profile and R&D statistics 51–52; innovation process 56–58; investment services 54; R&D activity 54–56; R&D activity in security and commodity broker services 60–62; R&D metrics for 58–60; retail banking 54; ten largest firms in 2003 64; virtual desktop 57–58; web services technology, case study 53–60 Financial Services Technology Consortium (FSTC) 61–62, 67 Food and Drug Administration (FDA) 83, 85 Frascati Manual 18 Gallouj, F. 9 general packet radio service (GPRS) 40 Global Services 79 Griliches, Z. 104–06, 108 Hagedoorn, J. 99 Hall, B.H. 107, 118 Hébert, R.F. 121n.8 Higgins, R.S. 109 Holemans, B. 109 Howells, J. 120n.1 (Chap.2) hybrid technology 40 IBM equipment technology 10, 69–72, 74–75, 78–79 IGEN Inc. firm 64 immediate service termination (IST) 40 information flows, for systems integration services industry 72–73 infrastructure technologies 8, 19–23, 36, 39, 40–44, 47, 50, 55, 61–62, 69, 71, 75, 100–01, 110, 114–15, 121n.3 (Chap.3) InGenium Research Inveresk 90

innovation and productivity growth, dimensions 97–110; of manufacturing and service-sector firms 97–100; public policies to enhance 111–19; R&D spending in 97–99; research joint ventures 99–100; see also separate entry Institute of Electrical and Electronics Engineers (IEEE) 41–43, 45 intellectual capital (IC) 54, 57–59, 63, 68, 70, 72, 74–80, 95–96, 119 Internet protocol (IP) 39–40 Investigational New Drug (IND) 87, 89 investment services 54; versus retail banking 55 Investment Technology Group Inc. 66 iso-technology curve 59–60 Jaffe, A.B. 113 Jankowski, J.E. 120n.4 Joint Research and Development Act of 1984, 122n.3 (Chap.7) JP Morgan Chase 68 Kanbur, S.M. 28 knowledge-intensive business services (KIBSs) 10 Lecht, G. 120n.4 LendingTree.com Inc. 66 Leyden, D.P. 107, 109–10 Lichtenberg, F.R. 106, 108–09 Link, A.N. 20, 99, 103, 104, 107–10, 121n.4 (Chap.4), 121n.8, 122n.2 (Chap.7), 122n.1 (Chap.8), 122n.2 (Chap.8) local area network (LAN) 39 local multipoint distribution service (LMDS) 40 Lynch, Merrill 61 Mairesse, J. 106 Mansfield, E. 104, 108 manufacturing firms: innovation in 18–21; innovation in, model 19; and nonmanufacturing firms, R&D statistics for 98; public good characteristics of 20; and service firms, comparison 3; and service-sector firms, innovation characteristics 97–100 Martin, S. 112 microarrays 95 Minasian, J.R. 105 Mintel International Group Ltd 60

Index 133

Quintiles Transnational 90

and productivity growth 97–9; see also individual entries R&D activities, in systems integration services industry 73–79; break\fix maintenance 75, 78; building new systems for clients 75, 78; categories 75; in diversified integrators 78–79; maintenance of code 75; in service-only integrators 77–78; services-only versus diversified firm’s R&D activities 77–79; support research 75, 78; systems configuration 78 R&D and productivity growth framework 101–05; 1950s and 1960s 105; disaggregate R&D studies 107; early 1970s 106; post-slowdown policy 107; regression results 104; theoretical model 101–05 R&D in service sectors 12–18; alternative definitions 16–17; classical definitions 12; customization 13; in financial services industry 51–52, 54–56; inflation-adjusted 14; representative systems integrators (1998), investments of 70; in security and commodity broker services 60–62; in systems integration services industry 73–4; of telecommunications industry 36–39; in wireless communications services 39–41; see also RD&T service industry RD&T service industry 81–96; biopharmaceutical industry 84–85; biotechnology industry 82–84; industry profile and R&D statistics 81–82; largest RD&T firms, by R&D expenditures 82 recombinant DNA technology 94 Rees, J. 99, 110 Research and Development Joint Venture Act of 1983, 122n.3 (Chap.7) research joint ventures (RJVs) 99–100; membership in 100 retail banking 54, 67; versus investment services 55

R&D activities: of CMO 94–95; of CRO 92–94; and drug development 90–92; identifying and measuring 117–18; measurement issues and implications for public policy 116–19; patenting servicesector innovations 118; R&D capital stock model 101; R&D-to-productivity growth relationship, empirical evidence 105; social and private rates of return to 113; spending, and innovation

Schankerman, M.A. 103, 106 Scherer, F.M. 103, 108 Schmoller, Gustav 30 Schumpeter, J.A. 19, 25, 31–35 Schumpeterian hypothesis 98 science based sectors 9, 18–23, 100, 110 Scott, J.T. 112, 122 nn.1, 2 (Chap.8) SEI Investments Company 63 service only firms 69–70, 73; R&D activity in 77–8

Miozzo, M. 9 Moch, D. 120n.4 Nadiri, C. 106 nanobiotechnology 95 National Cooperative Research Act (NCRA) 99, 116 National Cooperative Research and Production Act (NCRPA) 99 National Science Foundation (NSF) 4, 13–14, 17 network domain security (NDS) 40 network operator 42–50 New Drug Application (NDA) 89 Niteo Partners 67 Nobex 90 nondepository institutions 52 North American Industry Classification System (NAICS) 3, 36–38, 51–52, 64–66, 81, 83, 121n.1 (Chap.3) operation support system (OSS) 44 original equipment manufacturers (OEMs) 43–50, 121n.2 (Chap.4) Panopticon 26–27 Pavitt, K. 9 Personal communications services (PCS) 40 Pilat, D. 120n.4 Post-slowdown policy 107 Preeminence Act of 1991 115 Prevezer, M. 83 productivity growth 6, 9 product life-cycle management (PLM) 77 proprietary technology 9, 19–22, 58, 61, 66, 74, 76, 91–95, 100, 105, 116 protein engineering 94 public policies, to enhance innovation 111–19; government’s role in 111–14 Public Switched Telephone Network (PSTN) 39

134

Index

service provider 13, 36, 38–45, 48, 50, 63, 71, 85, 91, 117 service-sector innovations 6–35; in biopharmaceuticals over several phases of R&D 86; entrepreneur as innovator 23–25; and entrepreneurial activity 19; in financial services industry 56–58; innovation related characteristics 3; in manufacturing and service sectors, models 18–23; in manufacturing sector firms 18–21; in market 12; measurement 11–12; model 21; in organizational 12; in process 12; in product 12; public good characteristics 20; public policies to enhance 111–19; Schumpeter’s description 33–35; in service-sector firms 21–23; in systems integration industry 71–77; taxonomies 7–11; see also separate entry service sectors: industrial and institutional composition 2; innovation in 6–35, see also separate entry; R&D in 12–18, see also separate entry; systems integration in 8; technological maturity 10; see also financial services industry; systems integration services industry; telecommunications industry Siegel, D. 106, 108 Sirilli, G. 10 Sleuwagen, L. 109 Smith, Adam 26 Soete, L. 9 Solow, R.M. 101 Sombart, Werner 30–31 Standard Industrial Classification (SIC) 64 Standards Development Organization Advancement Act of 2004 (SDOAA) 99 Stanley, Morgan 61 Stigler, G.J. 109 Sundbo, J. 10 systems integration services industry 69–80; information flows for 72; innovation in 71–77; labor versus intellectual capital 75–77; production spectrum of 76; R&D activities 73–74; R&D investments of representative systems integrators (1998) 70; systems integration industry 69–70; systems integration services 70–71; see also separate entry Tassey, G. 18, 20, 110, 113, 121n.4 (Chap.2) 123n.4 technological change 7–11, 109

technology supply chain 41, 49 telecommunications industry 36–50; activity categories and research taxonomies 49; dispatch services 40; hybrid technology 40; improved accessibility 41; industry profile and R&D statistics 36–39; reliability 41; security 40–41; service providers 38; Wi-Fi/W-LAN 40; in wireless communications services 39–41; wireless data and internet access 40; see also separate entry Telecommunications Reform Act, US, 1996 39 Terlecky, N.E. 106–08 3G technology 40 total factor productivity 106, 110 universal value exchange (UVX) 62 US Patents and Trademark Office (USPTO) 118 US service sectors see service sectors van Reenan, J. 107 virtual private networks (VPNs) 48, 58 Voice over IP (VoIP) technology 39 Von Thünen, J.H. 28–30 Wallsten, S. 109 Wealth Management Workstation (WMW) 61 Weber, Max 30–31 WebFountain 79 web services technology, case study 53–60, 67–8 Weinstein, O. 9 Wi-Fi/W-LAN 40–49; additional supply chain participants 45–49; aggregators 44–5; case study 41–49; definition 42; description 41–43; development supply chain for 43; equipment manufacturers and service industry description 43; large-sized network operators, development activities by 49; network operators 44; small-sized network operators, development activities by 46 wireless communications services, R&D activities 39–41 wireless data and internet access 40 Wireless Internet service provider (WISP) 42 Zen-Bio 90