Dairy Policy Reform And Trade Liberalisation

%BJSZ1PMJDZ3FGPSNBOE 5SBEF-JCFSBMJTBUJPO ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT 03("/*4"5*0/'0...

Author: Pavel Vavra | Roger Martini

21 downloads 708 Views 907KB Size Report

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

Report copyright / DMCA form

DOWNLOAD PDF

%BJSZ1PMJDZ3FGPSNBOE 5SBEF-JCFSBMJTBUJPO

ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT

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

5IJTXPSLJTQVCMJTIFEPOUIFSFTQPOTJCJMJUZPGUIF4FDSFUBSZ(FOFSBMPGUIF 0&$%5IF PQJOJPOTFYQSFTTFEBOEBSHVNFOUTFNQMPZFEIFSFJOEPOPUOFDFTTBSJMZSFGMFDUUIFPGGJDJBM WJFXTPGUIF0SHBOJTBUJPOPSPGUIFHPWFSONFOUTPGJUTNFNCFSDPVOUSJFT

¦0&$% /PSFQSPEVDUJPO DPQZ USBOTNJTTJPOPSUSBOTMBUJPOPGUIJTQVCMJDBUJPONBZCFNBEFXJUIPVUXSJUUFOQFSNJTTJPO"QQMJDBUJPOTTIPVMECFTFOUUP 0&$%1VCMJTIJOH SJHIUT!PFDEPSHPSCZGBY 1FSNJTTJPOUPQIPUPDPQZBQPSUJPOPGUIJTXPSLTIPVMECFBEESFTTFEUPUIF$FOUSF GSBO¡BJTEhFYQMPJUBUJPOEVESPJUEFDPQJF SVFEFT(SBOET"VHVTUJOT 1BSJT 'SBODF DPOUBDU!DGDPQJFTDPN

3

FOREWORD This report is one of several studies that have been carried out under the Programme of Work of the OECD’s Committee for Agriculture, and in particular in the context of the activity on Assessing Future Agricultural Markets, Trade and Policies. Within this activity, studies were scheduled to provide assessments of the market, trade and welfare impacts of domestic and trade policy reform for selected commodities that currently receive very high support and protection. The report is a collection of three individual studies discussed and declassified by the Joint Working Party on Agriculture and Trade. The first two chapters analyse the trade and economic effects of the main policy measures applied to the dairy sector; specifically, the effects of milk price support measures and of milk quota systems. The support measures are then removed in the last chapter in order to assess the impact of international dairy trade liberalisation on production, consumption, trade, prices, income and welfare. The main authors are Pavel Vavra and Roger Martini, economists in the OECD Directorate for Agriculture. Joe Dewbre (OECD Directorate for Agriculture) and Nobunori Kuga, now of the Ministry for Agriculture, Forestry and Fisheries in Japan, contributed substantially in the early phases of the work on milk price support measures. Many colleagues in the OECD Secretariat and delegates from member countries furnished useful comments on earlier drafts of this report.

This page intentionally left blank

5

Table of Contents Chapter 1 TRADE AND ECONOMIC EFFECTS OF MILK PRICE SUPPORT MEASURES .................... 9 Introduction………….. ......................................................................................................................... 11 Milk price discrimination in OECD countries....................................................................................... 12 A stylised model of dairy pricing and trade........................................................................................... 17 Results of policy simulation analysis .................................................................................................... 21 Conclusions ........................................................................................................................................... 30 Annex 1.1 Algebraic version of graphical model .................................................................................. 35 Annex 1.2 Simulated effects of alternative pricing arrangements on dairy product markets – detailed Aglink results......................................................................................................................................... 41 References ............................................................................................................................................. 44

Chapter 2 TRADE AND ECONOMIC EFFECTS OF MILK QUOTA SYSTEMS ....................................... 47 Introduction ........................................................................................................................................... 49 Quota programmes – General overview ................................................................................................ 50 Theory of milk production quotas - Overview ...................................................................................... 52 Quota interactions with other policy objectives .................................................................................... 56 Modelling of milk supply in the presence of quota ............................................................................... 73 Conclusions ........................................................................................................................................... 76 Annex 2.1 Welfare implications of support price reduction versus quota level imposition.................. 82 Annex 2.2 Incentive to exchange quotas and emergence of quota value .............................................. 84 Annex 2.3 Long-run effects of quota imposition on farm assets at the farm level................................ 86 Annex 2.4 The development of the Norwegian milk quota system....................................................... 87 References ............................................................................................................................................. 91

Chapter 3 ANALYSIS OF INTERNATIONAL DAIRY TRADE LIBERALISATION................................. 95 Introduction ........................................................................................................................................... 97 The world dairy markets........................................................................................................................ 97 Agricultural policies supporting dairy production............................................................................... 102 International dairy trade liberalisation................................................................................................. 105 Summary.............................................................................................................................................. 128 Main conclusions................................................................................................................................. 131 Annex 3.1 Aglink results of dairy trade liberalisation scenarios – evaluating the market and trade impacts........................................................................................ 136 Annex 3.2 Sensitivity analysis............................................................................................................. 148 Annex 3.3 Main policy assumptions for dairy markets ....................................................................... 154 Annex A A DESCRIPTION OF THE AGLINK AND PEM MODELS....................................................... 159

6 Tables 1.1. Selected indicators of trade effects for comparing policy regimes................................................. 20 1.2. Average percentage changes of base scenario milk prices resulting from changes in manufacturing milk prices (LSP) and an assumed fluid milk premium (LFP)...................................................... 23 1.3. Production and consumption impacts of reductions in manufacturing milk prices and an assumed fluid milk premium, averages for 2002-2008 ................................................................................ 24 1.4. Indicators of economic costs and benefits in the PEM dairy model............................................... 27 1.1.1. Ranges of values for key components of formulas determining relative trade effects ................ 40 1.2.1. Federal Milk Order Principal Pricing Points, with Class I Differential....................................... 41 1.2.2. Percentage differences between the respective scenarios and the baseline (Consumption) ........ 42 1.2.3. Percentage differences between the respective scenarios and the baseline (Production) ............ 42 1.2.4. Percentage differences between the respective scenarios and the baseline (World dairy prices) 42 2.1. Impacts of quota increases on key variables (average changes from baseline for the EU) assuming constant government expenditures on subsidised exports ............................................................. 59 2.2. Impacts of quota increases on key variables (average changes from baseline for the EU) assuming constant volume of subsidised exports .......................................................................................... 60 2.3. Impact of a 1% increase in quota in the EU ................................................................................... 63 2.4. Doubling import quantity in Canada .............................................................................................. 70 2.5. Elasticity used in Aglink supply functions for Canada and the European Union........................... 75 2.6. Milk quota rent assumptions for the European Union (per cent of price) ...................................... 76 3.1. Unilateral liberalisation experiments............................................................................................ 123 3.2. Ranking of impacts of unilateral liberalisation experiments ........................................................ 123 3.3. Simultaneous liberalisation experiment........................................................................................ 125 3.4. Partial price transmission in the EU ............................................................................................. 127 3.1.1. Dairy policy reforms – Impact on the European Union............................................................. 137 3.1.2. Dairy policy reforms – Impact on Canada................................................................................. 138 3.1.3. Dairy policy reforms – Impact on the United States ................................................................. 139 3.1.4. Dairy policy reforms – Impact on Japan.................................................................................... 140 3.1.5. Dairy policy reforms – Impact on Mexico................................................................................. 141 3.1.6. Dairy policy reforms – Impact on Australia .............................................................................. 142 3.1.7. Dairy policy reforms – Impact on New Zealand ....................................................................... 143 3.1.8. Dairy policy reforms – Impact on Argentina............................................................................. 144 3.1.9. Dairy policy reforms – Impact on Brazil ................................................................................... 145 3.1.10. Dairy policy reforms – Impact on Rest of World .................................................................... 146 3.1.11. Dairy policy reforms – World.................................................................................................. 147 3.2.1. Sensitivity analysis of milk supply elasticity assumptions – results for the European Union... 148 3.2.2. Sensitivity analysis of milk supply elasticity assumptions – results for Canada ....................... 149 3.2.3. Sensitivity analysis of milk supply elasticity assumptions – World dairy prices ...................... 150 3.2.4. Sensitivity analysis of quota rent assumptions – results for the European Union ..................... 151 3.2.5. Sensitivity analysis of quota rent assumptions – results for Canada ......................................... 152 3.2.6. Sensitivity analysis of quota rent assumptions – World dairy prices ........................................ 153 3.3.1. Main policy assumptions for dairy markets (Agricultural Outlook 2003-2008) ....................... 154

7 Figures 1.1. Market effects of alternative milk price support measures............................................................. 18 1.2. Simulated impacts of reductions in manufacturing milk prices and an assumed fluid milk premium on US net import of butter, cheese and skimmed milk powder (changes from baseline) ............. 25 1.3. Simulated impacts of reductions in manufacturing milk prices and an assumed fluid milk premium on world market prices .................................................................................................................. 25 1.4. Estimated trade impacts of reductions in manufacturing milk prices and an assumed fluid milk premium......................................................................................................................................... 27 1.5. Simulated impacts of reductions in manufacturing milk prices and an assumed fluid milk premium on selected indicators of the economic benefits and costs of support ........................................... 28 1.1.1. Relative trade effects of milk price support measures................................................................. 38 1.2.1. Net exports for butter .................................................................................................................. 43 1.2.2. Net exports for cheese ................................................................................................................ 43 1.2.3. Net exports for SMP ................................................................................................................... 43 1.2.4. Net exports for WMP .................................................................................................................. 43 2.1. Quota imposition favours farm owners at the expense of input suppliers ..................................... 54 2.2. Interaction between quota level, domestic price, exports and government expenditures .............. 57 2.3. Relationship between price and quota holding export subsidies constant ..................................... 66 2.4. Income-compensating payments shift supply ................................................................................ 67 2.5. Impact on welfare of an increase in imports .................................................................................. 69 2.6. Domestic industrial milk quota and price response to increase imports ........................................ 72 2.1.1. The welfare implications of different policy options available to reduce large surpluses .......... 83 2.2.1. Development of a quota market and a value of quota ................................................................ 84 2.3.1. Long-run effects of a quota imposition at the farm level ............................................................ 86 3.1. Comparison of per capita consumption with per capita production of milk (in milk equivalents) 99 3.2. Production trends of major dairy products .................................................................................... 99 3.3. Consumption trends of dairy products in OECD and non-member economies ........................... 100 3.4. Trends in trade volumes and market shares for major dairy products and exporters .................. 101 3.5. Declining world dairy prices in real terms ................................................................................... 102 3.6. OECD average Producer Support Estimate for milk, 1986-2002................................................. 103 3.7. Producer Support Estimate for milk by country, 1986-88 and 2000-02 ...................................... 104 3.8. Impact of trade liberalisation on world dairy product output ....................................................... 118 3.9. Comparison of milk price and quantity adjustments in individual scenarios: The case of Canada ..................................................................................................................... 119 A.1.1.PEM Land Allocation Structure ............................................................................................... 163

This page intentionally left blank

9

CHAPTER 1 TRADE AND ECONOMIC EFFECTS OF MILK PRICE SUPPORT MEASURES

Abstract The analysis reported in this chapter estimates and compares market and trade effects of alternative milk price support measures, namely i) support due to trade measures – i.e. support attributable to a package of target prices, product support prices and measures at the border and ii) support due to domestic discriminatory pricing. Discriminatory pricing here refers to a government administered or sanctioned practice that leads to domestic market prices paid for some end-uses of raw milk (typically fresh milk products) to be higher by more than the additional marketing costs than prices paid for other raw milk end-uses (typically manufactured milk products). Under a price discrimination scheme, domestic fluid milk consumers partly support farmers through a higher fluid milk price. For the same ‘target’ producer support price, this enables a government to set support measures for the tradable manufacturing milk at a lower level in comparison to a government that does not operate price discrimination and price pooling. Market price support, whether the consequence of trade interventions in dairy product markets or the consequence of discriminatory pricing arrangements, leads to increased production (unless there are quota restrictions) and reductions in consumption, when compared to a situation without such support. In most cases, this will result in higher exports and lower imports. The mathematical analysis shows that on a dollar for dollar basis it is theoretically possible for milk price support resulting from discriminatory pricing to be as, or even more, trade distorting than milk price support resulting from explicit trade intervention in dairy product markets. The empirical results of the assumed change in policy parameters show that, as a practical matter, price-supporting discriminatory pricing arrangements are usually less trade distorting. However, compared to support due to trade measures, the price discrimination scheme imposes a higher burden on one particular group, fluid milk consumers.

This page intentionally left blank

11

Introduction In almost every OECD country, milk producers receive higher prices because governments intervene in the markets for raw milk and dairy products. Estimated rates of milk market price support are among the highest of all commodities monitored for the PSE. Governments intervene to obtain higher producer prices for raw milk using a package of mutually reinforcing domestic and trade policy measures. The typical package includes: 1) a target price for raw milk, 2) support prices for manufactured dairy products necessary to achieve that target price and 3) tariffs, tariff-rate quotas and export subsidies applied to imports or exports of tradable dairy products to defend the support prices. In a few countries, producer prices may be further enhanced using extra revenues generated via discriminatory pricing on the domestic market. In some countries, marketing quotas are used with other mechanisms to maintain market price support. To facilitate the presentation in this report, that part of milk price support attributable to a package of target prices, product support prices and trade measures will be referred to frequently as “support due to trade measures”. That part of milk price support that may be attributable to discriminatory pricing arrangements will be referred to simply as “support due to discriminatory pricing”, even though such arrangements themselves may require accompanying border measures. Both kinds of price support impose extra costs on domestic consumers and taxpayers and they distort trade leading to lower world market prices. The purpose of the analysis is to estimate and compare the effects of support due to trade measures with that of discriminatory pricing. Price support achieved through trade measures applied to tradable dairy products, e.g. butter, skimmed and whole milk powder, and cheese, results in domestic prices for those products that are higher than the corresponding world market prices for them. This drives up the prices dairy plants are willing to pay for the raw milk used to make those products which, through competitive domestic market price determination, then leads to higher prices paid for milk for all end uses. Discriminatory pricing arrangements, administered or sanctioned by the government, lead to prices paid for raw milk for some end uses (typically fresh milk products) that are higher than those paid for raw milk for other end uses (typically manufactured milk products). The additional revenue generated is then transferred back to farmers through a pooled or average price scheme. Generally speaking, in countries where governments intervene both in traded dairy product markets and via discriminatory pricing on the domestic market the overall level of support may be reduced, say, either by reducing trade interventions or by reducing the premium (and any associated tariff) charged domestic fluid milk consumers. The analysis to be reported in this chapter comprised comparisons of the effects of marginal changes in one or the other of these two types of intervention. The remainder of the chapter is organised as follows. The next section describes some general characteristics of milk price discrimination arrangements and their potential effects. It includes a brief review of past work. In the second major section, a stylised model of milk pricing and policy is used to derive some general, qualitative results concerning differences in the market and trade effects of alternative milk price support measures. The third section presents results of illustrative policy simulation experiments undertaken using the Aglink and PEM models to quantify policy effects on production, demand, trade and welfare. The final section of this chapter draws conclusions.

12

Milk price discrimination in OECD countries Price discrimination refers to the practice of selling the same product to different buyers at different prices. This can lead to an increase in market receipts if buyers can be segregated into distinct groups in which those least responsive to price (i.e. those with the lowest price elasticity of demand) are charged the highest price. Segregating consumers and charging them different prices is possible of course only if the seller – whether a private company, co-operative, government agency or quasi-government institution – has market power. In some countries, the government sets prices for different end-uses of milk by administrative fiat. Price differences between those various end-uses reflect inter alia the additional costs to deliver milk to fluid plants over the costs to market milk to manufacturing plants. In others, price premiums and discounts by end-use are determined by a state-trading agency or by a marketing institution (for example a co-operative) granted monopoly power by the government. The way buyers are segregated may also be different in different countries. The most common, and the main focus of this chapter, is an arrangement under which domestic consumers are grouped in different demand categories. In other cases the pricing arrangements may lead to differences in prices charged across export markets

Market-driven explanations for milk price differences This study is concerned with discriminatory pricing arrangements that cause the prices paid for raw milk purchased to be re-sold as fresh fluid milk to be higher than the price of milk purchased to manufacture dairy products. It is important to begin, however, by noting that some differences in the prices paid for fluid versus manufacturing milk might exist even if governments did not intervene at all. Generally, milk destined for fluid use has to meet more stringent sanitary standards than milk used for manufacturing dairy products. Historically, this difference in quality standards explained some differences in between prices of fluid and manufacturing milk. Today, however, most raw milk production in OECD countries would meet quality standards for fluid milk regardless of the end use for which it is purchased. It follows that, in the absence of any differences in quality and sanitary standards, raw milk sales in a particular region will be reallocated according to dairy processors bids so that the price of fluid milk and manufacturing milk will move towards convergence.1 Average prices for milk going to fluid use and milk going to manufacturing use (annual averages for example) are, amongst other factors, driven by consumer preferences for fresh – not reconstituted – fluid milk products, transportation and other marketing cost differences, and the daily, weekly, and seasonal patterns of milk production and consumption. Typically, in OECD countries, consumers prefer fresh milk to reconstituted milk from milk powders despite the potential cost advantage of the latter. Hence, consumer preference for fresh milk is a fundamental element of market demand. Supplying fresh fluid milk to meet consumer demand requires a higher price for milk and marketing services in regions of the country with insufficient milk supplies (Takayama and Judge, 1971). It follows that the higher price for fluid milk is to a large extent linked to transportation costs and seasonality of production issues which will be discussed in turn.

13

Transportation and other marketing costs Milk is a bulky, highly perishable commodity subject to bacterial contamination and, as such, cannot be stored on farms for any significant period of time. Raw milk must be handled under strict sanitary conditions and must be marketed quickly either for use in a fluid form or for processing into one of a wide range of storable manufactured dairy products. The considerable costs of transporting bulk raw milk suggests that the majority of milk to be consumed in fluid form will likely be produced and processed relatively close to the point of consumption. However, the costs of producing raw milk relatively close to the point of consumption, mostly near major urban centres, may be higher than the costs of producing milk elsewhere. The cost of the hired labour, animal feed, forage and land used in producing raw milk are typically higher in these more densely populated urban areas than in rural agricultural regions of a country. Because it is cheaper to transport manufactured dairy products than it is to transport raw milk, such production cost advantage may favour location of dairy manufacturing plants away from consuming centres. (Consider that the production of one kilogram of butter requires about 21 kg of milk and the production of 1 kg of cheese requires about 10 kg of milk.) The combination of higher transportation costs for raw milk, lower transportation cost for dairy products (in milk equivalent) and regional differences in production, may create a natural segregation of fluid milk and manufacturing milk markets creating a premium for those producers located nearby major consuming areas even in an unregulated market.2 To account for transportation costs in an analysis of milk pricing, a model addressing spatial variation in prices would need to be adopted. For example, the work of Pratt et al. from Cornell University (1998) gives an excellent insight into spatial and temporal variation of prices for raw milk, and for milk delivered to plants in different uses. Their model, the US Dairy Sector Simulator Model (USDSS), is a highly disaggregated network flow model that minimises the transportation and processing costs of transforming milk from farm centres into fluid milk and manufactured dairy products at consumption centres. The structure of the mathematical model is such that the shadow prices at each location are interpreted as relative spatial value. Thus, the relative spatial values of milk in fluid use, soft product uses, butter, non-fat dry milk and cheese in major consumption centres are solved for milk at every plant location, and for farm milk at each farm production centre as well. The model generates a fluid milk price surface that ranges up to USD 4.00 per hundredweight, and yielding a national weighted-average fluid milk price differential of about USD 2.47 per hundredweight (cwt) over manufacturing milk price levels. The USD 2.47 included USD 1.32 attributable to transportation costs alone, physically moving the milk from farms to plants. A basic differential of USD 1.15 was included as well. This basic differential was related, inter alia, to costs associated with timing and scheduling milk deliveries for fluid plants in addition to the simple average trucking costs between a farm and a plant. These costs include the need for additional quality control measures needed for fluid milk, additional storage facilities and adequate trucking capability to meet fluid milk processing schedules, and manufacturing capacity for milk when it is not needed in fluid use. The transportation cost issue has been incorporated and analysed in a number of spatial equilibrium model studies (Pratt et al, 1998, McDowell et al. 1988, McDowell et al. 1990, Nubern and Kilmer 1997, Cox and Chavas 2001). These studies all find that the presence of non-zero transportation costs generates market driven differences in manufacturing and fluid milk prices. For example, McDowell et al. (1988) estimated that, in the United States, market-driven differentials (per hundred-weight of raw milk) began

14 reflecting interregional transportation costs in the Kentucky-Tennessee region at USD 1.56 and in the Southern Plains region at USD 1.90, increasing to USD 4.08 in Florida. Kawaguchi, Suzuki, and Kaiser (2001) develop an annual interregional trade model that, under assumptions of perfect competition, generated fluid premiums averaging 0.32 with premiums of USD 1.55 and USD 1.95 per cwt generated in the Southeast and in Florida, respectively. This estimation used an annual model with interregional transportation costs, reflecting no seasonality of supply and demand, including no storage and transportation costs incurred by producer cooperatives to deliver milk as scheduled by fluid processors, and by assuming equal supply elasticities in all regions.

Seasonality of milk production Milk is produced continuously throughout the year, typically with marked seasonal variations. Milk production increases during the spring and early summer and contracts during the autumn and winter. In the high producing months the market price of raw milk usually falls somewhat. This is also the season when most of the extra raw milk production is being channelled into processing of manufactured dairy products. During the low producing months unit costs of raw milk production and market prices may rise significantly (both because productivity per cow falls and because feed and other costs of raw milk production may increase). Sending a smaller proportion of raw milk production for manufacturing dairy products accommodates production shortfalls. In sum, over the season, more milk will be sold at lower prices for processing into manufactured dairy products resulting in an annual average price paid for fluid milk greater than that paid for manufacturing milk. The amount of seasonal premium will be larger in those countries where the cost of production between the in-season and off-season period is greater. An extreme example is the case of New Zealand. Milk production in New Zealand is based on grazing with very low production cost during peak season, so dairy farmers follow the so-called curve of nature. Usually during the off-season months (about two months of the year) cows do not produce milk and manufacturing facilities substantially reduce their production or may shut down altogether. The farmers remaining in production to supply the fluid milk market are paid an out-of-season premium by the co-operative, a premium that can amount to as much as 60% of the milk price in the in-season period. However, as the domestic consumption of milk in New Zealand is only a fraction of “in-season” production the difference between the average price and the manufacturing milk price is negligible. In general, it could be said that the higher the share of the fluid milk market in a particular country, the higher might be the impact of seasonal premiums on the fluid and manufacturing milk prices. On the other hand, if the levels of production are not very different between off and in-season, likely owing to fairly constant cost structure regardless of the season, then the differences in prices over the year might not be very large. The seasonal variation in production has sometimes been used to justify milk market regulation; government intervention is deemed a critical element in order to stabilise prices and ensure adequate supplies of fluid milk at all times. Ippolito and Masson (1978) pointed out however, that there is a cost to ensuring “adequate” milk supplies by means of a constant fluid milk premium. These authors note that the fluid milk premium has to be set so as to ensure an adequate supply of milk during low-producing months. However, this price will far exceed the free market price in high-producing months and

15 thus consumers will have to pay an average price of fluid milk that is higher than it would otherwise be. In addition, the constant premium would lead to a higher producer price and consequently higher than otherwise production of processed dairy products in peak production months. Testuri, Kilmer, and Spreen (2001) using the model developed by Pratt et al. examined the seasonal variation in fluid milk differentials in the South-eastern United States for 1997. The study is an extension of Pratt et al. work that included monthly supply and demand information. The months of April and September represent months of lesser and greater milk scarcity. Estimated market differentials (per cwt) for Nashville ranged from USD 3.31 to USD 3.96, and for Miami ranged from USD 5.40 to USD 6.79. The Federal order Class I differentials are USD 2.60 for Nashville and USD 4.30 in Miami.3 The authors concluded “that the Class I price differentials should be changed from month to month instead of the same differential being used throughout the year as is the current practice.”

The price effects of discriminatory pricing, results from previous studies Milk price discrimination and pooling systems exist in a number of countries. In Canada and the United States premiums for various end-uses of milk are determined under a classified pricing system administered by a government agency. In Japan, although the government does not administer any milk prices, it establishes the regional marketing zones and regulates the distribution of milk. These regulations ensure that milk from lower cost regions cannot be transported to satisfy demands in higher priced fluid milk regions. Until recently fluid milk market regulations were also imposed in Australia and the United Kingdom. However, Australia deregulated its fluid milk market in 2000 (Box 1.1) and the United Kingdom abolished the classified pricing system in 1994. The impact of price discrimination in domestic milk markets has been analysed in Buxton (1977), Ippolito and Masson (1978), Dahlgran (1980), Helmberger and Chen (1994), Lippert (2001), Chavas and Cox (2001), Australian Bureau of Agricultural and Resource Economics (ABARE) (2001), the Australian Competition and Consumer Commission (2001), FAPRI (2003), USDA (2004a), and USDA (2004b). The analytical and empirical studies illustrate that price discrimination reduces fluid milk consumption and increases the amount of milk available for processing. In addition, the average (pooled) price will be higher than the producer price in the absence of a pricing scheme (holding other support measures constant), and therefore leads to higher production levels. Fluid milk consumers who pay higher prices lose from price discrimination, while consumers of manufactured dairy products likely gain as manufacturing milk prices might be reduced by the scheme. However, empirical studies suggest that the higher cost of fluid milk outweigh any benefit consumers gain from lower prices for manufacturing milk.4 However, all these results are conditional on the complexity of a particular market and regulatory framework. As price discrimination is usually accompanied by milk distribution restrictions, the impact on producers is region specific. These restrictions negatively affect producers in efficient regions while the opposite is true for producers in less-efficient regions. (See the studies by Helmberger and Chen; Chavas and Cox; and by ABARE). Restrictions that protect inefficient producers create sector-wide inefficiency in the sense that the same amount of total output could be produced at lower cost. On aggregate, as Dahlgran (1980) has argued, the average milk producer price might, in theory, be reduced as a result of the system of end-use pricing – a case which he

16 demonstrated using data for the United States. Lippert (2001) has identified the inefficiencies of operating the complex system of end-use pricing and milk supply management5 in Canada. Cox and Chavez (2001) develop an annual interregional model that includes 11 Federal orders and California. The model includes fluid milk as well as nine manufactured products for which milk components are allocated. The model is used to examine the welfare changes that reduced regulation would generate. In the scenario that would eliminate the dairy price support and both California and Federal order programs, the farm price of milk in the Upper Midwest increases from USD 12.79 to USD 13.28 (+USD 0.49) per hundredweight, while in Florida the price drops from USD 15.29 to USD 12.54 (-USD 2.75), and in the Southeast from USD 14.02 to USD 12.48 (USD 1.54). Dalgran (1980) estimated welfare shifts from deregulation in a 16-region model. Regional 1976 price effects are not reported, but are summarised as averages. The gap between average fluid and manufacturing milk prices on average falls from USD 1.17 per cwt. to USD 0.46 per cwt. This result suggests that 40% of the price difference is related to interregional transportation costs and regional market conditions. The analysis did not include costs incurred by producer cooperatives to meet the other additional fluid milk marketing costs. USDA (2004a), examined the economic effects of principal US dairy programs including the impact of Federal Milk Marketing Orders. The operation of FMMO was approximated by using FAPSIM model and by incorporating formulas that are used to set the minimum prices for each class. The study assumes that 48% of the current Class I price differential would exist after the program elimination. In the scenario, fluid prices decline almost 8% below baseline levels. Consumers respond to this price change by increasing their demand for fluid milk by approximately 2% above the baseline period. Lower milk production coupled with increased fluid milk use reduces the supply of milk available for manufactured dairy products, leading to higher manufactured milk prices. FAPRI (2003) estimated impact of removing US milk federal regulations assuming that after the FMMO elimination USD 0.50 per cwt (approximately 25%) of the fluid milk premium would remain. As fluid milk prices are reduced fluid consumption rises 2.5% in response. The results also show that the largest negative price effects on milk occur in the first few years of the analysis. Once supply adjustment occurs, milk prices return closer to levels found before federal order elimination. An interesting finding of the study is that regions with less than 20% fluid utilisation would face higher all milk prices with the elimination of federal orders while those with fluid utilisation in excess of 35% would face lower all milk prices. The impacts of removing the fluid milk regulations in Australia are discussed in ABARE (2001). The study draws attention to the regional impact of deregulation pointing out that producers in regions with high fluid milk utilisation (high cost regions) are likely to lose, whereas producers in the more efficient regions, formerly restricted from accessing major fluid markets, stand to benefit. Following the “real-world” experiment offered by domestic market liberalisation, the Australian Competition and Consumer Commission (2001) conclude that following deregulation, milk consumers are better off and that Australian processors and retailers had not captured the benefits to the detriment of consumers. The report also noted that after deregulation milk is likely to continue to be produced relatively close to fluid markets. However, there will be a shift in manufacturing milk production to lower cost dairy regions over time. This would lead to

17 a corresponding shift in the location of dairy manufacturing plants to those regions where dairy farmers are able to produce raw milk at competitive export prices, but still respecting higher transportation costs associated with providing fluid milk to domestic consumers. In the majority of the studies, discriminatory pricing arrangements are analysed in the context of a closed economy and not much attention has been paid to the impact of these arrangements on trade. Sumner (1999), is one of the very few studies analysing the trade distorting impact of discriminatory milk pricing arrangements. His study, focusing on the US Federal Milk Marketing Order system, shows clearly that US exports and imports of manufactured dairy products will vary directly with the size of the price premium charged to consumers of fresh milk products. Bouamra-Mechemache et al. (2002) evaluated the options for developing a price discrimination policy in the EU dairy sector. Their analysis shows that the EU price discrimination, without the EU quota system, would significantly affect world prices and trade due to the increase in output, resulting from the higher producer price under price discrimination. With the quota in place, the impact on trade is considerably less. The authors claim that as long as price discrimination does not involve price discrimination between domestic and export markets, it might be WTO-compatible and, as such, a domestic price discrimination policy could be a partial substitute for more traditional policy measures.

A stylised model of dairy pricing and trade The standard theoretical framework for analysing the market impacts of government intervention in milk pricing is developed in Buxton (1977), Ippolito and Masson (1978) and more recently in Sumner (1999) and Bouamra-Mechemache et al. (2002). Figure 1.1 constitutes a graphical representation of that framework. Annex 1.1 contains the algebraic version of the model. In this framework, there are only two end-use milk classes: fluid milk and manufacturing milk. Fluid milk is considered as non-traded with demand supplied exclusively from domestic production. Manufacturing milk is used entirely to manufacture tradable dairy products, the domestic supply of which could be greater (as in this illustration) or less than domestic consumption.

18 Figure 1.1. Market effects of alternative milk price support measures

^P

ª Dbp { « PD «¬

Df PfB

B

f

`

B

Q Bf Pm QsAB Q Bf º » QsAB »¼

S

a k Pd b

c

d

e

PmB Pw

i f

l

j

g

h DdA Dd

Qf

B

Qf

A

Qd

A

Qd

B

B

QsAB

Source: OECD Secretariat.

The line S in the diagram represents the total supply of raw milk (the marginal cost curve for milk production). There are two demand curves, Df representing the demand for fluid milk and DdA representing the combined demand for fluid and manufacturing milk. Demand for manufacturing milk is given by the difference between DdA and Df. Note that the slopes of the demand curves differ, reflecting a demand for fluid milk that is more inelastic than that for manufacturing milk. To simplify matters, it is assumed that in the absence of government interventions in milk pricing, the price received by producers and paid by purchasers would be the same regardless of whether the milk is to be used for fluid purposes or for manufacturing dairy products.6 Moreover, under these “free-market” assumptions the domestic price would be equal (in raw milk equivalent terms) to an appropriately defined world market reference price — labelled Pw in Figure 1.1. Now, assume there are two policy options for achieving a given producer price for milk — the price labelled Pd in Figure 1.1. Under the first policy option the government simply sets a flat support price that all purchasers of raw milk must pay. Since that price is above the associated world market price, Pw, the government would have to defend it through the imposition of trade measures — export subsidies (as in the present illustration) and tariffs/tariff rate quotas.7 The intersection of Pd and S determines the level of total milk production, QsAB. The price Pd implies fluid milk consumption and production of QfA. Manufacturing milk processors buy the rest of milk produced (QsAB – QfA) also at the price Pd. Part of the manufacturing milk production will be consumed domestically (QdA – QfA) and part will be exported (QsAB – QdA). If we assume that the quantity purchased will have to be exported at the prevailing world price Pw, then the per unit export subsidy will equal (Pd - Pw) and total expenditure on export subsidies would amount to the area ‘l’ + ‘j’ + ‘g’ + ‘h’. This is the financial

19 transfer to producers from taxpayers. The financial transfer to producers from consumers is represented by the area ‘b’ + ’d’ + ’c’ + ‘e’ + ’i’ + ‘f’. Under the second policy option the government achieves the same targeted producer price Pd by using a combination of a flat support price, PmB in Figure 1.1, and an administratively determined fluid milk premium. This premium, represented in the diagram by the difference between PfB and PmB, is the extra amount that purchasers of raw milk destined for fluid uses must pay. The price producers receive under this arrangement is the weighted average of Pf B and PmB where the weights are the quantities of milk going to each of the two end uses. In this example, the manufacturing milk price and administered fluid milk premium are set up such that producers receive the target support price Pd at the level of output QsAB. Since farmers face the same incentive price the level of total milk production is the same QsAB in both cases. Under the combined regime the government can increase producer prices either by increasing the fluid milk premium or by increasing the flat support price. This means that with discriminatory pricing, the same desired target price Pd can be achieved with manufacturing milk prices set at the lower level PmB as compared to the policy relying only on trade measures. This is because producers under a policy of price discrimination get a part of their price support in consequence of higher prices charged consumers of fluid milk. The diagram illustrates that in response to the increase in the fluid milk price caused by the introduction of the fluid milk premium the fluid milk consumption will fall to QfB, i.e. a decrease of (QfA – QfB). As a result of the higher fluid milk price and the shift in the starting point the combined demand curve DdA moves leftward to DdB. It follows that by lowering fluid milk consumption, more milk is left for manufacturing purposes (QsAB – QfB). At the same time, following the introduction of the fluid milk premium, domestic consumers of manufacturing products will face the lower price PmB. Accordingly, the domestic consumption of manufactured products is higher, and is equal to (QdB – QfB). The difference (QsAB – QdB) will be exported, attracting a per-unit export subsidy equal to (PmB - Pw) and a lower total expenditure on export subsidies - the amount shown by area ‘h’. Note that the area ‘j’ is effectively being “cross-subsidised” by domestic fluid milk consumers. The total transfer to producers from consumers that follows the introduction of the fluid milk premium can be split into two parts: a transfer due to the discriminatory pricing arrangements and a transfer associated with the flat support price. In Figure 1.1, the former is represented as area ‘a’ + ‘b’, and the latter is represented as areas ‘d’ + ‘e’ + ‘f’ + ‘g’. (The financial transfer from taxpayers to producers is represented as area ‘h’.) Note that since Pd is the weighted average of PfB and PmB, the area ‘a’ is equal to the area ‘c’ + ‘i’ + ‘l’ + ‘j’. 8 The unit market price support created by discriminatory pricing arrangements is now equal to the price gap between Pd and PmB. The unit market price support attributable to the flat support price is equal to the gap between PmB and Pw. When milk prices are supported only via the flat support price, fluid milk consumers enjoy greater consumer surplus by area ‘a’ + ‘k’ as compared to when the same amount of price support is achieved under discriminatory pricing. Conversely, consumers of manufactured products under discriminatory pricing benefit from greater consumer surplus as compared to the outcome obtained using trade measures alone (a result that is difficult to represent in the graph due to the shift of the demand curve.9) In effect, price support achieved using discriminatory pricing shifts the associated cost burden from

20 consumers of manufactured dairy products, and taxpayers if the country is a net exporter, to consumers of fluid milk products. It is important to note that the market and trade impacts are conditional on other policy measures operated. That is, the general analytical model developed above does not take in the account policy measures such as quota systems, which allow the trade distorting impact of market price support to be limited.10

Impacts of alternative policy regimes on export subsidies and trade Table 1.1 below synthesises the implications for market price support, implicit budgetary expenditures on export subsidies and net trade (production minus consumption) obtained in comparing a flat support price policy regime that relies exclusively on trade interventions with the regime combining trade measures with discriminatory pricing (a net exporter case). The row labelled ‘TM’ shows the level of market price support, implicit budgetary expenditures on export subsidies and net trade impacts for the trade-measures-only regime. Likewise, the second row shows these outcomes for the combined regime. The third row contains differences obtained by subtracting the results obtained in simulating the combined regime from those obtained in simulating the TM-only regime. Notice that the levels of market price support are, by design, identical between the two regimes. However, the implicit budgetary expenditures on export subsidies are lower under the combined regime. This result is unambiguous (within the limits and assumptions of the analytical framework developed as explained above): the implicit budgetary expenditures on export subsidies will generally be greater for a flat-price, trade measures only regime, than for one that combines trade intervention with price-supporting discriminatory pricing arrangements. The reason for this result is clear. Under the discriminatory pricing regime that sets fluid milk prices higher than would be generated by perfectly competitive market forces, consumers are made to pay a larger sum of money for a given aggregate quantity of milk, by “exploiting” the low price responsiveness of demand for fresh milk. The additional consumer expenditure then effectively finances some of the export subsidy that, under the flat price regime, is financed by the government. In principle, a discriminatory pricing arrangement can even be implemented such that consumers finance the entire (then implicit) export subsidy, and the government pays no export subsidy even though exports are sold on world markets below the price received by domestic producers. Table 1.1. Selected indicators of trade effects for comparing policy regimes Policy regime:

Market price support

TM b TM+DPA Difference Combined regime gives

(Pd-Pw)*Qs AB (Pd-Pw)*Qs -0No difference

a

AB

Implicit budgetary expenditures 11 on export subsidies AB A (Pd-Pw)* (Qs - Qd ) B AB B (Pm -Pw)* (Qs - Qd ) B AB A (Pd -Pm )* (Qs - Qd ) Smaller implicit budgetary expenditures on export subsidies

Net trade AB

A

(Qs - Qd ) AB B (Qs - Qd ) B A (Qd - Qd ) Smaller trade impact (ambiguous)

Notes: a) Trade measures only. b) Trade measures plus discriminatory pricing arrangements. Source: OECD Secretariat.

The implications of the two policy alternatives for the volume of trade itself, are not as straightforward. Figure 1.1 is drawn in such a way that less quantity has to be exported

21 under the combined regime. The reduction in exports (QdB – QdA) is due to the fact that, in the diagram, the increase in fluid milk price reduces the fluid consumption by less than the decrease in manufacturing milk price boosts the manufacturing milk consumption. However, in general terms, the outcome is ambiguous. In some circumstances, net trade could be greater with the combined regime. The result depends critically on the numerical values of certain economic parameters. Analysis with the algebraic version of the model in Figure 1.1 permits further insights into these relationships. That analysis is developed fully in Annex 1.1. The main findings are summarised below. In interpreting these results, recall that the comparisons are standardised on a given amount of market price support provided alternately either via a trade measure or via price discrimination. x

The key factors determining relative trade impacts are: the relative magnitudes of the elasticity of demand for fluid versus manufacturing milk, the initial trading status of the country, and the initial relative supported prices of fluid and manufacturing milk.

x

In one important special case — that of a country which is not a net exporter of dairy products the only one of these that matters is that the elasticity of demand for fluid milk be less (in absolute value) than the elasticity of demand for manufacturing milk. In this case, market support resulting from trade measures (tariffs and their equivalents) will always be more trade distorting than market price support due to discriminatory pricing.

x

In some other cases though, a lower elasticity of demand for fluid milk is not enough. For a net exporting country, market price support due to discriminatory pricing will be relatively less trade distorting than market price support due to trade measures only if exports represent a “small enough” share of total manufacturing milk use. (The specific condition for this case is developed in Annex 1.1.)

x

Moreover, the higher the initial gap between fluid and manufacturing milk prices the more likely that a marginal change in market price support due to a (further) increase in fluid milk prices will be less trade distorting than an equivalent increase in market price support due to trade measures.

x

Taken together these conditions narrow the field of situations where price discrimination is relatively more trade distorting to those of a large net exporting country with a low initial fluid milk premium.

Results of policy simulation analysis The purpose of the empirical analysis presented in this section is to quantify the analytical framework developed in the previous section. (Figure 1.1) and further elaborated in Annex 1.1. The empirical analysis is based on simulations with the OECD’s Aglink and PEM models. The main goal of the policy simulation experiments undertaken with these two models was to obtain ‘order of magnitude’ estimates of the impacts of marginal changes in milk price support price support provided either via a flat-price, trade measures regime or via discriminatory pricing. Although different in terms of commodity, country coverage and length of run, these two models share the same analytical framework — the one sketched out in Figure 1.1 (A description of the Aglink and PEM models is presented in Annex A). The Aglink analysis addresses potential commodity impacts. The PEM analysis focuses on economic costs and benefits — the potential welfare impacts. In both analyses, the policy experiments were aimed at measuring the effects of reducing the amount of

22 milk market price support provided.12 In both cases the comparison is of the impacts of two different ways governments might choose to reduce that support. In one, the government reduces support prices for manufactured dairy products by reducing the associated trade measures (tariffs or export subsidies as appropriate). In the other, the government achieves the desired reduction in milk market price support by reducing the fluid milk premium (and any associated border measures applied to fresh dairy product and fluid milk trade). Note that in the first kind of policy simulation experiment the fluid milk premium is not changed. Likewise, in the second kind of policy simulation experiment, the settings for dairy product prices and associated tariffs/export subsidies are left unchanged. The two scenarios are chosen such that the reduction in producer price is the same between them. The difference is in the impact on consumer prices. In one, the price reduction applies equally to all consumers and the extent of price discrimination is left unchanged; in the other the price reduction is restricted just to consumers of fluid milk and the extent of price discrimination is reduced. In the tables, graphs and discussion below the two scenarios are labelled LSP, for Lower Support Price for manufactured dairy products and LFP, for Lower Fluid Premium applying to fluid milk prices.

Aglink results The dairy component of Aglink covers production and consumption of milk and major dairy products in the principal OECD markets, encompassing both importers and exporters. In general, it would be possible to impose a hypothetical price support and price discrimination scheme in any country covered by Aglink in order to show the main consequences of the analysed support measures for dairy markets and trade. The empirical analysis was undertaken with the Aglink US module. This choice was based on the fact that both market price support and a fluid milk marketing program with pooling are used in the US and the relevant equations already exist in the model. As noted above, the simulations are carried out to quantify the analytical and mathematical concept developed above and in Annex 1.1. Hence, these scenarios are purely illustrative and do not intent to estimate the impact of classified milk price system adopted in the United States in reality. The analytical framework has been developed here, for the sake of transparency, with only two end-uses of milk whereas in reality in the United States the end-use milk pricing is more complex and refined (see USDA (2004c) and USDA (2004b). There is a disagreement among researcher whether the classified price system adopted in the United States includes policy subsidy element. While empirical research supports the existence of price differences caused by interregional transportation costs, there is no recent empirical research addressing differences in the marketing costs associated with servicing plants that manufacture dairy products such as cheese versus plants that process fluid milk. This lack of empirical research contributes to the disagreement. However, market-generated prices are higher than Federal order minimum Class I prices in every U.S. market, with differences averaging USD 1.28 per cwt in 2000, USD 1.15 per cwt in 2001, USD 1.44 per cwt in 2002, and USD 1.47 per cwt in 2003 (USDA\AMS Dairy Market Statistics, Annual Summary.) Here, the purpose of the analysis is not to determine whether an administrative premium exist in a particular country or not. Therefore, in the base scenario, the Aglink baseline fluid milk premium was increased marginally to instigate the policy fluid milk premium.13 The manufacturing milk price was left at the baseline level while the all milk producer price was recalculated to account for the fluid

23 milk premium increase.14 In the subsequent policy scenarios, this premium was then reduced to assess the market impacts. The analysis comprises development of two policy scenarios and then compares them to the base scenario. Following the scenario setting described above, the levels of milk and dairy product prices were recalculated from the base scenario levels in such a way as to achieve a required marginal decrease in producer price (to simulate a marginal decrease in milk producer price a 0.5% reduction has been used). In the simulation, the domestic market is allowed to clear at the recalculated prices through changes in trade flows. It is important to note that in the simulations no account is taken of the WTO limits on trade, in terms of subsidised exports or barriers to imports. Thus, the trade flows and their impact on world markets have to be considered only as indicative of the magnitude of possible changes. Annex 1.2 contains tables showing year by year comparisons of the two scenarios for several key price and quantity variables for each of the seven years in the baseline — 2002 to 2008. Tables 1.2.2 and 1.2.3 compare the 7-year averages of those results. Simulated changes in domestic prices15 are reported in Table 1.2.2, simulated changes in market quantities in Table 1.2.3. Table 1.2. Average percentage changes of base scenario milk prices resulting from changes in manufacturing milk prices (LSP) and an assumed fluid milk premium (LFP) Producer price (PD)

Fluid milk price (Pf)

Fluid milk premium (Pf - Pm)

Manufacturing milk price (Pm)

LFP scenario

-0.5%

-1.24%

-4.74%

0.0%

LSP scenario

-0.5%

-0.41%

0.0%

-0.56%

Source: OECD Secretariat.

Table 1.3 shows the relative changes from base scenario for each scenario with respect to milk and dairy product consumption and production. The table shows that in the LFP scenario the consumption of manufactured dairy products is the same as in the base scenario reflecting the fact that the prices of manufactured dairy products in this scenario are left at the baseline level. On the other hand the consumption of manufactured dairy products in the LSP scenario increases in response to the reduction of those prices. The reduction of fluid milk price in both scenarios (Table 1.2) has brought an increase in fluid milk consumption in both cases. The increase is relatively higher in the LFP scenario, which is to be expected given the greater size of the fluid milk price cuts. The reduction of milk production in response to the 0.5% decrease of producer prices is the same in both scenarios by definition and amounts to about 0.12% on average. The decrease in production of dairy products is influenced by the reduction in milk production but also by different increases in fluid milk consumption, which reduces the availability of milk for manufacturing purposes.

24 Table 1.3. Production and consumption impacts of reductions in manufacturing milk prices and an assumed fluid milk premium, averages for 2002-2008 Consumption impacts LFP

LSP

Production impacts LFP

LSP

% changes from base scenario Fluid milk

0.189

0.063

Raw milk

0.189

0.063

-0.122

-0.122

Butter

0

0.210

-0.262

-0.244

SMP

0

0.336

-0.350

-0.318

Cheese

0

0.168

-0.262

-0.204

Whey powder

0

0.079

-0.174

-0.136

Source: OECD Secretariat.

In both policy scenarios, fluid milk consumption increases and milk production falls. This results in a smaller volume of milk available for manufacturing and a consequent reduction in production of dairy products. These differences are manifested in changes of trade volumes summarised in Figure 1.2. (Figures 1.2.1. – 1.2.4. in Annex 1.2 depict traded volumes for each dairy product and for each of the seven years in the baseline for the baseline and the two policy scenarios.) Figure 1.2 indicates that the simulated quantity of net butter imports by the United States is increased in both scenarios. The LSP scenario shows a greater increase in butter net imports when compared to the LFP scenario. The difference is about one thousand tonne (kt) in absolute terms (42% in relative terms). An even larger increase in net imports is seen for cheese as is apparent from the size of the columns in the diagram. The LSP scenario shows greater increase when compared to LFP scenario by about 5 kt in absolute terms (30% in relative terms). The last set of columns in the diagram indicates that simulated skimmed milk powder (SMP) net exports have been reduced in both scenarios (in Figure 1.2 the net SMP exports are presented as negative net imports). The higher reduction is seen in the case of the LSP scenario and the difference when compared to LFP scenario is about 1.5 Kt in absolute terms (44% in relative terms). Results presented in Figure 1.2 make clear that milk market price support — whether resulting from the fluid milk premium or from higher manufactured dairy product prices, does have impacts on trade. These impacts are quantitatively different. A 0.5% decrease in average producer price achieved by reducing the price of supported manufactured products16 (LSP scenario) increases overall consumption (in milk equivalent) by a greater amount than is seen in the scenario which reduces the average producer price by means of reducing the fluid milk premium (LFP scenario). Given the fact that milk production is identical in both scenarios the differences in the fluid and manufacturing milk consumption translate directly into differences in net exports that have immediate consequences for world dairy market prices. The differences in results for simulated world market price impacts are shown in Figure 1.3 for both scenarios.17

25 Figure 1.2. Simulated impacts of reductions in manufacturing milk prices and an assumed fluid milk premium on US net imports of butter, cheese and skimmed milk powder (changes from baseline)

20

LSP

Thousand tons

15

LFP

10 5 0 -5 Butter

Cheese

SMP

Source: OECD Secretariat.

Comparing the impact across dairy products indicates that the highest percentage change in a world price is in the cheese market. The higher impact on cheese markets stems from the fact that the majority of manufacturing milk in the United States is used for the production of cheese. In fact, about three-quarters of all manufacturing milk are channelled to cheese production.18 It follows that the reduction in production in absolute terms is significantly higher for cheese than for the other products. The figure also shows that, as for net exports, world market price effects are greater for the LSP scenario. Figure 1.3. Simulated impacts of reductions in manufacturing milk prices and an assumed fluid milk premium on world market prices

% change

1.0 0.8

LSP

0.6

LFP

0.4 0.2 0.0 Butter

SMP

Cheese

WMP

Source: OECD Secretariat

In the earlier discussion of ‘in general’ effects of milk price support measures it was noted that for some specific situations the question of which support measure is most trade distorting could be answered definitively only with reference to the trading status of

26 the country and specific elasticity assumptions. The simulated result obtained here suggests a high degree of similarity in trade effects between the two with price support due to trade measures slightly more trade distorting than price support due to discriminatory pricing.

PEM results The Aglink analysis in the foregoing section focused on illustrating the price and quantity impacts of hypothetical reductions in producer prices achieved by reducing alternately the support prices for manufactured dairy products and the fluid milk premium. The PEM analysis reported in this section of the paper focuses on the effects on trade and welfare of hypothetical changes in the level of market price support provided alternately by those support mechanisms. Moreover, where the Aglink analysis was concerned mainly with supply, demand and price effects for individual manufactured dairy products at the final consumer, here the spotlight will be on aggregate effects measured in milk equivalents at the farm gate. That is, the “consumer” in PEM is the first consumer in the processing chain, who purchases from the dairy farmer. This is consistent with the definition of consumer used in the PSE database. It does prompt a caveat: To the degree that there is imperfect price transmission along the processing chain to the retail market, welfare results for the “consumer” in the model will be an imperfect proxy of the welfare of the final consumer of the dairy product. Specifically, model results will overestimate final consumer welfare gains in response to reduction in producer price support if some of those changes are captured as excess rents by processors or retailers.

Simulated net trade impacts of alternative market price support measures The first simulation experiment done using the PEM model was to evaluate the impact of reducing the level of any market price support provided by discriminatory pricing arrangements in Japan and in the United States. Accordingly, the simulated reduction in market price support is achieved by reducing an assumed fluid milk premium without changing trade measures. Consistent with the labelling scheme adopted for the Aglink analysis this simulation is called the “Lower Fluid Premium” (LFP) experiment. The second experiment simulated the impact of reducing the amount of support provided (implicitly) by trade measures applied to manufactured dairy products in the same two countries. Again in keeping with the Aglink labelling this experiment is called the “Lower Support Price” (LSP) experiment. In both experiments the simulated reduction in milk market price support was USD 100 million. Figure 1.4 shows the simulated impacts on net dairy product trade, converted to milk equivalents. When measured in milk equivalents both Japan and the United States were net importers of dairy products in 2001. (Since, by convention, net trade is measured as production minus consumption so that imports appear, as in Figure 1.4, as negative numbers.)

27 Figure 1.4. Estimated trade impacts of reductions in manufacturing milk prices and an assumed fluid milk premium (PEM simulated changes in the volume of net milk equivalent exports) Japan

United States

LFP

0.0 million tonne

LSP

-0.1 -0.12 -0.2

-0.17 -0.22 -0.26

-0.3

Source: OECD Secretariat.

PEM results in Figure 1.4 generally confirm the Aglink results for individual dairy products presented in the previous section: no matter the mechanism, reducing milk market price support increases imports or decreases exports. Likewise, these findings corroborate the earlier proposition, based on analysis with the stylised model, that trade measures will always be more distorting than price-supporting discriminatory pricing if the country in question is a net importer.

The economic costs and benefits of alternative milk market price support measures Table 1.4 presents and defines the PEM model indicators of the various categories of economic benefits and costs of support for the two categories of milk market price support featuring in the present study. The charts in Figure 1.5 display simulation results for these indicators comparing the two categories of milk price support. Consistent with the fact that in all the simulation experiments market price support is reduced, note that consumers gain while farm households and input suppliers lose. Estimated transfer efficiency ratios reveal the relative magnitude of these changes. Table 1.4. Indicators of economic costs and benefits in the PEM dairy model Economic indicators

Definition of measures

Taxpayer costs

Total change in costs incurred by government in paying for border measures * and consumer subsidies.

Consumer surplus

Change in consumer surplus.

Farm income

Change in return above opportunity costs to farm owned factors including land, cows, forage and silage.

Input supplier profits

Change in returns above opportunity costs earned by suppliers of purchased factors including purchased feed and hired labour.

Transfer efficiency

Change in farm income divided by the sum of the changes in taxpayer cost and consumer surplus.

Note: *Export subsidies would lead to positive costs for a net exporter whereas border measure costs would be negative for a net importing country where the government collects the tariff revenues. Source: OECD Secretariat.

28 Figure 1.5. Simulated impacts of reductions in manufacturing milk prices and an assumed fluid milk premium on selected indicators of the economic benefits and costs of support taxpayer costs Japan

100

United States

76

-21 -50.0

-34

million USD

million USD

64

64

0.0

-32

-41

50

LFP

-100.0 farm household incom e Japan

45

34 17

Fld*

Mnf**

Fld*

Japan

-50

input suppliers profits Japan

United States

-12

-1 0

LSP

Mnf**

Fld*

Mnf**

Fld*

Mnf**

United States

United States

0

0

-24 -26 LSP

-50

LFP

-54 -54

million USD

-11 million USD

* fluid milk consumers ** manufacturing milk consumers

consum er surplus

-11 -22

-50

-23 LSP LFP

-100

-100 transfer efficiency 1.0

0.5

0.42

0.47

LSP 0.28 0.31

LFP

0.0 Japan

United States

Source: OECD Secretariat.

The first diagram in Figure 1.5 depicts the impact of the two experiments on the taxpayer costs in the United States and Japan. The reduction in taxpayer costs is mainly due to a simulated increase in import tariff revenues in the two countries.19 (In this context, tariff revenues the government receives from, for example, auctioning rights to import protected dairy products, might helpfully be viewed as negative taxpayer costs.) Note that two factors influence the size of import tariff revenues: 1) the total volume of imports and 2) the price gap between domestically produced manufacturing milk and the world reference price. The results in Figure 1.5 suggest that the import quantities increase proportionally more than the corresponding reductions in the gap between the manufacturing milk and the world prices in all the experiments. Because consumer prices fall in both experiments, consumer surplus increases in both. However, as shown in Figure 1.5, the estimated impacts on the two milk markets are different. In the LSP experiments, the consumer prices for both fluid milk and manufacturing milk are reduced, benefiting both fluid milk and manufacturing milk consumers. In the LFP experiment, the fluid milk price is reduced without changing the gap between the manufacturing milk and the world prices confining the gains just to fluid

29 milk consumers. Indeed manufacturing milk consumers lose fractionally. The reason for this is the increase in the world reference price in response to the increase in the import quantity. The increase in the imports reduces supplies in the world dairy market, putting upward pressure on the world reference prices. This automatically increases the domestic manufacturing milk price and decreases consumer surplus of manufacturing milk. While qualitatively similar, there are some differences in the simulated results for the United States as compared to those for Japan. Note first, from Figure 1.4, that the simulated increase in milk equivalent imports is significantly greater for the United States, largely because the US dairy market is substantially larger. Obviously, in light of these changes in imports, the world market price effects and the associated reduction in surplus for consumers of manufacturing milk is greater in the LFP experiment for the United States simulation than for Japan. Moreover, there are differences in the simulated results that reflect the different characteristics of the milk markets in those countries. The ratio of fluid milk consumption to manufacturing milk consumption is significantly higher in Japan, while the ratio of the demand elasticity for fluid milk to that for manufacturing milk is significantly lower.20 There is another interesting difference in the results for the United States versus those for Japan. First, recall that in all the simulation experiments, the simulated reduction in milk market price support is the same. For Japan, the simulated increase in consumer surplus (the total for fluid milk and manufacturing milk consumers) due to a reduction in the fluid milk premium is less than the simulated increase in consumer surplus due to lower support prices for tradable dairy products. For the United States, although it is the other way around, consumers benefit just slightly more from the reduction in the fluid milk premium. In general terms, the implications for consumer surplus of discriminatory pricing versus flat price-trade measures based support are, as was the case for trade effects, ambiguous. The empirical result depends on the initial trading status of the country, initial levels of support and elasticities. Transfer efficiency refers to the proportion of total consumer and taxpayer costs of a support measure that translates into increased net income for farm households. Earlier work on the transfer efficiency of alternative support measures (OECD, 2002) reports that market price support is among the least efficient ways of improving incomes of farm households. Findings in that analysis suggested that, on average across the OECD and for all of agriculture, probably no more than 25% of the total costs paid by consumers and taxpayers under market price support measures can be counted as net gain for farm household income. The rest is lost either in the form of income gains by economic agents who were not the intended beneficiaries of support or in consequence of resources being diverted from other productive uses to the production supported commodities. Previous analyses of transfer efficiency focused on only one category of market price support — implicitly market price support that results from government intervention via trade measures. Notice that the estimated transfer efficiency of milk market price support is slightly higher than that obtained in earlier work considering agriculture in total, OECD-wide. Differences in two key economic parameters explain the results. For milk production, the share of farm owned factors is typically higher and the elasticity of supply of them is lower than for agriculture in total.21 These results also line up well with earlier findings regarding the trade effects.

30 The estimates of transfer efficiency for support due to discriminatory pricing are higher than for support due to trade measures. Since, by the design of the simulation experiment, the change in farm household income is roughly the same between the two simulation experiments; the difference is explained by differences in simulated impacts on taxpayer costs and consumer surplus. Once again, however, the differences are not large. Moreover, before any strong conclusions could be drawn in this regard some other cost considerations would need to be accounted for — especially potential extra administrative costs of price discrimination schemes as compared to price support arrangements based exclusively on trade interventions.

Conclusions This chapter has investigated trade and economic effects of milk price support measures. The market price support in this chapter was segregated into i) support due to trade measures – i.e. support attributable to a package of target prices, product support prices and trade measures and ii) support due to discriminatory pricing. Here, the discriminatory pricing does not refer to a discrimination between domestic and export markets but rather to a practice administered or sanctioned by a government that lead to prices paid for raw milk, for some end uses (typically fresh milk products) that are higher by more than the additional marketing costs than those paid for raw milk for other end uses (typically manufactured milk products) on the domestic market. The additional revenue generated by discriminatory pricing is then transferred back to farmers through a pooled (averaged) price. Although there are market-driven explanations for some of the difference between fluid and manufacturing milk (consumer preferences in terms of quality differences, transportation and marketing costs, and seasonality in milk production was noted), the administered price difference could be sufficiently large to provide an additional price support to producers. It follows, that domestic fluid milk consumers under price discrimination partly support farmers via such a higher set fluid milk premium or set fluid milk price which can enable a government to set support measures for the manufacturing (tradable) milk at a lower level (for the same “target” producer support price) in comparison to a government that does not operate price discrimination and pooling. An analytical framework was set up in this chapter to investigate the impact of support due to trade measures and due to discriminatory pricing on dairy markets and trade. The simulation analysis with a mathematical model (based in economic theory) of milk price determination shows that on a dollar for dollar basis it is theoretically possible for milk price support resulting from discriminatory pricing to be as or even more trade distorting than milk price support resulting from explicit trade intervention in dairy product markets. The empirical results of the assumed change in policy parameters obtained from simulations with Aglink and PEM suggest, however, that, as a practical matter, pricesupporting discriminatory pricing arrangements could be less trade distorting —but not by much. Moreover, such price-supporting arrangements are by design, unfair, imposing higher costs on one particular group of consumers. Both categories of milk price support cost consumers and taxpayers considerably more than the benefits they deliver in farm household income. It should, however, be noted that the results of this analysis are simulated model outcomes and thus have to be viewed within the limits of the tools applied and the hypotheses and assumptions postulated.22 The most important assumption is that markets are assumed to be perfectly competitive.

31 In summary, compared to a situation without market price support, market price support whether the consequence of trade interventions in dairy product markets or the consequence of discriminatory pricing arrangements, leads to increased production (unless there are quota restrictions) and reductions in consumption. In most cases this will result in higher exports and lower imports. Market price support due to the flat support price and trade measures regime is deemed to be relatively more distorting than market price support due to discriminatory pricing, however any associated reductions in trade distortion and export subsidies come at the expense of higher costs for fluid milk consumers. Box 1.1. Milk price deregulation in Australia Market milk regulations were originally introduced to guarantee year round supplies of fresh milk to Australian consumers. The regulations included farm gate and retail price controls on fresh milk sales as well as supply management arrangements, all on a regional basis. The fluid milk prices differed across regions despite the potential for inter-state trade, because an industry agreement ensured that inter-state fresh milk sales were priced to maintain the regulated farm-gate price. Although the farm-gate price of fresh milk varied from state to state, it was always well above the price of manufacturing milk. In 1995, a regulatory reform process began which stipulated that in each state only farm gate price controls would remain in place by January 1999. In July 1999 the review of market milk regulations in Victoria concluded there was no net public benefit from retaining farm-gate price controls. Victoria is the largest milk producing state in Australia, accounting for nearly two-thirds of total national milk production. Faced with the prospect of complete deregulation of such a large share of national milk production (in a region seen as having comparative advantage in milk production) governments in other producing states concluded that their own market milk regulations would be unsustainable. Accordingly, when the Victorian government announced the state milk marketing regulations would end on 1 July 2000, the individual state dairy industries and governments recognised that national deregulation of dairy industry was inevitable (OECD, 2001). The policy reform removed simultaneously the Dairy Market Support scheme and fresh milk regulations on 1 July 2000 to allow the market to determine milk prices. At the same time, a structural adjustment package was introduced in the form of the Dairy Industry Adjustment Act 2000 to help producers cope with the adjustment pressures and allow farmers to choose between adjusting to lower market returns versus leaving the industry (Australian Competition and Consumer Commission, 2001). The adjustment package is funded by a levy of 11 cents (USD 6 cents) per litre on all domestic sales of fresh milk. The levy partially replaced the implicit consumer tax inherent in the fresh milk regulations and was set at a level to ensure consumer prices would not rise after deregulation. It will remain in place for approximately eight years until the package is fully funded. The individual adjustment programs are called: Dairy Industry Adjustment Package (DIAP), Dairy Structural Adjustment Program (DSAP), Dairy Exit Program (DEP) and Dairy Regional Assistance Program (DRAP). Australian farm gate milk prices (AUS cents per litre) Manufacturing Fluid milk Average farm gate price

1995-96

1996-97

1997-98

1998-99

1999-00

2000-01

2001-02

26.0 49.9 31.1

24.0 51.1 29.6

23.9 52.0 29.4

22.5 51.5 28.5

22.0 52.3 26.2

29.0

30.2

Source: Department of Agriculture, Fisheries & Forestry, Australia (2002).

32

The table below illustrates the fluid, manufacturing and average (pooled) milk price received by farmers prior to deregulation (1995-96 until 1999-00) and after the deregulation (after 2000-01). It is apparent from the table that, prior to deregulation, there was a substantial difference between manufacturing and fluid milk market prices. The removal of state government controls of the farm gate supply and pricing of milk introduced an open market for fluid milk in Australia. Thus, after the milk market deregulation in July 2000, market forces removed the differential between ‘market’ and ‘manufacturing’ milk prices that had previously been imposed on markets by government policy. In the four years since deregulation, regional price variation has fallen across Australia, with a smaller change in the overall Australian average prices (ABARE, 2004). Market forces continue to generate prices that differ across regions, but with about half the variation that existed in the last four years of regulation. Farm gate milk prices in individual states both increased and decreased under deregulation (see table below). Australian farm gate milk prices by statea)

1996-97 1997-98 1998-99 1999-2000 2000-01 2001-02 2002-03

New South Wales c/L 37.5 36.4 35.1 32.3 29.1 32.5 32.8

Victoria c/L 25.3 24.2 24.4 22.1 29.3 33.3 24.8

Queensland c/L 38.7 38.4 38.5 36.8 30.6 34.5 34.8

South Australia c/L 28.7 28.7 29.1 28 27.7 31.5 30.3

Western Australia c/L 34.2 35.1 34.1 34.2 26.6 28.7 28.2

a/ Prior to 2000-01, prices are weighted average of both market and manufacturing prices. Source: ABARE (2004)

Tasmania c/L 22.8 22.6 23.7 20.9 25 32.7 25.9

Australia c/L 28.5 27.7 27.5 25.4 29 33 27.1

33

Notes 1.

Farmers would always sell milk to the highest bidder and processors would always buy from the lowest seller. Since milk is almost homogeneous in terms of quality with only minor adjustments in price for variations in content, efforts to sell for more or buy for less would fail; there would tend to be a single price for milk at a given place and time.

2.

It is possible that consumers would replace, within the limits of remaining quality differences, fresh fluid milk with reconstituted milk which would likely erase any differences between fluid milk and manufacturing milk price based on transportation costs. In addition, it is possible that evolution of pasteurisation (UHT) could affect transportation costs.

3.

The FMMO Class I differentials are added to a manufacturing milk price to obtain a minimum fluid milk price f.o.b. plants

4.

For the case of the United States, Ippolito and Masson (1978) estimate that the loss to consumers of fluid milk amounts to about USD 334 million while the gain to manufacturing milk consumers is about USD 120 million. Helmberger and Chen (1994) estimate the loss to fluid milk consumers in the United States to be USD 1000 million and the gain to manufacturing milk to be USD 600 million.

5.

Under the milk supply management system, the Canadian Milk Supply Management Committee (CMSMC) sets a national production target – the Market Sharing Quota (MSQ) – for industrial milk. The MSQ is set with the goal to achieve a domestic market balance in terms of butterfat, and is assigned to provinces largely on the basis of historical shares. The CMSMC monitors the evolution of the MSQ on a monthly basis based on the monthly production and demand situation to be more market responsive and avoid over-quota production. In addition to the MSQ, each province sets its own production target for fluid milk, and the entire milk quota – industrial and fluid together – is allocated to producers.

6.

However, as discussed above, the presence of transportation and other marketing costs and seasonal payments will in theory generate a market-driven fluid milk premium. The representation of this ‘natural’ premium would make the graphical analysis intractable. Nevertheless, the analytical framework remains valid if the demand schedules and administered prices as depicted in the diagram are viewed as net of transportation cost and seasonal premiums. (For further discussion see Ippolito and Masson.)

7.

Note that applying trade measures [import tariffs, tariff rate quotas (TRQ’s) and export subsidies] is analytically equivalent to supporting price by intervention buying.

8.

The area ‘a’ = ‘c’ + ‘i’ + ‘l’ + ‘j’ is equivalent to (PfB – Pd) QfB = (Pd – PmB) (QsAB – QfB). By rearranging the equation we get Pd QfB + Pd (QsAB – QfB) = PfBQfB + PmB (QsAB – QfB). By further simplifying we arrive at the formula for the average producer price that is: Pd = (PfBQfB + PmB (QsAB – QfB)) /QsAB.

9.

Given (PmB < Pd), the first term is bigger than the second term which implies that consumers of manufactured products under the discriminatory pricing arrangements increase their consumer surplus by: 'CS B

º ª QdB º ª QdA B B B B « ³ Dd ( x)dx Qd Q f Pm » « ³ DdA ( x)dx QdA Q fA PD » » « Q Bf » « Q Af ¼ ¬ ¼ ¬

34 10.

The consumer welfare impacts will be conditional on the degree of price transmission along the supply chain. For a discussion on price transmission issues, see Box 3.1 in Chapter 3. Analysis of international dairy trade liberalisation.

11.

Tariff revenues in the case of a net importer.

12.

Given the linear nature of the model, changing the exercise by increasing rather than decreasing producer price would not change the main points and findings of the analysis.

13.

The baseline used for the base and policy scenarios refers to the market and trade projections of the OECD Agricultural Outlook report (OECD, 2003).

14.

In the base scenario the fluid milk premium was increased by 5% while all milk producer price increased by about 0.5%.

15.

Differences between simulated price changes for fluid versus manufacturing milk reflect differences in the relative importance of these two demand categories in total milk production. Roughly two-thirds of total raw milk production in the US goes into production of manufactured dairy products. It follows that the proportion of milk used for fluid purposes is about one-third. Thus manufacturing milk has a relatively greater weight in calculating the pooled producer price.

16.

This mimics the effects of using just trade measures to accomplish the reduction.

17.

The domestic price of whey powder in the United States is used as a proxy for world whey powder price in Aglink, thus no impact on world whey price is reported.

18.

The allocation of US manufacturing milk in the model is based on regression analysis and historical shares of individual dairy products in the manufacturing milk markets.

19.

Consumer subsidies are also included in taxpayer costs. However, the rates of these subsidies are much smaller than the gap between the manufacturing milk and the world reference prices in the two countries. Accordingly, their relative contribution to the simulated change in total taxpayer costs is negligible.

20.

The fluid and manufacturing milk elasticities are –0.3 to –0.5 respectively in Japan, and –0.15 to –0.41 in the United States.

21.

This is even more so the case in countries where milk production quotas are in place. This may help to explain why reform of milk price support policy is typically a politically difficult matter.

22.

The hypotheses of the analytical framework are laid out in the first 5 paragraphs of the section on stylised model of dairy pricing and trade, endnote 7 and Annex 1.1.

35

Annex 1.1 Algebraic Version of Graphical Model

In this annex the expected trade effects of discriminatory pricing arrangements relative to trade measures are examined with a simple mathematical model for a milk market. The analysis employs the same simplifying assumptions underlying the graphical version of the model in Figure 1.1. of the main text. These are: x

There are only two end-use milk classes, fluid milk and manufacturing milk.

x

Fluid milk is produced and consumed domestically (i.e. no trade in fluid milk products).

x

Manufacturing milk is used to produce tradable dairy products.

x

Trade measures widen the gap between the implicit domestic and border prices of manufacturing milk.

x

Discriminatory pricing arrangements constitute the only source of a premium for milk for fluid use over manufacturing use.

x

Farmers receive a weighted average of the prices paid for milk destined for fluid uses and the prices paid for milk destined for manufacturing uses.

x

The country in question is small enough in world trade to have no or negligible influence on world trade. The supply, demand and price equations for this model are: Qs = S (Ps) (1) Qf = Df (Pf) (2) Qmd = Dm (Pm) (3)

Ps

Pf Q f Pm Qs Q f

Qs

(4) X = Qs - Qf - Qmd (5) The symbol Qs stands for quantity produced. S (Ps) is the milk supply function and Ps is a milk producer price. Qf is the demand for fluid milk, Df (Pf) the fluid milk demand function and Pf the fluid milk demand price. Similarly, Qmd is the demand for

36 manufacturing milk, Dm (Pm) the manufacturing milk demand function and Pm the manufacturing milk demand price. Since Ps is the weighted average of Pf and Pm, Ps can be written as in equation (4). X is the net exports of dairy products (milk equivalent term) which will be negative in the case of imports. All quantities and prices are considered in liquid units. Writing out the total differentials of equations (1) through (5):

S cPs dPs

dQs

HQ s dPs Ps

(6)

c D f Pf dPf

dQ f

K f Qf Pf

dPf

(7)

c Dm Pm dPm

dQdm

K m Qdm dPm Pm

(8)

dPs

§ · P Q f ¨1 K f m K f ¸ ¨ ¸ Pf Qs Q f © ¹ dP f § § P · P Qs ¨¨1 H m H ¸¸ Qs ¨¨1 H m Ps ¹ Ps © ©

· H ¸¸ ¹

dPm

(9) dX = dQs - dQf - dQmd (10) The symbols , f and m stand for the elasticities of milk supply, fluid milk demand and manufacturing milk demand respectively. Substituting equations (6), (7) and (8) in (10) and re-arranging, we have

dX dPs

HQs K f Q f dPf K m Qdm dPm Ps Pf dPs Pm dPs

(11) Equation (11) shows how a change in a milk producer price changes the volume of trade. When a trade measure is used to achieve a given increase in Ps the price gap between fluid milk and manufacturing milk (fluid milk premium) is assumed to remain at its initial level such that, dPf = dPm. This does not rule out the case examined in the previous section where the ‘initial level’ of the fluid milk premium was zero. It does allow though illustrating, as is done subsequently, the important dependence of trade effects on initial relative prices of fluid and manufacturing milk prices. In this case, equation (11) becomes

dX dPs

dPf dPm

dX dPs

trade

HQs §¨ dPf Ps ¨ dPs ©

dPf

m ·§ K Q ¸¨ f f K m Qd ·¸ ¸¨ P Pm ¸¹ f dPm ¹©

37

§ P · Qs ¨¨1 H m H ¸¸ Ps ¹ HQ s © Ps §Q P Qf ¨ s K f m K f ¨Q Pf © f

§ K f Q f K m Qdm · ¨ ¸ Pm ¸¹ · ¨© Pf ¸ ¸ ¹

(12) Consider now the case where the supposed increase in Ps is achieved only through an increase in the fluid milk premium without any change in the trade measures, i.e. dPm = 0. In this case, equation (11) becomes

dX dPs

dPm 0

dX dPs

DPA

HQs §¨ dPf Ps ¨ dPs ©

dPm

§ P · K f Qs ¨¨1 H m H ¸¸ Ps ¹ HQs © Ps · § P Pf ¨1 K f m K f ¸ ¸ ¨ Pf ¹ ©

·K f Qf ¸ ¸ Pf 0¹

(13) Subtracting equation (13) from (12) and re-arranging, we get the difference in trade impact due to trade measures and that due to discriminatory pricing: dX dPs trade

dX dPs DPA

m · ª § K m ·¸§¨ P f ·¸§¨ Qd ¨ « ¸°®1 K T 1 m f ¨ ¸ ¨ ¨ ¸ « Pm X Q ¸° K d ¹¯ ¹© ¬ © f ¹©

ª § K · P ° f T «1 ¨ m ¸® « ¨ K f ¸° Pm ¹¯ ¬ ©

§1 K ¨ f ©

· K ¸ f ¹

§ P · ½º ¨ 1 m ¸ °» ¨ P ¸ ¾» f ¹° © ¿¼

½°§ Q m · º d ¸» ¾¨¨ m ¸ °¿© X Qd ¹ »¼

(14) where is defined as

T

· § P K f Q s ¨¨1 H m H ¸¸ X Qdm Ps ¹ © § P P f Q f ¨1 K f m K f ¨ Pf ©

·§ Q ¸¨ s K Pm K f P f ¸¨ Q f ¹© f

· ¸ ¸ ¹

0 Pm d Pf ½ ° It is easy to prove that: and ¾ T 0 . Therefore whether trade measures 1 d K f 0 °¿ have greater or smaller impact on the volume of trade than discriminatory pricing arrangements depends only on the value in the square bracket in equation (14). That condition can be written as

38

dX

! dX

dPs trade dPs DPA

§K f · ¨¨ ¸¸ ©Km ¹

§ Pf · ¨¨ ¸¸1 K f K f © Pm ¹

Qdm m ! X Qd

(15)

The propositions in (15) are illustrated in Figure 1.1.1. The straight line Qdm/(X+ Qdm) I P ^3I3P I - I` is the border between two critical zones. Any points on the right-hand side of that line correspond to situations where the trade effect of trade measures is greater than the trade effect of discriminatory pricing arrangements. Any points on the left-hand side of that line indicate the situations where the trade effect of discriminatory pricing arrangements is greater than the trade effect of trade measures. Figure 1.1.1. Relative trade effects of milk price support measures

A (0 , 1 ) f}

C (1 , 1 )

f)-

K K

m)/{(P f/Pm)(1+

Q

m

d

/(X + Q

m

d

)=(

f

/

m

)/{(P f /P

m

)(1+ K f )- K f }

( f/

K K 4 5°

O

B (1 , 0 ) E x p o rt er

Im p o rt er

Q

m

d /( X+ Q

m

d)

Since our assumptions include ( f/ m)<1, i.e. that the elasticity of demand for fluid milk is always less than the elasticity of demand for manufacturing milk, all points must range from zero to one vertically.1 Horizontally, any points ranging from zero to one indicate net exporters; those exceeding one indicate net importers. Any points in Figure 1.1.1. may be divided into three cases based on the following criteria: the relative manufacturing and fluid milk prices; and the status of trade (importers and exporters).

39

Case 1: X<0 This is a case of net importers and the corresponding area in Figure 1.1.1. is the righthand rectangle area of the vertical line BC. The rectangle area is located in the right-hand side area of the border, indicating that the trade effect of trade measures is always greater than the trade effect of discriminatory pricing arrangements in the case of net importers of dairy products (in milk equivalent totals). Case 2: X>0 and Pf =Pm This is a case of net exporters with no initial price gap between fluid and manufacturing milk (no initial fluid premium). In this case, it is clear that equation (15) can be simplified as

§ K f · Qdm ¸¸ ¨¨ m dPs trade dPs DPA © K m ¹ ! X Qd dX

! dX

(15

This result means that the trade effect of a trade measure is greater than that for discriminatory pricing arrangements if the ratio ( f/ m) is less than the ratio (Qdm/(X+ Qdm)) and vice versa. The corresponding area for the former case in Figure 1.1.1 is the area AOC, and the area for the latter case is the area BCO. Case 3: X>0 and Pf >Pm This is a case of net exporters that currently have fluid premium. In this case, since (Pf /Pm)>1, the denominator of the left-hand term in the equation in (15) is greater than one,2 resulting in the value of the left-hand term being less than ( f/ m). Like Case 2, the percentage of exportation and the ratio between the two demand elasticities are key to determine the relative magnitude between the trade effect of trade measures and the trade effect of an increase in the fluid milk premium. As noted earlier, there is one other important relationship for this case — the ratio of the initial price of fluid milk to manufacturing milk. The bigger is this ratio, the higher is the possibility that the trade effect of an incremental change in trade measures is greater than the trade effect of an incremental change in the fluid milk premium. The black diamonds in Figure 1.1.1. correspond to data points constructed from the milk demand elasticities, milk quantity and price data that form part of OECD’s Aglink and PEM models.3 These particular results correspond to observed situations of seven major milk-producing countries in the OECD in 2001: Australia, Canada, European Union, Japan, Mexico, Switzerland and the United States. Table 1.1.1 summarises the ranges of key parameters observed across those seven countries. The ratio of the demand elasticity for fluid milk to that for manufacturing milk ranges in value between 0.30 to 0.78, indicating that, as expected, the demand for fluid milk is relatively price inelastic. The ratios of the demand for manufacturing milk to the total manufacturing milk production illustrate a wide variety in the trading status of dairy products amongst those countries. The minimum value (0.31) corresponds to the situation of a significant net exporter — Australia in this case while the maximum value (2.16) refers to the situation of a major net importer — Japan in this case.

40 Results in Figure 1.1.1. reveal that, in general, a given marginal change in price support achieved using trade measures will be more trade distorting than if that increase in support were to be achieved using discriminatory pricing arrangements. However, there is clearly the possibility of the reverse happening. This situation may, as stated above, occur for a country where the share of exportation is big; the elasticities of demand for fluid and manufacturing milk are close to each other; and the initial fluid milk premium is small. Table 1.1.1. Ranges of values for key components of formulas determining relative trade effects

minimum maximum average

K f / K m 0.30 0.78 0.49

(P f /P m )(1+ K f )- K f 1.00 1.29 1.10

( K f / K m )/{(P f /P m )(1+ K f )- K f } 0.25 0.78 0.45

Q

m

d /(X+Q

m

d)

0.31 2.16 1.16

Notes 1.

This is the case for major milk producing countries among the OECD, as shown in Table 1.2.

2.

Substituting 1 from the denominator yields: {(Pf/Pm)(1+Kf)-Kf }-1=(1+Kf)(Pf/Pm-1). Our assumptions in this case make that (1+Kf)(Pf/Pm-1)>0.

3.

The procedures used to obtain these data points are somewhat tedious and space does not allow their full documentation. The authors will be glad to share these, however, with interested readers.

41

Annex 1.2. Simulated Effects of Alternative Pricing Arrangements on Dairy Product Markets: Detailed Aglink Results As noted in the text, the US Federal Milk Marketing Orders are not specifically modelled in Aglink and the pooling of revenues from manufacturing and fluid milk markets in the model is represented only at the national level. Thus the implicit fluid milk premium does not correspond to any of the actual fluid milk price premiums observed. The FMMO Class I differentials at principal pricing points are illustrated in Table 1.2.1. Minimum prices (f.o.b. plants) for milk in fluid use are calculated by adding the Class I differentials to a manufacturing milk price. (Actual market-generated prices are typically higher in FMMO markets). The table clearly illustrates that the FMMO Class I differentials vary considerably across the United States depending on the marketing order region. The minimum Class I differentials reflect the spatial price variation as generated by Pratt et al., which modelled the US dairy industry as of 1995. Aglink baseline uses the manufacturing milk and all-milk prices as reported by USDA\NASS. The Aglink baseline fluid price is calculated implicitly from USDA\NASS manufacturing milk price and all milk price as follows: Fluid milk price = (all milk price * all milk quantity Manuf milk price * Manuf milk quantity) / fluid milk quantity.

Thus, there is only one implicit national fluid milk premium in Aglink. The all milk price is a weighted average f.o.b plant price, and therefore is calculated from marketgenerated prices, not FMMO minimum prices. Table 1.2.1. Federal Milk Order Principal Pricing Points, with Class I Differential Federal Milk Order

Principal Pricing Point

Major City in Principal Pricing Point

Principal Pricing Point differential (USD/100kg)

Northeast

Suffolk Co., MA

Boston

7.2

Appalachian

Mecklenburg, Co., NC

Charlotte

6.8

Southeast

Fulton Co., GA

Atlanta

6.8

Florida

Hillsborough, Co., FL

Tampa

8.8

Mideast

Cuyahoga Co., OH

Cleveland

4.4

Upper Midwest

Cook Co., IL

Chicago

4.0

Central

Jackson Co., MO

Kansas City

4.4

Southwest

Dallas Co., TX

Dallas

6.6

Arizona-Las Vegas

Maricopa Co., AZ

Phoenix

5.2

Western

Salt Lake Co., UT

Salt Lake City

4.2

Pacific Northwest

King Co., WA

Seattle

4.2

Notes:Class I differentials are added to a manufacturing milk price to obtain a minimum fluid milk price f.o.b. plants. Adopted from Federal Milk Order Market Statistics - 2003 Annual Summary http://www.ams.usda.gov/dyfmos/mib/2003_Ann_Sum.htm

The simulations with LFP and LSP scenario prices generated consumption, production and trade levels which were compared to the baseline projections. The results are presented in Table 1.2.2. and Table 1.2.3 as percentage difference between each scenario and baseline values for production and consumption of domestic milk and dairy products. Table 1.2.4 presents the percentage differences between the LFP scenario and the baseline and the LSP scenario and the baseline for the world dairy product prices. The

42 levels of net trade flows are graphically illustrated in Figure 1.2.1 for butter, in Figure 1.2.2 for cheese, in Figure 1.2.3 for SMP and in Figure 1.2.4 for whey powder. Given the marginal nature of the scenarios, the reported changes are relatively small (never above 1%), nevertheless, they clearly show the difference in the impacts the two price support measures have on the dairy markets. Table 1.2.2. Percentage differences between the respective scenarios and the baseline (Consumption) Consumption

Scenario

2002

2003

2004

2005

2006

2007

2008

Fluid milk

LFP

0.193

0.189

0.185

0.187

0.188

0.190

0.190

Fluid milk

LSP

0.066

0.064

0.062

0.062

0.062

0.061

0.061

Butter

LFP

0

0

0

0

0

0

0

Butter

LSP

0.192

0.208

0.205

0.211

0.217

0.219

0.217

SMP

LFP

0

0

0

0

0

0

0

SMP

LSP

0.308

0.307

0.296

0.339

0.391

0.356

0.354

Cheese

LFP

0

0

0

0

0

0

0

Cheese

LSP

0.168

0.178

0.166

0.163

0.165

0.168

0.167

Whey powder

LFP

0

0

0

0

0

0

0

Whey powder

LSP

0.086

0.086

0.081

0.077

0.075

0.073

0.071

Table 1.2.3. Percentage differences between the respective scenarios and the baseline (Production) Production

Scenario

2002

2003

2004

2005

2006

2007

2008

Milk

LFP

-0.012

-0.057

-0.097

-0.131

-0.161

-0.186

-0.208

Milk

LSP

-0.012

-0.057

-0.097

-0.132

-0.161

-0.186

-0.208

Butter

LFP

-0.097

-0.167

-0.226

-0.278

-0.321

-0.360

-0.387

Butter

LSP

-0.049

-0.128

-0.198

-0.261

-0.315

-0.362

-0.398

SMP

LFP

-0.062

-0.155

-0.261

-0.360

-0.449

-0.539

-0.621

SMP

LSP

-0.031

-0.109

-0.212

-0.318

-0.418

-0.522

-0.619

Cheese

LFP

-0.097

-0.167

-0.226

-0.278

-0.321

-0.360

-0.387

Cheese

LSP

-0.044

-0.114

-0.171

-0.220

-0.261

-0.296

-0.322

Whey powder

LFP

-0.064

-0.112

-0.152

-0.187

-0.214

-0.237

-0.255

Whey powder

LSP

-0.029

-0.076

-0.115

-0.148

-0.173

-0.196

-0.212

Table 1.2.4. Percentage differences between the respective scenarios and the baseline (World dairy prices) World Prices

Scenario

2002

2003

2004

2005

2006

2007

2008

Butter

LFP

0.066

0.105

0.117

0.131

0.137

0.139

0.138

Butter

LSP

0.167

0.218

0.218

0.233

0.240

0.243

0.241

SMP

LFP

0.020

0.050

0.072

0.089

0.101

0.112

0.121

SMP

LSP

0.069

0.102

0.112

0.121

0.130

0.138

0.147

Cheese

LFP

0.146

0.199

0.277

0.326

0.356

0.387

0.404

Cheese

LSP

0.326

0.326

0.400

0.435

0.464

0.493

0.506

WMP

LFP

0.028

0.056

0.073

0.084

0.089

0.093

0.095

WMP

LSP

0.060

0.108

0.110

0.114

0.117

0.119

0.119

43 Figure A2.1. Net exports for butter (negative for imports)

Figure A2.2. Net exports for cheese (negative for imports)

10 0

thousand tons

thousand tons

0

-10 -20

-100

-200

-30 2002

2003

2004

Butter base

2005

Butter ’Fluid’

2006

2007

2002

2008

2003

2004

Cheese base

Butter ’Manuf’

Figure A2.3. Net exports for SMP (negative for imports)

2005

2006

Cheese ’Fluid’

2007

2008

Cheese ’Manuf’

Figure A2.4. Net exports for WMP (negative for imports)

240

160

thousand tons

thousand tons

200 160 120 80

120 80 40

40

0

0 2002

2003

2004

2005

2006

2007

2008

2002

2003

2004

Whey powder base SMP base

SMP ’Fluid’

SMP ’Manuf’

Whey powder ’Manuf’ DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

2005

2006

2007

2008

Whey powder ’Fluid’

44

References ABARE (2004), A review of the Australian dairy industry, Hogan, J., Shaw, I., and P. Berry; December, 2004, Canberra, Australia. Australian Bureau of Agricultural and Resource Economics - ABARE (2001), The Australian dairy industry: Impact of an open market in fluid milk supply, Canberra, ACT, Australia. Australian Competition and Consumer Commission (2001), Impact of farm-gate deregulation on the Australian milk industry: study of prices, costs and profits, Australian Competition and Consumer Commission, Dickson, ACT, Australia. Bailey, K. (1997), Marketing and pricing of Milk and Dairy Products in the United States, Iowa State University Press. Bouamra-Mechemache, Z., J.P. Chavas, T. Cox, and V. Réquillart (2002), Price discrimination and EU dairy policy: An economic evaluation of policy options, Paper presented at the European Association of Agricultural Economists conference in Zaragoza, Spain, 28-31 August 2002. Buxton, B.M. (1977), A framework for evaluating the economic impact of classified pricing of milk, Staff paper P77-24, Department of Agricultural and Applied Economics, University of Minnesota, St. Paul. Cox, T.L. and J.P. Chavas (2001), “An interregional analysis of price discrimination and domestic policy reform in the US dairy sector”, American Journal of Agricultural Economics, 83 (1): 89106. Dahlgran, R.A. (1980), “Welfare cost and interregional income transfers due to regulation of dairy markets”, American Journal of Agricultural Economics, 62 (May): 288-296. Fallert, R.F. (1981), “Milk pricing - past, present, the 1980’s”. Journal of Dairy Science, 64 (6): 1105-1112. FAPRI (2003), The Effect on the United States Dairy Industry of Removing Current Federal Regulations, S. Brown, FAPRI-UMC Report #03-03 April, 2003. Gardner, B.L. (1984), “Price discrimination or price stabilization: Debating with models of US dairy policy”, American Journal of Agricultural Economics, 66 (5): 763-768. Helmberger, P. and Y.H. Chen (1994), “Economic effects of US dairy programs.” Journal of Agricultural and Resource Economics, 19 (December): 225–238. Hubbard, L.J. (1992), “Two-tier Pricing for Milk: A Re-examination”, Journal of Agricultural Economics, 43(3), 343-54. Ippolito, R. A. and R. T. Masson (1978), “The social cost of government regulation of milk”. Journal of Law and Economics 19 (1): 33-65. Kawaguchi, T., N. Suzuki, and H.M. Kaiser, “Evaluating Class I Differentials in the New Federal Milk Marketing Order System” Agribusiness (2001) Vol. 17(4): 527-538. Knutson, R.D. (1984), “Governments role in milk pricing - then and now – discussion”, American Journal of Agricultural Economics, 66 (5): 776-777.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

45 Lippert, O. (2001), The perfect food in a perfect mess: The cost of milk in Canada, Public Policy Sources, Number 52, The Fraser Institute, Vancouver, B.C., Canada. McDowell, H., A.M. Fleming, and F. Spinelli, (1990), US Milk Markets Under Alternative Federal Order Pricing Policies. Untied States Department of Agriculture, Economic Research Service, Agriculture and Trade Analysis Division, Commodity Economics Division, Staff Report No. AGES 9068 (November 1990). McDowell, H., A.M. Fleming, and R.F. Fallert (1988), Federal Milk Marketing Orders: An Analysis of Alternative Policies, Untied States Department of Agriculture, Economic Research Service, Agricultural Economic Report Number 598. Nash, E.F. (1961), “The Two-Tariff Milk Scheme”, Journal of Agricultural Economics, 14, 540551. Nubern, C.A. and R.L. Kilmer (1997), “Impact of Spatial Price Discrimination within Florida Dairy Co-operatives”, Agricultural and Resource Economics Review. 26 (1): 94-105. OECD (1996), Reforming Dairy Policy, Directorate for Food, Agriculture and Fisheries, Paris OECD (2001), Agricultural Policies in OECD Countries: Monitoring and Evaluation, Directorate for Food, Agriculture and Fisheries, Committee for Agriculture, Paris. OECD (2002a), The Incidence and Income Transfer Efficiency of Farm Support Measures, Directorate for Food, Agriculture and Fisheries, AGR/CA/APM(2001)24/FINAL, Paris. OECD (2002b), OECD Agricultural Outlook 2002-2007, Directorate for Food, Agriculture and Fisheries, Committee for Agriculture, Paris Pratt, J. E., P. M. Bishop, E. M. Erba, A. M. Novakovic, and M. W. Stephenson (1998), Normative Estimates of Class I Prices Across U.S. Milk Markets. Cornell Program on Dairy Markets and Policy, R.B. 98-05, 1998. Sumner D.A. (1999), “Domestic price regulations and trade policy: milk marketing orders in the United States” Canadian Journal of Agricultural Economics-Revue Canadienne D Agroeconomie, 47 (5): 5-16. Testuri, C.E., R.L. Kilmer, and T. Spreen (2001), “Seasonality of Class I Price Differential Estimates for the South-eastern United States” Journal of Agricultural and Applied Economics, Vol. 33(3): 591-604. Takayama, T. and G.G. Judge, Spatial and Temporal Price and Allocation Models. North-Holland Publishing Company, Amsterdam, 1971. USDA (2004a), Effects of U.S. Dairy Policies on Markets for Milk and Dairy Products. Price, J. Michael, U.S. Department of Agriculture, Economic Research Service, Technical Bulletin No. 1910, March 2004.

USDA (2004b), Economic Effects of U.S. Dairy Policy and Alternative Approaches to Milk Pricing, USDA report to Congress, July 2004. USDA (2004c), Federal Milk Order Market Statistics - 2004 Annual Summary. United States Department of Agriculture Marketing and Regulatory Programs, Agricultural Marketing Service Dairy Programs, http://www.ams.usda.gov/dyfmos/mib/2004_Ann_Sum.htm

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

This page intentionally left blank

47

CHAPTER 2 TRADE AND ECONOMIC EFFECTS OF MILK QUOTA SYSTEMS

Abstract The analysis in this chapter illustrates some specific market, trade and welfare implications of operating milk quota systems. Milk production quotas were typically introduced as a tool to control the growth of surplus production and budgetary expenditures in order to improve the political sustainability of high price support. They have also been used as a rural development policy and as a producer price stabilisation tool. Quota systems increase the transfer efficiency of market price support (MPS) by reallocating part of the price support benefits from input suppliers directly to farmers in the form of quota rent. Although consumption is restrained by high prices, quotas reduce the impacts of excess production resulting from MPS on trade and world markets. The impact of quota systems very much depends on the level at which the production limit is set and on the adjustments that are made in other policy tools. However, a quota system comes with its own set of problems, in particular those due to the inefficiencies that it may create, the costs that it imposes on consumers, the difficulties and costs of administration that may arise for governments, the difficulty in setting the quota at a level that would match production (or trade) under free trade conditions and the vested interests that it generates. The existence of quota systems also requires the continuation of high border measures, that is, quota systems allow a domestic market to be managed only if that market is isolated from external sources of supply. This, however, is uncertain in the context of multilateral trade reform. Furthermore, a quota right is a licence to sell milk at the supported price and as such it is an incomegenerating asset. With time, the value of quota, reflecting the difference between an underlying cost of production and the milk price at the quota level, becomes incorporated into the cost structure of dairy farms. Thus, quota imposition provides gains for initial beneficiaries, but subsequent generations can be locked into a higher cost structure. While initially a quota system might be seen as an attractive alternative, the vested interests and inefficient cost structures that are inherent to a quota may complicate reform efforts later on.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

This page intentionally left blank

49

Introduction The concept of production quotas or supply management in agriculture is not new and can be traced back to the early 20th century. Generally, a production quota is a limit imposed on the quantity produced. It could have the character of assuring the minimum required production level where under-production might be penalised. However, typically production quotas in OECD countries have a production restricting character and are combined with penalties for exceeding the quota limit. This chapter does not intend to present a detailed analysis of every possible aspect of operating a quota system. Indeed, imposing a limit on supply influences all facets associated with production. Quota could influence structural changes in agriculture, the structure of the dairy processing sector, welfare of producers and input suppliers (to some extent consumers),1 the value of assets in agriculture, production risks, the uptake of new technologies and, of course, production levels and trade. There is a large body of literature dealing with many individual issues related to quota. The quota system in this paper is treated and examined as one possible policy tool. Based on a description and analysis of the reasons for and impacts of milk quotas, the purpose of this paper is to contribute to the understanding of ways to reform or eliminate them. Generally speaking, milk quotas were often introduced to control the growth of surplus production and budgetary expenditures, to maintain market price support, and to provide price stability for dairy farmers. Quotas have thus been largely implemented as a second best alternative to reducing support that has allowed policy makers to continue using a high price support without necessarily aggravating budgetary problems. A quota is a powerful tool which gives policy makers (or producer groups sanctioned by government) a direct control of agricultural product supply. As such, a quota interacts with other policy tools in pursuing defined policy objectives. The importance of quota systems in the OECD dairy sector is demonstrated by the fact that currently more than half of all OECD milk production is governed by quotas. Each quota system implemented in an OECD country has its own special features and has evolved through time. A quota inherently assumes a value, which is likely to be its most important attribute from a policy reform standpoint. The quota is typically a licence to sell milk at the supported price and as such becomes valuable in its own right. The value of the quota, reflecting the difference between an underlying cost of production and a milk price, becomes incorporated into the cost structure of dairy farms with time. Thus, while initially a quota system is often seen as a viable and politically feasible tool, the vested interests and inefficient cost structures that are inherent to a quota may hinder reforms on price support later on. The concept of quota value is often misunderstood and its importance underestimated. This chapter presents basic fundamentals that try to facilitate an understanding of the emergence of quota values following its imposition. Important welfare effects of a quota system for owners of farm resources and suppliers of inputs, which are often neglected by standard textbook welfare economics, are evaluated analytically and empirically in the text. In the context of discussions of global liberalisation of dairy markets, the question arises how a supply managed dairy sector would respond if quotas were to be removed alongside border protection. However, the assessment of milk production potential in quota operating countries is complicated by the absence of historical evidence concerning DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

50 milk supply response as quotas in respective countries are in place already for a considerable amount of time. Given the importance of this response, the aspect of modelling dairy policy reform in the presence of production quotas is discussed here. The remainder of the chapter is organised as follows. The ensuing section presents a brief overview of general characteristics of milk quota systems implemented in OECD countries. The second major section discusses theoretical issues of milk production quotas focusing on the quota value. The third section examines analytically and empirically the quota interaction with other policy tools and impacts of changes in quota levels on markets, trade, prices, income and welfare using the Aglink and PEM models. The fourth section then describes the modelling of milk supply response in quota countries with the Aglink and PEM models which is important in the context of the dairy trade liberalisation analysis presented in Chapter 3. The final section of this chapter draws conclusions.

Quota programmes – General overview Milk production quotas were typically introduced as a tool to control the growth of surplus production and budgetary expenditures in order to improve the political sustainability of high price support. Quotas were also introduced with the view to stabilise prices and hence the income of farmers.2 Dairy quota systems vary considerably among OECD countries and are often governed by different mechanisms which can impact on milk production and trade in different ways. In the OECD area, a milk supply management system is operated in the European Union, Canada, Switzerland, Norway and Japan (Japan’s quota is operated by co-operatives).3 ,4 The section below briefly describes the main characteristics of milk production quota systems.5 Typically, the setting of production quotas relates to the amount of milk being shipped from dairy farms. So defined, the quantity limit is easier to observe than on farm production of milk. The total quantity supplied is usually bound at both national and individual farm level. However, dairy quota programmes might have a different coverage of supply. For example, the quota system previously implemented in Australia concerned fluid milk only while the total milk supply had not been bound.6 Market segmentation might be used to specify different quotas for fluid and manufacturing milk.7 This system operates in Canada, for example. Under the milk supply management system, the Canadian Milk Supply Management Committee (CMSMC) sets a national production target – the Market Sharing Quota (MSQ) – for industrial milk. The MSQ is set with the goal to achieve a domestic market balance in terms of butterfat, and is assigned to provinces largely on the basis of historical shares. The CMSMC monitors the evolution of the MSQ on a monthly basis based on the monthly production and demand situation in order to be market responsive and avoid over-quota production. In addition to the MSQ, each province controls its own production quota for fluid milk, and the entire milk quota – industrial and fluid together – is allocated to producers. When a quota system is implemented an important decision concerns setting of the actual level of quota. The quotas typically have a production limiting character where the reduction of total milk supply might be organised over a period of time progressively or regressively.8 A quota system usually provides for a mechanism of adjusting the quota amount in order to accommodate the evolution in internal and external markets. The quota adjustment might be conducted on a monthly or an annual basis, or might be subject to discretionary changes as they are considered appropriate, possibly in the context of less frequent, major revisions of underlying legislation. The former might be DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

51 seen in Canada while the latter in the European Union. For a quota system to be effective, compliance with the quota is a crucial factor. Non-compliance is always penalised, although the magnitude of penalties and stringency of enforcement differ across systems. Quota management is an important feature of a dairy quota programme. In some systems quotas are allowed to be marketed, in others, quotas are managed by administrative means. Under the marketable quota systems producers have the possibility to lease or sell/buy quota. The former practice permits a producer to temporarily lease (typically on an annual basis) additional quota and avert the possibility of penalty for over the quota production. On the other hand, a relatively less efficient producer may make more money by leasing quota out to more efficient competitors than by producing within quota. Selling or buying of quota represents the possibility to transfer quota permanently. It is often the case when quota is transferred between producers that it must be accompanied by farm assets such as land. In any case, many OECD countries’ quota regimes contain some restrictions that dissect the total into regional amounts that limit leasing and trading possibilities. In some systems, producers can transfer so-called used quota which the buyer can only use in the following marketing year, or so-called unused quota which can be utilised in the current marketing year. The difference in price of used and unused quota would be likely equal to the prevailing lease price. An example of quota management, by administrative means, is illustrated in Box 2.1. Another important feature of the quota system is its duration as given in the enabling legislation. Quota programmes may be of finite duration, with or without the renewal option, or of indefinite duration. As quota values become incorporated into the farm cost structure with time, the duration of the quota programme might itself influence the feasibility of programme termination or continuation. The theoretical issues of quota and emergence of quota value is explained in the following section. Box 2.1. Administrative quota management in France In France, quota is not managed on a commercial basis but by administrative decision, mainly at the level of the département. Quota-allocation decisions are based on priority criteria such as, for instance, the support of farmers with the lowest quota or new entrants. Hence, the administrative form of management attempts to keep down farm set-up and structural costs, as farmers do not have to bear the quota management costs they would incur in a commercial system. Moreover, under the French quota administration, the quotas are tied to land, which limits their tradability. Thus, it is not possible to sell quota independently from land (This does not prevent a market valuation of quota but this way quota price is small as the value of quota is carried via the land price). In addition, the levy charged for transfers leading to an increase in farm size allow the quota concentration to be limited and maintains the medium size of farm holdings. The administrative form of management contributes to a territorial development policy and is aimed at (i) preventing the abandonment of dairy farmland and maintaining dairy farming and its related services throughout the country, and (ii) limiting the regional concentration of farms. The administrative management of quota tries to avoid the social externalities resulting from concentration of production at farm level and in specific areas that might ensue if profitability was the sole criterion. The system thus endeavours to preserve the regional production spread, to maintain dairy farming in remote parts of the country such as mountainous areas and to slow down the drift away from traditional farming systems. In spite of that, in certain regions a concentration of dairy farming can, nevertheless, be observed, with smaller producers being crowded out. At the same time, there is also evidence of increased intensification (Rainelli and Vermersch, 1997) stemming from the fact that fixed quotas are capitalised into land prices.

Theory of milk production quotas - Overview The economic theory relating to how quotas impact markets or, more specifically, on supply, resource allocation and welfare is well established.9 The pernicious welfare effects of restrictions on production activities in the absence of externalities are well DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

52 known. Typically, the welfare effects of quotas are compared with the free market situation with the standard conclusion that quota systems are inefficient and cause considerable transfers from consumers to producers (for an example, see Veeman 1982), Harvey (1984) argued that in a political-economy context adoption of quotas (or supply management) may result in increased welfare as measured against a status quo policy that generates even greater distortions and misallocation of resources. Although Guyomard and Mahé (1994) agreed that, in a static approach, quotas could be a welfare improving policy instruments when price cuts are impossible to implement, they argued that a dynamic approach suggests that welfare gains to be expected from the corrective quota instruments are overestimated. The welfare implications of quota imposition as compared to price cuts, the two policy options available to reduce large surpluses, are illustrated in detail in Annex 2.1. Although the analysis presented in the annex shows that the net welfare gain of the support price cut is larger compared to the net welfare gain of quota limit imposition, the reduction in support prices for dairy products might be too large to be politically feasible. That is, restraining the gap between domestic production and consumption by means of support price cuts might cause politically unacceptable dislocation to the farm sector. On the other hand, quotas allow policy makers to bring the underlying trend in supply under direct control and at the same time permit them to use producer price as an instrument for the pursuit of other policy objectives without direct consequences for supply. Moreover, a reduction in price support levels keeps the marginal and average revenue of output elastic at the new (lower) price and does not change the supply incentives of support. Thus, in case the supply curve shifts to the right due to cost reduction and technological change, it inevitably causes production to increase at the existing level of support price and thereby continues to aggravate the surplus problem in the absence of a quota, despite the initial price support cut. Indeed the rightward shift of the supply curve would justify an additional reduction of support prices but the support price reduction is typically a politically sensitive issue.10 It follows that a quota system gives policy makers a direct control of agricultural product supply without the need to compromise the policy objective of high price support. As such, given that most studies show that farmers are generally risk-averse (Just (1974), Chavas and Holt (1990)), the package of price support and supply management offers producers that are risk-averse, greater price stability. However, with the imposition of a quota, several inherent quota features should be considered. Firstly, after quota imposition low-cost efficient milk production is impeded at the expense of high-cost inefficient production. That is, typically, quotas are distributed on the basis of historical production levels rather than efficiency criteria and all producers experience the same percentage cut in their production. As a consequence, some low-cost, efficient production is lost whilst some high-cost, inefficient production is maintained. The effect of a quota regime imposition on the efficiency of milk production is elaborated in detail in Annex 2.2. The analysis presented in the Annex 2.2 stipulates that when a quota system allows quota to be traded or leased, the efficient producers would lease or buy quota from less efficient producers and the rental price in a competitive market would be bid to a rate equal to the difference between support price and marginal cost. It should be noted that these results, where milk production is reduced while prices are held at the existing support levels, is analytically equivalent to the situation where a quota is imposed at the current production level but the support prices are increased subsequently. Oskam and Speijers pointed out that in general producers with low marginal costs will be willing to DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

53 buy or lease quotas from producers with higher marginal costs. They noted that pressure to allow for quota markets to develop and to increase quota mobility is inherent in a quota system and improving quota mobility by organising quota trade or leasing would speed up structural development. The analysis in Annex 2.2 also illustrates another important feature inherent in the quota – an emergence of quota value following quota system implementation. The value of quota, or the quota rent, can be understood as the discounted sum of the future stream of net benefits to producers arising from holding the quota. It could be argued that the value of quota is perhaps its most important attribute from a dairy policy reform standpoint. Dawson (1991) argues that the main criticism of quota systems concerns precisely the value. However, it should be stressed that quota carries a value whether it is tradable or not. Given the fact that producer support is usually tied to the quota, it is less profitable and often not feasible to supply milk without quota. For this reason, a quota right is an income-generating asset and assumes a value for the person who controls it. Thus, even if quota does not assume a price value, the quota has an implicit value that does affect decision-making when, for example, policy is debated or farm assets are transferred or sold. Given the fact that quota assumes a value, the producer welfare impacts of quota policies are not straightforward.11 If the quantity supplied at current support prices is restricted by a quota level, then one may argue that producers would loose part of their producer surplus (Annex 2.1). But would they? The simple analytical framework found in standard textbooks on welfare economics typically assumes that producer surplus accrues to the owner of relatively fixed assets (typically land, in the case of farmers) under the condition that supplies of variable factors are perfectly elastic. However, in reality, the supply of inputs is not infinitely elastic and therefore the producer surplus is distributed across farmers and other input suppliers. Thus, in the long run the benefits of market price support are shared by farmers’ own resources and by input suppliers. It is not easy to illustrate this phenomenon in a simple graph. Nevertheless, the following representation reveals important welfare implications of quota imposition - perhaps with repercussions also for prospective quota removal options. Figure 2.1 illustrates the partition of surplus on the basis of farmer owned resources and purchased input suppliers. For ease of exposition, it is assumed that 50% of surplus goes to farmer’s owned resources and 50% goes to input suppliers. 12

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

54 Figure 2.1. Quota imposition favours farm owners at the expense of input suppliers

Ps1 S

d Ps0 a

c z

x

PC

e

b

D

y

Q*

Q

Source: OECD Secretariat.

The figure shows that prior to quota imposition farmers would produce Q tonnes of milk at the price PS0 with the producer surplus equal to the sum of the areas a+b+c+x+y+z. By construction of the experiment the area a+b+c represents 50% of the total producer surplus and depicts return accruing to farmers’ own resources under the assumptions of this simple exercise. The area x+y+z represents the remaining 50% of the total producer surplus and are the returns accruing to input suppliers, again, reflecting the assumption explained in the previous paragraph. If trade in dairy products is fixed such that the domestic price of milk is set to clear the domestic market, then after applying quota of Q*, a price rise from Ps0 to Ps1 is required, and the quota assumes value corresponding to area a+x+d while returns to factors other than quota are reduced to area b+y. The quota system results in input suppliers losing an amount equal to area x+z. Farmers (to the extent that they hold relatively fixed assets, such as land) loose c but gain x, formerly input suppliers surplus, as part of quota rent.13 If domestic demand is determined by a target price, such that trade is determined by excess supply, then assuming the domestic price after application of a quota remains at Ps, quota value will be equal to a+x and return to factors is reduced to b+y at a marginal cost of Pc. Again, input suppliers see their returns reduced by x+z. Farmers lose factor rent a+c but gain a+x as quota rent; x is a transfer from input suppliers to producers because of the quota.14 In reality, the net gain to farmers is conditional on the share of surplus split between farmer’s owned factors of production and suppliers of purchased inputs and the size of the production restriction. From a political economy point of view, that owners of farm resources, following a quota imposition might gain part of the producer surplus of input suppliers is a positive development. After all, improving the welfare of farmers is a common goal of agricultural policy. Nevertheless, the benefits of quota in terms of

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

55 producer surplus will be in the long run capitalised into the value of quota.15 If quota is tied to land, the benefits will be capitalised into the value of land (see Annex 2.3.). This is indeed a general problem of any increase in farm net returns and is not unique to quota systems. The added complexity in quota systems is that the share of benefits flowing to owners of farm resources is magnified at the expense of input suppliers and the rent accruing to quota reduces the surplus accruing to traditional resources.16 17 The capture of input supplier surplus as quota rent explains, in part the high transfer efficiency of quota programmes. Transfer efficiency is defined as the ratio of farm income change to change in programme expenditure, in the form of either consumer or taxpayer costs. More generally, removing the ability of producers to react to price changes at the margin allows for related market price support policies to be highly transfer efficient, as this production response is a key determinant of transfer efficiency (OECD 2001). To illustrate the change in transfer efficiency of price support resulting from the imposition of a quota system, consider two alternatives using the setup in Figure 2.1. The first is an increase in price support (either as MPS or output support payments) from Pc to Ps0 without quota, and the second an increase from Ps0 to Ps1 with quota set at Q*. The first case, increasing price from Pc to Ps0, without quota, induces a production increase from Q* to Q, with a cost in terms of MPS level or required total payments equal to the area a+c+e+x+z. Of this, the producer gets a+c, the balance lost to input suppliers and deadweight, and the ratio (a+c)/(a+c+e+x+z) defines the transfer efficiency. In the second case, where support applied to raise prices from Ps0 to Ps1 with production fixed by quota at Q*, programme cost is equal to area d, and the increase in producer welfare (through quota rent) is also d, yielding a transfer efficiency of 1, the highest possible.18 When quotas are used to reduce supply the marginal cost of production of the last “quota” unit falls (compared to higher non-restricted production) and the intensity of use of fixed factors is reduced (Annex 2.3). As a consequence, the surplus earned by the “traditional” farm resources fall and the pure profit is bid up into the value of quota. Quota becomes an asset against which loans may be secured, possibly exacerbating the difficulties in reforming dairy policies involving a quota constraint. Moreover, with time, the quota value becomes incorporated into the farm cost structure, resulting in an increase in average costs so that in the long-run, as Dawson puts it, only the administrative nuisance of supply control remains. The report on feasibility of phasing out quota in the European Union goes even further stating that the system no longer achieves its major objective of stabilising and improving dairy farm incomes (Colman, 2002). This long-run effect brings into question the durability of the above-described transfer efficiency of policies involving quota constraints. If quota rents are not retained by current producers, but stay with those initially vested with the quota, then high transfer-efficiency of price support connected with quota restrictions is a transitory phenomenon. There are other important issues related to quota systems such as the impact on structural change or uptake of new technologies (Oskam and Speijers; Hennessy, 1995). Although Bailey (2004) argues that structural changes in the EU have been slower under the quota system than would otherwise have been the case, restructuring of the dairy sector has continued all the same, albeit with different geographical connotation. That is, while on average the number of farms and cows in the EU has continued to fall as yields improved over time, these numbers have remained more stable in less favoured areas. Hence, it could be argued that the presence of quotas, as they are typically tied to land, might preserve farming structures and allow farming practices in disadvantaged areas to prevail.19 Nevertheless, the question is, whether the policy objective to preserve farming DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

56 in less favoured areas could not be better achieved by direct measures as compared to a general quota system. The answer to such a question requires welfare analysis reflecting production jointness and presence of transaction costs. A different issue, although related to structural adjustment in the sector, concerns the situation where some producers completely cease milk production and continue to annually lease their quota. It is arguable whether the income from quota could accelerate the restructuring process and stimulate the cessation of production in these cases.20 However, it would be difficult to justify the permanent stream of income transfers from the industry to non-producing quota holders. Moreover, with a quota system in place it become “more expensive” to start dairying as this requires the purchase of quota which brings an important question of entry to the industry and the renewal of the dairying population. In this respect, nevertheless, the type of quota administration plays an important role and measures can be and have been adopted that subsidise or otherwise enable entry of new producers which have been discussed elsewhere.21

Quota interactions with other policy objectives While the sections above discussed the theoretical background of quotas and the emergence of quota value this section looks at quota as a policy tool and discusses the market and welfare impacts of changing levels of quota within a context of specific policy objectives. The setting of the level of quota represents an important policy decision which interacts with the effects of other policy tools. The impacts on world and domestic dairy markets of changes in the level of quota are conditional on the decision regarding other policy objectives. That is, the quota level has a direct influence on exports and government expenditure on subsidised exports within the objective of holding the supported domestic price unchanged. On the other hand the level of quota determines the cuts in domestic prices required to achieve other objectives such as holding exports or government expenditure on subsidised exports unchanged. The effect of quota policies on income is not obvious, as its impact is best seen as being part of a body of policies, and the impact attributable to the quota policy itself is dependent on the other polices of which that body is composed. A fruitful way of examining the impacts that quota polices have on indicators of interest such as farm income and production, is to look at how the mix of policies has to change in order to target a certain result for one of these indicators. The relationship between the quota and certain policy objectives might be illustrated using a simple diagram. Figure 2.2 schematically depicts the trade-off between the supported domestic price and the quota level given that the policy objective is to leave government expenditure on subsidised exports unchanged. Consider that at the initial level of the supported domestic price PS and milk quota Q* consumption equals quantity QDS of milk while Q*íQDS is exported with export subsidies equal to (Q*íQDS )×(PS í PW ). Holding the supported domestic price constant and increasing the quota to the new level QN* increases taxpayers costs (export subsidies) by (QN*íQ*)×(PS í PW ).22 Nevertheless, policy makers can, for a given level of quota, reduce the supported domestic price so that government expenditure (taxpayers’ costs) on subsidised exports would not be affected. Figure 2.2 illustrates that in order to keep government expenditure unchanged, the domestic price has to be reduced to a new level PN for which the lighter shaded area equals the darker shaded area in the diagram. That is, the export subsidies before and after quota increase must be equal; mathematically expressed (Q*íQDS )×(PS í DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

57 PW ) = (QN*íQDN )×(PN í PW ). Note that under the new price (PN) consumers will consume a higher quantity (QDN.).23 Figure 2.2. Interaction between quota level, domestic price, exports and government expenditures

P

S*

D

SN*

S

PS PN

PW S D

QDS QDN

Q*

QN*

Q

Source: OECD Secretariat.

A different scenario can be constructed to evaluate the increase of quota with the objective of holding the volume of dairy product exports constant. In the analytical framework of Figure 2.2 this scenario could be described as follows: “By how much would the price (PS ) have to be lowered to a new level (PN ) so that for a given increase in quota (from Q* to QN*) the volume of exports remains constant (Q*íQDS = QN*íQDN ).” Note that in this scenario the darker shaded area in Figure 2.2 would be smaller than the lighter shaded area, suggesting that government expenditure on exports would be reduced, by how much remains an empirical question. In order to evaluate numerically the relationship between the quota level, supported domestic prices and welfare under specific economic parameters and policy objectives, empirical analysis has been carried out for a representative OECD country using the Secretariat’s models: the partial equilibrium model Aglink and the policy evaluation model PEM. Aglink results of policy simulation analysis- Evaluating the market and trade impacts Aglink is a policy specific, partial equilibrium, dynamic model (see Annex A for the description of Aglink). The simulation experiments are conducted using the baseline data of the Agricultural Outlook baseline 2003-2008 published in OECD (2003). The dairy component of this model covers production and consumption of milk and main milk products in major OECD and several non-member economy markets, covering both DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

58 importers and exporters. Thus, the Aglink representation of the dairy sector allows to analyse a scenario impact on world markets in tradable dairy products that are explicitly modelled. Following the analytical example of Figure 2.2, the specific question to be addressed by the first empirical experiment is “how much would domestic prices have to be lowered to accommodate a given increase in milk quota while holding government expenditures on export subsidies constant?” As the level of the quota is exogenous in only one country/region in Aglink – the European Union (EU) – the EU module is used to set up the scenario.24 25 While Figure 2.2 depicts the analytics in terms of milk price and milk quantities, in reality milk is often priced and traded in the form of dairy products. The dairy market is also represented in this way in the Aglink model. Thus, theoretically there are a large number of permutations for adjusting individual dairy product exports while holding the overall government expenditure on exports constant. For the sake of transparency, the objective of holding the government expenditures on exports constant is achieved by holding the government expenditures on exports for each dairy product constant at the baseline level. To accommodate this assumption, the structure of the model has to be changed as follows. The dairy product prices, which are market clearing variables in the baseline, are changed to: PPi=XPi + BaseVALi / EXSi where i stands for individual dairy products, PP is the domestic dairy product price, XP is the dairy product world price, EXS is the volume of subsidised exports and BaseVAL represents government expenditures on exports in the baseline (calculated as EXSi×(PPi í;3i)).26 In this scenario, the volume of subsidised exports becomes the market clearing variable.27 The above equation determines the direct interaction between the volume of subsidised exports and prices for each quota level. Other policy instruments remain at their baseline level. This construction ensures that government expenditures on subsidised exports for each dairy product will be equal to the baseline level after the increase in the milk quota. However, it should be noted that holding government expenditures on exports constant does not guarantee that world prices would not be affected. The reduction in domestic prices would lower per unit export subsidies, allowing a greater volume of subsidised exports for the same amount of expenditure. Note that export subsidies are limited by the WTO both, in volume and value terms, which prevents any increase over these limits. In this respect, the scenario must be viewed as purely illustrative as no account is taken of the respective WTO limits on the volume of subsidised exports. In order to evaluate the market impacts of changes in the quota level, the experiment was undertaken with 1, 1.5 and 2% increases in milk production quota respectively. Table 2.1 provides results that address the initial question about how much domestic prices would have to be lowered to accommodate a given increase in milk quota while holding total expenditures on export subsidies constant. As illustrated in the table, the reduction in the EU internal prices of dairy products combined with fixed subsidised export expenditures allow greater exports for all dairy products. Thus, for example, if the milk quota were to be increased by 1%, then the required stability in export subsidy expenditures would be achieved by a simultaneous reduction in the butter price of more than 3% and an increase in exports of butter by 5.1%. The producer price of milk in this scenario would fall by 2.4%. As is apparent from the table, DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

59 world prices for all dairy products would be reduced due to the increased exports from the European Union, which is a dominant player on world dairy markets. Again, these results are purely illustrative as no account is taken of the respective WTO limits on the volume of subsidised exports. Table 2.1. Impacts of quota increases on key variables (average changes from baseline for the EU) assuming constant government expenditures on subsidised exports

Quantity

Domestic Prices

Subsidised Exports (Volume)

Government Expenditure on Subs. Exports

World Prices

% change

% change

% change

Milk

1.0

1.5

2.0

Butter

-3.1

-4.6

-6.0

Cheese

-1.4

-2.0

-2.7

SMP

-0.9

-1.4

-1.9

WMP

-1.2

-1.8

-2.4

Milk

-2.4

-3.6

-4.7

Butter

5.1

7.8

10.5

Cheese

2.3

3.5

4.6

SMP

4.5

6.8

9.2

WMP

2.7

4.0

5.4

Butter

0.0

0.0

0.0

Cheese

0.0

0.0

0.0

SMP

0.0

0.0

0.0

WMP

0.0

0.0

0.0

Butter

-0.6

-0.9

-1.2

Cheese

-0.3

-0.4

-0.6

SMP

-0.1

-0.2

-0.2

WMP

-0.6

-0.9

-1.2

Source: OECD Aglink model.

The results of the scenarios with 1.5 and 2% increase in quota are similar to those of the 1% quota increase. In the absence of the respective WTO limits on the volume of subsidised exports, the larger the increase in quota the larger the impact on the key variables presented in Table 2.1. The results illustrate that a 1.5% increase in the milk production quota would be neutral in terms of expenditure on subsidised export of dairy products if the supported domestic price is reduced by 3.6%. For the 2% milk quota increase, the milk producer price would have to be reduced by 4.7%. It is interesting to note that butter prices would have to be reduced substantially more than those for SMP. These results stem to some extent from the fact that in Aglink the EU demand for fat is specified as being less elastic than demand for non-fat solids. The results of the scenario where the policy objective is to increase quota while keeping export volumes fixed are reported in Table 2.2. The Aglink simulations are undertaken again with the EU model component of Aglink. Table 2.2 shows the impact of different increases in the level of quota (again 1, 1.5 and 2% respectively). Comparing the results of Table 2.1 and Table 2.2, the latter shows more profound cuts in dairy product and milk producer prices as exports are not allowed to increase relative to the first case

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

60 with fixed expenditure. For example to accommodate a 1% increase in quota, holding the export volume constant, would require a 3% reduction in the producer price of milk. Again, the results of the scenarios with 1.5 and 2% quota increases are similar in direction to those of the 1% increase but confirm that the larger the quota increase, the greater the impact on the key variables. Thus, the 2% increase in quota would result in a milk producer price reduction of 6%. As in the first experiment butter prices would be reduced the most, followed by those for WMP, cheese and SMP. Government expenditures on subsidised exports would be reduced for all dairy products with the highest reduction seen for butter, again followed by WMP, cheese and SMP. Table 2.2. Impacts of quota increases on key variables (average changes from baseline for the EU) assuming constant volume of subsidised exports

Quantity

Domestic Prices

Subsidised Exports (Volume)

Government Expenditure on Subs. Exports

World Prices

% change

% change

% change

Milk

1.0

1.5

2.0

Butter

-4.9

-7.3

-9.7

Cheese

-1.9

-2.8

-3.7

SMP

-0.2

-0.3

-0.4

WMP

-2.1

-3.1

-4.1

Milk

-3.0

-4.4

-5.9

Butter

0.0

0.0

0.0

Cheese

0.0

0.0

0.0

SMP

0.0

0.0

0.0

WMP

0.0

0.0

0.0

Butter

-8.7

-12.9

-17.0

Cheese

-3.4

-5.0

-6.7

SMP

-0.7

-1.0

-1.4

WMP

-6.6

-9.8

-12.9

Butter

0.3

0.4

0.5

Cheese

-0.1

-0.1

-0.2

SMP

0.0

-0.1

-0.1

WMP

-0.2

-0.2

-0.3

Source: OECD Aglink model.

As the volume of exports is held at the baseline level, the scenario could be expected to have a negligible impact on world dairy prices. However, as Table 2.2 indicates, the impact on world dairy prices is non-trivial. This result begs more explanation. After the United Kingdom accession to the European Union, New Zealand, which is the biggest exporter of butter in the World, was granted a market access quota for butter to the EU market. Thus, New Zealand producers benefit from their special ability to sell some of their butter at EU support prices rather than at lower world prices. It follows that the New Zealand butter export price is partly determined by the world butter price and partly by that on the EU domestic market. As the latter price falls in the Aglink scenario, it reduces the rent accruing to New Zealand producers and ultimately reduces the butter prices in New Zealand as well. Lower butter prices in New Zealand lead to a shift in dairy product output - and a small reduction in total butter production – with some consequences for world markets as shown in Table 2.2.28 DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

61 PEM results of policy simulation analysis — Investigating the connections between quota and welfare The Policy Evaluation Model (PEM) is a partial equilibrium static model including 5 major commodity categories and covering six countries plus a rest-of world component (see Annex A for a description of the PEM model). It is calibrated to a 2002 base year using primarily data from the OECD PSE database.29 Elasticity parameters are selected to give the model a medium-term (5-7 years) adjustment horizon. The dairy sector is represented in terms of raw milk equivalents. A single supply of raw milk is demanded by separate domestic markets for milk for fluid and industrial uses. Industrial milk (dairy products are expressed in raw-milk equivalent terms) is a tradable commodity in the model. The slope of the supply function is implicitly determined by the factor supply elasticities and the share of each factor in production, and the supply function is located by the specification of a unit quota rent which defines the difference between producer price and marginal cost at the initial equilibrium point. In this section, the PEM will be used to investigate the interactions between quota and other policies (primarily market price support) related to achieving specific policy outcomes. The focus of the analysis is on welfare impacts in the EU and Canada. All scenarios are combined with a sensitivity analysis that varies all parameters within their plausible ranges as identified in the reports by Abler (2001) and Salhofer (2001). In addition, unit marginal cost is also varied in a range between one half and double its base value. A Monte Carlo approach is used where all parameter values (plus unit marginal cost) are drawn from uniform distributions, and the scenario run using this new parameter set. This is repeated for 500 draws, and the extreme high and low values for all results are identified. These maximum and minimum results define the range of results that alternative but plausible parameter choices would yield. In many cases, this sensitivity analysis produces no variation in the results in each scenario. This is the case where the result is normally constrained (e.g. production under quota), constrained as part of the scenario (e.g. constant export subsidy levels), or immediately derived from such a fixed value. Where this is the case, the sensitivity results are omitted from the reported results in order to conserve space. Policy makers are interested in the impact of policy reform on welfare, and particularly farm income. In fact, policy reforms in the EU have often contained explicit provisions of compensation for any resulting reduction in farm income. The ambiguity of farm income changes for changes in quota was illustrated in Figure 2.1. As discussed above, total quota rent is a rectangle where the height represents the unit quota rent, i.e. the difference between the producer price and the shadow price (marginal cost of production at the quota level), times the quantity produced.30 Additional production increases total rent by increasing the “quantity” side of this rectangle (by the amount of the quota increase). However, any increase in production entails a move up the supply function, resulting in an increase in marginal cost as the price of farm inputs are bid up due to increased factor demand. This increase in cost reduces unit quota rent, shrinking the gap between the producer price and shadow price. Thus, for an increase in level of quota, the rectangle that defines total quota rent increases in one dimension and decreases in the other. Farm income is composed not only of total quota rent, but also of returns to factors of production owned by the farm household. Any production increase due to an increase in quota will raise these returns through increases in factor demand. In fact, part of the DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

62 reduction in unit quota rent (the vertical part of the quota rectangle) on total quota rent is returned to producers through their ownership of factors of production. The other part of this reduction in unit quota rent is received by suppliers of purchased inputs who are not farm households. The ability of purchased input suppliers to capture some of this increased surplus is the main cause of farm income losses due to quota changes. This sharing of total producer surplus between farm households and input suppliers was discussed above. Farm income stands alone as having components derived from quota as well as from factors of production. For other actors in the model, welfare changes are calculated as changes in consumer or producer surplus resulting from price changes and with reference to demand or supply functions. For example, the change in consumer welfare is the area defined by the change in consumer price and the demand function (e.g. area f in Figure 2.4.). Absent an exogenous shift in the consumer demand or input supply functions (none are considered here), there can be no welfare change without a price change in these markets. The terms “change in consumer surplus” or “change in producer surplus” are synonymous with consumer and supplier welfare changes. Each input in the model is represented by a single supply function. The change in welfare for input suppliers as a whole is presented by summing up the change in producer surplus for all purchased inputs in the model. For farm households, the change in producer surplus for each farm own factor is reported separately, though the method of calculation is the same in all cases. Taxpayers pay for the budgetary component of agricultural programmes. Their welfare changes as this expenditure changes. Accordingly, the change in taxpayer welfare is equal to the negative of the change in budgetary expenditure on agricultural programmes. If spending decreases, taxpayer welfare increases. This item is reported with a subheading specifically for changes in budgetary expenditure on export subsidy programmes, as this is a programme category of particular interest here. Total welfare change for the EU is the sum of the welfare changes for producers, consumers, input suppliers, and taxpayers. Changes in total welfare are driven by changes in prices in the input markets (for farmers and input suppliers), in the output market (for consumers and farmers), and by changes in budgetary expenditures on programmes (for taxpayers). Any changes in welfare in other countries or other sectors that result from the shocks that occur in each scenario are not considered. If the country in question is large enough to influence world prices, any increase in exports brought about by the quota increase will put downward pressure on world prices. This decline in world prices will be transmitted into the domestic market, pushing down domestic prices (assuming a constant domestic-world price differential that allows transmission of price changes from world to domestic markets). Any such reduction in domestic prices will also reduce unit quota rents.31

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

63 Table 2.3. Impact of a 1% increase in quota in the EU Increase of quota by 1% and…

Holding constant Increase of export volume by quota by 1% only reducing price

Commodity Market Impacts

3URGXFHU3ULFH¼WRQQH 0DUJLQDOFRVW¼WRQQH Quota rent rate (%) Net Exports (m tonnes)

:RUOG3ULFH¼WRQQH

~ result of experiment in levels ~ 299.1 304.2 252.9 252.9

304.2 243.4

221.4/284.5

221.4/284.5

221.4/284.5

0.2

0.2

0.2

212.2/274.6

0.2

9%/30%

5%/30%

6%/27%

10%/30%

8.1

7.0

7.4

7.4

8.1/8.1

7/ 7

7.4/7.4

7.4/7.4

169.3

170.3

169.9

169.9

169.3/169.3

170.3/170.3

169.9/169.9

169.9/169.9

~ result in % change ~ -4.5% -2.9% 0.8% 0.8%

Producer Price Marginal cost

-0.3% 0.8% 0.7/0.8

0.7%/0.8%

0.7%/0.8%

-3.5/ -2.6

Net Exports World Price

16.3% -0.5%

0.0% 0.0%

6.4% -0.2%

6.4% -0.2%

-164.2

82.6

~ million EUR ~ -10.1

-164.3/-164.2

82.6/82.6

-10.1/ -10.1

-1,167.7/ -974.6

-162.6 106.1

98.4 1 620.7

0.0 1 032.2

0.0 1 032.9

105.7/106.5

1620.7/1620.7

1,032.1/1,032.3

1,032.5/1,033.3

-95.5

-1 710.3

-1 084.9

0.0

-142.0/ -50.0

-1,756.9/ -1,664,3

-1,131.3/ -1,038.9

fixed

Economic Impacts Taxpayers of which export subsidies Consumers Farm Households of which farm owned of which farm land of which dairy livestock of which quota rents Input Suppliers Total 1. 2.

312.3 252.9

Holding constant export subsidy expenditures Holding constant and farm income by export subsidy reducing price and expenditures by reducing price compensating with direct payments

-2.9% -3.0%

-1 069.1

160.7

160.8

160.8

71.4

135.4/188.7

135.5/188.7

135.5/188.7

20.9/130.9

1.2

1.2

1.2

10.9

1.1/1.2

1.14/1.24

1.1/1.2

7.2/15.5

15.8

15.8

15.8

7.0

13.1/18.5

13.1/18.5

13.1/18.5

1.9/12.6

-273.2

-1 888.1

-1 262.6

-89.4

-345.0/ -202.0

-1,959.8/ -1,816.84

-1,334.2/ -1,191.4

-146.6/ -36.0

54.6

54.6

54.6

24.3

47.1/62.4

47.1/62.44

47.1/62.4

8.6/42.2

-100.0

46.8

-9.0

-11.8

-137.9/ -60.6

8.8/86.3

-47.0/30.5

-119.2/ 88.9

Items constrained constant in italic. Minimum-maximum ranges from sensitivity analysis shown below estimate.

Source : OECD PEM model

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

64

European Union Four experiments were carried out in the EU, each based on a 1% increase in quota, to illustrate the impact on welfare of quotas, and of quotas as part of a set of related policies. Table 2.3 presents the results of these experiments. Shown in this table are a set of market impacts of the experiment, including domestic and world prices, exports and marginal cost of production. It also shows changes in welfare for the different actors in the model; taxpayers, consumers, farm households, and input suppliers. Farm household welfare is disaggregated into returns to three categories of farm-owned inputs and total quota rent. Reported welfare results are restricted to the dairy sector; although there are clearly welfare spill-over effects in other markets, they will not be considered as part of this study.32 The first column of Table 2.3 shows the impact of a 1% increase in quota holding all other policy variables constant. Without a change in domestic price policy, domestic consumption will not change significantly, with the result that nearly the full amount of the quota increase will be exported. The increase in exports causes world prices to decline by half a percent. This change will be transmitted back into the domestic market, reducing domestic prices by approximately one-third of a per cent.33 The increase in domestic production has caused a movement along the supply function and a consequent rise in marginal cost of 0.8%. The combined effect of the domestic price change and increase in marginal cost has reduced unit quota rent from the baseline 20% to 19%. The increase in quota quantity is insufficient to compensate for the decline in unit quota rent, resulting in a reduction of EUR 273 million in the total value of quota. Returns to farm-owned inputs have increased by EUR 177 million, insufficient to fully compensate for the reduced value of quota. Input suppliers’ surplus increases because the price of inputs has increased. Demand for inputs is a derived demand based upon the quantity produced of the final output. Increasing quota increases the amount of milk supplied, and therefore the demand for inputs in milk production. The resulting change in input suppliers’ surplus amounts to EUR 55 million, and reflects the sharing of total producer surplus between these suppliers and farm households. The change in welfare for input suppliers is less than that for farm-owned inputs because of the relatively elastic supply of purchased inputs (recalling that change in input suppliers’ welfare is determined by changes in producer surplus from the supply function of these inputs). Total surplus change to both farm-owned and purchased inputs34 is around EUR 233 million, significantly below the change in total quota rents. The difference is largely due to the change in domestic prices induced by the world price change as discussed above. Export subsidies have to increase by EUR 162 million in order to fund the 16% increase in exports. This result highlights the role of quotas in limiting expenditures on export subsidy programmes. The sensitivity analysis for this scenario points out the importance of the unit quota rent in the results. The marginal cost displays significant variance, as its level is indeed part of the sensitivity analysis and can vary from between one half to double its base value. This has a strong influence on the total quota rents that form an important part of farm household welfare. This is common to the sensitivity analysis for all scenarios. The previous scenario demonstrates that quota increases by themselves tend to result in the increased production being exported, with an associated increase in expenditures on export subsidies. It should be noted that as in the Aglink scenario, the PEM experiments are illustrative and actual WTO limits on export subsidies are not taken into consideration. The second column of Table 2.3 adds to the previous scenario a restriction DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

65 on increasing exports. In this case, the increase in production caused by the quota expansion must be taken up by the domestic market in order to preserve current export levels. In order to obtain the required increase in domestic demand, the rate of market price support must be reduced to allow the domestic price to fall. In this experiment, the fall in domestic price amounts to 4.5%. This brings about significant benefits to consumers, in excess of EUR 1.6 billion, as well as benefiting taxpayers by reducing the amount of expenditure on export subsidies. This expenditure declines because, even though the trade volume is unchanged, the difference between the domestic and world price, and therefore the export subsidy rate, has fallen. The reduction in total quota rents is greater than the quota-expansion-alone scenario as this experiment results in a significantly greater fall in domestic prices, and therefore erodes total quota rents to a greater degree. Net welfare for the EU is increased, reflecting the reduced transfer to importers in the rest of the world resulting from subsidised EU exports. The Aglink results above investigated the change in domestic price support levels coincident with an increase in quota, given a restriction on growth in government expenditures on subsidised exports. A similar experiment is done here. It illustrates that domestic price support will be maintained at a higher level when compared to the previous scenario if the quota relaxation is coupled with a constraint on the amount of export subsidy expenditures. This experiment shows that a 1% increase in allowable quota production results in a fall of 2.9% in domestic prices (third column in Table 2.3). Total export subsidies are the product of exports and the domestic-world price differential; decreasing the domestic price will limit the expansion in exports caused by the quota increase (by stimulating domestic demand) and reduce the price differential. The decreased price differential allows exports to increase by 6.4% while keeping total expenditures on export subsidies constant. The ability to expand exports within fixed export subsidy expenditures is why the reduction in domestic price in this scenario is less than the constant-export volume scenario. This smaller increase in exports compared with the quota-only scenario limits the reduction in the world price of milk to 0.2%, less than half the decline in that scenario. Consumers are the main beneficiaries in this experiment. Policies that increase prices are designed to transfer income from consumers to producers, and the change in price support in this experiment reduces this transfer by over EUR 1 billion. The resulting reduction in total quota rent is even larger, EUR 1 260 million, the difference reflecting the proportion of production that is exported. Total reduction in farm welfare when factor income is included is EUR 1 084 million. Changes in factor income, both farm-owned and for suppliers of purchased inputs, are unchanged from the quota-only scenario as the change in production quantity is the same, governed by the quota shock in both cases. Figure 2.3 shows the relationship between domestic price and quota level that must hold if export subsidies are to be held constant. The rate of price decline required is such that the quota level is no longer binding on production after an increase of only 5.5%. That is, increasing quota by this amount requires a price reduction of 16% to hold export subsidies constant. The increased production also increases marginal cost by 5%, and the two taken together are enough to completely eliminate unit quota rents.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

66 Figure 2.3. Relationship between price and quota holding export subsidies constant

'RPHVWLF3ULFH¼WRQQH

290 285 280 275 270 265 260 255 250 122

123

124

125

126

127

128

129

Quota Level (m tonnes) Source: OECD PEM model

The decrease in domestic price required by the constraint on export subsidies has a sharply negative impact on farm income. As was mentioned at the outset, such changes in farm income resulting from policy reform raise questions of compensation. The experiment reported in the fourth column of Table 2.3 adds to the previous experiment by offering direct payments to dairy producers in an amount sufficient to hold farm household income constant. In the PEM, such direct payments are represented as a subsidy to pasture used by dairy producers (but not other users of pasture). Providing support of this kind reduces the cost structure of the farm operation, and therefore shifts the supply curve down and to the right (the movement from S to S’ in Figure 2.4). The shift in supply lowers marginal cost for a given level of production (by 3% in this case), and so increases the unit quota rent. Area c represents the increase in quota value resulting from this change in unit rents, while area d represents the increase in quota value resulting from the increase in quota level itself. If there is a decline in world price as a result of the increase in quota, the effect on total quota rents is equal to areas f+g, of which f is captured by domestic consumers as increased surplus.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

67 Figure 2.4. Income-compensating payments shift supply

Pd

S

g

f

a

MC0

Sq’

Sq

D

d

c

e Pw D S S’ 0

Qwp Q d

Qp

Q wd

Source: OECD Secretariat.

The result of this experiment is that even though the compensatory payment is based on the use of a farm input, most of the benefit to farm households is to maintain total quota rents. Why is this the case? The supply of dairy pasture in the model is quite elastic, while the binding quota makes the demand for pasture relatively inelastic. For this reason, the main effect of the payment is to reduce land price, and reducing dairy costs. This is seen in the significant reduction in producer marginal cost in this scenario compared to the others. The reduction in marginal cost leads to the benefits being capitalised into dairy quotas.35 Input suppliers benefit less under this scenario because the reduction in cost of land resulting from the subsidy induces a change in the input mix away from purchased inputs towards land use (extensification). In this experiment, the effect of the price reduction and the cost reduction on unit quota rents nearly washes out, reducing its absolute level by only about EUR 1/tonne and with almost no change in unit quota rents in percentage terms. Farm household factor income increases modestly (by amount e-c in Figure 2.4). The main welfare impact in this scenario is the significant net transfer from taxpayers to consumers of over EUR 1 billion. Sensitivity analysis reveals an important difference between the first and last two scenarios: The plausible range of values for change in total welfare spans zero in scenarios three and four. This ambiguity regarding possible total welfare effects does not exist in scenarios one and two. In the first scenario, net welfare is consistently negative, and consistently positive in the second. The greater complexity of the last two scenarios DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

68 makes the net welfare outcome less clear. It is worth noting that all of the components of net welfare have consistent signs under the sensitivity analysis; variation in their relative magnitudes drives the net welfare result. Reflecting the underlying hypothesis, the model results show that increasing quota by itself cannot be seen as a move toward increased market liberalisation as for the same level of producer price support, an increase in quota would require an increase in export subsidies (if permitted by the WTO limits), which would subsequently distort world prices further. Moreover, in the PEM experiments shown here, increasing quota (by itself or with an associated compensatory policy change) leads to a decrease in domestic welfare. Think about it this way: Quota policies exist largely to moderate the negative impact of price support policies on government expenditure and trade in order to enhance the political sustainability of high price support. Increasing quota is a reduction in that moderation and not as such a step toward liberalisation unless accompanied by an appropriate reduction in price support. Under these conditions, dairy quota reform can be achieved in a welfare-enhancing manner, as is evidenced by the second experiment. This result was also observed by Bouamra-Mechemache et al. (2002). They showed that in the presence of market distortions, an elimination of quota results in welfare losses.

Canada In the Canadian quota system for milk production, domestic prices and production quotas are jointly determined to balance domestic supply and demand, allowing for a set level of imports through TRQs. The Canadian dairy system provides quotas for fluid and industrial milk separately, the first being a provincial responsibility and the second federal. The total quota is allocated to producers as a single quantity. The price paid to farmers is a weighted average of prices paid by end users (i.e. fluid, manufacturing use) less marketing charges. In the experiments regarding the Canadian milk system, the policy innovation will be allowable imports, which is also controlled as part of Canadian dairy policy. The amount of allowed imports is doubled in both scenarios, from a base amount of 375 000 tonnes to 750 000 tonnes milk equivalent (Table 2.4) The first column of Table 2.4 contains a scenario where production quota (via industrial milk) is increased by 1% This provides an illustration of the transfer efficiency of policies including quota. A quota change in Canada involves, by design, a domestic price change but no change in trade volume. By examining the relative changes in consumer surplus (price support is via MPS, and therefore consumer-funded) and farm welfare, the transfer efficiency of a change in quota level is seen to be 40.3/42.3, or about 95%. While less transfer efficient than a price change with constant quota, which was shown to be 1, this shows that varying the quota level itself is still a very efficient way of transferring welfare to producers.36 Figure 2.5 illustrates the impacts of a policy change increasing allowable imports (via increasing a filled TRQ, for example). The domestic supply schedule is shown, with the vertical range representing the quota restriction on output, and the demand curve for industrial milk products. At the domestic price (set by policy), Pd, there are planned net imports equal to the area Qd-Qp. Increasing the amount of allowed industrial milk product imports without changing domestic quota requires that the domestic price be reduced sufficiently for domestic consumption to absorb the full amount of the additional imports. This reduction in the administered price combined with the increased quantity consumed increases consumer surplus by areas b+c+d. Area b is total quota rent that is transferred to consumer surplus through the price reduction, and area a is remaining total quota rent.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

69 Figure 2.5. Impact on welfare of an increase in imports D Sq

Pd

c

b

d

a

MC

D Pw S Qwp

0 Production

Domestic Consumption Source: OECD Secretariat.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

Qp

Qd

Imports

Q wd

70 Table 2.4. Doubling import quantity in Canada Hold farm income constant by changing

Commodity Market Impacts Producer Price (CAD/tonne) Marginal cost (CAD/tonne) Unit quota rent (%) Domestic Production (m tonnes) Net Exports (m tonnes) World Price (CAD/tonne)

1% increase in quota only

2x import quantity only

514.5

499.6

514.5/514.5

499.6/499.6

521.4/529.1

403.6

400.6

386.3

377.8

343.3/463.9

340.8/460.6

326.6/446.3

317.5/437.9

Industrial quota

Direct payments

~ result of experiment in levels ~ 524.7 499.6 499.6/499.6

22%

20%

26%

24%

9%/33%

8%/31%

16%/37%

12%/36%

8.3

8.2

7.8

8.2

8.3/8.3

8.2/8.2

7/8/7.9

8.2/8.2

-0.4

-0.8

-0.8

-0.8

-0.4/-0.4

-0.8/-0.8

-0.8/-0.8

-0.8/-0.8

239.3

239.3

239.3

239.3

239.3/239.2

239.3/239.3

239.3/239.3

239.3/239.3

-1%

-4%

1%

-4%

1%/1%

-4%/-4%

0.2%/1.7%

-4%/-4%

~ result in % change ~ Producer Price Marginal cost Domestic Production World Price

1%

0%

-4%

-6%

1%/1%

0%/0%

-4.3%/-3%

-7%/-5%

1%

0%

-5%

0%

1%/1%

0%/0%

-5.3%/-4.0%

0%/0%

0%

0%

0%

0%

0%/0%

0%/0%

0%/0%

0%/0%

~ million CAD ~

Economic Impacts Taxpayers Consumers Farm Households of which farm owned of which farm land of which dairy livestock of which quota rents Input Suppliers Total

1. 2.

0.0

0.0

1.5

-164.4

0/0

0/0

1.5/1.5

-193.4/-139.6

42.3

188.0

1.2

188.0

42.3/42.3

188/188

-31.2/25.4

188/188

-40.3

-169.6

0.0

0.0

-45.7/-35

-169.6/-169.6

fixed

fixed

16.9

0.0

-73.2

-13.0

13.9/20.3

0/0

-77.3/-69.6

-19.1/-6.9

0.1

0.0

-0.6

1.0

0.1/0.1

0/0

-0.7/-0.5

0.5/1.6

6.3

0.0

-28.5

-5.1

5.1/7.5

0/0

-23.0/-27.0

-7.4/-2.7

-63.7

-169.6

102.3

17.0

-73.2/-54.6

-169.6/-169.6

98.2/106.9

8.9/25.2

2.4

0.0

-10.6

-1.8

2.3/3.1

0/0

-11.1/-10.2

-2.3/-1.0

4.7

18.4

-7.8

21.7

-0.3/9.5

18.4/18.4

-41.6/14.6

-7.2/46.4

Items constrained constant in italic. Minimum-maximum ranges from sensitivity analysis shown below estimate.

Source : OECD PEM model.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

71 In the first experiment, shown in the second column of Table 2.4, only the change in allowed imports is considered. No welfare constraints are placed on the scenario, and quotas in both fluid and industrial markets remain constant. There is no change in the fluid market, as products in this market are non-tradable, and the quota supply in that market remains constant. The increase in allowed imports of industrial milk products increases the total supply on the domestic market, and so the price of industrial milk must be reduced to clear the market (Table 2.4 shows the blend, or all-milk average price). The result is that consumers benefit from reduced prices and farm households lose through the impact of the reduction in price on total quota rent. There is a small increase in net welfare of CAD 18.4 million that is equal to the increased consumer surplus derived from consumption of imported milk products (areas c plus d in Figure 2.5). The sensitivity analysis shows variation in unit quota rent (a variable included in the randomised set), but no variation in the change in total quota rent resulting from the scenario. This is because neither quota level nor marginal costs are changing as part of the scenario itself—the change in total quota rent is due entirely to changes in output price. The supply-management system has been the main form of policy support to the Canadian dairy sector for some time now. It is reasonable to assume that the quota system would be the probable tool to compensate producers for the change in allowable imports, if desired. The second experiment, shown in column 3 of Table 2.4, combines the increased import quantity with an adjustment of industrial quota such that farm welfare remains constant. The welfare-compensating change in quota responding to an increase in imports is to reduce quota. Why? Reducing quota limits the change in domestic price and maintains the unit quota rent. This is worth the trade-off with rents lost from reduced production. While the industrial price has to be reduced in this scenario, it is reduced less than was the case when only imports changed.37 The only effect of an increase in imports, as seen in the first experiment, is to reduce domestic prices enough to clear the market. This price decline reduces unit quota rent. The welfare-compensating change in quota policy is to reduce production to support domestic prices and counteract this effect. Industrial quota must be reduced nearly the full amount of the increase in imports. In fact, the unit quota rent received by producers has increased from the baseline value of 23% to 26% due to the decline in quantity produced.38 An interesting result is that the blend price has increased, despite a small reduction in industrial milk price and no change in the fluid price. This is because the weighting of the average price now contains more of the higher-valued fluid milk after the decline in industrial production. Reducing quota more and more in response to successively increased import competition remains the means by which the quota system may be used to compensate farm welfare right up to the point where quota is no longer binding. Quota ceases to become binding at some point because the reduction in domestic quota only moderates, rather than reverses, the decrease in domestic price required by the increase in imports. If the rate of reduction in domestic price exceeds the rate of reduction in marginal cost of production, which also falls as quota is reduced, then there is some point at which unit quota rents are eliminated and the industrial quota ceases to bind.39 This is shown in Figure 2.6, which graphs industrial quota level that holds welfare constant as imports are increased. The change in industrial milk price is also shown in Figure 2.6. This price falls continuously as imports are increased, demonstrating that the reduction in quota is always somewhat less than the increase in imports.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

72 Figure 2.6. Domestic industrial milk quota and price response to increased imports Canada, farm income held constant through quota reduction 400

6

5 300 4 250 200

3

150

industrial price blend price industrial milk supply

100

2

supply quantity (m tonnes)

Manufacturing milk supply price (CAD/tonne)

350

1

50 0

0 0.1

0.6

1.1

1.6

2.1 2.6 3.1 3.6 increase in imports (m tonnes)

4.1

4.6

5.1

5.6

Source: OECD PEM model.

An alternative approach to providing compensation to farm households for the change in allowed imports that is more reform-oriented would be to provide compensation in the form of direct payments. Column four of Table 2.4 presents the result of this experiment. The results are notionally similar to the EU experiment where compensation was provided by direct payments; the payments act to reduce the cost structure and thereby preserve total quota rents. Little net change in quota returns or returns from farm inputs is seen. Input suppliers’ surplus reduces marginally as farmers switch to using relatively more of the land input to which the support is applied. Taxpayers are also worse off, by the amount of the new direct payment, but this loss is less than what consumers gain from reduced prices and the outcome is a net social welfare gain. There is little doubt that the value of total quota rent plays an important role in determining total farm welfare where quotas exist. The experiments conducted for the EU and Canada demonstrate that welfare-compensating policy actions tend to have their impact on total quota rent rather than the returns to other farm assets, regardless of the way compensation is provided. The experiments also demonstrate that, other things equal, increasing quota is welfare reducing for producers. This implies that any reform of dairy sectors where quota policies are in place must recognise that quotas are part of a market price support system. In evaluating the welfare effects of quota systems it should be kept in mind that typically quota systems are analysed within the assumption of perfectly competitive markets. However, situations may exist where farmers are facing an oligopsonistic processing and retailing market structure. Where this is the case, an imperfect market at the processing and retail level would provide a more appropriate basis for comparison.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

73

Modelling of milk supply in the presence of quota A quota system and border measures supporting domestic prices could be seen to some extent mutually dependent policies. Simply removing production controls without also eliminating (or significantly reducing) market price support and related border measures would likely result in unsustainable overproduction. Conversely, keeping a quota system in place and removing market price support and border measures would make quotas irrelevant which, however, does not cause a problem of policy instability. Thus, the quota system has to be considered when evaluating a global dairy policy reform scenario. The impact of such a scenario, among other factors, depends heavily on the production response following the removal of milk production restrictions in milk quota countries. This is ever more important as dairy markets are relatively thin and some of the countries operating the quota system belong to the major players on international dairy markets. It follows, that correctly assessing the supply response potential after quota removal is crucial for the analysis. In constructing the supply functions in quota operating countries the assumptions made concerning unit quota rent and supply elasticities are the most important aspects to be decided. The assessment of milk production potential in quota operating countries is complicated by the absence of historical evidence as quotas in respective countries are in place already for a considerable amount of time. Thus, the supply schedule is not directly observable and it is impossible to point out the “true” quota rent, supply elasticity and underlying production potential in quota operating countries.40 There has been an intensive research and debate concerning the appropriate value of quota rent and a wide range of available values can be found in the literature.41 All these estimates are conditional on the method and assumptions used in the quota rent calculation (OECD, 2003b). Generally speaking, it seems that for the analysis of the medium to long run production potential in quota operating countries, the full marginal cost method should be considered in constructing the milk supply function. Thus, it is important to include all costs into the cost structure of the dairy sector, i.e. also those that are considered fixed in the short run such as labour and capital. Further it could be argued that using quota lease prices, which are annual prices of quota, is not appropriate as it overestimates the long run value of quota. This comes from the fact that the quota lease price reflects short-run considerations such as the threat of over quota production and the need to acquire annual quota to avoid the penalty. The price of quota bought or sold is the more appropriate one to use. However, even the quota rent observed from bought or sold quotas is often strongly influenced by short term market conditions reflecting the need of expanding producers. Hence, prices paid for quota on a segregated quota market for new entrants would seem to be more adequate to use in determining the quota rent. In assessing the quota rent from bough or sold quotas there is a need to consider the appropriate number of years of depreciation, the discount rate, the impact of reform on costs of production and on the price of land and the anticipation of producers about compensation.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

74 Milk supply functions in Aglink and PEM For countries which operate a quota system, the Aglink model has not specified the underlying milk supply function and milk production has been fixed at the quota level. In order to assess the impacts of trade liberalisation in the dairy sector supply functions for quota countries needed to be built into the model. The vertical position of the supply function can be determined from the quota rent or alternatively from the shadow price of quota (marginal cost of production at the quota level). However, this information provides merely a point on the supply curve of the respective country/region at a given moment in time. Thus, in order to identify the supply function it is also necessary to assume the elasticity. Given the multi-commodity and dynamic aspects of Aglink, short term and cross-price elasticities with respect to own price and to those for feed and beef also had to be assumed. Moreover, due to the interrelationship of dairy and beef markets, separate functions for dairy cows and milk yields had to be constructed such that their combined elasticity measured up to the targeted supply elasticity. The next paragraphs discuss the basis for these assumptions. In the context of the analysis of dairy policy reform, the Secretariat decided to employ in addition to the Aglink model also its PEM model, in order to examine the impact of dairy policy reform on production, consumption, income, welfare, trade, and world prices. In order to use the complementarities of the Aglink and PEM models, the PEM assumptions on quota rents and supply elasticities were used also in Aglink model.42 For Canada, the quota rent in PEM (and by construction in Aglink) is assumed to correspond to 23% of the price and the long run milk supply elasticity is 0.81, while in the case of the European Union the quota rent is set at 20% of the price and the long run milk supply elasticity is 1.23. The difference in supply elasticity between the two regions is due to the relatively higher intensity of land use in production in Canada, where the EU uses relatively more purchased inputs. Purchased inputs have a higher elasticity of supply, which carries over into the supply elasticity for milk. Other factors influencing the supply elasticity, such as factor substitutability and factor supply elasticities, are nearly identical between these regions; it is differences in factor shares that brings about the different supply elasticities. The parameters underlying the supply function were developed through consultants’ reports done as part of the PEM model development process. These reports, by Abler (1998) and Salhofer (1998), applied a meta-analysis to extensive literature reviews of research in the PEM regions to estimate elasticities of factor supply and substitution, as well as provide notional confidence intervals for these parameters.43 Taken together, these parameters and the model structure provide an implicit supply function, the long run elasticity of which is shown here in the next-to-last row of Table 2.5 which illustrates the short run and long run elasticity assumptions used in the construction of supply functions in Aglink. The estimates of quota rent were also made during the PEM pilot phase. In the case of Canada, an estimate was available in the literature (Moschini, 1986) For the EU and Swiss quota rents, agreement as to their value was reached between participants of the PEM working group, which included technical experts from all participating countries, including the EU, plus some non-participating countries. Subsequent to the pilot phase of the PEM, these were discussed in depth and judged reasonable by the Expert Group meeting on milk quotas organised by the Secretariat in September 2003. The participants of the Expert Group meeting noted the difficulties of estimating milk production response in quota operating countries particularly given the absence of historical evidence as milk quotas are in place already for a considerable amount of time. DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

75 The most common approaches to estimate quota rent have been identified; the acquisition of quota price from the market where quotas are tradable or alternatively the collection of micro-economic (farm-level) survey data to obtain an estimate of total costs and then derive the marginal cost to obtain the quota rent point (shadow price of quota). It was also noted that for the analysis of the medium to long run production potential in quota operating countries, the full marginal cost is the one to be considered in constructing the milk supply function. A wide range of assumptions on quota rents exists in the literature and it is difficult to pint point a single number.44 Nevertheless, the Expert Group meeting reached an agreement that the quota rent levels assumed in the PEM model were generally appropriate, as were the elasticities of supply used in the model. Table 2.5. Elasticity used in Aglink supply functions for Canada and the European Union Canada

European Union

0.082

0.078

Short term Milk cow/milk price Milk cow/feed price

í

í

Milk cow/beef price

í

0.06

Partial adjustment coefficient

0.885

0.93

Yield/milk price

0.1

0.12

Yield/feed price

í

í

Production/milk price

0.2

0.198

Production/feed price

í

í

Production/milk price

0.81

1.23

Production/feed price

í

í

Long term

Source: OECD Aglink model.

Different milk production potential assumptions across EU member states It is important to note that the European Union is treated as a single block in the Secretariat’s PEM and Aglink models which implies a certain limitation for the quota analysis and trade liberalisation scenario. The average quota rent for the EU, assumed in PEM and Aglink, may be substantially different from quota rents in individual EU member states. In fact, the quota rents differ substantially amongst EU countries, depending on the institutional set-up for quota allocation and tradability as well as on the efficiency of milk production. The variability is well shown in Table 2.6 which illustrates quota rent from three studies that are selected to show relatively high, average and low quota rent assumptions. Each study has used considerably different assumptions on quota rents and the table shows that within the broad categories of high, average and low quota rent assumptions there are large differences across the EU member states. The results of these studies are, as can be expected, to a large extent driven by the assumption on initial quota rent. Thus, for example, Lips and Rieder have concluded that the changes in raw milk production following a quota system abolition in the European Union differ significantly by country. The authors suggest that Ireland would strongly increase its milk production, countries such Denmark, Italy, Luxembourg, the Netherlands and Spain would show an expansion DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

76 of their raw milk production while countries such Germany, Greece, Portugal and Sweden would see a decline in quantity produced Comparing the magnitude of quota rents assumed in the above table it is apparent that on average the quota rent used in PEM and Aglink for the EU is smaller than those used in INRA-Wageningen (2002),45 is of a same (similar on average) magnitude as the ones used in Lips and Rieder (2003) and is larger than those used in Jensen and Frandsen (2003).46 Table 2.6. Milk quota rent assumptions for the European Union (per cent of price)

Austria Belgium Denmark Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Sweden United Kingdom

INRA-Wageningen Base 1998 45.9 31.8 41.6 24.4 35.2 45.3 36.8 49.1 36.8 29.2 36.0 26.9 37.5 15.2 42.6

Lips and Rieder Base 1997 17 20 26 15 22 20 0 31 23 18 23 0 24 10 27

Jensen and Frandsen Base 1997 9 11.9 5.3 1.6 10.5 12 2.6 7.2 5.3 11.9 16.6 5.9 5.8 2.2 8.9

Source: INRA-Wageningen (2002), Lips and Rieder (2003), Jensen and Frandsen (2003).

Nevertheless, as there are many assumptions underlying the quota rent calculations, the adopted value remains uncertain. Thus, given the fact that the specification of milk supply functions is of crucial importance for modelling dairy policy reform, it seems necessary to undertake a sensitivity analysis with respect to the possible range of the underlying supply function assumptions. A sensitivity analysis is thus conducted on the quota rent and supply elasticity assumptions adapted in Aglink in the trade liberalisation scenario. The results of the scenario and the sensitivity analysis results are presented in Chapter 3.

Conclusions The analysis in this chapter has illustrated some specific market, trade and welfare implications of operating milk quota systems. It should be kept in mind, however, that the analytical and empirical results presented in the paper have to be viewed within the hypotheses and assumptions adopted and within the limits and caveats of the models used. Quotas were typically introduced as a tool to control the growth of surplus production and budgetary expenditures in order to improve the political sustainability of high price support. Simultaneously, the objective of stabilising producer prices could be easier achieved within a system of production quotas. A quota imposition is a welfare reducing DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

77 instrument when compared to support price cuts. Nevertheless, when decision makers are unwilling to cut support prices, the implementation of quotas could be considered as a welfare improving policy instrument. The efficiency of production under quota depends to some extent on a particular quota administration system. For example, one way to rectify the inefficiencies related to quota imposition is to organise a quota trade so that producers with low marginal costs can buy or lease quotas from producers with higher marginal costs. On the other hand, quota trade can bring its own problems, such as increase in structural cost, industry concentration issues etc, so that the pros and cons of different administration systems need to be weighed. Nevertheless, from a dairy policy reform stand point, an important consideration is connected to the inherent nature of a quota system – the presence of a quota value. As the producer support is typically tied to the amount of quota, it is less profitable and often not feasible to supply milk without quota and therefore quota is an income-generating asset and assumes a value. With time the value of quota, reflecting the difference between an underlying cost of production and milk price, becomes incorporated into the cost structure of dairy farms. When evaluating a quota system, it is important to keep in mind that a quota is typically contingent on the existence of another policy, namely market price support, and, in many milk producing countries, that market price support is in turn often contingent on the presence of quota. Simply removing production controls without also eliminating market price support would likely be unsustainable; conversely, in the absence of a policy that raises domestic prices over world prices, there is little rationale for limiting the quantity that domestic producers may offer in the marketplace. Thus, quota interacts with the effects of other policy tools and impacts on markets and welfare within a context of specific policy objectives. If an objective is to hold the volume of subsidised exports or government expenditures on subsidised exports constant for a given increase in the quota level, then this increase has to be accommodated by a reduction in the domestic support price. Empirical results of experiments conducted with the Aglink model have shown the milk and dairy product price cuts required for a particular increase in quota for a given policy objective. The results, which are primarily driven by the size of demand elasticities, showed that the EU butter price would have to be reduced substantially more than the SMP price, while prices of cheese and WMP would be reduced somewhere in between the butter and SMP price cuts. The PEM empirical analysis of the relationship between quota level and price support in determining farm income in the European Union has shown that an expansion in domestic quota could be welfare-reducing for producers via erosion in the value of quota rents as marginal production costs rise and input suppliers recapture part of the producer surplus as demand for inputs increases. This erosion could be offset by an increase in the level of price support (or other means of delivering support) if the objective is to keep milk producers’ welfare unchanged. However, the experiments demonstrate that welfarecompensating policy actions tend to have their impact on quota rent rather than on the returns to other farm assets, regardless of the way compensation is provided. When the focus is on total welfare, the experiments show that dairy quota policy reform will be welfare enhancing if combined with an appropriate reduction in price support. The PEM analysis for Canada focused on a doubling in the import quantity, and also shows a loss in producer welfare due to a decline in domestic prices, which entirely DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

78 translates into a reduction in quota rents. But consumers gain from the drop in prices, with a small improvement in welfare overall. The experiments show that producer welfare can be maintained unchanged if the increase in imports is coupled with a reduction in the quota, or alternatively, a rise in compensatory direct payments. By increasingly reducing quota in response to successive amplifications in imports, farm welfare can be maintained up to the point where quota is no longer binding, and with marginal impacts on overall welfare. The same can be achieved if lower prices are compensated for by higher direct payments. This also results in a small overall welfare gain, but which comes at the expense of substantially increased costs to taxpayers. In summary, it could be said that quota represents a second best alternative that allows policy makers to continue a policy of high price support without necessarily aggravating budgetary problems. Quotas, as they are normally strictly enforced, also reduce the MPS impact on trade and world markets from excess production, although consumption is still limited by high prices. In most countries, quotas are thus an integral part of price support mechanisms, and exist to make price support sustainable from a budgetary point of view. Quotas exist in the context of market price support and their full impact depends on the other policies that are operated concurrently. Nevertheless, a quota system is unlikely to be considered as the best policy option. This is due to the inefficiencies that it may create, the cost that it imposes on consumers, the difficulties and costs of administration that may arise for governments, the difficulty in getting the information on the quota level that would match production (or trade) under free trade and the vested interests that it generates. The existence of quota systems also likely depends on the continuation of high border measures, which is uncertain in the context of multilateral trade reform. That is, quota systems allow a domestic market to be managed only if that market is isolated from external sources of supply. Quota imposition provides gains for initial beneficiaries, but subsequent generations can be locked into a higher cost structure, and the system then perpetuates itself. In other words, the value of quota gets built progressively into the farm cost structure and complicates the reform or removal process of the quota system later on. Thus, to quote Guyomard and Mahé (1994): “Quotas might be an attractive approach to reduce distortion in production, however, they do not benefit consumers and from a political economy stand point quotas are more likely to delay reforms on price support.”

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

79

Notes 1.

Although consumers are typically not affected directly by a quota system, the presence of quota may facilitate a continuation of high price support measures which indeed do influence consumers.

2.

In the closed economy-cobweb framework quota restricts production expansion when price is high, thus limiting a price fall in the following period.

3.

Switzerland has recently passed a new law which provides for the abolition of its milk quota from 2009 onwards.

4.

A description of individual supply management systems can be found in other sources: for the EU in EU Commission (2002), for Japan in Suzuki and Kaiser (1994), for Canada in Barichello (1999) (alternatively in Lipert (2001), for Switzerland in Swiss Federal Office for Agriculture (2001), for Norway see Annex G. A table of dairy quota programmes and their key features in OECD countries can be found in OECD (1990).

5.

For detailed discussion on key features of the OECD quota programmes see OECD, 1990.

6.

Fluid milk is defined here as milk sold in liquid form for drinking. Other terms used for fluid milk might include: liquid milk, drinking milk, table milk, and town milk.

7.

Manufacturing milk is milk used in production of dairy products (i.e. butter, cheese, SMP, WMP, casein etc).

8.

For example, the European Union implemented a series of milk quota cuts resulting in an 8% overall reduction in quotas over a nine-year period – 1984/85 to 1993/94 marketing years (Court of Auditors, 2001).

9.

See Alston (1980), Barichello (1981), Hubbard (1984), Harvey (1984), Barichello (1984), Burrell (1987), Moschini (1989), Burrell (1989), Dawson (1991), Oskam and Speijers (1992), Guyomard and Mahé (1994), Guyomard et al. (1996), Chen and Meilke (1996), Colman et al. (1998), Alston and Spriggs (1998), Turvey et al. (2002).

10.

Josling (1984) describes the output control and price reduction options in more general terms.

11.

When a quota is set at a level that is above quantity demanded domestically at set support prices, then the quota by itself has no direct consequence for consumers assuming that support prices are held constant.

12.

It is difficult to obtain the actual share. The shares estimated with the PEM model are presented later in this chapter.

13.

The example could be reversed to show that an increase in the quota level allows part of the producer surplus to be recaptured by input suppliers due to rising demand for purchased inputs so that milk producers may loose due to the quota rent erosion.

14.

In that case, when domestic demand is less than Q*, exports will need tax payer support given that world price would normally be less than PS0.

15.

For further discussion and some empirical evidence on capitalising government programme benefits to quota see Oskam and Speijers and Barichello (1996).

16.

If a quota is imposed at the current production level with a subsequent increase in support prices, the rent to quota will be increasing relative to rent accruing to traditional resources.

17.

In addition, the simplified analytical framework abstains from a complex relationship between milk producers and processors in quota operating countries. Milk and dairy processors, given their strong bargaining position, might also benefit from the presence of a milk quota system.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

80 18.

Note, the high transfer efficiency applies at the margin under binding quota.

19.

However, an experience of allowing quota trade among regions in Canada has shown rapid flow of milk quota and production to more efficient regions.

20.

Indeed the cessation of production might be better addressed by different measures as compared to sale of quota. As an example, might be taken the structural adjustment package introduced through Dairy Industry Adjustment Act 2000 to help producers cope with the adjustment pressures following the Australia dairy sector reform (ACCC (2001), ABARE (2001)).

21.

One option to encourage new producers is to place a siphon on the quota transfers by the market authority which then can be freely or at reduced cost distributed to new entrants (see Burrell (1989) and Swinbank and Peters (1990) for further discussion). In cases where quotas are not traded, when producer ceases production the corresponding quotas are reallocated back to the industry. See also Box 2.1 discussing the administrative quota management in France.

22.

The level of quota is proportional to taxpayers cost. For example, assuming half of the milk production is consumed domestically, a one percent increase in quota level translates to two percent increase in taxpayers’ cost (holding the support price constant).

23.

It should also be noted that for simplicity the figure represents a small country case which does not have a substantial impact on world markets and prices. For a large country the world price would have to be reduced in the diagram to reflect the impact of increased exports.

24.

In Canada the quota system is to some extent endogenous – the quota level is set contingent on domestic price support to hold subsidised exports at low levels.

25.

The EU is treated as a single block so that spatial effects due to industrial management are not taken into the account.

26.

World dairy prices in the Aglink baseline are defined as F.O.B. Northern Europe (OECD 2003a).

27.

In the EU component of Aglink intervention stocks are exogenous. It is therefore implicitly assumed that stocks are not a medium term solution to structural surpluses.

28.

Note, that the results measure the impact against a baseline of 2003 where the EU policy assumptions reflect the Agenda 2000. Under the agreed CAP reform, the additional 10% cuts in butter support prices (as compared to the Agenda 2000, see OECD (2004)) would reduce the price gap between the EU and New Zealand and also the positive world price impact estimated in the above scenario arising from the special New Zealand market access for EU butter.

29.

As is the case in the Aglink model, the EU is treated as a single region in the model.

30.

For ways to derive the shadow price in the presence of quota see Moschini (1986) or INRA-University of Wageningen (2002).

31.

If border protection is varied to maintain a constant domestic price in the face of a changing world price, this effect on quota rents would not occur.

32.

These spill-overs will likely be minor in scale compared with the own-sector impact of policy changes. A possible exception is the experiment where direct payments are made to producers. Such payments can bring about significant changes in land distribution.

33.

All scenarios include the possibility of price transmission from world to EU domestic prices. The lower the degree of price transmission, the smaller the reduction in domestic price, and so the lower the consumer benefit and higher the producer benefit (or lower producer loss) of the scenario.

34..

This figure is the sum of the returns to the three factors of production owned by the farm household (farm owned, land, dairy herd) plus returns to input suppliers as shown in Table 2.3.

35.

A similar experiment could have been carried out where compensation was in the form of a headage payment; this would have been reflected in the dairy livestock factor. In this case, the factor supply is

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

81

inelastic (elasticity of 0.5), but still only about one-third of the value of the payment accrues as rent to the livestock factor, with the balance supporting quota value. 36.

The argument regarding long-term durability of this transfer applies in this case as well.

37.

Since a quota reduction is welfare-enhancing for the producer, this would imply that the quota quantity lies between the optimal quantity for a monopolist and the quantity that would hold under perfect competition.

38.

Unit quota rent, when expressed in percentage terms as is done here, has the price in both the numerator and denominator UQR%=(P-MC)/P

39.

Fluid quota continues to bind throughout (fluid price is not changing).

40.

In the discussion concerning assumptions of milk supply functions the term quota rent refers to the unit quota rent (i.e. support price – shadow price) rather than to the total quota rent (i.e. (support price – shadow price) × quota quantity).

41.

See, for example, Moschini (1986), Guyomard et al. (1996), Booth et al. (1996), Colman et al, (2002), INRA & Wageningen (2002), Turvey et al (2003) and Jensen and Frandsen (2003). The estimation of supply response in regulated markets was very well summarised in a document by Larivière and Meilke presented at the meeting of the OECD Commodity Working Group on Meat and Dairy Products in 1998. A very good overview of various studies estimating milk quota elimination in Europe has been compiled by Salamon (2002).

42.

It is worth mentioning that the PEM demand elasticities come from the Aglink model.

43.

These confidence intervals can be used to form the basis of sensitivity analysis of PEM results.

44.

For example, in the case of Canada, Hickling (1990) calculated a considerably higher quota rent of 43% however Meilke et.al. (1996) have argued that a unit quota rent for Canada equal to 20% is a more reasonable estimate (based on 1993 Ontario industrial milk quota values). The different assumptions on quota rent adopted in the literature for the European Union are discussed in this study.

45.

One of the reasons might be the fact that the calculation of quota rents in these studies has not accounted for fixed cost of land, building etc. (see Annex of EU Commission (2002) for the comments).

46.

Van Tongeren (2002) has used an assumption of 20% quota rent for the European Union using a GTAP model.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

82

Annex 2.1 Welfare Implications of Support Price Reduction Versus Quota Level Imposition

In the case that a policy of support prices exists that require import tariffs and significant quantities of export subsidies, it can be shown analytically that lowering support prices as compared to imposing quota limits brings about higher net welfare gains. The domestic welfare implications of these two policy options available to reduce large gaps between domestic production and domestic consumption at support prices are illustrated in Figure 2.1.1. The figure depicts the impact of output reduction by means of a support price cut and by means of a quota imposition. In this simple analytical setting, the support price reduction is illustrated by a change in the support price from PS to a new level PN, which is assumed to remain above the world price PW. The price change would cause consumers to increase their consumption from quantity QDS to QDN and dairy farmers to decrease their milk production from quantity QSS to QSN by moving down the supply curve SS. The welfare implications of the support price reduction is an increase in consumer surplus by area a + b and a reduction in producer surplus by area a + b + c + d. As a result of both the price cut and the consequent decrease in surplus production, the amount of subsidised exports can be reduced from (QSS - QDS) to (QSN - QDN). Thus, the budgetary costs of disposing of the surplus through export refunds fall by area b+c+d+e+f+g. The net welfare gain of support price reduction is thus equal to area b + g + e + f.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

83 Figure 2.1.1. The welfare implications of different policy options available to reduce large surpluses

D

S

S*

PS a

b

PN

d

c

f e

g PW S D

QDS

QDN

QSN = Q*

QSS

Source: OECD Secretariat.

Under the second policy option, the desired level of output QSN is achieved by the imposition of a quota limit Q* which shifts the supply curve to SS* while support price remains at its original level of PS. At output level Q*, the supply curve is perfectly inelastic (vertical) as it is assumed that a severe penalty for exceeding the quota is applied, thus discouraging farmers from over-quota production. Consumers continue to face price PS at which they consume QDS and are unaffected by the implementation of the new policy. It follows that the change in consumer surplus is zero. The diagram implies that with the imposition of quota dairy farmers lose producer surplus equal to area d. This loss is an unavoidable consequence of the output restriction. Due to the quota limit, the surplus production decreases from (QSS - QDS) to (Q* - QDS) and the budgetary costs of disposing of the surplus are reduced by area d+e+f. It follows that the net welfare gain is equal to the area e+f. In comparing the two policy option the net welfare gain of the support price cut is larger by area b+g as compared to the quota limit imposition. Moreover, this result of greater welfare gains from reducing the support price relative to comparable quota reductions holds for any plausible elasticities of consumer demand and producer supply.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

84

Annex 2.2 Incentive to Exchange Quotas and Emergence of Quota Value

Following on the standard theory of the effect of milk production quota on asset values (Harvey,1984, Burrell, 1989, Dawson, 1991, and Colman et al, 1998), Figure 2.2.1 illustrates schematically the immediate impact of quota imposition on the dairy industry. The quantity of milk produced before the quota regime is represented by Q. The farmer’s revenue from sale of milk PS Q is shared between its variable resources (area below the supply curve SS) and its fixed resources (area above SS). The area above the cumulative cost incurred (area above SS) is conventionally defined as producer surplus. Figure 2.2.1. Development of a quota market and a value of quota

PS

S S

b

*

a

QR PC

c e d

S

QT

Q*

Q

Source: OECD Secretariat.

Consider the case that a quota system is introduced with quota limit imposed at Q* and no leasing or trade of this quota is allowed with respect to the initial allocation. Typically, quotas are distributed on the basis of historical production levels rather than efficiency criteria and all producers experience the same percentage cut in their production. As a consequence some low-cost, efficient production is lost from the industry whilst some high-cost, inefficient production is maintained. The implication of this process is shift of the supply curve to the left. The loss in producer surplus due to the leftward shift of supply curve is represented by area d+e+c. This loss is a consequence of the initial quota allocation inefficiency. The producer surplus is also decreased by area a, but this is the unavoidable consequence of introducing a supply control measure. DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

85 When the quota system allows quota to be marketable the efficient producers would lease or buy quota from less efficient producers and the rental price in a competitive market would be bid to a rate equal to the difference between support price and marginal cost, PS - PC, as shown in Figure 2.2.1. At this price, the quota market equilibrium would be struck and the quantity (Q* - QT) of quota would be traded. The producers who lease out or sell quotas would obtain a revenue equivalent to area b + c where b is the compensation for the loss of income to fixed resources and c is a profit. The producers who lease in or buy will gain b + c + d + e for the price of b+ c. The total rental value of quota in a period could be measured by the area (PS - PC) Q*. It follows from the illustration that in a competitive market the quota would move from less efficient producers to the most efficient producers and the supply curve would effectively regain its original position up to the quota limit thus eliminating the initial inefficiencies. To eliminate the initial inefficiencies, Colman (2000) estimated that for the United Kingdom, between 15 and 26% of total quota would have to be transferred from less efficient producers to more efficient producers for the industry to reach profit maximising allocation. For France, Guyomard et al. (1996) estimated that in the short-run between 11% and 16% of quota would had to be transferred while for Netherlands, Boots et al. (1997) estimated that an extra 10.2% of quota would had to be transferred in order to achieve full economic efficiency.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

86

Annex 2.3 Long-Run Effects of a Quota Imposition on Farm Assets at the Farm Level The impact of quota on farm assets might be illustrated using an example of a tenant farm. Following Harvey and Hubbard (1984), consider a tenanted farmer for whom the area below the support price, PS, level and above the long run marginal cost, LMC, is economic rent. Presumably, the landowner holds the most scarce asset and can demand a rent from the tenant which is equal to this economic rent. When a competitive sector is in the long run equilibrium, marginal cost (excluding rent) equals average cost (including rent) at the producer price (in the diagram long run equilibrium is represented by price support PS and production Q). The LAC1 in the diagram denotes the long run average cost including rents when all inputs are constant. Assume that the support price is maintained at PS while production is restricted to Q*. Figure 2.3.1. Long-run effects of a quota imposition at the farm level

LM C LAC 3 PS

LAC 1 C

B

LAC 2 PC

A

Q*

Q

Source: OECD Secretariat.

When output is reduced1 from Q to Q* the intensity of use of the fixed asset (land) is reduced and the rent accrued to the asset falls by area PS PC A B. This generates a pure profit equal to area PS PC A C as the LAC1 curve shift down along the LMC curve to LAC2. However, this profit would be bid up into the quota rent with the unit quota price equal to PS í3C . If the quota is attached to land the land rent will be increased by the area corresponding to quota rent (PS PC A C). Oskam and Speijers has illustrated that the value of quota can be determined as the difference between the price of land with and without quotas. Thus, in the long-run equilibrium the quota rent (one way or the other) will be incorporated in the cost structure and the LAC2 move vertically up to LAC3 eliminating pure profit.

1

Note that the economic rent created by a quota system are capitalised into the market values or implicit values of quotas in the alternate case noted before, as well. Thus, if a quota is imposed at the current production level and support prices are then increased, the economic rent that is generated will be absorbed by a rising value in the quota.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

87

Annex 2.4 The Development of the Norwegian Milk Quota System

Background Until recently, all dairy farmers in Norway have been organised as a cooperative. Consequently, all milk produced in Norway has been delivered to and sold through a nationally federated dairy cooperative. Since the 1930s, this cooperative has played a major role in the process of developing the Norwegian national agricultural policy. The regulatory model in Norwegian agriculture is built on the idea that farmers have a responsibility to produce milk within a limited amount of aggregate supply. The authorities set this amount. In this system, the farmers are responsible of the costs of any potential over-production of milk. The maximum price of milk is regulated and settled in the yearly Agricultural Agreement between the Government and The Norwegian Farmers Union and the Norwegian Small-holders Union. Through the last two decades, various regulatory schemes have been set up to align aggregate milk production with domestic milk demand in Norway. The design of the schemes has been influenced by several factors such as the market situation, international pressures, the political ambitions of domestic milk production, farm-based regional settlements, efficient production of consumer goods and the goal of having a multifunctional agricultural sector in order to maintain environmental and cultural values. Quotas, as a policy instrument to control the supply, were suggested as early as in 1930 and then again in 1956. In the early 1960s, the number of dairy producers that withdrew from the sector was high, and this led to less focus on the issue of quotas. An increased demand for milk in the years 1975-80 made a quota system seem unnecessary. In the late 1970s, a governmental committee was established to evaluate the dairy policy and it concluded that other policy instruments than quotas might be sufficient to regulate the market. The first quota-based regulatory scheme was set up in 1983. This was a reaction to the fact that surplus production in the dairy sector had become an increasing problem. Between 1978 and 1983, aggregate milk production grew by 10%. Although several small farms closed out, productivity growth and investments in farm development more than made up for the loss of production that this caused. Norway led an ambitious agricultural policy in the late 1970s, which led to a rapid growth in producer prices and subsidies. This stimulated investments and production in the agricultural sector. Further more, there was a dramatic fall in dairy demand between 1980 and 1982. In 1982, the production surplus of milk had grown to almost 300 million litres. The need for stronger production control seemed obvious and in 1983 a quota system was introduced. The system has been redesigned and changed several times since then and in the following, a description of the different systems is given.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

88

The two-price quota scheme (initial period, 1983-1990) The first quota-based scheme was a two-price quota system. The basic mechanism of this regulatory scheme was that if producers delivered milk without holding any quotas or in excess of the quota held, they were paid a lower price for their milk. Over the years, this price became a diminishing share of the ordinary milk price. Initial quotas were allocated to the milk producers according to the production level held between 1980 and 1982. At the time of implementing this system, the intention was to base the farmer quotas on a moving average of deliveries over the last three years, but this flexibility feature of the system was removed in 1984. However, farmers did have the opportunity to apply for exemptions from the main rules for allocating the quotas. The so-called “twoprice steering committee”, in which the parties of the Agricultural Agreement (the Government and the two farmers associations) were represented, was given the right to administer the rules and handle the exemptions. In 1984, structural and regionally diversified factors to modify quotas were introduced. To protect small-holders from the potential negative effects a quota could have on the income of marginal farms, the Small-holders Union wanted minimum quotas to be implemented. A minimum quota of 15 000 litres was introduced in 1984. This was later increased to 25 000 litres in 1986 and to 30 000 litres in 1987. For the Farmers Union it was more important to introduce a floor in accordance with prior production. In 1986 the argument that it was necessary to also protect the larger producers from large quota reductions won through. This led to the decision that no producer was to be given a quota of less than 84% of the initial quota in 1983. The many exemption rules made many producers eligible receivers of additional quotas. The most important exemption rules were those concerning investments in farm development and entry for new generations or new farms. Between 1983 and 1989, almost 50% of the farmers increased their quotas through various types of exemptions. The high number of exemption rules opened up for possibilities to exercise discretion when administering the system and evaluating the “productive resources” of a farm.

The quota buy-out scheme (1991-94) After almost a decade of practicing the quota system, the scope for redistributing quotas administratively on the basis of exemption rules dried up in 1990. Market demand for dairy products continued to decline and gave rise to a need for reducing the total domestic production as well. Dairy farmers faced lower production quotas each year while yield growth and technological development simultaneously contributed to increase the production capacity. The next regulatory instrument to be set up in the Norwegian dairy sector was “the quota buy-out scheme” (1991). The purpose of the buy-out programme was to give farmers incentive to reduce or quit production of dairy products. This was done by offering a financial grant (“the adjustment grant”) to the dairy farmers who voluntarily joined the programme and thus were obliged to withdraw from milk production for the next seven years. The number of farmers who joined the scheme was limited. In 1992 only 1.6 % of the total production quota had been withdrawn. These withdrawn quotas led to a small reduction in the domestic production since the quotas were not redistributed. Over the next years the quota buy-out scheme attracted even less interest, and in 1994 it was abolished. The system of reducing individual quotas by a factor in

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

89 order to coordinate the produced amount with the quantity demanded was used until a new quota system was implemented in 1996.

The quota buy and sell scheme (since 1996 - 2003) The buy-out scheme had been an ineffective instrument in the attempt to restructure the dairy sector. This trigged the introduction of a revised system in 1996; the so-called “quota buy and sell scheme”. One of the reasons for introducing this system was that economic theory argues that this system will reallocate the quotas to the most efficient farmers and thus contribute to make the agricultural sector more efficient. The buy and sell scheme was based on administratively set prices. The quantity of quotas (measured in litres) demanded was higher than the quantity supplied. Since a market with fixed prices could not be expected to balance quota supply and demand, the reallocation (sale) of quotas had to be administered in accordance with certain rules. In order to maintain the regional structure of dairy production, Norway was divided into six trading regions. Quotas supplied by farmers selling quotas, should primarily be redistributed within the region. The quota was redistributed by the following principles: All eligible dairy farmers were offered to buy a minimum level of 1 000 litres production quota. Beyond this minimum level, available production quantity was to be allocated to each applicant according to the total amount applied for. The amount bought could not exceed 20% of the previous quota. An upper limit was specified: The maximum production quota that could be held was 130 000 litres. There were some exemptions from this rule. In 1997, 25 million litres of the total quota that was sold by farmers to the state authority was not redistributed (sold) to the farmers that wanted to buy quotas. The quantity that was left to redistribute (sell out) by the state added up to approximately 2 000 litres per buyer. To enter the market as a new producer was only possible if one produced organic milk near dairies that processed organic raw milk or if farmers were forced to switch from sheep to dairy production due to predator problems. In 1997 the two-price system was changed to a levy-system. This meant that the farmers had to pay a levy nearly as high as the milk price for any milk produced beyond the quota system. The total consumption of milk has decreased over the last number of years. For example, in 1992, the Norwegians consumed 1 784 million litres while the quantity had decreased to approximately 1 500 million litres in 2002. The Buy and Sell Scheme has been the main policy instrument to reduce the surplus production of milk. Now, it seems as if the consumed quantity has stabilised at the level of 2002. Nearly all of the total quantity sold in the buy and sell scheme was redistributed (34.7 million litres) to farmers in 2002. It is expected that the consumption of milk in 2003 will increase slightly. The individual quotas were increased by a marginal factor both in 2002 and 2003.

The quota buy and sell scheme – 30% traded directly between farmers (2003 -) In 2002 the buy and sell-scheme was changed. As earlier, farmers selling their quota can choose to either sell all or nothing. What is new is that the state authority no longer buys the whole quota but only 70 % of it, still at a regulated price. The remaining 30% can be sold at a non-regulated market price directly to other producers who already hold a quota. This requires that the buyer is located within the same region as the seller.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

90 This system is implemented in 2003, and approximately 28 million litres will be sold this year. The total volume is redistributed (sold out), i.e. none of the quantity is held back in order to reduce the milk production. In areas where there is a surplus after the quotas have been redistributed, it is opened up for allowing new producers to enter the market. This is limited to the actual surplus within the region. These changes in the buy and sell-scheme only apply to cow-milk. Goat-milk is redistributed in accordance to the buy and sell-scheme established in 1996. The maximum quota at the single farm, after buying quota, is increased to 250 000 litres for cow-milk and 125 000 litres for goat-milk. Another mechanism in the quota policy system is the so-called joint operation. Joint operation is when two or more farmers holding separate quotas cooperate in order to produce milk. The sum of the individual quotas cannot exceed 750 000 litres of milk. The quotas still belongs to the individual farms but the milk is produced in a common production facility.

Local milk production and processing outside the quota system (initiated in 2003) The first of July 2003, the Government made it possible for small-scale local milk producers, to produce milk outside the quota system provided that they produce, process, market and sell the milk themselves. The production limit is 250 000 litres for cow-milk and 125 000 litres for goat-milk. Milk producers can also cooperate in processing milk, provided that the volume does not exceed 500 000 litres. The intension of this change is to diversify the milk products offered in the market, and to stimulate production of local high quality varieties of food.

Review of the quota system for goat milk (2003) In the Agricultural Agreement 2003, the Government and the farmer unions agreed to review the quota system for goat milk, and to propose possible changes to increase the room of manoeuvre for the producers delivering goat milk to dairies for industrial processing. A fundamental evaluation of the quota system for goat milk is a part of this review. This sector is facing challenges such as small producer communities and long transportation distances to the processing industry. This review is at an initial phase.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

91

References Abler, D.G., (2000), Elasticities of Substitution and Factor Supply in Canadian, Mexican, and US Agriculture, Report to the Policy Evaluation Matrix (PEM) Project Group, OECD, Paris. Alston, J.M. and Quilkey, J.J.(1980), ‘Insurance Milk”, Australian Journal of Agricultural Economics 24 (3): 283-290 Alston, J.M. and Spriggs, J. (1998), “Endogenous policy and supply management in a post-GATT world”. Canadian Journal of Agricultural Economics 31(1): 220-239. Australian Bureau of Agricultural and Resource Economics - ABARE (2001), The Australian dairy industry: Impact of an open market in fluid milk supply, Canberra, ACT, Australia. Australian Competition and Consumer Commission (2001), Impact of farm-gate deregulation on the Australian milk industry: study of prices, costs and profits, Australian Competition and Consumer Commission, Dickson, ACT, Australia. Bailey, A.S. (2004), Potential Benefits from “Latent” Structural Change Following the Removal of Milk Quotas for the European Dairy Farm Sector. Presented to the Agricultural Economics Society 78th Annual Conference, Imperial College, South Kensington, London 2-4 April 2004. Barichello, R.R. (1981), The Economics of Canadian Dairy Regulation. Economic Council of Canada. Technical Report No. E/12, Ottawa. Barichello, R.R. (1984), Analyzing an Agricultural Marketing Quota. Center Discussion Paper No. 454. Economic Growth Center. Yale University. Barrichello, R.R. (1996), “Capitalizing Government Program Benefits: Evidence of the Risk Associated with Holding Farm Quotas.” The Economics of Agriculture, Volume 2, Papers in Honor of D.Gale Johnson. J.A. Antle and D.A. Sumner eds, Chicago IL: The University of Chicago Press, pp 283-299. Barrichello, R.R. (1999), "The Canadian Dairy Industry: Prospects for Future Trade." Canadian Journal of Agricultural Economics. 47(5):45-55. Boots, M., Lansink, A.O. and Peerlings, J., (1997), “Efficiency loss due to distortions in Dutch milk quota trade.” European Review of Agricultural Economics 24 (1): 31–46. Bouamra-Mechemache, Z., Chavas, J.P., Cox, T. and V. Réquillart, (2002), “Partial Market liberalization and the efficiency of policy reform: the case of the European dairy sector.” American Journal of Agricultural Economics, 84(4):1003-1020) Burrell, A., (1987), “EC agricultural surpluses and budget control”. Journal of Agricultural Economics 38(1): 1–14. Burrell, A., (1989), Milk Quotas in the European Community. CAB International Chapter 8. Chavas, J.P., and Holt, M.T. (1990), Acreage Decisions Under Risk: The Case of Corn and Soybeans. American Journal of Agricultural Economics. 72: 529-538. DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

92 Chen, K. and Meilke, K. (1998), "The Simple Analysis of Transferable Production Quota: Implications for the Marginal Costs of Ontario Milk Production." Canadian Journal of Agricultural Economics. 46(1):37-52. Colman, D. (2000), “Inefficiencies in the UK Milk Quota System”, Food Policy 25 (1), 1-16. Colman, D. (2002) (Ed.), Phasing out Milk Quotas in the EU, Report to DEFRA, SEERAD, NAW, and DARDNI, Published by CAFRE, SES, U. of Manchester, and at Colman, D., Burton, M.P., Rigby, D.S. and Franks, J.R., 1998, Economic Evaluation of the UK Milk Quota Scheme. CAFRE, School of Economic Studies, University of Manchester, Manchester. Commission of the European Communities (2002), Report on Milk Quotas, Commission Working Document, Brussels. Court of Auditors (2001), “Special Report No 6/2001 on Milk Quotas together with the Commission’s Replies” European Communities Court of Auditors, Luxembourg. Dawson, P.J., (1991), “The simple analytics of agricultural production quotas.” Oxford Agrarian Studies 19(2), 127–130. Guyomard, H. and Mahé, L.P. (1994), “Is a production quota Pareto superior to price support only?” European Review of Agricultural Economics 21 (1): 31-36. Guyomard, H.,X, Delache, X., Irz, and Mahé L.P. (1996), “A microeconometric analysis of milk quota transfer: Application to French producers”, Journal of Agricultural Economics 47(2): 206-223. Harvey D.R. (1984), “Saleable quotas, compensation policies and reform of the CAP” in : K.J. Thomson & R.M. Warren (Eds) Price and Market Policies in European Agriculture. p 291204. Harvey, D.R. and Hubbard, L.J. (1984), “A comparative static analysis of the welfare impact of supply restricting marketing boards: a comment.” Canadian Journal of Agricultural Economics 32 p. 570-574. Hennessy, D.A. (1995), “Quotas, alternative technologies and immeserization.” Canadian Journal of Agricultural Economics, 43, 203-208. Hickling Management Consultants. (1990), International Competitiveness of Dairy Food Processing in Ontario and Quebec. Report prepared for Industry, Science and Technology Canada, Food Policy Task Force and Subsidy Analysis Branch, Ottawa. Hubbard, L.J. (1984), “The use of marketing quotas in the EC dairy sector”, in : K.J. Thomson & R.M. Warren (Eds) Price and Market Policies in European Agriculture. p 205-211. INRA-University of Wageningen (2002), Study on the Impact of Future Options for the Milk Quota System and the Common Market Organisation for Milk and Milk Products. Jansson, T. (2002), Consequences for agriculture, consumers and taxpayers of abolishing milk quotas in the EU. Paper presented at the 10th EAAE-Conference, August 28-31, 2002 in Zaragoza, Spain.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

93 Jensen, H. G. and Frandsen, S. E. (2003), Impacts of the Eastern European Accession and the 2003-reform of the CAP Consequences for Individual Member Countries, Working paper No. 11/03, Danish Research Institute of Food Economics (FOI) Josling, T. (1984), “US and EC farm policies: An eclectic comparison”, in : K.J. Thomson & R.M. Warren (Eds) Price and Market Policies in European Agriculture. p 2-19. Just, R.E. (1974), An Investigation of the Importance of Risk in Farmers’ Decisions. American Journal of Agricultural Economics. 56: 14-25. Kleinhanss, W., Manegold, D., Bertelsmeier, M., Deeken, E., Giffhorn, E., Jägersberg, P., Offermann, F., Osterburg, B., Salamon , P., (2002) Phasing out Milk Quotas - Possible Impacts on German Agriculture; Federal Agricultural Research Centre, Institute of Market Analysis and Agricultural Trade, Braunschweig. Lippert, O. (2001), The perfect food in a perfect mess: The cost of milk in Canada, Public Policy Sources, Number 52, The Fraser Institute, Vancouver, B.C., Canada. Lips, M. and P. Rieder (2003), “The abolition of the raw milk quota in the European Union – An analysis on a member country level.” Revised version of the paper “Endogenous adjusted output quotas – The abolishment of the raw milk quota in the European Union”, in the Proceedings of the 5th Conference on Global Economic Analysis, (2002) Volume 2: 4d1 – 4d13, Centre for Sustainable Development, Taiwan. Meilke, K., Sarker, R. and Le Roy, D. (1996), “Analyzing the potential for increased trade in dairy products: a Canadian perspective”. in Understanding Canada/United States Dairy Disputes - Proceedings of the Second Canada/U.S. Agricultural and Food Policy Systems Information Workshop, Ed. by R.M.A. Loyns, K. Meilke and R. D. Knutson, Published by: Department of Agricultural Economics and Farm Management, University of Manitoba Moschini, G., "Modeling the Supply Response of Supply Managed Industries: A Review of Issues." Canadian Journal of Agricultural Economics, 37(1989): 379-392. Moschini, G., “Modeling the effects of supply constraints on the Canadian agricultural sector: A dual approach. Ph.D. Dissertation. Department of Agricultural Economics and Business, University of Guelph, Guelph, Ontario. OECD (1990), The management of dairy quotas in OECD countries. Directorate for Food, Agriculture and Fisheries, Paris OECD (2001), Market Effects of Crop Support Measures, Directorate for Food, Agriculture and Fisheries, Committee for Agriculture, Paris OECD (2003), OECD Agricultural Outlook 2003-2008, Directorate for Food, Agriculture and Fisheries, Committee for Agriculture, Paris OECD (2004), Analysis of the 2003 CAP Reform, Directorate for Food, Agriculture and Fisheries, Committee for Agriculture, Paris Oskam, A.J. and Speijers, D.P. (1992), “Quota mobility and quota values”. Food Policy, 1992 (1): 41-52.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

94 Rainelli P. and Vermersch D. (1997), “Thematic network on CAP environment in the EU”, 26 p. contribution for CAP and the rural environment in transition, a panorama of national perspectives, edited by Brouwer F. and Lowe P., Wageningen Pers (1998), 355 p Salamon, P., Bertelsmeier, M., Jägersberg, P. and von Ledebur, O. (2002), “Modelling the Phasing Out of Milk-Quotas in Europe - An Overview”, 10th Congress of the European Association of Agricultural Economists, Zaragoza, August 28-31. Salhofer, K., (2000), Elasticities of Substitution and Factor Supply Elasticities in European Agriculture: A Review of Past Studies, Report to the Policy Evaluation Matrix (PEM) Project Group, OECD, Paris. Suzuki, N., and Kaiser, H.M. (1994), "Basic mechanism of Japanese dairy policy and milk market models: A comparison with Untied States dairy policy" Journal of Dairy Science 77: 17461754. Swinbank, A. and Peters, G.H. (1990), “Who pays a tax in kind?” Oxford Agrarian Studies 18: 123-132. Swiss Federal Office for Agriculture (2001), Agricultural Report 2001. Federal Office for Agriculture, Mattenhofstrasse 5, 3003 Bern. Turvey, C., Weersink, A. and Craig, M. (2002), “The value of dairy quota under a commercial export milk program”, Working Paper 02/12, Department of Agricultural Economics and Business, University of Guelph, Guelph, Ontario. Van Tongeren, F. (2002), “Forward-looking analysis of reforms of the EU dairy policy”, 10th Congress of the European Association of Agricultural Economists, Zaragoza, August 28-31. Veeman, M.M. (1982), ‘Social Cost of Supply Restricting Marketing Boards”, Canadian Journal of Agricultural economics 30:21-36.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

95

CHAPTER 3 ANALYSIS OF INTERNATIONAL DAIRY TRADE LIBERALISATION

Abstract This chapter briefly describes the basic characteristics of world dairy markets and presents an empirical analysis of dairy trade liberalisation. The empirical work was carried out using the Secretariat’s Aglink and PEM models in order to assess the impact of dairy policy reforms on production, consumption, trade, prices, income and welfare. The empirical results have to be viewed within the limits of these models and should be interpreted as giving broad indications of the possible direction and potential changes in markets, income and welfare due to liberalisation in the dairy sector, rather than as definitive forecasts of the outcomes. The simulation results of international dairy trade liberalisation suggest that there is potential for significant net welfare improvements, with consumers being the main beneficiaries and taxpayers also realising gains in reformed countries. The size of benefits to consumers, however, is subject to the degree of price transmission along the supply chain. It can also be expected that producers and exporters in developing countries gain from the reform, while consumers in these countries would face a reduction in welfare. Following international dairy trade liberalisation, world dairy prices would be lifted substantially, by 17 to 54%. While supply would shift towards more efficient areas, there would not be any significant change in total world milk production. In this respect, the assumption concerning the production potential in quota operating countries plays an important role in the analysis. Sensitivity analyses to test for this, however, confirm the results, and do not lead to fundamentally different conclusions. The results further suggest that market price support not only artificially depresses world market prices but also creates considerable distortions in price formation of dairy products. The analysis points to the logical fact that domestic price and supply adjustments can be expected to be highest if a country reforms its dairy policy unilaterally. As more countries join the reform process, adjustments become smaller and are least in the case of multilateral reform. This is also the reason why in this case the dairy sectors in countries where market price support is eliminated are ultimately not greatly affected.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

This page intentionally left blank

97

Introduction Milk producers, in virtually every OECD country and in many non-member economies, benefit from government interventions that boost the prices they receive for their raw milk production. Government support and protection for milk producers is also more widespread than for any of the other commodities for which the OECD calculates producer support estimates (PSE). As a result, milk is one of the most heavily protected agricultural commodities, with an average OECD-wide percent PSE in 2000-02 of 46%. The support to milk producers as measured by the PSE amounts to 16% of the total PSE as calculated for OECD countries (OECD 2003a).1 2 The majority of support to milk producers is delivered through market price support. In general, milk price support at the farm level is achieved either through trade measures (import tariffs, tariff rate quotas and/or export subsidies) applied to dairy products or through a combination of trade measures and discriminatory pricing arrangements. Even after the WTO Agreement on Agriculture, dairy trade continues to be distorted by average bound tariffs that are among the highest of all agricultural commodities, by a large number of tariff-rate quotas (TRQs), low minimum access requirements, a number of special safeguard provisions, and use of export subsidies and other export support. Consequently, the evaluation of the consequences of international dairy trade liberalisation is of a particular interest for OECD and non-member economies. Despite the high level of interest, the analysis of trade liberalisation in the dairy sector is a complex exercise and there is only a limited number of studies that have focused on this issue. The intricacies of the dairy sector, where from one input (raw milk, bulky and largely non-traded) many outputs (which are often joint products: cheese/whey powder, butter/SMP) can be produced and traded, complicates the analyses of dairy trade liberalisation. The individual policy measures applied in the dairy sector (trade measures, pricing arrangements and quota systems), analysed separately in Chapters 1 and 2 are simultaneously reduced or removed in order to assess the impact of international dairy trade liberalisation on production, consumption, trade, prices, income and welfare. To use the complementarities of Secretariat’s models the empirical analysis is carried out with the Aglink and PEM models. The Aglink analysis addresses potential commodity impacts. The PEM analysis focuses on economic costs and benefits — the potential welfare impacts. This chapter is organised as follows. The next section describes basic characteristics of world dairy markets and trends in trade in dairy products. That description is followed by a discussion of the composition of and trends in agricultural policies supporting dairy production. These two preceding sections provide a background for the analysis of dairy trade liberalisation. The empirical analysis is described in the fourth section of the paper, where the Aglink and PEM results are presented and discussed. The results are summarised, important caveats noted and main conclusions and lessons drawn in the final sections of this chapter.

The world dairy markets This section describes the basic characteristics of milk and dairy product markets and provides a background for the analysis of dairy trade liberalisation. Some specifics of milk production and markets are briefly explained. The focus is on a description of the DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

98 structure of dairy markets and on major trends in milk and dairy product production, consumption, and trade.

Production and consumption of milk and dairy products Milk is a bulky, highly perishable commodity subject to bacterial contamination. A given unit of milk contains a certain amount of milk fat, protein and other solids. The level of these components in milk changes only slowly over time. The components are used in dairy products in various, mostly fixed, proportions (for example, butter contains mostly milk fat, skimmed milk powder contains non-fat solids while different cheeses are produced with varying proportions of these components). Milk is produced continuously every day throughout the year. Milk production systems differ depending on the degree of integration with crops and its relation to land. In general, livestock production systems are differentiated into grazing, mixed farming and industrial systems (de Haan et al., 1997). Industrial systems where milk is produced by feeding animals concentrates made from grains and oilseeds are expensive in terms of energy used. Grazing is typically considered the low cost production system but it is subject to marked seasonal variations. A review of milk production developments over the last three decades reveals that world milk production trended upward from the 1970s until the mid-1980s, then declined mainly as a consequence of quota level reductions in the European Union and increased only modestly in the 1990s. The highest increases in milk production have occurred in non-member economies and in that part of the OECD area not subject to production quotas (mainly Oceania). Although historically the larger share of global milk production has been produced in the OECD area, world output is currently equally split between OECD countries and non-member economies, although the latter are expected to rapidly increase their share (OECD, 2003b). Geographically, the shift in milk production is directed mainly from North to South, and more specifically from Europe and North America to Asia, Latin America and Oceania. Despite the operation of a milk production-limiting quota system, the European Union remains the world largest dairy market and milk producer. In fact, following the 2004 enlargement the EU has considerably strengthened its already dominant position, with a total output of about 144 mt.3 The second largest milk producer is India with about 88 mt (of which 47 mt is buffalo milk), followed by the United States with 77 mt. Cow milk remains the most important milk produced, accounting for 84% of the total. Other types of milk (and their share of total world production) are buffalo milk (12%), goat milk (2%) and sheep milk (1%). It is worth noting that while in the last 10 years the production of cow milk has grown by less than 10% the production of buffalo milk has grown by almost 40% (IDF, 2003). Figure 3.1 compares per capita production and consumption (in milk equivalents) of dairy products. It is evident that, with the exception of Oceania and South America, the differences are small, resulting in equally small net import or net export positions, compared to total milk production. Part of the reason for the fact that local supply typically provides the majority of local dairy consumption stems from the bulky and perishable nature of milk, which limits the tradability of the product.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

99 Figure 3.1. Comparison of per capita consumption with per capita production of milk (in milk equivalents)

Oceania Western Europe

810 kg/pc

CEEC+CIS NAFTA South Asia South America Other Africa

per capita consumption (kg)

N.Africa&MidEast

per capita production (kg)

Central America S.E. Asia

0

100

200

300

400

Source: Date from IDF-CFCE (2003).

Global production trends over the last three decades for major dairy products, indicate that the production of skim milk powder (SMP) has somewhat declined, that of butter has remained relatively stable, while WMP and especially cheese production have substantially increased (Figure 3.2). Cheese is by far the most important dairy product and increasingly more milk is being channelled towards its production. These production trends have been to a large extent determined by changes in consumption. These changes in turn have been driven, in OECD countries, by per capita income growth but also by evolving nutrition and health concerns. Elsewhere, growth in demand for dairy products has been driven mainly by growing per capita income, changes in lifestyle, urbanisation and population growth. Figure 3.2. Production trends of major dairy products 20 Butter Cheese million tonnes

15

SMP WMP

10

5

0 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003

Source: OECD Secretariat. DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

100 Over the last two decades, cheese (a relatively income-sensitive product) has recorded a dramatic increase in consumption in OECD countries, but that in non-member economies has remained relatively flat (Figure 3.3). The share of world cheese consumption attributable to the OECD area has increased from 63% in 1982 to 78% in 2002 and is expected to grow further. Butter consumption on the other hand, has been slowly increasing in non-member economies, but has been on the decline in the OECD area where demand from restaurants and hotels is not enough to compensate for the fall off in consumption in private households. Figure 3.3. Consumption trends of dairy products in OECD and non-member economies 20 NME

16 million tonnes

OECD

12 8

Butter

Cheese

SMP

2002

1992

1982

2002

1992

1982

2002

1992

1982

2002

1992

0

1982

4

WMP

Source: OECD Secretariat.

World SMP consumption has been declining in OECD as well as non-member economies partly due to the reduction in feed use and partly due to the fact that other “non traditional” dairy products (whey protein concentrates, milk protein concentrates) are increasingly used as a source of milk solids. In addition, WMP powder has been continuously replacing SMP in the milk reconstitution market of non-member economies. The consumption of WMP in non-member economies has been increasing swiftly over the last decades and currently more than three quarters of all WMP is consumed there.

Trade in dairy products As has been the case in the past, the considerable cost of transporting bulk raw milk suggests that the majority of milk to be consumed in fluid form will continue to be produced and processed near the point of consumption. The differentiation of milk into largely non-tradable fluid (drinking) milk and manufacturing milk that is tradable in the form of milk products is a typical characteristic for the dairy sector and complicates the analysis of dairy markets and dairy policy reform. Despite technological developments in refrigeration and transportation, international trade in milk and milk products represents only about 5-7% of world production of cow milk (intra-EU trade excluded). In contrast, international trade in dairy products accounts for about 48% of the production of WMP, 27% of SMP, 10% of butter and 7% of cheese. It follows, then, that a relatively small change in the supply/demand balance of milk may have a substantial impact on traded dairy products. The relative thinness and volatility of

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

101 dairy markets is further enhanced by restrictions on market access through border measures and by export support. Trade in dairy products has increased in the last decade mainly for cheese, which saw trade soar by more than 50%, and WMP for which trade increased by 30% (Figure 3.4). On the other hand, trade in butter and SMP has been stagnating. This reflects an ongoing trend in world markets for dairy products from supply-led trade in bulk commodities (SMP, butter) to demand-driven trade in high value-added products, such as cheeses. Figure 3.4. Trends in trade volumes and market shares for major dairy products and exporters 1600 Rest of World New Zealand thousand tonnes

1200

EU Australia

800

400

0 1992 Butter

2002 Butter

1992 2002 Cheese Cheese

1999 SMP

2002 SMP

1992 WMP

2002 WMP

Source: OECD Secretariat.

As dairy products are typically consumed in the country of origin, world export markets have few big players – mainly the EU, New Zealand, and Australia (the United States is important on the SMP export market). Figure 3.4 illustrates that the dominance of these three countries on world dairy markets in the last decade has not weakened despite the increase in the volume of international trade. Nevertheless, reflecting quota constraints on milk production and limitations on the volume of subsidised exports under the Uruguay Round Agreement on Agriculture, the EU has lost a considerable share of world dairy markets to Australia and in particular to New Zealand. Dairy production in New Zealand is based on grazing with very low production costs. Around 97% of all milk produced is exported and the dairy industry accounts for almost a quarter of New Zealand’s total export earnings. While New Zealand’s share of world milk output is less than 3%, its share of world trade is more than 30% - and expanding. Similarly, Australia, with less than 2% of world milk production, supplies almost 20% of global dairy trade. On the import side the situation is much more fragmented. South America imports mostly whole milk powder, Mexico imports SMP, Africa imports SMP and WMP and the EU imports mainly butter and cheese from New Zealand. Whole milk powder, processed cheeses and feta are imported by Middle East countries. Although the US is exporting some of its cheddar production it is an importer of speciality cheeses. Asia is an increasingly important importer of WMP, butterfat and cheeses. Generally speaking, low value products are exported to developing countries while high value products are largely exchanged among high income countries.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

102 Trends in world dairy prices indicate that in real terms they have declined substantially since the early 1980’s (Figure 3.5). At that time world prices were to a large extent driven by subsidised exports from the European Union and there was a sharp decline in prices as the surplus problem in the EU mounted. Following the EU quota level cuts in the second half of the 1980s world prices temporary increased, later resuming the general declining trend. It is also striking that in the 1980s prices of fat (butter) were considerably higher than those for non-fat solids (SMP). The situation has been reversed in the 1990s and butter prices are currently substantially lower than the SMP counterparts. Figure 3.5. Declining world dairy prices in real terms

400 Butter Cheese SMP WMP

USD/Kt

300

200

100

0 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003

Source: OECD Secretariat.

Agricultural policies supporting dairy production In general, support levels for milk are higher than for most other commodities within countries and, with just a few exceptions, are high in most countries. Market price support (tariffs and export subsidies, administered prices) is the main form of support provided to milk producers. Many countries impose production-limiting quotas to control surpluses resulting from high support prices; in particular to manage public stocks and expenditures on export subsidies. The OECD gathers information regarding the level of support provided to producers through all types agricultural policy measures. The %PSE expresses the monetary value of this support as a share of gross farm receipts (market returns plus support). A notable feature of the %PSE for milk, calculated at the total OECD level, is the downward trend in support since the early 1990s, falling from a high of 59% in 1986-88 to 46% in 200002 (Figure 3.6).

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

103 Figure 3.6. OECD average Producer Support Estimate for milk, 1986-2002 Per cent of value of gross farm receipts %

70 60

P ayments bas edon income andmiscellaneous

50

P ayments bas edon input constraints

40

P ayments bas edon input us e

30 20

Market price support

10

P ayments bas edon historical entitlements P ayments bas edon area planted/animal numbers P ayments bas edon output

19 86 19 87 19 88 19 89 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 p

0 Market P rice Support

Source: OECD PSE/CSE database.

Along with rice and sugar, milk is one of the most highly supported commodities. Support for milk is significantly higher than that provided to other livestock products such as beef and sheepmeat. Between 1986-88 and 2000-02 there has been a reduction in the level of support provided to dairy producers in all countries except Norway where it has stayed the same, and in Hungary and Poland where it has increased.4 The reduction in support to dairy producers has been most significant in absolute terms in Australia, the Czech Republic, the European Union, Switzerland and the United States, with a reduction in the %PSE of more than ten percentage points, with the greatest proportional decrease having occurred in New Zealand (with currently the lowest %PSE in all OECD countries). Within the total OECD PSE there are significant variations between countries in the level of support provided to dairy producers (Figure 3.7). Support levels in 2000-02 were highest in Japan and Korea and the non-EU European countries of Iceland, Norway and Switzerland where over 70% of gross farm receipts for dairy producers are generated by support policies. In the European Union, Hungary, Canada, Mexico and the United States, it ranges between 45-55%. In the Czech Republic, Slovak Republic and Turkey, support averages just over 30%, while it has been very low throughout the whole period in New Zealand and Poland. The greatest reduction in support over the period took place in Australia, whose support to dairy producers was more than halved.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

104 Figure 3.7. Producer Support Estimate for milk by country, 1986-88 and 2000-02

Japan Ic e la n d N o rw a y S w itz e rla n d K o rea C anada U n ite d S ta te s OECD H u n g a ry (2 ) E u r o p e a n U n io n M e x ic o (2 ) 1 9 8 6 -8 8

T u rke y

2 0 0 0 -0 2

S lo v a k R e p u b lic (2 ) C z e c h R e p u b lic (2 ) A u s tr a lia P o la n d (2 ) N e w Z e a la n d -2 0 %

-1 0 %

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

Notes: 1) Countries are ranked according to 2000-02 levels. 2) For the Czech Republic, Hungary, Mexico, Poland and the Slovak Republic, 1991-93 replaces 1986-88. Source: OECD PSE/CSE database, 2003.

Although the market price differential has decreased, trade barriers continue to offer significant protection to dairy producers in most OECD countries. Market price support policies are designed to protect producers from lower and variable world market prices, insulating them from world market developments; and they have been effective in doing so. For example, in 1997, the average price received by OECD dairy farmers was 90% above the border reference price but in 1998 the difference increased to 130% when the reduction in border prices was not matched by a similar reduction in producer prices.5 The importance of market price support reflects the historical use of trade measures e.g. tariffs, import quotas and export subsidies in many OECD countries to protect dairy producers from lower priced traded products and to enable domestic pricing arrangements. In almost all instances, tariffs on dairy products are above the country average for all agri-food products and are among the highest on agricultural products. Average tariffs vary considerably between OECD countries: they are comparatively low in Australia and New Zealand, and comparatively high in Canada, the European Union, Japan, Norway, Poland and Switzerland. The most significant user of export subsidies on dairy is the European Union, accounting for 81% of the total during the period 19952000, with Switzerland accounting for a further 10% of total export subsidies. Nevertheless, export subsidies from the EU have been declining over the period 20002003.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

105 While there is some variation between countries in terms of the composition of support provided to dairy producers, the most distorting categories of support dominate. Market price support has traditionally been the most dominant support category in all OECD countries except in New Zealand and has remained so with only a few exceptions. However, market price support in Canada, the European Union, Norway and Switzerland has been accompanied by restrictions on the level of production, i.e. milk quotas. Payments based on input use constitutes the next most important category of support, with every OECD country implementing support measures to dairy farmers that are classified in this category (e.g. subsidies to improve manure storage facilities, fuel tax rebates). Payments based on output are relatively important in Iceland, Norway and the Slovak Republic; payments based on animal numbers in the Czech Republic, Norway and Switzerland; and payments based on historical entitlements in Australia and Switzerland. There have been some attempts to introduce or increase support provided through less production distorting measures and through those more directly targeted at environmental or farm income objectives. For example, support measures classified under payments based on historical entitlements have been introduced to the benefit of dairy producers in Australia, Canada, the Czech Republic, the European Union and Switzerland. Measures classified under payments based on input constraints or payments based on overall farm income have been either introduced or increased in many countries, but their overall significance remains very low in all cases. There have also been increases in the most distorting forms of support in some OECD countries between 1986-88 and 2000-2002. The importance of market price support measures in gross farm receipts has increased for dairy producers in Canada, Hungary, Norway, Poland and Turkey, although producers in Canada and Norway have been constrained by production quotas. Payments based on output have been introduced in the Czech Republic, Hungary and the United States but these are all relatively small. They have also been expanded in Iceland, the Slovak Republic and Switzerland although in all three countries quantitative limits are placed on production. Canada completed a gradual phase-out of such payments in the 2001/2002 dairy year. While both the level and percentage change has been small in some instances, the importance of payments based on inputs in gross farm receipts has increased in Australia, the European Union, Hungary, Japan, Korea, Norway and the Slovak Republic.

International dairy trade liberalisation As noted at the outset, despite the interest in the area of dairy policy reform, there are few studies that have undertaken a complex analysis of trade liberalisation in the dairy sector. A list of such studies may be found in Meilke and Lariviére (1999). The following paragraphs review the results of international dairy trade liberalisation according to some recent studies. FAPRI (2002), using a partial equilibrium dynamic model, estimated that under a full liberalisation scenario the net trade of all dairy products (butter, cheese, SMP, and WMP) would increase relative to baseline levels. Argentina, Australia, and New Zealand all gain while the European Union would lose market share in all dairy products compared to the baseline. EU butter and SMP exports would decrease significantly as the elimination of milk production quotas and intervention buying results in lower milk production, lower domestic prices, less butter and SMP production, and more domestic consumption of all

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

106 dairy products. World butter prices were estimated to increase by 40%, cheese prices by 22%, SMP prices by 30% and WMP prices by 26% on average over the baseline period. Zhu et al. (1998) used a hedonic spatial equilibrium model and showed that full trade liberalisation would reduce milk prices in Canada by 32%, in the EU by 26%, and in Japan by 36%. Milk prices would increase in Australia and New Zealand by 22% and 51% respectively. The impact on US milk prices was estimated to be relatively small – a reduction of 0.4%. Butter prices in New Zealand were estimated to increase by 76%, cheese prices by 45%, and SMP and WMP prices by 32% and 45% respectively. The study shows that milk production in the EU and Canada would increase by 2% and 3% respectively. These increases in milk production following the large cut in prices reflect the assumed high quota rents in this study. Larivière and Meilke (1998) used a non-spatial, multi-region model of the world dairy sector and noted that world dairy product prices would increase substantially under free trade. Their simulation results showed that world butter prices would increase by 32%, cheese prices by 44% and SMP prices by 15%. World milk production would increase by 0.8% while world production and consumption of butter and SMP was estimated to increase by 0.3% and 2.3% respectively, and cheese production and consumption would decline by 0.6%. Simulation results for Canada showed domestic milk price reduction of 36% and milk supply increase of 6.9%. Using a derivative of the OECD’s Aglink model, Shaw and Love (2001) estimated impacts of increasing market access and reducing export subsidies for dairy products. The study found that the value of world dairy trade would increase by USD 1.8 billion relative to a 1999 baseline under an increased market access scenario, with the value of milk production rising in Australia, New Zealand, and Argentina (with increases from 7 to 9% relative to the base), and declining in the EU and the United States (1.2% to 1.4% decline). With export subsidies reduced by half, domestic prices would fall and the value of milk production would decline in the EU and remain unchanged in the United States. The volume of EU milk production was assumed to be bound at the quota level. Langley et al. (2003) used a partial equilibrium dynamic model in which a full liberalisation scenario of the dairy sector showed milk prices increasing in Australia, New Zealand and Argentina by26%, 24% and 22% respectively, while they decreased in Canada, the EU 15, the US and Japan by 35% in the first and 8% in the three other countries. The world price of butter was estimated to increase by 58%, the cheese price by 30% and those for SMP and WMP by 9% and 18% respectively. The volume of trade would decrease for butter, SMP and WMP and increase for cheeses and other dairy products. However, products would be traded at higher prices and the value of dairy trade was estimated to increase by USD 2 billion. The medium-term effects of the experiment resulted in 3-4% lower production of raw milk and dairy products in heavily subsidised countries such as the EU. On the other hand, production in Canada would increase by 12%. The above studies all used different models in structure, data and parameters (i.e. country and product coverage, length of run, different assumptions on product supply and demand etc.).6 Thus, it is no surprise that the results differ across studies. The results are also subject to the caveats and limitations of modelling of such a significant change in the world dairy sector. A general discussion regarding the problems of modelling dairy trade liberalisation can be, for example, found in Meilke and Lariviére (1999). Nevertheless, all studies show significant increases in world dairy prices, important redistribution of milk production and trade among regions and sizeable impacts on DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

107 consumers and producers in reformed and other regions. The studies differed in the size of the estimated impacts and, importantly, also in the size and direction of milk supply adjustment in countries where policies were reformed in the presence of production quotas. An analysis of international dairy trade liberalisation has been also undertaken by the Secretariat as part of a broader project evaluating the consequences of dairy policy reform. In the analysis, individual market price support policy measures are simultaneously reduced or removed in order to assess the impact of dairy trade liberalisation on production, consumption, trade, prices, income and welfare. The empirical simulations are carried out using the Secretariat’s Aglink and PEM models (see Annex A for a description of the Aglink and PEM models). The results of the simulations are presented and discussed in turn. However, it should be noted that Aglink and PEM do not produce identical results. These models were developed for different purposes, and accordingly have different structures, country, product and time coverage. The simulation experiments are designed with the specific strengths and weaknesses of each model in mind, and their objectives are different. While Aglink and PEM results are not directly comparable, the two models have been used jointly in this analysis to complement the market and trade insights gained with Aglink with those from PEM results which focus more on income and welfare. Thus, it is important in reading the following results to keep in mind that the simulation experiments are not strictly comparable. For example, the welfare analysis conducted with PEM refers to market and trade results of PEM only and do not correspond to Aglink results. It is also important to be aware of the general limitations of analysing dairy trade liberalisation, such as those discussed in Meilke and Lariviére (1999). More specifically, the dairy products produced, consumed and traded in the Aglink and PEM models are assumed to be homogeneous. No account has been taken of different product attributes or different production practices, which can influence costs of production, product prices and consumer choice. Another limitation concerns the absence of risk in the analysis as risk and producers’ attitude toward risk influence production decisions. Another caveat is the fact that the rest of the world (ROW) is treated as a single block in the models. Many countries that are implicitly represented in the ROW module also use market price support and other policy measures. Given the size of the ROW, in terms of milk and dairy product production and consumption, the response of this region to trade liberalisation is crucial for the results of such a scenario.7 Further, given the high level of protection in the dairy sector the scenario of full trade liberalisation creates a shock to the model beyond what many would consider “marginal” or small. In addition, the models do not allow for structural change that can be expected to take place following such a shock. Moreover, the impact of trade liberalisation depends heavily on the production response subsequent to the removal of milk production restrictions in countries that employ a milk quota. Thus, the results are influenced by the milk supply response assumptions in quota operating countries. Finally, the results also depend on the price transmission through the supply chain. The usual assumption of perfect price transmission might be contested on the ground of increasingly high concentration at the processing and retail level. Many of these caveats are common to large world agricultural commodity models, reminding us that models are only an imperfect approximation of the reality. Some of the caveats will be addressed via sensitivity analysis and additional scenarios. The sensitivity analysis is useful in helping to understand results, and use of this tool is made here in DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

108 both the Aglink and PEM analyses. The primary benefit of such analysis is not to put a range on the results such that we are convinced that truth lies somewhere within, but rather to see if the basis upon which conclusions are drawn are sound. In any case, the conclusions drawn from the analysis in this paper take care to use the results indicatively rather than definitively. That is, the results are used to say something about the nature, possible direction and potential changes in markets, income and welfare due to liberalisation in the dairy sector, rather than to provide definitive forecasts.

Aglink results of dairy trade liberalisation scenarios- Evaluating the market and trade impacts Aglink is a partial equilibrium dynamic supply-demand model of world agriculture, developed by the OECD Secretariat in close co-operation with member countries. It represents annual supply, demand and prices for the principal agricultural commodities produced, consumed and traded in member countries. The dairy component of this model covers production and consumption of milk and main dairy products in major OECD and several non-member economies markets, covering both importers and exporters. Thus, the Aglink representation of the dairy sector allows the analysis of impacts on world markets for tradable dairy products where those markets are explicitly modelled. (see Annex A for a description of the Aglink model). The simulation experiments are conducted using the baseline data of the Agricultural Outlook 2003-2008 published in OECD (2003b) (See Annex 3.3 for the description of the policy assumptions for dairy markets for the baseline period 2003-2008). In general, agricultural markets are modelled in Aglink specifically to best capture individual policies and particular market settings relevant for each country. It follows that certain modifications to the model are required for a scenario such as a major dairy policy reform where support policies are eliminated or much reduced. For example, in quota operating countries represented in Aglink, a milk supply function had to be introduced, based on assumptions for quota rent and long term supply elasticity.8 All border measures were removed and all equations for subsidised exports of dairy products were eliminated. All domestic market clearing price identities were replaced by price transmission equations and trade becomes the market clearing identity. Since world prices in Aglink are based on the Northern European price, adjustments to reflect transport costs were implemented in modules other than the EU module. In the EU module an adjustment was needed in the price transmission equation of cheese, to reflect the different types of cheese used to denominate EU and world prices.9 Where it was appropriate, consumer subsidies were eliminated. Where domestic price discrimination existed, the fluid milk price was determined by the industrial milk price plus 50% of the baseline fluid milk premium.10 11 (Annex A explains in more detail the functioning of the dairy market in Aglink). For further details on various modifications to the Aglink model see technical report OECD (2004). Any border protection, tariffs, production quotas, support prices and consumer subsidies were eliminated and any fluid milk premium reduced by half. Only market price support policies were eliminated, direct payments were kept to their baseline value if they are exogenous or were calculated by the model where they are endogenous. Scenarios of a complete elimination of all market price support policies have been produced initially per country/region. These scenarios have been merged one by one in the next step. This process has allowed for the isolation of the contribution of each country’s reform to the overall impact on world markets. DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

109 The results of the trade liberalisation scenarios simulated with Aglink are presented in Annex 3.1. The tables in the annex show results in percent changes from baseline levels for prices, production and consumption of milk and dairy products. The trade figures are shown in thousand tons. The baseline levels refer to the last year values (2008) of the Agricultural Outlook baseline 2003-2008 published in OECD (2003b). The columns following the baseline levels illustrate the reform scenario results for a particular country or group of countries. The name of the reformed country or region appears always in the columns’ headings. Thus the results show the market impacts of removal of market price support in a particular country while all other countries would keep their policies at baseline levels.12 The regions (group of countries) in the tables are defined as follows: Europe í a scenario of dairy policy reform in all the European components of Aglink,13 NAFTA í a scenario of dairy policy reform in all NAFTA countries (United States, Canada, and Mexico), Atlantic í a scenario which combines reform in Europe and NAFTA countries. Finally, a scenario involving reforms in all Aglink countries was produced. The results are presented under the heading ALL and are highlighted in bold. For convenience, the ALL scenario is in the text referred as Aglink multilateral scenario (reform in countries represented in Aglink). It is important to keep in mind that the word “multilateral” is strictly connected to Aglink countries and cannot be considered in a broader sense as for many developing countries (represented in Aglink in a single block Rest of the World) dairy policies are not modelled and hence not reformed. Each column in the tables presents a “contribution” of a particular country/region to the process of dairy policy reform under which milk market price support is completely eliminated. The description of individual scenarios is summarised in Annex 3.2. As discussed above, such type of analysis is subject to certain caveats and limitations. Some of the limitations were also identified at the dairy experts meeting held in Paris in September 2003 (OECD 2003c). Most notably there is uncertainty about the price transmission to consumers of reduced farm and wholesale prices resulting from reform. Box 3.1 highlights some of the main issues related to the subject of imperfect price transmission. While this subject deserves further investigation, a detailed analysis of price transmission was outside the scope of the present study. Instead, a simplified sensitivity analysis was constructed to test the impacts of presumed imperfect price transmission for dairy products specifically modelled in Aglink. More specifically, the Aglink demand functions for dairy products and fluid milk uses wholesale or farm prices instead of consumer prices. The elasticities used in these demand functions are therefore a composite of the demand elasticities relative to retail prices and of the elasticities of transmission from wholesale (or farm) to retail prices. The demand elasticities used are affected by the relationship between wholesale and retail prices that existed over the period of estimation. By construction, it is implicitly assumed that this relationship will neither change over time nor following a policy reform. This assumption may be contested on the basis of the very high level of concentration of the retail food industry in most OECD countries. In the imperfect transmission scenario (labelled in the tables as Retail) it was assumed that consumers in countries affected by the reform would only benefit from half the decline in producer prices recorded in the scenario of simultaneous dairy policy reforms.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

110

Box 3.1. Price transmission along the supply chain - issues and findings The vertical adjustment to price shocks along the chain from producer to wholesale and to retail, and vice versa, is an important characteristic of the functioning of markets. As such, the process of price transmission through the supply chain has long attracted the attention of agricultural economists as well as policy makers. In this respect, a common concern of policy makers relates to the assertion that price reduction at the farm level is not fully transmitted down the supply chain, so that final consumers might not be able to benefit fully from this drop in prices. The implication of the incomplete pass-through of prices would be smaller effects on consumer welfare and an increase in rents for the firms in the downstream sector. A number of reasons are put forward in the literature that attempt to explain the asymmetries and imperfect passthrough of prices. For example, prices at the retail level may not adjust due to so called menu costs, which are costs associated with making frequent changes in retail prices and the uncertainty of whether the price shock is permanent or transitory (Ball and Mankiw, 1994). Response might also be asymmetric due to inventory management strategies, so that retailers may reduce their prices more slowly compared to reduction in farm prices to prevent running out of stock (Reagan and Weitzman, 1982). Wohlgenant (1985) also demonstrated that lags between retail and wholesale food prices can be explained by inventory behaviour of retailers. Alternatively, retailers selling perishable goods might be reluctant to raise prices in line with an increase at farm level given the risk that they are left with spoiled product (Ward, 1982). Azzam (1999), using a two-period model of spatially competitive retailers, has shown that asymmetry can occur even in a competitive environment due to intertemporal optimizing behaviour so that retail prices may rise relatively more than decline. Gardner (1975) pointed out that in addition to other causes, farm-retail price asymmetries might be the result of government intervention to support producer prices. Serra and Goodwin (2003) studied price transmission in the Spanish dairy sector and argued that scarcity of milk to some extent created by quota system may lead to a situation in which industries compete to increase both their access to milk quota and their retail market share, and thus may not pass the farm level price increase fully to the retail level. Many of these arguments relate to an adjustment problem at the retail level and prices which may be “sticky” in the short run, can be expected to adjust in the long run. However, imperfect price transmission in the long run is often seen as being an outcome of market power and oligopolistic behaviour. Many empirical studies have supported this hypothesis. Wann and Sexton (1992) modelled the California pear industry and showed retail price enhancement above the competitive norm in canned pear and fruit cocktail markets, although the hypothesis of competition in the raw pear input market was also rejected. The asymmetry in price adjustment due to market power, albeit at a local level, was also noted by Benson and Faminow (1985) who argued that that consumers' choice between food stores is based on locational convenience and presence of search cost might thus create locally imperfect markets. Gohin and Guyomard (2000) strongly rejected the hypothesis that French food retail firms behave competitively and illustrated that more then 20% and 17% of the wholesale-to retail price margins for dairy and meat products, respectively, can be attributed to oligopoly-oligopsony distortions. Abdulai (2002) illustrated that increases in producer prices of pork in Switzerland that result in the reduction of the marketing margin are passed on to retail prices faster than reductions in producer prices that lead to increases in the marketing margin. Similar results were found by Kinnucan and Forker (1987) who estimated price transmission for dairy products in the United States and showed that transmission elasticities for rising farm prices were larger that corresponding elasticities associated with falling farm prices, depending on the dairy product. London Economics (2003) investigated the links between retail and farm gate milk prices in the UK, Denmark, France and Germany. The study has found that in the UK a unit increase in the retail price of liquid milk is fully transmitted to the farm gate price whereas a unit increase in farm gate prices results in only a 0.56 increase in the price at retail and a unit decrease in farm gate price reduces the retail price by 0.71. In Germany, the study also found two-way price transmission, though rather imperfect. In Denmark, the study found no evidence of price transmission in any direction. In France, farm gate price changes were mainly but imperfectly transmitted to retail prices. The study associated some of the resulting differences among countries to differences in market structures, differences in the transmission of information or government interventions. A study on the impact of peanut policy changes in the United States concluded that data on farm-to-retail price margins for peanut butter indicate that farmers are capturing a smaller share of the retail price of peanut butter (about 20%) following the reform as compared with an average of about 25% between 1988-1998 (Dohlman et, al., 2004). However, here, it is important to note that a change in the farmer’s share of the retail margin does not necessarily point to imperfect price transmission and has to be evaluated against the development of other input costs. That is, costs related to processing, packaging, moving, advertising, storing and numerous other activities must be acknowledged. Thus, along with the potential that price transmission may be delayed, it may also be obscured by a variety of other costs that must be paid by consumers when they buy a retail product that is in some way derived from farm-produced goods.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

111 Although, market power was often identified as the main cause of imperfect price transmission, recent research shows that this does not always have to be the case. McCorriston et al. (2001) demonstrated that price changes can be greater or smaller than the competitive benchmark case depending on the interaction between market power and returns to scale. That is, if the cost function is characterised by increasing returns to scale, the influence of market power might be offset and the level of price transmission may increase relative to the competitive case. Weldegebriel (2004) also argued that the presence of oligopoly and oligopsony power does not necessarily mean imperfect price transmission. The author has illustrated that the functional forms of retail demand and farm input supply are key factors in determining the level of price transmission. Bettendorf and Verboven (2000), using a model of oligopolistic interaction, showed that the week transmission of coffee bean prices to consumer prices in Netherlands was due to a relatively large share of costs other than the costs of beans. The authors concluded that the market was relatively competitive. Empirical results of Holloway (1991) suggest that, during the period 1955-83, departures from competition in the retail markets of the major food groups have been relatively insignificant, that is results are not statistically different from the outcomes under perfect competition. Serra and Goodwin demonstrated that asymmetries were not present in the price transmission of highly perishable dairy products in Spain supporting the theory that market power could be consistent with symmetric price relationships. The issue of market power in a rapidly restructuring food supply chain and its impact on margins and price transmission remains an issue of further research. As noted above, price transmission is a particularly vital issue in the context of a major agricultural policy change as it affects the welfare impacts of reform on market participants. For example, McCorriston and Sheldon (1996) have estimated the price transmission and welfare impacts of tariff reduction following changes in the EU banana regime. The authors have noted that the higher the number of stages in the vertical market structure characterised by market power of the market actors, the likelier is a lower pass-through of price changes (the theoretical framework for measuring impacts of successive monopolies in a vertical market chain can be found in Cotterill, 2002). The simulation results showed that, compared to a competitive case, the consumer surplus increased by as much as 49% for the single-stage monopoly case, compared with only 24% in case of a the two-stage monopoly. As expected, the singleand two-stage Cournot oligopoly cases showed relatively higher increases in consumer welfare, with 71% and 39% per cent, respectively, when compared to the consumer surplus gain of the competitive case. These results, however, are based on various hypothetical scenarios rather than testing to determine which case is most likely. Sexton et al. (2003) have analysed specifically the implications of vertical market structure for trade liberalisation and market access with particular attention to impacts on developing countries. The authors reiterated that given that the food sector is most appropriately characterised by a successive oligopoly/oligopsony situation, the implication of reducing tariffs is different in magnitude than that implied by models that assume perfect competition. The authors also pointed out the rapidly increasing trend of consolidation and concentration in the food industry which might reduce the share of benefits from liberalisation within the food marketing chain that would accrue to developing countries. Although the results of price transmission between the farm and retail level differ among studies depending on the method used and the country and commodity covered, many of these studies point to some degree of imperfect pass-through of price changes along the supply chain. This suggest that an analysis of trade liberalisation that does not account for vertical market structure may likely over-estimate the benefits to consumers in reformed countries as the reduction in farm prices might be less then perfectly transmitted to final consumers. The pass-through of benefits depends very much on the character of the primary and processed commodities, demand and supply specificities, and the market structure, in particular the market power of actors at individual levels of the vertical supply chain. The subject of imperfect price transmission remains an important topic for further research that may become increasingly important in the context of further market concentration and trade liberalisation.

Another important analytical issue concerns the assumptions used in constructing specific supply functions in countries which currently operate a quota system, as the supply response in these countries is not directly observable for the period that the quota is binding. More detailed discussion on modelling of milk supply in the presence of quotas is available in Chapter 2. The importance of assumptions made concerning quota rent and supply elasticities for modelling supply response under quota were discussed there in detail. However, an array of plausible assumptions can be made concerning quota rent calculation, and there will always be uncertainty about the chosen value. Thus, it was deemed necessary to undertake a sensitivity analysis with respect to the possible range of quota rent and supply elasticity assumptions regarding the underlying supply function.14

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

112 The results of this sensitivity analysis are presented in Annex 3.2 and described after the discussion of the main results of the trade liberalisation scenarios.

Results for the European Union15 Table 3.1.1 illustrates the dairy trade liberalisation scenario results for the EU. The results indicate that the unilateral reform by the EU alone reduces the internal price of milk by as much as 16.5%. It should be noted that this is a considerable reduction as it has to be compared to the 2008 level which is already subject to a price decrease from the 2003 levels reflecting the CAP policy measures assumed in the baseline.16 Reform in other European countries does not have a substantial impact on the EU market. However, combining the reform process of Europe and NAFTA reduces the domestic milk price in the EU by 12.3% which is by 4 percentage points less than under the unilateral reform. Under the multilateral (ALL) reform scenario, the milk producer price would be reduced by 9.8%, which is almost by 7 percentage points less when compared to the unilateral reform case.17 The difference stems from the fact that EU prices are directly linked to world dairy prices in the scenarios and world prices increase substantially more under the Aglink multilateral reform as compared to the unilateral reform case. The impact of these scenarios on milk production reveals a similar pattern and shows a fall in production following the price decrease. At the end of the base period, milk production in the EU would decrease by 10.7% (unilateral reform) and 7.3% (Aglink multilateral reform) respectively. These results are sensitive to assumptions concerning quota rent as will be discussed later. The impact on dairy prices is consistent with the current butter/SMP price policy tilt in the EU which is in direct contrast with the world market price ratio for butter and SMP. 18 While the world market price of butter has recently been well below the SMP price (and is so also in the baseline), the opposite is the case for EU support prices. Consequently as a result of MPS elimination the butter price in the EU falls by 34% (unilateral reform) and 25% (multilateral reform) respectively while the SMP price increases following the substantial rise in the world price for SMP. As expected, consumption of dairy products would increase following the price reduction, except for SMP. SMP consumers not only face an increase in domestic market prices but also an elimination of the SMP consumer subsidy. Most of the milk is channelled to cheese production away from SMP and butter production. Following the reform process, the EU would change its status of a net exporter of dairy products to that of a net importer, with the exception of WMP. Consumption of WMP in the EU is relatively little with very low demand elasticities. It should also be noted that the net imports figure for SMP reported in Table 3.1.1 includes imports of non-fat solids.19 The results of the alternative scenario indicate that under the assumption of imperfect price transmission to consumers following dairy policy reform, the increase in consumption in the EU is necessarily smaller (by about half).20 It follows that this smaller increase in consumption reduces import demand for dairy products and world dairy prices increase by less when compared to the “full transmission” Aglink multilateral scenario case. Correspondingly, domestic producer prices (directly linked to world prices) are also lower and, as a result, milk and dairy products production declines.

Results for Canada Table 3.1.2. in the Annex presents the results for Canada. The direction of the results is similar to those presented in Table 3.1.1 for the EU. In comparing the results it shows DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

113 that milk and dairy producer prices would see much larger cuts in Canada than in the EU. However, again it must be noted that the cuts are relative to the baseline levels of 2008. These had already shown a 17% reduction in the implicit milk support price in the EU, while in Canada’s baseline the industrial milk target price was projected to continue to increase. So the price cuts following reform are from a relatively much lower level in the EU than in Canada at the end of the baseline. While price cuts are much more severe in Canada, milk production under the ALL multilateral scenario decreases only slightly. This reflects a lower supply elasticity and a higher quota rent assumption in Canada when compared to the EU. The sensitivity to the supply response assumptions is discussed later. The results for Canada also indicate that, similar to the EU, the largest drop occurs in butter prices. Consumption of milk and dairy products increases in response to substantially lower prices, with the exception of SMP. Consumption of SMP falls because the amount of SMP used for feed under the supply management structure is eliminated in all scenarios. A substantial quantity of milk is expected to be channelled to the production of other dairy products (mainly ice-cream). Given the fact that the price of sugar is much lower in Canada than in the United States (due to the US sugar support policy), Canada should enjoy a significant competitive advantage in the ice-cream market and most of the exports of other dairy products would be destined to the United States. Nevertheless, it should be noted that in order to properly balance Canadian milk production with that of dairy products, an increase in raw milk imports from the United States occurs in most of the scenarios (see net imports of milk in Table 3.1.2). The results for the imperfect price transmission scenario (Retail) follow the same direction as discussed in the case of the European Union. That is, consumption in Canada increases only by about half of the increase observed under the All scenario. Producer prices fall for milk as well as for all dairy products relative to the perfect price transmission (All) scenario case. Milk production falls by 2.6% as compared to a 0.8% decline in the All scenario.

Results for the United States Table 3.1.3 in the Annex presents scenario results for the United States. The results are very similar to those for the EU. The reductions in milk and dairy product prices (slight increase in the case of SMP) are almost comparable in both countries. Nevertheless, the decline in milk production in the United States is smaller when compared to the European Union. In the unilateral scenario, production is reduced by 11.6% while in the Aglink multilateral scenario the reduction amounts to 4.6% as compared to the baseline level. Butter prices in the unilateral scenario are estimated to fall by 39% which compares to a 27% fall under the Aglink multilateral scenario. On the other hand while SMP prices fall by 12% in the unilateral scenario, the price in fact increases by 4% in the Aglink multilateral scenario. These results show that there is a large difference in internal price changes for butter and SMP. As in the case of the EU, this stems from the butter/SMP price policy tilt which is much in favour of butter. Thus, as the butter price falls, production of butter is substantially cut which has a significant negative impact on the production of SMP. As the production of dairy products falls and consumption rises the United States becomes and net importer of dairy products with the exception of whey powder. Whey powder is produced jointly with cheese. As US cheese production falls in response to a large decline in prices, the supply DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

114 of whey powder drops as well. This triggers an increase in whey powder prices with some reduction in consumption. It should also be noted that the baseline fluid milk premium in this scenario was reduced by 50% (see note 10).

Results for Japan Although Japan reformed its dairy policy in 2001, its milk sector remains the most protected of all OECD countries, with the highest border measures (Figure 3.7). The majority of support to milk producers is classified as market price support so it could be expected that the elimination of market price support measures would have a substantial impact on the dairy sector in Japan. Table 3.1.4 illustrates that the elimination of market price support in Japan has generated a very large decline in milk and dairy product prices. Milk prices in the Aglink multilateral scenario fall by about 55% while butter and SMP prices fall by 74% and 59% respectively. Note that the decline in cheese prices is relatively small as cheese has not been a so-called designated supported product which is the case for butter and SMP. Nevertheless, the production of dairy products in Japan would not be sustainable at such low prices, although a minimum level of manufacturing milk processing was assumed even after the elimination of trade barriers. As a result of the large drop in production, net imports of dairy products and in particular butter and SMP would increase significantly. During the model simulations, milk production declined rapidly, as milk price fell, while fresh dairy product consumption was increasing strongly. This eventually created a situation where fresh dairy product consumption became larger than milk production.21 Two options were available to solve this problem. First, it could be assumed that reconstituted fluid milk would be produced from imports of WMP and as a result fluid milk prices in Japan would continue to be strongly determined by world prices of dairy products, and in particular WMP. The second option was to assume that Japanese consumers would, because of their strong purchasing power, refuse to consume reconstituted fluid milk. Under such conditions, the fluid milk price would be calculated through an internal market clearing price mechanism, as transport costs would act as a natural trade barrier. The latter option was used and the fluid milk premium charged on top of the manufacturing milk price has increased in the scenarios.22 Thus, although milk production in Japan would decline by about 20% it would survive and milk producers would continue to supply the Japanese fluid milk market.

Results for Mexico Support to milk producers in Mexico consists almost entirely of market price support. Table 3.1.5 shows that, as expected, the elimination of market price support measures generated large decreases in milk and dairy product prices. The milk producer price would decline by about 50%.23 In response to a significant drop in prices, milk production declined by about 24% and production of all dairy products was radically reduced. The large drop in prices on the other hand stimulated consumption of milk and dairy products. The consumption of milk powders has increased but the increase is relatively small when compared to that for other dairy products. This can be explained by the fact that Mexico currently operates a programme (LICONSA) which distributes reconstituted milk from milk powders to the poor. Even though this programme acts as a consumer subsidy it is in fact detrimental to milk producers as imports of milk powder under the programme and consequently reconstituted milk to some extent competes with domestic milk production. LICONSA DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

115 was not eliminated in the scenario since the objective of the programme is to support poor people and not dairy farmers. The budget of the LICONSA programme is assumed to remain at baseline levels. This, coupled with the increase in the weighted average world price of milk powders under these scenarios automatically reduces imports of milk powders and subsequent milk powder consumption through the LICONSA programme. Thus, although the drop in domestic prices stimulates domestic consumption, substantially less is distributed and consumed under the LICONSA programme as compared to the baseline. Nevertheless, despite the fact that imports under LICONSA are reduced, total imports of milk powder — and of other dairy products — increase following the elimination of border measures.

Impact on countries that operate without MPS policy. In general the results of the scenarios for countries that currently operate without MPS policies indicate that policy reform in a single reformed country brings them only a modest increase in producer prices and production and a decrease in consumption. As other countries join the dairy policy reform process, these impacts become substantially bigger. The results for Australia, New Zealand, Argentina and Brazil are presented in tables 3.1.6 – 3.1.9. As these countries already operate at world price levels, the results for each of them are to some extent similar. In the multilateral (ALL) policy reform scenario the milk producer prices would increase on average by about 25% in Australia, New Zealand and Argentina. Differences in price changes stem from differences in domestic markets and in the price trends projected in the baseline. The impact on production is, to some extent comparable, but varies according to differences in supply response in individual countries. The results in the tables confirm that Aglink multilateral (ALL) reform brings the highest increase in producer prices and production in countries that operate without MPS policy. In these countries, the estimated benefits to producers brought about by the reform process in Europe and NAFTA illustrates that the reform process in Europe (most notably in the European Union alone) brings substantially higher increases to milk producer prices in these countries than reform in NAFTA countries alone. This result reflects the dominant position of the European Union as the world’s largest dairy producer and exporter. While producers benefit, consumers in these countries are affected negatively by the reform process. The results illustrate the reductions in consumption in the various countries in response to the rise in prices. The figures also show much higher impacts of the reform on consumers in Argentina and Brazil when compared to those of New Zealand and Australia. As a result of a significant increase in world dairy prices, the consumption of these products decreases in Argentina and Brazil by nearly 20%. This is due to the fact that consumers in these countries are much more sensitive (have higher demand elasticities) to price changes than those in Oceania and hence would tend to loose more from global dairy trade liberalisation . In general, as milk and dairy prices increase in these countries, production increases and consumption decreases. As a result, exports increase significantly for all dairy products and most notably for cheese. Part of the reason for such an increase in cheese exports (particularly in Argentina and Brazil) is the fact that cheese consumption is relatively sensitive to price and represents the largest share of all dairy products DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

116 consumed. Thus, as prices rise, there is a relatively large drop in cheese consumption (in absolute terms) which releases a substantial amount of cheese for exports. Table 3.1.10 presents the results for all other countries not specifically accounted for in the analysis; in the Aglink model called ‘rest of world’ (ROW). The dairy sector there reacts directly to world prices and no policies are assumed in place. This is indeed a limitation as many countries implicitly represented in ROW do apply some type of policies, including market price support policies in some of them, although typically at lower levels of support when compared to majority of OECD countries. The table illustrates that milk production following the Aglink multilateral (ALL) scenario would increase, albeit not excessively (about 3%). Dairy product production, in particular SMP, increases more significantly. It should be noted that in the baseline a relatively small volume of SMP is produced (compared to the level of SMP production relative to butter production in Table 3.1.10 ) as the majority of skim milk is used in its fluid form. In the scenarios, following the increase in production of butter and (fluid) skim milk, and the reduction of (fluid) skim milk consumption, large quantities of skim milk become available for production of skim milk powder. It is also worth noting that dairy product production in the ROW represents a relatively small proportion of milk production. Thus, the increase in milk production, coupled with a substantial reduction in consumption is enough to change the ROW position from a net importer to a net exporter for all dairy products with the exception of WMP, which remains crucial in many countries for milk reconstitution.

Impacts on world markets Turning attention to the impacts of the reform on world dairy markets and on other agricultural commodities, Table 3.1.11 shows the results for world commodity prices, world milk and dairy product production, and world dairy trade. As expected, world prices for all dairy products increase significantly in all scenarios. In the multilateral (ALL) scenario the world butter price would increase by 57%, that for cheese by 35%, that for WMP by 17% and the SMP price by 21.5%. It should be noted, however, that these results are to a large extent influenced by model assumptions and limitations particularly the fact that the rest of the world module (accounting for one-third of world milk production) does not adequately represent the various policies in countries not specifically modelled in Aglink. The figures for other commodities indicate that as a result of dairy policy reform, world wheat and coarse grain prices would be slightly lower than under the baseline. This reflects the fact that the countries implementing dairy reform are often characterised by milk production systems which use more grain in feed ratios as compared to other countries in Aglink. The reduction of milk production in the reformed countries thus leads to lower demand for feed grains. World prices for oils would increase as a result of the demand substitution effect generated by higher butter prices. Higher oil prices lead to increased crush and larger meal production. But as meal demand weakens, with a decline in meal used for feed, world oilseed meal prices fall. The US beef price initially falls relative to baseline levels as milk production in the United States declines and more cows are culled. However, at the end of the baseline period the US beef price increases slightly as there is less milk production, a smaller dairy cow herd and thus proportionally less culled cows. By contrast, in the EU, the beef price declines at the end of the baseline period as more beef is produced. Given the fact that in DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

117 the EU beef and dairy cattle compete for the same land, the decline in EU milk production leads to an increase in the amount of suckler cows, which produce beef more efficiently than dairy cows. As a result, beef supply increases despite the decline in the number of milk cows. An interesting question is whether or not world milk production would increase under the dairy policy reform scenario. The answer is not obvious because both producers and consumers are affected by the reforms. Milk production in countries subject to the reform would fall but at the same time consumption would increase. On the other hand, production in countries without market price support would increase while consumption would fall. The impact on world prices and overall production results from a complex balancing of the dairy market. The results presented in Table 3.1.11 indicate that under the Aglink multilateral (ALL) scenario world milk production would be only slightly lower than in the baseline. Similarly, overall world production of dairy products would not see much change. Nevertheless, WMP production would decrease by about 2.7% while cheese production would increase by 2.1% (Figure 3.8). The different results for these two products are largely driven by changes in consumption following the reforms. WMP has been imported primarily by the non-member economies where it is used for milk reconstitution. In fact, 89% of global WMP consumption is realised in non-member economies. It follows that substantially higher world market prices for WMP, resulting from the reform, reduces demand and correspondingly imports into these areas. The table indicates that world WMP trade would decrease by about 3%. On the other hand, in the case of cheese, almost 80% of world cheese consumption comes from OECD countries. Thus, as a result of the dairy reform, the considerable reduction of internal prices for cheese in the reformed OECD countries would increase cheese consumption and trade in cheeses. The table indicates that trade in cheeses would increase by 25%. The results in the Annex Table 3.1.11 indicate that all world dairy product prices were estimated to increase following the reform, with the most striking increase occurring for butter. This raises an interesting point: market price support artificially depresses world market prices but also creates considerable relative distortions within the dairy market. That is, the policy decision in heavily protected countries to tilt butter/SMP support prices in favour of butter, favours the production of fat for which demand is declining.24 The excess fat not consumed domestically is then “pushed” onto the world market through the use of export subsidies, with the consequence of greatly depressing world butter prices. After the removal of market price support policies the relative fat/non-fat solids balance on dairy markets is corrected. In summary, the results of the Aglink dairy trade liberalisation scenarios presented in Annex 3.1 indicate two broad directions of adjustment. For countries currently applying market price support, the unilateral policy reform has a strong negative impact on domestic prices and production, positive for consumption. But as other countries join the dairy policy reform process the impact of the MPS elimination diminishes substantially. Figure 3.9 illustrates graphically the reduction in adjustment following the reforms in the case of Canada. The smaller the country (in the milk production sense), the larger the benefits it draws from other (larger) countries joining the dairy reform process. On the other hand, for countries that currently operate without MPS policy the opposite is typically true. That is, a policy reform in a single reformed country brings only a modest increase in producer prices and production DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

118 (decrease in consumption) in countries with little or no market price support. But as other countries join the dairy policy reform process these impacts become substantially more profound. Figure 3.8. Impact of trade liberalisation on world dairy product output

3 2 1 % 0 -1 -2 -3 Milk

FDP Butter Cheese SMP

WMP

Source: OECD Secretariat.

Figure 3.9. Comparison of milk price and quantity adjustments in individual scenarios: the case of Canada

0 -10 -20 % -30 -40 -50 Canada NAFTA Atlantic Milk Price

All

Retail

Milk Quantity

Source: OECD Secretariat.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

119

Sensitivity analysis of the quota rent and supply elasticity assumptions in Aglink As was noted above, the specification of milk supply functions is of crucial importance for modelling dairy policy reform. In this respect the assumptions made concerning quota rent and supply elasticities are of crucial importance when quota elimination is part of policy reform. The initial assumptions used in the Aglink scenarios are based on quota rent and supply elasticity values in the PEM model, and a sensitivity analysis was conducted on these assumptions. More specifically, the supply elasticities were initially increased by 10% and then decreased by 10% for the milk supply schedules of the EU and Canada respectively. In the case of quota rent assumptions, the quota rent was increased and decreased by 3%. Thus, for example for the EU (base quota rent of 20%), the sensitivity analysis was carried out with 17% and 23% quota rent (higher quota rent implies lower marginal cost of production at the quota level and thus more efficient milk production). The results of these sensitivity tests are presented in Annex 3.2 in Tables 3.2.1 to 3.2.3 for the elasticity assumptions and in Tables 3.2.4 to 3.2.6 for the quota rent assumptions. The tables show results in percent changes from baseline levels for prices, production and consumption of milk and dairy products. The trade figures are shown in thousand tons. The column following the baseline levels illustrates the results of the Aglink multilateral dairy policy reform (labelled as base scenario) and results in bold repeat the ones presented in Annex 3.1 under the heading ALL. The rest of the columns illustrate the impact of different assumptions for the EU and Canadian milk supply schedules respectively. The results in Table 3.2.1 show that the assumption of a 10% higher EU supply elasticity reduces milk production in the EU by an additional 0.8 percentage points. On the other hand milk production increases under the assumption of a lower elasticity when compared to the base scenario. Changing the supply elasticity in Canada has virtually no impact on the markets of the European Union, whereas the impact on the domestic market (presented in Table 3.2.2) is, as expected, more profound. Again, a higher elasticity induces lower production and vice-versa when compared to the base scenario. These results for the European Union and Canada stem from the fact that domestic prices following the reform fall below the shadow price of quota in both countries.25 In the case where prices would stay above the shadow price of quota, the scenario with a higher elasticity would indeed cause higher production response compared to the status quo. Table 3.2.3 illustrates the impact of different supply elasticity assumptions on world dairy prices. A higher EU supply elasticity, which, as noted above, reduces EU domestic milk and dairy product production, causes a marginal increase in world dairy prices when compared to the base scenario. The opposite is true for a lower EU elasticity assumption. The impacts on world dairy prices of different assumptions concerning the milk supply elasticity for Canada are negligible. The sensitivity analysis for the EU quota rent assumptions presented in Table 3.2.4 indicates that with a higher quota rent (quota rent of 23%), production in the EU would decrease by 5.9% following the multilateral dairy policy reform. This compares to a 7.3% reduction in the base scenario and a 8.6% drop in the scenario with a lower quota rent assumption (quota rent of 17%). It is also evident from Table 3.2.4 that the higher quota rent assumption in the European Union would result in deeper cuts in milk producer prices (í PRUHSURIRXQGFXWVLQGDLU\SURGXFWSULFHVDQGFRUUHVSRQGLQJO\ODUJHU increases in dairy product consumption. The different assumptions on the quota rent for Canada have only a small impact on EU markets. DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

120 Similarly, a change in the quota rent assumptions for Canada has a considerable impact on Canadian milk production. Table 3.2.5 illustrates that in the case of a higher quota rent assumption in Canada (quota rent of 26%) milk production there would increase by 1% following the multilateral dairy policy reform. In the case of the lower quota rent assumption (quota rent of 20%), milk production would decrease by 2.6% as compared to the baseline. The impact of different assumptions concerning the Canadian supply schedule has only a limited impact on internal Canadian prices as Canada plays a relatively minor role on world dairy markets. An interesting result comes with respect to cross-border trade of milk with the United States. While under the original supply function assumptions some milk would be imported from the United States to Canada, the scenario with a higher quota rent assumption (lower marginal cost of production) suggests that Canada would in fact export some milk to the United States. Table 3.2.6 illustrates the impact of sensitivity analysis of quota rent assumptions on world dairy prices. Given the size of the EU market, changes in the quota rent assumptions, as expected, impacts on world prices. Under the quota rent assumption of 23%, milk production in the EU is reduced less, which causes imports to decline and world prices for dairy products to decrease by about 1 to 2 percentage points relative to the base scenario. In the case of the quota rent assumption of 17%, the large drop in EU milk production results in an increase in EU imports, which strengthen world prices for dairy products by about 1 to 2 percentage points relative to the outcome of the scenario with the original assumptions. The impacts on world dairy prices of different assumptions concerning the supply schedule for Canada are relatively minor. The sensitivity analysis has shown that the supply schedule assumptions can have a considerable impact on estimated milk production. The quota rent which determines the vertical position of the supply curve is, in particular, critical: whether domestic prices following reform are above or below the shadow price of quota determines whether the country or region produces more or less milk when compared to the quota level observed prior to the reform. It should be noted that the range of quota rent in the sensitivity analysis exercise is to, by no means, imply a plausible range of quota rents for each country. The simulations with different quota rents are merely to test the sensitivity and robustness of results to a different set of assumptions. As discussed in the analysis of milk quota systems, a wide range of assumptions regarding quota exists in the literature. As expected studies that have used higher quota rent assumptions arrived at lower reduction or an increase in milk production in quota countries following the trade liberalisation and a quota system elimination. For example Larivière and Meilke (1998) used an assumption of 35% quota rent in Canada and estimated milk supply in Canada to increase by 6.9% under a trade liberalisation scenario. For a comparison, the estimated results of an alternative scenario with Aglink, using 10 percentage points higher quota rent in Canada (33% quota rent), indicates that milk production in Canada would increase by 5.6% under the ALL scenario. Similarly, in the case of increasing the EU quota rent by 10 percentage points (30% quota rent) milk production would decline in the EU by only 2.5% as compared to 7.3% reduction with 20% quota rent.

PEM results of dairy trade liberalisation scenarios- Evaluating the welfare impacts The Policy Evaluation Model (PEM) is a partial equilibrium static model including 5 major commodity categories and covering six countries (EU, USA, Mexico, Canada, DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

121 Switzerland and Japan) plus a rest-of world component (see Annex A for the description of the PEM model). It is calibrated to a 2002 base year using primarily data from the OECD PSE database.26 Elasticity parameters are selected to give the model a mediumterm (5-7 years) adjustment horizon. The dairy sector is represented in terms of raw milk equivalents. A single supply of raw milk is demanded by separate domestic markets for milk for fluid and industrial uses. Industrial milk (dairy products are expressed in rawmilk equivalent terms) is a tradable commodity in the model. The model is intended to represent price effects within a competitive market framework. As such, the welfare results presented below reflect the impact of changes in markets occurring through the price mechanism, and do not make any claims regarding welfare changes in their most general sense (including for example, externalities and other non-market effects, general equilibrium effects, effects of market failure, effects of structural change in production technology, and other social costs of adjustment). Some of the other mechanisms through which policy reforms may affect production and trade are considered in the “Decoupling: a conceptual overview” (OECD 2001). Results of the PEM model reflect the 2002 base year; thus are not represented as net of programme changes after that date. Thus, reported changes in production, prices, and welfare do not reflect programme changes made after the base year but within the adjustment period of the model, such as the recent 2003 CAP neither reform nor the full implementation of Agenda 2000 in the EU, for example. Annex A contains a fuller description of the PEM model. As is the case with all equilibrium models, scenarios leading to large percentage changes in prices are subject to the caveat that the elasticity parameters in the model are assumed accurate only for prices “close” to the baseline equilibrium. In order to address this, all scenarios are combined with a sensitivity analysis that varies all parameters within their plausible ranges as identified in the reports by Abler (2001) and Salhofer (2001). In addition, unit marginal cost is also varied in a range between one-half and double its base value. A Monte-Carlo approach is used where all parameter values (plus unit marginal cost) are drawn from uniform distributions, and the scenario run using this new parameter set. This is repeated for 500 draws and the extreme high and low values for all results are identified. These maximum and minimum results define the range of results that alternative but plausible parameter choices would yield. In many cases, this sensitivity analysis produces no variation in the results in each scenario. This is the case where the scenario dictates the result; in all of these scenarios, support is eliminated, so there is no variation in the change in taxpayer welfare or quota rent. Taxpayer costs and quota rent will always change from the base value to zero. Only when sensitivity analysis identifies some variation are the results shown. This is to enhance readability of the tables. Support to milk production in the six regions represented in the PEM as measured by the %PSE ranges from 45% in Mexico to 74% in Japan. The monetary value of this support ranges from USD 1.1 billion in Mexico to USD 18.4 billion in the European Union and totals some USD 36 billion for the entire PEM region. As is typical of support provided to the dairy sector, most of this is in the form of market price support. They represent some of the most protected agricultural markets in the OECD and the potential gains from reform of their dairy sectors are correspondingly large. In this section the results of simulation experiments using the PEM to investigate the potential of dairy reform to improve welfare will be presented.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

122 In the first set of experiments, dairy policies were eliminated unilaterally in each PEM region. That is, support was reduced to zero in one region only, leaving all others’ policies unchanged (Table 3.1). In all scenarios the elimination of price support is sufficient to make quota no longer a binding constraint in the model, thereby “eliminating” the quota restriction. Since quotas are endogenous in the model, no special modifications are required in moving from a situation where the quota is binding to one where this is not the case. In addition to market price support, support in the form of consumption subsidies and output support were also eliminated. Policies are represented in the PEM according to the categorisation in the PSE and are reflected as price wedges in the market where they have their initial incidence. Since each region represents a separate experiment, the results shown in the table are not additive and do not in any way indicate global net results. In each experiment, production is reduced, and the country becomes a net importer of dairy products. Table 3.1 presents the commodity market impacts as the outcome of the experiment in level terms, and below that in terms of percentage change. Table 3.1. Unilateral liberalisation experiments

Canada Commodity Market Impacts Domestic Production (m tonnes) Net Exports (m tonnes) World (border) Price (USD/tonne)

European Union

6.5

96.2

5.83/7.32

86.61/106.47

Japan Mexico Switzerland ~ result of experiment in levels ~ 5.6 4.8 1.6 5.58/5.63

4.59/4.89

1.4/1.91

United States 55.0 54.71/55.42

-5.7

-29.2

-15.1

-8.8

-2.7

-25.0

-6.41/-4.95

-37.84/-19.84

-15.1/-15.09

-8.94/-8.64

-2.9/-2.39

-25.31/-24.68

156.3

188.2

160.0

159.6

161.6

193.5

155.9/156.55

182.48/195.23

162.09/162.22

158.99/159.22

161.58/161.99

185.46/186.19

-41

-36

-32

-41

-72

-27

-42.6/-39.76

-38.18/-33.86

-32.48/-30.81

-41.57/-41.48

-71.99/-71.92

-23.39/-23.05

~ % change ~ Producer Price Domestic Production Net Exports World Price Economic Impacts Taxpayers of which export subsidies Consumers

-21

-21

-33

-52

-51

-29

-28.95/-10.78

-28.75/-12.41

-33.28/-32.79

-53.47/-50.4

-57.34/-41.73

-29.1/-28.18

1429

-517

248

306

-2247

3346

1220.9/1609.7

-641.52/-383.89

247.75/247.9

299.78/313.63

-2418.2/-2010.21

3300.67/3387.13

3

18

6

3

1

21

2.3/2.73

13.72/21.66

5.8/5.89

3.02/3.17

1.3/1.56

20.9/21.38

0 0 1504

2067 941 11590

234 0 6300

377 52 1431

3286 0 2014

~ USD millions ~ 16 0 1427

1499.4/1508.41 10825.86/12443.57 6269.83/6331.22 1424.76/1427.96 1430.41/1432.47

Farm Households of which dairy capital of which farm land of which dairy livestock of which quota rents

-999

-10842

-1134.29/-836.62

-12920/-8617.26

Net impact Net impact as % of initial value of sector (valued at world price)

-545

-824

1985.33/2044.54

-1764

-1269.9/-1112.69 -550.66/-538.01 -1003.55/-643.58 -1813.76/-1716.32

-193

-2728

-664

-265

-291

-1011

-287.67/-92.99

-3861.45/-1531.19

-717.99/-606.63

-269.47/-260.5

-353.83/-225.62

-1065.54/-963.64

-107.2

-739.3

-223.7

-168.0

-105.6

-384.2

-133.62/-66.48

-922.1/-498.35

-231.33/-216.18

-169.23/-165.68

-105.87/-104.11

-392.66/-375.28

-72.6

-251.1

-304.2

-112.4

-123.4

-368.8

-107.65/-34.68

-373.17/-133.24

-333.68/-276.26

-115.16/-109.43

-149.89/-96.54

-383.5/-356.21

-625.7

-7123.1

0.0

0.0

-304.7

0.0

0/0

0/0

-567.59/-40.33

0/0

-937.11/-313.33 -10708.18/-3537.62

Input Suppliers

-1192

-44.9

-1127.9

-85.2

-107.6

-126.8

-1639.0

-69.76/-20.24

-1713.32/-577.03

-93.12/-76.8

-111.06/-104.3

-162.64/-90.37

-1674.94/-1599.09

460.0

1687.5

5256.5

789.2

857.0

1896.8

353.8/593.61

1001.35/2613.11

5176.14/5336.09

785.09/794.03

709.6/1010.93

1880.01/1918.64

36.8

8.7

409.9

51.9

163.6

16.0

28.31/47.5

5.13/13.4

403.67/416.14

51.61/52.2

135.43/192.94

15.88/16.21

Source: OECD PEM model.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

123 The presence of a binding quota makes the supply function more inelastic on average. This is because the supply function contains a vertical component where the quota is binding and supply does not change in response to price changes. The larger this vertical component is (larger quota rent), the more inelastic the total supply response will be. For example, the percentage price decline in Canada is nearly double that of the United States, but the supply response is the same in percentage terms. As might be expected, the most dramatic market results are seen in countries with the highest initial rates of support. Switzerland, Canada, and Mexico see the highest reduction in domestic milk prices, while production is most strongly impacted in Switzerland, Mexico, and the United States. Scale of production is much more important in determining the effects of reform on world prices, with the United States and the European Union having a more significant impact on world prices, followed by Japan (Table 3.2). Table 3.2. Ranking of impacts of unilateral liberalisation experiments Rank of impact of experiment (in % change) %PSE Ranking (2002) Highest

Lowest

1 2 3 4 5 6

Japan Switzerland Canada United States EU Mexico

Quantity Produced

Producer Price Switzerland Mexico Canada EU Japan United States

Mexico Switzerland Japan United States EU Canada

World Price United States EU Japan Canada Mexico Switzerland

Net Impact (% of value) Japan Switzerland Mexico Canada United States EU

Source: OECD PEM model.

In these results, the country with the highest initial percentage PSE, Japan, sees a relatively moderate impact in production and prices, especially when compared with countries such as Mexico, with a relatively low %PSE for milk. The relatively moderate production change seen in Japan is explained by the assumption in the model that fluid milk is non-tradable and so fluid demand must be supplied by domestic producers. For this reason, domestic production must always at least equal domestic fluid demand. Japan is the only country for which this becomes a binding constraint in the model,27 with all domestic production dedicated to fluid use, and all manufacturing milk demand supplied by imports. This requirement for domestic fluid supply-demand balance drives the results for that country. It is also the reason that the ranges of possible results indicated by sensitivity analysis are quite narrow; the fluid demand function is the major determinant of supply, making the results insensitive to changes in other parameters. In an autarkic market for fluid milk, world price does not play a role in price formation (the domestic fluid market is essentially isolated from the rest of the world). Domestic fluid milk is still able to command a premium of 170% over the world price because of the assumed natural barrier to fluid trade. This premium moderates the effect of liberalisation of dairy policy on producer price, production, and welfare.28 Japan is seen as having the greatest potential benefits from liberalisation. The net improvement in welfare calculated in the experiment is just over four times the initial value of the entire dairy sector, measured as domestic production before liberalisation DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

124 multiplied by world price.29 This results from the disproportionate gain in consumer welfare. Switzerland has the next-greatest potential benefit, at over one and a half times the initial value of the sector, followed by Mexico at 52% of the value of the sector. Because of the size of the dairy sector in the European Union, the potential benefits of reform amount to about 9% of the value of the sector. In this set of experiments, Mexico is seen to consistently rank higher in terms of impact than its level of support as measured as %PSE or its size would indicate. Nearly all support to dairy producers in Mexico is in the form of market price support, and there are no quota restrictions in the country. Thus, elimination of support has a relatively greater impact on producer price, and there are no quantitative restrictions moderating the production response to this price change. Conversely, Japan, which benefits from the natural barrier to trade related to fresh milk, sees the highest net benefit to reform with the second-least impact on producer price relative to other countries. The next experiment carried out using the PEM model was a liberalisation of the dairy sector in all regions at the same time (Table 3.3). The main difference in this experiment when compared with the previous one-country-at-a-time experiments is that the effect on the world price is more dramatic. All six regions reduce production and become net importers, for a total change in demand on the world market of nearly 51 million tonnes. This causes world prices to rise by 42%. By comparison, the most significant change in world price in the unilateral liberalisation experiments was 21%, related to liberalisation in the United States. A key determinant of the world price post-liberalisation is the ability of the rest of the world to supply additional milk to the market. The “Rest-of-World” is represented in the PEM by a ROW supply function and a ROW demand function (whose elasticises are estimated using the Aglink model). The 42% increase in world price is required to draw an increase of 34 million tonnes in net exports from other countries. Structural change or capacity limitations on exporters could change this result. For example, if Australia and New Zealand were unable to significantly increase output, the world price would be higher and net ROW exports lower. On the other hand, if the higher world price resulted in the emergence of new major milk exporters (perhaps Brazil and Argentina), the world price could be lower. Caution is warranted in interpreting the results of a shock of this magnitude. As was seen in the Aglink results, multilateral reform moderates the changes in the domestic milk markets of protected countries when compared with a unilateral reduction in support, and results in reduced gains for consumers as well as reduced losses for producers.30 and lower net welfare benefits in each country. Moreover, the multilateral reform scenario does not include liberalisation of policies in the countries contained in the ROW component. More widespread liberalisation would have the impact of moderating the impact on the represented sectors even further. Once again, this effect is most pronounced in Mexico. Total welfare in all regions increases by some USD 7 billion, an amount equal to about one fifth of the market value of dairy production in the six regions, valued at world prices before liberalisation. Total welfare in the Rest of World will also increase because of the change in world price, as the region is a net exporter.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

125 Table 3.3. Simultaneous liberalisation experiment Canada

European Union

Japan

8.1

116.2

5.6

7.9

2.0

68.0

207.9

7.07/9.27

105.23/128.15

5.59/5.63

7.46/8.38

1.72/2.33

62.63/73.18

202.4/214.19

Commodity Market Impacts Domestic Production (m tonnes) Net Exports (m tonnes)

Mexico

Switzerland

United States

Total 6 regions

~ Simulation result, levels ~

-1.9

-4.3

-13.0

-4.7

-2.0

-8.0

-33.9

-3.11/-0.53

-14.49/6.66

-13.26/-12.78

-5.25/-4.12

-2.27/-1.61

-14.34/-2.07

-37.11/-30.16

218.2

World Price (USD/tonne)

209.81/225.67

~ Change, % ~ Domestic Production Net Exports

-1.5

-4.4

-33.0

-19.4

-39.0

-11.8

-9.0

-13.81/13.04

-13.43/5.43

-33.18/-32.74

-24.34/-14.99

-47.49/-29.05

-18.83/-5.15

-11.39/-6.23

408.4

-162.1

199.7

115.3

-1 660.2

1 006.7

40.12/730.04

-307.39/-4.63

194.64/205.66

90.86/142.89

-1913.51/-1391.25

185.69/1875.31

41.7

World Price

36.27/46.47

~ Change, USD millions ~

Economic Impacts Taxpayers Dairy Consumers Crops Consumers Farm Households

0.0

2 066.8

233.9

15.5

377.0

3 286.2

5 979.4

782.9

7 123.1

5 263.6

635.9

1 143.3

-532.0

14 416.7

705.85/871.08

6221.95/8142.43

1111.55/1179.44

-1103.28/115.05

12629.36/16442.99

52

433

118

85

0

480

1 167

-659.1

-8 065.9

-1 197.4

-315.6

-773.1

5131.52/5405.97 541.82/742.38

-760.44/-541.31 -9637.08/-6303.36 -1274.22/-1118.62-371.43/-261.72 -941.24/-598.74

of which dairy capital of which dairy pasture land of which dairy livestock of which quota rents

Net Impact Net impact as % of initial value of sector (valued at world price)

-11 893.4 -14072/-9436.47

-18.5

-685.9

-667.1

-137.4

-256.5

-482.0

-2 247.4

-157.53/-0.53

-2058.42/6.66

-720.15/-12.78

-166.22/-4.12

-333.07/-1.61

-738.42/-2.07

-3476.95/-30.16

-224.7

-119.6

-7.8

-193.4

-85.04/114.55

-544.53/275.15

-231.8/-217.67 -138.81/-100.78

-102.9

-218.1

-866.6

-105.46/-96.94

-299.06/-124.14

-1170.44/-454.07

-7.1

-63.6

-305.6

-58.5

-108.9

-182.2

-726.0

-59.78/52.21

-197.2/70.45

-334.48/-278.36

-70.17/-47.39

-139.25/-77.31

-268.93/-100.45

-813.53/-620.02

-625.7

-7 123.1

0.0

0.0

-304.7

0.0

-8 053.5

0/0

0/0

-567.59/-40.33

0/0

-11832.74/-4064.35

-937.11/-313.33 -10708.18/-3537.62

Dairy Input Suppliers

-882.3 -1296.32/-477.82

-31.5

-374.9

-85.7

-64.2

-102.0

-873.3

-1 531.6

-57.64/-2.38

-940.66/170.33

-93.41/-77.17

-74.95/-54.29

-139.44/-66.56

-1126.63/-627.05

-1872.69/-1203.76

4 214.4

271.7

92.2

749.1

53.1/180.61

624.89/1084.97

4042.87/4422.39 236.01/315.7

645.3

998.6

6 971.2

501.56/803.43

955.35/1088.52

6667.48/7380.9

7.38

3.84

328.66

17.86

123.15

8.44

3.27

4.25/14.45

3.2/5.56

315.29/344.89

15.52/20.75

95.72/153.33

8.07/9.2

3.13/3.46

Source: OECD PEM model.

The results show the effectiveness of the quota in the European Union at restricting supply; despite a reduction in producer price of 22%, production has reduced only by 4%. Canada, which also maintains quota restrictions on production, sees a drop in production of only 2%. Switzerland also has restrictions on production, but the decline in producer price is enough to provoke a decline in production of 40%.31 The decline in production in Japan at 33% may be considered moderate; without the natural premium (of close to 100% in this scenario) afforded to fluid milk due to its non-tradability, the price decline there would be closer to 75%. Not shown in the table, net changes in world production and consumption of dairy products are modest, with production, consumption, and trade changes in the rest of world being roughly equal in magnitude and opposite in sign to the results in the modelled regions.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

126 The results of the sensitivity analysis show a greater degree of variation for countries which make use of quota systems. Uncertainty regarding the unit quota rent drives this result. The unit quota rent determines the marginal cost, which locates the supply function. A high unit quota rent implies a lower supply function. That is, a high quota rent means the cost structure of production is lower, and so a greater level of production at a given price once the quota system is removed. Varying unit quota rent (marginal cost) in the sensitivity analysis effectively varies the location of the supply function. In Canada and the European Union, this means that domestic production may increase or decrease as a result of liberalisation, depending on the assumed level of unit quota rent. The price decline in Switzerland is sufficiently large, so that the change in production is unambiguously downward. The European Union, moreover, may see its net exports increase or decrease as a result of reform. This has an impact on net welfare, which may also increase or decrease as a result of reform in these countries. Alternative uses of crops compete with dairy producers, who demand grains and oilseed meal for feed. Eliminating support to dairy can bring significant benefits to these consumers as the reduction in demand for dairy feed and fodder reduces the prices for these commodities. In the European Union and the United States, where the dairy sectors are relatively large, these benefits are significant and make up around half of the net benefits of reform in those countries. In each country, with the exception of the United States, consumers are the main beneficiaries of reform through lower domestic prices for milk. In the United States, market price support is relatively less important (though still significant) as a proportion of total support. Domestic prices to consumers in the United States rise because the elimination of consumer subsidies combined with the rise in world prices is greater than the price reduction due to the elimination of domestic price support (plus the fact that 50% of the fluid milk premium is assumed to remain after liberalisation). On the other hand, US taxpayers benefit from the elimination of these subsidies. Because of its lack of quota restrictions (which tend to concentrate benefits to producers at the expense of input suppliers) and relatively large use of purchased inputs, the United States is also the only country where input suppliers face the same order of welfare change as do farm households. By contrast, in the European Union, the loss of quota rents results in farm households (and quota holders in particular) being the main losers in dairy policy reform. Box 3.1 described the potential for price transmission along the retail chain to be less than perfect. This possibility was considered in the Aglink scenarios, where a passthrough of 50% was assumed. This experiment regarding retail market power is repeated here using the EU as an illustrative example.32 In this example, domestic producer price and world price are aligned by removal of border measures and other support policies (this is the shock introduced to the model), but it is assumed that domestic retailers who purchase from domestic and foreign suppliers are able to alter their margin by an amount equal to half of any upstream price changes (50% price transmission). The scenario considered here is unilateral elimination of domestic support in the EU; the results however are quite general and apply equally to the case of multilateral liberalisation. The results are presented in Table 3.4.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

127 Table 3.4. Partial price transmission in the EU

Unilateral liberalisation with 50% retail price transmission Commodity Market Impacts

Unilateral liberalisation with 100% price transmission

Difference between two scenarios

~ Simulation result, levels ~

Domestic Production (m tonnes)

93.0

95.9

-3%

Net Exports (m tonnes)

-25.9

-29.4

-12%

World Price (USD/tonne)

184.1

189.4

-3%

~ Change, levels ~ Net Exports (m tonnes)

-37.6

-42.2

Domestic Production

-24%

World Price

15% 18% ~ Change, USD millions ~

~ Change, % ~

Economic Impacts Taxpayers

-21%

2 066.8

2 066.8

0%

Dairy Consumers

5 398.6

11 566.2

-53%

Retailer Rents

6 606.3

0.0

-

-11 193.7

-10 876.6

3%

-2 981.3

-2 742.4

9%

-811.2

-755.8

7%

Farm Households of which Dairy capital of which dairy pasture land of which dairy livestock

-274.9

-252.0

9%

-7 126.4

-7 126.4

0%

Dairy Input Suppliers

-1 163.5

-1 052.6

11%

Net Impact

1 714.5

1 703.8

1%

of which quota rents

Source: OECD PEM model.

When support policies are eliminated, partial price transmission through the retail market means that the consumer price falls less than it would otherwise. This means that fewer dairy products are imported than would have been the case, with the result that the world price is lower than it would otherwise have been. Producers, who receive the world price, respond to this by producing less. Farm households are worse off compared to the full-transmission case because of the relative decline in producer price; they lose some producer surplus that would have accrued to farm owned inputs in the alternative full-transmission case (all quota rents are lost in both cases). Consumer welfare increases by half the amount of the previous scenario, owing to the more modest price reduction in the partial price transmission scenario. By altering their margin in response to downstream price changes, retail intermediaries are able to obtain economic rents.33 These rents are defined as the amount of per-unit additional margin times quantity sold. In this case the additional unit margin is the differential between producer and consumer prices. These rents amount to USD 6.603 billion, significantly more than the change in consumer welfare of DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

128 USD 5.398 billion in this scenario.34 Since consumers and retailers essentially share the price change 50-50, why are these values not the same? One reason is the slope of the demand function; consumer surplus has a triangular component reflecting a declining marginal benefit to additional consumption, while intermediary unit rents are the same for all units sold. Further, it is not just a matter of retailers sharing in consumer rents. When intermediaries alter their margin and retarding demand, it has the effect of reducing the price received by domestic and foreign producers (via world prices). Accordingly, some of their rent is at the expense of domestic and foreign producers and input suppliers. This transfer from foreign producers notwithstanding, the deadweight losses that arise from intermediaries’ action on their margins results in a smaller increase in EU welfare compared with the full price transmission scenario (i.e. a relative decrease). While intermediaries obtain rents at the expense of consumers, domestic and foreign producers, and input suppliers, it is the domestic consumers that bear the majority of the cost of retail margin manipulation. Why? Trade movements limit domestic producer losses from the partial transmission of prices. Consider the case of a small country: if changes in trade had no impact on world price, domestic producers facing that world price would be indifferent to retailers’ ability to change their margin and the subsequent effect on consumer demand and trade. Conversely, if there were no possibilities for trade, domestic supply would need to contract by the same amount as did domestic demand, requiring a larger reduction in producer price and a larger welfare loss for producers. The EU is large enough that changes in net trade affect world prices, but this price change is small relative to the domestic consumer price change. In the scenario shown here, the impact on domestic producers of the reduction in world price amounts to an additional loss of 3% relative to the alternative liberalisation scenario. The most significant impact of partial price transmission is the transfer between two different domestic constituencies: consumers and owners of intermediary enterprises. The difference in the net production and welfare impacts is small. The effect on net trade is proportionally more significant (a 13% reduction in net imports), as the reduction in domestic consumption compared with the full-transmission scenario is manifested primarily as a reduction in imports.

Summary The empirical analysis of dairy trade liberalisation presented in this document was carried out using the Secretariat’s Aglink and PEM models in order to assess the impact of dairy policy reforms on production, consumption, trade, prices, income and welfare. The empirical results have to be viewed within the limits of the models used (the model limitations are discussed in the text and are summarised later in this section). The dairy trade liberalisation results of Aglink and PEM are not strictly comparable as these models have different structures, country, product and time coverage. The simulation experiments were tailored to each model with the specific strengths of each model in mind. The results of the experiments are complementary in that the Aglink scenarios provide insights in market and trade outcomes of liberalisation, while the PEM results enlighten in particular the income and welfare effects. It is also important to bear in mind that the simulation results have an indicative (normative) rather than definitive character. That is, the results give broad indications of the nature, possible direction and potential changes in markets, income and welfare due to liberalisation in the dairy sector, rather than definitive forecasts of the outcomes.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

129 The results of the dairy trade liberalisation scenario in which market price support is removed shows that for countries currently applying MPS support the unilateral policy reform has a strong negative impact on domestic prices and production (positive for consumption). As other countries join the dairy policy reform process the impact of the MPS elimination diminishes substantially. On the other hand, for countries that currently operate without MPS policy the results are equivalent, though obviously going in the opposite direction. That is, in general the results of the scenarios for countries that currently operate without MPS policy indicate that a policy reform in a single reformed country brings them only a modest increase in producer prices and production (decrease in consumption). As other countries join the dairy policy reform process these impacts become more profound. As expected, the removal of all market price support measures is estimated to have a substantial impact on world dairy markets. The results indicate that world prices for all dairy products would increase significantly. In the multilateral scenario where policies are reformed in countries covered in Aglink, world butter prices would increase by 57%, cheese prices by 35%, WMP prices by 17% and those for SMP by 21.5%. PEM results show that the implicit world price of milk would increase in the PEM multilateral scenario by 46%. The Aglink results also suggest that market price support not only artificially depresses world market prices but also creates considerable distortion within the dairy market milk components balance. That is, the policy decision in heavily protected countries to tilt butter/SMP support prices in favour of butter, depresses world fat (butter) prices more than those for non-fat solids. Following the removal of market price support policies the relative fat/non-fat solids balance is re-established in dairy markets. Total world production of milk and dairy products is estimated not to be altered significantly following the reform process, but a shift occurs to more efficient regions. The results indicate that world milk production would be reduced by only 0.2%. Butter production would remain virtually unchanged while WMP production would be reduced by 2.7%. SMP and cheese production would increase by 1.2% and 2.1% respectively. Following the reform, less butter and WMP would be traded. The reduction in butter trade stems primarily from large cuts in subsidised exports from countries where MPS is eliminated, while the drop in WMP trade reflects an increase in WMP prices and lower demand from NMEs where WMP is used for reconstitution. Cheese trade and consumption in countries where MPS is eliminated would increase significantly following the reform. Potential welfare gains from liberalisation estimated with the PEM model are in the order of USD 8 billion for the six countries included in the model. The beneficiaries of liberalisation are taxpayers and consumers, from whom dairy policies are designed to provide transfers to producers. Efficiency losses from these transfers mean that eliminating them provides a net social welfare gain, of the amount mentioned above. Japan stands out as a region where this gain is particularly pronounced. The welfare gains to consumers in that country are four times the amount lost to producers by eliminating transfers caused by price support. The PEM cannot pronounce on welfare changes in other regions, but it seems likely that the increase in world prices would be welfareimproving for producers in regions without border protection, and welfare-reducing for consumers in such regions. While the results of the international dairy liberalisation scenarios in this study are to some extent comparable to results of other studies, the analysis itself is subject to caveats DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

130 and shortcomings. Some of the caveats were addressed via a sensitivity analysis and through production of an alternative scenario. Other caveats, often common to large world agricultural commodity models, remain. Some of the caveats relate to model specification, some to the scenario design and others to the complexity of milk production and trade. First, while scrutinising the results it is important to keep in mind that the policy reform analysed here relates to the dairy sector only. The support in other sectors remains at baseline levels and this may have an important bearing on the final outcomes of the reform scenarios presented in this report, in particular where beef and dairy production are closely integrated, as is the case for many European countries. It follows that in the context of multilateral and multi-commodity agricultural policy reform, the polarisation of results between reformed countries and countries currently without any market price support might be different and in fact very likely reduced. In theory, it might be argued that the loss in relative profitability of the dairy sector would be smaller if all sectors are reformed compared to the case were support is only eliminated for dairy. That is, the increase in relative profitability would stem not only from a reduction in commodity prices that are competing with the dairy sector but also from cost reductions likely to be associated with the reduction in feed prices. In the case of countries currently without market price support the opposite could be expected. As the world prices of feed and other commodities rise, the dairy sector would become relatively less profitable when compared to the scenario assessing dairy reform only. Further study would be required to give more insights into the impacts of a multi-commodity reform. It should also be kept in mind that the dairy products produced, consumed and traded in the Aglink and PEM models are assumed to be homogeneous. No account has been taken of different product attributes, sanitary standards or different production practices such as environmentally friendly farming, respecting animal welfare etc., which can influence costs of production, product prices and consumer choice. Moreover, although results of the analysis show that adjustment in milk production in reformed countries is not excessive and their dairy sectors would continue to remain in business, the results focus on country or regional aggregates. At this level of aggregation it is not possible to infer the level of adjustments required at the farm level. Structural changes at the farm level might be indeed substantial. Another limitation concerns the absence of risk in the analysis. Producers in countries which are not currently exposed to world price fluctuations and are protected by a minimum price programme would probably consider the market conditions of the postreform as being more risky and prices relatively more unstable. The opposite might be true from the perspective of producers not currently protected by MPS measures, although the future stability of world market prices after trade liberalisation remains subject to debate. The impact of trade liberalisation depends heavily on the production response subsequent to the removal of milk production restrictions in countries that employ a milk quota. This is even more important as dairy markets are relatively thin and some of the countries operating a quota system are major players on the international milk and dairy product markets. Thus, effectively the results of international dairy liberalisation are influenced by the milk supply response assumptions in quota operating countries, notably those on quota rent and supply elasticities. Another caveat is the fact that the rest of the world is treated as a single block in the models. Many countries that are implicitly represented in the ROW module use market DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

131 price support and other policy measures. Given the size of the ROW, in terms of milk and dairy product production and consumption, the response of this region to trade liberalisation is crucial for the results of such a scenario. If the multilateral reform simulation would include the impact of liberalisation in a large number of developing countries which are currently represented in the single “rest of the World” module, the adjustment needed in the represented countries in this paper might be even smaller. Unfortunately, to disaggregate the developing world into smaller regions is very costly due to the lack of consistent and reliable data for these countries. Finally, given the high level of protection in dairy sector the scenario of full trade liberalisation create a shock to the model beyond what many would consider “marginal” or small.35 Price movements outside a certain range can call into question the continuing validity of the elasticity parameters in the model, with a corresponding impact on accuracy. As an example, in the case of demand functions, the issue of market saturation might become important. Moreover, the models do not allow for structural change. That is, while relative factor intensity may be varied according to relative price changes, the underlying production functions do not change. Thus, the results presented here do not represent a long-term equilibrium, but rather a medium-term perspective on reform. However, as such they do provide a useful indication of the winners and losers from reform, as well as an idea of the potential scale of its benefits.

Main conclusions The market, trade and welfare effects of international dairy trade liberalisation have been analysed empirically using the Secretariat’s models, Aglink and PEM. The empirical results have to be viewed within the limits of the models and are subject to the usual modelling caveats. Nevertheless, they do provide important insights into the potential of liberalising dairy markets. Thus, the results should be viewed as conveying something about the nature, possible direction and potential changes in markets, income and welfare due to liberalisation in the dairy sector, rather than offering definitive forecasts. The main findings and lessons that could be drawn from the analysis of international trade liberalisation presented in this chapter can be summarised as follows: x

Dairy trade liberalisation would have the potential for significant net welfare improvements, with consumers being the main beneficiaries and taxpayers also realising gains. The size of benefits to consumers, however, is subject to the degree of price transmission along the supply chain.

x

The elimination of distorting market price support policies results in only little change in total world milk production, but world dairy prices would be lifted substantially (by 17% to 54%) while supply would shift towards more efficient areas.

x

Supply adjustments in milk supply in reformed countries would be relatively not large. However, while their dairy sectors would not be affected greatly at the aggregate level, substantial structural adjustment may take place at the farm level.

x

Producers and exporters in developing countries would gain large benefits from the reform, while consumers in these countries would face a substantial increase in prices.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

132 x

The price and supply adjustment would be higher if a country reforms its dairy policy unilaterally. As more countries would join the reform process, adjustment would be smaller and would be least in the case of multilateral reform.

x

World-wide multi-commodity reform would likely result in further reductions in adjustment pressures in the dairy sector.

x

Distortions in price formation of dairy products (tilt) would be eliminated.

x

The assumption concerning the production potential in quota operating countries plays an important role in the analysis. However, sensitivity tests confirm the results, and do not lead to fundamentally different conclusions.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

133

Notes 1.

For example, in Canada and the United States milk producers receive 36% and 24% respectively of all producer support which compares to 17% and 10% in the European Union and Japan (2000-02 average, OECD 2003a)

2.

Note, that recent reform of the EU CAP will likely have an impact on PSE measures for the EU (see OECD 2004b).

3.

The milk quota allocated to the 10 new EU members represents about 16% of the total EU quota.

4.

Note that in Poland, despite the increase, the PSE level is relatively low.

5.

As measured by the Producer Nominal Protection Coefficient (NPCp), an indicator of the nominal rate of assistance to producers measuring the ratio between the average price received by producers (at the farm gate), including payments per tonne of output, and the border price (measured at the farm-gate level).

6.

There are other studies that have estimated partial trade liberalisation scenarios such as BouamraMechemache et al. (2002) or Donnellan and Westhoff (2001).

7.

The Rest of World module represents about 33% of world milk production in the Aglink specification and about 60% of world milk production in the PEM specification.

8.

In order to use the complementarities of the Aglink and PEM models, the PEM assumptions on quota rents and supply elasticities were implemented in Aglink. The modelling of milk supply functions with Aglink and PEM has been discussed in Chapter 2.

9.

The cheese price in the European Union Aglink component is for Emmental while the world price indicator is for cheddar. The difference in the price of these two cheeses has been consistently 30% and this factor was used in the price transmission equation.

10.

EU butter intervention stocks are declining very rapidly in the baseline and were left at the baseline level. SMP intervention stocks go to zero already in the baseline. In the US, the baseline butter and SMP stocks were assumed to decline faster in the trade liberalisation scenarios (see technical report OECD (2004) for more detailed description of Aglink modifications).

11.

The retention of 50% of the fluid milk premium is arbitrary and was chosen in the absence of tangible estimates of the natural fluid milk premium. The presence of natural (market driven) premium arises due to market driven factors such as consumer preferences, transportation costs and seasonality (for more detailed discussion see Chapter 1. The 50% of the baseline fluid milk premium was added on top of the manufacturing milk price (in this scenario linked to the world prices of dairy products) to approximate the level of the fluid milk price. This procedure was carried out whenever the baseline data indicated a difference between fluid and manufacturing milk.

12.

For the dairy markets baseline policy assumptions see Annex 3.3.

13.

Complete elimination of all market price support measures in the EU 15, Poland, Hungary and Russia. (Note, that since the analysis was completed Poland and Hungary joined the European Union).

14.

Another limitation concerns the absence of risk as an explanatory variable in the Aglink milk supply functions. Several scenarios were undertaken under the assumption that farmers are risk averse and that a downward shift in reformed countries supply could be expected. However, the assumption of the shift

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

134 was arbitrary and the results are not reported here. Nevertheless, as expected, the results showed that world prices would increase more under these scenarios. 15.

It should be noted that EU in the 2003-2008 baseline used in the simulations refers to EU 15. The policy assumptions for the EU are those of Agenda 2000 (see Annex 3.1)

16.

The 2003-2008 baseline (OECD 2003b) assumes a 15% cut in EU support prices for butter and SMP.

17.

Again, it should be noted that the word “multilateral” is used in the text to reflect a scenario where countries represented in Aglink reform their dairy policies rather than all countries in general (see the definition of the Aglink scenarios in Annex 3.1).

18.

As butter consists mainly of fat and SMP mainly of non-fat solids components of milk, butter/SMP tilts do not alter the target or support price for milk because the sum of the two products' values remains unchanged.

19.

The SMP production in the EU module is calculated as a residual of the not-fat solids market. In the scenarios, in order to prevent a model output that would yield negative SMP production (due to a fall in milk production and a large increase in production of fresh dairy products), imports of non-fat solids (i.e. whey powder concentrates or milk protein concentrates) were introduced to ensure a minimum level of SMP production. The non-fat solids import variable was introduced on a SMP equivalent basis in the SMP world market clearing price identity.

20.

In the case of SMP the larger proportion of consumption decrease comes from the elimination of consumption subsidies.

21.

It should be noted that already in the baseline more than 60% of milk production is used for fluid milk and fresh dairy product use.

22

Note that initially the fluid milk premium was reduced by 50% in line with other countries using the domestic pricing arrangements

23.

It should be noted that in estimating domestic milk prices, the large absolute processing margin had to be reduced in this scenario as it is at a relatively high level (compared to producer price) in the baseline (for details see a technical report OECD (2004).

24.

In all fairness it should be noted that the EU has acknowledged its butter/SMP policy tilt and in its CAP reform reduced support price of butter considerably more that SMP price, somewhat reducing the disparity.

25.

This fact also partly explains the larger decline in EU milk production when compared to that Canada in the base scenario, and which reflects the model assumption of an EU milk supply elasticity of 1.23, which is higher than the 0.81 milk supply elasticity assumed for Canada.

26.

The EU is treated as a single region in the model.

27.

While all countries in the model become net importers as a result of liberalisation, domestic production level still exceeds the portion of domestic demand that is for fluid products. Note, that the net import position of specifically modelled countries in PEM is balanced by increase in net exports from rest of the world (ROW in PEM includes countries such as New Zealand, Australia, Argentina).

28.

One may question whether consumer preference for fresh milk would be sufficient to sustain a premium of this size over the importable substitutes such as UHT or reconstituted milk. Regardless, this points out the potential importance of consumer preference in the effects of reform.

29.

The initial value of the sector is presented used as a way to find a comparable metric for the benefits of reform. Other potential options would be value of consumption or GDP, the point being finding an appropriate measure of economic scale to compare against.

30.

Bearing in mind that the purpose and result of market price support policies is to create a transfer from consumers to producers.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

135

31.

Since under full liberalisation, quotas are no longer binding, these results will be sensitive to the estimate of quota rent, which locates the supply function.

32.

This is not to imply that retail price transmission is different in the EU compared with any other region. The basic insights of this experiment are generally applicable to all regions (but note the large-country vs. small country difference discussed below).

33.

Were retailers to maintain a constant margin in terms of percentage of the producer price (for example), the same effect would hold. What is important here is that the margin changes in absolute terms.

34.

These rents do not include any change in the intermediaries’ income from any other source, such as changes in producer surplus. This market is not represented in the model, which explains both the simple structure of price transmission assumed and the absence of any calculations for this sector other than these rents.

35.

It is unclear exactly what the size of this critical band is; acceptable price changes of from +- 10% to +35% have been cited. See Market Effects of Crop Support Measures (OECD 2001) for a discussion.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

136

Annex 3.1 Aglink Results of Dairy Trade Liberalisation Scenarios: Evaluating the Market and Trade Impacts

Legend for the tables The tables show results in percentage changes from baseline levels for prices, production and consumption of milk and dairy products. The trade figures are shown in thousand tons. The baseline levels refer to the last year values (2008) of the Agricultural Outlook baseline 2003-2008 published in OECD (2003a). The columns following the baseline levels illustrate the results of dairy policy reform for a particular country or group of countries. The name of the reformed country or region appears always in the columns’ headings. Thus the results show what would be the market impact if only that country or region would removed its market price support while all the other countries would keep their policies at baseline levels. The reformed regions (group of countries) in the tables are defined as follows: x x x x x x

Europe í D VFHQDULR RI D GDLU\ SROLF\ UHIRUP LQ DOO WKH (XURSHDQ FRPSRQHQWV RI Aglink (a scenario of a complete elimination of all market price support measures in the EU 15, Poland, Hungary and Russia ) NAFTA íDVFHQDULRRIDGDLU\SROLF\UHIRUPLQDOO1$)7$FRXQWULHV8QLWHG6WDWHV Canada, Mexico) Atlantic íDVFHQDULRZKLFKFRPELQHVUHIRUPLQ(XURSHDQG1$)7$FRXQWULHV ALL íDVFHQDULRLQYROYLQJUHIRUPVLQDOl Aglink countries was produced, the results are highlighted in bold. ..&NAFTA and .&Europe (Table 3.1.4) are scenarios of Japan reform combined with reforms in respective regions. For countries not applying MPS (such as New Zealand) the scenarios indicate an impact on these countries of reform in specific region. Supplemental scenario is defined as follows:

x

Retail íLVDQLPSHUIHFWWUDQVPLVVLRQVFHQDULRZKHUHLWZDVDVVXPHGWKDWFRQVXPHUV of countries affected by the reform would only benefit from half of the decline in price recorded in the scenario of simultaneous dairy policy reforms.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

137 Table 3.1.1. Dairy policy reforms – Impact on the European Union

Dairy policy reforms

Baseline

EU

Europe

Internal prices EUR/t Milk

Atlantic

All

Retail

-9.8

-12.1

% changes from baseline 256

-16.5

-16.0

-12.3

Butter

2917

-33.5

-31.0

-22.0

-25.3

-26.6

Cheese

3899

-17.0

-15.9

-10.6

-10.0

-12.5

WMP

2280

-14.4

-14.2

-9.2

-7.5

-9.2

SMP

1943

4.8

3.4

1.1

7.5

5.3

122023

-10.7

-10.4

-8.9

-7.3

-9.7

Butter

1718

-46.9

-45.8

-40.3

-33.7

-38.6

Cheese

7665

-0.5

0.1

1.6

1.1

-0.2

-2.8

-5.0

-4.5

-87.5 -87.5 % changes from baseline

-87.5

Production kt Milk

WMP SMP Consumption kt Fresh dairy prd.

% changes from baseline

915

-10.4

-10.3

803

-87.5

-87.5

40935

7.9

7.7

5.7

4.5

2.7

Butter

1730

4.0

3.7

2.4

2.9

1.4

Cheese

7407

15.4

14.2

9.0

8.4

5.1

WMP

450

2.1

2.1

1.3

1.1

0.6

SMP

769

-31.7

-31.2

-30.4

-32.7

-31.9

Butter

-23

851

827

711

606

665

Cheese

-257

920

784

290

288

135

WMP SMP

-465 -34

-360 981

-361 976

-433 875

-414 417

-420 725

Net imports kt

Thousand tons

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

138

Table 3.1.2. Dairy policy reforms – Impact on Canada

Dairy policy reforms

Baseline

Canada

NAFTA

Internal prices CAD/t Milk

Atlantic

All

Retail

-27.9

-29.7

% changes from baseline 648

-46.9

-30.5

-28.5

Butter

6712

-61.8

-51.4

-42.9

-45.3

-46.2

Cheese

8826

-47.3

-36.3

-34.0

-33.5

-35.3

Other dairy products

4380

-21.5

-17.4

-14.9

-15.0

-15.4

SMP

5465

-46.9

-49.0

-45.2

-41.7

-42.9

8233

-15.8

-5.3

-1.2

-0.8

-2.6

Butter

85

-55.3

-33.9

-26.8

-25.3

-24.8

Cheese

344

-47.0

-5.6

-5.0

-5.1

-5.7

42.7

40.4

45.6

-47.0 -52.8 % changes from baseline

-31.5

Production kt Milk

Other dairy products SMP Consumption kt Fresh dairy products

% changes from baseline

894

20.1

45.1

104

-90.4

-61.2

2877

23.2

13.9

12.9

12.6

5.8

Butter

93

93.5

64.7

47.0

51.6

20.4

Cheese

355

37.7

25.1

23.3

22.6

10.3

Other dairy products

877

22.7

4.8

3.5

3.7

1.2

SMP

90

-47.0

-45.9

-46.1

-46.4

-46.9

Butter

8

141

96

74

77

48

Cheese

12

307

120

112

110

68

Other dairy products

-17

2

-378

-368

-347

-414

SMP Milk

-15 0

80 0

3 392

-6 125

-12 145

-28 168

Net imports kt

Thousand tons

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

139

Table 3.1.3 Dairy policy reforms – Impact on the United States

Dairy policy reforms

Baseline

USA

NAFTA

Internal prices USD/t Milk

Atlantic

All

Retail

-12.7

-14.6

% changes from baseline 321

-17.1

-15.6

-13.3

Butter

3115

-39.4

-35.3

-23.9

-27.1

-28.4

Cheese

3141

-16.3

-14.2

-11.1

-10.5

-13.0

Whey powder

646

17.2

11.0

14.5

21.4

25.0

SMP

2091

-12.4

-13.5

-2.2

4.0

1.9

81326

-11.6

-10.3

-5.1

-4.6

-6.9

Butter

573

-33.0

-31.9

-16.2

-13.7

-15.8

Cheese

4635

-15.0

-12.7

-6.4

-6.2

-9.3

-4.2

-4.1

-6.2

-20.4 -24.2 % changes from baseline

-24.0

Production kt Milk

% changes from baseline

Whey powder

632

-10.1

-8.6

SMP Consumption kt

489

-48.5

-47.4

25969

4.1

3.8

3.4

3.3

Fresh dairy products

1.7

Butter

567

24.0

20.7

12.9

14.9

7.2

Cheese

4786

6.7

5.8

4.4

4.1

2.5

Whey powder

499

-11.9

-10.3

-6.7

-6.6

-8.5

SMP

434

9.9

10.8

1.6

-2.8

-0.7

-4

319

294

160

157

126

Cheese

152

1168

1018

658

636

703

Whey powder SMP

-133 -65

-128 225

-130 224

-140 70

-140 33

-136 60

Net imports kt Butter

Thousand tons

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

140

Table 3.1.4 Dairy policy reforms – Impact on Japan

Dairy policy reforms

Baseline

Japan

..& NAF

Internal prices Y/t

..& EUR

All

Retail

% changes from baseline 80529

-52.3

-53.8

-54.3

-55.0

-63.2

Milk, fluid

85399

-55.0

-56.4

-56.8

-57.5

-64.4

Butter

954618

-85.1

-78.4

-76.9

-73.7

-74.2

Cheese

429835

-22.8

-9.9

-11.2

-8.1

-9.6

SMP

520839

-60.7

-60.9

-58.4

-59.1

-59.8

8422

-20.7

-20.1

-19.7

-19.4

-19.3

Butter

87

-99.9

-99.9

-98.8

-98.8

-82.7

Cheese

37

-99.7

-99.7

-99.7

-99.7

-99.7

Other dairy products SMP

412 175

-99.8 -99.7

-99.8 -99.7

-99.8 -98.6

-99.8 -98.6

2.9 -82.6

Milk, all

Production kt Milk

% changes from baseline

Consumption kt Fresh dairy products

% changes from baseline 5419

21.1

22.0

22.3

22.8

9.7

Butter

88

92.0

68.0

64.3

57.0

16.4

Cheese

264

19.7

7.6

8.7

6.0

3.5

SMP

231

142.7

143.9

127.7

132.4

34.5

Net imports kt

Thousand tons 0

168

147

143

136

86

Cheese

227

315

283

286

279

273

WMP SMP

0 56

45 559

45 562

45 523

45 534

45 280

Butter

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

141

Table 3.1.5 Dairy policy reforms - Impact on Mexico

Dairy policy reforms

Baseline

Mexico

NAFTA

Internal prices MXN/t

Atlantic

All

Retail

% changes from baseline

Milk

3543

-58.6

-51.5

-50.2

-49.6

-50.9

Butter

33798

-46.8

-33.5

-21.8

-25.1

-26.4

Cheese

48335

-49.6

-37.8

-35.6

-35.1

-36.9

WMP

61828

-63.5

-61.5

-61.1

-60.4

-61.1

Production kt

% changes from baseline 10266

-30.0

-26.3

-23.7

-23.3

-24.6

Butter

20

-100.0

-100.0

-100.0

-100.0

-100.0

Cheese Milk powders

175 229

-57.0 -81.0

-46.9 -77.3

-38.4 -72.6

-37.3 -71.6

-35.1 -70.0

4252

9.2

7.4

7.1

7.0

2.9

Butter

74

101.5

57.9

32.1

38.5

17.5

Cheese

238

23.1

15.6

14.2

14.0

6.3

Milk powders

390

7.1

5.7

3.6

2.5

-6.1

Milk

Consumption kt Fresh dairy products

% changes from baseline

Net imports kt

Thousand tons

Butter

54

148.9

116.7

97.6

102.3

86.8

Cheese Milk powders

63 163

217.9 375.6

182.4 361.7

164.1 342.7

161.7 336.2

139.7 298.7

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

142

Table 3.1.6 Dairy policy reforms – Impact on Australia

Dairy policy reforms

Baseline

Europe

NAFTA

Internal prices AUD/t

Atlantic

All

Retail

% changes from baseline

Milk

356

26.8

12.5

30.4

33.6

30.6

Butter

2682

37.1

32.0

52.2

46.6

44.3

Cheese

4323

21.9

22.6

27.4

28.2

25.0

SMP

2949

20.0

0.9

16.8

25.0

21.7

Production kt

% changes from baseline 12909

11.8

4.4

12.6

13.8

12.1

Butter

163

33.8

-1.3

31.4

36.7

33.3

Cheese

423

18.0

16.9

20.5

19.4

16.8

WMP

281

-4.0

2.6

0.9

2.7

1.1

272

32.4

-0.2

34.3 29.9 % changes from baseline

31.2

2240

-6.4

-3.2

-7.1

-7.7

-7.1

Butter

56

-6.0

-5.2

-7.9

-7.3

-6.9

Cheese

222

-18.2

-18.7

-21.9

-22.4

-20.3

WMP

65

0.0

0.0

0.0

0.0

0.0

SMP

50

-3.0

-0.1

-2.6

-3.7

-3.2

Milk

SMP Consumption kt Fresh dairy products

Net exports kt

Thousand tons

Butter

107

166

108

163

171

165

Cheese

202

318

315

337

334

318

WMP SMP

216 223

205 312

223 222

218 305

223 318

219 309

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

143

Table 3.1.7 Dairy policy reforms – Impact on New Zealand

Dairy policy reforms

Baseline

Europe

NAFTA

Internal prices NZD/t

Atlantic

All

Retail

% changes from baseline

Milk

520

19.1

12.5

24.6

25.7

23.1

Butter

3719

25.3

30.0

42.5

36.3

33.5

Cheese

4559

24.0

24.8

30.2

31.1

27.5

WMP

4546

7.9

11.4

13.1

15.0

13.0

SMP

4490

14.7

0.7

12.4

18.4

16.0

17388

7.5

4.7

9.2

9.7

8.4

Butter

497

12.6

0.4

14.0

14.7

13.2

Cheese

394

21.2

26.0

24.2

23.2

20.2

WMP SMP

610 441

-11.4 11.5

0.4 -0.4

-8.8 11.7

-7.4 12.6

-7.7 11.4

Production kt Milk

% changes from baseline

Consumption kt

% changes from baseline

Butter

30

-5.8

-6.7

-8.9

-7.9

-7.4

Cheese

37

-5.6

-5.8

-6.9

-7.0

-6.3

WMP

4

0.0

0.0

0.0

0.0

0.0

SMP

34

-3.4

-0.2

-2.9

-4.1

-3.6

Net exports kt

Thousand tons

Butter

447

512

451

519

523

515

Cheese

361

446

465

458

454

442

WMP SMP

606 407

537 458

608 405

553 459

561 464

559 458

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

144

Table 3.1.8 Dairy policy reforms – Impact on Argentina

Dairy policy reforms

Baseline

Europe

NAFTA

Internal prices Pesos/t

Atlantic

All

Retail

% changes from baseline 558

19.6

9.3

22.3

24.9

22.9

Butter

14805

23.6

20.7

33.5

29.9

28.4

Cheese

10220

28.1

31.6

36.7

37.7

33.6

WMP

10956

7.8

12.4

13.6

15.5

13.6

SMP

9677

16.6

1.0

14.1

21.3

18.8

9842

11.6

4.6

12.6

14.0

12.4

Butter

47

11.6

-4.0

10.1

12.6

11.5

Cheese

454

19.2

14.8

22.3

23.4

20.6

WMP SMP

202 45

3.2 19.7

-1.9 -2.7

4.1 17.0

6.0 19.7

5.0 18.3

-19.8

-19.0

Milk

Production kt Milk

% changes from baseline

Consumption kt

% changes from baseline 41

-16.3

-14.5

-21.6

Cheese

430

-14.6

-16.0

-18.0

-18.4

-16.8

WMP

121

-5.7

-8.7

-9.4

-10.5

-9.4

SMP

24

-11.6

-0.8

-10.1

-14.3

-12.9

Butter

Net exports kt

Thousand tons

Butter

6

18

10

19

20

19

Cheese

24

174

160

203

209

190

WMP SMP

81 22

94 33

88 21

101 32

106 34

102 33

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

145

Table 3.1.9 Dairy policy reforms – Impact on Brazil

Dairy policy reforms

Baseline

Europe

NAFTA

Internal prices R/t

Atlantic

All

Retail

11.0

10.1

% changes from baseline 10.0

458

7.8

6.7

Milk

25005

7.8

4.2

8.8

9.6

8.2

Fresh dairy products

14704

-0.6

-1.4

-1.1

-1.1

-1.1

Butter

95

18.3

2.2

18.3

21.1

18.8

Cheese

540

26.4

19.9

30.8

32.3

28.1

WMP

329

7.7

4.6

10.5

12.6

10.5

SMP

69

64.2

7.4

63.1

75.0

66.7

% changes from baseline 0.2 0.0 -1.7

Milk producer price Production kt

Consumption kt Fresh dairy products

% changes from baseline

0.1

14694

0.6

Butter

98

-15.6

-13.9

-20.7

-19.0

-18.2

Cheese

570

-15.1

-16.6

-18.7

-19.1

-17.5

WMP

385

-8.0

-7.5

-11.4

-13.2

-11.6

SMP

91

-14.2

1.6

-11.6

-16.8

-15.0

Net exports kt

Thousand tons

Butter

-2

30

13

35

36

33

Cheese

-30

199

173

243

254

222

WMP SMP

-56 -22

0 35

-12 -19

22 32

36 44

23 37

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

146

Table 3.1.10 Dairy policy reforms – Impact on Rest of the World

Dairy policy reforms

Baseline

Europe

NAFTA

Production kt

Atlantic

All

Retail

% changes from baseline 219153

2.6

1.2

2.9

3.1

2.9

Butter

4132

10.8

4.7

12.0

13.0

12.0

Cheese

2308

12.3

9.9

14.6

15.1

13.8

WMP

496

6.8

5.6

9.0

9.9

9.0

SMP

427

165.6

67.4

180.3

199.6

184.6

Fresh dairy products

113691

-5.0

-2.8

-5.8

-6.2

-5.8

Skim milk Butter

61471 4534

-1.1 -3.7

-0.1 -3.3

-0.9 -4.8

-1.3 -4.4

-1.2 -4.3

Cheese

2549

-11.4

-9.4

-12.9

-13.2

-11.6

WMP

1734

-2.2

-3.4

-3.7

-4.2

-3.7

SMP

1094

-2.2

-0.4

-2.1

-2.7

-2.5

Milk

Consumption kt

% changes from baseline

Net exports kt

Thousand tons

Butter

-401

211

-58

313

336

289

Cheese

-241

333

227

425

445

372

-1 238 -667

-1 166 64

-1 151 -375

-1 129 125

-1 116 215

-1 129 148

WMP SMP

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

147 Table 3.1.11. Dairy policy reforms – World

Dairy policy reforms

Baseline

Europe

NAFTA

World prices USD/t

Atlantic

All

Retail

64.3

57.4

54.6

% changes from baseline

Butter

1374

45.3

Cheese

1991

25.7

28.9

33.6

34.5

30.8

WMP

1789

8.5

13.6

14.8

16.9

14.8

SMP Wheat Maize Rice Oilseeds Oilseed meals Vegetable oils Beef, USA Beef, EU (E/t) Beef, Mercosur (pesos/t) Pork, USA Poultry, USA

1705 145 110 256 243 164 559 2577 2641 1038 1287 874

16.8 -1.6 -0.1 -0.3 -1.0 -3.3 2.9 -0.7 -0.6 -1.9 -0.9 -0.7

1.0 -0.3 -0.9 -0.2 -0.3 -2.2 2.9 1.1 -0.1 -1.2 -0.7 -0.6

14.2 -1.6 -0.8 -0.4 -1.2 -4.3 4.2 0.0 -0.7 -2.3 -1.1 -1.0 % changes from baseline

21.5 -1.5 -1.0 -0.4 -1.1 -4.1 3.4 0.4 -0.5 -2.4 -1.0 -1.0

19.1 -1.8 -1.0 -0.5 -1.6 -4.9 3.5 0.4 -1.1 -2.3 -1.4 -1.3

Milk

615016

0.0

-0.8

-0.6

-0.2

-1.3

Fresh dairy products

39.7

Production kt 246658

-1.3

-1.0

-1.5

-1.5

-2.2

Butter

8470

-1.4

0.2

-1.1

0.1

-1.6

Cheese

19095

3.0

-0.5

2.3

2.1

0.3

WMP

4002

-2.1

-1.7

-2.2

-2.7

-3.1

SMP

3390

6.8

-0.8

2.3

1.2

-0.6

World exports kt

% changes from baseline

Butter

816

-3.8

-0.5

-3.0

-1.3

-3.3

Cheese

1345

9.1

33.1

23.4

25.3

19.7

WMP SMP

1435 1049

-12.4 6.0

1.1 -8.1

-3.5 0.4

-3.0 4.5

-3.9 3.0

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

148 ANNEX 3.2 Table 3.2.1. Sensitivity analysis of milk supply elasticity assumptions – results for the European Union Dairy policy reforms Internal prices E/t Milk Butter Cheese WMP SMP Production kt Milk Butter Cheese WMP SMP Consumption kt Fresh dairy products Butter Cheese WMP SMP Net imports kt Butter Cheese WMP SMP

Baseline

Base scenario

256 2917 3899 2280 1943

-9.8 -25.3 -10.0 -7.5 7.5

122023 1718 7665 915 803

-7.3 -33.7 1.1 -5.0 -87.5

40935 1730 7407 450 769

4.5 2.9 8.4 1.1 -32.7

-23 -257 -465 -34

606 288 -414 417

Base+10% elast. EU

Base+10% elast. CAN

Base-10% elast. CAN

-10.6 -25.5 -10.2 -7.8 6.5

-9.8 -25.3 -10.0 -7.5 7.5

-9.8 -25.3 -10.0 -7.5 7.5

-6.5 -31.9 1.5 -4.2 -87.5

-7.3 -33.7 1.1 -5.0 -87.5

-7.3 -33.8 1.0 -5.0 -87.5

4.9 2.9 8.7 1.1 -32.4

4.5 2.8 8.4 1.1 -32.7

4.5 2.9 8.5 1.1 -32.7

575 266 -422 420

606 285 -414 417

606 291 -414 417

Base -10% elast. EU

% changes from baseline -9.1 -25.0 -9.8 -7.2 8.5 % changes from baseline -8.1 -35.5 0.6 -5.9 -87.5 % changes from baseline 4.1 2.8 8.2 1.0 -33.0 Thousand tons 635 309 -406 415

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

149 Table 3.2.2. Sensitivity analysis of milk supply elasticity assumptions – results for Canada Dairy policy reforms Internal prices CAD/t Milk Butter Cheese Other dairy products SMP Production kt Milk Butter Cheese Other dairy products SMP Consumption kt Fresh dairy products Butter Cheese Other dairy products SMP Net imports kt Butter Cheese Other dairy products SMP Milk

Baseline

Base scenario

648 6712 8826 4380 5465

-27.9 -45.3 -33.5 -15.0 -41.7

8233 85 344 894 104

-0.8 -25.3 -5.1 40.4 -47.0

2877 93 355 877 90

12.6 51.6 22.6 3.7 -46.4

8 12 -17 -15 0

77 110 -347 -12 145

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

Base+10% elast. EU

Base+10% elast. CAN

Base-10% elast. CAN

-28.1 -45.4 -33.7 -15.1 -42.2

-27.9 -45.2 -33.5 -15.0 -41.7

-28.0 -45.3 -33.5 -15.0 -41.7

-0.9 -25.8 -5.2 40.7 -47.9

-1.5 -25.4 -5.1 40.4 -47.1

0.0 -25.3 -5.1 40.5 -47.0

12.7 52.0 22.8 3.7 -46.3

12.6 51.6 22.6 3.7 -46.4

12.7 51.7 22.7 3.7 -46.4

78 110 -349 -11 145

77 110 -346 -12 204

77 110 -347 -12 80

Base -10% elast. EU

% changes from baseline -27.8 -45.1 -33.3 -14.9 -41.2 % changes from baseline -0.7 -24.9 -5.1 40.2 -46.1 % changes from baseline 12.6 51.3 22.5 3.6 -46.4 Thousand tons 76 109 -345 -13 144

150 Table 3.2.3. Sensitivity analysis of milk supply elasticity assumptions – World dairy prices Dairy policy reforms World prices USD/t Butter Cheese WMP SMP

Baseline

1374 1991 1789 1705

Base scenario 57.4 34.5 16.9 21.5

Base+10% elast. EU

Base -10% elast. EU

% changes from baseline 57.9 34.9 17.3 22.6

56.8 34.2 16.5 20.4

Base+10% elast. CAN

Base-10% elast. CAN

57.4 34.6 16.9 21.5

57.4 34.5 16.9 21.5

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

151 Table 3.2.4. Sensitivity analysis for quota rent assumptions – results for the European Union Dairy policy reforms Internal prices E/t Milk Butter Cheese WMP SMP Production kt Milk Butter Cheese WMP SMP Consumption kt Fresh dairy products Butter Cheese WMP SMP Net imports kt Butter Cheese WMP SMP

Baseline

Base scenario

256 2917 3899 2280 1943

-9.8 -25.3 -10.0 -7.5 7.5

122023 1718 7665 915 803

-7.3 -33.7 1.1 -5.0 -87.5

40935 1730 7407 450 769

4.5 2.9 8.4 1.1 -32.7

-23 -257 -465 -34

606 288 -414 417

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

EU quota rent of 23%

CAN quota rent of 26%

CAN quota rent of 20%

-8.6 -24.9 -9.6 -7.0 9.0

-9.9 -25.3 -10.0 -7.5 7.5

-9.8 -25.2 -9.9 -7.5 7.5

-8.6 -36.6 0.3 -6.4 -87.5

-7.3 -33.8 1.0 -5.0 -87.5

-7.3 -33.7 1.1 -5.1 -87.5

3.9 2.8 8.1 1.0 -33.2

4.5 2.9 8.5 1.1 -32.7

4.5 2.8 8.4 1.1 -32.7

653 321 -402 414

606 294 -414 417

605 282 -414 417

EU quota rent of 17%

% changes from baseline -11.1 -25.7 -10.4 -8.0 6.0 % changes from baseline -5.9 -30.8 1.8 -3.6 -87.5 % changes from baseline 5.1 2.9 8.8 1.1 -32.2 Thousand tons 557 253 -426 421

152 Table 3.2.5. Sensitivity analysis for quota rent assumptions – results for Canada Dairy policy reforms Internal prices CAD/t Milk Butter Cheese Other dairy products SMP Production kt Milk Butter Cheese Other dairy products SMP Consumption kt Fresh dairy products Butter Cheese Other dairy products SMP Net imports kt Butter Cheese Other dairy products SMP Milk

Baseline

Base scenario

648 6712 8826 4380 5465

-27.9 -45.3 -33.5 -15.0 -41.7

8233 85 344 894 104

-0.8 -25.3 -5.1 40.4 -47.0

2877 93 355 877 90

12.6 51.6 22.6 3.7 -46.4

8 12 -17 -15 0

77 110 -347 -12 145

EU quota rent of 23%

CAN quota rent of 26%

CAN quota rent of 20%

-27.7 -45.0 -33.2 -14.8 -40.9

-28.0 -45.3 -33.5 -15.0 -41.7

-27.8 -45.2 -33.4 -15.0 -41.7

-0.6 -24.6 -5.1 40.0 -45.6

1.0 -25.3 -5.1 40.6 -47.0

-2.6 -25.4 -5.1 40.3 -47.1

12.5 51.1 22.4 3.6 -46.4

12.7 51.7 22.7 3.7 -46.4

12.6 51.6 22.6 3.7 -46.4

76 109 -343 -13 144

77 110 -348 -12 0.2

77 110 -346 -12 286

EU quota rent of 17%

% changes from baseline -28.2 -45.5 -33.8 -15.2 -42.6 % changes from baseline -1.0 -26.1 -5.2 40.9 -48.5 % changes from baseline 12.8 52.2 22.9 3.8 -46.3 Thousand tons 78 111 -350 -10 146

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

153 Table 3.2.6. Sensitivity analysis for quota rent assumptions – World dairy prices Dairy policy reforms World prices USD/t Butter Cheese WMP SMP

Baseline

1374 1991 1789 1705

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

Base scenario 57.4 34.5 16.9 21.5

EU quota rent of 23%

EU quota rent of 17%

% changes from baseline 56.5 34.0 16.3 19.8

58.2 35.1 17.6 23.2

CAN quota rent of 26%

CAN quota rent of 20%

57.3 34.5 16.9 21.5

57.4 34.6 17.0 21.6

154

Annex 3.3 Table 3.3.1 Main policy assumptions for dairy markets (Agricultural Outlook 2003-2008)

ARGENTINA Dairy export tax AUSTRALIA (a) Domestic support payment (b) CANADA Milk target price (b) Butter support price SMP support price Dairy subsidy Cheese tariff-quota in-quota tariff out-of-quota tariff Subsidised export limits cheese SMP EU15 (c) (d) Milk quota (e) Milk target price Butter intervention price SMP intervention price Tariff-quotas butter in-quota tariff out-of-quota tariff cheese (f) in-quota tariff out-of-quota tariff SMP in-quota tariff out-of-quota tariff Subsidised export limits (a) butter cheese SMP other milk products JAPAN (c) Direct payments (m) Deficiency/direct payment ceiling (g) Milk guaranteed price (b) standard transaction price (h) deficiency payment (i) Butter stab. indicative price SMP stab. indicative price Cheese tariff (j) Tariff-quotas Butter in-quota tariff out-of-quota tariff SMP in-quota tariff out-of-quota tariff WMP in-quota tariff out-of-quota tariff

Average 1997-01

2000

2001

2002

2003

2004

2005

2006

2007

2008

0

0

0

5

5

5

5

5

5

5

1.1

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

56 5,491 4,533 2.46 20 2 254

57 5,558 4,670 1.58 20 1 246

58 5,711 4,826 1.50 20 1 246

59 5,870 4,945 0.33 20 1 246

60 5,986 5,038 0.00 20 1 246

61 6,104 5,083 0.00 20 1 246

62 6,230 5,225 0.00 20 1 246

64 6,354 5,321 0.00 20 1 246

65 6,480 5,415 0.00 20 1 246

66 6,608 5,507 0.00 20 1 246

10 47

9 45

9 45

9 45

9 45

9 45

9 45

9 45

9 45

9 45

118 0.319 3,282 2,055

119 0.319 3,282 2,055

120 0.319 3,282 2,055

120 0.319 3,282 2,055

120 0.319 3,282 2,055

120 0.319 3,282 2,055

121 0.310 3,200 2,004

122 0.292 3,036 1,901

122 0.274 2,872 1,798

122 0.265 2,790 1,747

kt pw % % kt pw % % kt pw % %

85 63 153 77 43 108 59 36 95

87 66 144 102 42 96 68 35 88

87 66 144 102 42 96 68 35 88

87 66 144 102 42 96 68 35 88

87 66 144 102 42 96 68 35 88

87 66 144 102 42 96 68 35 88

87 66 144 102 42 96 68 35 88

87 66 144 102 42 96 68 35 88

87 66 144 102 42 96 68 35 88

87 66 144 102 42 96 68 35 88

kt pw kt pw kt pw kt pw

420 346 288 1,013

399 321 273 958

399 321 273 958

399 321 273 958

399 321 273 958

399 321 273 958

399 321 273 958

399 321 273 958

399 321 273 958

399 321 273 959

JPY/kg kt pw JPY/litre JPY/litre JPY/litre ’000 JPY/t ’000 JPY/t %

.. 2,374 .. .. .. .. .. 34

.. 2,400 74 64 11 910 524 31

10 2,270 .. .. .. .. .. 31

10 2,270 .. .. .. .. .. 30

10 2,270 .. .. .. .. .. 30

10 2,270 .. .. .. .. .. 30

10 2,270 .. .. .. .. .. 30

10 2,270 .. .. .. .. .. 30

10 2,270 .. .. .. .. .. 30

10 2,270 .. .. .. .. .. 30

2 35 605 116 17 260 0 24 345

2 35 679 116 16 275 0 24 377

2 35 679 116 16 275 0 24 377

2 35 679 116 16 275 0 24 377

2 35 679 116 16 275 0 24 377

2 35 679 116 16 275 0 24 377

2 35 679 116 16 275 0 24 377

2 35 679 116 16 275 0 24 377

2 35 679 116 16 275 0 24 377

2 35 679 116 16 275 0 24 377

% AUDc/kg CADc/litre CAD/t CAD/t CADc/hltr kt pw % % kt pw kt pw mt pw EUR/litre EUR/t EUR/t

kt pw % % kt pw % % kt pw % %

For notes, see end of the table.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

155 Table 3.3.1 (cont’d)

KOREA Tariff-quotas Butter in-quota tariff out-of-quota tariff SMP in-quota tariff out-of-quota tariff WMP in-quota tariff out-of-quota tariff MEXICO Butter tariff Tariff-quotas cheese in-quota tariff out-of-quota tariff SMP in-quota tariff out-of-quota tariff Liconsa social program RUSSIA Butter tariff Cheese tariff UNITED STATES (k) Milk support price (b) Target price (l) Butter support price SMP support price Butter tariff-quota in-quota tariff out-of-quota tariff Cheese tariff-quota in-quota tariff out-of-quota tariff Subsidised export limits (a) butter SMP

kt pw % % kt pw % % kt pw % % % kt pw % % kt pw % % MXN mn % % USDc/litre USDc/litre USD/t USD/t kt pw % % kt pw % % kt pw kt pw

Average 1997-01

2000

2001

2002

2003

2004

2005

2006

2007

2008

0.3 40 89 0.8 20 176 0.4 40 176

0.3 40 89 0.9 20 176 0.5 40 176

0.4 40 89 0.9 20 176 0.5 40 176

0.4 40 89 0.9 20 176 0.5 40 176

0.4 40 89 1.0 20 176 0.5 40 176

0.4 40 89 1.0 20 176 0.6 40 176

0.4 40 89 1.0 20 176 0.6 40 176

0.4 40 89 1.0 20 176 0.6 40 176

0.4 40 89 1.0 20 176 0.6 40 176

0.4 40 89 1.0 20 176 0.6 40 176

8

6

4

2

0

0

0

0

0

0

9 50 132 90 0 132 3,195

9 50 131 90 0 131 3,334

9 50 129 90 0 129 3,425

9 50 128 90 0 128 3,410

9 50 126 90 0 126 3,395

9 50 125 90 0 125 3,380

9 50 125 90 0 125 3,364

9 50 125 90 0 125 3,349

9 50 125 90 0 125 3,334

9 50 125 90 0 125 3,319

20 15

20 15

20 15

20 15

20 15

20 15

20 15

20 15

20 15

20 15

23 0.0 1,492 2,228 12 9 112 132 12 84

22 0.0 1,454 2,227 13 9 117 135 12 84

22 0.0 1,701 2,079 13 9 117 135 12 84

22 38.5 1,957 1,947 13 9 117 135 12 84

22 38.5 2,315 1,764 13 9 117 135 12 84

22 38.5 2,315 1,764 13 9 117 135 12 84

22 38.5 2,315 1,764 13 9 117 135 12 84

22 0.0 2,315 1,764 13 9 117 135 12 84

22 0.0 2,315 1,764 13 9 117 135 12 84

22 0.0 2,315 1,764 13 9 117 135 12 84

26 78

21 68

21 68

21 68

21 68

21 68

21 68

21 68

21 68

21 68

a) Year ending 30 June. b) For manufacturing milk. c) Year beginning 1 April. d) Prices and payments in market Euro’s -see Glossary of Terms. e) Total quota, EU15 starting in 1995. f) Calendar year minimum access for Australia, New Zealand and Canada before 1995. g) Manufacturing milk eligible for deficiency/direct payments. h) Paid to producers. i) Difference between transaction price and guaranteed price. j) Excludes processed cheese. k) Year beginning 1 January. l) The counter-cyclical payment is determined as a 45% difference between the target price and the Boston class I price. m) In addition to direct payments, a compensation payment is paid - equal to 80% difference between the market price and the base price (the average price of the past three years).

Note : The source for tariffs and Tariff Rate Quotas (except Russia) is AMAD (Agricultural market access database). The tariff and TRQ data are based on Most Favoured Nation rates scheduled with the WTO and exclude those under preferential or regional agreements, which may be substantially different. Tariffs are averages of several product lines. Specific rates are converted to ad valorem rates using world prices in the Outlook. Import quotas are based on global commitments scheduled in the WTO rather than those allocated to preferential partners under regional or other agreements. Source: OECD Secretariat.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

156

References Abdulai, A. (2002), “Using threshold co-integration to estimate asymmetric price transmission in the Swiss pork market” Applied Economics, 34, 679-687. Azzam, A. M. (1999), “Asymmetry in rigidity in farm-retail price transmission”, American Journal of Agricultural Economics, 81, 525–533. Ball, R. and Mankiw, G. (1994), “Asymmetric price adjustment and economic fluctuations”. Economic Journal 104, 247-261. Benson, B. L. and Faminow, M. D. (1985), “An alternative view of pricing in retail food markets”, American Journal of Agricultural Economics, 67, 296–305. Bettendorf, L. and Verboven, F. (2000), “Incomplete transmission of coffee bean prices in the Netherlands”, European Review of Agricultural economics, 27, 1-16. Bouamra-Mechemache Z.; J.P. Chavas; T. Cox and V. Réquillart (2002), “EU dairy policy reform and future WTO negotiations: a spatial equilibrium analysis”, Journal of Agricultural Economics. 53(2):4-29. Cotterill, R.W. (2002), Who Benefits from Deregulated Milk Prices: The Missing Link is the Marketing Channel. Food Marketing Policy Issue Paper No. 26 April 2002 Food Marketing Policy Center, Department of Agricultural and Resource Economics, University of Connecticut. de Haan D, H. Steinfeld and H. Blackburn (1997), Livestock and the environment: finding a balance. http://www.fao.org/docrep/x5303e/x5303e00.htm Study report coordinated by World Bank, FAO,USAID. Dohlman, E., Hoffman, L., Young, E. and McBride, W. (2004), Peanut Policy Change and Adjustment Under the 2002 Farm Act, Economic Research Service, USDA, Washington, http://www.ers.usda.gov/publications/OCS/Jul04/OCS04G01/ocs04G01.pdf Donnellan, T. and Westhoff, P. (2001), World Dairy Trade Reform: Perspectives from Europe and the USA. Paper presented at the International Dairy Federation World Summit, Auckland, New Zealand, October 28th - November 1st 2001. FAPRI (2002), The Doha Round of the World Trade Organization: Appraising Further Liberalization of Agricultural Markets Food and Agricultural Policy Research Institute Iowa State University and University of Missouri-Columbia Working Paper 02-WP 317 November 2002. Gardner, B.L. (1975), “The Farm-Retail Price Spread in a Competitive Food Industry”, American Journal of Agricultural Economics 57, 383-406. Gohin, A. and Guyomard, H. (2000), “Measuring market power for food retail activities: French evidence”, Journal of Agricultural Economics 51, 181-195. Holloway, G.J. (1991), "The Farm-Retail Price Spread in an Imperfectly Competitive Food Industry", American Journal of Agricultural Economics, 71, 979–989. IDF (2003), World Dairy Situation 2003, Bulletin of the International Dairy Federation, 384/2003. IDF-CFCE (2003), “World market situation and outlook for dairy products”, presentation of the IDF president Philippe Jachnik to the 4th International Federation of Agricultural Producers’ conference in Pretoria, South Africa, 10.05.2003. DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

157 Kinnucan, H. W. and Forker, O. D. (1987), “Asymmetry in farm-retail price transmission for major dairy products”, American Journal of Agricultural Economics, 69, 285–292. Langley, S., Blayney, D., Stout, J., Somwaru, A., Normile, M.A., Miller, J. and Stillman, R. (2003), A Trade Liberalization in International Dairy Markets. USDA-ERS, Washington, DC. Paper prepared for presentation at the American Agricultural Economics Association Annual Meeting, Montreal, Canada, July 27-30, 2003. Larivière, S. and Meilke, K. (1999), An Assessment of Partial Trade Liberalization on Canada, the US, and the EU – 15 dairy industries, paper prepared for the Policy Research Symposium National and Trade Dairy Policies Implications for the Next WTO Negotiations Kansas City, October 8-9 1999. London Economics (2003), Examination of UK Milk Prices and Financial Returns, Report prepared for The Milk Development Council, February 2003. McCorriston, S. and Sheldon, I.M. (1996), “Trade Policy in Vertically-Related Markets”, Oxford Economic Papers 48: 664-672. McCorriston, S., Morgan, C. W. and Rayner, A. J. (2001), “Price transmission: the interaction between market power and returns to scale”, European Review of Agricultural Economics, 28, 143–159. Meilke, K. and Larivière, S. (1999), The problems and Pitfalls in Modeling International Dairy Trade Liberalization. International Agricultural Trade Research Consortium, Working Paper 99-3. OECD (2001a), Decoupling: a conceptual overview, Directorate for Food, Agriculture and Fisheries, Committee for Agriculture, Paris. OECD (2001b), Market Effect of Crop Support Measures, Directorate for Food, Agriculture and Fisheries, Committee for Agriculture, Paris. OECD (2003a), Agricultural Policies in OECD Countries: Monitoring and Evaluation, Paris. OECD (2003b), OECD Agricultural Outlook 2003-2008, Directorate for Food, Agriculture and Fisheries, Committee for Agriculture, Paris OECD (2004a), Dairy trade liberalisation with Aglink, Technical paper describing the changes needed to produce trade liberalisation scenarios with AGLINK. Available on http://www.oecd.org/findDocument/0,2350,en_2649_33783_1_119669_1_1_1,00.html

OECD (2004b), Analysis of the 2003 CAP Reform, Directorate for Food, Agriculture and Fisheries, Committee for Agriculture, Paris Reagan, P. and Weitzman, M. (1982), “Asymmetries in price and quantity adjustments by the competitive firm”, Journal of Economic Theory, 27, 410–420. Serra, T. and Goodwin, B.K. (2003), “Price transmission and asymmetric adjustment in the Spanish dairy sector”, Applied Economics, 2003, 35, 1889–1899. Sexton, R.J., Sheldon, I., McCorriston, S. and Wang, H. (2003), Analyzing Vertical Market Structure and Its Implications for Trade Liberalization and Market Access, International Agricultural Trade Research Consortium, Working Paper No. 03-8 Shaw, I. and Love G. (2001), Impacts of Liberalising World Trade in Dairy Products, ABARE Research Report 01.4, Canberra. Wann, J.J. and Sexton. R.J. (1992), “Imperfect competition in multi-product food industries with an application to pear processing”, American Journal of Agricultural Economics, 72, 980–990. Ward, R. W. (1982), “Asymmetry in retail, wholesale, and shipping point pricing for fresh vegetables”, American Journal of Agricultural Economics, 62, 205–212.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

158 Weldegebriel, H.T. (2004), “Imperfect Price Transmission: Is Market Power Really to Blame?”, Journal of Agricultural Economics . 55 (1), 101-114. Wohlgenant, M.K. (1985), "Competitive Storage, Rational Expectations, and Short-Run Food Price Determination", American Journal of Agricultural Economics, 67, 739–748. Zhu, Y, Cox, T. and J-P. Chavas (1998), A Spatial Equilibrium Analysis of Trade Liberalization and the U.S. Dairy Sector, Department of Agricultural and Applied Economics, University of Wisconsin-Madison, Final Report for the NRI grant # 94-37400-0966. http://aae.wisc.edu/globalMt/nri-gatt/index.html.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

159

Annex A A Description of the Aglink and PEM Models

Aglink model Aglink is a partial equilibrium dynamic supply-demand model of world agriculture, developed by the OECD Secretariat in close co-operation with member countries. It represents annual supply, demand and prices for the principal agricultural commodities produced, consumed and traded in member countries. The overall design of the model focuses particular attention to the potential influence of agricultural policy on agricultural markets in the medium term. Development on the basis of the agricultural economics literature, existing member country models, and on formal Bilateral Reviews has resulted in a model specification which reflects the views of participating member countries, subject to constraints which uniformity across country modules requires. Thus, agricultural markets are modelled specifically to best capture individual policies and particular market settings relevant for each country. Individual country modules modelled in Aglink are calibrated on baseline projections, received from member countries via a so-called questionnaire reply system. The country modules are then merged and the entire model (~ 2800 equations) is solved simultaneously to generate the commodity baseline. Model characteristics, key factors and model assumptions related to the Aglink model used in the development of the Agricultural Outlook 2003-2008 baseline (OECD, 2003b) and in empirical simulations carried out in this report are described below.

General characteristics and assumptions Aglink is a “partial equilibrium” model for the main OECD agricultural commodity markets relative to supply, consumption and prices. Non-agricultural markets are not modelled, and are treated exogenously to the model. Feedback to the macro-economy is not accounted for. This may be particularly important for Rest of World countries in which agriculture is often a significant part of the domestic economy. Certain markets, such as sheepmeat, fish and wool are also not modelled or incompletely modelled. World markets for agricultural commodities are competitive. Buyers and sellers do not behave as if they had market power, and market prices are determined through a global equilibrium in supply and demand. Domestically produced and traded commodities are viewed to be perfect substitutes by buyers and sellers. In particular, importers do not distinguish commodities by country of origin. Countries/regions modelled endogenously in Aglink are: Argentina, Australia, Brazil, Canada, China, the European Union 15, Hungary, Japan, Korea, Mexico, New Zealand, Poland, Russia, Rest of World, Uruguay, and the United States. Rest of World module is specified without any policy measures in place. Countries/regions accounted for DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

160 exogenously are: Czech Republic, Norway, Other Independent States, Slovakia, Switzerland and Turkey. The main commodities modelled by Aglink are: Barley, Feed barley, Beef and veal, Butter oil , Butter, Casein, Coarse grains, Cheese, Eggs, Fresh dairy products, Lamb, Maize, Milk, Concentrated milk, Manioc, Milk powder, Mutton, Non ruminant meat, Other cereals, Other dairy products, Vegetable oils, Oilseed oil, Oilseed meal, Oilseeds, Oats, Pigmeat, Palm oil, Potatoes, Poultry meat, Rice, Rapeseed oil, Rapeseed, Ruminant meat, Rye, Soybean, Special crops, Sunflower, Sunflower oil, Sunflower meal, Sheepmeat, Soybean oil, Soybean meal, Skim milk, Skim milk powder, Sorghum, Vegetable oil, Whole milk powder, Wool, Wheat and Whey powder. Aglink simulates market determination of equilibrium prices for most of its commodities. For these commodities it is assumed that a market price must adjust to equate exactly total demand, including carry-over, to total supplies, including carry-in. Each market uses a specific world reference price. In Aglink, considerable effort was made to retain a calendar year basis for all data. This was not possible for many series, particularly for crops and for dairy. The functional relationships linking supply and demand to prices in Aglink are in most cases linear in the logarithms of the variables. Equation coefficients are partial elasticities. In developing Aglink, an attempt has been made to obtain up-to-date estimates of these elasticities. Many of these new elasticities come from, or are based on, models currently in use in member countries. Some are the result of econometric analysis initiated by the Secretariat, through consultants or by Secretariat staff. Where world market and domestic producer and consumer prices are linked, that link is represented through price equations which are linear in world market prices, converted to local currency terms, margins approximating transportation costs and quality differentials, and border measures -- tariffs, taxes, subsidies etc. In Aglink, trade for each country by commodity pairing is given one of three possible treatments. In a few cases, the level of imports or exports, either bilateral or in total, can be set exogenously. This may be the case, for example, where a trade quota or an access agreement applies. In a few other cases certain bilateral trade links are reflected, for example, poultry trade between the United States and Canada. Finally, and most commonly, trade is the residual of a supply-utilisation identity equation. In these cases it is the modeller’s responsibility to identify simulated exports or imports above export limits or below import access.

Dairy markets specific characteristics and assumptions The dairy component of Aglink covers production and consumption of milk and main dairy products in major OECD and several non-member economies markets, covering both importers and exporters. Thus, the Aglink representation of the dairy sector allows the analysis of impacts on world markets for tradable dairy products where those markets are explicitly modelled. As for other commodities in Aglink, dairy markets are modelled specifically to best capture individual policies and particular market settings relevant for each country. Milk production in Aglink is expressed as the product of milk cow inventory and milk yield. In Canada and the EU, milk production is determined by the setting of the production quota. Since output prices do not guide producer decisions, price elasticities of milk supply have not been defined for these countries. A 'shadow price' of milk supply in DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

161 quota countries has to be identified in order to specify an underlying supply function in these countries. This is essential for modelling a scenario which involves a substantial policy change or, alternatively, a total elimination of a quota system. The milk production link to the beef sector in Aglink is based on a theory of supply in which producers invest in breeding stock by retaining cows and heifers from slaughter when the capital value of these animals exceeds their current market value. The capital value of a beef-breeding cow is a function of the expected income stream earned from future sales of calves. The higher the expected value of future beef and milk production the greater the investment in the breeding herd. The retention for breeding lowers the availability of animals for slaughter in the short run. Thus, to the extent that current beef prices influence expectations of future beef prices, there exists the possibility of a negative elasticity of beef supply response in the short run. In Aglink, the equations corresponding to investment demand for beef cows link ending inventories to expected producer prices, feed costs and other factors. The beef and milk production equations link supply in a particular year to the breeding inventories in earlier years and to producer prices for beef and competing products and to costs. Dairy supply is modelled on the assumption that the value of milk components (fat, non-fat solids) will tend to equalise across products. Thus, if demand for a product made primarily from one of the components grows relative to demand for products made from the other then the relative value of components would adjust. That is, a unit of fat in cheese would have the same value as unit of fat in WMP or butter, after adjusting for processing costs. Thus, only butter and SMP prices are typically used as proxies for fat and non-fat solids prices. Typically in Aglink, butter production and SMP production are residuals of the market-clearing for milkfat and non-fat solids, respectively. The production of cheese and WMP are logit functions that depend on the price of that good relative to the input cost. This last term is calculated on the basis of the butter and SMP prices and the shares of milkfat and non-fat solids in the various products. In the dairy market, as is the case for other commodities, where world dairy prices and domestic producer and consumer prices are linked, that link is represented through price (transmission) equations which are linear in world market prices, converted to local currency terms, margins approximating transportation costs and quality differentials, and border measures. In several countries, that have a large domestic dairy market and operate with border protection measures, a domestic market clearing price is assumed. In these cases, typically, the trade equations are linked to the evolution of domestic policy and market prices and limits set by the WTO. The world market reference prices for dairy sector are specified as follows: the world prices of butter, cheese, SMP and WMP are the FOB Northern Europe prices denominated in US dollars. The world casein price is approximated by New Zealand casein export price. The world whey powder price is approximated by the US whey powder wholesale price.

PEM model The Policy Evaluation Model (PEM) provides a stylised representation of production, consumption, and trade of milk, and major cereal and oilseeds crops in six OECD countries: Canada, the European Union, Japan, Mexico, Switzerland, and the United DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

162 States.1 The PEM allows for a stylised version of existing and hypothetical policies in the participant countries. The purpose of the PEM is to provide a closer connection between measurement of support as done using the PSE and quantitative analysis of the impacts and distribution of such support. In constructing the PEM, three main sets of assumptions were required: 1) those relating to the basic structure of supply and demand response, 2) those relating to the underlying data and the elasticities, and 3) those relating to the primary incidence of support measures on prices and quantities. Economic theory and results of previous studies guided analysts’ choices about the structure of the model, the data and economic parameters to use. The classification of support measures in the PSE guided choices about their primary incidence. The starting point for analysis of policy effects for the PEM is the Producer Support Estimate (PSE). There are eight main categories in the PSE, one for market price support and seven for different kinds of budgetary payments, distinguished by implementation criteria. The PSE data conveys two kinds of information necessary for PEM analyses. First, the PSE indicates the level of, and changes over time in the level of monetary transfers from consumers and taxpayers to farmers resulting from agricultural policies. Second, support estimates are classified according to the way the associated policy measure is implemented thereby highlighting the ‘initial incidence’ of the support measure for analytical purposes. Each of the main kinds of support defined in this classification appears in the model with a specific differentiated “initial incidence” on producer and consumer incentive prices. The country ‘modules’ of the PEM were all developed according to a common structure. Policy experiments were carried out using a model linking these individual modules through world price and trade effects. Commodity supply is represented through a system of factor demand and factor supply equations. Excepting the rest of world module, there are equations representing demand and supply response and prices for at least four categories of inputs used to produce these crops in the study countries. The factor demand equations reflect the usual assumptions of profit maximisation constrained by the production relationship. Supply response corresponding to a medium term adjustment horizon of approximately five years is reflected in the values assumed for the price elasticities of factor supplies and the parameters measuring the substitutability of factors in production as well as the factor shares. No factor is assumed to be completely fixed in production, but land and the other farm-owned factors are assumed to be relatively more fixed (have lower price elasticities of supply) than the purchased factors. Likewise, no factor is assumed freely mobile, but purchased inputs are assumed relatively more mobile (a higher elasticity of supply) than the farm-owned factors. Most supply parameters needed for the model come from systematic reviews of the empirical literature by external consultants. (D. Abler (2000) and K. Salhofer (2000)). Both reviews were commissioned by the Secretariat to obtain objectively plausible values of the parameters. 2 Each of the country modules has two farm-owned factors: land and a residual “other farm owned factors”. The set of purchased factors covered in each country includes, at the least, fertiliser and a residual “other purchased factors”. In the PEM, land is assumed heterogeneous, but transformable between one use and another. The farmer acts to maximise profits by allocating land across its possible uses (wheat, coarse grains, oilseeds, rice, other arable uses, milk or beef pasture, other agricultural uses) according to a transformation function.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

163 The land transformation function is assumed separable for different categories of use such that the land allocation problem facing the farmer is solved in successive stages. First, the producer chooses to allocate land to rice, other agricultural uses, or to a group of uses including all other arable and pasture uses. This group is then allocated in the second stage between pasture, cereals and oilseeds, and other arable uses. Finally, the cereals and oilseeds group is allocated between wheat, coarse grains, and oilseeds (Figure A.1.1). Figure A.1.1 PEM Land Allocation Structure All land

V1

Pasture, field crops

V2

Pasture

V2

V1

Rice

V1

Other/misc.

V2

Other Arable

Cereals/Oilseeds

V3 Wheat

V3 Coarse Grains

V3

Oilseeds

At each of these stages a constant elasticity of transformation (CET) function is used to describe how uses may be allocated. That is, at each level in this decision-making process the transformability of land is the same, but this rate differs between levels. The parameter of the CET function, V, determines the mobility of land between uses at each stage. As we move downward through this land allocation framework, land becomes more similar in use and therefore more easily fungible between uses. We expect V3>V2>V1 in general. Commodity demand equations in the PEM models relate domestic consumption of outputs to prices (at the farm level). Co-movement of prices may occur even when policy measures are targeted directly to only one or two commodities because wheat, coarse grain, oilseeds and rice may be substitutes in both production and consumption.3 Moreover, depending on the degree to which crops are substitutes in demand, comovement in their prices may lead to small “net” changes in quantity demanded for any one crop and thus in their total. That is, the total demand for crops may be highly price inelastic. The PEM does not represent in a fully comprehensive manner the specifics of support programmes applying to each individual commodity in each one of the participant countries. Rather, the aim is to represent the “incidence” of support measures in the same way that “incidence” is implied by the classification of support measures for the PSEs. In this system, support measures are classified according to the main or primary condition that producers must meet in order to be eligible for the support. Usually, knowledge of the conditions of eligibility of a particular support measure, as revealed by its classification in the PSE, will be enough to infer its “initial incidence”.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

164 In order to undertake policy simulation experiments the model must be calibrated for a specific base year using the data in the PSE database. This calibration includes all quantities produced, consumed and exported in each country and each commodity of the model, the set of world and domestic prices and the amounts of the different kinds of support creating price wedges. Land quantities are taken from FAO data and other inputs quantities are defined using quantity or constant price volume indexes. Input prices are derived then from cost shares and factor quantities.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

165

Notes 1.

The European Union is treated in the model as a single region. A version of the PEM model incorporating beef production and trade is currently under development by the Secretariat.

2.

Although the own and cross-price elasticities of crop supply are not explicit parameters in the PEM crop models, their values can be calculated from knowledge of the elasticities of factor supply, factor substitution and factor shares.

3.

Cross-elasticities of demand are assumed to exist between the crop commodities, but not between milk and beef or between these livestock commodities and crops. This assumption is driven primarily by data availability.

DAIRY POLICY REFORM AND TRADE LIBERALISATION – ISBN-92-64-01159-5 © OECD 2005

0&$%16#-*$"5*0/4 SVF"OESÏ1BTDBM 1"3*4$&%&9 13*/5&%*/'3"/$& 1 *4#/o