WPS5165 Policy Research Working Paper 5165 Krueger/Schiff/Valdés Revisited Agricultural Price and Trade Policy Reform in Developing Countries since 1960 Kym Anderson The World Bank Development Research Group January 2010 Policy Research Working Paper 5165 Abstract A study of distortions to agricultural incentives in 18 percent of world agricultural output and employment. developing countries during 1960­84, by Krueger, Schiff This paper summarizes the methodology used in the and Valdés (1988; 1991), found that policies in most new study (pointing out similarities and differences of those developing countries were directly or indirectly with those used by the OECD and by Krueger, Schiff harming their farmers. Since the mid-1980s there has and Valdés), compares a synopsis of the indicators from been a substantial amount of policy reform and opening Krueger, Schiff and Valdés and the new study for the up of many developing countries, and indicators of that period to 1984, summarizes the changing extent of price progress have been made available recently by a new distortions across countries and commodities globally study that has compiled estimates for a much larger since then, and concludes by evaluating the degree of sample of developing countries and for as many years as distortion reduction over the years since 1984 compared possible since 1955. The new study also covers Europe's with how much still remains, according to the results of a transition economies and comparable estimates for global economy wide model. high-income countries, thereby covering more than 90 This paper--a product of the Development Research Group--is part of a larger effort to understand the implications of changes in agricultural trade policies for economic welfare and poverty reduction. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The author may be contacted at kym.anderson@adelaide.edu.au. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Krueger/Schiff/Valdés Revisited: Agricultural Price and Trade Policy Reform in Developing Countries since 1960 Kym Anderson School of Economics University of Adelaide Adelaide, SA 5005, Australia Phone +61 8 8303 4712 kym.anderson@adelaide.edu.au Keywords: Agricultural price distortions, trade policies, developing countries JEL codes: F13, F59, H20, N50, O13, Q18 This article draws on results from a recent multi-country World Bank research project (see www.worldbank.org/agdistortions). Financial assistance from World Bank Trust Funds (particularly those provided by the governments of Japan, the Netherlands and the United Kingdom), and the Australian Research Council is gratefully acknowledged, as are the contributions of the country case study contributors to the Agricultural Distortions project, computational assistance by a team of research assistants led by Ernesto Valenzuela, research collaboration with Johanna Croser and Peter Lloyd, and helpful feedback from commentators on numerous conference and seminar papers over the past year and from referees. Views expressed are the author's alone and not necessarily those of the World Bank or its Executive Directors, nor the countries they represent, nor of the institutions providing funds for this research project. Forthcoming in Applied Economic Perspectives and Policy 1(2), 2010. 2 Krueger/Schiff/Valdés Revisited: Agricultural Price and Trade Policy Reform in Developing Countries since 1960 Two decades ago, a major World Bank study of distortions to agricultural incentives in 17 developing countries plus Portugal was published by Krueger, Schiff and Valdés (1988; 1991). That study covered roughly the period 1960-84, which for many developing countries was the first 25 years of independence from a colonial power. It found that policies in most of those developing countries were harming their farmers, either directly via such things as taxes on agricultural exports or indirectly via manufacturing protection or overvalued exchange rates. Since the mid-1980s there has been a substantial amount of policy reform and opening up of many developing countries, but no systematic quantitative monitoring of those policy changes. To help fill this lacuna, a study by the World Bank has revisited this issue and provides indicators for a much larger sample of developing countries and for as many years as possible since 1955. The new study also covers European economies in transition from socialism and, for completeness, it extends estimates for high-income countries back three decades prior to the start of comparable estimates from 1986 by the Organization for Economic Cooperation and Development (OECD). In so doing it covers more than 90 percent of world agricultural output and employment, with the focus countries accounting for 96 percent of global GDP. Some of the policy developments of the past half century have happened quite suddenly and been transformational. They include the end of colonization around 1960, the implementation of the Common Agricultural Policy in Europe from 1962, the floating of exchange rates and associated liberalization, deregulation, privatization and democratization in the mid-1980s in many countries, and the opening of China in 1979, Vietnam in 1986, and Eastern Europe following the fall of the Berlin Wall in 1989 and the demise of the Soviet Union in 1991. Less newsworthy and hence less noticed are the influences of policies that change only gradually in the course of economic development as comparative advantages evolve, but they too have made a substantial impact on the global economy. The present paper is structured as follows. It begins with some background comments before summarizing the methodology used in the World Bank's new study, pointing out its 3 similarities and differences with those used by the OECD and by Krueger, Schiff and Valdés (K/S/V). It compares a synopsis of the indicators from K/S/V and the new study for the period to 1984, before summarizing the changing extent of price distortions across countries and commodities regionally and globally since then. The third section reports results from a global economy wide modeling exercise aimed at quantifying the trade and welfare effects of the reduction in price and trade distortions over the years since 1984 compared with the prospective effects of removing remaining distortions to agricultural and other merchandise trade. The final section concludes with some observations on what might influence the prospects of such reform in the years ahead. Background For decades agricultural protection and subsidies in high-income (and some middle-income) countries have been depressing international prices of farm products, which lowers the earnings of farmers and associated rural businesses in developing countries. The 1958 Haberler report to Contracting Parties to the General Agreement on Tariffs and Trade forewarned that such distortions might worsen, and indeed they did between the 1950s and the early 1980s in East Asia (Anderson, Hayami and Others 1986). Such policies depress international prices for farm products, thereby adding to global inequality and poverty because three-quarters of the world's poorest people live in poorer countries and depend directly or indirectly on agriculture for their main income (World Bank 2008). In addition to this external policy influence on rural poverty, the governments of many developing countries have directly taxed their farmers over the past half-century. A well-known example is the taxing of exports of plantation crops in post-colonial Africa (Bates). At the same time, many developing countries chose also to overvalue their currency, and to pursue an import-substituting industrialization strategy by restricting imports of manufactures. Together those latter measures indirectly taxed producers of other tradable products in developing economies, by far the most numerous of them being farmers (Krueger, Schiff and Valdés 1988; 1991). Thus the price incentives facing farmers in many developing countries have been depressed by both own-country and other countries' agricultural price and international trade policies. 4 This disarray in world agriculture, as Johnson described it in the title of his seminal 1973 book, means there has been over-production of farm products in high-income countries and under-production in low-income countries. It also means there has been less international trade in farm products than would be the case under free trade, thereby thinning markets for these weather-dependent products and thus making them more volatile. Using a stochastic model of world food markets, Tyers and Anderson (table 6.14) found that instability of international food prices in the early 1980s was three times greater than it would have been under free trade in those products. During the past 25 years, however, numerous countries have begun to reform their agricultural price and trade policies, typically as part of a broader reform agenda. That has raised the extent to which farm products are traded internationally, but not nearly as fast as globalization has proceeded in the non-farm sectors of the world's economies.1 To what extent have reforms of the past two decades reversed the above-mentioned policy developments of the previous three decades? Empirical indicators of agricultural price distortions (called Producer Support and Consumer Subsidy Estimates, or PSEs and CSEs) have been provided in a consistent way since 1986 by the Secretariat of the OECD (2008) for its 30 member countries. However, until now there have been no comprehensive time series rates of assistance to producers of nonagricultural goods to compare with those PSEs, nor do they tell us what happened in those advanced economies in earlier decades ­ which are of more immediate relevance if we are to see how the two groups of countries' policies developed during similar stages of development. As for developing countries, almost no comparable time series estimates have been generated since the Krueger, Schiff and Valdes (1988) study, which covered the period 1960-1984 for just 17 developing countries plus Portugal.2 An exception is a recent set of estimates of nominal rates of protection generated by the International Food Policy Research Institute (IFPRI) for key farm products in China, India, Indonesia and Vietnam since 1985 (Orden et al.). The OECD (2009) also has released PSEs for Brazil, China and South Africa as well as several other East European countries. 1 In the two decades to 2000-04, the value of global exports as a share of GDP rose from 19 to 26 percent, even though most of GDP is nontradable governmental and other services, while the share of primary agricultural production exported globally, including intra-European Union trade, rose from only 13 percent to just 16 percent (World Bank 2007 and FAO, as summarized in Sandri, Valenzuela and Anderson). 2 A nine-year update for the Latin American countries in the Krueger, Schiff and Valdés sample by the same country authors, and a comparable study of seven central and eastern European countries, contain estimates at least of direct agricultural distortions (see Valdés 1996; 2000). The Krueger, Schiff and Valdés (1991) chapters on Ghana and Sri Lanka have protection estimates back to 1955, as does the study by Anderson, Hayami and Others for Korea and Taiwan (and Japan, and much earlier in the case of rice). 5 The World Bank's new Database of Agricultural Distortions (Anderson and Valenzuela) complements and extends those efforts by OECD and IFPRI and the seminal K/S/V study. It builds on them by providing similar estimates for other significant (including many low-income) developing economies, by developing and estimating new, more comprehensive policy indicators, and by providing estimates of nominal rates of assistance (NRAs) for non-agricultural tradables to compare with those for the farm sector.3 This new database includes estimates for 75 countries that together account for between 90 and 96 percent of the world's population, farmers, agricultural GDP and total GDP (table 1). The sample countries also account for more than 85 percent of farm production and employment in each of Africa, Asia, Latin America and the transition economies of Europe and Central Asia, and their spectrum of per capita incomes ranges from the poorest (Zimbabwe and Ethiopia) to among the richest (Norway).4 NRAs and consumer tax equivalents (CTEs) are estimated for more than 70 different farm products, with an average of nearly a dozen per country. In aggregate the coverage represents around 70 percent of the gross value of agricultural production in the focus countries,5 and just under two-thirds of global farm production valued at undistorted prices over the period covered. Not all countries had data for the entire 1955-2007 period, but the average number of years covered is 41 per country.6 Of the world's 30 most valuable agricultural products, the NRAs cover 77 percent of global output, ranging from two-thirds for livestock, three-quarters for oilseeds and tropical crops, and five-sixths for grains and tubers. Those products represent an even higher share (85 percent) of global agricultural exports. Having such a comprehensive coverage of countries, products and years offers the prospect of obtaining a reliable picture of both long-term trends 3 These estimates and associated analytical narratives are discussed in far more detail in a global overview volume (Anderson), and the detailed developing country case studies are reported in four regional volumes covering Africa (Anderson and Masters), Asia (Anderson and Martin 2009a), Latin America (Anderson and Valdés) and Europe's transition economies (Anderson and Swinnen). 4 The only countries not well represented in the sample are those in the Middle East and the many small ones, but in total the omitted countries account for less than 4 percent of the global economy (made up of 0.2 percent from each of Sub-Saharan Africa and Asia, 0.9 percent from Latin America, and the rest from the Middle East and North Africa). 5 Had seven key mostly-nontraded food staples (bananas, cassava, millet, plantain, potato, sweet potato and yam) been included for all instead of just some developing countries, their product coverage would have risen from around 70 to 76 percent; and had those staples had an average NRA of zero, they would have brought the weighted average NRA for all covered agriculture in developing countries only about half of one percentage point closer to zero each decade over the sample period (Anderson, table 12.10). 6 By way of comparison, the seminal multi-country study of agricultural pricing policy by Krueger, Schiff and Valdés (1988; 1991) covered an average of 4.3 products for 23 years to the mid-1980s for each of its 18 focus countries that together accounted for 6 percent of global agricultural output; and the producer and consumer support estimates of the OECD (2008) cover 22 years for its 30 countries that account for just over one-quarter of the world's agricultural output valued at undistorted prices. 6 in policies, and annual fluctuations around those trends, for individual countries and commodities as well as for country groups, regions, and the world as a whole. North America and Europe (including the newly acceded eastern members of the European Union (EU)) each account for one-third of global GDP, and the remaining one- third is shared almost equally by developing countries and the other high-income countries. When the focus turns to just agriculture, however, developing countries are responsible for around three-fifths of value added globally, with Asia accounting for over half of that lion's share. The developing countries' majority becomes stronger still in terms of global population and even more so in terms of number of farmers, almost three-quarters of whom are in Asian developing countries. Hence there is a vast range of per capita incomes and agricultural land per capita, and thus agricultural comparative advantages, across the country groups listed in table 1. Asia has had much faster economic growth and export-led industrialization than the rest of the world: since 1980, Asia's per capita GDP has grown at four times, and exports nearly two times, the global averages, and the share of Asia's GDP that is exported is now one-third above that for the rest of the world and for Latin America and far above that for Africa. Asia's GDP per capita is now half as high again as that of our focus African countries, although still only one-third that of Latin America. However, in the earlier half of our time series Asia was poorer than Africa and hence the poorest of the country groups in table 1. By 2000-04 just 12 percent of Asian developing country GDP came from agriculture on average. That contrasts with Africa where the share for our focus countries ranges from 20 to 40 percent, and with Latin America and Europe's transition economies where it is down to 6 percent (and to just 2 percent on average in high-income countries). The share of employment in agriculture remains very high in Asia though, at just under 60 percent ­ which is the same as in Africa and three times the share in Latin America and Eastern Europe, although more farmers work part-time on their farms in Asia than in other developing countries so these data understate the productivity of labor on Asian farms. By contrast, less than 4 percent of workers in high-income countries are still engaged in agriculture (Sandri, Valenzuela and Anderson 2007). Hence both own-country and rest-of-world distortions to agricultural incentives are of great importance to not only African but also Asian developing country welfare, inequality and poverty. 7 Methodology for Measuring Price Distortions7 The study's methodology focuses mainly on government-imposed distortions that create a gap between a country's domestic prices and the prices of like tradable products at the country's border (or, in the case of nontradable farm products, what they would be in the absence of domestic price subsidies or taxes). Since it is not possible to understand the characteristics of agricultural development with a sectoral view alone, not only are the effects of direct agricultural policy measures (including distortions in the foreign exchange market) examined, but also those of distortions in non-agricultural tradable sectors. Specifically, the NRA for each farm product is computed as the percentage by which government policies have raised gross returns to farmers above what they would be without the government's intervention (or lowered them, if NRA<0).8 Included are estimates of the output-price equivalent of product-specific input subsidies. A weighted average NRA for all covered products for a country is derived using the value of production at undistorted prices as product weights. Those production weights are also used to obtain weighted average NRAs across countries for individual or sets of products. While most of the focus is on agricultural producers, we also consider the extent to which consumers are taxed or subsidized. To do so, we calculate a CTE by comparing the price that consumers pay for their food and the international price of each food product at the border. Differences between the NRA and the CTE arise from distortions in the domestic economy that are caused by transfer policies and taxes/subsidies that cause the prices paid by 7 Only a brief summary of the methodology is provided here. For details see Anderson et al. or Anderson (Appendix A). 8 Such a distortion creates an economic cost to society which can be estimated using welfare measures techniques such as those pioneered by Harberger, who notes that this focus allows a great simplification in evaluating the marginal costs of a set of distortions: changes in economic costs can be evaluated taking into account the changes in volumes directly affected by such distortions, ignoring all other changes in prices. In the absence of divergences such as externalities, the measure of a distortion is the gap between the price paid and the price received, irrespective of whether the level of these prices is affected by the distortion. Other developments that change incentives facing producers and consumers can include flow-on consequences of the distortion, but these should not be confused with the direct price distortion estimated here. If, for instance, a country is large in world trade for a given commodity, imposition of an export tax may raise the price in international markets, reducing the adverse impact of the distortion on producers in the taxing country. Another flow-on consequence is the effect of trade distortions on the real exchange rate, which is the price of traded goods relative to non-traded goods. Neither of these flow-on effects are of immediate concern, however, because if the direct distortions are accurately estimated, they can be incorporated as price wedges into an appropriate country or global economy-wide computable general equilibrium (CGE) model which in turn will be able to capture the full general equilibrium impacts (inclusive of real exchange rate effects) of the various direct distortions to producer and consumer prices. Such price wedges are provided for 2004 by Valenzuela and Anderson, and are used in a global CGE model by Valenzuela, van der Mensbrugghe and Anderson (results from which are summarized below). 8 consumers (adjusted to the farmgate level) to differ from those received by producers. In the absence of any other information, the CTE for each tradable covered farm product is assumed to be the same as the NRA from border distortions, and the CTE for nontradable farm products is assumed to be zero. The value of consumption at undistorted prices is used to obtain product weights to generate weighted average CTEs across products or countries.9 To the NRA for covered products is added a `guesstimate' of the NRA for non- covered products (on average around 30 pecent of the total) and an estimate of the NRA from non-product-specific forms of assistance or taxation, including for inputs.10 Since the 1980s some high-income countries' governments have also provided so-called `decoupled' assistance to farmers but, because that support in principle does not distort resource allocation, its NRA has been computed separately and is not included for direct comparison with the NRAs for other sectors or for developing countries. Each farm industry is classified either as import-competing, or a producer of exportables, or as producing a nontradable (with its status sometimes changing over the years), so as to generate for each year the weighted average NRAs for the two different groups of tradable farm products. We also generate a production-weighted average NRA for nonagricultural tradables, for comparison with that for agricultural tradables via the calculation of a percentage Relative Rate of Assistance (RRA), defined as: (1) RRA = 100*[(100+NRAagt)/(100+NRAnonagt)-1] where NRAagt and NRAnonagt are the percentage NRAs for the tradables parts of the agricultural (including non-covered) and non-agricultural sectors, respectively.11 Since the NRA cannot be less than -100 percent if producers are to earn anything, neither can the RRA (since the weighted average NRAnonagt is non-negative in all our country case studies). And 9 Again this is valued at the farmgate level, following the OECD's approach. The CTE so estimated is probably larger in ad valorem terms than it would be had it been estimated at the retail level, but may be smaller in dollar terms depending on the nature of markups along the value chain. 10 Not all country authors were able to estimate all farm input subsidies, and, following the OECD, no authors included subsidies to water use. India has relatively large input subsidies, but even there they added only a few percentage points to the NRA. In some cases input subsidies would have been more or less than offset by restrictions on imports of intermediate inputs (or by imperfect domestic competition in their provision, e.g. by para-statal monopolies). Hence their fuller estimation for other countries is unlikely to have made much difference to the aggregate NRA for developing country agriculture, particularly in poorer countries where only the wealthiest farmers are major users of modern inputs. Had the focus been on effective rates of assistance to value added, the extent of rate under-estimation could have been greater, but that is irrelevant for the present purpose where the focus is on nominal rates. 11 Farmers are affected not just by prices of their own products but also by the incentives nonagricultural producers face. That is, it is relative prices and hence relative rates of government assistance that affect producer incentives. More than seventy years ago Lerner provided his Symmetry Theorem that proved that in a two-sector economy, an import tax has the same effect as an export tax. This carries over to a model that also includes a third sector producing only nontradables (Vousden, pp. 46-47). 9 if both of those sectors are equally assisted, the RRA is zero. This measure is useful in that if it is below (above) zero, it provides an internationally comparable indication of the extent to which a country's sectoral policy regime has an anti- (pro-)agricultural bias. This approach is not well suited to analysis of the policies of Europe's or Asia's former socialist economies prior to their reform era, because prices then played only an accounting function and currency exchange rates were enormously distorted. During their reform era, however, the price comparison approach provides as valuable a set of indicators for them as for other market economies of distortions to incentives for farm production, consumption and trade, and of the income transfers associated with interventions.12 In addition to the mean NRA, a measure of the dispersion or variability of the NRA estimates across the covered farm products also is generated for each economy. The cost of government policy distortions to incentives in terms of resource misallocation tend to be greater the greater the degree of substitution in production. In the case of agriculture which involves the use of farm land that is sector-specific but transferable among farm activities, the greater the variation of NRAs across industries within the sector then the higher will be the welfare cost of those market interventions. A simple indicator of dispersion is the standard deviation of the covered industries' NRAs. Anderson and Neary show that it is possible to develop a single index that captures the extent to which the mean and standard deviation of protection together contribute to the welfare cost of distortionary policies. That index recognizes that the welfare cost of a government-imposed price distortion is related to the square of the price wedge, and so is larger than the mean and is positive regardless of whether the government's agricultural policy is favoring or hurting farmers. In the case where it is only import restrictions that are distorting agricultural prices, the index provides a percentage tariff equivalent which, if applied uniformly to all imports, would generate the same welfare cost as the actual intra- sectoral structure of protection from import competition. Lloyd, Croser and Anderson show that, once NRAs and CTEs have been calculated, they can be used to generate such an index even in the more complex situation where there may be domestic producer or consumer taxes or subsidies in addition to not only import tariffs but any other trade taxes or subsidies or quantitative restrictions. They call it a Welfare Reduction Index (WRI). Such a measure is the percentage agricultural trade tax (or uniform NRA and CTE) which, if applied equally to all agricultural tradables, would generate the same reduction in national economic welfare as the 12 Data availability also affects the year from which NRAs can be computed. For Europe's transition economies that starting date is 1992 (2000 for Kazahkstan), for Vietnam it is 1986 and for China it is 1981. 10 actual intra-sectoral structure of distortions to domestic prices of tradable farm goods. They show also that, if one is willing to assume that domestic price elasticities of supply (demand) are equal across farm commodities, then the only information needed to estimate the WRI, in addition to the NRAs and CTEs, is the share of each commodity in the domestic value of farm production (consumption) at undistorted prices. To obtain dollar values of farmer assistance and consumer taxation, we have taken the country authors' NRA estimates and multiplied them by the gross value of production at undistorted prices to obtain an estimate in US dollars of the direct gross subsidy equivalent of assistance to farmers (GSE). These GSE values are calculated in constant dollars, and are also expressed on a per-farm-worker basis. Likewise a value of the consumer transfer is derived from the CTE, by assuming the consumption value is the gross value of production at undistorted prices divided by the self-sufficiency ratio for each product (production divided by consumption, derived from national volume data or the Food and Agriculture Organization's (FAO's) commodity balance sheets). These transfer values are helpful for generating an estimate of the contribution of each policy instrument to the overall NRA, and the trade data that provide the self-sufficiency ratio helped each country author attach a trade status to each product each year (bearing in mind also the likely impact of the NRAs and CTEs on the observed self-sufficiency ratio). Once each farm industry is classified either as import-competing, or a producer of exportables, or as producing a non-tradable, it is possible to generate for a given year the weighted average NRAs for the two different groups of tradable farm industries. They can then be used to generate an agricultural trade bias index defined as: 1 NRAag x ( 2) TBI 1 1 NRAag m where NRAag m and NRAag x are the average NRAs for the import-competing and exportable parts of the agricultural sector (their weighted average being NRAag t ). This index has a value of zero when the import-competing and export sub-sectors are equally assisted, and its lower bound approaches -1 in the most extreme case of an anti-trade policy bias. Part of the anti-trade bias in developing countries in the past was the result of government intervention in the domestic market for foreign currency. The most common arrangement was a dual exchange rate, whereby exporters had to sell part or all of their foreign currency to the government at a low price. This effectively taxed and thus 11 discouraged production of exportables. At the same time it created an artificial shortage of foreign currency so that potential importers bid up its purchase price, which had the same effect as an import tax and thus encouraged import-competing production (Dervis, de Melo and Robinson). The size of these effective if implicit trade taxes depends on the extent to which the government purchase price differs from what would be the free-market equilibrium price, the price elasticities of demand for and supply of foreign currency, and the retention rate (the extent of the requirement to sell a portion to the government). In some countries there were more-complex multiple exchange rates, whereby traders of some products were subject to more favorable treatment than others. In estimating NRAs in developing countries, participants in the Agricultural Distortions project endeavored to include the effects of these implicit trade taxes, and to show how much impact they had on the NRAs and RRA. The practice was rife in newly independent developing countries in the 1960s and 1970s, but was gradually phased out over the 1980s and early 1990s as part of overall macroeconomic policy reform initiatives.13 Anderson and Neary also show that it is possible to develop a single index that captures the extent to which import protection reduces the volume of trade. Once NRAs and CTEs have been calculated, Lloyd, Croser and Anderson show how they can be used to generate a more-general Trade Reduction Index (TRI) that allows for the trade effects also of domestic price-distorting policies, and regardless of whether they (or the trade measures) are positive or negative. Such a measure is the percentage agricultural trade tax (or uniform NRA and CTE) which, if applied equally to all agricultural tradables, would generate the same reduction in sectoral trade volume as the actual intra-sectoral structure of distortions to domestic prices of tradable farm goods. They show also that, if the domestic price elasticities of supply (demand) are equal across farm commodities, then the only information needed to estimate the TRI, in addition to the NRAs and CTEs, is the share of each commodity in the domestic value of farm production (consumption) at undistorted prices. 13 There were many other reasons for exchange rate movements which affected the international competitiveness of farmers, but they also affected producers of other tradable products. Where they are due to the actions of international borrowers and lenders (as in the 1990s in Latin America ­ see Quiros and Opazo), they cannot be interpreted as price distortions in the same way as captured in NRA estimation. Where misalignment arises because of government macroeconomic policy, such as a delay in the adjustment of a fixed exchange rate when the government increases its borrowing from abroad, symmetric treatment of any such "overvaluation" during a heavy borrowing period would require taking into account exchange rate "undervaluation" during periods of low foreign borrowing or repayment of foreign debt. For these reasons, we do not follow K/S/V in including deviations of real exchange rates from benchmark values unless these deviations arise from direct distortions as with dual or multiple exchange rates. 12 Needless to say, there are numerous challenges in applying the above methodology, especially in less developed economies with poor-quality data. Ways to deal with the standard challenges are detailed in Anderson et al. and the country-specific challenges are discussed in the analytical narratives in the regional and global volumes listed in footnote 3 above. The NRAs and CTEs are similar to the PSEs and CSEs computed by OECD (2008), except that each of the OECD's measures is expressed as a percentage of the distorted rather than the undistorted price. Thus, it is lower than the comparable NRA or CTE, and has a maximum value of 100 percent. The OECD does not attempt to estimate rates of distortion to prices of non-covered farm products, thereby implicitly assuming they are the same as the average for the roughly 70 percent of farm production that is covered by direct price comparisons. Nor does the OECD take into account distortions to non-farm sectors or to the market for foreign exchange, and it does not estimate indicators such as the WRI and TRI. The estimates by Krueger, Schiff and Valdés (1988; 1991) distinguish for each country a `direct' and `indirect' rate of `protection' as measures of distortion to agricultural incentives. Their `direct' rate is not identical to the agricultural NRA described above, but is the closest for comparative purposes.14 Their `indirect' rate is the number of percentage points by which the `direct' rate for each product, or the production-weighted average for a country's covered farm products, should be reduced because of the adverse macroeconomic influence on farmer incentives of that country's non-farm policies (most notably protection to the manufacturing sector and overvaluation of the country's currency). Their `total protection' rate, therefore, is not identical to the above RRA, but again it is the closest for comparative purposes. It is not identical to the RRA partly because the K/S/V measure is an attempt to estimate econometrically the indirect effect on farm distortions of those non-farm policies, whereas the RRA explicitly uses an estimate of the NRA for non-farm tradable sectors alongside the estimated NRA for the tradable farm sector and both of those NRAs explicitly incorporate an estimate of the trade-taxing effect of multiple exchange rates. Since there are now plenty of sectoral and economy wide models of national and global markets available, the study summarized in Anderson leaves it to modelers to determine how much the estimated domestic price distortions influence a country's real exchange rate and the international relative price of farm products (see footnote 8 above). The other important 14 Formally, K/S/V's `direct protection' measure is the ratio of (a) the difference between the relative producer price and the relative border price and (b) the relative adjusted border price measured at the equilibrium exchange rate and in the absense of all trade policies, where the `relative price' refers to the price of the farm product relative to the price of all non-farm products. 13 differences between the K/S/V study and that summarized in Anderson are that the former's product sample is smaller, its country sample is smaller (in particular, it omits the biggest developing countries of China, India and Indonesia), and it provides only unweighted averages of distortions to farmer incentives across its developing countries. Distortions to Agricultural Incentives in Developing Countries before 1985 We turn now to summarizing first the stylized facts that emerged from K/S/V for the period to 1984 as compared with the findings of Anderson and Valenzuela's (2008) compilation and aggregation of NRAs and related indicators. We begin by focusing on just those developing countries included in the K/S/V sample, and then show how much the new estimates for the fuller sample of 41 developing countries differ from those for the more-limited K/S/V sample of 17 countries. We leave until the following two sections a discussion of the estimates for more-advanced economies pre-1985, and of the period since the K/S/V era for both sets of countries. The key empirical findings from the study by Krueger, Schiff and Valdés (1988; 1991) and their authors' detailed country case studies are based on the estimates shown in table 2, for 4 groups of countries classified according to their level of national per capita income at the time. The most important findings, based on the unweighted average estimates across developing countries for the entire period from 1960 to 1984 roughly (see exact years for each country in note b of table 2), are: The direct rate of assistance (DRA) to farmers, due to agricultural policies, was negative (average DRA of -8 percent), but tended to be more negative the lower a country's per capita income (as low as -23 percent for the lowest-income group, but +24 percent for the highest-income group); Even more important were non-agricultural policies, particularly manufacturing protection, which on average were three times as harmful to farmers as agricultural policies; 14 Thus direct plus indirect policy influences mean that developing country farmers faced prices that were an estimated 30 percent below what they would have been without distortionary farm price, trade and exchange rate policies on average; and Within the agricultural sector, the producers of exportables tended to be taxed by agricultural policies (average DRA of -12 percent) and those producing import- competing farm products tended to be protected (average DRA of 16 percent) but, when the indirect impact is included (which reduces the DRA by 23 percentage points), the TRA for both sub-sectors were negative on average for the 17 countries and for all but Group IV countries (South Korea and Portugal). For the K/S/V sample of 17 developing countries and the period 1960-84, there was therefore a severe anti-trade bias in farm policies and also a severe anti-agricultural bias that was reinforced by non-farm policies, particularly manufacturing protection; and the biases against farmers ­ especially those capable of exporting ­ tended to be greater the lower the national per capita income. The unweighted NRAs and RRAs in the recent World Bank project, for the same developing countries and years, are shown in part (b) of table 2. The estimates of agricultural NRAs, however, include a bigger sample of covered products (more than twice as many as in K/S/V including livestock products which were mostly ignored in K/S/V). The total agricultural NRAs (but not the NRAs for exporting and import-competing farmers) also include, unlike in K/S/V, non-product-specific assistance and guesstimates of assistance to the roughly 30 percent of the value of farm products that has not been included in the new study's explicit price comparison exercise. Recall too that the NRA estimates incorporate the trade-taxing effects of multiple exchange rates; hence they can be expected to have more of an anti-trade bias than K/S/V's DRA measure. The new NRA and RRA estimates reinforce the conclusions from K/S/V for the period to 1984. Specifically, the new agricultural NRAs are very similar to the comparable DRAs (both averaging -8 percent, and within 4 percentage points for the four income groups). Second, non-agricultural policies were even more important in depressing the new RRA than agricultural policies, with their NRA averaging 34 percent compared with the average NRA for agriculture of -8 percent. Third, the estimated direct plus indirect policy influences on farmers' incentives on average are very similar in the two studies: a TRA of -30 percent by K/S/V, and an RRA of -29 percent from the new study. And fourth, as anticipated (because of the inclusion of the impact of multiple exchange rates), within the agricultural 15 sector the new NRA for producers of exportables is more negative than the DRA (average NRA of -25 compared with a DRA of -12 percent). However, the new NRA for producers of import-competing farm products is lower rather than higher than the DRA except for Group IV countries. This is mainly because of the broader product coverage in the new dataset, plus the estimated presence of import subsidies for some food staples in Zambia. Nevertheless, the trade bias index for each of the four income groups is shown in the final column of table 2 to be more negative based on the new NRAs than on K/S/V's DRAs, such that the average over the 17 countries is one-quarter larger for the new estimates (-0.30 compared with -0.24). The comparison between parts (a) and (b) of table 2 thus suggests the new agricultural NRAs and RRAs are indeed similar in magnitude to the K/S/V's DRA and TRA. That gives us confidence to ask two further questions. One, to be delayed until the next section, is how have distortions in those 17 developing countries changed since the mid-1980s? The other is: How much do those average NRAs and RRAs for just 17 countries to 1984 change when the new database's fuller sample of 41 developing countries is included?15 Table 3 includes 5- year average NRAs by geographic region for the full time series, where it is again apparent that the NRAs tend to be higher, the higher a region's income per capita (indicated in column 1). It is also apparent that the NRA trend over the period 1960 to 1984 was flat for each of the three developing country regions. For developing countries as a whole during 1960-1984, their weighted average NRA was -22 percent, which compares with an unweighted average NRA (and DRA) of -8 percent in the K/S/V sample of countries for most of that period. This inclusion of more developing countries in the sample, including from Sub-Saharan Africa but especially China, suggests K/S/V underestimates the DRA for developing countries. Table 4 shows the NRAs for the farm sector's import-competing and exportable sub- sectors, together with the trade bias index. Again the trend to 1984 in the weighted average NRA for each of the two sub-sectors for the full sample of developing countries is flat. But note that the degree of anti-trade bias in the agricultural NRAs is greater for the full sample than it was for the K/S/V sample of 17 developing countries: the NRA averages for exportables is -44 percent and for importables is 13 percent for the full sample, compared with -25 and 7 percent, respectively, for the sample of just 17 countries. Thus the anti-trade bias index for the full sample is shown in the final column of table 2 to be much greater for the full sample than for the 17 countries: -0.50 compared with -0.30 (or -0.24 according to K/S/V's DRAs). 15 Turkey is not included in the developing country grouping hereafter, but rather with the European transition economies. 16 Table 5, which includes NRAs for non-farm tradable sectors, reveals that for the full sample the RRA too is lower than for the K/S/V sample of 17 countries in the period to 1984. Latin America and Asia had very high rates of manufacturing protection in that period, and they were especially high in China and India which were not included in the K/S/V study. Since those two are large economies, the weighted average NRA for all developing country producers of non-farm tradables is estimated to be 47 percent for the 1960-84 period, generating a weighted average RRA of -49 percent compared with the unweighted average rate of -29 percent for the K/S/V sample (or -30 percent based on K/S/V's TRA). Together these new findings suggest the broad qualitative conclusions drawn from the K/S/V study of two decades ago would not have altered had they included more products and more countries in their sample. However, with a bigger sample they would have been able to stress their policy implications even more forcefully, as the estimated magnitudes of the anti- agricultural and anti-trade bias indicators both would have been larger by two-thirds. Distortions to Agricultural Incentives in High-Income Countries pre-1985 Tables 3 to 5 also show the new project's estimated weighted average NRAs and RRAs for high-income countries, which include all the significant economies of Western Europe plus Australia, Canada, Japan, New Zealand and the United States, from which several points are worth stressing. First, the agricultural NRAs were already more than 20 percent by the latter 1950s, and they doubled over the period to 1984 (dipping only slightly in the mid-1970s when international food prices spiked upwards). This contrasts markedly with the developing country average NRA of below -20 percent in that era. Second, even exporting farmers in high-income countries were assisted, although much less so than import-competing farmers who enjoyed an NRA average that was more than three times that of import-competing farmers in developing countries. And third, with declines in manufacturing protection in high-income countries, their RRA average rose even more than their agricultural NRA average, from 14 percent in 1955-59 to 38 percent in 1980-84 and 51 percent in 1985-89. Together these estimates mean that farmers in developing countries were harmed in the K/S/V era not only by their own countries' agricultural and non-farm policies but also ­ and increasingly from the latter 1950s to the latter 1980s ­ by competition in world markets from high-income countries that was enhanced by those countries' pro-agricultural policies. 17 Distortions to Agricultural Incentives since 1985 The bottom panel of table 2 provides NRA and RRA estimates post-1984 for the 17 countries in the K/S/V sample. Comparing them with the middle panel reveals that those countries reduced substantially their taxation of export agriculture, raised their protection of import- competing agriculture and as a result their overall agricultural NRA switched from an average of -8 percent in 1960-84 to 10 percent in 1985-2004. Meanwhile, the NRA for non- farm tradables fell by two-thirds, such that the RRA for this sample rose from -29 percent to 1 percent. The anti-agricultural bias in those 17 developing countries thus disappeared on average, although the anti-trade bias within their farm sectors increased slightly (trade bias index rose from -0.30 to -0.38). These broad findings are true also for the bigger sample of 41 developing countries (with the exception of the anti-trade bias which diminished), even though the magnitudes are generally larger ­ see final row of middle and bottom panels of table 2. To focus on just the covered farm products for which direct price comparisons have been made, figure 1 summarizes the trends in NRAs and reveals a marked difference in the levels of support to import-competing versus exportable farm products. Exportables in developing countries were taxed heavily from the late 1950s until the mid-1980s but then that taxation was gradually phased out (although some taxes remained in 2000-04, for example in Argentina). Importables, by contrast, have been assisted increasingly throughout the past five decades in developing countries on average (even though some import subsidization of staple foods occurred from time to time in low-income countries), and the long-run fitted trend line has almost the same slope for developing countries as for high-income countries (compare the upper and lower graphs in figure 1).16 The net effect of all the explicit and implicit trade taxes and subsidies, together with domestic taxes and subsidies on tradable farm products, is that the NRA for exportable farm products is typically well below the NRA for importables, so that the trade bias index, as defined in the methodology section above, is negative. Table 4 shows that the agricultural 16 Both developing and high-income country NRAs for import-competing farm products rise in the latter 1980s. This is because of the slump in international food prices in 1986 that was far from fully transmitted to domestic markets, and so lowered the denominator far more than the numerator of the NRA calculation. 18 trade bias index has steadily become less negative since the late 1980s for the developing country group, but mainly because of the decline in agricultural export taxation and in spite of growth in agricultural import protection. The two sub-sectors to which that trade bias index's NRAs refer (exportable and import-competing farm products, respectively) are not equal contributors to overall farm production, however, so the TBI when weighted across numerous products/countries is not a perfect indicator. It also ignores distortions to consumer prices which need not be identical to producer distortions. A superior indicator is the trade reduction index discussed in the methodology section above. The TRI associated with NRAs and CTEs for covered agricultural products has fallen substantially from its peak in the mid-1980s for Africa and Asia, as it has for high-income countries (figure 2). That is, the considerable extent of decline in the anti-trade bias in farm policies indicated by the trade bias index is confirmed by the TRI measure. The fall in the TRI has been more because of the fall in national mean NRAs than in their variance, however. The regional average NRAs hide a great deal of diversity across products and countries, including within each region. One way of summarizing the within- country NRA diversity across products is to calculate the standard deviation around the mean NRA for all covered farm products each year. Even when that is averaged over whole geographic regions, the diversity is still evident, and it has not declined much since the K/S/V era for Africa and Latin America (from 34 and 49 percent in 1965-1984, respectively, to 29 and 40 percent in 1985-2004) and it has risen for Asian developing countries (from 50 to 61 percent ­ see Anderson, table 1.6). This has important welfare implications, because the cost of government policy distortions to incentives in terms of resource misallocation tend to be greater the greater the degree of substitution in production (Lloyd), which is high in the case of agriculture where farm land is sector-specific but transferable among farm activities. The increase in the RRA for developing countries began slowly in the 1970s but accelerated over the 1980s and 1990s. Indeed the RRA was slightly above zero by the end of the 20th century (table 5 and figure 3). That is, the removal of the anti-agricultural bias in developing countries has been a gradual process, but it is nonetheless remarkable that in just the one generation since the K/S/V era that bias has all but disappeared except in Africa. Slightly over half of the rise in the RRA for developing countries since the mid-1980s is due to falls in protection to producers of non-farm tradable goods, suggesting that much of the reduction in relative prices faced by farmers over the past two decades can be attributed to general trade liberalization rather than to farm-specific policy reform. 19 Governments in the past tried to alter not only the trend level of farm prices but also to reduce their year-to-year fluctuations. Typically this was done by varying the restrictions on international trade according to seasonal conditions domestically and changes in prices internationally. Effectively this involves exporting domestic instability and not importing instability from abroad. When many countries indulge in such insulating behavior it `thins' international markets for farm products, making them more volatile and thereby encouraging even more countries to insulate. To see how much that type of intervention has changed since the K/S/V era, table 6 reports the average across focus countries of the percentage point deviation each year of national NRAs for 12 key farm products around their trend value, for the sub-periods before and from 1985. For the majority of products that indicator is lower in the latter period, in both developing and high-income countries.17 This is yet another way in which distortions to agricultural incentives for developing countries have diminished since the mid-1980s ­ but note the important exceptions of rice and wheat in table 6. How has the importance of different policy instruments changed since the K/S/V era? Traditionally in developing countries, trade measures at the border (export and import taxes or subsidies and their equivalent from quantitative trade restrictions and multiple exchange rates) have been the dominant forms of intervention. Table 7 shows the various contributions of different policy measures to the overall estimated NRAs as of 1981-84 and 2000-04. In the earlier period, trade measures accounted for more than three-quarters of the total agricultural NRA for developing (and also high-income) countries. In the latter period, trade measures were much less of a contributor in developing countries, and most of that came from import barriers whereas in the earlier period it came mainly from export barriers. Production taxes have also declined substantially. What is now more important in developing countries, both relatively and absolutely, are net subsidies to farm inputs and other non-product-specific assistance. The most notable case is India, where large subsidies to fertilizer, water and power for irrigation add several percentage points to India's agricultural NRA (Anderson, Ch. 10). Trade measures are responsible for an even larger share ­ over 90 percent ­ of the distortion to consumer prices of food, since direct domestic consumer subsidies or taxes, as distinct from the indirect ones provided by border measures, are relatively rare (lower half of table 7). The dominance of trade measures in both consumer tax equivalents (CTEs) and 17 That this indicator tends to be much less in developing than high-income countries is mainly a reflection of the fact that the absolute values of the agricultural NRAs tend to be smaller in developing countries (see table 5). 20 NRAs for agricultural products means we should expect those two indicators to be highly correlated. And indeed that is the case: for all focus countries, all covered products and all available years in the panel set, the coefficient of correlation between farm product NRAs and CTEs is 0.93. Finally, how are the above policy reforms reflected in the welfare reduction index? This single indicator captures the partial equilibrium welfare effect of each country's regime of price distortions for covered agricultural products in place at any time (while ignoring non- covered farm products and indirect effects of sectoral and trade policy measures directed at non-agricultural sectors). The WRI measure reflects the welfare cost of agricultural price- distorting policies better than the NRA or CTE because it includes the distortions on both sides of a market and it recognizes that the welfare cost of a government-imposed price distortion is related to the square of the price wedge. It thus captures the disproportionately higher welfare costs of peak levels of assistance or taxation, and is larger than the mean and is positive regardless of whether the government's agricultural policy is favoring or hurting farmers. In this way the WRI goes some way towards indicating what a computable general equilibrium (CGE) can provide in the way of estimates of the welfare effects of the price distortions captured by the product NRA and CTE estimates, while having the advantage of providing an annual time series of this sectoral indicator. The WRI five-year results in figure 4 indicate a fairly constant tendency in developing countries for their covered products' policies to reduce economic welfare from the 1960s to the mid-1980s, but thereafter that indicator nearly halves in the 1990s. This pattern is generated by different policy regimes in the different country groups though, as the WRI has the desirable property of correctly identifying the welfare consequences that result from both positive and negative assistance regimes, and the larger the variance in assistance levels the greater the potential for resources to be used in activities which do not maximize economic welfare and hence the larger the WRI. One consequence is that the WRI for Africa spikes in the mid-1980s ­ in contrast to the NRA, which moves close to zero. The reason is that while Africa was still taxing exportables it had moved (temporarily) from low to very high positive levels of protection for import-competing farm products when international food prices slumped in 1986 (table 4). At the aggregate level African farmers received almost no government assistance then (NRA close to zero), but the welfare cost of its mixture of agricultural policies as a whole was at its highest according to the WRI. Another consequence is that for developing countries its average WRI in the years 1995 to 2004 is around 20 percent even though its average NRA for covered products in those years is close to zero (see 21 figure 1(a)), again reflecting the high dispersion across product NRAs ­ particularly between exportables and import-competing goods ­ in each country. By way of summary of both the WRI and TRI estimates, table 8 provides the mean and the log-linear regression growth rate of each of those indicators for the K/S/V era and for the period since 1984. It shows them separately for the K/S/V countries broken down into their income groups, as well as for all developing countries in the new database and for high- income countries. Several points are worth noticing from this table. First, in terms of the mean WRI and TRI, the 17 K/S/V countries have almost the same values as the fuller sample of 41 developing countries for the period 1960-84, at around 44 and 24 percent, respectively. The two samples differed in terms of growth in those indexes over those years, however: the trade- and welfare-reducing effects of policies in the smaller K/S/V sample increased 3 or 4 times faster than in the fuller sample of 41 developing countries, suggesting the more-limited sample would have exaggerated the growth in those indexes for developing countries. Second, the K/S/V sample was not very representative of the fuller developing country sample in the more-recent 1985-2004 period: the mean WRI and TRI are each more than half as large again in the latter period as in the former period for the K/S/V countries, whereas for the fuller sample those means fell by roughly one-fifth. That contrast is clear also in the rates of (negative) growth of the indexes over the latter 25 years, which fell much faster in the full sample than in the sample of just 17 countries. Third, there is a U shape in the mean WRI and TRI values across income groups: they become lower as one moves from the lowest income group to Group II and then Group III but then are highest for Group IV. This is consistent with the decline in the negative agricultural NRA as one goes from Group I through to Group III and then the move to a large positive NRA for Group IV (see middle panel of table 2 above). That U shape is similar in the later period except the means for Group III are lower (its policies are less welfare- and trade- reducing than in the earlier period) and those for Group IV are higher (its policies are more than twice as welfare- and trade-reducing as those of the earlier period). Fourth, by region it is only in Latin America that the trade-reducing aspect of agricultural policies has diminished substantially, and it is more in Africa than in Latin America that the welfare-reducing aspect of agricultural policies has diminished. For Asia both indexes are similar in the two periods, but that hides much diversity of reform experiences within the region, with protection growth in such countries as South Korea offsetting the dramatic reforms in such countries as China. 22 And fifth, the mean WRI and TRI in the earlier period were half as large again for high-income countries as for the 41 developing countries (and 2.5 times larger for the European Union), and that gap became even wider by the more-recent period. This is reflected too in the faster increase in these indexes during the early period and their slower decline (especially for non-EU countries) in the later period. Economy-Wide Effects of Reforms since 1984 and of Remaining Policies It is clear from the above that there has been a great deal of change over the past quarter of a century in policy distortions to agricultural incentives throughout the world, and considerable diversity in the rates and types of change. In addition to the anti-agricultural and anti-trade biases of policies of many developing countries being reduced since the K/S/V era, export subsidies of high-income countries have been cut and some re-instrumentation toward less inefficient and less trade-distorting forms of support, particularly in Western Europe, has begun. However, protection from agricultural import competition has continued to be on an upward trend in developing countries, if one ignores the latter 1980s when the limited transmission of the slump in international food prices to domestic markets led to NRAs spiking upwards. What, then, have been the net economic effects of agricultural price and trade policy changes around the world since the early 1980s? Also, how do the effects on farm incomes and economic welfare in developing countries compare with the effects of those price distortions still in place as of 2004? Valenzuela, van der Mensbrugghe and Anderson use a global economy-wide model (the World Bank's Linkage model ­ see van der Mensbrugghe) to provide a combined retrospective and prospective analysis that seeks to assess how far the world has come, and how far it still has to go, in removing the disarray in world agriculture. It quantifies the impacts both of past reforms and current policies by comparing the effects of the above NRA and CTE distortion estimates for the period 1980-84 with those of 2004. Several key findings from that economy-wide modeling study, summarized in table 8, are worth emphasizing. First, the policy reforms from the early 1980s to the mid-2000s is estimated to have improved global economic welfare by $233 billion per year, and removing the distortions remaining as of 2004 would add another $168 billion per year. This suggests 23 that in a global welfare sense the world moved three-fifths of the way towards global free trade in goods over that quarter century. Second, developing countries benefited proportionately more than high-income economies (1.0 percent compared with 0.7 percent of national income) from those past policy reforms, and would gain nearly twice as much as high-income countries by completing that reform process (an average increase of 0.9 percent compared with 0.5 percent for high- income countries). Of those prospective welfare gains from global liberalization, 70 percent would come from agriculture and food policy reform. This is a striking result given that the shares of agriculture and food in global GDP and global merchandise trade are only 3 and 6 percent, respectively. The contribution of farm and food policy reform to the prospective welfare gain for just developing countries is slightly greater, at 72 percent. Third, the developing countries' share of the world's primary agricultural exports rose from 43 to 55 percent, and its farm output share from 58 to 62 percent, because of those reforms, with rises in nearly all agricultural industries except rice and sugar. Removing remaining goods market distortions would boost their export and output shares to 64 and 65 percent, respectively. Fourth, for developing countries as a group, net farm income (value added in agriculture) is estimated to be 4.9 percent higher than it would have been without the reforms of the past quarter century, which is more than ten times the proportional gain for non- agriculture. If policies remaining in 2004 were removed, net farm incomes in developing countries would rise a further 5.6 percent, compared with just 1.9 percent for non-agricultural value added. As well, returns to unskilled workers in developing countries ­ the majority of whom work on farms ­ would rise more than returns to other productive factors from that liberalization. Why Does This Matter? Where to from Here? The degree of distortions to K/S/V mattered in the late 1980s because policies of many developing countries at that time were harming their economies and especially their farmers. Since farm households were much poorer on average than non-farm households, these policies were not only national welfare-reducing but also contributing to inequality and poverty. The above comparison of K/S/V results and those of the new World Bank study 24 reported in Anderson deepens our understanding of that 1960-1984 period of history and of the subsequent 20 years in the following ways: Had K/S/V had the same broader range of covered products, the larger sample of developing countries and the greater variety of indicators as in the new study, it would not have altered the earlier study's key conclusions but it would have enabled the authors to stress their policy implications even more forcefully, as the estimated magnitudes of the anti-agricultural and anti-trade bias indicators would have been larger by about two-thirds;18 The new measures of distortions to farmer incentives in high-income countries confirm that developing country farmers were also being harmed increasingly by rich-country policies during that period to the mid-1980s; Since the mid-1980s, many developing countries have undertaken national policy reforms that have reduced substantially the inter-sectoral bias against agriculture and, within the farm sector, the anti-trade bias of the past ­ and more so, and at a faster pace, for the fuller sample of developing countries than for the K/S/V sample; Nonetheless, many distortions remain within the agricultural sector even in those countries with RRAs close to zero, and even where import restrictions are the main distortionary measure (suggesting tariffs are far from uniform, not to mention subsidies); In a global welfare sense the world moved three-fifths of the way towards global free trade in goods over the quarter century since the early 1980s which, while impressive and gratifying, means there is still another two-fifths of the way to go before these wasteful policies are finally abandoned; Developing countries have benefited proportionately (as a share of their GDP) more than high-income economies from those policy reforms, and would gain nearly twice as much as high-income countries by completing that reform process (of which 72 percent of those prospective gains to developing countries would come from agriculture and food policy reform); and 18 Indeed even the new results for 41 developing countries may under-state the degree of anti-agricultural bias in policies, since those 41 are larger, richer and less agrarian than the non-focused developing countries. A new study using simple political econometrics suggests the agricultural NRA of the latter group in 2000-04 could be 10 percentage points lower than for the 41 focus countries. Even so, their share of all developing country agriculture is sufficiently small that their inclusion would have lowered the aggregate NRA for developing countries by 10 percentage point, from 9 to 8 percent (Anderson et al., table 2.13). 25 Net farm income (agricultural value added) in developing countries is estimated to be 5 percent higher than they would have been without the reforms of since the early 1980s, which is more than ten times the proportional gain to non-agricultural households, and if policies remaining in 2004 were removed those net farm incomes would rise a further 6 percent. Together, these findings suggest both inequality and poverty could be further alleviated by such reform, given that three-quarters of the world's poor are in farm households in developing countries (World Bank 2008).19 Furthermore, those latter results are from a comparative static economy-wide model and so underestimate the gains by ignoring the dynamic gains that typically accompany market liberalization. Ideally the reform processes of the past quarter century would continue, boosting global economic growth, reducing inequality within and between countries, and alleviating poverty. If the convergence of national RRAs towards zero continues (from below by most developing countries and from above by higher-income countries), there would continue to be a re-location of global farm production (in global share terms) from high-income to developing countries, reversing the policy distortion-driven opposite trend in the quarter century prior to the mid-1980s. Whether international food prices would rise or fall would depend on the relative size of the two groups of countries and which had the larger RRA change (bearing in mind that some export restrictions still remain, including in Argentina). According to the global modeling exercise reported in table 9, the net change in international prices would be very small if all goods market distortions as of 2004 were removed globally. Those results also suggest international markets would be `thicker', so their volatility from year to year would be less, further boosting global food security. That rosy scenario would imply that the early 1960s to the mid-1980s was an aberrant period of welfare-reducing policy divergence (negative and very low RRAs in newly independent developing countries, positive and rising RRAs in most high-income countries) that has given way to growth-enhancing, welfare-improving and inequality- and poverty- reducing reforms during which the two country groups' RRAs, like their NRAs, are converging towards zero. In this view the reforms could be seen as the result of learning from the differing growth experiences of more- and less-open developing economies. 19 A new set of economy-wide national and global modeling studies that uses the NRA agricultural distortion estimates in Anderson finds that removal of those and distortions to other tradable goods markets does indeed lower estimated inequality and poverty. See Anderson, Cockburn and Martin. 26 An alternative interpretation of history is that it is the most recent 25-year period of RRA changes that is aberrant. The RRA declines in high-income countries, according to this alternative view, are associated more with, in the case of the EU, its 1992 Single Market initiative and subsequent EU enlargements than with external reform pressure from other World Trade Organization (WTO) members,20 and with the fact that the high protection rates of the mid-1980s represent a temporary spike above trend caused by the very low international commodity prices then, and conversely for the low rates in 2007-08 reported by the OECD. As for the rise of developing country RRAs in this alternative view, that simply follows the example of higher-income countries and will not stop when those RRAs reach zero. Inspection of the NRAs in figure 1 for exporting and import-competing sub-sectors of developing country agriculture reveals that the convergence of aggregate NRAs to near zero is mainly with respect to the exporting sub-sector. NRAs for import-competing farmers in developing countries, by contrast, are positive and (if one ignores the latter 1980s when international food prices spiked downwards) are trending upwards over time. Moreover, when the RRA is plotted against the log of real per capita income and straight regression lines are estimated for developing and high-income countries, they both slope upward and at the same rate (figure 5). True, the intercept on the vertical axis for the developing countries' trend line is lower than that for high-income countries. Nonetheless, in developing countries there are few signs of a slowdown of the upward trend in agricultural protection from import competition over the time period studied.21 On the contrary, there are numerous signs that developing country governments want to keep open their options to raise agricultural NRAs in the future, particularly via import restrictions. One indicator is the high tariff bindings developing countries committed themselves to following the Uruguay Round: as of 2001, actual applied tariffs on agricultural products averaged less than half the corresponding bound tariffs for developing countries of 48 percent, and less than one-sixth in the case of least-developed countries (Anderson and Martin 2006, table 1.2). Another indicator of agricultural trade reform reluctance is the unwillingness of many developing countries to agree to major cuts in bound agricultural tariffs in the WTO's on-going Doha Round of multilateral trade negotiations. More than that, the current negotiations have 20 See Swinnen. As explained by Josling, the budgetary cost of continuing with the EU's past levels of support would have sky-rocketed following the EU membership expansion eastwards, with little if any of those extra payments going to the traditional lobbyists for the CAP. 21 True, applied tariffs were lowered or suspended as a way of dealing with the international food price spike in 2008, but initial indications are that this, and the food export taxes or quantitative restrictions imposed that year by numerous food-exporting developing countries, lasted only until international prices returned close to their trend levels in 2009 (as happened after the price hike of 1973-74 and the price dip of 1986-87). 27 brought to prominence a new proposal for agricultural protectionism in developing countries. This is based on the notion that agricultural protection is helpful and needed for food security, livelihood security and rural development. This view has succeeded in bringing "Special Products" and a "Special Safeguard Mechanism" into the multilateral trading system's agricultural negotiations, despite the fact that such policies, which would raise domestic food prices in developing countries, may worsen poverty and the food security of the poor (Ivanic and Martin). These two alternative interpretations of history have profoundly different implications for the future. The first suggests that the WTO's Doha Round of multilateral trade negotiations is likely to conclude with substantial cuts to agricultural tariff and subsidy bindings that lock in recent reforms and go close to relegating protectionism in agricultural markets to history. In that case world food price trends would simply depend on whether improvements in farm versus nonfarm technologies could keep pace with the growth in global demand for farm products. That was certainly possible in the 20th century (see Pfaffenzeller, Newbolt and Rayner) but, given the pace of climate change and the recent growth in demand for biofuels, it may be more of a challenge in the 21st century especially if much of the world continues to shun genetically modified food. In particular, the emerging economies of China and India would become more food import-dependent as they continue to rapidly industrialize, should their RRAs cease rising and instead stay at their present near- zero levels. The other interpretation of history ­ one that views as normal a movement from taxing to subsidizing farmers as an economy develops ­ suggests the Doha Round will struggle to reach an ambitious reform outcome in agriculture, and that developing countries will make use of the legal wiggle room they have allowed themselves in their WTO bindings to follow Japan, Korea and Taiwan into higher levels of agricultural protection. In that case international food prices would rise less than in the first scenario, but domestic food prices in developing countries, particularly for importables, would rise relative to international prices. If this is the more realistic interpretation of history, it places much more weight on the role of the economics profession in contuning to expound the virtues of governments keeping out of markets that would otherwise function well. 28 References Anderson, James E. and J. Peter Neary (2005), Measuring the Restrictiveness of International Trade Policy, Cambridge MA: MIT Press. Anderson, Kym (ed.) (2009), Distortions to Agricultural Incentives: A Global Perspective, 1955-2007, London: Palgrave Macmillan and Washington DC: World Bank. Anderson, Kym, John Cockburn and Will Martin (eds.) (2010), Agricultural Price Distortions, Inequality and Poverty, Washington DC: World Bank. Anderson, Kym and Johanna L. Croser (2009), National and Global Agricultural Trade and Welfare Reduction Indexes, 1955 to 2007, Supplementary database at www.worldbank.org/agdistortions. Anderson, Kym, Johanna L. Croser, Damiano Sandri and Ernesto Valenzuela (2010), "Agricultural Distortion Patterns Since the 1950s: What Needs Explaining", Ch. 2 in Anderson, Kym (ed.), The Political Economy of Agricultural Price Distortions, Cambridge and New York: Cambridge University Press (forthcoming). Anderson, Kym, Yujiro Hayami and Others (1986), The Political Economy of Agricultural Protection: East Asia in International Perspective, London: Allen and Unwin. Anderson, Kym, Marianne Kurzweil, Will Martin, Damiano Sandri and Ernesto Valenzuela (2008), `Measuring Distortions to Agricultural Incentives, Revisited', World Trade Review 7(4):1-30. Anderson, Kym and Will Martin (eds.) (2006), Agricultural Trade Reform and the Doha Development Agenda, London: Palgrave Macmillan and Washington DC: World Bank. Anderson, Kym and Will Martin (eds.) (2009), Distortions to Agricultural Incentives in Asia, Washington DC: World Bank. Anderson, Kym and Will Masters (eds.) 2009, Distortions to Agricultural Incentives in Africa, Washington DC: World Bank. Anderson, Kym and Johan Swinnen (eds.) 2008, Distortions to Agricultural Incentives in Europe's Transition Economies, Washington DC: World Bank. Anderson, Kym and Alberto Valdés (eds.) 2008, Distortions to Agricultural Incentives in Latin America, Washington DC: World Bank. Anderson, Kym and Ernesto Valenzuela (2008), Global Estimates of Distortions to Agricultural Incentives, 1955 to 2007, core database at www.worldbank.org/agdistortions 29 Bates, Robert H. (1981), Markets and States in Tropical Africa: The Political Basis of Agricultural Policies, Berkeley: University of California Press. Dervis, Kemal, Jaime de Melo and Sherman Robinson (1981), 'A General Equilibrium Analysis of Foreign Exchange Shortages in a Developing Country', Economic Journal 91: 891-906. FAO (2007), FAOSTAT, available at www.fao.org (accessed June). Haberler, Gottfried (1958), Trends in International Trade: A Report by a Panel of Experts, Geneva: General Agreement on Tariffs and Trade, October. Hayami, Yujiro (2007), `An Emerging Agricultural Problem in High-Performing Asian Economies', Policy Research Working Paper 4312, World Bank, Washington DC. Ivanic, Maros and Will Martin (2008), `Implications of Higher Global Food Prices for Poverty in Low-Income Countries', Agricultural Economics 39: 405-16. Johnson, D. Gale (1973), World Agriculture in Disarray, London: St Martin's Press (revised in 1991). Josling, Timothy (2009), `Western Europe', Ch. 5 in Distortions to Agricultural Incentives: A Global Perspective, 1955-2007, edited by K. Anderson, London: Palgrave Macmillan and Washington DC: World Bank. Krueger, Anne O., Maurice Schiff and Alberto Valdés (1988), `Agricultural Incentives in Developing Countries: Measuring the Effect of Sectoral and Economy-wide Policies', World Bank Economic Review 2(3): 255-72, September. Krueger, Anne O., Maurice Schiff and Alberto Valdés (1991), The Political Economy of Agricultural Pricing Policy, Volume 1: Latin America, Volume 2: Asia, and Volume 3: Africa and the Mediterranean, Baltimore: Johns Hopkins University Press for the World Bank. Lerner, Abba (1936), `The Symmetry Between Import and Export Taxes', Economica 3(11): 306-13, August. Lloyd, Peter J. (1974), `A More General Theory of Price Distortions in an Open Economy', Journal of International Economics 4(4): 365-86, November. Lloyd, Peter J., Johanna L. Croser and Kym Anderson (2009), `Global Distortions to Agricultural Markets: New Indicators of Trade and Welfare Impacts, 1960 to 2007', Policy Research Working Paper 4865, World Bank Washington DC, March. Orden, David, Fuzhi Cheng, Hoa Nguyen, Ulrike Grote, Marcelle Thomas, Kathleen Mullen and Dongsheng Sun (2007), Agricultural Producer Support Estimates for Developing 30 Countries: Measurement Issues and Evidence from India, Indonesia, China and Vietnam, IFPRI Research Report 152, Washington DC: International Food Policy Research Institute. OECD (2008), Producer and Consumer Support Estimates (online database accessed at www.oecd.org for 1986-2007 estimates; and OECD files for estimates using an earlier methodology for 1979-85). OECD (2009), Agricultural Policies in Emerging Economies: Monitoring and Evaluation 2009, Paris: Organization for Economic Co-operation and Development, March. Pfaffenzeller, Stephan, Paul Newbolt, and Anthony Rayner (2007), `A Short Note on Updating the Grilli and Yang Commodity Price Index', World Bank Economic Review 21(1): 151­63. Quiroz, Jorge and Luis Opazo (2000), "The Krueger, Schiff, Valdés Study 10 Years Later: A Latin American Perspective", Economic Development and Cultural Change 49(1): 181-96, October. Sandri, Damiano, Ernesto Valenzuela and Kym Anderson (2007), `Economic and Trade Indicators, 1960 to 2004', Agricultural Distortions Working Paper 02, World Bank, Washington DC. Posted at www.worldbank.org/agdistortions. Schiff, Maurice and Alberto Valdés (1991), The Political Economy of Agricultural Pricing Policy, Volume 4: A Synthesis of the Economics in Developing Countries, Baltimore: Johns Hopkins University Press for the World Bank. Swinnen, Johan F.M. (ed.) (2008), The Perfect Storm: The Political Economy of the Fischler Reforms of the Common Agricultural Policy, Brussels: Centre for European Policy Studies. Tyers, Rodney and Kym Anderson (1992), Disarray in World Food Markets: A Quantitative Assessment, Cambridge and New York: Cambridge University Press. Valdés, Alberto (1996), `Surveillance of Agricultural Price and Trade Policy in Latin America During Major Policy Reforms', World Bank Discussion Paper No. 349, Washington DC, November. Valdés, Alberto (ed.) (2000), `Agricultural Support Policies in Transition Economies', World Bank Technical Paper No. 470, Washington DC, May. Valenzuela, Ernesto and Kym Anderson (2008), "Alternative Agricultural Price Distortions for CGE Analysis of Developing Countries, 2004 and 1980-84", Research Memorandum No. 13, West Lafayette IN: Center for Global Trade Analysis, Purdue 31 University, December. Freely downloadable at https://www.gtap.agecon.purdue.edu/resources/res_display.asp?RecordID=2925 Valenzuela, Ernesto, Dominique van der Mensbrugghe and Kym Anderson (2009), `General Equilibrium Effects of Price Distortions on Global Markets, Farm Incomes and Welfare', Ch. 13 in Distortions to Agricultural Incentives: A Global Perspective, 1955-2007, edited by K. Anderson, London: Palgrave Macmillan and Washington DC: World Bank. van der Mensbrugghe, Dominique (2005), "Linkage Technical Reference Document: Version 6.0." Unpublished paper, World Bank, Washington, DC. http://www.worldbank.org/prospects/linkagemodel Vousden, Niel (1990), The Economics of Trade Protection, Cambridge and New York: Cambridge University Press. World Bank (2007), World Development Indicators, Washington DC: World Bank. World Bank (2008), World Development Report 2008: Agriculture for Development, Washington DC: World Bank. 32 Figure 1: Nominal rates of assistance to exportable, import-competing and all covered agricultural products,a high-income and developing countries, 1955 to 2004 (%) (a) Developing countries 90 70 50 30 10 -10 1965-69 1970-74 1975-79 1980-84 1985-89 1990-94 1995-99 2000-04 -30 -50 Import-competing Exportables Total (b) High-income countries plus Europe's transition economies 90 70 50 30 10 -10 1955-59 1960-64 1965-69 1970-74 1975-79 1980-84 1985-89 1990-94 1995-99 2000-04 -30 -50 Import-competing Exportables Total a Covered products only. The total also includes nontradables. Source: Author's derivation, using data in Anderson and Valenzuela 33 Figure 2: Trade Reduction Indexes for covered tradable farm products, by region, 1960 to 2007 (%) a. Africa, Asia, and Latin America 60 40 20 0 1960-64 1965-69 1970-74 1975-79 1980-84 1985-89 1990-94 1995-99 2000-04 Africa Asia Latin America b. Developing countries, high-income countries, and European transition economies 70 50 30 10 -10 1960-64 1965-69 1970-74 1975-79 1980-84 1985-89 1990-94 1995-99 2000-04 2005-07 Developingcountries Europe'stransition econs. H e igh-incom countries Source: Lloyd, Croser and Anderson 34 Figure 3: Nominal rates of assistance to agricultural and non-agricultural tradable products and relative rate of assistance, all focus developing countries,a 1965 to 2004 (%) 100 80 60 40 percent 20 0 1965-69 1970-74 1975-79 1980-84 1985-89 1990-94 1995-99 2000-04 -20 -40 -60 NRA non-ag tradables NRA ag tradables RRA a Weighted averages across countries, using agricultural production valued at undistorted prices as weights. Source: Author's derivation, using data in Anderson and Valenzuela 35 Figure 4: Welfare Reduction Indexes for covered tradable farm products, by region, 1960 to 2007 (%) (a) Africa, Asia and Latin America 80 60 40 20 0 1960-64 1965-69 1970-74 1975-79 1980-84 1985-89 1990-94 1995-99 2000-04 Africa Asia Latin America (b) Developing countries, high-income countries and Europe's transition economies 100 80 60 40 20 0 1960-64 1965-69 1970-74 1975-79 1980-84 1985-89 1990-94 1995-99 2000-04 2005-07 Developing countries Europe's transition econs. High-income countriesb Source: Lloyd, Croser and Anderson 36 Figure 5: Relationships between real GDP per capita and RRA, all 75 focus countries, 1955 to 2007 400 Relative Rate of Assistance (%) 300 200 100 0 -100 -1 0 1 2 3 Ln real GDP per capita HIC RRA obs HIC fitted values DC RRA obs DC fitted values Coefficient Standard error R2 DCs 0.26 0.02 0.17 HICs 0.28 0.03 0.14 Source: Author's derivation with country fixed effects, using data in Anderson and Valenzuela 37 Table 1: Summary of NRA/CTE/RRA coverage statistics, World Bank agricultural distortions project Number and size of countries Number % of 2000-04 global: Popn. Ag GDP Africa 21 11 7 Asia 12 51 37 Latin America 8 7 8 SUB-TOTAL, all DCs 41 69 52 European transition econs 14 7 7 High-income countries 20 14 33 TOTAL 75 92 92 Number of years covered Maximum Av. per country Africa 51 43 Asia 53 42 Latin America 51 39 SUB-TOTAL, all DCs 53 43 European transition econs 47 17 High-income countries 53 52 TOTAL 51 41 Number of products covered Maximum Av. per country Africa 44 8 Asia 35 8 Latin America 27 10 SUB-TOTAL, all DCs 59 9 European transition econs 25 12 High-income countries 39 15 TOTAL 74 11 Total number of NRA ests. (years and products) Total Av. per country Africa 7318 348 Asia 3546 296 Latin America 2881 360 SUB-TOTAL, focus DCs 13745 335 European transition econs 2847 203 High-income countries 13377 669 TOTAL, focus countries 29969 400 Source: Author's derivation based on data in Anderson and Valenzuela 38 Table 2: Estimates by K/S/V of the direct, indirect and total rates of assistancea to farmers in 17 developing countries, by income group,b and comparable NRA and RRA estimates, 1960-1984 and 1985-2004 (%, unweighted averages across countries) (a) K/S/V estimates, circa 1960-84 Direct rate of assistance to farmers Indirect Total Trade rate of rate of bias Import- Exportable All farm Income group assistance assist- indexe (poorest first): competing products incl. to ance nontradables farmersc (TRA) (DRA) Group 1 18 -21 -23 -29 (-26) -52 -0.33 Group II 10 -16 -12 -25 (-35) -37 -0.24 Group III 14 2 -0 -16 (-23) -16 -0.11 Group IV 28 1 24 -14 (-14) 10 -0.21 All 17 countries 16 -12 -8 -23 (-29) -30 -0.24 (b)Anderson and Valenzuela estimates (circa 1960-84) Nominal rate of assistance to farmers Nominal Relative Trade rate of rate of bias Import- Exportable Total (incl. Income group assistance, assist- indexe (poorest first): competing nontradables)d non-agric ance Group 1 -16 -50 -22 14 -44 -0.40 Group II 4 -26 -13 49 -38 -0.29 Group III 12 -7 -4 19 -21 -0.14 Group IV 40 0 26 13 18 -0.29 All 17 countries 7 -25 -8 34 -29 -0.30 All 41 countriesf 13 -44 -22 47 -49 -0.50 (c)Anderson and Valenzuela estimates (1985-2004) Nominal rate of assistance to farmers Nominal Relative Trade rate of rate of bias Import- Exportable Total (incl. Income group assistance, assist- indexe (poorest first): competing nontradables)d non-agric ance Group 1 3 -45 -21 10 -35 -0.47 Group II 30 -10 5 17 -8 -0.31 Group III 38 -5 2 7 -4 -0.31 Group IV 122 7 87 2 101 -0.52 All 17 countries 38 -15 10 12 1 -0.38 All 41 countriesf 26 -16 1 15 -14 -0.33 39 a The three rates of assistance shown here are what Schiff and Valdes call `direct protection', `indirect protection' and `total protection'. Apart from rounding errors, column 3 is the production-weighted average of columns 1 and 2 and an unreported direct rate of assistance for nontradable farm products, and column 5 is the sum of columns 3 and 4. b Group 1 is Cote d'Ivoire (1960-82), Ghana (1955-77) and Zambia (1966-84); Group II is Argentina (1960-84), Colombia (1960-83), Dominican Rep. (1966-85), Egypt (1964-84), Pakistan (1960-86), Philippines (1960-86), Sri Lanka (1960-85), Thailand (1962-84) and Turkey (1961-83); Group III is Brazil (1969-83), Chile (1960-83) and Malaysia (1960-83); and Group IV is Rep. of Korea (1960-84) and Portugal (1960-84). In the full K/S/V set of countries, Morocco is included in Group II and the total, but its exclusion makes no more than 1 percentage point difference to rows 2 and 5 of the K/S/V unweighted averages. It is therefore excluded here to aid comparison because Morocco was not included in the more- recent study reported in Anderson. c Numbers in parentheses are that due to manufacturing protection, which accounts for most of the indirect rate of assistance. d Includes also non-product-specific assistance to farmers and guesstimated NRAs for non- covered products (neither of which are included in the first two columns). In deriving the RRA, the NRA for just agricultural tradables is used. e Trade Bias Index, TBI = (1+NRAagx/100)/(1+NRAagm/100) ­ 1, where NRAagx and NRAagm are the weighted average percentage NRAs for the exportable and import- competing parts of the agricultural sector shown in columns 1 and 2, with weights based on production valued at undistorted prices; and similarly for part (a) of the table using DRAs. f The full sample of 41 focus developing countries reflected in the subsequent tables. Source: Schiff and Valdés (tables 2.1 and 2.2) and author's derivation from Anderson and Valenzuela. 40 Table 3: Nominal rates of assistance to agriculture,a focus countries, 1955 to 2007c (%) 1955-59 1960-64 1965-69 1970-74 1975-79 1980-84 1985-89 1990-94 1995-99 2000-04 2005-07 Africa (14% of global per capita GDP) -14 -8 -11 -15 -13 -8 -1 -9 -6 -7 na Asia (20% of global per capita GDP) -27 -27 -25 -25 -24 -21 -9 -2 8 12 na Latin America (64% of global per capita GDP) -11 -8 -7 -21 -18 -13 -11 4 6 5 na All developing countries -26 -23 -22 -24 -22 -18 -8 -2 6 9 na Eastern Europe and Central Asiab (48% of global per capita GDP) na na na na na na na 10 18 18 25 High-income countries (540% of global per capita GDP) 22 29 35 25 32 41 53 46 35 32 17 All focus countries (wted. average): 3 5 6 0 2 5 17 18 17 18 na Source: Author's derivation, using data in Anderson and Valenzuela a. Weighted average for each country, including non-product specific assistance as well as authors' guesstimates for non-covered farm products (but not decoupled assistance), with weights based on gross value of agricultural production at undistorted prices. Estimates for China pre-1981 and India pre-1965 are based on the assumption that the nominal rate of assistance to agriculture in those years was the same as the average NRA estimates for those countries for 1981-84 and 1965-69, respectively, and that the gross value of production in those missing years is that which gives the same average share of value of production in total world production in 1981-84 and 1965-69, respectively. Developing country and world aggregates are computed accordingly. b The transition economies of Central and Eastern Europe and the former Soviet Union are not included in the high-income or developing country aggregates. 41 Table 4: Nominal rates of assistance to agricultural exportables, import- competing products, and the trade bias index,a focus regions, 1955 to 2007 (%) 1955- 1960- 1965- 1970- 1975- 1980- 1985- 1990- 1995- 2000- 2005- 59 64 69 74 79 84 89 94 99 04 07 Africa NRA agric. exportables na -30.1 -38.4 -42.6 -42.6 -35.0 -36.7 -35.8 -26.1 -24.6 na NRA agric. imp-comp na 18.6 11.8 1.9 14.5 13.2 58.3 5.2 9.8 1.6 na Trade Bias Index na -0.41 -0.45 -0.44 -0.50 -0.43 -0.60 -0.39 -0.33 -0.26 na Latin America NRA agric. exportables na -20.4 -12.8 -27.0 -25.2 -27.1 -25.0 -10.5 -3.5 -4.6 na NRA agric. imp-comp na 26.3 8.7 -2.8 1.1 13.6 5.1 19.4 12.5 20.6 na Trade Bias Index na -0.37 -0.20 -0.25 -0.26 -0.36 -0.29 -0.25 -0.14 -0.21 na South Asiac NRA agric. exportables na -37.5 -37.2 -30.0 -36.1 -27.9 -20.6 -15.8 -12.0 -6.2 na NRA agric. imp-comp na 39.2 41.2 39.4 45.1 37.9 63.3 25.1 14.5 26.5 na Trade Bias Index na -0.55 -0.56 -0.50 -0.56 -0.48 -0.51 -0.33 -0.23 -0.26 na China and Southeast Asiac NRA agric. exportables na -55.5 -55.1 -51.8 -50.1 -50.0 -41.0 -20.8 -2.2 0.1 na NRA agric. imp-comp na -10.3 -8.9 -9.4 -2.6 0.5 15.1 3.3 13.3 12.3 na Trade Bias Index na -0.50 -0.51 -0.47 -0.49 -0.50 -0.49 -0.23 -0.14 -0.11 na Developing countriesc NRA agric. exportables na -46.5 -44.6 -45.4 -43.9 -41.4 -35.8 -18.7 -5.5 -3.0 na NRA agric. imp-comp na 12.7 13.5 7.8 12.8 16.5 37.7 22.6 22.0 23.0 na Trade Bias Index na -0.53 -0.51 -0.49 -0.50 -0.50 -0.53 -0.34 -0.23 -0.21 na European transition econs. NRA agric. exportables na na na na na na na -3.2 -1.0 -1.0 15.2 NRA agric. imp-comp na na na na na na na 32.5 35.4 35.7 32.3 Trade Bias Index na na na na na na na -0.27 -0.27 -0.27 -0.13 High-income countries NRA agric. exportables 4.2 7.4 13.5 10.3 11.3 12.1 22.3 15.9 8.1 6.9 2.9 NRA agric. imp-comp 31.2 45.9 50.2 36.5 47.4 58.1 71.4 62.4 53.9 50.7 30.8 Trade Bias Index -0.21 -0.26 -0.24 -0.19 -0.24 -0.29 -0.29 -0.29 -0.30 -0.29 -0.21 World c NRA agric. exportables na -23 -20 -23 -25 -24 -17 -7 -1 0 na NRA agric. imp-comp na 35 37 27 34 38 57 43 38 36 na Trade Bias Index na -0.43 -0.42 -0.39 -0.44 -0.45 -0.47 -0.35 -0.28 -0.26 na Source: Author's derivation, using data in Anderson and Valenzuela a. NRAs for non-covered products are included here (unlike in figure 1.3). b. Trade Bias Index, TBI = (1+NRAagx/100)/(1+NRAagm/100) ­ 1, where NRAagx and NRAagm are the weighted average percentage NRAs for the exportable and import-competing parts of the agricultural sector, with weights based on production valued at undistorted prices. TBIs shown here are calculated using the regional 5-year averages of NRAagx and NRAagm. c. Estimates for China pre-1981 and India pre-1965 are based on the assumption that the nominal rate of assistance to agriculture in those years was the same as the average NRA estimates for those countries for 1981-84 and 1965-69, respectively, and that the gross value of production in those missing years is that which gives the same average share of value of production in total world production in 1981-84 and 1965-69, respectively. The developing country and world averages are computed accordingly. 42 Table 5: Nominal rates of assistance to agricultural and nonagricultural tradables, and the RRA,a by region, 1955 to 2007 (%) 1955- 1960- 1965- 1970- 1975- 1980- 1985- 1990- 1995- 2000- 2005- 59 64 69 74 79 84 89 94 99 04 07 Africa NRA agric. na -13.3 -19.6 -25.0 -22.1 -13.5 -0.3 -15.4 -8.7 -12.0 na NRA non-agric. na 3.7 2.7 1.5 5.7 1.6 9.2 2.7 2.0 7.3 na RRA na -15.2 -21.4 -26.0 -25.9 -13.1 -8.3 -17.1 -10.4 -18.0 na Latin America NRA agric. na -11.4 -9.3 -23.0 -19.0 -12.9 -11.2 4.4 5.5 4.9 na NRA non-agric. na 26.9 31.3 27.8 23.3 18.5 16.8 7.3 6.6 5.4 na RRA na -30.2 -30.9 -39.8 -34.2 -26.6 -24.0 -2.7 -1.0 -0.5 na South Asiab NRA agric. na 4.1 4.4 9.7 -7.7 1.8 47.1 0.2 -2.4 12.7 na NRA non-agric. na 114.4 117.8 81.7 57.8 54.6 39.9 18.6 15.0 10.1 na RRA na -51.5 -51.9 -39.8 -41.6 -33.3 5.1 -15.5 -14.9 3.4 na China and Southeast Asiab NRA agric. na -43.6 -42.6 -40.1 -35.7 -34.5 -27.8 -12.0 4.9 7.1 na NRA non-agric. na 36.5 36.5 33.7 30.8 20.6 23.3 19.8 9.6 5.5 na RRA na -58.7 -58.0 -55.2 -50.8 -43.4 -41.6 -26.4 -4.2 1.5 na Developing countriesb NRA agric. na -24.0 -27.3 -31.9 -25.5 -21.0 -15.6 -3.9 4.0 7.4 na NRA non-agric. na 58.3 60.0 45.8 37.3 34.6 27.0 16.7 9.8 6.3 na RRA na -52.0 -54.5 -53.3 -45.8 -41.3 -33.6 -17.6 -5.3 1.1 na European transition econs. NRA agric. na na na na na na na 10.0 18.3 16.1 17.0 NRA non-agric. na na na na na na na 9.8 5.5 4.6 2.7 RRA na na na na na na na 0.1 12.2 11.0 13.9 High-income countries NRA agric. 23.0 30.9 36.8 26.5 34.7 43.0 55.5 48.2 36.6 33.9 18.3 NRA non-agric. 7.5 8.5 7.7 5.4 3.6 3.4 3.2 2.5 1.7 1.3 -0.7 RRA 14.3 20.6 27.1 19.9 30.1 38.3 50.6 44.6 34.3 32.1 19.2 Worldb NRA agric. na 5.6 7.6 0.8 2.6 5.7 18.7 19.7 18.4 18.6 na NRA non-agric. na 19.0 20.5 16.1 13.7 10.0 9.8 7.6 6.0 4.0 na RRA na -11.3 -10.7 -13.2 -9.8 -3.6 8.1 11.3 11.8 14.0 na Source: Author's derivation, using data in Anderson and Valenzuela a. The RRA is defined as 100*[(100+NRAagt)/(100+NRAnonagt)-1], where NRAagt and NRAnonagt are the percentage NRAs for the tradables parts of the agricultural and non-agricultural sectors, respectively. b. Estimates for the RRA for China pre-1981 and India pre-1965 are based on the assumption that the agricultural NRAs in those years were the same as the average NRA estimates for those countries for 1981-84 and 1965-69, respectively, and that the value of production in those missing years is that which gives the same average share of value of production in total world production in 1981-84 and 1965-69, respectively. Developing and world country aggregates are computed accordingly. 43 Table 6: Deviation of national NRA around its trend value,a 12 key covered farm products,b developing and high-income countries, 1965-84 and 1985-2004 (NRA percentage points) Developing countries High-income countries 1965-1984 1985-2004 1965-1984 1985-2004 Grains, oils, sugar Rice 32 64 66 229 Wheat 33 47 80 91 Maize 36 33 53 58 Soybean 46 117 75 61 Sugar 53 66 179 173 Tropical cash crops Cotton 38 33 42 28 Coconut 22 20 na na Coffee 41 27 na na Livestock products Milk 76 69 239 190 Beef 45 52 128 127 Pigmeat 81 60 92 77 Poultry 109 74 164 197 a Deviation is computed as the absolute value of (residual ­ trend NRA) where trend NRA in each of the two sub-periods is obtained by regressing NRA on time. b Unweighted average of national deviations. Source: Authors' derivation, using data in Anderson and Valenzuela 44 Table 7: Contributions to total agricultural NRA and CTE from different policy instruments,a developing and high-income countries, 1981-84 and 2000­04 (%) 1981-84 2000­04 All High- All High- (a) NRA developing income developing income Border measures countries countries countries countries Import tax equivalent 6 34 8 24 Export subsidies 1 2 1 1 Export tax equivalent -20 0 -3 0 Import subsidy equivalent -2 0 -1 0 ALL BORDER MEASURES -15 36 5 25 Domestic measures Production subsidies 1 2 1 1 Production taxes -5 0 -1 0 Net subsidies to farm inputs 1 3 2 2 Non-product-specific assistance 1 1 2 5 ALL PRODUCTION SUPPORTS -2 6 4 8 Decoupled payments to farmers 0 6 0 11 TOTAL NRA (including decoupled) -17 48 9 44 (b) CTE Border measures Import tax equivalent 10 46 10 32 Export subsidies 1 2 1 1 Export tax equivalent -22 0 -2 0 Import subsidy equivalent -3 0 -1 0 ALL BORDER MEASURES -14 48 8 33 Domestic measures Consumption subsidies -1 0 -1 -6 Consumption taxes 0 0 1 0 ALL CONSUMPTION MEASURES -1 0 0 -6 TOTAL CTE -15 48 8 27 a In the absence of data, we assume the share of input tax/subsidy, domestic production tax/subsidy and border tax/subsidies for non-covered farm products is the same as that for covered farm products. The first period begins in 1981 because that was the first year for which estimates for China are available. b All table entries have been generated by dividing the Gross Subsidy Equivalent of all (including decoupled) measures by the total agricultural sector's gross production valued at undistorted prices. Source: Author's derivation, using distortion data in Anderson and Valenzuela 45 Table 8: WRI and TRI averages and growth rates,a developing and high-income countries,b 1960-1984 and 1985-2004 (%) 1960-1984 1985-2004 K/S/V income group Mean (%) Growth rate (% p.a.) Mean (%) Growth rate (% p.a.) (poorest first): WRI TRI WRI TRI WRI TRI WRI TRI Group 1 45 23 0.8 1.1 42 24 -0.7 0.1 Group II 38 12 -0.1 0.0 37 16 -0.4 -0.4 Group III 31 15 0.6 0.5 23 7 -1.8 -0.9 Group IV 63 42 2.2 3.1 155 128 0.6 0.3 17 K/S/V countries 44 23 0.9 1.2 64 44 -0.6 -0.2 All 41 developing countries 44 25 0.2 0.4 34 21 -1.4 -1.4 Africa 43 25 0.5 0.9 33 23 -1.4 -1.6 Asia 51 22 0.0 0.0 51 24 -1.5 -1.5 LAC 35 19 0.3 -0.1 30 11 -1.0 -0.5 All high-income countries 65 32 0.6 0.3 73 37 -1.3 -0.7 EU15 110 55 0.2 0.0 75 42 -2.2 -1.3 Other high-income 33 15 1.0 0.6 72 34 -0.7 -0.3 a. The average annual compound growth rates are the beta coefficients from a regression of the log variable on time for the period shown. In order to obtain a natural logarithm, the WRI indicator used is not as a percentage but rather as a coefficient, defined as (1 + WRI/100); and similarly for the TRI. b. See table 2 for the classification of K/S/V's 17 developing countries in the 4 income groups shown. Source: Author's calculations based on estimates in Anderson and Croser 46 Table 9: Effects of reforming global goods markets between 1980-84 and 2004, and of removing remaining price and trade distortions as of 2004 Reform from Move to free 1980-84 to 2004 trade as of 2004 Global econ welfare, $b (%) $233b (0.8%) $168b (0.6%) DCs' econ welfare, $b (%) $73b (1.0%) $65b (0.9%) DC share of global ag output 58% 62% 62% 65% DC share of global ag exports 43% 55% 55% 64% % rise in DC ag (nonag) VA 4.9% (0.4%) 5.6% (1.9%) % rise in international agricultural 13% <1% and food prices Source: Valenzuela, van der Mensbrugghe and Anderson