WPS5603 Policy Research Working Paper 5603 Would Freeing Up World Trade Reduce Poverty and Inequality? The Vexed Role of Agricultural Distortions Kym Anderson John Cockburn Will Martin The World Bank Development Research Group Agriculture and Rural Development Team March 2011 Policy Research Working Paper 5603 Abstract Trade policy reforms in recent decades have sharply countries. The global LINKAGE model results suggest reduced the distortions that were harming agriculture in that removing those remaining distortions would reduce developing countries, yet global trade in farm products international inequality, largely by boosting net farm continues to be far more distorted than trade in nonfarm incomes and raising real wages for unskilled workers in goods. Those distortions reduce some forms of poverty developing countries, and would reduce the number of and inequality but worsen others, so the net effects poor people worldwide by 3 percent. The analysis based are unclear without empirical modeling. This paper on the Global Trade Analysis Project model for a sample summarizes a series of new economy-wide global and of 15 countries, and nine stand-alone national case national empirical studies that focus on the net effects of studies, all point to larger reductions in poverty, especially the remaining distortions to world merchandise trade on if only the non-poor are subjected to increased income poverty and inequality globally and in various developing taxation to compensate for the loss of trade tax revenue. This paper is a product of the Agriculture and Rural Development Team, Development Research Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The author may be contacted at wmartin1@worldbank.org. 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 Would Freeing Up World Trade Reduce Poverty and Inequality? The Vexed Role of Agricultural Distortions Kym Anderson University of Adelaide and CEPR kym.anderson@delaide.edu.au John Cockburn Laval University and PEP jcoc@ecn.ulaval.ca Will Martin World Bank Wmartin1@worldbank.org JEL codes: D30, D58, D63, F13, O53, Q18 Keywords: Poverty, income inequality, price distortions, farm trade policy Author contact details: Kym Anderson School of Economics University of Adelaide Adelaide SA 5005, Australia Phone +61 8 8303 4712 Fax +61 8 8223 1460 kym.anderson@adelaide.edu.au This is a product of a World Bank research project on Distortions to Agricultural Incentives. The authors are grateful for helpful comments from workshop participants and referees, and for funding from World Bank Trust Funds provided by the governments of the Netherlands (BNPP) and the United Kingdom (DfID) and from the Australian Research Council. The views expressed are the authors' alone and not necessarily those of the World Bank and its Executive Directors, nor the countries they represent, nor of the institutions providing funds for this research project. 2 Abstract Trade policy reforms in recent decades have sharply reduced the distortions that were harming agriculture in developing countries, yet global trade in farm products continues to be far more distorted than trade in nonfarm goods. Those distortions reduce some forms of poverty and inequality but worsen others, so the net effects are unclear without empirical modeling. This paper summarizes a series of new economy-wide global and national empirical studies that focus on the net effects of the remaining distortions to world merchandise trade on poverty and inequality globally and in various developing countries. The global LINKAGE model results suggest that removing those remaining distortions would reduce international inequality, largely by boosting net farm incomes and raising real wages for unskilled workers in developing countries, and would reduce the number of poor people worldwide by 3 percent. The analysis based on the Global Trade Analysis Project model for a sample of 15 countries, and nine stand-alone national case studies, all point to larger reductions in poverty, especially if only the non-poor are subjected to increased income taxation to compensate for the loss of trade tax revenue. Would Freeing Up World Trade Reduce Poverty and Inequality? The Vexed Role of Agricultural Distortions [Running title: Would Freeing Up World Trade Reduce Poverty?] Kym Anderson, John Cockburn and Will Martin1 1. INTRODUCTION For decades, earnings from farming in many developing countries have been depressed by a pro-urban, anti-agricultural bias in own-country policies as well as by governments of richer countries favoring their farmers with import barriers and subsidies. Both sets of policies reduced national and global economic welfare, inhibited economic growth, and added to inequality and poverty because no fewer than three-quarters of the world's billion poorest people still depend directly or indirectly on farming for their livelihood (World Bank 2007). During the past two to three decades, numerous developing country governments have reduced their sectoral and trade policy distortions, while some high-income countries also have begun reforming their protectionist farm policies. Yet myriad policy measures continue to distort world food markets, and in many and complex ways (Anderson 2009). In some developing country settings they raise food prices for consumers and the earnings of farm households, in other settings they lower them; but in most situations there is a mixture of 1 KYM ANDERSON is George Gollin Professor of Economics at the University of Adelaide, Australia, JOHN COCKBURN is a Professor of Economics at Laval University in Quebec, Canada and co-director of the Poverty and Economic Policy (PEP) research network, and WILL MARTIN is a Research Manager in the Development Research Group at the World Bank in Washington DC. This is a product of a World Bank research project on Distortions to Agricultural Incentives. The authors are grateful for helpful comments from workshop participants and referees, and for funding from World Bank Trust Funds provided by the governments of the Netherlands (BNPP), the United Kingdom (DfID), the Multi-donor Trust Fund for Trade, and from the Australian Research Council. The views expressed are the authors' alone and not necessarily those of the World Bank and its Executive Directors, nor the countries they represent, nor of the institutions providing funds for this research project. 2 winners and losers in both rural and urban areas, not least because many farm households receive some of their income from non-farm sources. The only feasible option for discerning the net impacts of price-distorting policies on poverty and inequality is to undertake quantitative analysis using economy-wide models with up-to-date price distortion data and ideally detailed household information on the earning and spending profiles of different groups of people, both rural and urban. The need for undertaking poverty and inequality analysis remains strong, notwithstanding the contributions of policy reforms over the past quarter-century. Partly as a result of those policy reforms and the consequent growth of incomes in many developing countries, the number of people living on less than $1 a day nearly halved over the 1981- 2005 period, and their share of the global population fell from 42 to 16 percent (Table 1). Yet that number of extremely poor people was still almost 900 million in 2005, and it may have risen above that following the eruption of the global financial crisis that began in 2008. Moreover, most of the improvement has been in Asia (especially China), while in Sub- Saharan Africa the incidence of poverty was little lower in 2005 than in 1981, at around 40 percent (amounting to 300 million people in 2005). Despite the success of China, it still had over 100 million people on less than $1 a day in 2005, 90 percent of whom were rural. And in India the number of extreme poor remains stubbornly close to 300 million ­ and 74 percent rural, even with large subsidies to their farmers. Less pressing than extreme poverty but nonetheless still important to the welfare of individuals is the extent of income inequality. In the past it was just inequality at the local level that affected individuals' utility, but the information and communication technology revolution has increased awareness of income differences not only within local regions but also nationally and internationally. At the national level, there are concerns about rural-urban inequality as well as inequality within each of those broad geographic zones. Within rural 3 areas, for example, differences in incomes can be vast between landless unskilled farm workers, subsistence farmers, the larger commercial farmers, and non-farm workers in rural towns. Assessing what has happened to the world's income distribution in recent decades depends on one's focus. Milanovic (2005) points to three possibilities. One is intercountry inequality, which compares country-level average incomes where each country has an equal weight in the world distribution regardless of population size. In that case income distribution appears to have become more unequal. The second is international inequality, which still compares country average incomes but this time weighting by the populations of countries. In that case income inequality appears to have decreased, although mostly due to the fast growth in populous China and India (Bourguignon, Levin and Rosenblatt 2004; Atkinson and Brandolini 2004). And the third possible focus is global inequality, which involves comparing individual incomes regardless of country of citizenship, thus taking into account within-country inequality which is ignored by the international inequality approach where individuals are deemed to earn their country's average income. Rapid growth in the large emerging economies has tended to offset the increase in inequality within countries and so, by this last definition, global inequality appears to have remained roughly constant since the late 1980s.2 In the light of the evidence currently available, the question this paper focuses on is: how much scope is there to further reduce poverty and inequality in the world, and in specific developing countries, by getting rid of remaining distortions to incentives facing producers and consumers of tradable goods unilaterally or globally? 2 A study by Sala-i Martin (2006) found that GDP per capita disparities between countries have shrunk as economies have converged. See also the analyses based on household survey data rather than GDP per capita, such as by Milanovic (2002, 2006). A recent review of the global poverty and inequality evidence is available in Ferreira and Ravallion (2009). 4 Empirical studies undertaken as background for the World Trade Organization's on- going Doha round of multilateral trade negotiations suggest that in 2001, when that round was launched, policy-driven distortions to agricultural incentives contributed around two- thirds of the global welfare cost of merchandise trade barriers and subsidies (see, e.g., Anderson and Martin 2005). While such empirical studies did not have access to comprehensive estimates of distortions to farmer and food consumer incentives in developing countries other than applied tariffs on imports, a more recent study that draws on a new database of distortions to agricultural incentives has confirmed that earlier result: Valenzuela, van der Mensbrugghe and Anderson (2009) suggest agricultural price and trade policies as of 2004 accounted for 70 percent of the global welfare cost of those and other merchandise trade policies. This is a striking result, given that the shares of agriculture and food in global GDP and trade are only 3 and 6 percent, respectively. The contribution of farm and food policies to the welfare cost of global trade-distorting policies for just developing countries is estimated by those authors to be even greater, at 72 percent ­ of which two-thirds is due to policies of developing countries themselves. Even so, the estimates of price distortions that went into that modeling study show that many developing countries protect their less-competitive farmers from import competition, so some farm households might be hurt if all markets were opened (Anderson 2009, Ch. 1). The World Bank's recent study of price distortions (Anderson 2009) shows that the rate of assistance to farmers relative to producers of non-farm tradables has fallen by one- third for high-income countries since the latter 1980s (from 51 to 32 percent) while in developing countries it has all but disappeared (rising from -41 percent in the early 1980s to +1 percent in 2000-04). The latter trend for developing countries is mainly because of the phasing out of agricultural export taxes, since assistance via import restrictions has risen over the period shown. In both high-income and developing countries there remains a large gap 5 between their nominal rates of assistance (NRAs) for import-competing and exporting agricultural industries, as well as a continuing large gap (albeit smaller than in the 1980s) between the relative rates of assistance in the two groups of countries. In the light of that evidence, the above question to be addressed here can be expressed more specifically, for any developing country of interest, as: how important are its own policies compared with those of the rest of the world in affecting the welfare of the poor in that country, and what do agricultural policies in particular contribute to those outcomes? Clear answers to this question can guide countries in their national policymaking and as they negotiate bilateral and multilateral trade agreements. Now is an appropriate time to address this multi-faceted question for at least two policy reasons. One is that the World Trade Organization (WTO) is struggling to conclude the Doha round of multilateral trade negotiations, and agricultural policy reform is once again one of the most contentious issues in those talks. The other is that poorer countries are striving to achieve their United Nations-encouraged Millennium Development Goals by 2015, the prime ones being the alleviation of hunger and poverty. There are also several analytical reasons as to why now is the time to focus more thoroughly on this issue. One is that methodologies to address it have advanced at a rapid pace recently, involving micro-simulation modeling based on household survey data in conjunction with economy-wide computable general equilibrium (CGE) modeling. Prominent examples include the studies in Hertel and Winters (2005, 2006) and in Bourguignon, Bussolo and da Silva (2008). Household income information is increasingly important for poverty and inequality analysis because farm households and rural areas of developing countries are rapidly diversifying their sources of income beyond what agricultural land and farm labor can generate, including from part-time off-farm work and remittances (Otsuka and Yamano 2006, Otsuka, Estudillo and Sawada 2009). Hence the earlier close correspondence 6 between net farm income or agricultural GDP and farm household welfare is fading, even in some low-income countries (Davis, Winters and Carletto 2009). Frequently, many of the poor, including the rural poor, are net-buyers of staple foods, causing them to be adversely affected--at least in the short run--by increases in prices of staple foods. Second, the compilation of national household surveys that are comparable for cross- country analysis has progressed rapidly such that there are now recent surveys for more than 100 countries available at the World Bank. That dataset (www.worldbank.org/prospects/gidd) has already begun to be used in conjunction with the World Bank's LINKAGE model of the global economy to assess global income distribution issues (e.g., Bussolo, De Hoyos and Medvedev 2008). Third, the World Bank has recently compiled a very comprehensive new global database that updates and expands substantially our understanding of the distortions to agricultural incentives in developing countries in particular.3 Those estimates have since been expressed so as to make them usable in national and global economy-wide models (Valenzuela and Anderson 2008). They differ from the usual ones employed by trade modelers of developing country policies in that they are based on direct domestic-to-border price comparisons rather than (as with the GTAP dataset, see Narayanan and Walmsley 2008) on applied rates of import tariffs and other key border measures. A first attempt to exploit those new methodologies and databases has recently been undertaken to assess the relative impacts on national, regional and global poverty and inequality of agricultural and non-agricultural trade policies at home and abroad. This paper summarizes and draws lessons from the papers that have emerged from that research project. At the outset it should be made clear that agricultural and trade policies are far from the first-best policy instruments for achieving national poverty or income distribution 3 That distortions database is documented fully in Anderson and Valenzuela (2008) and is based on the methodology summarized in Anderson et al. (2008) and detailed in Appendix A of Anderson (2009). 7 objectives; that is the prerogative of domestic social welfare and income tax policy measures. However, if empirical studies reveal that national trade-related policies are worsening particular countries' poverty or inequality, they provide yet another reason ­ on top of the usual national gains-from-trade reason ­ for those countries to reform their policies unilaterally. Should the inequality and poverty alleviating effects of national trade-related policy reforms be contingent on the rest of the world also reforming, that provides a further reason for that country to participate actively in promoting multilateral trade negotiations under the World Trade Organization (WTO). And should global modeling studies reveal that multilateral trade reform would alleviate global inequality and poverty, it underlines the importance of bringing the WTO's Doha Development Agenda (DDA) expeditiously to a successful conclusion with ambitious reform commitments. On the other hand, a negative finding (e.g., that trade liberalization would increase poverty in a particular country) need not be a reason to shun welfare-enhancing trade reform. Rather, such a result could be used to provide guidance as to where tax or social programs need to be targeted so that all groups in society can share in the economic benefits from such reform (see Ravallion 2009). Global reform results could also provide bargaining power to developing countries seeking aid-for- trade side payments to alleviate any increase in poverty projected to result from multilaterally-agreed trade reform. The paper begins with an outline of the analytical framework and the common empirical methodology adopted by the global and national case studies being summarized. It then compares modeling results from both global and national models, before mentioning some caveats and drawing out policy implications. The findings are based on two studies that each use a global model to examine the effects of farm and non-farm price and trade policies on global poverty and its distribution within and across many identified countries, plus nine 8 individual developing country studies spanning the three key regions of Asia (where nearly two-thirds of the world's poor live), Sub-Saharan Africa and Latin America. 2. ANALYTICAL FRAMEWORK In order to adequately capture poverty and inequality effects of price-distorting policies, careful consideration must be given to its impacts on household income and expenditure. Many farm households in developing countries rely on the farm enterprise for virtually all of their income, and in the world's poorest countries the share of national poverty concentrated in such households is large. The fact that the poorest households in the poorest countries are concentrated in agriculture means those households are likely to benefit from farm producer price increases engendered by trade policy reform, other things equal. However, the outcome is not certain because poor households also spend the majority of their income on staple foods (Cranfield et al. 2003), so if food prices rise as a consequence of reform then this adverse effect on their household expenditure may more than offset the beneficial effect on them of higher earnings. The urban poor also would be adversely affected by a rise in consumer prices of staple food. However, it is possible that a trade reform that induced a rise in food prices may also raise the demand for unskilled labor (according to the relative factor intensities of production in the economy's expanding sectors), which ­ depending on how intersectorally mobile is labor ­ could raise the income of poor households more than it raises the price of their consumption bundle. The approach adopted in the present study to operationalize the above theory is a variant on the path-breaking approach pioneered by Hertel and Winters (2005, 2006) in their study of the poverty consequences of a prospective Doha round agreement under the WTO. The present study contrasts with that earlier one in three respects. First, here the focus is on 9 the impacts of agricultural domestic and trade policies, distinguishing them from the impacts of other merchandise trade policies. A second distinction is that we examine inequality as well as poverty. Third, we focus on the effects of current policies, i.e. of full (not partial) global liberalization, whereas the Hertel and Winters study focuses mainly on the multilateral partial reform proposals that were on the table as of 2005. The country case studies examine not just multilateral trade reform but also unilateral reforms that individual developing countries might implement. The effects of unilateral action are compared with what full liberalization abroad would generate, so as to be able to assess the relative importance domestically for each nation of own-country policies as distinct from those of other countries (over which the country has influence only indirectly via trade negotiations). The national CGE models are able on their own to estimate the effects of unilateral reform of agricultural or all merchandise trade-distorting policies. For the national modeler to estimate the effects of other countries' policies, however, requires input from a global model. The World Bank's LINKAGE model is chosen here for that purpose. It too is calibrated to 2004, based on Version 7 of the GTAP global protection database (Narayanan and Walmsley 2008) apart from the replacing of its applied agricultural tariffs for developing countries with the more comprehensive set of national price distortion estimates from Valenzuela and Anderson (2008). There are various ways of transmitting the results derived from a global CGE model such as LINKAGE to a single-country CGE model. Like Hertel and Winters (2006), we adopt the approach developed by Horridge and Zhai (2006). For imports, Horridge and Zhai propose the use of border price changes from the global model's simulation of rest-of-world liberalization (that is, without the focus developing country). For the focus developing country's exports, the shift in its export demand curve following liberalization in the rest of the world is given in percentage changes by x=(1/).q where x is the percentage vertical shift 10 in the export demand curve, is the elasticity of substitution between the exports of country i and those from other countries, and q is the percentage change in the quantity of exports under the scenario with liberalization in the rest of the world excluding the focus country. All the CGE models referred to below are comparative static, and they assume constant returns to scale and perfectly competitive homogeneous firms and product markets. In all cases other than South Africa (and to a smaller extent for Nicaragua), unemployment is assumed to be unaffected by the trade policy regime. These assumptions are imposed simply because of insufficient data and empirical evidence to impose alternative ones across all the countries being modeled. This use of a standard set of assumptions reduces the risk that differences across countries in results are driven by different assumptions about investment behavior or the degrees of monopolistic competition, firm heterogeneity, economies of scale, or aggregate employment response to trade policy changes (see Helpman, Itskhoki and Redding 2010). Such specifications typically lead to underestimation of the net national welfare gains that would accrue from trade reform though (Francois and Martin 2010). In particular, without dynamics the models will not generate a growth dividend from freeing up markets or from eventual productivity/efficiency gains from trade. Since economic growth is the predominant way in which poverty is reduced in developing countries (see the literature review in Ravallion 2006), the absence of dynamics implies that the results from this study almost certainly underestimate the potential poverty alleviating consequences of liberalization ­ and might in some situations indicate poverty increases when in fact they would be decreases once the growth consequences are incorporated. All the country case studies, and one of the two global modeling studies surveyed below, make use of household survey data in addition to a social accounting matrix (SAM). The SAM is the basis for the data in the CGE model, while the household survey data are used in micro-simulation modeling. 11 Typically the experiments are performed in two stages. The first stage involves the imposition on the national CGE model of the policy shock (either unilateral liberalization or an exogenous shock to border prices and export demand provided by the LINKAGE model). This generates changes in domestic product and factor markets. The consequent changes in consumer and factor prices are then transmitted to the micro-simulation model to see how they alter the earnings of various household types (according to the shares of their income from the various factors) and their cost of living (according to the shares of their expenditure on the various consumer products). That in turn provides information on changes in the distribution of real household incomes and hence in inequality, and in the number of people below any chosen poverty line such as US$1 a day. All country case studies ran a common set of simulations so as to make it possible to compare the inequality and poverty effects in each country of own-country versus rest-of- world policies affecting markets for agricultural (including lightly processed food) goods versus other merchandise. The other global study referred to in the next section uses the same 2004 global protection dataset but implements global reform shocks using a slightly different global model for each of 15 developing countries but with national household survey data attached in order to undertake micro-simulations. In most cases additional simulations were also run, often to illustrate the sensitivity of the results to key assumptions pertinent to that particular country case study. Even though the models surveyed here are all standard perfectly competitive, constant-returns-to-scale, comparative static, economywide CGE models, they nonetheless differ somewhat in order to capture important realities (such as labor market characteristics or data limitations) in their particular setting. However, to ensure their comparability, they all aimed to conform to a common set of factor market assumptions and closure rules in addition to using 2004 as their base and undertaking a common set of simulations using the same 12 global distortions dataset. Specifically, all modelers assumed the following: a fixed aggregate stock of factors (including no international mobility), with the exception of labor in the Nicaraguan and South African studies where some aggregate employment responsiveness to trade policy is allowed because of high unemployment in the baseline; possibly some sector- specific capital and labor, but most capital and labor types are assumed to be intersectorally mobile with a common flexible rate of return or wage; and land is assumed to be specific to the agricultural sector but mobile across the different crop and livestock activities within that sector. The key agreed macroeconomic closure rules that each case study aimed to adopt are a fixed current account in foreign currency, to avoid foreign debt considerations, and fixed real government spending and fiscal balance, so as to not affect household utility other than through traceable changes in factor and product prices and taxes. Fiscal balance is achieved by using a uniform (generally direct income) tax to replace net losses in revenue from abolishing sector trade taxes and subsidies. 3. SYNOPSIS OF EMPIRICAL FINDINGS: GLOBAL MODEL RESULTS This section summarizes the results from two global models (denoted LINKAGE and GTAP).4 The following section then brings together the results from nine more-detailed national case studies, before the lessons learned from both sets of analyses are drawn together. It would be surprising if all the studies came to the same conclusions, but the strength of this blend of 4 Results were also generated by Bussolo, De Hoyas and Medvedev (2010) making use of the global LINKAGE model and combining this with the newly developed Global Income Distribution Dynamics (GIDD) microsimulation tool (Bussolo, De Hoyos and Medvedev 2008). However, since the key inputs into the microsimulation from the LINKAGE model are just the labor income changes, those results are not directly comparable with the other studies surveyed here. 13 somewhat different global and national models is that it is more likely to expose the various determinants of the measured effects in different settings than if only a single type of model was employed. (a) LINKAGE Model Results Anderson, Valenzuela and van der Mensbrugghe (2010) use the World Bank's global LINKAGE model (van der Mensbrugghe 2005) to assess the market effects of the world's agricultural and trade policies as of 2004 on individual countries and country groups, so as to be able to say something about international inequality (in the Milanovic (2005) sense, taking into account the economic size of countries) and about poverty (using a simple elasticities approach). This model also provides the basis for estimating the effects of rest-of-world policies on the import and export prices and demand for the various exports of any one developing country, for use by each of the nine country case studies discussed in the next section. The LINKAGE model results suggest that developing countries would gain nearly twice as much as high-income countries in welfare terms if 2004 agricultural and trade policies were removed globally (an average welfare increase of 0.9 percent, compared with 0.5 percent for high-income countries ­ bottom of column 1 of Table 2). Thus in this broad sense of a world of just two large country groups, completing the global reform process would reduce international inequality. The results vary widely across developing countries, however, ranging from slight losses in the case of some South Asian and Sub-Saharan African countries that would suffer exceptionally large adverse terms of trade changes, to an 8 percent increase in the case of Ecuador (whose main export item, bananas, is currently 14 heavily discriminated against in the EU market where former colonies and least developed countries enjoy preferential duty-free access). Bearing in mind that three-quarters of the world's poorest people depend directly or indirectly on agriculture for their main income, and that farm sizes are far larger in high- income than in developing countries, the LINKAGE study also looks at the extent to which agricultural and trade policies in place as of 2004 reduced rewards from farming in developing countries and thereby added to international inequality in farm incomes. It finds that net farm incomes in developing countries would rise by 5.6 percent, compared with 1.9 percent for non-agricultural value added, if those policies were eliminated (bottom of final two columns of Table 2). This suggests that inequality between farm and nonfarm households in developing countries would fall from such reform, notwithstanding the notable exception of India. The large reduction in agricultural GDP in India, as in higher-income countries, reflects in part the fact that import-competing farmers in India currently enjoy considerable protection at the border. In high-income countries net farm incomes would fall by 15 percent on average, compared with a slight rise for real non-farm value added. These results suggest inequality between farm and nonfarm households within high-income countries, as also in India, would increase if no compensating domestic measures were taken. They also suggest inequality between farm households in developing countries and those in high-income countries would reduce substantially. These inequality results would not be very different if only agricultural policies were to be removed (compare columns 2 and 3 of Table 2), underscoring the large magnitude of the distortions from agricultural, as compared with non- agricultural, trade policies. This study also reports that unskilled workers in developing countries ­ the majority of whom work on farms ­ would benefit most from reform (followed by skilled workers and then capital owners), with the average change in the unskilled wage over all developing 15 countries rising 3.5 percent when deflated by the aggregate Consumer Price Index or CPI (column 4 of Table 3). However, the most relevant consumer prices for the poor, including those many poor farm and other rural households who earn most of their income from their labor and are net buyers of food, relate just to food and clothing. Hence deflating by a food and clothing price index rather than the aggregate CPI provides a better indication of the welfare change for those workers. As shown in the final column of Table 3, for all developing countries the real unskilled wage over all developing countries would rise by 5.9 percent with that deflator. That is, inequality between unskilled wage-earners and the much wealthier owners of capital (human or physical) within developing countries would reduce with full trade reform. The above results for real factor rewards and net farm income suggest that poverty, as well as international and intra-developing country inequality, could be alleviated globally by agricultural and trade policy liberalization. The authors of that study go a step further to explicitly assess reform impacts on poverty even though the LINKAGE model has only one single representative household per country. They do so using the elasticities approach, which involves taking the estimated impact on real household income and applying an estimated income to poverty elasticity to estimate the impacts on the poverty headcount index for each country. They focus on the change in the average wage of unskilled workers deflated by the food and clothing CPI, and assume those workers are exempt from the direct income tax imposed to replace the lost customs revenue following trade reform (a realistic assumption for many developing countries). Under the full merchandise trade reform scenario, Table 4 reports that extreme poverty (the number of people surviving on less than US$1 a day) in developing countries would drop by 26 million relative to the baseline level of just under one billion, a reduction of 2.7 percent. The proportional reduction is much higher in China and in Sub-Saharan Africa, 16 each falling around 4 percent. It is even higher in Latin America (7 percent) and South Asia other than India (10 percent). By contrast, the number of extreme poor in India (though not in the rest of South Asia) is estimated to rise, by 4 percent.5 Under the more moderate definition of poverty--those living on no more than US$2 per day--the number of poor in developing countries would fall by nearly 90 million compared to an aggregate baseline level of just under 2.5 billion in 2004, or by 3.4 percent (notwithstanding the number in India below $2 a day still increasing, but by just 1.7 percent). (b) GTAP Model Results Hertel and Keeney (2010) draw on the widely used global economy-wide model of the Global Trade Analysis Project (GTAP). Their study adopts the same price distortions as the other studies surveyed here, and runs the same scenarios, but generates its own world price changes from the GTAP model for the multilateral trade reform scenarios. Those prices changes alter border prices for the various countries in the GTAP model, a subset of which have attached to them detailed household survey data. This permits the authors to say something about poverty impacts across a range of diverse economies using an internally consistent framework that captures the distributive effects of all factor income changes. This multi-country study focuses on 15 developing countries: five Asian (Bangladesh, Indonesia, Philippines, Thailand, and Vietnam), four African (Malawi, Mozambique, Uganda, and Zambia), and six Latin American countries (Brazil, Chile, Colombia, Mexico, Peru, and Venezuela). Overall, it concludes that removing current farm and trade policies globally would tend to reduce poverty, and primarily via agricultural reforms (Table 5). The unweighted average for all 15 developing countries is a headcount decline in extreme poverty 5 The rise in India is partly because of the removal of the large subsidies and import tariffs that assist Indian farmers, and partly due to the greater imports of farm products raising the border price of those imports. 17 (<$1 a day) of 1.7 percent. The average fall for the Asian sub-sample is twice that, however ­ and it is in Asia where nearly two-thirds of the world's extremely poor people live (although their sample did not include China and India). Turning to their results for specific countries, it is the agricultural-exporting developing countries in the sample, namely Chile, Thailand and Vietnam, where the most poverty alleviation would occur (column 3 of Table 5). The majority of the 15 countries studied experience small poverty increases from non-agricultural reforms, although the unweighted average across the fifteen countries suggests a slight decrease, primarily due to a strong decline in Vietnam (column 2 of Table 5). These GTAP model results are close to the LINKAGE model results in the first part of this section of the paper. The authors explore the relative poverty-friendliness of agricultural trade reforms in detail, examining the differential impacts on real after-tax factor returns of agricultural versus non-agricultural reforms. Their analysis is extended to the distribution of households by looking at stratum-specific poverty changes. They find that the more favorable impacts of agricultural reforms are driven by increased returns to peasant farm households' labor as well as higher returns for unskilled work off-farm. They also find that liberalization of food grain markets represents the largest contribution to poverty reduction, and that removing import tariffs in those commodity markets dominates the poverty-increasing impacts of subsidy removal by high-income countries. The final column of Table 5 reports the percentage change in the national poverty headcount when the poor are not subject to the income tax rise required to replace trade tax revenue following trade reform. This assumption represents a significant implicit income transfer from non-poor to poor households and thus generates a marked difference in the predicted poverty alleviation. Trade reforms go from being marginally poverty reducing in most of the 15 cases to being poverty reducing in all cases and by a considerable magnitude. 18 It reduces the poverty rate by roughly one-quarter in Thailand and Vietnam, for example. Overall, the regional and total average extent of poverty alleviation is around four times larger in this scenario than when the poor are also assumed to be levied with income taxes to replace lost trade tax revenue. The unweighted average poverty headcount reduction for the three regions shown in the final column of Table 5 are in line with the population-weighted averages from the LINKAGE model reported in Table 4 above with a similar tax-replacement assumption: the latter's 30 percent for Asia excluding China and India and 6.8 percent for Latin America are above the GTAP model's 14 percent and 5.7 percent, while their 3.8 percent for Sub-Saharan Africa is just below the 4.5 percent obtained for the Hertel and Keeney sample. 4. SYNOPSIS OF EMPIRICAL FINDINGS: NATIONAL MODEL RESULTS We turn now to see how the results from nine more-detailed individual country case studies compare with the above results from global models.6 Like the two global models, they focus on price-distorting policies as of 2004, even though the database for their CGE models and their household survey data typically date back a little earlier in the decade. They all include more sectoral and product disaggregation than the global models, and have multiple types of 6 The nine national studies are for Brazil (Ferreira Filho and Horridge (2010), China (Zhai and Hertel 2010), Indonesia (Warr 2010a), Mozambique (Arndt and Thurlow 2010), Nicaragua (Sanchez and Vos 2010), Pakistan (Cororaton and Orden 2010), Philippines (Cororaton, Corong and Cockburn 2010), South Africa (Herault and Thurlow (2010), and Thailand (Warr 2010b). Results were also generated for Argentina (Cicowiez, Diaz- Bonilla and Diaz-Bonilla (2010), but they are not included here because they are based on a household survey that unfortunately did not include rural areas. 19 households and types of labor. All of the national studies include micro-simulations drawing on model results, as in the above GTAP global models. The national results for real GDP and household consumption suggest that GDP would increase from full global trade reform, but only by 1 or 2 percent, in all nine countries. Given falling consumer prices, real household consumption would increase by considerably more in most cases. Generally these numbers are a little larger than those generated by the global LINKAGE model. They share the feature of the global models of probably underestimating the poverty-alleviating benefits of trade reform, given the broad consensus in the literature that trade liberalization increase growth, which is in turn a major contributor to poverty alleviation. The comparative Tables 6 and 7 summarize the national results for the incidence of extreme poverty and income inequality, respectively, resulting from own-country, rest-of- world or global full liberalization of agricultural or all goods trade. Some authors ran only six of the nine simulations shown in this table, but those that ran all nine found their results to sum up almost exactly, to one decimal place. We therefore have inferred the three missing results in the other country studies by assuming that the agriculture-only and nonagriculture- only results sum to the all-goods reform results. The inferred numbers are shown in italics in Tables 6 and 7. In each case the total effects on poverty7 and inequality are subdivided into rural and urban. One should not necessarily expect the unweighted averages of the poverty results for each region to be similar to those generated by Hertel and Keeney (2010), because only five of the national case studies were included among the 15 countries sampled by Hertel and Keeney. Nonetheless, the latter's unweighted averages of national poverty effects for each of the key developing country regions are reported in brackets in the last 4 rows of Table 6(c), to 7 Using national or $1 a day poverty lines, except for China for which results are available only for $2 a day. 20 make it easy to compare with the unweighted regional averages for the national case studies. In all but three of those twelve comparisons for global liberalization (agricultural, non- agricultural and all merchandise), the projected regional average poverty reductions from global liberalization are larger from the available sample of national case studies than from Hertel and Keeney's 15-country sample. As for the individual country results, poverty is reduced in all countries by both global agricultural and, with the exception of the Philippines, non-agricultural liberalization (Table 6(c)). When all merchandise trade is liberalized, the extent of reduction ranges from close to zero to about 3.5 percentage points, except for Pakistan where it is more than 6 points.8 On average nearly two-thirds of the alleviation is due to non-farm trade reform, with the important exception of Brazil where agricultural reform is the major contributor to its large pro-poor outcome. The latter result is despite the presence of tariff protection for Brazil's poor import-competing farmers, and is a consequence of the increase in demand for unskilled labor following liberalization, which evidently outweighs the poverty impact of removing farm tariffs. The contribution of own-country reforms to the fall in poverty appears to be equally as important as rest-of-world reform on average, although there is some considerable cross-country divergence in the extent of this for both farm and non-farm reform. The poverty alleviation is sub-divided in parts (a) and (b) of Table 6 into rural and urban sources. A glance at the final column of that part of the table reveals that rural poverty is cut much more than urban poverty in every case. That is true for both farm and non-farm trade reform, and for own-country as well as rest-of-world reform. Since the rural poor are much poorer on average than the urban poor, this would lead one to expect trade reform to reduce inequality also. 8 The Pakistan results were generated assuming replacement of trade taxation with a rise in direct income taxes. Only urban, non-poor households pay direct taxes in Pakistan, so the removal of tariffs decreases the after- tax incomes of the urban non-poor and means the benefits of trade reform go mainly to the poor. 21 Indeed, the results at the bottom of Table 7(c) for this sample of countries show that inequality would decline in all three developing country regions following full trade liberalization of all goods, or just agricultural products, and both for own-country and rest-of- world reform. The effect of non-farm trade reform on its own is more mixed, providing another reason to urge trade negotiators not to neglect agricultural reform in trade negotiations. Rest-of-world and global agricultural reform both lead to a reduction in inequality in every country in the sample except Thailand (plus the Philippines slightly for global reform), whereas unilateral agricultural reform reduces (or leaves constant) inequality in a small majority of countries with China, the Philippines and Thailand being the exceptions (but the latter effects are small). Non-farm global reform increases inequality slightly in three countries. In the case of Indonesia the inequality-increasing impact of non- farm reform more than offsets the egalitarian effect of farm trade reform, whereas both types of reform increase inequality in the case of the Philippines and Thailand. Inequality within the rural or urban household grouping is not altered very much by trade reform as compared with overall national inequality (compare parts (a) and (b) with part (c) of Table 7). This underlines the point that trade reform would tend to reduce urban-rural inequality predominantly rather than inequality within either region. Several of the national studies investigate impacts of reforms that could complement trade reforms, most notably different approaches to deal with the elimination of trade tax revenues. If these revenues can be recouped through taxes that do not bear on the poor, then the impacts of reform for poverty reduction are more favorable. The China study focuses on the important issue of reducing the barriers to migration out of agriculture, by improving the operation of land markets and reducing the barriers to mobility created by the hukou system. These measures, and international trade liberalization that increases China's market access, 22 are found to reduce poverty such that a combination of these measures would benefit all major household groups. 5. WHAT HAVE WE LEARNED? As found in previous studies, whether based on ex post econometrics (as in Harrison 2007) or ex ante economy-wide simulation (as in Hertel and Winters 2006), so this study also finds mixed results that are not easy to summarize, particularly with regard to the poverty effects. There is nonetheless a high degree of similarity in the most important sign: the estimated national extreme poverty effect of freeing all merchandise trade globally. It happens to be the effect for which there is the most overlap between the studies summarized above: all but two of the 32 cases shown indicate that overall global trade reform would decrease poverty. This beneficial impact of full liberalization of merchandise trade on the world's poor would come more from agricultural than non-agricultural reform; and, within agriculture, more from the removal of substantial support provided to farmers in developed countries than from developing country policy reform. According to the economywide models used in the present study, such reform would raise real earnings of unskilled workers in developing countries, most of whom work in agriculture. Their earnings would rise relative to both unskilled workers in developed countries and other income earners in developing countries. This would thus reduce inequality both within developing countries and between developing and developed countries, in addition to reducing poverty. According to the LINKAGE model results, the number of extremely poor people in developing countries (on less than $1 a day) is estimated to fall by 2.7 percent with global opening of all goods markets, and by 4 percent in China and Sub-Saharan Africa, but to rise by 4 percent in India (or by 1.7 percent if the more moderate $2 a day poverty level is used). 23 The 15-country results from the GTAP model suggest that the poverty-reducing effects would be somewhat larger, and the nine national case studies all find global trade liberalization to be poverty alleviating, regardless of whether the reform were to involve only agricultural goods or all goods, with the benefit coming roughly equally from reform at home and abroad. The latter studies also find that rural poverty would be cut much more than urban poverty in all cases, whether from reform at home or abroad and whether or not it included non-farm goods. Global trade liberalization would reduce international inequality as between developing and high-income countries, both in total and for just farm households, according to the LINKAGE model. But it cannot be guaranteed that every developing country would be better off unless there is a strong economic growth dividend from reform (not captured in the comparative static modeling used in these studies). Full trade liberalization of all goods, or just of agricultural products, also would cause inequality to decline within each of the three developing country regions covered by our sample of countries, and both for own-country and rest-of-world reform. Inequality within the rural or urban household grouping would not alter much following full trade reform, suggesting that trade reform's predominant impact would be to reduce urban-rural inequality. The mechanism through which governments adapt to the fall in tariff revenue is also shown to be crucial. If it is assumed (realistically) that the poor do not have to bear any of the burden of replacing trade taxes, instead of sharing it proportionately, the estimated degree of poverty alleviation is about four times greater in the 15 countries studied with the GTAP model. Results from the two global analyses indicate that removing remaining agricultural policies would have much stronger impacts on poverty and inequality than would non- agricultural trade reforms. A weighted average across the nine country case studies would 24 probably come to a similar conclusion. The nine national case studies also shine some light on the relative importance of domestic versus rest-of-world reform for those countries: the contribution of own-country reforms to the fall in poverty appears to be equally as important as rest-of-world reform on average, although there is some considerable cross-country divergence in the extent of this, both for farm and non-farm reform. 6. CAVEATS The impacts of agricultural and other trade reform are complex, simultaneously affecting product and factor markets, government budgets and external trade. The studies included in this survey provide a range of ex ante modeling perspectives, including both global and national models. Considerable attention has been devoted to capturing poverty effects through the use of recent micro-simulation and poverty elasticity approaches, and to using the same price distortion estimates, the same global model for getting rest-of-world border shocks for the nine national models, and similar behavioral assumptions, tax replacement assumptions and model closures. Nonetheless, there is ample scope for further exploration of this issue through additional comparisons, including in the form of drilling down into each modeling result to explore its origins. The reforms considered here refer only to liberalization of goods trade. Freeing global trade in services is also likely to bring gains to most national economies, including their farmers (Francois and Hoekman 2010). Freeing international capital flows would add to those gains (Hoxha, Kalemli-Ozcan and Vollrath 2009), as would freeing the international movement of low-skilled labor from developing to higher-income countries (Walmsley and Winters 2005; World Bank 2005). How those reforms would interact with farm and other goods trade reforms, in terms of their impacts on global poverty and inequality, is bound to be complex and so awaits the development of more-sophisticated global simulation models. 25 Another key challenge that remains is to capture the growth effects of liberalization and, in particular, their general equilibrium distributive (poverty and inequality) consequences. This area of research has only recently begun to be addressed in the empirical literature, building on the gains made in the theoretical endogenous growth literature in the 1990s. Existing partial equilibrium analysis strongly suggests that the trade-growth-poverty nexus is important, possibly much more important than the static re-allocative impacts captured in the current set of studies. There is reason to believe that, once dynamics are included in models, they will reinforce the basic finding of this study that agricultural and other merchandise trade policy reform is poverty and inequality reducing. A further modeling change is to introduce a stochastic dimension so as to capture changes in the probability of falling into poverty. This is important if greater openness changes the risk of food price spikes: an upward spike could cause a food-deficit household to starve, for example. General equilibrium empirical modeling that contains sufficient sector and household detail to be useful for poverty analysis, even without a dynamic component, is still in its infancy. However, this field may develop rapidly in response to the demand for climate change studies, an early prototype being Ahmed, Diffenbaugh and Hertel (2009). There is huge scope also for exploring empirically the possible effects of complementary domestic reforms that could accompany agricultural price and trade policy reforms. This is illustrated in the China case study by Zhai and Hertel (2010), which showed that if labor market reform were to accompany trade reform the poverty alleviation would be several times greater. Even in the extreme case of India, the latter reforms would probably not increase poverty if more-efficient transfer mechanisms were in place and high-payoff infrastructure investments were made during a phasing out of agricultural producer supports. The politics of having first-best domestic policies in place are not necessarily any less complex than those associated with trade policies, however, which underscores the need for 26 comprehensive political economy analysis that does not limit its focus just to border policy measures.9 7. POLICY IMPLICATIONS The above empirical findings have a number of policy implications. First and foremost, the generally attractive results in terms of poverty and inequality alleviating effects from trade policy reforms, whether unilateral or multilateral, provide yet another reason as to why it is in the interests of countries to seek further liberalization of national and world markets. Second, a recurring theme in the national case studies is that the benefits in terms of poverty and inequality alleviation, in addition to the standard aggregate real income gains associated with trade liberalization, are generally much greater from global reform than from just own-country reform. In the Indonesia study, for example, unilateral trade liberalization is expected to reduce poverty only very slightly, but liberalization by the rest of the world is expected to lower poverty very substantially. In the Philippines, domestic reform alone from current levels of protection may marginally increase poverty rates, whereas rest-of-world liberalization would almost fully offset that (and more than offset it in the case of just agricultural reform). Third, the results of this set of studies show that the winners from trade reform would be found among the poorer countries and the poorest individuals within countries. However, it is also clear that even among the extreme poor, some could lose out. Hence the merit of compensatory policies, ideally ones that focus not on private goods but rather on public goods that reduce under-investments in pro-growth factors such as rural human capital. 9 A beginning has been made to political econometric analysis of the World Bank's agricultural distortions database in Anderson (2010). 27 Fourth, the strongest benefits would come from agricultural reform, underscoring the economic and social importance of securing reforms for that sector in addition to manufacturing, notwithstanding the political sensitivities involved. There are more-direct and hence more-efficient domestic policy instruments that could meet governments' poverty and hunger Millennium Development Goals than trade policies, but generally they are more of a net drain on treasury finances. This is particularly so for those governments of low-income countries which still rely heavily on trade tax revenue. One solution to that dilemma is to expand aid-for-trade funding as part of official development assistance programs. Finally, the finding from most of the national case studies that domestic reform on its own can be a way of reducing poverty and inequality suggests that developing countries need not hold back on national reforms while negotiations in the World Trade Organization's Doha Round and other international accords continue. It also suggests that developing countries have little to gain, and potentially much to lose from a poverty alleviating perspective, from negotiating exemptions or delays in national reforms in the framework of WTO multilateral agreements. 28 REFERENCES Ahmed, S.A., N.S. Diffenbaugh and T.W. Hertel (2009), `Climate Volatility Deepens Poverty Vulnerability in Developing Countries', Environmental Research Letters, 4, 8 (available at www.iop.org/EJ/erl, doi:10.1088/1748-9326/4/3/034004). Anderson, K. (ed.) (2009), Distortions to Agricultural Incentives: A Global Perspective, 1955-2007 (London: Palgrave Macmillan and Washington DC: World Bank). Anderson, K. (ed.) (2010), The Political Economy of Agricultural Price Distortions (Cambridge and New York: Cambridge University Press). Anderson, K., J. Cockburn and W. Martin (eds.) (2010), Agricultural Distortions, Inequality and Poverty (Washington DC: World Bank). Freely available as an e-book at www.worldbank.org/agdistortions Anderson, K., M. Kurzweil, W. Martin, D. Sandri and E. Valenzuela (2008), `Measuring Distortions to Agricultural Incentives, Revisited', World Trade Review, 7, 4, 1-30. Anderson, K. and W. Martin (2005), `Agricultural Trade Reform and the Doha Development Agenda', The World Economy, 28, 9, 1301-27. Anderson, K. and E. Valenzuela (2008), Global Estimates of Distortions to Agricultural Incentives, 1955 to 2007, data spreadsheets at www.worldbank.org/agdistortions Anderson, K., E. Valenzuela and D. van der Mensbrugghe (2010), `Global Poverty Effects of Agricultural and Trade Policies Using the LINKAGE Model', Ch. 2 in Anderson, Cockburn and Martin (2010). Arndt, C. and J. Thurlow (2010), `Mozambique', Ch. 10 in Anderson, Cockburn and Martin (2010). 29 Atkinson, A.B. and A. Brandolini (2004), `Global World Inequality: Absolute, Relative or Intermediate', Paper prepared for the 28th Conference of the International Association for Research in Income and Wealth, Cork, Ireland, 22-28 August. Bourguignon, F., M. Bussolo and L.P. da Silva (eds.) (2008), The Impact of Macroeconomic Policies on Poverty and Income Distribution: Macro-Micro Evaluation Techniques and Tools (London: Palgrave Macmillan and Washington DC: World Bank). Bourguignon, F., V. Levin, and D. Rosenblatt (2004), `Declining International Inequality and Economic Divergence: Reviewing the Evidence Through Different Lenses', Economic Internationale, 4Q, 13-26. Bourguignon, F. and C. Morrison (2002), `Inequality Among World Citizens: 1820-1992', American Economic Review 92(4): 727-44. Bussolo, M., R. De Hoyos and D. Medvedev (2008), `Is the Developing World Catching Up? Global Convergence and National Rising Dispersion', Policy Research Working Paper 4733, World Bank, Washington DC, September. Bussolo, M., R. De Hoyos and D. Medvedev (2010), `Global Income Distribution and Poverty in the Absence of Trade Distortions Using the GIDD Model', Ch. 3 in Anderson, Cockburn and Martin (2010). Chen, S. and M. Ravallion (2010), `The Developing World is Poorer Than We Thought, But No Less Successful in the Fight Against Poverty', Ch. 10 in Anand, S., P. Segal and J.E. Stiglitz (eds.), Debates on the Measurement of Global Poverty (London and New York: Oxford University Press). Cicowiez, M., C. Diaz-Bonilla and E. Diaz-Bonilla (2010), `Argentina', Ch. 12 in Anderson, Cockburn and Martin (2010). Cororaton, C., E. Corong and J. Cockburn (2010), `Philippines', Ch. 8 in Anderson, Cockburn and Martin (2010). 30 Cororaton, C. and D. Orden (2010), `Pakistan', Ch. 7 in Anderson, Cockburn and Martin (2010). Cranfield, J.A.L., P.V. Preckel, J.S. Eales and T.W. Hertel (2003), `Estimating Consumer Demand Across the Development Spectrum: Maximum Likelihood Estimates of an Implicit Direct Additivity Model', Journal of Development Economics, 68, 289-307. Davis, B., P. Winters and G. Carletto (2009), `A Cross-Country Comparison of Rural Income-Generating Activities', mimeo, Agricultural Development Economics Division, Food and Agriculture Organization, Rome, January. Ferreira Filho, J. and M. Horridge (2010), `Brazil', Ch. 13 in Anderson, Cockburn and Martin (2010). Ferreira, F.H.G. and M. Ravallion (2009), `Global Poverty and Inequality: A Review of the Evidence', Review of Economics and Statistics 51: 436-44. Francois, J.F. and B. Hoekman (2010), `Services Trade and Policy', Journal of Economic Literature, 48, 3, 642-92, September. Francois, J.F. and W. Martin (2010), `Ex Ante Assessments of the Welfare Impacts of Trade Reforms with Numerical Models', Ch. 13 (pp. 379-434) in H. Beladi and E.K. Choi (eds.), New Developments in Computable General Equilibrium Analysis for Trade Policy (London: Emerald Group Publishing). Harrison, A.E. (ed.) (2007), Globalization and Poverty (Chicago: University of Chicago Press). Herault, N. and J. Thurlow (2010), `South Africa', Ch. 11 in Anderson, Cockburn and Martin (2010). Helpman, E., O. Itskhoki and S.J. Redding (2010), `Inequality and Unemployment in a Global Economy', Econometrica, 78, 4, 1239-83, July. 31 Hertel, T.W. and R. Keeney (2010), `Inequality and Poverty Impacts of Trade-related Policies Using the GTAP Model', Ch. 4 in Anderson, Cockburn and Martin (2010). Hertel, T.W. and L.A. Winters (2005), `Estimating the Poverty Impacts of a Prospective Doha Development Agenda', The World Economy, 28, 8, 1057-71. Hertel, T.W. and L.A. Winters (eds.) (2006), Poverty and the WTO: Impacts of the Doha Development Agenda (London: Palgrave Macmillan and Washington DC: World Bank). Horridge, M. and F. Zhai (2006), `Shocking a Single-Country CGE Model with Export Prices and Quantities from a Rest of the world Model', pp. 94­104 in Hertel and Winters (2006). Hoxha, I., S. Kalemli-Ozcan and D. Vollrath (2009), `How Big are the Gains From International Financial Integration?' NBER Working Paper 14636, Cambridge MA, January. Milanovic, B. (2002), `True World Income Distribution, 1988 and 1993: First Calculation Based on Household Surveys Alone', Economic Journal, 112, 476, 51-92. Milanovic, B. (2005): Worlds Apart: Measuring International and Global Inequality (Princeton: Princeton University Press). Milanovic, B. (2006), `Global Income Inequality: What It Is and Why It Matters', World Economics, 7, 1, 131-57, January-March. Narayanan, B.G. and Walmsley, T.L. (eds.) (2008), Global Trade, Assistance, and Production: The GTAP 7 Data Base, West Lafayette: Center for Global Trade Analysis, Purdue University. Downloadable at http://www.gtap.org. Otsuka, K., J.P. Estudillo and Y. Sawada (eds.) (2009), Rural Poverty and Income Dynamics in Asia and Africa (London and New York: Routledge). 32 Otsuka, K. and T. Yamano (2006), `Introduction to the Special Issue on the Role of Nonfarm Income in Poverty Reduction: Evidence from Asia and East Africa', Agricultural Economics, 35 (supplement), 373-97. PovcalNet (2008), http://go.worldbank.org/7X6J3S7K90, World Bank, Washington DC, accessed May. Ravallion, M. (2006), `Looking Beyond Averages in the Trade and Poverty Debate', World Development, 34, 8, 1374-92. Ravallion, M. (2009), `Bailing Out the World's Poorest', Challenge, 52, 2, 55-80. Ravallion, M., S. Chen and P. Sangraula (2007), `New Evidence on the Urbanization of Poverty', Population and Development Review, 33, 4, 667-702. Sala-i Martin, X. (2006), `The World Distribution of Income: Falling Poverty and Convergence, Period', Quarterly Journal of Economics, 121, 2, 351-97. Sanchez, M. and R. Vos (2010), `Nicaragua', Ch. 13 in Anderson, Cockburn and Martin (2010). Valenzuela, E. and K. Anderson (2008), `Alternative Agricultural Price Distortions for CGE Analysis of Developing Countries, 2004 and 1980-84', Research Memorandum No. 13, Center for Global Trade Analysis, Purdue University, West Lafayette IN, December. Freely downloadable at https://www.gtap.agecon.purdue.edu/resources/res_display.asp?RecordID=2925 Valenzuela, E., D. van der Mensbrugghe and K. Anderson (2009), `General Equilibrium Effects of Price Distortions on Global Markets, Farm Incomes and Welfare', Ch. 13 in Anderson (2009). van der Mensbrugghe, D. (2005), `LINKAGE Technical Reference Document: Version 6.0', Unpublished, World Bank, Washington DC, January 2005. Accessible at www.worldbank.org/prospects/linkagemodel 33 Walmsley, T. and L.A. Winters (2005), `Relaxing the Restrictions on the Temporary Movement of Natural Persons: A Simulation Analysis', Journal of Economic Integration, 20, 4, 688- 726. Warr, P.G. (2010a), `Indonesia', Ch. 6 in Anderson, Cockburn and Martin (2010). Warr, P.G. (2010b), `Thailand', Ch. 9 in Anderson, Cockburn and Martin (2010). World Bank (2005), Global Economic Prospects 2006: Economic Implications of Remittances and Migration (Washington DC: World Bank). World Bank (2007), World Development Report 2008: Agriculture for Development (Washington DC: World Bank). Zhai, F. and T. Hertel (2010), `China', Ch. 5 in Anderson, Cockburn and Martin (2010). 1 Table 1: Global poverty and inequality, by region, 1981 to 2005 (number and percent of people on less than $1/day in 2005 PPP) 1981 1987 1993 1999 2005 Share of Index of poor (%) income who are inequality rural, (Gini 2002 coefficient) 2004a No. of people (million): Sub-Saharan Africa 157 202 247 299 299 69 n.a. East Asia and Pacific 948 598 600 425 180 85 0.37 of which China 730 412 444 302 106 90 0.36 South Asia 387 384 341 359 350 75 0.35 of which India 296 285 280 270 267 74 0.33 Latin America and Caribbean 27 35 34 40 28 34 0.52 Rest of world 9 9 15 23 22 50 n.a. WORLD 1528 1228 1237 1146 879 74 n.a. East+South Asia's share of world 87 80 76 68 60 Share of population (percent): Sub-Saharan Africa 40 42 44 46 39 East Asia and Pacific 69 39 36 24 10 of which China 74 38 38 24 8 South Asia 42 37 29 27 24 of which India 42 36 31 27 24 Latin America and Caribbean 7 8 7 8 5 WORLD 42 30 27 23 16 a Gini coefficient is the population-weighted cross-country average of national Gini coefficients in the region for the nearest available year to 2004. Source: Chen and Ravallion (2010) except for rural share (Ravallion, Chen and Sangraula 2007) and Gini coefficient (PovcalNet 2008). 2 Table 2: Effects of full global liberalization of agricultural and all merchandise trade on national economic welfare and real GDP, by country and region, using the LINKAGE model (percent change relative to benchmark data) All sectors' Agricultural All sectors' policies policies policies Economic Agric Non-ag Agric Non-ag welfare(EV) GDP GDP GDP GDP East and South Asia 0.9 -0.3 0.7 0.5 2.9 of which China 0.2 2.8 0.2 5.7 3.0 India -0.2 -6.1 1.4 -8.3 -0.3 Africa 0.2 0.1 0.8 -0.9 0.0 Latin America 1.0 36.3 2.8 37.0 2.3 All developing countries 0.9 5.4 1.0 5.6 1.9 Eastern Europe &Central Asia 1.2 -4.4 0.3 -5.2 0.3 All high-income countries 0.5 -13.8 0.2 -14.7 0.1 World total 0.6 -1.0 0.4 -1.2 0.5 Source: LINKAGE model simulations from Anderson, Valenzuela and van der Mensbrugghe (2010). 3 Table 3: Effects of full global merchandise trade liberalization on reala factor prices, by country and region, using the LINKAGE model (relative to the benchmark data, percent) Nominal change deflated by Nominal change in unskilled aggregate CPI wages deflated by: Food Capitalb Landb and Skilled user user Aggregate Food clothing wages cost cost CPI CPI CPI East and South Asia 3.4 3.0 -1.8 3.2 4.6 4.8 Africa 4.7 4.3 0.1 4.4 5.8 6.9 Latin America 1.4 1.9 21.1 4.5 2.4 4.1 All developing countries 3.0 2.9 1.6 3.5 5.5 5.9 Eastern Europe & Central Asia 3.2 2.6 -4.5 1.7 4.2 4.5 High-income countries 1.0 0.5 -17.9 0.2 3.3 3.3 World total 1.3 1.2 -3.1 0.9 3.6 3.8 a Real changes deflated by a Consumer Price Index or CPI. b The user cost of capital and land represents the subsidy inclusive rental cost. Source: LINKAGE model simulations from Anderson, Valenzuela and van der Mensbrugghe (2010). 1 Table 4: Effects of full global merchandise trade liberalization on the incidence of extreme poverty using the LINKAGE model Average Baseline Change in number of Change in number unskilled headcount poor from baseline of poor from wage New levels, $1/day New levels, $2/day levels baseline levels change, Number Number reala $1/day $2/day Headcount of poor, Headcount of poor, $1/day, $2/day, $1/day, $2/day, (%) (%) (%) (%) million (%) million million million % % East Asia 4.4 9 37 8 151 34 632 -17 -52 -10.3 -7.6 China 2.1 10 35 9 123 34 440 -5 -12 -4.0 -2.7 Other East Asia 8.1 9 50 6 29 42 192 -12 -40 -30.1 -17.1 South Asia -1.9 31 77 32 454 78 1124 8 8 1.8 0.7 India -3.8 34 80 36 386 82 883 15 15 4.2 1.7 Other South Asia 4.0 29 94 26 68 92 241 -8 -7 -9.9 -2.7 Sub Saharan Africa 5.3 41 72 39 287 70 508 -11 -14 -3.8 -2.7 Latin America 4.1 9 22 8 44 21 115 -3 -6 -6.8 -4.7 Middle East & North Africa 14.3 1 20 1 3 13 40 -2 -19 -36.4 -32.7 Developing country total 5.9 18 48 18 944 46 2462 -26 -87 -2.7 -3.4 Developing excl. China 6.5 21 52 20 820 50 2022 -21 -74 -2.5 -4.7 East Europe & Central Asia 4.5 1 10 1 4 9 43 -0 -4 -6.8 -8.0 a Nominal unskilled wage deflated by the food and clothing CPI Source: LINKAGE model simulations from Anderson, Valenzuela and van der Mensbrugghe (2010). 2 Table 5: Effects of full global liberalization of agricultural and all merchandise trade on the incidence of extreme poverty using the GTAP model (percentage point change using $1 a day poverty line) Alternative tax Default tax replacement replacement (poor are exempt) Agriculture-only Nonagriculture- All merchandise All merchandise reform only reform reform reform Asia Bangladesh -0.3 0.5 0.3 -5.3 Indonesia -1.1 0.5 -0.6 -5.2 Philippines -1.4 0.4 -1.0 -6.4 Thailand -11.2 0.9 -10.3 -28.1 Vietnam -0.5 -5.3 -5.7 -23.6 Africa Malawi -1.6 -0.3 -1.9 -5.6 Mozambique -1.2 0.2 -1.0 -4.3 Uganda -0.0 0.1 0.1 -6.0 Zambia -0.0 0.1 0.1 -2.0 Latin America Brazil -2.5 0.4 -2.2 -10.0 Chile -4.8 0.1 -4.6 -12.3 Columbia -0.7 0.6 -0.1 -4.1 Mexico 0.8 0.4 1.1 -0.5 Peru -0.6 -0.2 -0.8 -5.2 Venezuela 0.2 0.7 0.9 -2.1 Unweighted averages: -Asia -2.9 -0.6 -3.5 -13.7 -Africa -0.7 0.1 -0.7 -4.5 -Latin Amer -1.3 0.3 -1.0 -5.7 -All 15 DCs -1.7 -0.1 -1.7 -8.0 Source: Hertel and Keeney (2010, Table 4.5). 3 Table 6: Impact of reform on the incidence of extreme poverty, national case studies (percentage point change using national or $1 a day poverty line) (a) rural poverty Base Agriculture-only reform Nonagriculture-only reform All merchandise reform (%) Unilateral R of W Global Unilateral R of W Global Unilateral R of W Global China($2/day) 58 0.3 -1.4 -1.1 0.2 -0.5 -0.3 0.5 -1.9 -1.4 Indonesia 29 0.1 -1.1 -1.1 -0.2 -3.2 -3.3 -0.1 -4.3 -4.4 Pakistan 38 -1.4 -0.1 -1.5 -6.2 -1.1 -7.1 -7.6 -1.2 -8.6 Philippines 49 0.0 -0.6 -0.3 0.6 -0.3 0.2 0.6 -0.9 -0.1 Thailand 30 0.3 -1.6 -1.3 -3.8 0.7 -3.1 -3.5 -0.9 -4.4 Mozambique 36 -1.6 0.0 -1.6 -0.5 -1.5 -2.0 -2.1 -1.5 -3.6 South Africa 17 -0.3 -0.3 -0.7 -0.8 0.0 -0.8 -1.1 -0.4 -1.4 Brazil Nicaragua 63 -0.7 0.3 -0.4 -0.6 -0.3 -0.9 -1.3 0.0 -1.3 (b) urban poverty Base Agriculture-only reform Nonagriculture-only reform All merchandise reform (%) Unilateral R of W Global Unilateral R of W Global Unilateral R of W Global China($2/day) 3 0.0 0.0 0.0 0.0 -0.1 -0.1 0.0 -0.1 -0.1 Indonesia 12 -0.1 -0.3 -0.4 -0.1 -1.7 -1.8 -0.2 -2.0 -2.2 Pakistan 20 -2.4 -0.1 -2.7 4.7 -1.4 3.1 2.3 -1.5 0.4 Philippines 19 0.8 -0.9 -0.2 1.2 -0.7 0.3 2.0 -1.6 0.1 Thailand 6 0.0 -0.8 -0.7 -3.3 0.2 -3.2 -3.3 -0.6 -3.9 Mozambique 37 -0.5 0.0 -0.5 -0.4 -1.3 -1.7 -0.9 -1.3 -2.2 South Africa 4 -0.1 -0.2 -0.3 -0.4 0.0 -0.4 -0.5 -0.2 -0.7 Brazil Nicaragua 27 0.3 -0.5 -0.2 -1.0 1.4 0.4 -0.7 0.9 0.2 4 Table 6 (continued): Impact of reform on the incidence of extreme poverty, national case studies (percentage point change using national or $1 a day poverty line) (c)total poverty Base Agriculture-only reform Nonagriculture-only reform All merchandise reform (%) Unilateral R of W Global Unilateral R of W Global Unilateral R of W Global China($2/day) 36 0.2 -0.8 -0.6 0.1 -0.4 -0.3 0.3 -1.2 -0.9 Indonesia 23 -0.0 -0.8 -0.8 -0.1 -2.7 -2.8 -0.1 -3.5 -3.6 Pakistan 31 -1.6 -0.1 -1.8 -3.6 -1.2 -4.6 -5.2 -1.3 -6.4 Philippines 34 0.4 -0.6 -0.1 0.7 -0.3 0.2 1.1 -0.9 0.1 Thailand 14 0.1 -1.1 -0.8 -3.5 0.4 -3.3 -3.4 -0.7 -4.1 Mozambique 36 -1.3 0.0 -1.3 -0.4 -1.4 -1.8 -1.7 -1.4 -3.1 South Africa 10 -0.2 -0.3 -0.5 -0.6 -0.1 -0.6 -0.8 -0.3 -1.1 Brazil 31 -0.5 -2.3 -2.8 -0.4 -0.1 -0.5 -0.9 -2.4 -3.5 Nicaragua 41 -0.1 -0.2 -0.3 -0.9 0.8 -0.1 -1.0 0.6 -0.4 Unweighted averages: -Asia 28 -0.2 -0.7 (-2.9)-0.8 -1.2 -0.8 (-0.6)-2.2 -1.5 -1.6 (-3.5)-3.0 -Africa 32 -0.8 -0.2 (-0.7)-0.9 -0.5 -0.7 (0.1)-1.2 -1.3 -0.9 (-0.7)-2.1 -Latin Am. 36 -0.3 -1.3 (-1.3)-1.6 -0.7 0.4 (0.3)-0.3 -1.0 -0.9 (-1.0)-2.0 -All 9 DCs 43 -0.4 -0.6 (-1.7)-1.0 -0.9 -0.6 (-0.1)-1.5 -1.3 -1.2 (-1.7)-2.6 a Numbers in italics for individual countries are implied assuming linearity holds; numbers do not always add because of either rounding or interaction effects Source: Country case studies in Parts II to IV of Anderson, Cockburn and Martin (2010) plus (in the case of the unbolded numbers in brackets in the final 4 rows), from Hertel and Keeney (2010) as reported in the last 4 rows of Table 5 above. 5 Table 7: Impact of reform on the incidence of income inequality, national case studies (percentage point change in Gini Coefficient) (a) rural Base Agriculture-only reform Nonagriculture-only reform All merchandise reform (%) Unilateral R of W Global Unilateral R of W Global Unilateral R of W Global China 0.32 0.0 -0.2 -0.2 0.0 0.0 0.0 0.0 -0.2 -0.2 Indonesia 0.29 0.0 0.0 0.0 0.1 0.0 0.1 0.1 0.0 0.1 Pakistan 0.26 -0.1 -0.0 -0.1 0.3 0.0 0.3 0.2 -0.0 0.2 Philippines 0.43 0.2 -0.1 0.1 0.3 0.0 0.1 0.5 -0.1 0.2 Thailand 0.33 0.0 0.5 0.5 0.4 0.0 0.4 0.4 0.5 0.9 Mozambique South Africa 0.63 -0.1 -0.1 -0.2 -0.3 0.0 -0.3 -0.4 -0.1 -0.5 Brazil Nicaragua (b) urban Base Agriculture-only reform Nonagriculture-only reform All merchandise reform (%) Unilateral R of W Global Unilateral R of W Global Unilateral R of W Global China 0.26 0.0 0.1 0.1 0.0 -0.1 -0.1 0.0 0.0 0.0 Indonesia 0.36 0.0 -0.1 -0.1 0.3 0.3 0.6 0.3 0.2 0.5 Pakistan 0.40 -0.1 -0.0 -0.1 -1.9 0.0 -1.9 -2.0 -0.0 -2.0 Philippines 0.48 0.3 -0.2 0.1 0.1 0.0 0.1 0.4 -0.2 0.2 Thailand 0.15 0.1 0.6 0.7 0.5 0.0 0.5 0.6 0.6 1.2 Mozambique South Africa 0.62 -0.1 -0.1 -0.2 -0.5 0.0 -0.5 -0.6 -0.1 -0.7 Brazil Nicaragua 6 Table 7 (continued): Impact of reform on the incidence of income inequality, national case studies (percentage point change in Gini Coefficient) (c)total Base Agriculture-only reform Nonagriculture-only reform All merchandise reform (%) Unilateral R of W Global Unilateral R of W Global Unilateral R of W Global China 0.44 0.1 -0.4 -0.3 0.0 -0.1 -0.1 0.1 -0.5 -0.4 Indonesia 0.34 0.0 -0.1 -0.1 0.2 0.2 0.4 0.2 0.1 0.3 Pakistan 0.34 -0.1 -0.0 -0.2 -3.2 -0.1 -3.1 -3.3 -0.1 -3.3 Philippines 0.51 0.3 -0.2 0.1 0.1 0.0 0.1 0.4 -0.2 0.2 Thailand 0.34 0.1 0.7 0.8 0.4 0.0 0.4 0.5 0.7 1.2 Mozambique 0.48 -1.2 -0.1 -1.3 -0.3 0.2 -0.1 -1.5 0.1 -1.4 South Africa 0.67 -0.1 -0.1 -0.2 -0.4 0.0 -0.4 -0.5 -0.1 -0.6 Brazil 0.58 -0.2 -1.4 -1.6 0.1 -0.1 0.0 -0.1 -1.5 -1.7 Nicaragua 0.53 -0.1 0.1 0.0 -0.1 -0.2 -0.3 -0.2 -0.1 -0.3 Unweighted averages: -Asia 0.39 0.1 -0.0 0.1 -0.5 0.0 -0.5 -0.4 -0.0 -0.4 -Africa 0.58 -0.7 -0.1 -0.8 -0.4 0.1 -0.3 -1.0 -0.0 -1.0 -Latin Am. 0.56 -0.2 -0.7 -0.8 0.0 -0.2 -0.1 -0.2 -0.8 -1.0 -All 9 DCs 0.59 -0.2 -0.2 -0.4 -0.3 -0.0 -0.3 -0.5 -0.2 -0.7 a Numbers in italics are implied assuming linearity holds; numbers do not always add because of either rounding or interaction effects Source: Country case studies in Parts II to IV of Anderson, Cockburn and Martin (2010).