Institut für Agrarpolitik und Marktforschung der Justus-Liebig-Universität Giessen Arbeitsbericht Nr. 46 SARAH MAAS What Explains EU Food Aid? An Econometric Analysis Gießen 2008 Bestell-Nr. 08/3 Anschrift des Instituts: Senckenbergstr. 3 35390 GIESSEN Tel. Nr. 0641/99-37020; Fax: 0641/99-37029 email: Sekretariat.Marktlehre@agrar.uni-giessen.de Abstract This study investigates the determinants of EU food aid allocation among recipients for the period 1993-2003 from a political economy perspective. Thereby the analysis differentiates between emergency and non-emergency food aid which is, to the author’s best knowledge, not covered by preceding studies. In a first step, following previous studies investigating the political economy of aid allocation, a hypothetical market for food aid supply and demand is constructed. In order to explain the allocation of food aid among recipients several factors that may determine food aid supply and demand are considered. Beside humanitarian and development objectives, the examination includes commercial and external interests as well as bureaucratic factors in order to account for the fact that the EU has often been criticized to use food aid for commercial and political rather than for humanitarian purposes. In a second step, empirical evidence concerning the validity of the theoretically derived factors is provided. Using two different econometric approaches commonly applied in the foreign aid literature – the two-part and the Tobit model – the analysis finds that EU food aid is, to a large extent, determined by humanitarian objectives. Most importantly, both emergency and non- emergency food aid allocation of the EU strongly respond to the recipient’s average calorie supply per capita. Additionally, the results show that non-emergency food aid is targeted towards countries with low economic and non-economic well-being. In spite of the clear humanitarian orientation, EU food aid is also determined by external interests and bureaucratic inertia. As concerns the former a positive relationship between food aid donations of other donors and those of the EU suggests that the EU competes with other donors for political influence in the recipient countries. No evidence is found that EU food aid is allocated with a view to reward countries with sound policies and institutions as the widespread discussion about the role of good governance in aid allocation suggests. Instead, EU emergency food aid allocation is biased towards non-democratic countries meaning that the EU, regardless of the recipient’s ideological view, responds to the needs of countries which may have a lower capacity to effectively help its people in case of food shortages. Finally, neither the common belief that emergency food aid is particularly responsive to recipients’ needs nor the hypothesis that emergency food aid is intrinsically political as compared to non-emergency food aid is confirmed by the analysis.   Table of Contents List of Figures ............................................................................................................... III List of Tables ................................................................................................................ IV List of Abbreviations ..................................................................................................... V 1 Introduction ............................................................................................................. 1 1.1 Purpose of the Study ........................................................................................................... 1 1.2 Organization of the Study .................................................................................................... 3 2 Food Aid as an Aid Instrument .............................................................................. 4 2.1 Food Aid in Official Development Aid ............................................................................. 4 2.2 Types of Food Aid ................................................................................................................ 8 2.2.1 Program Food Aid ....................................................................................................... 10 2.2.2 Project Food Aid .......................................................................................................... 11 2.2.3 Emergency Food Aid ................................................................................................... 13 2.3 Sources of Food Aid .......................................................................................................... 14 3 Food Aid in the EU ................................................................................................ 16 3.1 Trends in EU Food Aid ...................................................................................................... 16 3.2 Historical Development of EU Food Aid Policy .................................................................. 22 3.2.1 EU Food Aid Policy until 1996 ..................................................................................... 22 3.2.2 EU Food Aid Policy Reform of 1996 ........................................................................... 24 3.3 Institutional Framework of EU Food Aid Policy .................................................................. 25 4 Political Economy of EU Food Aid ...................................................................... 28 4.1 Public Choice Theory – A Political Market of Food Aid ..................................................... 28 4.2 Determinants of Food Aid Supply ...................................................................................... 32 4.2.1 Humanitarian and Development Interests ................................................................... 33 4.2.2 Surplus Disposal as Driving Factor ............................................................................. 34 4.2.3 Commercial Interests .................................................................................................. 35 4.2.4 External Interests ........................................................................................................ 36 4.2.5 Bureaucratic Interests ................................................................................................. 40 4.3 Determinants of Food Aid Demand ................................................................................... 41 4.3.1 Humanitarian Determinants ........................................................................................ 42 4.3.2 Economic Determinants .............................................................................................. 42 4.3.3 External Determinants ................................................................................................. 44 4.3.4 Bureaucratic Determinants .......................................................................................... 44 4.4 Impact of the Recent Rise in World Food Prices on EU Food Aid Market Equilibrium ...... 45 5 Empirical Analysis of EU Food Aid ..................................................................... 49 5.1 Estimation Methods ........................................................................................................... 49 5.1.1 Ordinary Least Squares .............................................................................................. 50 5.1.2 Binary Choice Models ................................................................................................. 52 5.1.3 Tobit Model ................................................................................................................. 55 5.2 Data and Specification of the EU Food Aid Model ............................................................. 58 5.2.1 Dependent Variable .................................................................................................... 58 5.2.2 Explanatory Variables ................................................................................................. 59 5.2.2.1 Recipients’ Needs ............................................................................................................... 59 5.2.2.2 Good Governance ............................................................................................................... 61 5.2.2.3 Donor Interests .................................................................................................................... 62 5.2.3 General Remarks on the Model Specification ............................................................. 66 5.3 Overview of Empirical Results of Precedent Studies on EU Food Aid .............................. 68 5.4 Empirical Results of Own Estimation ................................................................................. 70 5.4.1 Empirical Results for Project and Program Food Aid .................................................. 70 5.4.1.1 Project and Program Food Aid: Estimation Results of the Tobit Model .............................. 70 5.4.1.2 Project and Program Food Aid: Estimation Results of the Two-Part Model ....................... 73 5.4.2 Empirical Results for Emergency Food Aid Allocation ................................................ 78 5.4.2.1 Emergency Food Aid: Estimation Results of the Tobit Model ............................................. 78 5.4.2.2 Emergency Food Aid: Estimation Results of the Two-Part Model ...................................... 82 5.5 Discussion and Synthesis of the Results ........................................................................... 86 6 Conclusion ............................................................................................................ 91 List of References ....................................................................................................... 94 Appendix .................................................................................................................... 103    II List of Figures Figure 1: Food Aid and Total ODA 1975-2006 .................................................................................... 5 Figure 2: Food Aid Deliveries to Recipient Regions in 1993-2007 ....................................................... 6 Figure 3: Food Aid Deliveries of Major Donor Countries in1993-2007 ................................................. 7 Figure 4: Development of the Three Categories of Food Aid 1995-2007 ............................................ 9 Figure 5: Development of the Three Categories of EU Food Aid 1988-2007 .................................... 17 Figure 6: Development of EU Project and Program Food Aid in Major Regions 1988-2006 ............. 18 Figure 7: Development of EU Emergency Food Aid to Major Regions 1988-2006 ............................ 20 Figure 8: Development of EU Food Aid Procurement 1988-2007 ...................................................... 22 Figure 9: Determination of the Optimal Development Aid Budget for Donor Countries ..................... 30 Figure 10: Effect of Shifts of the Supply and Demand Curve on the Food Aid Market Equilibrium .... 31 Figure 11: Effects of Rising World Food Prices on the Food Aid Market Equilibrium ........................ 45 Figure 12: Interdependencies in Food Aid Allocation between Recipient Countries .......................... 48 Figure A.13: Development of EU Program Food Aid in Major Regions 1988-2006 ......................... 104 Figure A.14: Development of EU Project Food Aid in Major Regions 1988-2006 ............................ 104 III List of Tables Table 1: Share of Food Aid in EU Development Aid .......................................................................... 16 Table 2: Allocation of EU Project and Program Food Aid across Major Recipients in 1988-2006 ..... 19 Table 3: Allocation of EU Emergency Food Aid across Major Recipients in 1988-2006 .................... 21 Table 4: List of Variables ................................................................................................................... 64 Table 5: Expected Signs of the Explanatory Variables of EU Food Aid ............................................. 65 Table 6: Tobit Model Estimation Results for EU Project and Program Food Aid Allocation ............... 71 Table 7: Marginal Effects of Significant Variables of the Tobit Model on EU Project and Program Food Aid Allocation (based on equation 3) ......................................................................... 73 Table 8: Two-Part Model Estimation Results for EU Project and Program Food Aid Allocation ........ 74 Table 9: Marginal Effects of Significant Variables of the REF Probit Model on EU Project and Program Food Aid Allocation .............................................................................................. 77 Table 10: Tobit Model Estimation Results for EU Emergency Food Aid Allocation ........................... 78 Table 11: Marginal Effects of Significant Variables of the Tobit Model on EU Emergency Food Aid Allocation (based on equation 11) ....................................................................................... 82 Table 12: Two-Part Model Estimation Results for EU Emergency Food Aid Allocation ..................... 83 Table 13: Marginal Effects of Significant Variables of the RE Probit Model on EU Emergency Food Aid Allocation ...................................................................................................................... 85 Table 14: Summary of Results: Significant Impacts of the Explanatory Variables on EU Food Aid ... 87 Table A.15: List of Countries ........................................................................................................... 103 Table A.16: Allocation of EU Project and Program Food Aid across Recipients in 1988-2006 ........ 105 Table A.17: Allocation of EU Emergency Food Aid across Recipients in 1988-2006 ...................... 107 Table A.18: Data Sources ............................................................................................................... 109 Table A.19: Descriptive Statistics .................................................................................................... 110  IV List of Abbreviations ACP Countries from Africa, Caribbean, and Pacific (Lomé member states) Bn Billion CAP Common Agricultural Policy CIS Commonwealth of Independent States DG Directorate General EC European Commission ECHO European Community Humanitarian Office EM-DAT Emergency Event Database EU European Union FAC Food Aid Convention FAO Food and Agriculture Organisation of the United Nations FE Fixed Effects (Panel Model) GEE Generalized Estimation Equations GM Genetically Modified HDI Human Development Index i.i.d. independently and identically distributed IMF International Monetary Fund INTERFAIS International Food Aid Information System kg kilograms LDC Least Developed Country LDV Lagged Dependent Variable LIFDC Low Income, Food Deficient Country LPM Linear Probability Model MDG Millennium Development Goals Mio Million ML Maximum Likelihood NGO Nongovernmental Organization ODA Official Development Assistance OECD/DAC Development Assistance Committee of the Organisation for Economic Cooperation and Development OLS Ordinary Least Squares PL 480 Public Law 480 PQLI Physical Quality of Life Index RE Random Effects (Panel Model) SIPRI Stockholm International Peace Research Institute t tons Tsd Thousand UK United Kingdom V VI UN United Nations UNDP United Nations Development Programme UNHCR United Nations High Commissioner for Refugees UNICEF United Nations Children’s Fund US United States USD US Dollar WFP World Food Programme WTO World Trade Organisation What Determines EU Food Aid? An Econometric Analysis 1 Introduction 1.1 Purpose of the Study Over one decade after the World Food Summit in 1996 and the establishment of the United Nations Millennium Development Goals (MDGs) virtually no progress in reducing hunger has been made, even though the sources of hunger are widely known today. The main problem of hunger is not a capacity for food production insufficient to provide everyone an adequate diet but instead an unequal distribution of resources, i.e. a lack of purchasing power of the poor. In order to solve this disparity problem, resources in form of food aid are redistributed from industrialized towards developing countries. However, even the considerable amount of food aid allocated to developing countries during the past four decades did not help to alleviate hunger in these countries. The consequences are shrinking food aid levels as well as losses in public and political attention devoted to food aid during the past decade. Today, the development of world food markets poses a new challenge to the fight against hunger. Increasing food demand in developing and transition countries, especially in China and India, the redirection of staple crops, such as maize and wheat, from food uses to biofuel production in the European Union (EU) and United States (US) and increasingly frequent occurrence of droughts, floods and storms as an outcome of climate change have led to a strong rise of world food prices in the recent past. As soaring world prices for agricultural commodities (occasionally called “agflation”) limit the purchasing power, many people in developing countries are exposed to an increasing risk of food insecurity, causing riots such as the so-called Tortilla- Crisis in Mexico in December 2007. The following excerpt illustrates the consequences of these developments for the poorest. “A wave of food price inflation is moving through the world, leaving riots and shaken governments in its wake. For the first time in 30 years, food protests are erupting in many places at once. […] The measures of today’s crisis are misery and malnutrition. […] Those on $1 a day are cutting back on meat, vegetables and one or two meals, so they can afford one bowl. The desperate – those on 50 cents a day – face disaster.” THE ECONOMIST (2008) In view of this development, the topic of food aid is once again attracting the attention of not only academics and practitioners of international development, but also governments of donor countries are increasingly under pressure. As concerns the latter, in recognition of the burden INTRODUCTION the rise of world market prices poses on developing countries the European Commission (EC) and the US raised the food aid budget by €117 Mio and US$770 Mio, respectively (EC 2008a, FAO 2008a). This increase in financial support certainly is essential, but is it enough to provide the most vulnerable and food insecure people adequate access to food? An effective provision of food aid is an essential precondition to reduce hunger and food insecurity in developing countries. However, even though industrialized countries have considerable capacity to help those in need, there is widespread belief that food aid is not efficiently targeted at the poorest and most food-insecure countries as donors follow objectives other than humanitarian concerns, such as trade and geopolitical interests. Especially the US and the EU as the largest bilateral donors have been criticized by scientists as well as practitioners for their mostly self-serving food aid systems. With a view to future challenges to food security and scarce resources, inefficient food aid systems will be even less tolerated in the future than they were in the past. As the second largest food aid donor the EU has the moral obligation and the capability to contribute to the reduction of hunger and poverty in developing countries, but only if most vulnerable and food insecure people are addressed. In this context, this study investigates the effectiveness and efficiency of EU food aid, or in other words whether EU food aid targets those countries most in need of food aid. Thus, this study aims at finding out whether the EU actually allocates food aid with respect to postulated – humanitarian – allocation criteria necessary to meet the nutritional needs of the very poor, or whether self-interests as for instance commercial and external objectives are at the center of EU food aid allocation. The thesis has been stimulated by the study of PRINZ (1994), but extends that study in several aspects. First, additional economic and political factors that might determine the allocation of food aid are considered in the analysis. Second, PRINZ neglects the panel structure of the data and merely estimates pooled regression models. This shortcoming is corrected by the econometric analysis at hand which employs Ordinary Least Squares and Limited Dependent Variable models explicitly taking into account the panel data structure. Moreover, a larger coverage of countries is possible due to improved data availability for former Soviet states in the time period considered in this study (1993-2003). Finally, the analysis is augmented by disaggregating total food aid flows into emergency and non-emergency (i.e. project and program) food aid as the objectives of these types of food aid clearly differ from one another. This proceeding further aims at testing the widespread hypothesis that emergency food aid is more responsive to recipients’ needs and to a lesser extent motivated by donors’ interests than project and program food aid. As it is rather assumed on theoretical considerations than based 2 INTRODUCTION 3 on empirical evidence this hypothesis is important to examine, especially in view of the rising importance of emergency food aid vis-à-vis project and program food aid. 1.2 Organization of the Study The thesis is organized as follows. Chapter 2 gives detailed information about the role and importance of food aid in total development aid as well as major food aid donors and recipient regions. Further, main characteristics of different types of food aid – i.e. emergency, project, and program food aid – are reviewed, including a discussion of its virtues and limitations in reducing poverty and food insecurity. Chapter 3 illustrates the particular food aid system of the EU. First, the development of food aid supply, again separated into emergency, project, and program food aid, during the past 20 years is presented. Afterwards, the development of EU food aid policy and the regulative system are reviewed and put into the context of the development of world agricultural markets and agricultural policy. In Chapter 4, EU food aid is subject to a theoretical examination. More precisely, a political market for EU food aid is constructed by means of Public Choice Theory. Similar to economic theory, it is postulated that food aid allocation can be explained by food aid supply and demand. Thus, several determinants of food aid supply of donor countries and food aid demand of recipients are established. The chapter concludes with an application of the theory in order to explain the impacts of the recent food price rise on the food aid market equilibrium. Based on these theoretical considerations, Chapter 5 empirically investigates which factors determine EU food aid allocation by applying econometric techniques. To this end, the econometric methods used in the analysis are first briefly introduced. After providing detailed information about the model specification and data used in the analysis the estimation results are presented. This is followed by a synthesis and a profound discussion of the estimation results as well as an evaluation of the virtues and limitations of the analysis. The thesis concludes with a summary of the main findings and implications. 2 Food Aid as an Aid Instrument Today as in the past, economic well-being dramatically differs across countries. On one side of the scale, the rich countries’ populations live in abundance of resources – although certainly not entirely free from poverty – whereas, on the other side of the scale, a large share of the population in poor countries suffers from hunger and malnutrition. Helping poor countries in alleviating hunger and poverty is thus a moral obligation for rich countries. For this reason a considerable amount of food aid is distributed from industrialized to developing countries. However, as food aid is only a fraction of total foreign aid this chapter sets food aid in the perspective of Official Development Assistance (ODA) and gives an introduction to basic characteristics and principles of the food aid system on the global level. First, the importance and effectiveness of food aid as an instrument of ODA is discussed followed by a brief presentation of major food aid donors and destinations. Going into more depth, the different food aid types – emergency, project, and program food aid – and food aid procurement are then illustrated, including their major virtues and shortcomings in reducing hunger and poverty. 2.1 Food Aid in Official Development Aid It is well known that enough food is produced globally to provide each person an adequate diet for a healthy life. Aggregate figures of food availability, however, hide the fact that many people lack the access to food sources to meet their nutritional needs. Recent estimates of the Food and Agriculture Organization of the United Nations (FAO) find that 852 Million (Mio) people – corresponding to one-eighth of the world population – suffer from hunger (FAO 2006a). The FAO established the goal of a per capita intake of 2 350 calories per day which is currently not met in approximately 60 countries (MAXWELL and BARRETT 2005 p. 8). Although food insecurity is often visible at the country level, more often it is not. More precisely, even if the calorie supply in a country as a whole is in principle sufficient to provide everyone an adequate diet, particular groups of the population still may suffer from hunger. This is for example the case in the Democratic Republic of Congo, Somalia, Afghanistan, and Burundi where more than two-thirds of the population were undernourished during the period 1988-2000 (FAO 2002). Basically, the reason for this problem is an unequal distribution of resources both between and within countries. If poor people lack the purchasing power to compensate chronically insufficient production and short-term production shortfalls by food purchases on local and international markets the consequence is hunger. Since developing countries often do not have the means, and sometimes not the will, to help their food insecure population, rich countries are obliged to support poor countries in overcoming this problem. In order to alleviate poverty and hunger, FOOD AID AS AN AID INSTRUMENT Figure 1: Food Aid and Total ODA 1975-2006 0 20 40 60 80 100 120 140 Total ODA Food Aid Bn USD Time Source: OECD/DAC (2008).  considerable efforts in form of ODA are made – with a sharply increasing volume in the past decade as shown in Figure 1. Food aid is a subcategory of ODA which is defined by BARRETT and MAXWELL (2005 p. 254) as the “provision of food commodities for free or on highly concessional terms to individuals or institutions within one country by foreign donors”. As such food aid is characterized by the (1) international sourcing of (2) resources allocated on concessional terms (3) in form of or for the provision of food. This definition unambiguously distinguishes food aid from other assistance programs, intra-national food transfers and commercial international food trade (BARRETT and MAXWELL 2005 p. 5). One important aspect of the above definition of food aid is that a considerable grant element is required to qualify food deliveries as food aid. To provide “considerable grant element” does not necessarily and against conventional wisdom mean that food aid is allocated for free. Food aid may also be provided as loans, as for instance parts of US food aid, meaning that food aid is a transfer of resources on at least concessional terms from high income countries to developing countries. Thus, food aid fundamentally is an entry into the recipients’ balance of payments in order to stimulate its development and economic growth (BARRETT 1998, WFP 2008). In spite of the high public awareness of food aid as compared to other forms of foreign aid which is merely due to its visibility and public interest in emergency situations, the share of food aid in total ODA is very small. Figure 1 shows that – contrarily to the dramatic rising trend in ODA 5 FOOD AID AS AN AID INSTRUMENT since 2000 – the amount of food aid deliveries decreased since the early 1990s. That is, after averaging 6.6% of total ODA in the period 1975-2000 food aid deliveries worldwide declined to only 1.8% in 2001-2006. Also the share of food aid in world food production and commercial food trade is of minor importance accounting for only 0.2% of cereal production and 2% of cereal trade in 2007 (WFP 2008, note that the share refers to cereal not food production and imports). As food aid is an instrument of development aid with special focus on relieving food insecurity and hunger 90% of food aid goes to low income food deficient countries (LIFDCs) and 61.5% to Least Developed Countries (LDCs). However, even in the group of LIFDCs food aid deliveries only amount to 0.4% of total cereal production, but to 5.5% of LIFDCs cereal imports in 2007 (WFP 2008). On average, the share of food aid products in total food consumption equaled 2.5% in developing countries in 2001-2003. However, this share may be much larger for selected economies. In Eritrea, for instance, the share of food aid in total food consumption reached 45.7% in 2001- 2003 implying that food security in selected LIFDCs depends to a substantial degree on food aid (FAO 2006b). At the regional level Asia was historically the largest recipient of food aid.1 With its rapid Figure 2: Food Aid Deliveries to Recipient Regions in 1993-2007 0 1 2 3 4 5 6 7 8 SSA Asia EE & CIS LA & Caribbean NA & ME 6 SSA=Sub-Saharan Africa, LA & Caribbean=Latin America and the Caribbean, EE & CIS= Eastern Europe and the CIS, NA & ME=North Africa and Middle East. Source: WFP/INTERFAIS (several years).                                                              Mio t Time 1 Please note that food aid data used throughout this study are obtained by the International Food Aid Information System (INTERFAIS) database and has been provided courtesy of the World Food Programme (WFP). Cereals are specified in grain equivalents, other commodities in actual quantities. FOOD AID AS AN AID INSTRUMENT economic growth in the 1960s and 1970s, Asia was replaced as the largest food aid recipient by Sub-Saharan Africa in the mid-1970s which suffered from economic stagnation and civil strife (BARRETT and MAXWELL 2005 pp. 10). As illustrated in Figure 2, this constellation remained in principle unchanged until today although the share of food aid delivered to Sub-Saharan Africa increased in recent years from 28% in 1993 to its all time high of 60% in 2006 (with 4.1 Mio tons (t) this is not the highest amount absolute terms). This increase, however, comes at other regions’ expense, most notably Asia whose share in total food aid decreased from 40% in 1998 to merely 18% in 2006, corresponding to 1.2 Mio t. The end of the Cold War in the early 1990s led to a sharp increase of food aid deliveries to the former Soviet Union, thereby temporarily cutting down food aid flows to Africa and Asia. However, in the following years the amount of food aid to Eastern Europe and the Commonwealth of Independent States (CIS) fell rapidly, with a sole exception in 1999 when the Russian ruble crisis again triggered large food aid deliveries to Eastern Europe. While the regions of destination varied considerably over time, the donors of food aid remained largely unchanged. Figure 3 presents major bilateral food aid donors in 1993-2006. Although food aid is given by many industrialized countries, the major share is concentrated on few donors. The US is historically the largest donor of food aid with food aid shipments accounting for 95% of total food aid in the mid-1960s. The US share in total food aid declined to one-half to two-thirds in the 1980s, averaging 72% in the period 1996-2006. Even though US food aid flows in total terms declined steadily since 1999, its share in total food aid reached 75% in 2006. This Figure 3: Food Aid Deliveries of Major Donor Countries in1993-2007 0 2 4 6 8 10 12 USA EU Canada Japan Australia Mio t Time Source: WFP/INTERFAIS (several years). 7 FOOD AID AS AN AID INSTRUMENT means that US food aid decreased at a slower pace than other donors’ food aid contributions. Food aid donations of the EU as the second largest donor averaged 15.5% in 1996-2006. However, since the mid-1990s EU food aid donations were considerably reduced as a result of major reforms of its food aid policy, which will be reviewed in greater detail in Chapter 3. Other important bilateral food aid donors are Canada and Australia, averaging 4.7% and 3.1% of total food aid flows in the period 1996-2006. Further, Japan delivers 7.9% of total food aid which is primarily headed towards its Asian neighbor countries (WFP/INTERFAIS several years, COLDING and PINSTRUP-ANDERSEN 2002 p. 198). Even though food aid may in principle enhance food security – given that it is targeted to vulnerable and food insecure households, supports agricultural infrastructure and is allocated to avoid adverse price effects – during the past two decades food aid faced sharp criticism which finally led to the decline of food aid in total development aid. As compared to private, well- functioning markets, food aid is considered as an inefficient and unsustainable instrument for both reducing poverty and food insecurity and addressing shortfalls in food availability in developing countries (CLAY et al. 1998, CLAY 2000, DEL NINNO et al. 2007). Moreover, COLDING and PINSTRUP-ANDERSEN (2002) point out that hunger is a problem of mutually interactive factors which cannot be solved by food aid alone due to its limited capacity in combating the causes of hunger, i.e. poverty. However, even though the ability of food aid to reduce structural food insecurity caused by poverty is questionable, the importance of food aid in brief periods of localized food shortages which cannot be corrected by local markets should not be undervalued (GILLIGAN and HODDINOTT 2007). Regardless of its minor share in total development aid, BARRETT and MAXWELL (2005 p. 6) emphasize that food aid is an important instrument to “make a big difference at the margin by relieving shortfalls in food availability that contribute directly to widespread, often acute hunger, malnutrition and undernutrition […]”. Thus, the virtue of food aid is short-term relief for highly vulnerable and food insecure people in poorly functioning markets. Obviously, there is a common understanding in the literature that food aid effectiveness largely depends on the purpose of food aid which is strongly associated with the type of food aid. In order to evaluate EU food aid allocation it is thus important to be aware of the basic concepts, virtues and limitations of the different types of food aid – emergency, project, and program food aid – which will be discussed in the following sections. 2.2 Types of Food Aid So far food aid has been treated as a homogenous instrument of development aid. In fact, food aid is commonly differentiated in emergency, project, and program food aid. Project and program 8 FOOD AID AS AN AID INSTRUMENT food aid (also called development or non-emergency food aid) target at long-term development and poverty reduction in the recipient country. In contrast, emergency food aid (also called humanitarian or relief food aid) is seen as an instrument of relief in humanitarian catastrophes caused by natural or man-made disasters (GABBERT 2000 p. 29, BARRETT and MAXWELL 2005 p. 13). Figure 4: Development of the Three Categories of Food Aid 1995-2007 0 2 4 6 8 10 12 14 16 9 Program Project Emergency Source: WFP/INTERFAIS (2008).  Traditionally, program food aid was by far the most important type of food aid until about 1990. In the 1990s, however, substantial criticism of program food aid led to a sharp decline in program food aid associated with an increasing importance of emergency food aid. As a consequence, emergency food aid replaced program food aid as the major type of food aid in 1996, with the exception of 1999 where the ruble crisis led to massive program food aid deliveries to Russia. In fact, Figure 4 reveals that even project food aid, which historically made up the smallest fraction of food aid, overtook program food aid in 2001. Due to its continuing downward trend the share of program food aid in total food aid merely equaled 14% in 2007, while emergency and project food aid accounted for 62% and 24%, respectively (WFP 2008). In order to understand these dramatic changes it is necessary to have a closer look at major virtues and shortcomings of emergency, project, and program food aid.2                                                              Mio t Time 2 The following treatment is restricted to the most commonly mentioned aspects of food aid and is thus far from being exhaustive. For a more detailed treatment see for instance COLDING and PINSTRUP-ANDERSEN (2002).  FOOD AID AS AN AID INSTRUMENT 2.2.1 Program Food Aid Program food aid, which was first introduced by the US in 1954 under Public Law (PL) 480, constituted the first type of official food aid. Program food aid primarily aims at reducing poverty in developing countries. It is directly given to the recipient’s government on a bilateral basis (government-to-government) either as a grant or on concessional terms. The share of food aid on concessional terms is, however, almost negligible as the share of program food aid given as grants increased from 61.7% in 1972 to 98% in 2006. Only the US continues to supply program food aid as loans while the EU and other donors deliver program food aid exclusively as grants (COLDING and PINSTRUP-ANDERSEN 2002, WFP 2007). Program food aid may be provided as food aid in kind or in form of cash transfers. If the former is the case, the recipient’s government has the possibility to directly transfer commodities as gifts in kind to poor and food insecure households or to monetize food aid, which means that the commodities are sold on the market with the revenue usually going to the government. In the case of monetization, food aid in kind is essentially the same as cash transfers which are budgetary resources used to generate revenues in domestic currency in form of counterpart funds. These serve as a source of public revenue and may reduce balance-of-payment deficits of the recipient country. The recipient’s government can use these additional resources either to provide public goods, for instance physical infrastructure, agricultural research or health and education services, or to redistribute income to poor population groups (HERRMANN and PRINZ 1993, BARRETT 1998, GABBERT and WEIKARD 2000, COLDING and PINSTRUP-ANDERSEN 2002). Program food aid typically includes formal agreements which oblige the recipient to channel resources into uses approved by the donor (“conditionality”). These agreements, however, are unlikely to impose any restriction on the use of food aid. In fact, in the case of cash transfers and monetization of food aid in kind recipients are free to use program food aid donations without significant restrictions due to the fungibility of budgetary resources. That is, resources can be re- directed from the intended and agreed purpose to other purposes (BARRETT and MAXWELL 2005 p. 15, HAGEN 2006). Consequently, BARRETT and MAXWELL (2005 p. 13) argue that program food aid essentially is “nothing else than foreign aid provided in the form of food”. Unfortunately, program food aid is characterized by several shortcomings, which made it a frequent object of criticism and finally led to the sharp decline of program food aid since the 1990s. First, on the microeconomic level program food aid may cause adverse producer incentives, the so-called disincentive effects. By adding to the domestic food supply, food aid deliveries in kind push local food prices down thereby reducing the returns and competitiveness of agricultural production. This consequently leads to a decreasing food production in recipient 10 FOOD AID AS AN AID INSTRUMENT countries and lower farm incomes. If one additionally takes into account that the poorest parts of the population often live in rural areas and crucially depend on agricultural production, program food aid may rather hurt than support the poorest, most vulnerable and food insecure groups (SCHULTZ 1960, MAXWELL and SINGER 1979, MAXWELL 1991 pp. 66, CLAY et al. 1998). Recent studies, however, question the existence of disincentive effects arguing that program food aid has no or only minor impacts on local production (BARRETT 2002a, LOWDER 2004, ABDULAI et al. 2005). Second, program food aid is considered a cost-ineffective transfer of resources as the trans- action cost of food aid in kind, which include transport, handling, storage and administration, are likely to be higher than transaction cost of aid donated in cash. Generally speaking, the cost- effectiveness of program food aid is determined by the source of food aid, which will be discussed in greater detail in Section 2.3. CLAY et al. (1996) finds that EU program food aid is most cost-effective when the sources of food are either commercial imports arranged by the recipient or purchased nearby the affected region (i.e. triangular transactions or local purchases). Conversely, food aid procured on EU markets is less likely to be cost-effective. Therefore, if more flexibility in sourcing or budget support instead of in kind deliveries are allowed, a substantial reduction in costs, or an increase in food aid contributions at equal expenses is possible (COLDING and PINSTRUP-ANDERSEN 2002). Third, it is argued that recipients’ governments often use the financial resources given as food aid for activities other than distributing food to food insecure people. In this case program food aid fails to meet its target to increase the food security of the most vulnerable groups of the recipients’ population while at the same time benefiting the political powerful (BARRETT and MAXWELL 2005). Despite the strong criticism of program food aid in general, the role of program food aid in alleviating poverty and food insecurity depend on several factors, most importantly on the demand and supply situation of food staples, economic policies in the recipient country, the targeting, and on the source of food aid commodities (HERRMANN and PRINZ 1993, JAYNE et al. 2001, BARRETT 2002b). A conclusive evaluation of program food aid thus requires a more detailed assessment of the particular arrangement of food aid and the specific conditions in each recipient country. 2.2.2 Project Food Aid The major objective of project food aid is to support particular projects related to poverty- reduction and disaster-prevention activities, thereby improving the nutritional situation of 11 FOOD AID AS AN AID INSTRUMENT vulnerable, often remote, parts of the population. Most commonly known projects supported by project food aid are food for work programs, school feeding initiatives, and supplemental feeding programs for mothers and children (GABBERT 2000 p. 30). As project and program food aid aim at long-term reduction of poverty and food insecurity both types of food aid share several characteristics. Similar to program food aid, project food aid may be donated as cash transfers or as food aid in kind which is either freely distributed to vulnerable and food insecure groups or sold on the market (WFP 2007). Thus, project food aid supports local interventions rather than governments and is therefore more frequently distributed in kind than monetized. Due to its use in local interventions project food aid is usually provided multilaterally through international organizations and nongovernmental organizations (NGOs), most importantly the World Food Programme (WFP) (COLDING and PINSTRUP-ANDERSEN 2002). However, since program food aid today is also channeled through international organizations and NGOs and project food aid is also given to recipients’ governments, it is increasingly difficult to differentiate between both types of food aid (BARRETT and MAXWELL 2005 p. 13). Despite its strong similarities, the stable amount of project food aid, as opposed to the sharp decline in program food aid, in recent years implies that project food aid has some desirable properties when compared to program food aid. Most importantly, project food aid is often distributed in kind and thus particularly suitable in situations where food is more valuable than cash transfers, that is, in acute or chronic food insecurity due to insufficient supply and poorly functioning markets. But even in efficient markets it is possible that food aid in kind is preferable to cash transfers as project food aid has proved to be more appropriate in targeting vulnerable groups. Also, if properly targeted direct distribution of food aid commodities to households particularly benefit rural households and has less adverse effects on those who are excluded from food aid which is due to the fact that in kind deliveries have weaker effects on the demand side than cash transfers (BASU 1996, ARNDT and TARP 2001). Finally, given that women usually control food whereas cash is more likely to be managed by men COLDING and PINSTRUP- ANDERSEN (2002) argue that food aid in kind strengthens the power of women over household resources. This is expected to have a positive impact on food security and child nutrition as evidence suggests that women distribute household resources more equitable between household members than men. 12 Regardless of its merits, project food aid is also subject to criticism. First, similar to program food aid, project food aid may have disincentive effects by adversely affecting food prices and farmers’ income. Second, food transfers are expensive, logistically difficult to handle and thus often ill-timed. Finally, no improvement in food security can be expected if the composition of food products do not meet the (micro-)nutritional needs of the target group or if supplementary FOOD AID AS AN AID INSTRUMENT factors, such as hygiene, nutrition knowledge and adequate health care, are required but not provided (COLDING and PINSTRUP-ANDERSEN 2002). The effectiveness and efficiency of project food aid in reducing poverty and improving food security is thus largely determined by the particular design and implementation of each project. 2.2.3 Emergency Food Aid The general intention of emergency food aid is to support people affected by some kind of acute emergency such as floods, droughts, hurricanes, economic shocks, or civil war and conflicts, and who are not able to meet their nutritional needs for a limited period of time. Food aid increases food security among those affected by natural or man-made disasters, including refugees and internally displaced persons, and serves as an important instrument of emergency relief and post-crisis rehabilitation. In the case of natural disasters, well-targeted food aid may also help to prevent emergency-induced migration and political instability (BARRETT 1998, GABBERT and WEIKKARD 2000, COLDING and PINSTRUP-ANDERSEN 2002, BARRETT and MAXWELL 2005 p. 14). Emergency food aid is clearly distinct from project and program food aid as it is exclusively provided as gift and generally aims at short-term relief. The share of emergency food aid in total food aid experienced a sharp increase at the end of the 1990s. Basically, this expansion reflects two developments. Besides the sharp reduction of program food aid, an increasing frequency of man-made and natural disasters, with the latter usually being accredited to climate change, more often requires short-term food aid deliveries in order to reduce the burden of affected people (CHRISTENSEN 1999, GABBERT 2000 p. 32, ADDISON 2002, BARRETT and MAXWELL 2006). In contrast to program and project food aid, which may have serious negative impacts on the economy of the recipients, most importantly disincentive effects, emergency food aid is usually exempted from this criticism. This is due to the fact that food aid is essential in natural and man- made emergencies when people do not have access to food neither by own production nor by purchasing food on the market. Even though it is commonly acknowledged that emergency food aid has no or only minor adverse effects on the recipient’s economy and agricultural sector, it nevertheless shares some shortcomings of project and program food aid. Emergency food aid is rather costly, often ill-timed and does not take into account the needs of the recipient countries in an adequate manner, especially with a view to the provision of micronutrients. Additionally, even though well-intended, humanitarian food aid during war may prolong and intensify conflict if the relief falls in the hands of belligerents. However, only 2% of food aid channeled through the WFP is lost this way (STEWART 1998, ADDISON 2002). Also, the impacts of emergency food aid 13 FOOD AID AS AN AID INSTRUMENT on the recipient countries are controversially discussed lately. Recent case studies of GELAN (2007) and CHABOT and DOROSH (2007) find disincentive effects of food aid on domestic agricultural production in countries receiving a large share of food aid for emergency relief, e.g. Ethiopia. Nonetheless, it is commonly agreed upon that emergency food aid is more demand- driven and responsive to the needs of the targeted groups than program and project food aid (NEUMAYER 2005). Altogether, as already mentioned, the effectiveness of food aid on the recipient’s economy crucially depends on the sourcing of food aid commodities. As an assessment of food aid is not possible without taking into account its procurement, the following section will briefly address the major types of procurement and its development. 2.3 Sources of Food Aid Commodities used as food aid may be procured in three different ways. The largest share of food aid is procured as direct transfers, meaning that food commodities originating from donor countries are shipped to the recipients. Commodities may also be purchased in surplus regions of the recipient country itself – so-called local purchases – or sourced in third countries, which are mostly developing countries other than the recipient – known as triangular transactions (CLAY and BENSON 1991 pp. 143). The dominance of direct transfers is due to its role as a convenient opportunity to dispose agricultural surpluses of donor countries which was the main purpose of the US and the EU food aid as the largest bilateral donors prior to the early 1990s. However, with major modifications in both agricultural and food aid policy of the EU in the mid-1990s the structure of food aid procurement slowly began to change (BARRETT and MAXWELL 2005 pp. 55). Moreover, the increase in procurement outside the donor country is further supported by the shift from program towards emergency food aid. That is to say, triangular transaction and local purchases are more likely to be linked to emergency and relief responses because of problems and shortages associated with the ad-hoc sourcing of appropriate commodities (CLAY 1994 p. 20). Although direct transfers still made up for 59% of total food aid (3.5 Mio t) in 2007, its share in total food aid recently declined. While the EU almost eliminated direct transfers with only 1% in 2007, virtually all of US food aid is still procured in the US (99.3% in 2007). As the amount of direct transfers declined, local purchases and triangular transactions gained in importance and most recently accounted for, respectively, 17% and 24% of total food aid (1.0 and 1.4 Mio t in 2007) (WFP 2008). 14 FOOD AID AS AN AID INSTRUMENT 15 From the donors’ as well as the recipients’ point of view this is a welcomed development since commodities purchased from inside or nearby the affected, food insecure region may benefit for the recipient country – or other developing countries in the case of triangular transactions – as well as the donor in several ways. First, delivery lags are reduced as commodities are not shipped from the donor country to the recipient but purchased locally. Second, transportation costs are reduced leading to higher cost efficiency in procurement and delivery and, third, locally bought food usually suits the tastes and nutritional habits of the local population. Finally, local purchases and triangular transactions stimulate demand thereby supporting farmers and the agricultural sector in the region where the purchases are made. This minimizes or even eliminates the potentially disincentive and devastating effects of food aid in the recipient region or country (CLAY and BENSON 1991 p. 143, PRINZ 1994 p. 21, ZERBE 2004, BARRETT and MAXWELL 2005 p. 59). To sum up, food aid is a complex and dynamic system. The effects and effectiveness of food aid depend on myriad factors, most importantly the type and procurement of food aid. In recent years major changes in food aid donations highlights two basic concerns. On the one hand, the rise in emergency food aid mirrors the inappropriateness of food aid as a single means of long- term poverty reduction, which may even have adverse effects on the recipient’s agricultural sector, whereas the importance of food aid as short-term relief was underscored in the past. On the other hand, the increasing share of triangular transactions and local purchases in times of shrinking agricultural surpluses implies that donor interests other than development assistance may play a significant role in food aid allocation, as for example agricultural sector and geopolitical interests. This is supported by the fact that major food aid donors represent the world’s leading food exporters and geopolitical influential and powerful democracies (BARRETT and MAXWELL 2005 p. 12). 3 Food Aid in the EU The previous chapter already indicates that food aid allocation generally does not only serve recipients’ needs but also multiple donor interests. As this chapter will show, the EU is no exemption from this. However, in order to make inferences about the determinants of EU food aid it is inevitable to understand its genesis and its structure as well as how it has evolved over time. Therefore, based on the previous introduction to the global food aid system and its basic concepts, virtues, and limitations, this chapter reviews the particular food aid regime of the EU. First, the development of EU food aid flows is illustrated, followed by a discussion of EU food aid policy and regulation. Thereby, the objectives of the EU in form of stated and non-stated intentions of food aid are highlighted. One remark is appropriate at this point. There are basically two forms of European food aid. The first form refers to the collective allocation of food aid of all EU members and is officially granted by the EC. As for the second form, single EU member states independently provide food aid to developing countries which is, however, not covered in this study as food aid of individual member states is not subject to the EU regulation but to the sovereignty of the member states. Instead, focus is given to the first form which is usually labeled EC food aid. However, throughout this study it will be referred to as EU food aid. 3.1 Trends in EU Food Aid Due to sharp criticism of food aid, in particular food aid delivered by the US and the EU, and falling agricultural surpluses during the past two decades, EU food aid was subject to major changes in both the total amount of food aid provided and its structure. Even though food aid deliveries during the period 1976-2006 averaged 15.5% of total EU foreign aid, Table 1 shows Table 1: Share of Food Aid in EU Development Aid Year Total ODA (Mio USD) Food aid (Mio USD) Share of food aid in ODA (%) 1978 716.33 219.35 30.62 1982 862.11 326.65 37.89 1986 1 413.41 182.69 12.93 1990 2 495.68 399.88 16.02 1994 4 396.37 n.a. n.a. 1998 4 461.55 363.54 8.15 2002 5 101.89 317.20 6.22 2006 9 366.68 275.67 2.94 Ø 1976-2006 3 337.29 297.20 15.50 Ø 1976-1990 1 270.54 276.88 25.05 Ø 1991-2006 5 274.86 361.44 7.49 n.a. = not available. Source: OECD/DAC (2008).   FOOD AID IN THE EU that the share of EU food aid actually fell from almost 40% of EU foreign aid in 1982 to merely 3% in 2006. The development of EU food aid in total EU development aid is thus consistent with the development of food aid at the global level. Figure 5: Development of the Three Categories of EU Food Aid 1988-2007 17 Source: WFP/INTERFAIS (2008).  0 500 1000 1500 2000 2500 3000 3500 4000 Project Program Emergency After reaching its peak with 3.8 Mio t in 1992 the total amount of EU food aid declined to only 0.6 Mio t in 2002. Figure 5 presents the development of food aid over the past 19 years decomposed into emergency, project, and program food aid. Over the whole period considered emergency food aid was the most important type of food aid with a share of 48.6% of total food aid, followed by program food aid with 36.9% and project food aid with 14.5%. However, the importance of each type of food aid was subject to major changes over the period, which is largely owed to a sharp decrease of the amount of program food aid (not at least triggered by the criticism of program food aid) whereas the amount of emergency and project food aid remained rather stable. Consequently, the share of program food aid fell behind that of emergency food aid in mid-1990s with the exception of 1999 where, again, a huge amount of program food aid was shipped to Russia in view of the ruble crisis. In order to analyze the patterns and determinants of EU food aid allocation a more comprehensive understanding of the structure of EU food aid is essential. With a view to the upcoming econometric analysis the Time Tsd t FOOD AID IN THE EU remainder of this section will present the development of EU food aid separated into two groups – emergency as well as project plus program food aid.3 The total amount of project and program food aid together shows large variations with the highest amount of 2.3 Mio t supplied in 1992. The high degree of volatility of project and program food aid is surprising since both types of food aid aim at long-term poverty reduction which would imply rather constant food aid flows. If the high degree of variability also is present at the country level, the unpredictability of aid flows may lead to losses in effectiveness (CELASUN and WALLISER 2008). Since 2000 the amount of project and program food aid was drastically reduced reaching 0.05 Mio t in 2004. Only recently, program food aid began to increase again, which is, however, solely caused by deliveries to the CIS. Coincidentally, the revitalization of program food aid to the CIS came along with the accession of the ten Central and Eastern European Countries to the EU. Figure 6 shows that important recipient regions of project and program food aid are Sub- Saharan Africa with 32% and Asia with 15% of total project and program food aid deliveries. However, with 36% of project and program food aid on average the CIS and Eastern Europe is the most important recipient region. An interesting relationship is to be mentioned with respect to Figure 6: Development of EU Project and Program Food Aid in Major Regions 1988-2006 0 200 400 600 800 1000 1200 1400 1600 SSA Asia LA & Caribbean EE & CIS NA & ME 18 SSA=Sub-Saharan Africa, LA & Caribbean=Latin America and the Caribbean, EE & CIS= Eastern Europe and the CIS, NA & ME=North Africa and Middle East. Source: WFP/INTERFAIS (2008).                                                              Tsd t Time 3 In the following project and program food aid are not treated separately but as one group as the econometric analysis in Chapter 5 examines project and program food aid jointly. Please note that in the following the expression project and program food aid actually refers to project plus program food aid. FOOD AID IN THE EU the dominance of the CIS as program and project recipient. In the aftermath of the Cold War the EU faced international criticism accusing the EU to divert a considerable part of development aid to the former communist states and thus, to withdraw important resources from other regions in need, such as Sub-Saharan Africa. The leading position of the CIS in Figure 6 seems to confirm this criticism as the amount of program food aid sharply increased after the fall of the Berlin Wall, whereas no food aid at all was delivered to the communist Soviet states during the Cold War. This relationship suggests that political interests strongly influence the allocation of food aid to particular countries (for a separate illustration of project and program food aid see Figures A.13 and A.14 in the appendix). However, the regional approach hides the fact that single countries within one region receive a much higher quantity of food aid than other countries of the same region. Therefore, a detailed record of project and program food aid deliveries to recipients in total as well as in per-capita terms in the period 1988-2006 is given in Table 2. For briefness, table 2 only includes the 15 most important recipients in terms of total food aid. An extension of this table including all recipients can be found in table A.16 in the appendix. Table 2 shows that CIS countries, most prominently the Russian Federation (rank 1), Poland (3), Albania (5), and Georgia (8), received huge amounts of food aid (summed over all years 1988 to 2006). Even though the EU provided food aid to the CIS in only a few years, these deliveries exceeded shipments to other countries that received food aid for a much larger part of the period. In the most extreme case Poland amounts to the third largest recipient of project and program food aid in total terms by receiving Table 2: Allocation of EU Project and Program Food Aid across Major Recipients in 1988-2006 Recipient country Total food aid Per-capita food aid Nb. of years receiving food aid Total Average p.a. Total Average p.a. (Tsd t) (kg) Russian Federation 1 836.16 96.64 12.53 0.66 6 Bangladesh 1 719.13 90.48 13.06 0.69 14 Poland 1 434.46 75.50 37.42 1.97 2 Egypt 1 135.56 59.77 17.97 0.95 11 Albania 1 076.23 56.64 340.23 17.91 3 Ethiopia 1 046.53 55.08 17.07 0.90 16 Mozambique 880.64 46.35 52.26 2.75 16 Georgia 853.59 44.93 172.65 9.09 4 Azerbaijan 622.15 32.74 79.91 4.21 5 Romania 523.91 27.57 23.33 1.23 4 Lithuania 509.04 26.79 142.57 7.50 2 Armenia 475.97 25.05 147.57 7.77 8 Latvia 466.70 24.56 188.66 9.93 2 Kyrgyzstan 364.42 19.18 76.84 4.04 6 Malawi 361.48 19.03 33.05 1.74 17 Source: Own calculation using data from WFP/INTERFAIS (2008). 19 FOOD AID IN THE EU huge amounts of food aid in only two years. This bias towards CIS countries does not change significantly when considering per-capita food aid. Most notably, Albania received 17.91 kilograms (kg) per capita p.a., followed by Latvia and Georgia with 9.93 kg and 9.09 kg per capita p.a., respectively. Further, the findings by, among others, PRINZ (1994 p. 40) and NEUMAYER (2003a p. 32, 2005) that countries with small population tend to receive more food aid in per capita terms than populous countries seems to be supported by Table 2. Similar to the case of project and program food aid, the total amount of emergency aid varied considerably during the past two decades with the largest supply of 1.5 Mio t in 1992 and the lowest supply of 0.4 Mio t in 1989 and 0.7 Mio t in 1996. The volatility of emergency food aid is anticipated as emergency food aid is designed as a local and short-term response to disasters. With respect to the regions and countries of destination, however, emergency food aid is clearly distinct from project and program food aid, suggesting that emergency food aid follows different allocation criteria than project and program food aid. As shown in Figure 7, the major recipient region is Sub-Saharan Africa which accounts on average for 69% of total emergency food aid during the period 1988-2006. The total amount of emergency food aid delivered to Sub-Saharan varies strongly with the highest amount of 1 Mio t delivered in 1991. Further important destinations are the CIS and Asia which on average receive 14% and 10% of EU emergency food aid, respectively. Consistent with the general pattern of food aid distribution, a considerable amount of EU emergency food aid was provided to the CIS following the breakup Figure 7: Development of EU Emergency Food Aid to Major Regions 1988-2006 0 200 400 600 800 1000 1200 SSA Asia LA & Caribbean EE & CIS NA & ME Tsd t Time SSA=Sub-Saharan Africa, LA & Caribbean=Latin America and the Caribbean, EE & CIS= Eastern Europe and the CIS, NA & ME=North Africa and Middle East. Source: WFP/INTERFAIS (2008). 20 FOOD AID IN THE EU Table 3: Allocation of EU Emergency Food Aid across Major Recipients in 1988-2006 Recipient country Total food aid Per-capita food aid Nb. of years receiving food aid Total Average p.a. Total Average p.a. (Tsd t) (kg) Ethiopia 3 487.42 183.55 56.88 2.99 19 Sudan 1 125.40 59.23 36.23 1.91 19 Serbia and Montenegro 700.28 36.86 94.62 4.98 13 Rwanda 596.16 31.38 79.83 4.20 13 Korea, Dem. Rep. 512.22 26.96 23.27 1.22 9 Eritrea 485.68 25.56 135.46 7.13 12 Malawi 470.67 24.77 43.03 2.26 13 Angola 419.21 22.06 32.17 1.69 19 Mozambique 362.91 19.10 21.54 1.13 11 Somalia 336.05 17.69 48.20 2.54 16 Zimbabwe 329.71 17.35 27.63 1.45 14 Afghanistan 298.88 15.73 24.85 1.31 13 Russian Federation 278.84 14.68 1.90 0.10 13 Occupied Palest. Territory 273.08 14.37 n.a. n.a. 19 Bangladesh 268.88 14.15 2.04 0.11 10 n.a. = not available. Source: Own calculation using data from WFP/INTERFAIS (2008). of the Soviet Union. However, after the peak in 1992 with 293 Thousand (Tsd) t (19.6%), emergency food aid delivered to CIS countries declined to 6.9 Tsd t in 2006, which corresponds to merely 1.9% of EU emergency food aid. The remaining regions – Latin America and the Caribbean as well as North Africa and the Middle East – receive a much smaller amount of emergency food aid. Table 3 illustrates that at the country level total emergency food aid deliveries differ enormously across countries with Ethiopia and Sudan receiving by far the largest amount. Due to the large variation in population size across countries the amount a country receives in total terms apparently does not allow any predictions about food aid available in per capita terms. In per capita terms Eritrea and Liberia receive the largest amount of emergency food aid with 7.13 and 5.21 kg, respectively. Again, an extended version of Table 3 can be found in the appendix (Table A.17). Finally, another subject to examine in greater depth is the development of EU food aid procurement. As already mentioned in Section 2.3 the sources of food aid changed considerably in the past decade as a result of major reforms of the EU’s food aid and agricultural policy. Figure 8 shows that the share of direct transfers started to decrease in 1992 which can be attributed to both a reduction in agricultural surpluses in the EU and the recognition of potential benefits of other forms of procurement (CLAY et al. 1994 p. 19). The decline of direct transfers was associated with an increasing share of local purchases and triangular transactions. Nowadays, EU food aid is mostly purchased from inside or nearby the affected, food insecure 21 FOOD AID IN THE EU region with local purchases and triangular transactions accounting for 34% and 65% of EU food aid in 2007, respectively. This is in sharp contrast to the US which sources only negligible parts of food aid in the recipient country and third countries (less than 0.5%) (WFP 2008). Figure 8: Development of EU Food Aid Procurement 1988-2007 0 500 1000 1500 2000 2500 3000 3500 Tsd t Time Direct Transfers Triangular Transactions Local Purchases Source: WFP/INTERFAIS (2008).  3.2 Historical Development of EU Food Aid Policy To understand the dynamic development of EU food aid outlined in the previous section it is essential to have a closer look on the evolvement of EU food aid policy during the past four decades. Starting in the mid-1960s, the EU advanced to the second largest bilateral food aid donor after the US. However, budget constraints in industrialized countries, reforms of the agricultural policy and the enclosure of agricultural issues in the trade negotiations of the World Trade Organisation (WTO) led to considerable changes of EU food aid policy in 1996. In order to make inferences about potential determinants of EU food aid allocation, the following sections give an overview of the development of EU food aid policy including major changes of the 1996 reform. 3.2.1 EU Food Aid Policy until 1996 EU food aid policy started in the 1960s when the US, which was the first to implement the concept of food aid under PL 480 in 1954, imposed pressure upon the then-European Community to share the “food aid burden”. The EC followed the request of the US in 1967 by 22 FOOD AID IN THE EU establishing EU food aid under the Wheat Trade Agreement and the Food Aid Convention (FAC) specifying annual minimum levels of food aid. The incentive to approve the US claim was, however, not solely driven by humanitarian concerns. Instead, it is commonly acknowledged that EU food aid policy of the early years was primarily motivated by the domestic agricultural sector and therefore strongly linked to the Common Agricultural Policy (CAP) (CATHIE 1981 p. 25, BARRETT and MAXWELL 2005 pp. 55). The EU’s CAP aimed at increasing farmers’ incomes by a complex support system including, inter alia, import protection and export subsidies. Thereby the CAP constituted the most expensive EU policy accounting for half of its budget in the 1960s. By giving the prospect of higher world market prices, a contribution to the food aid system was seen as an opportunity to reduce farmers’ payments and thus to relax the CAP’s burden on the EU budget (BARRETT and MAXWELL 2005 pp. 55). Further, the fact that the EU for the first time experienced nonmarketable cereal intervention stocks in 1967 surely contributed to the establishment of EU food aid (PRINZ 1994 p. 32). In the 1970s and 1980s, EU market intervention and farm support programs led to the production of large food surpluses thereby further depressing world market prices. This development had a strong impact on EU expenditures under CAP due to increasing payments required to guarantee farmers a high and stable income and to keep ever larger intervention stocks (CHRISTENSEN 1999, MAAS and SCHMITZ 2007). To dispose food surpluses in developing countries constrained by food shortages was thus a possibility to get the CAP expenditures under control by reducing costly intervention stocks and relieving the surplus situation on domestic food markets. The tight surplus situation on the EU market thus led to increasing amounts of food aid donations in the 1980s and resulted in the advancement of the EU to the second largest food aid donor. Unfortunately, with the disposal of agricultural surpluses being a primary motivation of EU food aid, the amount of food aid shipments depended on the world market situation. In times of strong demand and high food prices on international markets more food was sold and less provided as food aid. This had adverse effects on food insecure people in developing countries since food aid supply declined exactly when poor household could least afford to purchase food on the market and therefore food aid demand was high (COLDING and PINSTRUP-ANDERSEN 2002). 23 In the late 1980s and the 1990s two major concerns led to a rethinking in the EU concerning its agricultural policy, with major consequences for EU food aid. On the one hand, expenditures for European farmers under CAP steadily increased and made up 70% of the EU’s budget in the 1980s, despite a shrinking importance of agriculture in the Gross Domestic Product (GDP) and employment. On the other hand, agricultural policy continued to negatively affect world market FOOD AID IN THE EU price level and stability. This was a significant burden to food exporting developing countries by raising the risk of agricultural production and lowering farmers’ competitiveness, thereby increasing poverty in rural areas. The combination of internal budgetary concerns and external criticism, most strongly expressed in the WTO negotiations initiated several reforms of the CAP starting with the MacSharry reform in 1992. By reducing price support and introducing direct payments and land set-aside programs the reform was a first step towards an increasing market orientation of agricultural production. The result was a cutback of agricultural surpluses manifesting in shrinking intervention stocks in the 1990s, which considerably contributed to the reduction in food aid donations (BARRETT 1998, COLDING and PINSTRUP-ANDERSEN 2002, TANGERMANN 2006, MAAS and SCHMITZ 2007). Yet, the CAP was not the only policy under strain. Also the food aid system itself faced public criticism because of its mainly supply-driven character. Particularly the incorporation of food aid in the Doha-Round of the WTO trade negotiations set the policy under pressure. In 1996, EU food aid policy was finally subject to significant changes which will be discussed below. 3.2.2 EU Food Aid Policy Reform of 1996 In the mid-1990s, EU food aid policy moved from a supply-driven towards a more demand- based approach of food aid which more strongly takes into account recipients’ needs. In recognition that the causes of poverty are complex and to enhance coherence with EC development strategies, program and project food aid were integrated in the conceptual framework of an overall poverty reduction strategy, thus responding to the widespread criticism that food aid cannot solve the problem of hunger and poverty on its own (EC 2001, YOUNG and ABBOTT 2008). Moreover, with respect to its different objectives and to assure a prompt reaction to disasters, emergency food aid was fully separated from non-emergency food aid. One important change in EU food aid policy was the decreasing amount of food aid procured within the EU. This is to be seen with a view to the permanently lower grain stocks as a result of the ongoing CAP reform which was started by the MacSharry reform and continued with the Agenda 2000 and the Mid-Term Review in 2003. As the EU no longer needed to hold large cereal reserves, food aid commodities had to be purchased on domestic and international markets (COLDING and PINSTRUP-ANDERSEN 2002). This resulted in an increase in local purchases and triangular transactions which accounted for 99% of EU food aid in 2007, to the benefit of both the EU and recipient countries (ZERBE 2004, WFP 2008). The major change of EU food aid policy, however, was the transformation from food aid in kind shipments to cash transfers for the following reason. If food aid does not serve the aim of 24 FOOD AID IN THE EU surplus disposal but is purchased on the international market, it clearly competes with other, perhaps more cost-effective aid instruments. In cases where an improvement in food security does not necessarily depend on deliveries of commodities in kind, for instance in non- emergency situations and well-functioning markets, it is thus reasonable to reconsider the use of food aid as development instrument (CHRISTENSEN 1999, COLDING and PINSTRUP-ANDERSEN 2002). Consequently, if possible program food aid is provided in form of financial assistance which gradually replaces food aid in kind due to desirable impacts of cash transfers on cost- efficiency, local food markets, regional trade relations and economic development, and its flexibility to address country-specific dimensions of food insecurity. Food aid in kind is mainly used in emergency situations and in cases where it is best to target particularly vulnerable groups (EC 2001 pp. 10, EC 2004 p. 18). As a consequence, food aid donations in kind declined from 89% of total EU food aid in 1993 to 39% in 2000 (EC 2001 p. 19). Yet, the change from food aid in kind to cash transfers is not a panacea. Several problems remain. One important problem still unsolved is the strong variation in food aid supply. The supply of food aid is dependent on the budget situation of the donor countries on the one hand and on world market prices on the other hand. As EU food aid was transformed from a surplus- based into a budget-based program, food aid tends to decline in times of soaring world food prices due to increasing cost of providing food aid. With respect to budget constraints high world market prices of food aid commodities therefore translate into lower tonnage of food aid donations (COLDING and PINSTRUP-ANDERSEN 2002). Also, food aid is subject to the price volatility on world markets causing a greater variability of food aid donations. As a consequence, food aid deliveries are procyclical rather than countercyclical which is in sharp contrast to the fact that a declining food availability and security in recipient countries could be prevented by countercyclical food aid deliveries (BARRETT and HEISEY 2002). In order to make inferences about the determinants of EU food aid allocation it is not sufficient to account for observable allocation decisions and the development of food aid policy. Also, stated intentions of EU food aid as well as legislative and institutional requirements need to be considered. Therefore, the complex regulatory system of EU food aid is briefly reviewed in the next section. 3.3 Institutional Framework of EU Food Aid Policy Based on the Common Market Organization of Cereals, the early EU food aid policy was characterized by a strong link to agricultural trade arrangements through the FAC, stressing that surplus disposal was a major purpose of EU food aid. This is emphasized by Council 25 FOOD AID IN THE EU Regulation 67/120 which explicitly forbids the purchase of cereals in third countries if the interventions stocks contain sufficient inventory for food aid deliveries. In fact, the Regulation states that food aid commodities should originate from surplus regions to assure relief and should be purchased either on the market or directly from intervention stocks (PRINZ 1994 pp. 45, BARRETT and MAXWELL 2005 pp. 55). When EU food aid policy was increasingly confronted with criticism in the early 1980s, food aid policy was modified by several regulations with the aim to guarantee a better adaption of EU food aid to recipients’ needs. To enhance the cooperation and coordination of the EU and NGOs with respect to emergency food aid allocation a network of European NGOs for food aid, EuronAid, was founded in 1981 (PRINZ 1994 pp. 51). Moreover, Council Regulation 3331/82 issued in 1982 was the first Regulation to define the, mostly humanitarian, objectives of EU food aid. These include the improvement of food security of target groups, the intervention in emergencies and disasters, the support of recipients’ progress in food production and the contribution to a balanced economic and social development in recipient countries. The regulation explicitly specified three allocation criteria, i.e. basic food requirements, income per capita and the current account balance (CLAY et al. 1994 p. 35, PRINZ 1994 p. 47). Subsequent regulations, inter alia, define conditions to meet transport costs, the choice of commodities (besides cereals, skim milk powder and oilseeds are delivered to a measurable extent) and the implementation of early warning systems. Finally, a complex system consisting of international agreements and EU regulations determined annual food aid contributions with the competence being split between the Directorate General (DG) Agriculture and the DG Development. Unfortunately, the complexity of the food aid system and the division of competences resulted in conflict between both DGs and finally led to time-delayed reaction to emergencies and disasters (PRINZ 1994 pp. 51). 26 In order to assure a more timely and effective distribution of food aid and to improve poor control mechanisms of the generation and use of counterpart funds the procedure of food aid provision was significantly simplified in 1987. These changes, however, did not solve the problem of inefficient distribution of competences within the EC, so that the reform process was continued. EU emergency aid was reorganized and fully separated from project and program food aid in 1992. More precisely, the EC initiated the European Community Humanitarian Office (ECHO) which was henceforward responsible for the entire organizational and financial execution of EU emergency food aid in order to guarantee a timely response to emergencies by accelerating the process of food aid giving (PRINZ 1990, PRINZ 1994 p. 57, CLAY et al. 1996). According to the EC emergency food aid is distributed with a view to provide short-term food security in countries facing food crises. This includes the distribution of food to refugees and internally displaced FOOD AID IN THE EU 27 persons, assist nutritional recovery, support initial rehabilitation and recovery efforts as well as restore basic livelihoods and improve resilience to future shocks of vulnerable groups (EC 2008b). As concerns program and project food aid the responsibility for development operations was completely shifted from DG Agriculture to DG Development, stressing the increasing humanitarian orientation of EU food aid policy (EC 2004). The reform of EU food aid policy was finally completed in 1996 with the adoption of the Council Regulation 1292/96, replacing seven previous regulations. Basically, EU project and program food aid moved towards a holistic approach with the major objective being “to overcome problems of temporary food shortages, to manage post-crisis situations for food recovery and finally to address structural food security problems” (EC 2004 p. 6). The new regulation extends the list of allocation criteria, so that EU food aid is intended to be delivered with regard to: - the food situation and food shortages in the recipient country using human development and nutritional indicators; - per-capita income and the existence of particularly poor population groups; - social indicators of the welfare of the population; - the recipient’s balance-of-payments situation; - the economic and social impact and financial cost of the proposed operation; - the existence of a long-term policy on food security in the recipient country (EC 1996). To conclude, the food aid system of the EU has changed drastically since its establishment in 1967. While EU food aid was strongly influenced by agricultural sector interests first, with the primary intention to dispose surpluses, humanitarian objectives gained importance in subsequent years. This development was accompanied by major changes in the type and procurement of food aid, most importantly a sharp increase in emergency food aid, local purchases and triangular transactions. With major modifications in 1996, food aid policy intends to support long-term reduction of poverty and food insecurity as well as to respond effectively and coherently to recipients’ needs. However, with regard to the observed allocation of food aid illustrated in Section 3.1 – for instance the considerable amount of food aid flows to the CIS – doubts on a purely humanitarian oriented food aid supply remain. It is thus worthwhile to explore whether or not EU food aid actually succeeds in realizing well-intended statements and objectives. 4 Political Economy of EU Food Aid The EU provides food aid exclusively on a grant basis without receiving any monetary reward for its food aid contributions. This raises the question what motivation is behind the EU providing food aid to less wealthy countries at all? Why does the EU provide food aid to some countries every year in large amounts while others seldom receive food aid? Food aid allocation most likely is not a one-dimensional decision but serves many different objectives. Although the EC specified humanitarian criteria for food aid allocation, the previous chapter has cast doubt on whether these stated intentions are actually followed or whether there are other criteria which determine food aid allocation but are not stated by the EC. In fact, there is a widespread belief that, besides humanitarian and development concerns, donors also pursue economic as well as political and strategic interests in food aid allocation. Unfortunately, according to the TINBERGEN rule (1967) an effective policy requires one policy instrument for each objective. If food aid as a policy instrument serves several objectives – including others than humanitarian concerns – this will have a negative impact on its effectiveness as not those countries most in need receive food aid but countries that are economically and politically important for the EU. Thus, the multidimensional decision-making process fundamentally determines the effectiveness of EU food aid in reducing poverty and food insecurity abroad. In order to examine the determinants of EU food aid allocation an analytical framework of institutions and incentive structures is required and will be presented in the following section. To this end, potential determinants of food aid from the supply side, i.e. the donor’s point of view, as well as from the demand side, i.e. the recipient’s point of view, will be explored theoretically. Finally, the theoretical considerations will be illustrated on the basis of the recent price rise of food products on the world market. 4.1 Public Choice Theory – A Political Market of Food Aid In order to explain EU food aid allocation and the preceding decision-making process a Political Economy approach is chosen. In general, Political Economy is a theoretical construct to analyze the behavior of political institutions using Public Choice Theory, which is the application of economic theory to nonmarket decision-making, or in other words the application of economics to political science (MUELLER 1976). As such Public Choice Theory combines political and economic elements in order to analyze the behavior of political institutions, such as governments, political parties, voters or particular interest groups, in the presence of economic incentives. Well known applications of Public Choice Theory are, for instance, the median voter POLITICAL ECONOMY OF EU FOOD AID theorem and the “political business cycle” (MUELLER 2001). The Public Choice approach postulates that political decisions of individuals and institutions are determined by the underlying incentive structure implying that rational, self-interested decision-makers decide as to maximize their own utility. Changes in the decision-makers’ behavior are thus induced by changes in expected cost and expected benefit of a particular action, or between competing actions. The choice finally made by the decision-maker reveals his or her preferences (PRINZ 1994 p. 62). In leaving the realm of traditional rational actor models basically three subcategories of public choice models are relevant to analyze the allocation of food aid – the theory of interest groups, the theory of political voter systems, and the theory of bureaucracies. First, the theory of interest groups based on the work of OLSON (1965), which was later enhanced by BECKER (1983), postulates that the action of (organized) interest groups, as for instance the farm lobby, has a measurable impact on policy outcomes (MUELLER 2005 pp. 475). An application of this theory to foreign aid allocation is given by LAHIRI and RAIMONDOS-MØLLER (2000) approving the importance of lobbying of ethnic groups in ODA allocation. Furthermore, the theory of political voter systems first examined by BUCHANAN and TULLOCK (1962) deals with rational based constitutions as individuals and groups are able to determine the rules laid down in the constitution. This theory is appropriate to the examination of EU food aid allocation as the “rules of the game”, that is to say, the particular voting rules applied within the EU (e.g. majority or unanimity rule) determines the policy outcome to a considerable extent (PRINZ 1994 p. 62, MUELLER 2005 pp. 615). Finally, the political economy of bureaucracies based on NISKANEN (1971) offers important insights with respect to the allocation of food aid. NISKANEN argues that bureaucracies and public administrative bodies act to the benefit of its decision-makers (the so-called rational actor models). This egoistic behavior is in sharp contrast to the traditional point of view that administrative bodies act and decide in a selfless manner with regard to maximize public welfare (MUELLER 2005 p. 359). However, delving deeper into the causes of human motivation, recent work of FREY and STUTZER (2000a, 2000b) derived approaches to explain behavior which is inconsistent with rational actor models, such as intrinsic motivation and happiness. Therefore, both egoistic, purely self-interested motivation of political actors and altruistic interests, as for example the ideological orientation of EU decision-makers, may influence aid allocation (ROUND and ODEDOKUN 2003). 29 As already mentioned, the EU provides food aid entirely on a grant basis and does not receive any monetary reward for its food aid contributions. Therefore, the very fact that food aid actually is provided indicates that there has to be some non-monetary benefit for the EU. But since the amount of aid given is limited, there also has to be some cost. POLITICAL ECONOMY OF EU FOOD AID On the one side, the EU benefits by achieving specific objectives, which may be based on self- interest of the EU, such as external interests, or purely altruistic as aid per se is seen as a desirable objective and is thus approved by the voter. On the other side, food aid is a transfer of resources from the EU to the recipient. If food aid is provided in form of financial transfers the cost merely confine to the amount of cash donations. If food aid is donated in kind the EU agrees to bear additional cost for transport, procurement and distribution. These budgetary costs are strongly associated with the political cost of food aid or, following BEENSTOCK (1980) the “political rancor”. Since budgetary resources are limited food aid competes with other expenditures. Especially in times of tight budgets within the donor country, such as high unemployment or large net budget deficits, voters may resent high expenditures abroad which are also needed at home. Governments are less willing in this situation to use relatively scarce budgetary resources to donate food aid (BEENSTOCK 1980). Figure 9: Determination of the Optimal Development Aid Budget for Donor Countries 30 Source: PRINZ (1994 p. 95) following BEENSTOCK (1980 p. 140). Development aid Z2 r1 Political resistance to development aid r2 q 2 q 1 Z1 R1 R2 This relationship is illustrated in Figure 9 where Z represents the political preference function and R the restriction function with respect to the budget assigned to food aid. The concave progression of Z indicates that decision-makers are willing to accept additional political opposition only if the associated increase in food aid is disproportionally high. The political POLITICAL ECONOMY OF EU FOOD AID restriction function R gives the government’s trade-off between food aid as a desirable objective and political resentments. R is convex suggesting that, given a fixed total budget, the marginal opposition increases with a rise of development aid. The optimal allocation of food aid is found at the osculation point of the preference and the restriction function which corresponds to the total amount q1 of development aid provided in the initial situation. If the budgetary situation in the donor country tightens, resistance of opponents of development aid will increase from R1 to R2. As a result, merely the preference function Z2 can be reached, so that the amount of food aid agreed upon in budget negotiations shrinks from q1 to q2 (BEENSTOCK 1980, PRINZ 1994 pp. 95). Similar to classical economic theory, public choice theory postulates that the actual amount of food aid provided depends on both the supply and the demand side. The decision over EU food aid allocation is thus subject to a set of economic incentives within a particular institutional framework in the recipient and donor country. Following PRINZ (1994), a political market for EU food aid where the EU supplies and the recipient countries demand food aid can be postulated in order to address the question why food aid contributions strongly vary over time and across countries. Figure 10 shows such a political market for food aid where the political price of food aid on the supply side is equivalent to the marginal cost of food aid. On the demand side the political price of food aid refers to the marginal benefit of an increase in food aid provided. Figure 10: Effect of Shifts of the Supply and Demand Curve on the Food Aid Market Equilibrium A B 31 Source: Own illustration following PRINZ (1994 pp. 70, 98). Food aid donations Political price of food aid S D q q‘‘ q‘ S‘‘ S‘ p p‘‘ p‘ Food aid donations Political price of food aid S D q q‘‘ p‘ D‘‘ D‘ p‘‘ p q‘ POLITICAL ECONOMY OF EU FOOD AID As shown in Figure 10A, the supply curve of food aid has an upward slope indicating that a higher amount of food aid provided is associated with increasing marginal cost. If the supply curve shifts to the left the political cost increase from p to p’, leading to a reduction of the amount of food aid contributions from q to q’. In contrast, if the supply curve shifts to the right the political cost of food aid decline from p to p’’ and the amount of food aid provided increases from q to q’’. Thereby, different shapes of the supply curve may either reflect the change of marginal political cost of food aid to one particular recipient over time, or may refer to different marginal cost across recipient countries. As concerns food aid allocation across countries, the positive slope of the supply curve implies that a higher amount of food aid delivered to a particular recipient is associated with an increase in marginal cost. This is due to the fact that an increasing amount of food aid provided to one recipient is associated with declining deliveries to other countries. This probably results in a loss of goodwill of these countries which may negatively affect the achievement of the donor’s objectives (PRINZ 1994 p. 69). On the demand side illustrated in Figure 10B, food aid benefits the recipients due to an increase of the resource stock. However, the negative slope of the demand curve indicates that the benefit of food aid declines with increasing amounts of food aid. First, the capacity of existing infrastructure in developing countries to discharge, store and transport food aid may be limited. Second, an increasing amount of food aid leads to higher dependency on donor countries and is thus accompanied by a loss of autonomy as the capability of donors to influence the recipients’ decisions increases. If the demand curve shifts to the right the political benefit increases from p to p’’, leading to a rise of the amount of food aid receipts from q to q’’. If the demand curve shifts to the left the political benefit decreases from p to p’, leading to a decline of the amount of food aid receipts from q to q’. Similar to the supply side, different demand curves may either reflect the change of marginal political cost of accepting food aid of one donor over time or different marginal costs of accepting food aid from different donors. Differences in demand for food aid across recipients may, for instance, depend on whether the ideological view of the recipient and the donor matches (PRINZ 1994 p. 71). 4.2 Determinants of Food Aid Supply As outlined in the previous section the quantity of food aid provided is determined by both food aid supply and food aid demand. Different objectives and interests of donors and recipients shaping the curves are reviewed in the next sections, starting with the supply side. Each factor considered leads to a shift of the supply curve as illustrated in Figure 10A. If the factor increases the marginal political cost of food aid contributions, the supply curve shifts to the left, thus 32 POLITICAL ECONOMY OF EU FOOD AID reducing the amount of food aid delivered, whereas a reduction of marginal political cost leads to a shift of the supply curve to the right, associated with an increasing amount of food aid. The following sections discuss hu