Tunisia - From universal food subsidies to a self-targeted program

Report No. 15878-TUN Republic of Funisia From Universal Food Subsidies to a Self-Targeted Prog,ram June 1996 Document of the World Bank  Currency and Equivalent Units Currency Unit = Tunisian Dinar (TD) ITD = 1,000 millimes US$1.00 = TDO.9697 (estimated average 1995) TD1.000 = US$1.0312 Fiscal Year January 1 - December 31 Glossary of Abbreviations AES Subsidy already eliminated (under reform program) AIDS Almost Ideal Demand System ASAL II Second Agricultural Structural Adjustment Loan BW Bread Wheat BMI Body Mass Index CCGC Central Field Crops Cooperative (Cooperative Centrale des Grandes Cultures) CGC General Compensation Fund (Caisse Generale de Compensation) COCEBLE Central Cooperative for Wheat (Coop&rative Centrale de B1e) CST Tunisian Sugar Complex (Complexe Sucrier de Tunisie) DHS Demographic and Health Surveys Project DW Durum wheat EFRSL Economic and Financial Reform Support Loan ES Subsidy eliminated under reform scenario FAO Food and Agriculture Organization INS National Statistics Institute (Institut Nationale de la Statistique) LAINO A new private dairy processor for sterilized local fresh milk. LF Local Fresh Milk NF Needy Families Direct Assistance Program OC National Cereal Authority (Office des Cerdales) OCT Tunisian Trade Organization (Office du Commerce de Tunisie) ONH National Oil Board (Office National de l'Huile) PAE Per Adult Equivalent PS Specific Weight of flour. (Poids Specifique) SAL Structural Adjustment Loan SRM Sterilized-Reconstituted Milk STIL Tunisian Dairy Products Company (largest) (Societe Tunisienne des Industries Laitieres) STS Tunisian Sugar Company (Societe Tunisienne de Sucre) UHT Ultra High Temperature UTSS Tunisian Union of Social Solidarity (Union Tunisienne de Solidarite Sociale)  REPUBLIC OF TUNISIA FROM UNIVERSAL FOOD SUBSIDIES TO A SELF-TARGETED PROGRAM EXECUTIVE SUMIMARY Overview i. Through the Caisse Generale de Compensation (CGC), the Tunisian Government has subsidized the consumption of basic food-stuffs and a variety of other items since 1970. Subsidies have been paid on an unrestricted basis, meaning that the transfers have been available to anyone who chooses to purchase subsidized commodities in any quantity desired. ii. While CGC subsidies had always made a substantial contribution to the welfare and nutritional intake of the poor, by the 1980s it became apparent that the universal subsidy program had become too costly. The Tunisian Government was then faced with a common policy dilemma in reforming its subsidy program: how to reduce budgetary costs, in a politically acceptable way, while protecting low income groups. Rather than switching to an entirely different method of transferring income to the poor, the Government sought to fine-tune the existing framework of price subsidies by shifting subsidies to food products which are primarily consumed by lower-income households. With this approach, a type of self-targeting, subsidized products are still available to all, but they are selected specifically to discourage the rich from consuming them. iii. The transition from universal subsidies towards a more self-targeted program has been a continuing process rather than a discrete policy change. The rationale for this study is to evaluate the impact of CGC food subsidies, particularly on the poor, and to undertake a mid-term assessment of the on-going reform program. The analysis provides a basis for policy-makers in Tunisia to sharpen targeting mechanisms and offers lessons to other countries regarding the practical issues involved in implementing self-targeting reforms. Impact of the Universal Subsidy Program iv. Since its inception, the central objectives of the CGC subsidy program have involved redistributing income toward the poor and protecting the purchasing power and nutritional status of low- income groups. To some extent, the universal subsidy program' was successful in meeting these goals. It was progressive in relative terms, contributing over five times more to the purchasing power of the I/The "universal subsidy" program covers the period until 1990, after which the Government launched a series of reforms designed explicitly to improve the targeting of CGC subsidies. Under the universal program, subsidies were placed on a broad range of products, and the subsidized commodities were generally the only variety of the product available on the market. During this period (1970-90), the program experienced virtually universal "take-up" across income groups; hence the term "universal." The program after 1990, however, is referred to as the "reform" program or the "self-targeted" program due to the efforts of policy-makers to introduce quality differentiation in order to improve the targeting of food subsidies.  ii Executive Summary poor than to that of the rich (as a share of total expenditures).2 This is not surprising since most CGC subsidies have been placed on food products which, by Engel's Law, generally constitute a larger share of total spending by lower-income consumers than by the more well-to-do. In 1990, expenditures on subsidized products accounted for close to nine percent of the total expenditures of the lowest-income quintile, as compared to less than two percent for the wealthiest quintile.3 Moreover, roughly 60 percent of total caloric intake and 73 percent of protein consumption of the poor were derived from subsidized foods.4 v. Despite these benefits, by the mid-1980s, it was clear that the universal subsidy program was inefficient and costly. It was inefficient because it subsidized a broad range of products available to all Tunisians regardless of need. In fact, the wealthiest-income group benefitted twice as much as the poorest-income group from the program in absolute terms.' The program was costly because it claimed a large share of Government resources: by 1984, outlays on subsidies hovered around four percent of GDP and ten percent of total Government expenditures. Alternative Courses of Reform vi. The high and rising costs of the program, combined with inefficiencies and substantial leakages to the non-poor, nade an overhaul of the universal subsidy system an urgent priority. In the early 1980s, Tunisian policy-makers began exploring ways in which to reform the program. An initial attempt to reduce the budgetary costs of the CGC was made at that time, and subsidies on several food items were eliminated, effectively doubling their prices. However, violent riots, which erupted in response to these efforts, forced officials to rescind the measures and delayed the adoption of significant reforms until the end of the decade. vii. Because a complete elimination of the program no longer seemed like a feasible option from a political standpoint, the Government examined alternatives to better target the transfers to the poor. Targeting was considered an attractive option because it reduces leakages of program benefits to the non-poor and concentrates public expenditures on those who need them the most. viii. Traditional targeted programs, such as direct-assistance schemes, however, generally require specific selection of the target group. As a result, they often entail administrative costs, experience implementation difficulties and suffer from a lack of political support. Administrative costs are associated with, inter alia, identifying beneficiaries, screening applicants, maintaining eligibility lists, ann 198S, food subsidies contributed 5.5 times more to the total expenditures of the lowest-income group than the highest; in 1990, they accounted for 5.4 times more of the purchasing power of the lowest-income quintile than the highest; by 1993, they contributed to 7.2 times more. I/Population is ranked by total expenditure per adult equivalent which is used as a proxy for income throughout the analysis. I/INS Household Expenditure Survey, District of Tunis component (1990). Intake derived from purchased subsidized products only (excludes on-farm consumption). I/On a per capita basis. INS Household Expenditure Survey (1985).  Executive Summary iii and preventing fraud. Implementation difficulties often arise in selecting the beneficiaries and keeping up-to-date eligibility lists due to a lack of information regarding individual or household incomes. Without such information, these programs tend to be plagued by leakages to the non-poor and exclusion of eligible beneficiaries. Political support for well-targeted programs can be weak due to the narrow range of beneficiaries which usually excludes powerful and vocal groups within the population. Attempts to implement other targeted direct-assistance programs in Tunisia have not been able to avoid these problems.6 As a result, the Tunisian Government did not consider traditional direct-assistance schemes (such as food stamps) as feasible alternatives to the CGC food subsidy program. ix. Other targeting mechanisms were also explored, but did not prove to be appropriate candidates for reforming the food subsidy program. For example, geographic targeting was rejected because, in most areas, distinctions between neighborhoods were too obscure to make it effective. The Reform Program: From Universal Subsidies to Self-Targeting x. Given political, economic, and administrative constraints, the foundations provided by the universal food subsidy system favored reforms to reduce costs and improve the distribution of subsidies across income groups within the existing framework of food price subsidies over those requiring an entirely new institutional structure to channel transfers to intended beneficiaries. In this context, a reform program was developed and incorporated into the VIII Development Plan (1991-96). Its explicit goals aimed to reduce CGC expenditures with the least possible impact on the lower-income households. The primary components of the reform program included: (i) improving the targeting of CGC intervention toward the poor; (ii) gradually adjusting prices to progressively reduce and eliminate subsidies on certain products; and (iii) reducing unnecessary production and distribution costs for subsidized products. xi. A particularly innovative aspect of the Tunisian reform program has been its reliance on self-selection mechanisms to improve the distributional incidence of subsidies. Self-targeting occurs when benefits are available to all, but the program is specifically designed so that the non-poor elect not to participate. While other targeted programs require social workers or other agents to undertake some kind of screening (such as individual or group assessment) to determine eligibility, with self-targeting the decision to participate is made by individuals themselves. xii. The principal device used to promote self-targeting in Tunisia is quality differentiation. Designing a self-targeted food subsidy program using quality grading involves examining household expenditure data to determine whether there are significant differences in consumption across income groups. If the poor consume a different basket of goods from wealthier consumers, this basket can be selected for subsidization. Using existing survey data to identify goods that should benefit from subsidies allows self-targeted programs to economize on information costs by avoiding the cumbersome task of assessing individuals' income levels to determine eligibility. In practice, however, consumption patterns may not differ significantly across income groups. This does not mean that self-targeting is not feasible, 6/Direct-assistance schemes include the Needy Families Program and the Union Tunisienne de Solidarite Sociale, which is responsible for low-income food ration programs and cash transfers to the elderly and handicapped. These programs are quite small in comparison to the CGC program and would not compensate the poor for an elimination of food subsidies. The World Bank (April 1993).  iv Executive Summary but rather that it may require some creativity to invent "inferior" subsidized goods which are unattractive to wealthier consumers and unsubsidized high-quality alternatives ("superior" goods) to siphon off the demand by the rich. xiii. As a first step in improving the targeting accuracy of the Tunisian program, subsidies were eliminated on goods clearly consumed disproportionately by the rich (such as meat). Remaining subsidies were focused on basic staples, and currently all of the subsidized products, with the possible exception of certain types of milk, are consumed in substantial quantities by the poor. xiv. The reform program in Tunisia extended self-targeting efforts still further by developing a range of goods that represent distinct "quality" grades to consumers of different incomes. Two new approaches to self-targeting are currently being implemented (a summary of self-targeting measures is presented in Box A below). xv. The first tactic for extending self-targeting, a variation of the "inferior goods" approach, involves shifting subsidies to narrowly-defined items within a product line that are perceived by consumers to be of a lower quality because they possess certain unattractive features in their packaging or ingredients. Although the intrinsic value of these products is preserved, these perceived "inferior" characteristics discourage consumption by wealthier households.7 Examples are discussed in paras. xvii- xxi below. xvi. The second method, the "superior goods" approach, is appropriate when a government has traditionally subsidized a particular commodity and no other varieties or qualities are available on the market at cost (usually because of official marketing regulations or restrictions). In Tunisia, the quality of subsidized products has generally been reasonable, but not exceptional. Since markets have otherwise been tightly controlled by state marketing boards, the subsidized product has often been the only variety available on the market. To reduce subsidy costs while maintaining benefits to the poor, authorities have liberalized the sale of higher-quality versions of these products, which are sold at cost and attract the demand from wealthier households who then consume less of the subsidized products. This "superior goods" approach is consistent with the pursuit of a general liberalization strategy designed to increase efficiency in the subsector, but has the added benefit, in this context, that it promotes self-targeting. Several examples are discussed below. xvii. Data from the 1990 household expenditure survey clearly indicated which cereals products justified continued subsidies under the CGC reform program and which were candidates for removal from the program. In 1990, baguettes were consumed virtually exclusively by the rich; wealthier households also consumed disproportionately more pasta, couscous and pastry flour. Although the lowest-income quintile consumed slightly more gros paing than the richest, expenditure patterns for gros pain did not differ significantly across quintiles, and middle-income households spent the most on this type of bread. I/It is important to note that it is perceived inferior characteristics, and not any real downgrading of nutritional value, that make a particular item suitable for self-targeting. In fact, some "inferior" products are actually nutritionally richer than their "superior' counterparts, although they possess certain features (such as color, texture, taste or packaging) that make them unattractive to wealthier consumers. 8/Gros pain is a larger loaf bread made from the same type of traditional baking flour as baguettes.  Executive Summary v Semolina was the only existing cereals product that was clearly consumed more by the poor than the rich in both absolute and relative terms. The CGC reform program responded to these expenditure patterns by eliminating direct subsidies on baguettes, pasta, couscous and pastry flour.9 Household survey data for 1993 indicate that, despite these measures, the rich did not substitute consumption in favor of semolina or gros pain, which remain subsidized under the CGC program, and the reforms were indeed successful in improving the targeting of cereals subsidies. In addition, the Government recently applied the "superior goods" approach by liberalizing the sale of unsubsidized high-quality varieties of bread that had not previously been allowed on the market but which appeal to high-income consumers.'