Turkey - Price stabilization and food security policies in grains

This internal workirg paper is prepared for STAFF USE ONLY, The view expressed are not necessarily those of the World Bank. AGP66 TURKEY PRICE STABILIZATION AND FOOD SECURITY POLICIES IN GRAINS AGREP Division Working Paper No. 66 Prepared by: David Bigman (consutant) Economics and Policy Division Agriculture and Rural Development Department March 1983  TURKEY PRICE STABILIZATION AND FOOD SECURITY POLICIES IN GRAINS by David Bigman 1. Incroduction 1. Price policies for food and feed grains in general and wheat in particular touch tbe vital interests of most Turkish house&odds. The following statistics are indicative: roughly b6 percent of the Turkish population works in the agricultural sector, most of whom produce some grain for sale or for hboe consumption; value added in grains comprises about one-third of the agricultural GDP; wheat alone occupies about 45 percent of the total sown area and accounts for 19 percent of the wholesale price index, and bread accounts for 11-12 percent of the consumer price index. This annex evaluates the performance of these policies and examines the direct and indirect effects oo consumer3, producers, the fiscal budget and the balance of payments. The main tool of analysis is a simulation model of the wheat sector in Turkey. The structural and behavioral parameters of the model are based on real data; the policy decisions in the model are policies and practices existing in Turkey, adjusted to the specific structure of the simulation model. Alternative policies are aiso examined with the model, and their efficiency is compared with existiLig policies. 2. The primary goals of highly intensive government involvement in the wheat market are to maintain wheat prices at levels that would remunerat2 producers fcr their production costs; ernsure a more desirable wheat distribution betweea surplus and deficit areas; and prevent sharp increases in consumer prices. The Turkish government has tried to achieve these goals through the direct regulatioo of commodity prices; by providing fertilizers, capital and arher farm inputs at subsidized prices; by establishing and providing mar eting services such as transport, storage and processing; and, finally, by close regulation of imports and exports. 3. The ultimate responsiblity for those programs is vested in the Council of Ministers. The Ministries of Commerce, Agriculture and Forestry, and Finance as well as the State Planning Organization, concerned parastatals, sales cooperatives and other agencies act in an advisory capacity to the Council on various aspects of policy, particularly the annual exercize to determine the farm procurement prices. Various State Economic Enterprises (SEE) and Unions of Sales Co peratives carry out the procurement and the marketing of the c-nodities concerned. Thus, the Soil Products Office (TMO), an SEE which operates under the nominal control of the Ministry of Commerce, intervenes in grains on behalf of the government. The report does not attempt to examine the whole range of government interventions, however. Neither government subsidies for farm inputs nor the periodic but temporary interventions of TXO aimed at preventing pockets of scarcity are considered. Instead, the report focuses on price determination, procurement, storage and trade processes from a national and a long-term, rather than a short-run and regional: perspective.  4. Although the share of TMO purchases in total wheat production comprises only about 10 percent (although it has reached 20 percent in some years), or about one-third of the marketed surplus, their role in the wheat market is pivotal. The basic procurement price, whicb the government announces each June, becoces the effective support price for producers, since the T10 is required to procure any quantity at the announced price. 'he procurement price, therefore, is a major factor affecting the entire market each year. Sales of the TM0 to local millers at a reduced price are made on an allocation basis, and they are mostly if not solely designated for bread proauction. The sale price may at times be partially subsidized. 'This aftects the price of bread to consumers, especially in urban areas. In 1974/7), for instance, the ThM sale price was 40 percent ana in 1978/79 it was 20 percent below the support purchase price. Losses suffered by TM10 are covered by the'government and effectively can be regarded as subsidies to breadl. Since 1960, however, government's intentions are to set the sale price above the procurement price thereby abolishing this direct subsidy. Nonetheless, there is still an indirect element of subsidy in TU10 sales through storage and transport operations, and its access to Central Bank rediscounts at highly preferential rates. 5.- Special attention in the report is given co trade activities and trade policies of the Turkish government. Wheat exports became fairly significant in the mid to late 1970s, as whear production was deliberately encouraged following the trauma of the world-'iie wheat shortage in 1972-74. In 1978 and 1979, for instance, Turkish wheat exports were around 2 million MT or about 15 percent of total wheat production. These subsequently were curtailed with the onset of the recent economic crisis, and the country became a net importer, tlough marginally so, in 1981 and 1982. Export and iiaport operations have been carried out with the primary intention of ensuring adequate stocks so that government agencies will be able to perform the function of regulating and stabilizing the domestic market. Although private exports of wbeat has recently been authorized, THO will remain involved in wbeat trade and the government will secure its goals through tariffs, subsidies or quota restrictions. b. In general, a country's economic policies relating to the agricultural sector are typically determined by broad guidelines and overall strategies, but their actual execution on a year-to-year or even day-to-day basis is largely determined by the specific conditions existing at that period of time. 'The general strategies accord with long-range goals; the specific day-to-day tactics respond to any short-term change in the production and marketing environment. 'Thus, for example, a country may adopt the strategy of promoting the free market or allowing comparative advantage considerations to determine crop selection. Nevertheless, a year of poor domestic harvest or unfavorab1'. world markets may still force radical government intervention if the far-.;.s' subsistance or Lhe availability of food to low-income consumers is ena5,,c ' red.  7. The simulation model used in this paper explicitly recognizes the nature of the economic environment within which the Turkish government operates. The model accounts for changes in government policies in response to environmental changes. It thus provides the means for evaluating the effects ot these policies over an extended period of time and fur examining the degree to whicb these short-run tactics indeed complement the long-run strategies and tally with the general goals of ton; economy. 8. 'T'he paper contains two types of analyses. The first evaluates the specific price policies and other interventions presently practiced in Tu'rkey. 'lhis analysis is based on actual 1981 production and price figures as well as on _he estimated growth rates for wheat production, per capita wheat consumption and the population. 'ibe second is a more general comparative analysis of alternative policy regimes. To avoid reference to any specific year and permit _ core general review, ihe latter analysis is based on index numbers instead of actual figures and the pe:formance of the various policies is evaluated in rei3tion to the outcomes under free trade. 9. Aware of cne diversity of governme.it goals with respect to food and other staple foods, the report refrains from resorting to a single performance measure (e.g. tota' economic gains). Instead it presents the effects of the various policies over a wide range of performance criteria. II. *Model, Data and Parameters 10. 'the analysis is based on a simulation model of the agricultural sector.l/ The model accounts for short-run stochastic variations in agricultural production due to varying climatic conditions as well as with long-run growth trends in production and consumption. ilonte Carlo simulations yield random "weatber" events in the home country (Turkey) and abroad. For each ranaom event the state variables of the model are calculated according to the specified de-ision rules and the state variables of the previous years. The process is repeated a large number of times and the statistics collected yield the expected values and the frequency distribution of the performance measures. 11. As mentioned ir the introduction, the analysis is divided into two parts: one evaluating the specific price policies and orher intervention practices in Turkey, and the other a more general analysis of alternative policies and policy combinations and under different scenarios. The former, Turkey-specific aaalysis use actual 1981 data of wheat consumption and production in Turkey ano financial magnitudes are given in Turkish lira (IL) in terms of 1981 prices. The latter, more general analysis expres.es the results in terms of index numbers rather than actual figures in order to avoid reference to any specific year. Accordingly, the iniLial levels of both output and price bave been normalized to be 100. Financial magnitudes 1/ See Bigman and Reutlinger (1979a, 1979b); Bigman (198U, 1981, !'82a, 198Zb, 1982c) for further references and a more detailed aescription of che model.  (economic gains, taxes, rev-enues) are then axpressed as percentages of total domestic expenditures on wheat in a normal year and quantities (output, imports, exports) are expressed as percentages ot the mean output. The (real) rate of interest in all the financial calculations is assumed to be 5 oer_ent. The values of the behavioral parameters in the model are generally ',ased on estimates obtained in previous studies or the Turkish wheat sector (Im:ohoroglu and Kasnakoglu (1979) and (198U); Gurkan (1979); Somel (1979)), estima,es prepared for this report 1/ and on TMO estimates. 12. The simulation experiments themselves were ccnducted over a time horizon of 3U years, each experiment consisting of 300 iterations of a 3U-year sequence of ranco. production events. The total sample therefore consists of 9,OUu oDservatioois arawn at r, :viom from a normal probability distribution. 13. Following is a brief description of :-ne structure of the model and the s?ecific parameters assumed in the analysis. Consumption 14. Consumers in the country are divided into 3 income groups (low-income urban consumers, 3idcle and high income urban consumers, and rural consumers). A separate demand function is specified for each group and total demand is the sum of the demands of the 3 groups. This division allows an examirnation of policies that are targeted on low-income urban consumers only. 15. Table I lists the specific demand parameters assumed in the simulation experiments. Table 1: Wheat Demand Parameters Middle- and Low-Income High-Income Urban Urban Rural Total Consumers Consumers Consumers Population Mean quantity consuimd as percentage of total consumption 20 30 50 100 Price elasticity of demand at the mean price Upper segment - p p .40 .10 .30 .26 Lower Segment - P< P .55 .20 .30 .32 Where P is the mean price. As noted above, the mean price is normalized to be 100 in the more -eneral policy evaluation section. In the dynamic analysis, the population is assumed to gDow at an annual rate of 2.1 percent. 1/ These estimates have been prepared by Haluk Kasnakoglu, Tercan Baysan and Aslan Gurkan of the Department of Economics, the Middle East Technical University, Ankara, and World Bank staff.  -5- Production 16. Production in the country is stochastic. Producers are assumed to determine the area cultivated on the basis of the expecti!d price. They are also assumed to, be risk-averse, and adversely respond to the risk associated with price variability. rhe specific supply function ass;umed in the analysis has the form 1/. A -St = n.