Yugoslavia - Adjustment policies and development perspectives (Vol. 3 of 3) : Methodological and statistical annex

Report No.3954-YU Yugoslavia: Adjustment Policies and Development Perspectives F (In Three Volumes) Volume Ill: Methodological anrd Statistical Annex November 2, 1982 Country Programs Department I Europe, Middle East and North Africa Region FOR OFFICIAL USE ONLY Document of the World Bank This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  Currency,Equivalents 1/ 1976 1 US Dollar = 18.19 Dinars 1 Dinar 0.055 US Dollars 1977 1 US Dollar = 18.30 Dinars 1 Dinar 0.055 US Dollar 1978 1 US Dollar = 18.64 Dinars 1 Dinar 0.054 US Dollar 1979 1 US Dollar = 19.00 Dinars 1 Dinar 0.053 US Dollar 1980 1 US Dollar = 24.91 Dinars 1 Dinar 0.040 US Dollar 1981 1 US Dollar = 35.51 Dinars 1 Dinar 0.028 US Dollar 1/ Period average exchange rates. The dinar has not been maintained within announced margins since July 12, 1973. Following a devaluation on October 22, 1982, the rate stood at 63.51 Dinars per US Dollar. Glossary of Abbreviations BOAL Basic Organization of Associated Labor COAL Complex (or Composite) Organization of Associated Labor CGE Computable General Equilibrium CIFER Community of Interest for Foreign Economic Relations CoI Community of Interest DOD Debt Outstanding and Disbursed DRC Domestic Resource Cost GMP Gross Material Product ICOR Incremental Capital-Output Ratio KBP Kosovska Banka Pristina LDR Less Developed Region(s) MDR More Developed Region(s) MLT Medium- and Long-Term OAL Organization of Associated Labor QR Quantitative Restriction SSE Small Scale Enterprise(s) YBIEC Yugoslav Bank for International Economic Cooperation Throughout this report, the term 'region' is used to refer to Yugoslavia.'s six republics and two autonomous provinces.  FOR OFFICIAL USE ONLY YUGOSLAVIA ADJUSTMENT POLICIES AND DEVELOPMENT PERSPECTIVES This report is based on the findings of a World Bank economic mission which visited Yugoslavia in June 1981. The mission consisted of the following: Suman Bery, Chief of mission Petros Aklilu (Agriculture) Deepak Bhattasali (Foreign Trade and Finance) Kosara Gavrilovic (Translator and Interpreter) Paul Harrison (Consultant) (Agriculture) Frederick Kilby (Employment and Regional Development) Francois Laporte (Industry) Surinder Malik (Industry) Leslie Manison (IMF) (Foreign Trade and Finance) Mieko Nishimizu (Total Factor Productivity; Regional Development) Sherman Robinson (Adjustment Policies; CGE Model) Laura Tyson (Consultant) (Adjustment Policies; CGE Model) Assistance to the mission's work in Washington was provided by Jeffrey Lewis and Mukaila Ojelade. Ann Pepper had primary responsibility for document production. A draft of the report was discused with official and academic bodies in Yugoslavia in June and July 1982, and additional material has been included based upon that visit. This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  YUGOSLAVIA ADJUSTMENT PERFORMANCE AND DEVELOPMENT PERSPECTIVES Table of Contents VOLUME III: METHODOLOGICAL AND STATISTICAL ANNEX Appendix I A Summary Description of the Computable General Equilibrium (CGE) Model for Yugoslavia .... ............ 1 A. Introduction ......... .............................. 1 B. The CGE Static Model Formulation .... ............... 1 Product Markets and Factor Supplies .... ........ 2 Foreign Trade and Foreign Exchange .... ......... 8 Income and Product Demand ..... ................. 15 Market Clearing Conditions .................. 17 C. The Dynamic CGE Model .............................. 17 D. Trade Policy and Resource Allocation: Some Empirical Results ............................ 17 Measurement and Interpretation of Domestic Resource Cost ..... ................... 18 The Interpretation and Measurement of Market Clearing and Shadow Exchange Rates ......................................... 20 The Contribution of Various Factors to the Depreciation of the Equilibrium Exchange Rate Over the 1976-80 Period ..... ............. 21 Sectoral Market and Social Profitability Rates: Measurement and Interpretation ........ 22 Appendix II Mathematical Presentation of the CGE Model .... .......... 28 A. Introduction ......... .............................. 28 B. Equations of the Flexible Exchange Rate Model ....... 28 The Operation of Markets ..... .................. 33 C. Fixed Exchange Rates and Import Rationing .... ...... 34  - 1i - Table of Contents kcontinued) Appendix III Measurement and Decomposition of Growth in Total Factor Productivity .................................... 35 Appendix IV A Note on Data Classification ........................... 44 STANDARD TABLES ......................................... 45 Statistical Appendix 49 Table No. Title Population, Employment A 1.1 Population, 1970-81 A 1.2 Employment and Labor Force, 1970-81 A 1.3 Social Sector Employment, 1970-80 A 1.4 Social Sector Employment by Region, 1975-80 A 1.5 Structure of Registered Job Seekers, 1970-80 A 1.6 Monthly Earnings Per Worker in the Social Sector, 1970-81 A 1.7 Personal Incomes in the Private Agricultural Sector, 1970-80 A 1.8 Annual Personal Incomes Per Worker in the Social Sector and Incomes Per Active Worker in Private Agriculture, 1970-80 National Accounts A 2.1 GDP by Expenditure at Current Market Prices, 1971-80 A 2.2 GDP by Expenaiture at Constant 1972 Prices, 1971-80 A 2.3 GDP by Industrial Origin at Current Factor Cost, 1971-80 A 2.4 GDP by Industrial Origin at Constant 1972 Factor Cost, 1971-80 A 2.5 Gross Material Product by Industrial Origin at Current Market Prices, 1970-80 A 2.6 Gross Material Product by Industrial Origin at Constant 1972 Prices, 1970-80 A 2.7 Growth in Gross Material Product by Industrial Origin, 1976-80 A 2.8 Distribution of Gross Fixed Investment by Activity and Sector, 1970-80 A 2.9 Distribution of Realized Investment in Fixed Assets in the Social Sector by Activity, 1971-80  - iii - Table of Contents (continued) Table No. Title Balance of Payments A 3.1 Balance of Payments, 1970-81 A 3.2 Services and Transfers, 1970-81 A 3.3 Balance of Payments with Convertible Area, 1970-80 A 3.4 Merchandise Imports and Exports by Area, 1970-81 A 3.5 Composition of Exports by End-Use, 1970-81 A 3.6 Composition of Imports by End-Use, 1970-81 A 3.7 Exports According to SITC Classification, 1970-81 A 3.8 Imports According to SITC Classification, 1970-81 A 3.9 Export and Import Price Indexes According to SITC Classification, 1970-80 A 3.10 Imports by Import Regime, 1975-80 A 3.11 Average Tariffs by Customs Positions, 1975-80 External Debt A 4.1 External Debt, 1971-80 A 4.2 Structure of External Debt Disbursed and Outstanding, 1971-81 A 4.3 Publicized Eurocurrency Borrowings, 1972-80 A 4.4 Export Credits Extended by Yugoslavia, 1971-80 Public Finance A 5.1 Consolidated Revenues and Expenditures of the Public Sector, 1975-80 A 5.2 Revenues of the Public Sector, 1975-80 A 5.3 Budget of the Federation, 1975-82 Monetary Statistics A 6.1 Money Supply by Sector, 1971-80 A 6.2 Domestic Credit by Sector, 1971-80 A 6.3 Selected Interest Rates, 1977-81 Agricultural Statistics A 7.1 Agricultural Production and Growth Rates by Crop A 7.2 Agricultural Growth Rates by Republic and Province, 1965-85 A 7.3 Agricultural Land Use, 1959-85 A 7.4 Ownership of Cultivable Land, 1959-85 A 7.5 Arable Land Use A 7.6 Actual and Planned Crop Area, 1976-80  - iv - Table of Contents (continued) Table No. Title A 7.7 Yields for Selected Agricultural Products, 1969-85 A 7.8 Actual and Planned Agricultural Production by Sector, 1976-85 A 7.9 Livestock Population, 1976-85 A 7.10 Livestock Productivity A 7.11 Agricultural Machinery, 1959-79 A 7.12 Chemical Fertilizer Consumption - Active Substance A 7.13 Agricultural Commodity Prices, 1979-81 A 7.14 Purchases from Agricultural Producers, 1976-79 A 7.15 Average Producer Prices Paid in Individual and Social Sectors, 1977-80 A 7.16 Price Comparisons - Major Yugoslav Crops, 1981 A 7.17 Exports and Imports of Food and Agricultural Commodities, 1978-79 A 7.18 Agroindustry Capacities, 1981-85 A 7.19 Net Income of Rural Households A 7.20 Per Capita Consumption of Major Food Products, 1969-78 A 7.21 British Standard Labor Requirements for Private Agriculture in Yugoslavia, 1980 A 7.22 British Standard Labor Requirements for Total Yugoslav Livestock, 1980 Industrial Statistics A 8.1 Industrial Production, 1970-80 A 8.2 Growth in Industrial Production, 1952-80 A 8.3 Structure and Growth of Industrial Production, 1952-80 A 8.4 Production in the Industrial Sector, 1976-85 A 8.5 Industrial Employment, 1965-80 A 8.6 Investment in Fixed Assets in the Industrial Sector, 1971-80 A 8.7 Distribution of Investment in Fixed Assets Within the Industrial Sector, 1971-80 A 8.8 -Structure of Trade in the Industrial Sector, 1975-81 A 8.9 Export and Import Price Indices, Industrial Sector, 1975-80 A 8.10 Structure of Industrial Exports and Contribution to Export Growth, 1975-80 A 8.11 Industrial Exports, Selected Subsectors, 1976-80 A 8.12 Structure of Industrial Imports and Contribution to Import Growth, 1975-80 A 8.13 Industrial Imports, Selected Subsectors, 1976-80 A 8.14 1981-85 Plan Projections for Merchandise Exports A 8.15 Plan Projections for Merchandise Imports  v Table of Contents (continued) Table No. Title Regional Statistics A 9.1 Composition of Population, 1948-80 A 9.2 Indices of Income Differentials of Social Sector, 1980 A 9.3 Regional Distribution of Transfers in the 1976-80 and 1981-85 Plans A 9.4 Planned LDR Growth Targets, 1981-85 A 9.5 Selected Regional Indicators Price Statistics A 10.1 Yearly Changes in Price Indices, 1971-80 A 10.2 Price Trends in Yugoslav Industry, 1970-1979  APPENDIX I: A SUMMARY DESCRIPTION OF THE COMPUTABLE GENERAL EQUILIBRIUM (CGE) MODEL FOR YUGOSLAVIA A. Introduction I.1 The model of Yugoslavia used in this report is in the family of multisector computable general equilibrium (CGE) models that have been developed during the past few years as instruments of planning and policy analysis for developing economies. 1/ The CGE model for Yugoslavia, like CGE models designed and estimated for other developing economies, traces the evolution of sectoral outputs and relative prices over time in response to changing conditions of supply and demand on individual markets. The model embodies the view that economic performance is the outcome of decentralized decisions made by producers and consumers in response to market indicators. Such indicators, in turn, are influenced by government policy choices, international economic conditions, and other variables that are exogenous from the point of view of producers and consumers who respond to them. It is important to emphasize at the outset that this view of the economy does not imply that markets are perfect or that the decision making of producers and consumers is necessarily guided by neoclassical rules of profit maximization or utility maximization. Instead, the specification of the model explicitly recognizes the existence of rigidities and imperfections in the economy and attempts to model them formally. I.2 The emphasis on markets and market-clearing mechanisms in the CGE model reflects the view that the independent behavior of decentralized producers and consumers plays a significant role in influencing economic performance. In this way, the model diverges from traditional planning models, such as input-output or linear programming models, in which either fixed technological coefficients and/or central planning of economic activity constrain or eliminate decentralized choices. Decentralized decision making in response to market indicators is fundamental to the institutions of the Yugoslav economic system, and it is difficult to analyze alternative development strategies without a framework that captures the main institutions and behavioral rules of the system. The CGE model provides such a framework. B. The CGE Static Model Formulation I.3 The overall CGE model for Yugoslavia, like the CGE models for other economies, is divided into a static, annual general equilibrium model and a separate between-period model which provides the necessary intertemporal linkages. The static model is described in some detail in 1/ See K. Dervis, J. de Melo and S. Robinson, General Equilibrium Models for Development Policy, (Cambridge; Cambridge University Press, 1982).  - 2 - this section, after which the structure of the intertemporal model is presented. A mathematical presentation of the equations of the static model is provided in Appendix A.II. Product Markets and Factor Supplies I.4 There are 18 sectors or product markets in the Yugoslav CGE model. These sectors were chosen to reflect the priorities of the 1976-80 Social Plan. They are based on the Yugoslav sectoral classification scheme embodied in the 1976 input-output table, augmented to include the nonproductive sector. The sectoral classification scheme of the CGE model and the correspondence to the 1976 input-output table are presented in Table 1. Table 1: SECTORAL CLASSIFICATION SCHEME USED IN THE CGE MODEL FOR YUGOSLAVIA CGE Classification , Input-Optput Table (Sector No.) Agriculture Agriculture, fishery and irrigation (35) Forestry (36) Electrical energy Electricity (1) Coal Extraction of coal (2) Processing of coal (3) Oil and gas Extraction of oil and gas (4) Oil refineries (5) Ferrous metals Iron ore mining (6) Iron and steel basic industries (7) Nonferrous metals Nonferrous ore mining (8) Nonferrous basic industries (9) Smelting, alloying and refining (10) Nonmetallic minerals Extraction of nonmetallic minerals (11) and construction materials Nonmetallic mineral products (12) Building materials (21) Scrap and waste (46) Machines and fabricated Fabricated metals (13) metals Machinery (14) Transportation equipment Transport equipment (15) and shipbuilding Shipbuilding (16) Electrical equipment Electrical apparatus, appliances and supplies (17)  -3- Table 1 (continued) CGE Classification Input-Output Table (Sector No.). Chemicals and paper Basic chemicals (18) Processing of chemicals (19) Paper and paper products (24) Textiles, leather, rubber, Sawmill and wood boards (22) wood, printing and Furniture and fixtures (23) miscellaneous Textile yarns and fabrics (25) manufacturing Finished textiles (26) Leather and fur (27) Leather footwear and fancy goods (28) Rubber (29) Printing and publishing (33) Other manufacturing industries (34) Food processing Food processing and animal feed (30) Beverages (31) Tobacco (32) Construction Housebuilding (37) Civil and hydraulic engineering (38) Installations and finishing work (39) Infrastructure Transport and communications (40) Public utilities (44) Trade Wholesale and retail trade (41) Crafts and other productive Catering and tourism (42) services Arts and crafts (43) Nonproductive services Not included in input-output table. I.5 The CGE model assumes that each sector produces a single good, ana for each such good the model formalizes conditions of supply and demand. Since only 2 of the 18 sectors included in the Yugoslav CGE model-- electricity and nonproductive services--are treated as producing nontradeable goods, it is useful to begin the discussion with the model's specification of behavior for tradeable goods. 1/ The model assumes that 1/ According to the figures in the 1976 input-output table, electricity is a tradeable good since both import and export transactions are recorded. These transactions are very small, however, and they do not show up in the trade statistics reported in the Statistical Yearbook of Yugoslavia and other publications. Because of their small size and the data problems associated with them, the trade transactions of the electricity sector in the model were suppressed and electricity was treated as a nontradeable good.  -4- domestically produced tradeable goods and imports are imperfect substitutes. For each tradeable sector, the model defines a "composite" good which is an aggregation of the domestic good and the imported substitute. The trade aggregation function for each good is given by a constant elasticity of substitution (CES) function, in which an important parameter is the elasticity of substitution between the domestic good and the imported substitute. Elasticities of substitution vary across sectors, with smaller elasticities reflecting greater differences between the domestic good and the imported substitute and hence greater difficulties in substituting one for the other. In the Yugoslav model, the elasticity of substitution between imports and domestic goods was assumed to be quite low in all sectors, except for agriculture, food processing, consumer goods, infrastructure (transportation) and other productive services (tourism and travel abroad). 1/ The lowest elasticities were applied to raw material sectors because imports in such sectors are largely purchases of items for which there are limited domestic substitutes, at least in the short to medium run. I.6 Assuming that users of a particular good seek to minimize the cost of acquiring it, the "composite good" formulation for tradeable goods yields the result that the derived ratio of the imported substitute to the domestically produced good is a function of the ratio of their prices. Following the small-country assumption, the world price of the imported substitute is assumed fixed by world market conditions over which Yugoslavia has no influence, and the supply of imports to Yugoslavia is infinitely elastic at that price. The domestic price of the imported substitute is determined by the world price, the tariff rate, the premium rate, and the exchange rate. 1.7 In the current version of the Yugoslav CGE model, it is assumed that the small-country assumption applies on the export side as well, so that the prices at which Yugoslavia can sell its exports are determined by world prices. Consistent with this specification, the model implicitly assumes that exports are largely a function of domestic supply conditions rather than of foreign demand conditions. In the current version exports are set at exogenous target levels. 2/ The model allows for divergences between domestic and export prices and hence for implicit variations in incentives to export on the supply side. I.8 Finally, in the model, the absolute price level is set exogenously and is given by a weighted index of sectoral prices. Because the CGE model focuses on the determination of relative prices among sectors and real variables in the economy, the inflation rate of this aggregate price index is given to the model exogenously. The model does not solve 1/ See data tables at the end of this appendix. 2/ Export supply in each sector can be also specified as a function of the ratio of the export price to the domestic prices.  - 5 - for the inflation rate but rather for relative prices and sectoral outputs given an exogenously specified time path of the aggregate price index. For historical runs over the 1976-80 period, the actual rate of inflation of the gross domestic product deflator was used as the measure of domestic inflation. For forward runs over the 1981-85 period, the rate of inflation was exogenously projected to fall from 37% in 1981 to 13% by 1985. I.9 Domestic prices influence supply and demand decisions in all sectors. On the supply side, it is necessary to specify for each sector of production the technological conditions and behavioral rules that influence supply conditions in response to changing prices. Technological conditions are given by a production function for each sector. The production functions chosen allow for substitution between labor and capital inputs. 1/ Value-added production functions for all sectors, except the nonproductive service sector, were estimated using data series on real value-added, the real value of the capital stock and the number employed in each sector over the 1965-78 period. The production function for the nonproductive service sector was estimated using the share of labor payments in total value-added in that sector. Finally, the technology of intermediate input use is given by either; (1) fixed coefficients (from the 1976 input-output table) or (2) a neoclassical function of aggregate value added (labor and capital) and aggregate intermediate inputs. 2/ 1.10 There are several additional things to note about the production function formulation used in the model. Sectoral capital stocks are fixed within periods, reflecting the assumption that capital cannot be moved across sectors once installed. Sectoral capital stocks are updated period to period by subtracting depreciation and adding net investment in the dynamic part of the CGE model. In the model, capacity utilization is a function of the overall degree of foreign exchange and import rationing in the economy and of the dependence of individual sectors on imported intermediate inputs (see section D for a more detailed discussion of the mechanism). Finally, in the sectors in which there are assumed to be substitution possibilities between aggregate intermediate inputs and aggregate value added, there are also assumed to be adjustment costs that impose a lage in adjustment to changes in relative input prices. I.11 In two sectors in the model -- agriculture and nonproductive services -- output supply is simply determined by available inputs and is 1/ A CES production function specification was used for the electricity sector and Cobb-Douglas functions for the others. See Appendix II for a detailed discussion of alternative specifications of production technology that can be used. 2/ A Cobb-Douglas formulation is used in the following sectors: ferrous metals, nonferrous metals, chemicals and paper, and food processing.  - 6 - given by the production function. In historical runs of the model, levels of inputs and outputs are set at their actual levels. In forward runs of the model, input supplies and hence output levels in these sectors are projected. exogenously according to plan projections and targets. In the remaining sectors, however, output supply depends not only on available inputs and production technology but also on the behavioral rules guiding the output and employment decisions of individual producers. 1.12 In the static CGE model, the capital stock is fixed at the sectoral level, and variations in the level of output supply depend on variations in the level of variable inputs. The model requires the specification of behavioral rules that influence producer choice about the levels of variable inputs to be utilized and of output to be produced under different market conditions. In CGE models for other countries the usual practice is to assume profit-maximization as the behavioral rule guiding input and output choice. In this approach, labor (or different types of labor, if more than one type is identified in the model) is assumed to be the variable input, and producers choose their demand for labor and hence their supply of output to maximize their profits. For a given set of product market prices and a given wage (or a given set of wages, one for each type of labor identified), labor demand is set where the wage equals the marginal value product of labor. Output supply is then determined via the production function given the available capital stock and the desired labor demand. According to this specification of firm behavior, changes in product prices and/or labor costs give rise to changes in labor demand and output supply. In this way, the model captures the factors underlying the supply responsiveness of the economy--or the way in which output supply depends on market prices. i.13 In the CGE model for Yugoslavia, the assumption of profit-maximization is replaced by a set of rules that attempt to capture the behavior of self-managed firms in Yugoslavia during the 1976-80 period. This specification incorporates two basic features of such behavior: rigidities in labor employment levels that limit the supply responsiveness of the firm to changes in product market conditions, and payment rules that produce a divergence between the marginal value product of labor and the personal income that labor receives. The specification of firm behavior in the model succeeds in capturing these features, albeit in a somewhat ad hoc fashion. 1.14 According to the rules of behavior adopted in the model, output decisions of self-managed firms are guided by a form of constrained profit-maximization. In making their output decisions, firms are assumed to treat a portion of their labor supply as fixed in the short run. This assumption reflects the fact that under the Yugoslav system, layoffs or reductions in employment levels are either severely restricted by law or are inconsistent with the likely behavioral rules of self-management. The model captures this institutional feature of Yugoslav reality by assuming that only a portion of total labor can be varied in the short run in response to changes in market indicators.  -7 - I.15 The model assumes that decisions about the amount of variable labor to be used are guided by a "planning" or "accounting" wage. For a given set of product market prices and a given planning wage, the demand for variable labor is set where the planning wage equals the marginal value product of variable labor. The distinction between variable and fixed labor has the effect of reducing the employment variability and supply responsiveness in the model relative to what they would be in a model based on unconstrained profit maximization. With a portion of its labor supply fixed in the short run, the short-run supply curve of the self-managed firm is steeper than the short-run supply curve of the profit-maximizing firm--i.e., changes in demand give rise to relatively larger changes in prices and relatively smaller changes in quantities in the self-managed firm than in the unconstrained profit-maximizing firm. 1.16 In both historical and forward runs of the model, the actual level of total employment in the productive, nonagricultural sectors is set exogenously. In historical runs, total employment is set at its actual level year by year while in forward runs it is set year by year at the level projected by the plan. Within each productive, nonagricultural sector, the distribution of employment between variable and fixed labor is also set exogenously, with lower skill and educational categories treated as variable labor and higher skill and educational categories treated as fixed labor. Combining exogenous information about total employment and its sectoral distribution between fixed and variable labor, the model calculates the total supply of variable labor available during each period. The demand for variable labor is derived by assuming that firms maximize "accounting profits" while valuing labor at the planning wage, and the model solves for the average planning wage that induces firms collectively to hire the total supply of variable labor available during each period. In this formulation, the total employment of variable labor is exogenously specified, but its sectoral composition varies in response to the hiring behavior of individual firms. The planning wage is solved for endogenously to equate the total demand for variable labor with its exogenously specified employment. 1.17 In contrast to variable labor, which is assumed to be mobile across firms and sectors, fixed labor is assumed to be sector-specific just as capital is. Once the amount of variable labor in each sector is determined, it is combined with the fixed labor supply and the fixed capital stock according to the sectoral production functions to determine the level of output in each sector. The sectoral production functions, along with the market prices and the planning wage for variable labor generated in the model solution, also generate estimates of the marginal value products of fixed labor and capital in each sector. Because both fixed labor and capital are immobile across sectors, their marginal value products differ across sectors. The marginal value product of fixed labor in each sector is interpreted as an accounting or planning wage for fixed labor in each sector while the marginal value product of capital in each sector is interpreted as a measure of the accounting rental rate or profitability of capital in each sector.  - 8 - 1.18 It should be noted that the estimated planning wages for variable and fixed labor, as well as the rental rate on capital, are defined gross of all taxes, except turnover taxes, and gross of depreciation. In order to go from planning wages and rental rates to observed payments made to different types of labor and to observed enterprise saving (retained profits), it is necessary to consider the income distribution decisions of the self-managed firm in greater detail. I.19 The model of firm behavior distinguishes between the planning wage that the firm uses in deciding the level of labor utilization and the actual personal incomes paia to firm members. While employment decisions are guided by the planning wage, workers receive a personal income that can differ from the planning wage and hence from their marginal value product to the firm. The difference between the average personal income paid to labor and the planning wage reflects the share of firm members in firm net income, exclusive of the planning wage. In this formulation, the actual personal income received by an enterprise worker can be thought to consist of two parts: a planning wage component and a component that reflects the worker's share in the firm's net income. Only the first part of this payment reflects the current marginal value contribution of the worker to the firm's output. Since workers also contribute to the firm's capital stock over time by their collective enterprise saving decisions, the second part of the total personal income payment - what might be called a profit share - reflects payments of the returns to capital to the workers who are its collective owners. I.20 The size of the second component of personal incomes depends on both the legal and contractual obligations affecting the distribution of enterprise income and on the firm's decision about what part of its disposable net income to distribute to workers and what part to retain in enterprise funds for investment. In the model, these shares are given by various fixed coefficients. Ihe share of various legal and contractual obligations (including depreciation) in total firm income (gross value-added), and the shares of saving and personal incomes in net disposable firm income (gross value-added minus depreciation and all taxes and contractual obligations) are fixed (and are estimated for each sector on the basis of historical information). It is important to note that enterprise income distribution rules can be adjusted in the model to reflect policies designed to influence firm decisions about saving and personal income payments. Given the importance of enterprise saving rates to the total accumulation (saving) ability of the economy and given concern with alternative policy measures to enhance enterprise saving, alternative model runs with different assumptions about enterprise distributional decisions can be useful for policy analysis. Foreign Trade and Foreign Exchange 1.21 In its current form, the CGE model treats exports as exogenous. In historical runs, the dollar values of sectoral exports are set at the values given in the Yugoslav foreign trade statistics. In forward runs, these values are set according to the projections of sectoral exports given in the economic balances used as the basis for the 1981-85 federal plan.  -9- 1.22 In the model, total foreign exchange available for imports depends not only on export earnings but also on available foreign capital flows. The CGE model distinguishes four different sources of capital flows: net remittance flows; net earnings from service exports other than construction, infrastructure and tourism (the earnings of which appear as export earnings of these sectors in the model); net long-term, medium-term and short-term capital flows minus interest payments; and changes in official foreign exchange holdings. In historical runs of the model, each of these capital flows is set at its actual annual dollar value for each year during the 1976-80 period. In forward runs, the dollar values of these flows are set exogenously according to the projections in the economic balances used as the basis for the 1981-85 federal plan. I.23 As noted earlier, desired imports by sector of origin depend on the relative prices of imports and imperfect domestic substitutes. The model solves for the total desired demand for foreign exchange for imports by summing desired imports across all sectors. Total desired demand for foreign exchange for imports is then compared with the total foreign exchange available for imports from export earnings and from foreign capital flows. In model simulations with a flexible exchange rate, the model allows the exchange rate to adjust until the desired demand for foreign exchange for imports equals the available supply. As an illustration, a flexible exchange rate simulation over the 1976-80 period solves for the exchange rate that would have equated the foreign exchange desired by importers to the available supply given by the actual values of export earnings and capital flows during the period. In the model, the exchange rate affects the desired demand for foreign exchange by influencing the domestic currency prices of imports--an increase in the dinar-dollar exchange rate raises these import prices, reducing the desired demand for imports and hence the demand for foreign exchange. I.24 In model simulations with a fixed exchange rate, the exchange rate is set exogenously during each period and does not adjust to equate the supply and demand of foreign exchange. In historical, fixed exchange rate simulations, the exchange rate is set at the official period-average parity each year. In forward, fixed exchange rate simulations the nominal exchange rate is projected to adjust according to different assumptions about the path of the real (price-deflated) exchange rate over the 1981-85 plan period. In all fixed exchange rate simulations, the exchange rate is set below its market-clearing level, reflecting the persistence of an overvalued dinar and excess demand for foreign exchange. With an overvalued exchange rate, the desired demand for imports in the model exceeds the available supply of foreign exchange, and the model has to specify rules whereby the available supply of foreign exchange is rationed among competing users. 1.25 Two kinds of rationing mechanisms which operated to varying degrees in Yugoslavia during the 1976-80 period are included in the model. Simple quantitative restrictions on either the allowable foreign exchange available for particular types of imports or for particular types of importers represent a fixprice mechanism for allocating foreign exchange.  - 10 - Under the fixprice or quantity rationing rules formulated in the CGE model, actual imports by sector of origin are restricted to some fraction of desired imports. This fraction varies by type of import in accordance with Yugoslav quantitative restrictions which are stricter on imports of capital and consumer goods than on imports of raw materials and industrial inputs. The impact of quantity rationing by sector of destination--that is, by the sector purchasing imports--depends on the composition of its import purchases and the quantitative restrictions applied to them. Under the quantity rationing regime, the domestic currency price of an import depends on the fixed exchange rate and remains fixed despite an overall shortage of foreign exchange. In this sense, the quantity rationing regime is a fixprice mechanism for allocating foreign exchange. 1.26 An alternative to the fixprice quantity rationing regime is the flexible-price or flexprice rationing scheme, also called the premium rationing scheme. The distinguishing feature of the flexprice mechanism is that potential users of foreign exchange who do not have access to foreign exchange in the amounts they need are able to pay a premium over the official exchange rate to obtain the foreign exchange they need. As noted in the text the operation of such a mechanism became possible following the creation of the communities of interest for foreign economic relations (CIFER) in 1978. Under a premium rationing scheme, the purchaser of an import evaluates the import at the official tariff-ridden price, marked up by the premium over the official exchange rate that must be paid to obtain the needed foreign exchange. A flexprice rationing mechanism presupposes the existence of a functioning secondary market for foreign exchange in which enterprises allocated foreign exchange (at the official rate) for import licences for their own import needs or retaining foreign exchange from their export earnings can sell foreign exchange at a premium above the official exchange rate. In the absence of legal, regional and other institutional barriers, the premium rate on foreign exchange transactions should rise just enough to make the demand for foreign exchange equal to the available supply. I.27 The CGE model allows for the coexistence of both fixprice and flexprice rationing, but it is necessary to specify exogenously the relative importance of fixprice versus flexprice rationing over time. 1/ In historical runs of the model, it was assumed that through 1978, quantity rationing was used to allocate foreign exchange while its significance declined in 1979 and 1980, as premium rationing grew in importance. With this assumption and given exogenously specified information about changes 1/ Technically, in the model, sectoral quantity rationing rates are specified exogenously and the additional premium rate needed to clear the foreign exchange market is determined endogenously. Alternatively, the quantity rationing rate can be solved endogenously, given exogenously specified premia rates. See Appendix II for further discussion.  -~ 11 - in the severity of quantitative restrictions on different categories of imports over time, the model yields both an overall quantity rationing rate -- defined and measured as the ratio of actual total imports to desired total imports -- and an overall premium rate required to equate the demand and supply for foreign exchange in the presence of existing quantitative restrictions. These rates are reported in Table 4.7 in the text. 1/ In forward runs of the model, alternative assumptions about the relative importance of quantity and premium rationing, and about the sectoral composition of quantitative restrictions, were tried. As an illustration, the forward run based on the original 1981-85 plan (the Plan I run) shows a slight increase in the severity of quantitative rationing -- the overall ratio of actual to desired imports falls from .90 in 1981 to .87 in 1985. This increase in quantitative rationing reflects model assumptions of increases in quantitative restrictions on imports of machinery, transportation equipment and electrical equipment, in line with plan targets to promote further import substitution in these sectors. 1.28 To understand the effects of fixprice and flexprice rationing of foreign exchange on economic performance, it is necessary to understand the distortions in the domestic prices of imports which they produce. The effects of flexprice rationing on the domestic price of imports are straightforward. The premium rate that emerges is the same in its effects on domestic prices of imports as an additional uniform tariff or import surcharge. A pure flexprice system is equivalent to setting such a uniform charge on import prices just high enough to make the demand for foreign exchange equal its supply. In a mixed fixprice-flexprice system, import prices are premium-ridden but by assumption the premium rate is not high enough to equate the supply and demand for foreign exchange. Some excess demand for foreign exchange remains, and even at premium-ridden prices, some importers are quantity constrained in the sense that they cannot purchase all the imports they desire. The resulting gap between actual and desired imports suggests that rationed importers would be willing to pay an additional charge, over and above the premium rate, to obtain the imports they desire, reflecting in value terms the severity of the remaining quantity rationing. This additional charge, which is called the value of the quantity restriction (VQR) in the model, varies between types of import depending on the severity of the quantity restriction applied. As an illustration, because quantity restrictions on the imports of capital goods and consumer goods are stricter than those applying to the imports of raw materials, the VQRs for imports of capital and consumer goods are larger than the VQRs for imports of raw materials. Given exogenously specified information about the sectoral composition of quantitative restrictions and 1/ The extent of exchange rate disequilibrium in the base year is essentially assumed since it reflects the base-year parameters and data. The quantitative rationing rate in 1976 is set exogenously and is calculated as an import-weighted average of exogenously specified sectoral rationing rates in that year. The sectoral rationing rates in turn are based on a comparison of sectoral import/output and import/value-added ratios in 1976 with trend values of these ratios over earlier historical periods.  - 12 - the model solution of the overall quantity rationing rate in a given year, the model calculates the VQR for each kind of import good (for each import by sector of origin). 1.29 The existence of both actual premia that some importers have to pay over the official exchange rate to get the foreign exchange they need and potential premia arising from quantitative rationing reflect rents arising in the system from import rationing. In turn, the existence of rents introduces the possibility of inefficiency and waste as demanders of imports seek to acquire desired foreign exchange through nonmarket means. Such "rent-seeking" behavior can have a real resource cost associated with it, as resources that might otherwise be used in production are redirected to such activities as lobbying for quantity allocations in communities of interest, negotiating complex agreements among groups of enterprises for the distribution of foreign exchange, and the like. Even if actual rent-seeking behavior does not develop to the full extent consistent with the existence of rents, it is of great interest to obtain measures of the rents generated by the rationing of foreign exchange. Such measures reflect the incentives for rent-seeking behavior and the resulting degree of strain on the resource allocation mechanism. As the incentives for rent-seeking behavior strengthen, the likelihood that resources will be wasted and output lost in the pursuit of rents increases. 1.30 To obtain a measure of the incentives for rent-seeking behavior, it is first necessary to specify the likely distribution of rents across agents in the economy. The model assumes that in the first instance these rents are distributed across sectors of production (sectors of destination in an input-output sense) according to their import purchases. Thus the rents available to an individual output sector depend on the rents that can be earned on its total purchases of imported inputs. The rent realizable from the purchase of a particular type of import is given by the difference between the value of the import at the official exchange rate plus tariffs and its value further marked up by the premium rate and the VQR applying to it. Calculating the rent on each type of import purchased in this way and summing across all imports purchased yields the total rents realizable by each sector from its total import purchases. The implication is that the "chaseable rents" in a given sector depend on both the magnitude and the composition of its import purchases -- in general, the larger its purchase of imports and the greater its purchase of imports that are subject to high levels of quantitative restrictions (and hence that have high VQRs), the greater the incentives for a particular sector to waste resources in the pursuit of foreign exchange. 1/ 1/ In this discussion, rents on import purchase are distributed by sectors of destination - that is, by the sectors purchasing imports rather than by the sectors in which imports originate. This is the formolation adopted for all rents realizable from purchases of imports of capital and intermediate goods. In contrast, rents realizable on imports of consumer goods are distributed by sector of origin - that is, by the sectors in which such imports originate in proportion to their share in total imports of consumer goods. Such consumer goods imports are relatively quite small.  - 13 - I.31 Table 2 contains estimates of rental income distributed across sectors in the manner described here. Table 2. RENTAL INCOME FROM QUANTITY AND PREMIUM RATIONING AS A PERCENT OF VALUE ADDED 1976 1977 1978 1979 1980 Agriculture 2.19 1.04 1.77 2.91 2.98 Electrical energy 20.03 13.96 21.72 29.64 33.71 Coal 8.73 5.37 6.85 8.37 8.27 Oil and gas 39.38 13.57 18.21 55.05 122.32 Ferrous metals 37.60 18.19 23.83 34.15 40.43 Nonferrous metals 18.34 11.36 16.62 21.45 23.17 Nonmetallic minerals and construction materials 9.52 4.31 5.72 6.62 7.52 Machinery and fabricated metals 15.84 9.27 12.92 16.07 17.59 Transportation equipment and shipbuilding 27.27 16.11 22.40 29.75 30.69 Electrical equipment 21.03 12.29 17.57 23.24 25.20 Chemicals and paper 25.73 14.87 19.08 29.04 30.41 Textiles, leather, rubber, wood and miscellaneous manufacturing 10.64 5.90 9.29 11.20 10.73 Food processing 8.12 4.84 9.71 16.05 16.64 Construction 7.23 3.11 5.04 6.09 7.69 Infrastructure 9.17 5.30 7.44 9.54 9.21 Trade 1.96 1.00 1.52 1.79 1.80 Crafts and other productive services 3.57 1.98 2.95 3.83 3.76 Primary (agriculture and food processing) 3.29 1.69 3.09 4.85 4.89 Industry 17.86 9.78 13.70 17.90 19.37 Services Li 4.57 2.48 3.69 4.46 4.47 Average 8.29 4.33 6.39 8.24 8.59 /1 Includes construction, infrastructure, trade, crafts and other productive services. The results indicate, as expected, that the overall degree of disequilibrium in the foreign exchange market, as measured by the behavior of rents arising from the rationing of foreign exchange, increased sharply in 1979 and 1980 as a result of the developing foreign exchange crises and the increasing overvaluation of the dinar. The results also indicate, as anticipated, that rental income as a percentage of value-added is greater  - 14 - for the more import-dependent sectors; that is, for sectors that depend most on purchases of imports of raw material and capital goods for production. Finally, the results suggest that the incentives for rent-seeking behavior in the system due to the value of "chaseable" rents were quite dramatic over the entire period. The strength of rent-seeking by the end of the 1976-80 period is also suggested by the fact that more and more domestic agents began to require foreign exchange in lieu of domestic currency on domestic transactions. I.32 Even with good will and social pressure, the power of these incentives must have encouraged rent-seeking behavior in a variety of forms, consistent with the institutional constraints facing economic actors. For example, rent-seeking undoubtedly made the process of reaching self-management agreements about foreign exchange and import rationing more time-consuming and divisive than would otherwise have been the case. Furthermore, many of the agreements involving the transfer of foreign exchange from one agent to another at the official exchange rate must have contained elaborate schemes whereby the actual price of the transfer reflected at least some of the effective premia. 1.33 This disequilibrium in the foreign exchange market and the rent-seeking behavior that it encourages can result in output losses due to the diversion of resources from production to rent-seeking activities and due to temporary shortfalls and delays in the supply of imported inputs. The model captures these effects in a somewhat ad hoc manner by assuming that the value of the wasted resources in each sector is a fraction of the rents accruing to the sector. In the case of rent-seeking behavior, this assumption implies that enterprises spend money to acquire rents, which is reasonable. If output losses are due to import bottlenecks, the model is assuming that such bottlenecks are proportional to the rents. While arbritrary, such an assumption is defensible since the rents are large~st in the most import-dependent sectors, which are also the sectors one would expect to be most disrupted by shortages. 1.34 In the preceding discussion of rents it was assumed that all rents accrued to the purchasers of imports. Such an assumption is reasonable for a system in which all available foreign exchange is distributed to import purchasers who then either use it for their own import needs or sell it at a premium to other importers. The situation is actually somewhat different in the Yugoslav system because of the rentention rights of exporters. According to the retention scheme, exporters are allowed to keep a variable fraction of their foreign exchange earnings. Given the excess demand for foreign exchange in the economy, this access to foreign exchange confers some of the rents arising from the rationing of foreign exchange on exporters. Under the circumstances, the actual distribution of rents among producers in the economy becomes very complex, depending on import purchases, export performance, the actual extent of discretionary control exporters exercise over the use of their retention quotas, and the extent of the secondary market for foreign exchange in the system. In its current version, the CGE model for Yugoslavia attempts to capture the effects of the interaction of these  - 15 - factors by assuming that exporters are allowed to retain effective discretionary control over a certain fraction of their earnings which they can use for their own import purchasers or sell at a premium rate to other import purchasers. Consistent with this assumption, the model distributes a fraction of total rents from import purchasers to exporters, the fraction depending on the assumed effective retention rate. As the discussion in the text indicates, the effective retention rate is assumed to vary over time depending on the extent of ad hoc intervention in exporters' control over their retained foreign exchange. I.35 The effective retention rate is used in the model to estimate the subsidy on exports provided by the retention scheme. This subsidy reflects the accrual of rents to exporters from their discretionary control over scarce foreign exchange. Formally, the subsidy rate on exports deriving from the retention scheme is calculated as a fixed percentage markup over the value of exports at the official exchange rate, where the markup is measured as the effective retention rate times the overall premium rate. As an example, if the effective retention rate is 20% and the premium rate is 10% (as in the model solution for 1980), then the subsidy rate on exports from the retention scheme is estimated at 2% of the ainar price. This calculation reflects the fact that the appropriate valuation of an additional dollar of retained foreign exchange is given by the premium rate, since the exporter always has the option of transferring his excess foreign exchange on the secondary market at this rate. Applying the estimated subsidy rate to sectoral export earnings yields estimates of total rental income accruing to exporters by sector, and these sectoral estimates can be used to analyze the distribution of rents among import purchasers and exporters in the system. I.36 In closing this section, it is important to emphasize that the export retention scheme permits exporters to capture some of the rents associated with their earnings of foreign exchange, and this in turn partially redresses the bias against exports that emerges from an overvalued exchange rate and reliance on fixprice and flexprice rationing schemes. If the effective retention rate were 100% and if the premium rate on foreign exchange were high enough to equate the supply and demand for foreign exchange without quantitative restrictions, then the system would be equivalent in its effects to a flexible exchange rate regime and there would be no bias against exports. In the Yugoslav case, where effective retention rates are limited and where the secondary market for foreign exchange is not extensive enough to produce a market-clearing premium rate, the net effect of the retention scheme is to reduce but not to eliminate the bias against exports in the system. Income and Product Demand 1.37 To complete the static CGE model, it is necessary to develop the demand side of the model for domestic production. This is accomplished by tracing through the incomes generated in the productive sectors of the economy and the various demands they induce. There are three classes of recipients of income in the model: productive enterprises in the  - 16 - agricultural and nonagricultural sectors; households; and the government or nonproductive sector. The model contains an elaborate set of accounting and behavioral rules to determine how the value-added or factor incomes generated in the course of production are distributed among these different income recipients. Basically, the distribution of value added depends on: (1) enterprise decisions about the distribution of their net income (dohodak) between personal incomes and various funds; and (2) the various taxes that influence the distribution of value added between the productive and nonproductive sectors. Once the model specifies enterprise saving decisions (as discussed earlier) and the various tax payments (porezi iz licnih dohodaka, porezi iz dohotka, etc.) from enterprise income, the distribution of income among enterprises (in the form of retained earnings), households, and the government is derived from a number of accounting identities. I.38 To obtain demands for output, the model specifies saving and expenditure behavior for the three types of income recipients identified. The income remaining with agricultural and nonagricultural enterprises after they have distributed personal incomes to their workers is assumed to be saved and made available for investment expenditure in fixed or working capital. The income flowing into the nonproductive sector from various taxes and other contributions from households and enterprises is used for four basic purposes: (1) to finance government expenditures on the output of the productive sectors; (2) to finance government transfer payments to households; (4) to finance direct government interventions in the economy; and (4) to finance investment. The fourth use of the government's funds is equivalent to gross government savings in the model. Finally, households are assumed to save a fraction of their disposable (after tax) income and to spend the rest on the output of the productive and nonproductive sectors. I.39 In the model, total investment activity in the economy is determined by the available supply of savings, including both domestic and foreign savings. The total supply of investable resources is distributed among different sectors according to exogenously specified shares. In historical runs, the sectoral distribution of investment is determined by actual sectoral shares in investment estimated over the 1976-79 period. In forward runs, sectoral investment shares are set exogenously to reflect plan targets and priorities. In the model, the sectoral shares of investment are an important policy variable, capturing the effects of policy-induced allocations of capital among priority sectors and barriers to capital mobility across sectors. The implications of different investment strategies on economic performance can be analyzed and contrasted by comparing the model results under alternative assumptions about the sectoral distribution of investment funds. 1.40 The allocation of investment shares by sector specifies the distribution of investment demand by sector of destination; that is, by the sector in which additions to the capital stock actually occur. Investment by sector of origin in turn refers to the demand for the output of a particular sector that results from the investment activity or additions to capital occurring in sectors of destination and is obtained through the use of a capital coefficients matrix.  - 17 - 1.41 In addition to investment demand for output, the model distinguishes intermediate demand and total consumption demand, encompassing both personal and government or collective consumption demand. Intermediate demand is determined through the use of the input-output matrix. Consumption demand is determined by using expenditure share parameters to divide up total household expenditure into demand for the output of each sector. Finally, collective or government consumption demand by sector is given by dividing up the total expenditure of the nonproductive sector according to the sectoral shares of collective expenditure in the input-output table. Market Clearing Conditions I.42 According to the CGE model formulation, for a given set of sectoral domestic prices and the exchange rate, it is possible to specify both sectoral supplies and sectoral demands. Market clearing requires that prices adjust until the resulting sectoral supplies and demands are equal. I.43 The task of the static, within-period CGE model is to find a set of prices such that these product market clearing conditions are satisfied and such that the total demand for foreign exchange is consistent with the available supply of foreign exchange. A solution repreRents an economywide equilibrium in product markets and in the foreign exchange market, given exogenously specified market constraints and exogenously specified availabilities of labor and capital. It is important to emphasize that the model does not assume that these resources are allocated among sectors according to the laws of supply and demand. Such an assumption does not seem warranted in the case of Yugoslavia, where capital markets and labor markets as traditionally conceived do not exist. In contrast to factor markets, however, product markets do exist, and it is the fundamental assumption of the CGE model that supply and demand decisions on these markets respond to prices. Given the specification of factor allocation mechanisms and of foreign exchange rationing, the model solution for product markets does not represent a neoclassical free-market equilibrium. But it does represent a market equilibrium solution constrained by behavioral and institutional specifications believed to represent a realistic picture of the Yugoslav economy. C. The Dynamic CGE Model I.44 In general, the role of the dynamic component of the CGE model is to update all the exogenous variables entering the static model. The most important exogenous variables that need updating include: world prices of imports and exports; the nominal exchange rate; the domestic price index; the sectoral allocation of investment funds, and the resulting changes in sectoral capital stocks; the growth of the different categories of domestic labor; and net foreign capital inflows. Some variables (remittance flows, household consumption parameters, the growth of domestic labor supplies) are updated using simple time trends, growth rates or exogenous projections. Others are updated to reflect assumptions about the likely effects of policy instruments. In historical runs of the model the  - 18 - exogenous variables are set year by year to match their actual historical values. In forward runs of the model, the projections either embody the plan's own assumptions (in the Plan I run) or alternative assumptions as described in Chapter V of the text. D. Trade Policy and Resource Allocation: Some Empirical,Results 1.45 This section provides more detail on the methodology underlying some of the uses of the CGE model in analyzing the Yugoslav economy reported in the main text. The main use of the model in the report has been to analyze issues of trade policy. In general, simulations of the model are used to isolate the effects of alternative assumptions on the workings of the economy. The model thus provides a framework for exploring the economywide impact of different policy regimes and development strategies. Measurement and Interpretation of Domestic Resource Cost 1.46 One of the objectives of the CGE analysis is to provide measures of the extent of the bias against exports and in favor of import substitutes arising from the Yugoslav trade and foreign exchange regime. An overall measure of the extent of this bias is provided by domestic resource cost (DRC) calculations. Formally, the DRC for an exporting or import-substituting activity is defined as the domestic cost of labor and capital resources per unit of foreign exchange earned by that activity. The foreign exchange is earned either directly in the case of an export activity or indirectly by saving foreign exchange in the case of an import substituting activity. In either case, the foreign exchange earned is measured net of the foreign exchange used for direct and indirect intermediate import requirements of the activity. Formally, the DRC for import substitution in sector i is given by: value-added per unit of domestic output in sector i (dinars) DRC. net foreign exchange saved by substituting a unit of domestic output of good i for imports (dollars) while the DRC for exports in sector i is given by: value-added per unit of domestic output in sector i (dinars) DRCe = net foreign exchange earned by exporting a unit of domestic output of good i (dollars) In both cases, value-added per unit of domestic output provides a measure of the total labor and capital costs per unit of output. In both cases, the DRCs have the dimensions of a dinar-dollar exchange rate, showing the dinars of domestic resource costs required per dollar of foreign exchange earned.  - 19 - 1.47 For a particular sector, differences between the DRC for import substitution and the DRC for exports reflect the biases in incentives that result from the trade and foreign exchange regimes. The DRCs for exports and import substituting activities within a sector will be the same when the effective exchange rate is the same for exports and imports in that sector. This condition requires that the nominal exchange rate augmented by tariffs and import premia on the import side equal the nominal exchange rate augmented by export subsidies on the export side; formally, EER = ER (l+tmi+PR)(l+VQRi) = EERi = ER (l+tei) where: i refers to particular sector production; EERe = is the effective exchange rate for imports; PR I is the economywide premium rate on foreign exchange; VQRi is the the value of the quantitative restriction on imports; EERi is the effective exchange rate for exports; tmi is the tariff on imports and tei is the subsidy rate on exports. If this condition holds and if domestic producers are indifferent between producing for export or for domestic markets, then the net foreign exchange earning from both export and import-substituting activities within a sector will be the same. Because the costs of production for either activity within a sector are the same, this implies that the DRCs for both activities will be the same under these conditions. I.48 Across sectors, differences in DRCs reflect both sectoral differences in the effects of the trade and exchange regimes -- tariffs, quantitative restrictions, premia and export subsidies -- and sectoral differences in indirect taxes, labor and capital costs. Sectoral differences in factor returns and indirect taxes cause distortions in market incentives generated in the domestic market. These effects can be taken into account in the DRC calculations by measuring labor and capital costs at estimated economywide shadow prices for labor and capital. To the extent that the estimated shadow prices of resources reflect their social scarcity, the use of such prices converts the DRC calculations from private indicators of the incentives facing individual producers at actual market prices to indicators of the social costs of the incentive system which has emerged. I.49 In the absence of an economywide objective function to generate shadow prices of labor and capital, the analysis in the text makes approximations to these prices based on a solution of the CGE model. Using estimated sectoral production functions and the model of firm behavior discussed earlier, the CGE solution provides estimates for the marginal value products of variable and fixed labor and the marginal value product of capital in each sector. These estimates provide measures of intersectoral differences in factor productivity and hence intersectoral differences in resource costs under the actual market conditions on which the CGE solution is based. To eliminate the effects of such differences in the shadow price DRC calculations, an average shadow wage and an average rental rate are calculated from the sectorally differentiated marginal  - 20 - value products of labor and rental rates and are used as approximations for the shadow prices of labor and capital. On the assumption that capital and labor resources are of different kinds and are not mobile between agriculture, manufacturing and services, three pairs of average wage and rental rates are calculated: an average wage and rental rate for agriculture, set equal to the wage and rental rate in the CGE solution; an average wage and rental rate for manufacturing set equal to the average wage and rental rates in all manufacturing sectors in the CGE solution; and an average wage and rental rate for services set equal to the average wage and rental rates in all service sectors in the CGE solution. These average wage and rental rates are used as approximations for the shadow prices of labor and capital in agriculture, all industrial sectors and all service sectors respectively in the DRC calculations. The use of these approximations eliminates intersectoral differences in DRC's due to intersectoral differences in labor and capital costs among manufacturing sectors and among service sectors, but not between agriculture, manufacturing sectors and service sectors. 1.50 The DRCs based on shadow prices of labor and capital in the text incorporate the approximations described here. Intersectoral differences in these shadow-price DRCs give an indication of the intersectoral distortions in incentives caused by the trade and exchange regime. In a neutral trade and exchange regime, the shadow price DRCs for export and import substitutes would be the same within each sector and across all manufacturing sectors and all service sectors, although they would differ between agriculture, manufacturing and services. The Interpretation and Measurement of Market_Clear,ing and Shadow~ Exchang,e Rates i.51 In the CGE model, the market-clearing exchange rate is defined as the nominal exchange rate (dinars per dollar) that equates the supply and demand for foreign exchange in a given period, given a set of tariffs and export subsidies, given flows of foreign capital and reserve accumulation, given domestic and world inflation rates and given existing taxes and product and factor market distortions. As thus defined, there is nothing necessarily optimal or sustainable about a market-clearing exchange rate. It is simply the exchange rate that equates the supply and demand for foreign exchange under these given conditions, thereby eliminating excess demand for foreign exchange and the need for either fixprice or flexprice rationing schemes and the retention schemes that often accompany them. The market-clearing exchange rate in a given period would not be normal or sustainable if, for example, the supply of foreign exchange during that period reflects abnormal or unsustainable factors, such as abnormally large export earnings from a transitory export surge or unsustainably large inflows of foreign aid or credit. Because the market-clearing exchange rate is estimated given distortions in price signals from existing tariffs and subsidies, existing taxes, and existing imperfections in product and factor markets, it is also not an appropriate measure of the shadow price of foreign exchange. Estimates of the shadow price of foreign exchange can however be obtained from estimates of the equilibrium exchange rate when  - 21 - the latter estimates are based on CGE runs in which such distortions have been eliminated by assumption. I.52 A flexible exchange rate simulation of the CGE model for the base period 1976-1980 yields estimates of the equilibrium exchange rate that would have equated the supply and demand for foreign exchange under the actual historical conditions of that period, including the actual dollar value of exports, the actual inflows of foreign capital, and the actual pattern of reserve flows. An alternative set of estimates based on a more normal pattern of reserve behavior, in which reserves are set to cover two months of imports, provides a path of "normal" or sustainable exchange rates which serves as a benchmark or reference against which the counterfactual scenarios described in experiments B-1 through B-4 are analyzed. The Contribution of Various Factors to the Depreciation of the Equilibrium Exchange Rate over the 1976-80 Period 1.53 The historical flexible exchange rate simulation indicates that under the actual conditions prevailing over the 1976-80 period, the equilibrium exchange rate would have depreciated from an estimated 22.33 in 1976 to an estimated 33.47 in 1980, assuming that sufficient reserves were held to cover two months of imports of goods and services in each year. In the absence of the indicated depreciation, the growing excess demand for foreign exchange was reflected instead in the growing use of fixprice and flexprice rationing and the growing rental income associated with exchange rate disequilibrium. Experiments B-1 through B-4 identify four factors that contributed to the indicated depreciation in the equilibrium exchange rate over the 1976-80 period: (1) the shortfall in export growth over the 1976-80 period; (2) the inflation differential between Yugoslavia and its trading partners; (3) the oil price shock; and (4) the decline in real net remittances. Comparing the estimated equilibrium exchange rate for 1980 in the absence of each of these factors with the estimated equilibrium exchange rate in their presence allows one to assess their relative contribution to Yugoslavia's growing foreign exchange difficulties during the 1976-80 period. This assessment is carried out with the estimates of the equilbrium exchange rates for 1980 contained in Table 4.11 in the text. Using the historical base period simulation as a benchmark, each of the experiments B-1 through B-4 progressively removes one of the four causal factors in order to explore its separate contribution to the estimated depreciation of the exchange rate. For example, comparing the 1980 equilibrium exchange rate of 33.467 from the historical simulation with the 1980 equilibrium exchange rate of 29.532 from experiment B-1 suggests that the shortfall in export growth relative to trend over the 1976-80 period would have warranted a depreciation of 33.467/29.532 or 13.3%. Similarly, comparing the 1980 equilibrium exchange rate from experiment B-1 with the 1980 equilibrium exchange rate from experiment B-2 suggests that the differential between Yugoslavia's inflation rate and that of its trading partners necessitated a depreciation of 29.532/24.367 or 21.2%. Continuing with experiments B-3 and B-4, the depreciations resulting from the oil price shock and the real decline in net remittances are as follows:  - 22 - Oil price shock: 24.367 or 5.9% 23.010 Real net remittance decline 23.010 or 7.3% 21.444 These results can be combined to calculate the percentage contributions of each of the four identified factors to the overall difference between the "shock-ridden" equilibrium exchange rate for 1980 in the presence of these factors (given by the 1980 equilibrium exchange rate from the historical simulation) and the "shock-free" equilibrium exchange rate for 1980 in their absence (given by the 1980 equilibrium exchange rate from experiment B-4) using the following relationship: 33.467 = (1.133)(1.212)(1.059)(1.073) (21.444) = (1.560) (21.444) The percent change between the "shock-ridden" and "shock-free" equilibrium exchange rates for 1980 is obtained by taking logarithms of both sides of this expression and subracting log (21.444) from log (33.467). The shares of the logarithms of each of the four terms in parenthesis on the right hand side expressed as a percentage of the log of their product (log 1.560) provides a measure of the percentage contribution of each of the four identified factors to the total difference between the "shock-ridden" and "shock-free" equilibrium exchange rate for 1980. These contributions are given in the text as: export shortfall 28.1% inflation differential 43.2% oil price increase 12.9% net remittance slowdown 15.8% Sectoral Market and Social Profitability Rates: Measurement and Interpretation I.54 Tables 4.2 and 4.3 in the text make use of estimates of market and social profitability rates to assess the priority investment strategy pursued under the 1976-80 plan. Market profitability rates are measured by the gross marginal value product of capital in each sector. As described earlier, these estimates are generated by the CGE solution given the production functions and behavioral assumptions on which the model rests. It is important to emphasize that the gross marginal value product of capital does not correspond to any accounting measure of profitability or net income used to assess the performance of the Yugoslav firm. Such measures depend on the taxes, contractual obligations, and distributional decisions of each firm. As a result accounting concepts, such as net income and accumulation, diverge from the estimated marginal value product of capital.  - 23 - I.55 The social profitability measures in Table 4.3 attempt to estimate the marginal value product of capital in each sector in the absence of market price distortions caused by turnover taxes, tariffs, export subsidies, quantitative import restrictions, and premium rationing and in the absence of sectoral differences in labor's marginal value product. The social profitability measures are generated in a CGE solution for 1980 that incorporates the following features: the exchange rate is allowed to adjust to equate the demand and supply of foreign exchange, thereby eliminating quantity and premium rationing; all tariffs, export subsidies and turnover taxes are set equal to zero; and the marginal value product of labor (or the planning wage) is equalized across all industrial sectors and across all service sectors, although it is allowed to differ between industry and services (reflecting the assumption of perfectly mobile and undifferentiated labor across industrial sectors and across service sectors but not between them). These features eliminate the effects of product market and labor market distortions on the return to capital, thereby providing approximate measures of the economic return to capital in each sector. These measures can then be used as rough indicators of an economically appropriate sectoral allocation of investment, although as noted in the text, such considerations as externalities and dynamic comparative advantage need also to be taken into account.  - 24 - APPENDIX I Table 3 PERCENTAGE COMPOSITION OF BASE DATA - 1980 GROSS DOMESTIC VALUE CAPITAL SECTOR OUTPUT IMPORTS EXPORTS DEMAND ADDED LABOR STOCK 1 AGRICULTURE 13.00 4.08 8.04 9.55 15.71 41.70 5.30 2 ELECTRICITY 1.66 0.00 0.00 1.30 1.12 0.44 10.84 3 COAL 0.96 0.38 1.79 0.19 1.07 0.79 1.22 4 OIL AND GAS 2.22 0.81 15.02 0.43 0.35 0.07 1.27 5 FERROUS METALS 2.92 1.47 7.01 0.20 0.69 0.47 2.15 6 NON-FERROUS METALS 1.58 2.85 2.10 0.15 0.76 0.37 2.32 7 NONMETALLIC MINERALS 2.77 2.03 3.09 0.42 2.97 2.06 2.36 8 MACHINES AND FAB MET 5.51 8.68 22.51 5.82 4.37 3.13 3.49 9 TRANSPORT AND SHIPS 3.68 6.39 8.62 3.87 2.27 1.26 2.04 10 ELECTRICAL MACHINERY 2.68 5.74 6.39 2.65 1.67 1.40 1.20 II CHEM,PAPER,AND WOOD 3.18 5.54 11.77 1.30 1.71 1.07 4.80 12 TEXTILES,LEATHER,RUB 10.19 17.62 5.36 7.77 8.25 10.10 5.23 13 FOOD PROCESSING 8.25 7.15 2.87 8.64 3.00 2.02 4.20 14 CONSTRUCTION 10.95 1.81 0.00 18.93 7.28 5.28 3.40 15 INFRASTRUCTURE 7.13 18.98 3.40 4.41 8.33 5.00 17.70 16 TRADE 9.67 3.61 0.00 12.49 17.39 6.48 4.00 17 CRAFTS & OTHER SERV. 7.52 12.86 2.03 8.25 9.53 5.84 5.06 i8 NONPROD SERVICE(GOV) 6.15 0.00 0.00 13.64 13.52 12.51 23.42 ALL SECTORS 100.00 100.00 100.00 100.00 100.00 100.00 100.00  APPENDIX I Table 4 TRADE AND PRODUCTION PARAMETERS - 1980 SECTOR SKAP SEX RMD AMPT TM RHOC RHOD 1 AGRICULTURE 36.45 2.55 9.87 12.16 13.51 2.00 1.00 2 ELECTRICITY 43.36 0.00 0.00 24.13 0.00 0.90 0.44 3 COAL 59.56 3.23 29.91 19.11 11.69 0.80 0.32 4 OIL AND GAS 99.19 2.98 108.61 44.86 0.00 0.50 0.37 5 FERROUS METALS 65.73 4.08 38.90 23.41 18.75 0.70 0.32 6 NON-FERROUS METALS 54.49 14.63 24.21 19.26 16.00 0.70 0.38 7 NONMETALLIC MINERALS 67.67 5.98 18.50 21.82 16.00 0.30 0.28 8 MACHINES AND FAB MET 55.80 12.80 72.84 26.61 19.71 0.85 0.32 9 TRANSPORT AND SHIPS 59.37 14.14 42.45 27.51 25.61 0.85 0.32 10 ELECTRICAL MACHINERY 56.81 17.45 45.00 25.59 23.10 0.87 0.10 ii CHEM,PAPER,AND WOOD 74.82 14.18 67.13 27.99 18.70 0.50 0.32 12 TEXTILES,LEATHER,RUB 42-60 14.06 9.51 13.31 22.19 1.10 0.26 13 FOOD PROCESSING 67.46 7.05 5.83 9.70 14.70 2.00 0.32 14 CONSTRUCTION 61.67 1.34 0.00 11.83 0.00 0.90 0.14 15 INFRASTRUCTURE 58.97 21.65 9.48 19.60 0.00 1.10 0.16 16 TRADE 50.97 3.04 0.00 14.78 0.00 0.90 0.32 17 CRAFTS & OTHER SERV. 55.03 13.90 4.87 12.32 0.00 1.10 0.61 18 NONPROD SERVICE(GOV) 30.01 0.00 0.00 0.00 0.00 0.90 0.10 ii ---------------------------------------------------------------------__------__--------------- EXPLANATION OF COLUMNS ____________ _ _______ SKAP = CAPITAL SHARE IN VALUE ADDED (%) SEX = EXPORTS/GROSS OUTPUT (%) RMD = IMPORT/DOMESTIC SUPPLY (%) AMPT = IMPORTED/TOTAL INTERMEDIATES (%) TM = TARIFF RATE (%) RHOC = TRADE SUBSTITUTION ELASTICITY RHOD = CAPITAL-LABOR SUBSTITUTION ELASTICITY  APPENDIX I Table 5 1980 INPUT OUTPUT COEFFICIENTS AGRICULT ELECTRIC COAL OIL AND FERROUS NON-FERR NONMETAL MACHINES TRANSPOR ELECTRIC COLUMN 1 2 3 4 5 6 7 8 9 10 ROW 1 AGRICULTURE 0.3010 0.0072 0.0090 0.OQO1 0.0016 0.0028 0.0039 0.0019 0.0010 0.0011 2 ELECTRICITY 0.0029 0.0523 0.0268 0.0091 0.0293 0.0420 0.0235 0.0112 0.0054 0.0126 3 COAL 0.0001 0.2838 0.1991 0.0000 0.0653 0.0203 0.0110 0.0062 0.0007 0.0018 4 OIL AND GAS 0.0160 0.1181 0.0176 0.7673 0.0357 0.0271 0.0924 0.0109 0.0089 0.0085 5 FERROUS METALS 0.0005 0.0105 0.0189 0.0103 0.5148 0.0106 0.0157 0.1788 0.0945 0.0695 6 NON-FERROUS METALS 0.0001 0.0008 0.0004 0.0002 0.0150 0.5776 0.0041 0.0309 0.0235 0.0870 7 NONMETALLIC MINERALS 0.0020 0.0101 0.0040 0.0035 0.0887 0.0420 0.1382 0.0192 0.0062 0.0272 8 MACHINES AND FAB MET 0.0065 0.0302 0.0387 0.0223 0.0198 0.0402 0.0295 0.2224 0.1518 0.1176 9 TRANSPORT AND SHIPS 0.0052 0.0023 0.0081 0.0007 0.0011 0.0035 0.0040 0.0054 0.3208 0.0008 10 ELECTRICAL MACHINERY 0.0008 0.0453 0.0034 0.0026 0.0017 0.0044 0.0064 0.0227 0.0327 0.2556 11 CHEM,PAPER,AND WOOD 0.0245 0.0048 0.0138 0.0170 0.0054 0.0225 0.0321 0.0191 0.0191 0.0539 12 TEXTILES,LEATHER,RUB 0.0051 0.0103 0.0170 0.0024 0.0033 0.0112 0.0202 0.0207 0.0338 0.0262 13 FOOD PROCESSING 0.0406 0.0000 0.0000 0.0001 0.0002 0.0005 0.0000 0.0001 0.0000 0.0000 14 CONSTRUCTION 0.0015 0.0097 0.0051 0.0013 0.0014 0.0020 0.0085 0.0091 0.0018 0.0069 15 INFRASTRUCTURE 0.0093 0.0403 0.0327 0.0581 0.0416 0.0516 0.0526 0.0334 0.0225 0.0308 16 TRADE 0.0191 0.0461 0.0294 0.0292 0.0295 0.0324 0.0446 0.0620 0.0506 0.0506 17 CRAFTS & OTHER SERV. 0.0116 0.0198 0.0721 0.0269 0.0531 0.0233 0.0370 0.0324 0.0279 0.0445 18 NONPROD SERVICE(GOV) 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 19 SUB-TOTAL 0.4470 0.6917 0.4961 0.9512 0.9073 0.9141 0.5239 0.6863 0.8013 0.7946 CHEM,PAP TEXTILES FOOD PRO CONSTRUC INFRASTR TRADE CRAFTS & NONPROD COLUMN 11 12 13 14 15 16 17 18 ROW 1 AGRICULTURE 0.0637 0.0823 0.4482 0.0147 0.0021 0.0112 0.0258 0.0000 2 ELECTRICITY 0.0264 0.0089 0.0061 0.0052 0.0143 0.0106 0.0111 0.0000 3 COAL 0.0084 0.0014 0.0019 0.0006 0.0077 0.0010 0.0014 0.0000 4 OIL AND GAS 0.0557 0.0097 0.0105 0.0163 0.0869 0.0156 0.0170 0.0000 5 FERROUS METALS 0.0049 0.0032 0.0010 0.0416 0.0060 0.0009 0.0076 0.0000 6 NON-FERROUS METALS 0.0130 0.0021 0.0011 0.0044 0.0005 *0.0000 0.0014 0.0000 7 NONMETALLIC MINERALS 0.0444 0.0046 0.0090 0.1387 0.0061 0.0034 0.0049 0.0000 8 MACHINES AND FAB MET 0.0302 0.0199 0.0138 0.0359 0.0177 0.0096 0.0303 0.0000 9 TRANSPORT AND SHIPS 0.0014 0.0005 0.0007 0.0047 0.0664 0.0040 0.0159- 0.0000 10 ELECTRICAL MACHINERY 0.0057 0.0029 0.0015 0.0204 0.0045 0.0041 0.0078 0.0000 11 CHEM,PAPER,AND WOOD 0.4652 0.0737 0.0259 0.0083 0.0051 0.0102 0.0100 0.0000 12 TEXTILES,LEATHER,RUB 0.0340 0.3595 0.0136 0.0648 0.0290 0.0230 0.0314 0.0000 13 FOOD PROCESSING 0.0128 0.0221 0.2324 0.0000 0.0003 0.0098 0.157,3 0.0000 14 CONSTRUCTION 0.0024 0.0026 0.0012 0.1784 0.0053 0.0063 0.0035 0.0000 15 INFRASTRUCTURE 0.0617 0.0306 0.0245 0.0683 0.1258 0.0339 0.0279 0.0000 16 TRADE 0.0578 0.0329 0.0352 0.0729 0.0462 0.0131 0.0450 0.0000 17 CRAFTS & OTHER SERV. 0.0417 0.0173 0.0174 0.0268 0.0736 0.0302 0.0432 0.0000 18 NONPROD SERVICE(GOV) 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 19 SUB-TOTAL 0.9295 0.6742 0.8440 0.7021 0.4974 0.1868 0.4415 0.0000  APPENDIX I Table 6 1980 CAPITAL COEFFICIENTS AGRICULT ELECTRIC COAL OIL AND FERROUS NON-FERR NONMETAL MACHINES TRANSPOR ELECTRIC COLUMN 1 2 3 4 5 6 7 8 9 10 ROW I AGRICULTURE 0.2486 0.0133 0.0000 0.0000 0.0000 0.0042 0.0016 0.0000 0.0000 0.0000 2 ELECTRICITY 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 3 COAL 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 4 OIL AND GAS 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 5 FERROUS METALS 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 6 NON-FERROUS METALS 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 7 NONMETALLIC MINERALS 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 8 MACHINES AND FAB MET 0.2177 0.5134 0.4393 0.3338 0.4229 0.4423 0.4044 0.4261 0.3685 0.4403 9 TRANSPORT AND SHIPS 0.0456 0.0086 0.0298 0.0128 0.0320 0.0301 0.0296 0.0270 0.1445 0.0224 10 ELECTRICAL MACHINERY 0.0564 0.1330 0.1138 0.0865 0.1096 0.1145 0.1048 0.1103 0.0954 0.1141 1i CHEM,PAPER,AND WOOD 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 112 TEXTILES,LEATHER,RUB 0.0058 0.0069 0.0069 0.0069 0.0069 0.0069 0.0069 0.0069 0.0067 0.0069 13 FOOD PROCESSING 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 14 CONSTRUCTION 0.3591 0.2451 0.3303 0.4803 0.3489 0.3224 0.3731 0.3498 0.3068 0.3364 i5 INFRASTRUCTURE 0.0034 0.0040 0.0040 0.0040 0.0040 0.0040 0.0040 0.0040 0.0039 0.0040 16 TRADE .0.0303 0.0361 0.0361 0.0361 0.0361 0.0361 0.0361 0.0361 0.0353 0.0362 17 CRAFTS & OTHER SERV. 0.0332 0.0395 0.0396 0.0396 0.0396 0.0395 0.0395 0.0396 0.0387 0.0397 18 NONPROD SERVICE(GOV) 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 t