Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Bulletin Number 95-3 March 1995 ECONOMIC DEVELOPMENT CENTER I ^777^Oda POLITICAL ECONOMY PERSPECTIVES ON THE 1994 TURKISH ECONOMIC CRISIS: A CGE MODELING ANALYSIS Erinc Yeldan ECONOMIC DEVELOPMENT CENTER Department of Economics, Minneapolis Department of Agricultural and Applied Economics, St. Paul UNIVERSITY OF MINNESOTA POLITICAL ECONOMY PERSPECTIVES ON THE 1994 TURKISH ECONOMIC CRISIS: A CGE Modeling Analysis A. Ering Yeldan University of Minnesota, St. Paul, visiting Bilkent University, Ankara February, 1995 Research for this paper was supported by a project grant from Harb-Is (Union for Defense Industry Workers of Turkey). I wish to acknowledge my indebtedness to Terry Roe, Xinshen Diao, Korkut Boratav, Tuncer Bulutay, Merih Celasun, Bilsay Kurug, Ahmet Ertugrul, Ahmet Tdktik, and to members of the Harb-Is Research Department for their invaluable comments and suggestions; and to Ahmet K6se and Alper Ydmaz for their diligent research assistance. Needless to say, none of them bear any responsibility on the possible errors and/or emissions of the propositions that the paper claim. POLITICAL ECONOMY PERSPECTIVES ON THE 1994 TURKISH ECONOMIC CRISIS: A CGE Modeling Analysis ABSTRACT This paper investigates the role of the deterioration of the Turkish public sector balances in the latter half of the 1980's and the evolution of the economic crisis with the aid of a computable general equilibrium model. The theoretical basis of the model utilized in the paper rests upon the Walrasian and the Structuralist/Keynesian macro foundations. The model recognizes 15 production sectors, 3 socio-economic income groups and a central government; and simulates the production, income distribution, consumption, capital accumulation and foreign trade processes of the national economy within a simultaneous system of algebraic equations. The distinguishing feature of the model is that, it accommodates monopolistic mark-up pricing rules in the industrial sectors, and endogenously solves for capacity utilization and unemployment level through Keynesian mechanisms of effective final demand. The model investigates the evolution of the crisis under three main headings: (i) the role of the financial crisis and the unprecedented deficit in the public sector balances; (ii) the roles of the foreign borrowing strategy of the state and the short term foreign capital inflows on balance of payments and the foreign exchange rate; and (iii) the role of the political-economic relations of income distribution and inflationary processes emanating from real wage increases and non-competitive pricing behavior in the industrial sectors. The general equilibrium results of the model underscore the importance of intra-class relations of income distribution and conflict in the evolution of price movements in the Turkish economy; and suggest that the sources of the crisis lie in the historical role of the administrative interventions of the state towards protection of the capitalist and rural incomes, which would otherwise be squeezed out in favor of wage-labor in the late-1980's. 2 POLITICAL [O.,IONOMY PERSPECTIVES ON THE 1994 TURKISH ECONOMIC CRISIS: A CGE Modeling Analysis Since January 1994, the Turkish economy has suffered from a major financial crisis which had been followed by a real contraction. Real Gross Domestic Product has declined more than 5 percentage points over the year, with inflation rate soaring to almost 150%, and the number of unemployed increasing by at least 600 thousand additional jobless people. It is a generally well-accepted fact today, that the main sources of the crisis can be traced down to the culminating pressures of large fiscal deficits on the part of the state. As a matter of fact, with observations of Public Sector Borrowing Requirement (PSBR) reaching around to 13% of the GDP by the end of 1993,' it should be regarded as no surprise that the fiscal imbalances would eventually erupt into a major financial disequilibrium in the way we have witnessed over 1994. In this paper, rather than merely summarizing these commonly known facts once again, I will attempt to analyze the basic underlying processes of the crisis from a political economy perspective of class conflict and pressure, together with conflicting interests on income distribution and accumulation. Thus, instead of regarding the realized fiscal imbalance of the state as a given fact -as if it originated in a vacuum with some abstract (a-historical) dynamics on its own- the paper proposes to investigate the underlying structural causes of the fiscal-cum-financial crisis from the point of view of pressures culminating from conflicting private interests over partitioning of national output. I believe it is only through such a class-based perspective of conflicting claims, that the very sources of the crisis can be dis-integrated and be depicted in a historical perspective. To this end, the paper employs a "Computable General Equilibrium" (CGE) model in order to investigate the evolving patterns of the crisis in the Turkish macro-economic context. The theoretical basis of the model rests upon the Walrasian general equilibrium system of optimizing agents facing endogenous price signals. In an attempt to reflect various structural features and bottlenecks thought to prevail in the Turkish economy -especially in the manufacturing sector- the model accommodates various market imperfections and rigidities borrowed from the structuralist/Keynesian tradition. Utilizing the CGE model as an "economics laboratory", I investigate the factors leading to the fiscal disequilibria and finally to the culminating crisis through explicit modeling of the strategic behavior of the antagonistic private agents in the factor and product markets. Our strategy will be as follows: first the model will be calibrated to the Turkish historical path 1990-1993 (section II) so as to generate the evolution of the 1994 crisis; and then, I will try to disentangle the various socio-economic processes behind this historical path via a series of counter-factual simulations (section III). In what follows, I introduce the main features of the model in section I, and reserve section IV for summary comments and conclusions. The formal algebraic equations of the CGE model are listed in the Appendix of the paper. I. The Computable General Equilibrium Model I-1. The General Structure and Dimensions The CGE model employed in this study is based on the recent 1990 Input-Output Table of the Turkish economy published by the State Institute of Statistics (SIS, 1994a). For the complete macro structure, it utilizes the 1990 Social Accounting Matrix which is discussed in full detail in K6se & Yeldan (1994). Formally, it is built around 15 production sectors; 5 socio-econoiiic classes; 3 behaviorally distinct private households and a central government. It seeks to represent the economic activities of its distinguishing agents via a simultaneous system of algebraic equations, which is brought into equilibrium with the aid of a set of endogenous prices, wage rates and the foreign exchange rate. The theoretical structure of the model is based on a synthesis between the elements of the traditional Walrasian general equilibrium framework (Dervis, de Melo & Robinson, 1982), and the structuralist (Taylor, 1981; 1991) schools of thought in an attempt to reflect a closer and more realistic depictions of the Turkish economy. Based on this synthesis, the distinguishing feature of the macro closure utilized in the model is a series of macro adjustments on income distribution, foreign exchange, and fiscal expenditures so as to create the necessary pool of aggregate savings to finance a predetermined level of investment expenditures. Within this adjustment process, in order to sustain the required level of savings, three mechanisms are put to work: (i) within 4 classic Kaldorian income transfer is called for from rural and the private sector worker households with lower saving propensities to high-saving capitalist classes; (ii) a portion of private savings is claimed by the state as coercive, forced savings in order to finance its fiscal deficit; and finally (iii) any insufficiency of the aggregate domestic funds is closed by alignments in the foreign rate of exchange and by foreign transfers. In the manufacturing industries, where a significant degree of concentration and monopoly power are thought to be exercised, product prices are set through fixed producer mark-ups over variable costs. Consequently, in the manufacturing sector the level of output supplied is determined by the aggregate final demand, given the mark-up For agriculture and services, given the lack of empirical induced market price. evidence on market structures, marginal cost pricing rules are hypothesized along with a neo-classical production function to determine the output supply. Given the mark-up based price structure, the model is able to capture the endogenous inflation pressures originating from structural rigidities and conflicting claims of various social classes on national output. The possibility of mark-up pricing results in both a distorted price structure and an income composition favoring profit/rent appropriators as discussed extensively in Boratav (1991) and Boratav, Tiirel & Yeldan (1994). The model distinguishes five socio-economic classes derived from their respective categories of income formation: Rural class earns the agricultural value added; industrial labor, receives manufacturing and energy sectoral wages; while services labor receives the services wages as income. The two urban capitalist classes, industrial and commercial/financial, capture the respective sectoral (residual) profits. A major limitation of the model, with respect to overall income generation, is its restricted capability to account for generation of financial/rentier incomes based on arbitrage possibilities in the banking and the financial sectors. Evidence suggests that with extensive liberalization in the financial markets and the capital account in the late 1980's, there has been a shift in the secondary relations of distribution favoring rentiers and the financial centers (see, e.g., Boratav, Tiirel & Yeldan, 1994; Yeldan, 1995). However, given its Walrasian structure, the model focuses exclusively on the real sphere of the domestic economy and does not have the appropriate mechanisms to capture financial flows and the consequent patterns of rentier income generation. 2 In this setting, the so-called profit income of the financial bourgeoisie, as distinguished in the model, is derived from the value added of the commercial and financial services sector by deducting wage payments. Thus, the financial sources of income can only be captured to the extent they emanate from transactions with the real sectors of the economy. In the model the private agents and the public sector are distinguished as two different bodies carrying out decisions on savings and composition of final demand. Within the private sector, the functional sources of income are channeled into 3 prototype private household groups: rural, urban-labor and urban-capitalist. It is through these three behaviorally distinct households that private saving and consumption activities are represented. 1-2. Production Technology and the Labor Market On the production, it is assumed that output of each sector is produced by a single representative firm. The primary inputs consist of aggregate (physical) capital and labor. Labor input is conceptualized into two categories of rural and urban, which are employed, respectively, in agriculture and the non-agricultural (industry and services) sectors. The "capital" input is a heterogeneous entity for each sector, made up by the fixed capital composition coefficients matrix, known in the literature as the B-Matrix. Stock of capital, once installed, is regarded fixed and sector-specific. Yet, at the end of each period, sectoral aggregate physical capital stocks are updated with the addition of fixed investments, net of depreciation. Following Celasun (1986) and Dervis et. al. (1982), the gestation lag of new fixed investments is regarded as one year. Sectoral depreciation allowances are compiled from the "depreciation costs" line of the 1990 Input-Output Table. The model regards aggregate labor supplies as fixed and does not account for workleisure choices at the worker-household level. Consequently, producers face an infinitely inelastic labor supply schedule independent of the wage rate. Labor demand decisions, on the other hand, are sensitive to real wage movements, and are derived from the marginal productivity conditions. In the urban labor market, nominal wage rate is exogenously set from outside of the model in an attempt to account for the rapid advances in industrial labor remunerations observed during 1990-1993. Thus, the urban labor market clears through quantity adjustments on employment, with a portion of urban labor remaining unemployed. For the agricultural sector, however, given lack of sound 6 data, the model resorts to flexible movements of the rural wage rate to clear the rural labor market. In the production sphere, the most important distinguishing attribute the model is with regards to its specification of the pricing strategy in the manufacturing sectors. In manufacturing, final product price is determined by imposing a fixed mark-up rate over the average variable costs. This specification is designed to reflect the oligopolistic market power which diverts the market price from the marginal cost, yet guarantees the producer a steady and given level of profit income. In response to increased wage and intermediate input costs, the producer secures its level of profits via application of the fixed mark-ups, thereby passing increased production costs on to the final consumers, inflating the overall producer price index level. This mechanism generates an important source of structural inflation in the manner described in Taylor (1991) and Bourguignon, de melo & Suwa (1991). Figures 1 and 2 depict these processes in the manufacturing sectors. Given the wage rate Wo in Figure 1, and the average variable costs as AVCo in Figure 2, the imposition of the mark-up, p, sets the product price at PX; and the quantity demanded can be read 0 from the aggregate demand schedule as To. An exogenous increase in the wage rate to Wi, pushes the average variable costs to AVCi, leading to a higher final price and a lower amount of quantity demanded. With lagging aggregate demand, output supply falls and, as specified in Figure 1, unemployment increases. [Insert Figures 1 and 2 approximately here] Under these behavioral assumptions, the sectoral production technologies are represented by constant elasticity of substitution (CES) production functions with capital and labor as primary inputs. The elasticities of substitution are adopted from the previous CGE modeling applications on the Turkish economy. As a general rule, the capital and labor substitution elasticity is regarded to be higher in agriculture as compared to industry and services. In the manufacturing sectors, where mark-up induced pricing behavior is thought to be prevalent, market supply is determined by the endogenous rate of capacity utilization. In this manner, denoting capital by K, labor by L, the underlying production function by F(K,L), and the capacity utilization rate by U, output supply becomes, 7 wage 1 W0 Lemand tmployment Figure 1. The urban-labor market pr VC 0 T T I, quantity u Figure 2. Oligopolistic pricing in the manufacturing industries Xs = U F(K,L). In this fashion, under the pre-determined price, product market equilibrium is obtained by capacity adjustments, which in turn lead to reduced employment and supply at the sectoral level. 1-3. Determination of Foreign Trade The model regards world prices as exogenously given (the small country assumption). Domestic import and export functions are derived through the so-called Armington commodity specification of the traditional CGE modeling exercises. Accordingly, within each traded sector, the domestically produced good (DC), the import (M), and the export (E) are differentiated from each other by way of imperfect substitutability. Product differentiation, in this context, is given from outside the model by way of independent specification of an elasticity of substitution (in the case of import) and of transformation (in the case of export). The import and export functions of the economy are derived, then, as solutions to the respective optimization problems of the consumers and the producers, given the relative price signals. Determination of imports at the sectoral level is depicted in Figure 3. Domestic good and sectoral import form a consumption composite along the convex isoquant yielding the final consumer a given level of satisfaction. Given the import-domestic good relative price ratio pM, cost minimizing amount of import is Mo. If the relative price of import increases to Pm (by way of, say, increased tariffs or currency depreciation), imports are reduced to Mi. Determination of exports follow the same principles. In Figure 4, combinations of supply to the domestic market, DC, and export, E, are given by the concave transformation possibility frontier. Faced with a relative export-domestic good price ratio PE, producer maximizes its revenues at the export allocation Eo. An increase rate of foreign exchange (depreciation) leads to a higher export to domestic price ratio, PE, increasing sectoral exports to the level Ei. [Insert figures 3 and 4 approximately here] Given the aggregate demand for imports and export earnings in foreign currency 8 MO M1 composite good schedule MO1%.. = P%.,I AA oumestii guuu Figure 3. Determination of import demand domestic good Figure 4. Determination of export supply determined in this manner, the balance of payments constraint faced by the economy becomes: P'M = PE. + REMIT + FSAV + GBOR where REMIT denotes exogenous (in)flows of workers' remittances from abroad and GBOR denotes government borrowing from the international finance centers. Residual foreign savings to close the current account deficit is given by FSAV. These three quantities are set at their historically realized levels in the model experiments. The balance of payments equilibrium is maintained by endogenous adjustments of the (nominal) exchange rate. With complete liberalization of the capital account and the shift to a fully convertible foreign exchange rate in 1989, this approach is definitely the most appropriate one in the Turkish context. 1-4. Income Generation and Fiscal Balances Private household incomes are generated from sources of value added income, transfers from the central government budget, and exogenous transfers from the rest of the world. Private savings are determined via fixed saving propensity parameters out of private disposable incomes; the residual income being spent in final consumption expenditures. Sectoral allocation of consumption expenditures are given by fixed shares (with the underlying assumption that preferences are Cobb-Douglas) derived from the 1990 Input-Output Table. Another distinguishing attribute of the model is its detailed description of the public accounts so as to trace out the inflationary pressures emanating from the fiscal (im)-balances of the state. The public disposable income consists of income taxes on the private sector, tariffs on imports, and the production and corporate taxes levied on the producers. Aggregate public expenditures, on the other hand, are composed of public consumption and investment, current transfers to the private sector, subsidies on exports, and debt repayments to the rest of the world. In line with the macro closure imposed in the model, the public savings rate and the aggregate level of public investment are regarded as exogenous policy instruments, given their historically realized values. In general, planned public savings fall short of 9 the thereby emerging public savingand the realized investment expenditures; investment deficit is covered through foreign borrowing plus forced transfers from the private savings pool. This mechanism entails, on the one hand, excess demand disequilibria in the product markets leading to inflationary pressures; and, on the other hand, leads to real appreciation of the domestic currency through induced inflows of foreign borrowing and short term direct portfolio investments. 3 The macro-closure of the model can be summarized more formally as follows: Let GREV, GDTOT and GSAV denote, respectively, the public disposable income, public consumption expenditures and public savings. Then, GREV = GDTOT + GSAV Here, GDTOT is taken exogenous at its current (nominal) historical value, and GSAV is determined via application of the historically realized public saving rate out of public With aggregate public investments, GINV, pre-determined (again at its income. historical value), the realized saving-investment deficit of the public sector becomes: GSIDEF = GINV - GSAV This deficit is the main source of the public sector borrowing requirement which, in turn, is to be financed from the private savings surplus (HHSAV-PRINV), and net inflows from the foreign economy (FSAV-ER). Thus, the overall macro-equilibrium of the domestic economy becomes: GSIDEF = HHSAV - PRINV + FSAV-ER This mechanism -which blends elements of neo-classical closure on the private sector together with a Kaldorian pre-determined investment level on the part of the public sector- has a dual implication for the domestic economy: on the one hand, it generates inflation of the overall price index, contracting private real incomes; and on the other, through depleting investable funds, it reduces the productive potential of I believe that this structure the economy, restricting it to a lower growth path. captures quite successfully the most recent Turkish reality with culminating inflationary pressures at rates of 60 to 70 percent per annum; faltering and erratic rates of growth with boom-and-bust cycles of real production; and stagnant exports and 10 an unprecedentedly high current account deficit accompanied by currency appreciation (in purchase-price-parity terms). This summarizes our discussion of the main attributes and distinguishing features of the CGE model. We now turn to applications to the Turkish historical growth path 1990 through 1993 in the next two sections. H. Historical Simulations of the Model, 1990-1993 Here I first utilize the model to track the historical 1990-1993 path of the Turkish economy. It is important to note at the outset here, that one cannot expect for any model to generate the "real life" on a one-to-one basis. Obviously, any modeling exercise is but only a gross approximation of the reality, with the most important variables of interest to be focused at a more detailed fashion, leaving others of secondary interest to be outside the realm of the model's coverage. In this exercise then, we focus first and foremost, on the variables and economic relations that deem primary importance for our purpose at hand, namely that of analytically depicting the underlying structural processes and main mechanisms which have culminated to the 1994 crisis. To this end, the most important variables of interest that one will need to track closely, offer themselves quite naturally to the list of: the annual rate of domestic inflation, the exchange rate, grossrnational product and sectoral production, wages, employment, fiscal balances, foreign trade transactions, and the macro aggregates such as consumption, savings and investments. The values of the historically realized levels of the a-forementioned variables between 1990 and 1993 are to be contrasted below with the model's simulation values in Tables 1, 2 and 3, and the Figures 5, 6, and 7. The historical validation exercise is carried out in 3 steps as summarized in Celasun (1986: 44): first, the model's algebraic equations are calibrated to a social accounting matrix of a given base-year data set (1990). Here the values of various share and shift parameters of the algebraic equations are generated so as to reproduce the base-year data set as a "solution" of the model. Thus, the "base-year" solution values of the model exactly replicate the SAM values of 1990. Beginning from this solution set, in the second stage, the historically observed values of various exogenous variables and policy parameters over the relevant time span, 1990-1993, are fed into the model as given data. In the third stage, the technical parameters of the model are re-adjusted, 11 if necessary, in an iterative manner until the necessary "convergence" of the endogenous variables with their historical values is achieved. The values of various public sector aggregates that are taken as exogenously given under the historical validation exercise are listed in Table 4. Here, the realized ratios of the public sector's consumption, savings and investment to the aggregate GNP are utilized as given policy variables; and are parametrically changed in the next section, in our quest for the answer to "what would likely to have happened if the respective policy actions had assumed different values?". These "counter-factual" model simulations will provide valuable insights on the role played by the government's policy actions on the culminating crisis. In the foreign economic sphere, inflows of "foreign capital" and "workers' remittances" are set at their given values. With respect to the world prices of traded commodities, a uniform rate of 5% foreign inflation is assumed. In the labor markets, following the "supply of civilian labor" line of the SIS Labor Statistics, domestic labor supply is increased at a rate of 1.5% per annum. Finally, in the non-agricultural sectors, an annual rate of 3% growth of neutral total factor productivity is assumed. We take the model's historical performance in turn. II-1. The Historical and the Model Simulation Values of Domestic Prices, Exchange Rate and The Macro Aggregates, 1990-1993 The three most important prices for our analysis are the overall domestic price level, the foreign rate of exchange (TL/US $), and the urban-labor wage rate. Among these, the first two are solved endogenously by the model, whereas the nominal wage rate is regarded as given by non-economic factors. The 1990-1993 historical and the simulated values of the price level and the exchange rate are given in Table 1. The domestic price level reflected in the model is the "producer price index" and is a weighted average of the gross sectoral producer prices. The foreign exchange rate reflects the "annual average" of the daily domestic currency values of the US dollar. In the model, this variable is endogenously solved to close the balance of payments accounts. As can be observed both variables trace out their historical values closely, the highest discrepancy being observed in the 1991 rate of foreign exchange with a divergence of 9%. 1991 is the year when foreign savings turned into negative and the remittances declined by %12. Yet, in the same year, the 12 current account balance was observed to yield a surplus of $258 million. Thus this asymmetric movement in the balance of payments results in the observed divergence as the In the remaining two years, model tries to account for these opposite trends. discrepancy of the simulated value of the exchange rate is less than 0.5%. [Insert Table 1 approximately here] Figure 5 portrays the urban-labor market. As stated above, nominal wages are given exogenously. The historical wage rate developments are adapted from the SIS Manufacturing Industry Surveys, and, in general, represent the "organized/formal" labor wages. In this regard, by implementing the manufacturing wage figures over to the whole economy, the historical data carries a bias towards exaggerating wage cost increases in the specified period. Consequently, in the model simulations, the urban wage rates are raised by 40% in real terms over the three years. In nominal values this figure amounts to an increase of 465%. [Insert Figure 5 approximately here] The rate of unemployment is depicted in reference to the left-hand axis of Figure 5. The simulated unemployment ratio exceeds the official historical rate by as much as 4 percentage points; but generally follows the realized trend. Considering that there exists important differences in the definitions of the unemployment concept used in the model vis-a-vis the official statistics, and the dubious quality of the labor statistics overall, the model solutions in the labor market were regarded as "sufficiently close" to the historical realizations. In fact, the production figures as reported by the model track their historical values very closely; and reproduce the value of gross national product both in net factor prices and in producer values. As can be observed from Table 2, the only discrepancy in the nominal GNP is observed in 1993, at the rate of 1.5%. Components of expenditures on the national product are listed in the remaining lines of Table 2, where the model's expenditure structure is found to compare closely with the historical structure. 4 [Insert Table 2 approximately here] 13 Table 1. Prices, Wages and the Foreign Exchange rate 1990 Price Level (a) Foreign Exchange TL Urban Wage index(b) Price Indexes: Agriculture Mining Industry Model 1991 Historical Model 1992 Historical Model 1993 Historical 100 2607 100 158 4565 202 159 4181 202 258 6887 353 257 6882 353 415 10934 565 408 11007 565 100 100 100 146 163 172 151 163 155 255 263 267 245 262 253 421 426 448 398 413 395 Notes: a. Producer Price index b. In nominal prices (regarded as exogenous to the model) Table 2. Gross National Product and the Macro Aggregates GNP (a) GNP(f.p)(a) Model/Hist GNP Composition of GNP (%) Pub Disp Income Public Consumption Public Savings Public Investment Public Sav-lnv Deficit Private Private Private Private Disp Income Consumption Savings Investment Foreign Deficit (b) 1990 395.3 361.4 1991 Model Historical 621.1 622.5 561.7 562.4 1.0 1992 Model Historical 1072.1 1069.1 969.1 967.5 1.0 1993 Model Historical 1788.7 1818.3 1638.6 1617.2 1.0 13.5 10.2 3.3 8.6 5.2 13.1 12.9 0.5 8.8 8.4 11.7 11.6 0.5 7.6 7.3 12.5 13.5 -1.0 7.4 8.5 11.6 12.6 -1.0 6.8 7.9 10.3 13.7 3.4 6.9 10.4 86.5 69.1 16.8 14.9 88.2 69.4 18.7 10.3 88.3 72.0 16.3 8.5 88.9 70.1 18.8 12.3 88.4 73.3 15.1 9.4 90.7 71.2 19.5 14.1 3.3 -0.1 -0.2 2.1 2.1 5.3 Notes: a. SIS, new GNP series, in current trillion TL. b. Balance of payments, "capital movements" item. Sources: Department of Treasury; SIS; Boratav et.al (1994). -3.4 6.9 10.3 5.4 Figure 5. Manufacturing Wages and Unemployment Unemployment rate f) 0 12 10- Wage Rate rI - k cn I UU Wage rate -IMM-N-ummm 140 . 120 8 - 100 80 6 U Model D Historical 60 40 2 20 0 - -- -- - 1990 1991 1992 Years - 1993 0 1-2. The Historical and the Model Simulation Values of Foreign Trade Variables, 19901993 Variables pertaining the foreign trade relations are documented in Table 3. We observe that even though "import" variable is tracked closely, the model's "export" figure suffers from higher discrepancies vis-a-vis to its realizations. (See Figures 6 and 7). The foreign currency values of aggregate exports generally track the historical path, yet, a band of close to $ 1 billion dollars are kept between the simulated and the realized values. Here the important data problem was to convert the export magnitudes which were reported in Turkish Lira units in the SIS 1990 Input-Output Table, into foreign currency equivalents by using a single (yearly average) foreign rate of exchange. This was then complemented with export data obtained from other sources (mainly from the Treasury Trade Statistics) for the rest of the period. In reality, the foreign exchange rate adjusts daily, and thus, the aggregate value of foreign trade transactions recorded in domestic versus foreign currency denominations is expected to yield differences, especially when there is a need to reconcile different data sources. In a similar vein, the composition of exports reported in the SIS Input-Output Table was not consistent with the similar data reported by the Department of Treasury Trade Statistics, which was the main source of foreign trade data for the post-1990 period. In particular, exports of primary agriculture were substantially lower in the SIS data. However, in order not to violate the overall consistency of the 1990 SAM balances reflected from the SIS, I had to accept the divergencies of the Input-Output data as the given structure. [Insert Table 3 approximately here] [Insert Figures 6 and 7 approximately here] The values of the exogenous variables along the historical path are documented in Table 4. Table 5, in turn, lists the sectoral mark-up rates implemented throughout the validation exercise, along with the endogenous solution values of the "capacity utilization rates". In the course of the historical path, based on the evidence of Istanbul Chamber of Industry (1993) data, the mark-up rates are adjusted upwards by 20% between 1990 and 1991; and held at their respective rates onwards. 14 Figure 6. Ex orts 16 Historical Model 1412- - 10CO en o 8 - 5m 6-420 1990 t 1991 Years I i i 1992 1993 Figure7. Imports 30- Historical Model 25 •n Wo. 15 m 105 01990 - 1 1991 1992 Years 1993 Table 3. Foreign Trade Exports (mill $) Composition of exports (%) Agriculture Industry 1990 12959 1991 Model Historical 13593 11320 1992 Model Historical 14715 13425 1993 Model Historical 15349 14379 6.2 18.1 81.9 9.1 90.9 19.7 80.3 92.5 14.8 85.2 93.8 15.1 84.9 22302 Imports (mill $) Composition of imports (%) 5.9 Agriculture 12.9 Consumer manufact 81.2 Producer manufact 21909 21047 24596 22872 28388 29429 4.2 3.8 13.8 4.5 13.1 82.4 82.4 5.2 13 81.8 4.3 12.7 83.1 13.2 82.6 7.5 5.8 13.9 80.3 Table 4. Exogenous Variables under the Historical Simulation (In nominal prices) 1990 34225.80 Public Investment 43083.50 Public Consumption 16239.20 Current Transfers 3.47 Public Savings/GNP (%) 3.12 Public Factor Income/GNP (%) -4280.70 Public Foreign Borrowing (a) 12928.20 Foreign Deficit (b) 3248.00 Workers' Remittances (c) Hicks-neutral productivity growth (%) 1993 1992 1991 54761.30 80178.30 27606.60 0.44 0.65 1326.80 -456.60 2819.00 3.00 79403.80 144802.10 46931.30 124664.10 246163.50 131407.60 -1.04 -3.42 -0.08 1571.90 0.03 1936.30 95126.80 2790.00 3.00 21351.40 3008.00 3.00 Notes: a. Net foreign borrowings item in the central budget accounts of the Treasury. b.Balance of payments, capital movements item; converted into TL using the midyear foreign exchange c. Million US$ Table 5. The Mark-up Rates and Capacity Utilization in Manufacturing Industry (%) Food Processing Textiles, Clothing Chemicals Petroleum Products Soil Products Metals Machinery Mark-up Rates 1993 1990 22.2 26.7 37.4 31.1 36.1 30.1 42.8 35.6 58.1 21.1 29.9 Capacity Utilization Rate 1992 1991 1993 95.1 91.5 92.1 99.9 89.7 87.9 89.8 69.7 95.3 90.6 90.3 95.6 83.1 85.4 85.9 25.4 89.3 92.5 92.3 35.9 86.1 89.4 88.9 99.4 [Insert Table 4 approximately here] [Insert Table 5 approximately here] Model results indicate that the highest rate of capital utilization is realized in the petroleum products industry, with the lowest rate being observed in chemicals and the soil products. It has to be noted here, that the capacity utilization figures as reported by the model do not suggest any propositions with regards to sectoral managerial activity and/or marketing difficulties as such, but should be taken mainly as a "technical" variable to close the product market commodity balances. With the final price pre-determined by the fixed mark-ups, the marketed output is adjusted by the "utilization" parameter in an attempt to bring forth equilibrium with aggregate demand. Thus, the size of the capacity utilization rate should be interpreted as an indicator of the overall vigor of the domestic economy along the lines of Keynesian effective demand, rather than as a marketing problem of stock management. III. Evolution of the Crisis and Its Components In this section, I will utilize the 1990-1993 simulated growth path as a benchmark to investigate quantitatively the processes behind the 1994 crisis via a set of counterfactual scenarios. I take up three interrelated issues: (i) the role of the financial crisis and the unprecedented deficit in the public sector balances; (ii) the roles of the foreign borrowing strategy of the state and the short term foreign capital inflows on balance of payments and the foreign exchange rate; and (iii) the role of the political-economic relations of income distribution and inflationary processes emanating from real wage increases and non-competitive pricing behavior in the industrial sectors. I-1. Fiscal Balances of the State: Experiment El Over the analyzed period, the most striking indicator over the public sector balances pertain to its borrowing requirements. As a ratio of GNP, the PSBR increased rapidly from 7.4% in 1990 to reach 10.9% in 1991, and has shown a rigid stability at that plateau. Concurrently, volume of public savings deteriorated severely, and fell 15 from 3.5% of GNP in 1990, to 0.4% in 1991, and to -3.5% in 1993. This collapse was also reflected on public investments, which experienced a decrease of 3.8% annually in real terms over 1990-1993. In addition to these developments, net factor earnings of the public sector -an item which consists mainly of net revenues of the state enterprises- fell from their GNP ratio of 5% in 1985, to 3.1% in 1990, and to 0.6% in 1991, with being practically zeroed out in the remaining of the period. Thus, the public sector has suffered an income loss of close to 5% of the GNP in just five years. On the other hand, the 1993 value of the current transfers, which is the major expenditure item in the fiscal accounts, exceeded its 1990 level by 70% in real terms, amounting to an increase of 6.6-folds in nominal prices. Had these adverse developments not occurred in the public sector balances, what would their consequences be for the domestic economy? How would the macro balances differ from the realized path? In this first experiment, I will take these questions entirely at an hypothetical level in the abstract, to study quantitatively the contribution of the deterioration of the fiscal balances on the evolution of the crisis. Under Experiment El, I first increase the public savings-GNP ratio to its 1980's average of 4%, from its realized value of -1.3% over 1991-1993; and reduce the annual rate of increase of nominal public consumption expenditures from 78% to 45%. Likewise, the rate of growth of current transfers is reduced to 45% in nominal terms; and finally, to compensate for the revenue losses of the state, net factor earnings of the public sector is increased to its 1980's average of 3% of the GNP, from the 1990-1993 average of 0.3%. The results of the experiment are documented under the heading El in Tables 8, 9 and 10. The simulated rate of inflation (in producer prices) is portrayed in Figure 8. Accordingly, under the assumptions of the experiment, the annual rate of price increase is calculated to average 35% over the 1990-1993 period, down from its historically realized rate of 60%. This result is further complemented by positive developments in employment, which increases by 2 index points; and in production, where real GNP increases by 3.1% over its 1993 value. [Insert Figure 8 approximately here] The experiment results reveal an annual increase of 19% in the public disposable income in real terms. This is possible because of the implemented parametric increases in the net factor earnings and also due to the increased tax earnings as a result of 16 Figure 8. Domestic Inflation Rate: Experiment E1 Arr 43U 400 350 300 . 250 ' 200 El 150 100 50 0 1990 1991 1992 Years 1993 expansion in the economic activity. With explicit assumptions on public expenditures calling for real reductions through the experiment, the public saving-investment gap is. reduced by 26.3% compared to its 1993 value, attaining a ratio of 7.4% to the GNP. [Insert Tables 6, 7 and 8 approximately here] At this juncture we should ask the following: "Is it possible to make such an isolated assessment of fiscal balances of the state, independent from all the historical socio-economic developments of the period?" In order words, could the Turkish state have replicated the model's technical results which were obtained under a set of exogenous assumptions? This question invokes the political economy issues of income distribution and real income generation within the private sector. Table 8 suggests some interesting generalizations on this issue. For instance we observe that, under the assumptions of the experiment, the real income level of the urban-capitalist household falls by %4.2 in 1991, and by 5.2% in 1993, as compared to its historical values. In similar vein, rural incomes also suffer a loss of 3.1%; with the urban-workers becoming the only group which becomes slightly better off. The major reason why the former groups suffer a loss is because of the reduction of the current government "transfers" to the private sector and diversion of factor incomes in favor of the public sector. The worker household maintains its income mainly due to the fact that nominal wages are assumed constant at their historical levels throughout the experiment. Consequently, under the assumptions of the experiment while public income improves, private incomes erode. One can underline the experiment results in the following manner: in the Turkish economy under the conditions of the early 1990's, during when urban real wages scored rapid increases, and the production and export performance of the domestic economy gave faltering signals, the rapid erosion of the fiscal balances of the state served for the protection of the urban-capitalist and rural incomes in real terms. This protection amounted to as much as 5.9% for the urban-capitalists, and to 3.1% for the rural households. The costs of this "protection" to the domestic economy entailed the rapid surge of the domestic price inflation from its possible annual rate of 35% to 60%, and the loss of aggregate investment growth of 17.8%, and export revenues of 3.8%. Thus, I advance the following proposition: in the Turkish economy the public sector fiscal crisis of the early 1990's have served for the crucial function of sustaining and re-distributing the private capitalist incomes. At the micro level, this meant 17 Table 6. Macro Economic Indicators: Experiments E1 & E2 Historical (a) 1991 1993 1990 3I95775.3 392807.6 430427.5 GNP 12928.2 -288.7 22891.2 Foreign Deficit 56827.3 52441.3 44506.8 Public Disp Income Private Disp Income 340072.9 346507.4 391147.2 Aggregate Consumption 43083.5 50706.3 59236.5 Public 273501.1 272685.1 309041.8 Private Domestic Savings 13743.8 1735 -14729.7 Public 66571.8 73822.3 82105.4 Private Aggregate Investment 34225.8 34632.1 29999.1 Public 59017.9 40636.5 60267.9 Private 11187.9 11320.5 14379.1 Exports (c) 24817.9 21909.4 28388.2 Imports (c) Experiment El (b) 1991 1993 0.4 3.1 -3.5 10.7 18.7 19.5 -3.2 -2.3 Experiment E2 (b) 1993 1991 -7.3 -0.6 -100.1 -2.1 -5.3 -2.2 -7.7 -0.7 -8.4 -2.1 -10.2 -3 -2.3 -0.7 -5.8 -7.3 809.3 -3.1 120.5 -4.2 -0.6 -0.9 7.3 -9.4 0.2 14.1 1.2 0.9 16.2 18.2 3.8 2.4 -2.3 0.3 0.7 -1.6 -5.7 -46.1 14.6 -23.9 Notes: a. In Real 1990 prices, billions TL. b.Percentage changes over the "historical" values. c. Millions US$. Table 7. Equilibrium in the Public Sector 1990 Revenues from Taxes Indirect taxes Tariffs Income Taxes Corporate Taxes Net Factor Income Transfer Expenditures (c) Public Disposable Income Public Sav-lnv Deficit Net Foreign Borrowing Memo: Public Sav-lnv Def/GNP (% Historical (a) 1991 1993 Exp El (b) 1991 1993 Exp E2 (b) 1991 1993 23601.6 13494.0 24243.4 6686.2 12400.0 19188.2 56827.3 24361.2 13198.5 25568.2 6212.2 2559.1 20797.5 52441.3 25723.8 15541.4 28974.7 6427.3 1575.8 35714.5 44506.8 1.9 4.2 -1.6 0.9 362.4 1.2 18.7 7.2 12.9 -2.2 3.9 745.1 -30.9 19.5 20482.0 -4376.7 32897.0 1326.8 44728.7 1936.3 -24.2 3.5 -26.8 -2.4 -6.3 10.6 -100.0 -100.0 5.2 8.4 10.4 6.3 Notes: a. In Real 1990 prices, billions TL. b.Percentage changes over the "historical" values. c. Includes payments on export subsidies 7.4 0.0 0.4 -0.6 -1.1 -0.6 -1.7 -2.2 8.2 -3.8 7.9 -8.2 -11.9 -6.5 1.2 -5.3 10.5 Table 8. Production, Employment and Private Incomes: Experiments E1 & E2 1990 Real output (a) Agriculture 96529.4 Energy 19966.4 Consumer manuf 108056.6 Producer manuf 127691.0 Construction 51112.2 Services 260953.3 Unemployment rate (%) 6.3 Real Wages (c) Rural 822.3 Urban (d) 9514.2 Private Disposable Income (a) Rural 71932.0 Urban-worker 93605.2 Urban-capitalist 167006.4 Historical 1991 1993 Experiment El (e) 1991 1993 Experiment E2 (e) 1993 1991 99377.7 20436.5 106150.7 116705.8 40716.2 273736.8 11.8 97704.5 24290.5 116168.6 139457.5 488896.3 313790.3 10.3 0.0 -0.1 0.2 1.9 12.8 -0.2 11.8 1.2 2.9 3.3 5.6 28.7 1.0 8.8 0.0 -0.6 0.1 -1.2 -1.9 -0.2 12.4 -0.5 -0.9 13.0 -15.8 -38.1 -2.1 13.1 830.4 12429.4 1022.6 13242.5 -2.4 0.0 -2.6 0.0 -0.7 0.0 -2.8 0.0 70449.5 82515.2 107707.8 125831.4 168350.1 182800.7 -1.9 0.4 -4.2 -3.1 0.6 -5.9 -0.6 -0.2 -1.2 -0.8 -6.1 -11.9 Notes: a. In real 1990 prices, billions TL b. Mining, electricity, gas and water c. Aggregate annual wage payments in real 1990 prices, thousands TL. d. For workers employed in industry and services. e. Percentage changes over the "historical" simulation. administration of an extensive scheme of production and export subsidies and current transfers, and controlling the prices of state enterprises which strategically produce mostly intermediate inputs in an attempt to generate cost-savings to the private capital. Through this process, conducive conditions could have been generated for the private capital to "absorb" the rapid increases in the labor remunerations. As extensively discussed in Yeldan (1995), the fiscal deficit of the Turkish state between 1990 and 1993 does not necessarily imply a problem of "bureaucratic mismanagement" in the abstract; but is a reflection of the administrative and socioeconomic policies on the part of the public sector, which were deemed necessary to sustain the generation of economic surplus for the private capital. The state has used its taxation-cum-subsidy policies and the prices (losses) of its production enterprises as the strategic instruments of this historical maneuver, and financed its fiscal deficits via forced savings by way of price inflation and increased foreign indebtedness. In the following experiment, I take the issue of the effects of foreign borrowing and of the inflows of short term foreign capital on the domestic economy. 111-2. Foreign Borrowing and Short Term Capital Inflows: Experiment E2 Turkey has liberalized its capital account and declared full convertibility of its domestic currency, the Lira, in August 1989. This policy had significant repercussions in the financial economy. For the monetary authority it basically meant the loss of control over the equilibrium relation of the real rate of interest with the foreign exchange rate, as these variables came under the direct scrutiny of the foreign markets, responding to the arbitrage opportunities on an international scale. As discussed extensively in Uygur (1994) and Boratav (1994), for a small developing economy operating under this regime, it becomes imperative that the domestic rate of real interest has a return higher than the return on holding foreign exchange; as this is the only way to prevent currency substitution and dollarization of the domestic economy. Turkey, as well, faced this constraint in its financial sector, and beginning 1990, foreign capital movements were observed to display sensitivity with respect to the differentials on the rate of interest and foreign exchange. For instance according to the calculations in Boratav, Tiirel & Yeldan (1994), in 1990 the rate of return of the domestic interest relative to the foreign exchange was 25%. The foreign capital inflows in 1990 amounted $3 billions. In 1991, the relative return of the interest rate fell to -3.3%, and in the same year the economy witnessed a loss of $3.02 billions. After this 18 year, return on the interest rate successively exceeded the rate of depreciation of the foreign exchange, and led to short term foreign capital inflows amounting to $1.4 billions in 1992, and to $3.8 billions in the first 7 months of 1993. According to the balance of payments data of the same period in 1993, the aggregate amount of foreign capital transferred into the domestic economy reached to $9.3 billions, or to 5.6% of the GNP. Thus, in this period the foreign capital inflows are observed to serve for the dual function of covering the public sector deficits as well as of expanding the import capacity of the domestic economy. The state was able to finance its deficits through borrowing at high interest rates, which, in turn, stimulated the financial (speculative) capital inflows. Yet the costs of this fragile equilibrium were reflected on the sectors producing the traded goods as a consequence of currency appreciation and faltering export demand. As, a version of this "Dutch Disease" phenomenon has spread, imports expanded rapidly, and the ratio of exports to imports fell to 0.516 in 1993, from its 1989 value of 0.736. I investigate the quantitative effects of these developments in the domestic economy under experiment E2. Consequently, the purpose of this experiment can be summarized as the quest to "what would the inflationary consequences and the macro balances be, had the possibilities of foreign capital inflows were restricted in the 1990's?". To this end, the experiment imposes zero net foreign borrowing and foreign transfers to the economy in the 1991-1993 period. All other factors which were regarded exogenous under the historical base-simulation -such as the fiscal deficits, real wage increases, and the assumed technological growth rates, etc.- are kept intact. The analysis of this experiment is based on the reports in Tables 6, 7, and 8 above, and the Figures 9 and 10. The simulated solution path of the inflation rate under the experiment is portrayed in Figure 9. Accordingly, the elimination of foreign capital transfers into the economy has inflationary consequences, increasing the annual level of producer price index by 4 percentage points over its historical realization. In the meantime, as compared to the 1993 levels, the gross national product falls by 7.3%; the ratio of unemployed increases by 3 index points to 13.1%; and the real disposable incomes fall in the public sector by 5.3%, and in the private sector by 7.7%. Thus, under the assumptions of experiment E2, the domestic economy drifts into real contraction and price inflation. 19 [Insert Figure 9 approximately here] Results displayed in Table 8 complement this picture, indicating large scale income In comparison to end of 1993 values, urbanlosses for the private households. capitalist incomes decline by %12; worker incomes by 6%, and the rural incomes by 0.8%. Consequently, aggregate private consumption expenditures are reduced by 7.3%, and the private savings by 9.4%. Thus, given the hypotheses of rapid real wage increases, high fiscal deficits, and the technical specifications of the growth and distribution parameters of the model, we can summarize the findings of the current experiment as follows: inflows of foreign capital, which are attracted by the arbitrage differential between the domestic rate of interest and foreign exchange, have played an important role in sustaining both the public and private disposable incomes in the Turkish economy in the 1990's. Given the above mentioned developments of the period, foreign inflows enabled the domestic economy to overcome the threat of a massive contraction with accelerating inflation. Figure 10 portrays the path of one of the most crucial instruments of this process -the nominal rate of foreign exchange- in comparison to both the historical run and the simulations under experiments El and E2. Elimination of foreign capital inflows (experiment E2) naturally leads to depreciation of the currency in order to equilibriate the balance of payments accounts. Accordingly, the equilibrium nominal exchange rate reaches to 16,450 TL/$ in 1993, as compared to its realized rate of 10,934 TL/$. Per contra, under experiment El, reduction of the fiscal deficits results in deflation of the price level and leads to appreciation of the currency. The El experiment suggests that the 1993 rate of equilibrium foreign exchange be 7,190 TL/$. [Insert Figure 10 approximately here] Not surprisingly, the major consequence of currency appreciation of the post-1989 period was stagnation of the overall export performance, together with the massive expansion of the import volume. As the simulation results of experiment E2 reveal, the loss of export earnings reaches to 14.2%, while the increase in import demand amounts to 23.9% as compared to 1993 values. Thus the foreign liberalization and the consequent capital inflows served to cover the large deficits of the public sector on the one hand, and stimulated the consumption propensities of the private sector, on the other. Yet this short-lived invigoration did not lead to a sustainable expansion of real 20 Figure 9. Domestic Inflation Rate: Experiment E2 450 Experiment E2 Historical 400 350 300 K 0, '0 250 a. 200 150 100 50 0 1990 1991 1992 1993 Years Figure 10. Foreign Exchange Rate: Experiments E1 & E2 Of~A fA IOUUU Experiment E2 16000 14000 12000 Historical Co 10000 Ci 8000 Experiment E1 6000 4000 2000 0 1990 1991 1992 Years 1993 cate mechanism which was based production, and was met h increased imports. This on a fragile balance between the rates of domestic interest and the ioreign exchange collapsed by the end of 1993, leading to financial chaos, and to a severe contraction of the real GNP in the months to follow. III-3. Wage Costs and the Oligopolistic Market Behavior: Experiments 3 & 4 In many popular writings, wage costs are regarded as the main source of price inflation in the Turkish economy. However, technically speaking, the existing data -in spite of its rudimentary nature- is far from providing support for this claim. For instance, based on rough calculations derived from the manufacturing industry survey statistics, manufacturing real wage rate has decreased by 1.8% between 1986 and 1987, yet in the same year, the annual rate of consumer price changes has accelerated from 34.5% to 38.9%; and in 1988 jumped to the plateau of 75.4%. In contrast, in the same The first year, the wage rate continued to fall in real terms at a rate of 7.2%. significant increase in the real wage rate is observed in 1989, with a rate of 25%. However, in the same year, the annual change in the consumer price index loses 6 percentage points and tapers to 69.6%. In the post-1990 period, the inflation rate is observed to "stabilize" at the rate of 60%, yet wage costs surge rapidly to reach in 1992, for example, to a level which is twice their level of 1989 in real values! Thus, candid observations on the quantitative indicators do not seem to lay support to the hypothesis that the main inflationary dynamics of Turkey are to be found with the wage costs. To analyze this hypothesis in a more formal manner, we will study both the cost-push and the demand-creation effects of wage increases in our general equilibrium framework under experiment E3. Under this experiment, real wages are kept constant at their 1990 level, and all other public sector variables are given at their previously assumed levels. However, anticipating the expected cost savings on public employee wage bill, I resort to change the public consumption expenditures item to an endogenous entity along the experiment. With this technical change in the closure rule of the model, I regard aggregate public employment in the public services sector as a given constant; thereby decline in the wage costs reduce the public consumption expenditures on a direct basis. In the previous simulations, the model regarded public consumption expenditures as nominally fixed, and determined the level of employment in the public services sector endogenously, given the wage rate. The question, whether it is the level 21 of public employment or the wage-fund (public consumption) that is prior determined in the public services sector is a technical one open to empirical debate. Here, I propose to follow the first route in an attempt to strengthen the possible links between wage costs, public expenditures and price inflation. The simulated inflation rate under these assumptions is displayed in Figure 11. Maintaining the real wage rate at its 1990 level results in a reduction of the 1993 value of the producer price index from 415 points, to 318 index points, lowering the annual rate of price inflation by 13 percentage points to 47%. It has to be reminded that the current variant of the experiment entails an upper bound on the possible inflation savings, as wage cost reductions are directly matched by expenditure savings in the public sector accounts. Thus, under the other possible extreme (which I do not report here because of space constraints) where the original closure rule of the model was kept intact, and the public services sector had to accommodate the wage reductions by increased public employment (no savings on public consumption expenditures), the experiment led to a decline of only 6 percentage points in the annual inflation rate. [Insert Figure 11 approximately here] Another important issue that needs closer analysis in this respect is related to the role of "returns to the capital-owner" in setting cost dynamics. On discussions about the "cost-push" aspects of price inflation, current debates have generally focused solely on wage-costs of labor, ignoring the pressures emanating form the profit/rent recipients. Yet, the pricing processes based on the imposition of independent mark-ups over average variable costs seem to be a widely-resorted strategy in the Turkish manufacturing. Through this mechanism, the classical trade-off between the wage-profit frontier is effectively surmised as the economic returns to the capitalist are protected by way of mark-ups, with increases in costs being reflected onto the final consumers via price adjustments. Mark-up induced pricing behavior in the Turkish manufacturing industry is studied at length in Sahinkaya (1993), Boratav et. al. (1994), Boratav (1991), Karaaslan (1989), and Ozmucur (1986). Kaytaz at. al. (1993), on the other hand, analyze the noncompetitive structures in the manufacturing sectors directly via measures of "concentration". Further, data on "concentration of the manufacturing industry" has now become part of the Turkish "official" statistics, and is being routinely published by the SIS. (See SIS, 1994b). A general synthesis that emerged from these studies is that, 22 Figure 11. Domestic Inflation: Wage Costs and Oligopolistic Pricing 450 Historical 400 350 Experiment E3 300 Experiment E4 a) 250 -J a) U, h. 200 0. 150 100 50 0 1990 1992 1991 1993 Years Figure 12. Wage Costs and Oligopolistic Profits: Manufacturing Industry 160 Olinnoolistic ffits 140 Wages 120 npetitive 'rofits 100 Q* C 80 60 40 20 0 1990 1991 1992 Years 1993 there exists a fairly stable relationship between the wage share and the mark-up rates in the manufacturing sectors. Boratav et. al. (1994) examine the behavior of industrial mark-ups in the post-1980 trade liberalization era, and conclude that the removal of trade restrictions and foreign liberalization have not yet brought forth the expectations of the theory, prognasticating a fall in the mark-ups. To the contrary, the producers seem to have alleviated the pressures of foreign competition via intensified rent-seeking on various other aspects of macro policy making, as the average mark-up rate in industry is observed to rise secularly after 1985, to reach the plateau of 40% in 1990. (See Boratav et. al., Table IV.3). Kaytaz et. al. (1993) and Yeldan (1995) also suggest disaggregated data supporting those conclusions. Furthermore, it has to be noted that the above arguments pertain mostly to aggregate manufacturing where about a third of value added is produced within the public It is a widely accepted fact that public pricing administration is enterprise system. more often a politically regulated process; and the behavior of mark-ups should be regarded to be mostly under the influence of the private directives. Such oligopolistic pressures on the price dynamics are analyzed through the simulation of experiment E4, and are further to be contrasted with the results obtained under E3. Formally, experiment E4 extends the assumptions of E3 by the additional hypothesis of competitive, marginal-cost pricing behavior throughout the whole economy. These two experiments are discussed jointly through the rest of this section. Figure 11 discloses that under the competitive pricing configuration, the implemented cost savings on real wages generate a further reduction of 6 percentage points in the annual rate of inflation, dampening the rate of growth of producer prices to 41%. In other words, the experiment E4 suggests that the mark-up based oligopolistic pricing strategy in the manufacturing sector has contributed as much as 6 percentage points to Confronted with rising wage-costs, the overall inflation in the Turkish economy. industrialists could have sustained their profits by inflating the final prices of the products they supply and led to a process which I will refer under "profit/rent inflation". This technical analysis is further portrayed in Figure 12, where the evolution of the profit income against the real wage movements is contrasted under the (historical) The figure oligopolistic versus the (hypothetical) competitive pricing environments. plots the historically realized wage rate changes, admitting jumps of 30% and 7.6% per annum in real terms during the post-1990 period. The historical simulation path on profits reveals that "oligopolistic profits" follow the real wage path, maintaining the 23 real rate of return to capital. Per contra, had the cap i:ncreases under a competitive configuration, the profit rz consequent price inflation would be lower. -owner faced the real wage would have fallen and the [Insert Figure 12 approximately here] This process is displayed at the sectoral level in Table 9. Under the "historical" column by 1993, against the the real wage index increases of 139.8, the average manufacturing profit index reaches to 152.4 points. Thus, the model solutions suggest that, under the "historical" simulation, assuming a value of 1.0 in 1990, the profit/wage ratio increases to 1.314 in 1993. Under the oligopolistic conditions of E3, elimination of real wage increases leads the aggregate profit index to reach 158.9 points. However, under the competitive setting of experiment E4, one could expect the index of (competitive) profits to be 138.5, and that the price index be pulled down to 287 index points. [Insert Table 9 approximately here] The macro consequences of the experiments E3 and E4 are displayed in Table 10, and the sectoral production and income generation results are listed in Table 11. One of the most interesting observations of the experiment E3 is that elimination of real wage increases brings forth a 1.3% decline of GNP as compared to its 1993 historical level. Per contra, under the same wage assumptions, but with a competitive pricing environment of E4, gross national product is observed to increase by 1.5% over its 1993 value. Hence, comparison of E3 and E4 suggest that the economic costs of oligopolistic pricing behavior in the Turkish manufacturing are not restricted to profit/rent induced price This outcome is the result two inflation, but entail real output losses, as well. factors: first it is due to the lower capacity utilization rates realized under the oligopolistic environment (See Table 12); second, its lower employment capacity generates lower private incomes as compared to the competitive setting, causing aggregate expenditures to falter. [Insert Tables 10, 11, and 12 approximately here] This process has also significant implications for the public sector disposable 24 Table 9. Real Wages, Profits and Prices in the Manufacturing Industry, Experiments E3 & E4 Real Wages Real Profits Food Processing Textiles Forestry Products (a) Chemicals Petroleum Products Soil Products Metals Machinery Profit/Wage Ratio (b) Domestic Price Level 1990 100.0 Historical 1991 1993 130.7 139.7 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 1.000 100.0 121.4 121.9 79.9 121.7 128.8 107.9 113.9 110.9 1.284 158.1 Experiment E3 1991 1993 100.0 100.0 138.4 139.7 99.2 137.2 150.2 124.6 133.0 133.1 1.314 415.6 Experiment E4 1991 1993 100.0 100.0 141.6 148.9 114.3 137.9 160.4 128.2 138.4 136.3 1.241 318.5 123.5 127.1 85.2 121.7 133.9 111.4 116.9 111.0 1.176 118.9 116.5 149.5 126.9 87.0 329.6 123.7 147.8 77.7 1.215 287.0 105.8 128.6 90.3 74.7 121.0 59.8 79.1 30.2 0.902 113.2 Notes: a. Competitive pricing is assumed under this sector, b. Ratio of aggregate profits to aggregate wage costs. Table 10. Macro Economic Indicators: Experiments E3 & E4 (Percentage Changes over the Historical Solution Values) GNP Foreign Deficit Public Disp Income Private Disp Income Aggregate Consumption Public Private Domestic Savings Public Private Aggregate Investment Public Private Exports Imports Experiment E3 1991 1992 1993 -0.9 -1.2 -1.4 -4.2 6.2 7.1 -9.5 -12.5 -18.1 0.4 0.4 0.6 -9.8 -0.4 -11.6 -0.6 -13.9 -0.3 -0.9 3.3 1.2 4.2 1.4 4.1 32.9 -22.0 2.6 1.5 24.7 -17.8 3.7 2.2 30.5 -6.5 3.9 2.2 Experiment E4 1991 1992 0.2 -0.6 -13.2 10.2 -7.8 -10.1 1.6 0.8 -8.2 1.1 -9.4 0.0 0.2 3.6 0.6 4.0 39.6 -27.3 7.3 4.2 51.4 -23.1 6.5 3.8 1993 1.5 13 -14.4 3.3 -10.4 2.3 -1.4 6.9 44.7 -8.2 8.7 4.6 Table 11. Production, Employment and Private Incomes: Experiments E3 & E4 (Percentage Changes over the Historical Solution Values) Experiment E3 1992 1993 19911 Real Output Agriculture Energy Consumer manufacturin Producer manufacturing Construction Services Unemployment rate (%) Real Wages Rural Urban Private Disposable Income Rural Urban-Worker Urban-Capitalist 0.(0 7.;2 4.1 1 4.43 10.(3 1 -0.1 7.f3 0.0 11.0 5.6 6.2 13.8 7.5 5.9 0.1 Experiment E4 1991 1992 5.5 6.3 12.5 0.2 10.5 11.1 15.0 16.2 11.8 1.1 12.8 10.8 1993 8.0 14.6 11.5 15.1 11.9 18.8 2.0 4.0 1.8 1.0 4.7 6.9 18.9 1.6 5.4 1. 3 -30.U7 2.4 -7.6 2.6 1.0 7.5 -30.7 4.9 -7.6 6.6 1.0 5 2.%r -13.13 8.f5 3.3 -17.3 3.5 -15.9 10.6 8.4 -11.8 7.4 5.7 -16.0 10.0 6.0 -13.6 13.7 11.1 Table 12. Capacity Utilization in the Manufacturing Industries (%) Food Processing Textiles, Clothing Forestry Products (*) Chemicals Petroleum Products Soil Products Metals Machinery Man Industry Average Historical 1991 1993 89.7 95.4 90.6 87.9 100.0 100.0 90.3 89.8 95.6 99.4 85.9 83.1 89.3 92.3 86.1 88.9 91.7 91.3 Experiment 1991 95.3 91.7 100.0 89.8 94.7 85.9 90.6 88.0 92.0 El 1993 89.8 91.5 100.0 88.6 98.3 93.3 95.6 92.7 93.7 Experiment 1991 95.6 91.5 100.0 89.3 95.4 82.5 90.6 86.1 91.3 E2 1993 91.1 100.0 100.0 84.4 90.7 77.4 86.4 79.8 88.7 Experiment 1991 93.3 92.1 100.0 90.8 98.0 85.4 90.2 86.0 91.9 Note: * Due to the high share of public ownership with administered price controls, the forestry products sector is not subjected to mark-up pricing. E3 1993 88.4 92.1 100.0 91.9 100.0 91.2 95.9 91.9 93.9 income generation. With real wage increases eliminated, worker incomes are reduced, thereby reducing the tax revenue base of the state. In the meantime, reduced wage incomes mean reduced aggregate expenditures, causing a further fall on the state's sources of excise taxes. Consequently, compared to 1993, public disposable income falls by 18.1% under experiment E3, and by 14.4% under E4. This result reveals one important attribute of wage-labor: even though it is a cost item on the production side, wage is also a "revenue" item on the expenditure side. The model solutions underscore the importance of this latter function of wage earnings in sustaining aggregate demand in the economy, and further maintain that, under conditions of oligopolistic markets, its effectiveness is limited. IV. Summary and Conclusions In this study I have investigated the post-1990 evolution of the Turkish economy within a structuralist Walrasian macro-economic model which emphasizes the underlying processes of production - income distribution - consumption and foreign trade in a general equilibrium framework. The model is first used to simulate the historically realized growth path of the Turkish economy over the period 1990-1993; and then it is utilized as a "social laboratory" to analyze a series of "counter-factual" policy environments so as to dis-integrate the underlying mechanisms behind the 1994 crisis. The counter-factual policy simulations of the model reveal the importance of the public sector fiscal deficits as a major factor generating inflationary dynamics in the Turkish economy in the analyzed period. Accordingly, under the parametric assumptions which bring the fiscal balances of the state to their 1980's average -with more modest expenditure patterns and lower PSBR to GNP ratios- the model solutions suggest the annual rate of inflation to fall from 61%, to 35%; and that the real national product to increase by 3.1% over its 1993 value. However, the issue of whether the foregoing technical assumptions of the model could have been realized in the Turkish economic setting of the 1990's remains a dubious question, open to debate. For, in this regard the model solutions underscore the crucial function of the state in sustaining and re-distributing capitalist incomes in response to the intensified conflicting claims on national income through the 1990's, which culminated into rapid real increases in remunerations of the wage labor, and 25 At the micro level, the state is faltering output production and declining exports. observed to interfere the domestic output markets through an extensive scheme of production and export subsidies and current transfers; and through effective price controls on the state enterprises which strategically produce mostly intermediate inputs in an attempt to generate cost-savings to the private capital. Through this process, conducive conditions could have been generated for the private capital to "absorb" the rapid increases in labor costs. However, this mechanism meant significant losses for the public sector and generated strong stagflationary pressures for the domestic economy, raising the domestic price inflation from its possible annual rate of 35% to 60%, and leading to a loss of aggregate investment growth rate of 17.8%, and export revenues of 3.8%. The role of foreign capital inflows was the second issue analyzed with the aid of With full capital account liberalization in 1989, Turkey received the model. significant foreign inflows of short term capital, which were attracted by the arbitrage differential between high domestic real interest rates and the revalued foreign exchange rate. Under the experiment in which this process was reversed and the foreign borrowing was set to zero in net terms, the domestic economy was observed to suffer real contraction, with an accelerating inflation rate. GNP declined by 7.3%, and the urbancapitalist and worker incomes fell by 12% and 6%, respectively. Inflation rate rose by additional 4 percentage points per annum. Hence, in retrospect it can be argued that, the short term foreign capital inflows served to postpone the evolving economic crisis by providing timely financial resources to the public sector to (partially) cover its growing deficits, and by stimulating the private consumption demand, it generated short term sparks in the growth rate of the economy which meant mini boom and bust cycles over the period. Yet, the fragile balance upon which this mechanism rested was clearly not sustainable, and its collapse in the end of 1993 meant a severe contraction in output and financial chaos. The experiment which analyzes the effects of real wage increases on the price inflation reveal that by keeping the real wages at their 1990 level, the reduction in the inflation rate can be expected to lie between 6% and 13%. This would depend on whether wage reductions will have direct cost savings on the public expenditures. However, these abstract results of the experiment do not take account of the pressures emanating from the oligopolistic pricing manufacturing industry. behavior argued to be prevalent in the Complementing the assumption of wage reductions with the competitive marginal cost pricing behavior, the model solutions suggest that the costs 26 of mark-up induced profit/rent inflation in the Turkish economy entail additional price inflation of 6% per annum, and a loss of 1.5% in the real gross national product. Thus, oligopolistic pricing power in manufacturing has both inflationary price and also contractionary output consequences for the aggregate domestic economy. In conclusion, based on the political economy perspective of class conflict and political pressure, the current model underscores that conflicting claims of various social classes on national output have to be regarded as important sources of disequilibria in the domestic economy; and that the erosion of fiscal balances of the public sector is but a mere reflection of the state's intervention in the factor and the product markets in an attempt to resolve the distributional conflict in favor of capital, during a period of real wage increases and faltering production and trade performance. 27 Notes 1) For further detailed descriptive analysis of the evolution of the crisis, see Yeldan (1995), Boratav, Tiirel & Yeldan (1994) and/or Uygur (1994). 2) For modeling exercises which link the financial and the real spheres of the macro economy in the CGE context, see Yeldan (forthcoming), and Sak & Yeldan (1993). Taylor (1991) also documents a series of financial-linked CGE models in the structuralist tradition. 3) In the absence of a full fledged financial system recognized by the model, I -namely foreign borrowing and had to regard these last two magnitudes short term capital inflows- as exogenously given at their historical realizations in the simulation experiments. However, their effects on the exchange rate remain non-neutral through the balance of payments constraint. 4) In this part of the historical validation exercise, the major difficulty was to overcome the discrepancy between the two officially reported GNP series for the Turkish economy. Recently, the State of statistics changed its method of estimation of the GNP from the "production" side, and has re-created a "new" serie which exceeded the aggregate GNP serie reported by the State Planning Organization (SPO) by as much 25% on the average. The SPO's serie is calculated from the "expenditure" side, and has the advantage of reporting the macro expenditures on national product. In the popular Turkish press the SIS serie is referred to as the "new GNP", whereas the SPO's is known as the "old serie". The data utilized in the model reflects the SIS's GNP values calculated from the production side. To arrive at the macro composition of expenditures, the SPO's ratios are imposed on the SIS's GNP aggregates. Thus, the observed divergences on certain expenditure items in Table 2 are mostly reflections of the shortcomings related with this calculation. This approach is obviously not all too satisfactory, but until the "new" serie is complemented by further data on the macro expenditure side, there does not seem to be any other alternative. 28 ferences Boratav, K. 1994. "Iktisadi Kriz Uzerine Bazi Gozlemler" Iktisat, Isletme ve Finans, 9(100), Temmuz: 22-30. _. Gercek Yay. 1991. 1980'li Yillarda Turkiye'de Sosyal Smiflar ve Bliisiim. Istanbul: _., O. Tiirel & E. Yeldan (1994) "The Turkish Economy Towards the Next Milennium: A Balance Sheet, Problems and Prospects" Project Report prepared for UNU/Wider, mimeo. Bourguignon, F., de Melo, J. & A. Suwa (1991) "Modeling the Effects of Adjustment Programs on Income Distribution" World Development, 19(11): 1527-1544. Celasun, M. 1986. "A General Equilibrium Model Of The Turkish Economy, SIMLOG-1" METU Studies in Development, 13 (1,2):29-64. Dervis K., J. de Melo & S. Robinson. 1982. General Equilibrium Models for Development Policy, London: Cambridge U. Press Istanbul Chamber of Industry (ISO). Istanbul. 1993. 150 Biiyiik Ozel Isletme, Yay. No.1993/3, Karaaslan, M. 1989. "Pricing in the Turkish Non-Agricultural Sector" Journal of Contemporary Management, No.3, March:42-57 Kaytaz, M., S. Altm & M. Gines. 1993. "Tiirkiye Imalat Sanayinde Yogunlasma: 1990" TMMOB,1993 Sanayi Kongresi Bildiriler Kitabi, 1. Cilt, Yay. No.160. Kose, A. & A. E. Yeldan (1994) "Cok Sekt6rlii Hesaplanabilir Genel Denge Modellerinin Veri Tabani Uzerine Notlar: Tiirkiye 1990 Sosyal Muhasebe Matrisi", Ankara, mimeo. _1991b. "Prices and Income Distribution" Bogazigi University Research Papers, No. ISS/Ec 91-17. Sak, G. & E. Yeldan (1993) "Reflections on Asset-Backed Securitization in Turkey: Results of a Financial CGE Model" METU Studies in Development, 2o(3): 325-356. Sahinkaya, S. 1993. "Imalat Sanayiinde Sekt6rel Isgiicii Verimliligi, Reel Ucretler ve Gayn Safi Karlar veya Mark-up Oranlan" Toplum ve Bilim, No.4, Nisan:27-51. State Institute of Statistics (SIS) Statistical Bulletin, Turkey, various years, Ankara. _____________________ Ankara. Manufacturing Industry Surveys, various years, 29 Ecoo____my_______(1994a) Input-Output Structure Economy, 1990. Ankara. of the Turkish Ind_____________(1994b) Concentration of the Turkish Manufacturing Industry, 1985-1989. Ankara. Taylor, L. (1981) Structuralist Macroeconomics, NewYork: Basic Books Taylor, L. 1991. Socially Relevant Policy Analysis: Structuralist CGE Models for the Developing World, Massachusettes: MIT Press. Uygur, E. 1994. "Tiirkiye'de Ekonomik Kriz: Olusumu, Seyri ve Gelecegi" Iktisat, Isletme ve Finans, 9(100):42-54. Yeldan, A. E. (forthcoming) "Financial Liberalization and Fiscal Repression in Turkey: Results of a Financial CGE Model" Journal of Policy Modeling. . (1995) "Surplus Creation and Extraction under Structural Adjustment: Turkey, 1980-1992" Review of Radical Political Economics, in press. 30 Appendix: Documentation of the CGE Model This Table gives a formal presentation of the CGE model used in the paper. The documentation adheres to the following legend: upper case letters without a bar are endogenous variables; those with a bar denote variables that are treated exogenous. Lower case and Greek letters refer to policy variables or structural parameters. Subscripts i and j refer to sectors, k refers to labor types and h to household categories. Nonlinear relations are not explicitly spelled out, as their forms are part of the traditional CGE folklore. The C(.) and d(.) refers to the composite and CET functions; and f(.) refers to CES production functions. Prices PM. = PWMI.-(1 + tm.)*ER PE = PWE..(1 + te.)-ER PC. = PDi-(DCi/CCi) + PMi (Mi/CC i) PX. = PD.* (DC./XS.) + PE. (Ei/CCi) PVA. = PX (1 - tn)S. a.ji -PC J PK = TPC-b.. SJ PLEV = E PC. 1 1i 31 endogenous variables PM. : domestic price of import good PE : domestic price of export good PC. : composite good price PX. :producer price PVA : value added (net) price PK. : capital good price PD. : domestic good price PLEV : aggregate price level ER : exchange rate (TL/US $) exogenous variables and parameters tm. : tariff rate on imports te. : export subsidy rate tn : indirect tax rate a. : Leontieff input-output coefficient : capital composition coefficient b. (i : price index weights PWM. : world price of import PWE. : world price of export Output supply and the Factor markets PX. = (1 + i = Ti)-AVC i XS i = U.ifi(K, Lk) ) = Wik PVA.*i (aXSi/aL i k UNEMPk =E - L k 32 manufacturing sectors RP.i = PVA-XS i AVC. i= L W kLLD k ik ik W-LD + .a.PC.XS Wik ik j k j endogenous variables XS. output supply Ui capacity utilization rate (i= manufacturing) D L ik labor demand UNEMPk : unemployed labor RP : aggregate sectoral profits AVC.1 : average variable costs Exogenous variables and parameters Ti :price mark-up (i= manufacturing) K : capital stock nominal wage rate Wik Ek : labor supply Income generation and saving-investment balance Y"R = PVA -XS A A + RR -I YH = W W. LD + (1- -ERE+ TR A -].ER MR) + TR i Y" = RP. - FKTRYG + TR i i>A K 33 GREV = TARIFE + DOLVER + HHVER + KAPVER + FKTRYG + GBOR-ER - EXSUI TARIFE = t tm.I -PWM. *MER I I DOLVER = HHVER = KAPVER = tn.1Ii PX.-iXS l l Lj htx h hYh ctx - RPi i>A FKTRYG = rfg*GDP EXSUB = te.*PWE.*E. ER GREV GSAV + GREV = GSAV + GDTOT HHSAV = h GSAV = h Y.H (1 -htxh)h h h -YGDP GSIDEF = GIF - GSAV = HHSAV - PRINV + FSAV-ER 34 endogenous variables : private household incomes : public disposable income GREV TARIFE : aggregate import tariff revenue DOLVER: aggregate indirect tax revenue HHVER : aggregate income tax revenue KAPVER: corporate income tax revenue YH R,W,K FKTRYG : net factor income of the public sector EXSUB HHSAV GSAV GSIDEF PRINV GDP : aggregate export subsidies paid : aggregate private savings : aggregate oublic savings : public sector saving-investment deficit : aggregate private investment , gross domestic product exogenous variables and parameters MR : share of rural household in workers' remittance inflows htxh : income tax rate rh : private household saving rate : public saving rate : public investment rate M : workers' remittances TRh : current transfers GDTOT: aggregate public consumption exoenditures GIF : aggregate public investments GBOR : net public borrowing from abroad ctx : corporate income tax rate rfg : ratio of public sector net factor income to the GDP FSAV : foreign savings (deficit) 35 Final demand and product market equilibrium I I I XD. = di(DCi,E) CC. = c.(DCiM) = CD.+ GD.+ V.+ ID min P i M. + P? DC i 1 subject to CC i = c(DC,M) i i max PE iE + P? Dc 1 i subject to XD. = d.(DC,E.) PWEOi II E-EOi( PWEi V. = aij XS J CD = (cdi/PCi) [ Y-( 1 - htxh)l ( ) h GD = (gdi/PCi) GDTOT GD pserv = W ur (i r public services) LD pserv DKP = (kp./PK.) PRINV 36 DKG = (kg./PK) GIF ID. = b.. (DKP + DKG) XS I = XD.I CC = CD.+ GD.+ V.+ ID PW• i-M = 1 1 I 1 1 PWEi.E i + RM + GEOR + FSAV endogenous variables DC. : domestic good E : Sectoral exports M. Sectoral imports CC. : Armingtongan composite good CDi : Sectoral private consumption GD. : Sectoral public consumption V intermediate input demand DKP. : private investment demand by destination DKG. public investment demand by destination ID. : aggregate investment demand XD. : aggregate domestic demand PWE : world price of export good exogenous cd. gd. kpi kg. variables and parameters : private consumption shares : public consumption shares : private investment by destination share : public investment by destination share PWEO.: world price of competitive exports ?. •:export demand elasticity 37 4- C 00 .2 o 0 NM CnD Sct 0. D So CO tO oCN T- >% 00 j ( a CMN CN qt N C0 r- LO CM m iO r- M 4t T- M m r*- oo o I 'M to 0) CO ) N-Co CM CO 00 M-oo0 l cn T- 0 CNo CO <qo Tt M CoV CD cn 0 0 C1 Co CD 0m +C O E_ o)I CL M orT co o) o 0) C5 I^ co .qtco h( 0Om 0 19t ltt\ e (0 LO Q^) 0 Y) 0 < 0m - cN. r-- . eo. OCCO lO CO ^ 0 CO lo r* r^^t ~ 0 LO C C <D t oo00 P,-0) r- C ^6oo(o) o b^ =o + co C6 Lo; CO <o) q0 co it o 6Cdc .0co q-t . a .eo. c . o. Lr6 m) O) n CO id . LO CV) CO 0. Co D co c 00 m G ) 'S 00^L r-, C CN LO Vi CO I W.^ EO T -D cn 0 oo r* r*-cn 0) r4%) CO CY) CO CO CO C tC O I% ff^ t- cor So OC D r- o M 0 0 3om - CLIO0L CmO 00 LO C0)r-1 C CO CN c OlO ( 5DM 00 o I I O CN CD O cn CO C C4 eg - Cm N I c - c6') c CO M C)NON CM avs h U) - CM c) - LO Co (D 0o V) *s rc lcN O C c0 I. c. CL 0 It i Cd r00) CN C mom r'- 25 00 or). o. c . od- O 2+ O r* r . 0 CD O ho - - -r-----NCO lC a r^ C r S-rýrs r-.oc 6s dc oi ai c6 Lo ai co I ) 0C r% co 0) E m CY) M 00 oo*co r C)o co) c o0)N 0 - 3-I N qe N cn u)eodo 000Lo cN oDeo)r FcN V-00 00 r- qt rt CO o 0 U C) r%L + LO0) a CY) >zt- d m aiI eme wod1 ol~ll~ e w wom Ic OIOI oInIo IC, 01~IOIL C CYaNCMi(N(NI <r; C 0 CCO5 L m E c 0 0§ 0 0 a.4-.- S 0CLo0 4 -1 , E Uc-..o >' - -a ca <MLL U.I Ccn UI-a. CD -- o C) (L 0, (D O( CO. ( s 0 o 0 x LU 0 00000000000000 o c E S0000000000 .2 ooooooooooooooo Q w E a o- M M C C IIn CO -r----ddddd-oddd 00 cr^<cu i 000 cMO <0 <0 0 0 C50 _ __ __ o5§ * ooooooooooooooo ooooooooooooooo =5 i- ddddddddo 0e00 0000006 0 I- Sx --- O 2 c Su 0 L0 0 00000 6 6 00 00000 • o0 .- £ M -M -^ OD 0 tITw 4- .... C C ^ N O N °. O 0OO | | OOO0 ^ ^^ ^^ij-L M 60 o -*---- o COO r *CO0ors LC)T-0 qt CO l i CY0)C a)0)NOD r .C 'cf .nSgEo Co O0) i felig~~lul ~llls!-^l <_O^Zji^ __ 0 om~00000 om . 0 0 CO) dddd 0 c)o - a-- I000011 Vv N C 0oCm 00 M T LL I C c rC 0: I co Jl I aft ---C rCC Ca d 0 C( CN C oe (Q I... o cc C) c f% ooq r*-- cm I,. C 0 0 I (in -- ti --· (b ·~qt~: 4r0 II CD ui U) in 0) C0 -- ul O 0 O I- LO LO 0) II cD II 0 0 Co 0, rz C, C 01 9 n I- Co cb L CN ~tf~ r 0 oJ LO -- s 04 4 --- -- CO )V i 0 Co c 0 0 0 0) Cv) o E E 0 rC im E I (C) N co 0) N V) C4 I I 0 0 I- D N 0 D N U) § U- ui CD L 0) .in 00 Li) to CO - CO -f O I§' tos i a' ICU'(= U) '0 in' Go 0. U) (0 0D U) 0 U) E I,5 E o 0 - CD Go In I0 m (0 1_ <r N 8 n N 0 C Go r CO U 1m -> (0 V) ,D i, =g I- Cd oCO 'U U) 0 16 0 0 a. U) 0I - N a' 0) ~in (U C 0 0 (U M0 U) «r (Q Q, a' I Q E 5 (U (U a' 0~ a' /) 0 (U 'U I C) 0 o w5 c e)0 eAco Sc e C C o0 00 00 N 0o 00 ) r%: 10 ocVLn 00Oco q-o0 'iCo 0 cV oo(De 0E) D O00o i4 66i Cd i4 CL (6 m P, --.. o) OC \JU O DCE)OCD m S00 1o0 o CV) (D aNo M 10 (6 f%:(6 Vo cD o 0 0o C ) C NCe) C C6d06 6d roc)0NIV Dof) oort-c P- c) o *E) T-0 CrooO C v r-*-- ONvcOCE 0)0) 0) 04 V) Cu) cN U) 0m o C 00M Sc Oo ( C CO cV) 0NNG 0 o0 CM NC Ct h G Go 0 0 a 00 w00 0 0 C 0 CC o 4 lifts co o >5 * OO mi (0)0 '- a COC 0 Go 00 GoO e ON 00 0oco »E) CdO > ' CLoE 00 - 0 0 0) 0 P- Q c '- 0 CO c )- 'CM o 0 t C o co '2- U) 00 U N- ) N U) C 0 0 c0) O)i tI U 'C CO O 0 ... r- a CO I m CE s i o I I I I I i I oII RECENT BULLETINS 89-5 McGuire, Mark F. and Vernon W. Ruttan, "Lost Directions: U.S. Foreign Assistance Policy Since New Directions," August. 89-6 Coggins, Jay, "On the Welfare Consequences of Political Activity, "August. 89-7 Ramaswami, Bharat and Terry Roe, "Incompleteness in Insurance: An Analysis of the Multiplicative Case," September. 89-8 Rosenzweig, Mark and Kenneth Wolpin, "Credit Market Constraints, Consumption Smoothing and the Accumulation of Durable Production Assets in Low-Income Countries: Investments in Bullocks in India," September. 89-9 Pitt, Mark and Mark Rosenzweig, "The Selectivity of Fertility and the Determinants of Human Capital Investments: Parametric and Semi-Parametric Estimates," October. 89-10 Ruttan, Vernon, "What Happened to Political Development," November. 90-1 Falconi, Cesar and Terry Roe, "Economics of Food Safety: Risk, Information, and the Demand and Supply of Health," July. 90-2 Roe, Terry and Theodore Graham-Tomasi, "Competition Among Rent Seeking Groups in General Equilibrium," September. 91-1 Mohtadi, Hamid and Terry Roe, "Political Economy of Endogenous Growth," January. 91-2 Ruttan, Vernon W., "The Future of U.S. Foreign Economic Assistance," February. 92-1 Kim, Sunwoong and Hamid Mohtadi, "Education, Job Signaling, and Dual Labor Markets in IDedig Countries," January. 92-2 Mohtadi, Hamid and Sunwoong Kim, "Labor Specialization and Endogenous Growth," January. 92-3 Roe, Terry, "Political Economy of Structural Adjustment: A General Equilibrium - Interest Group Perspective." April 1992. 92-4 Mohtadi, Hamid and Terry Roe, "Endogenous Growth, Health and the Environment." July 1992. 93-1 Hayami, Yujiro and Vernon W. Ruttan, "Induced Technical and Institutional Change Evaluation and Reassessment." February 1993. 93-2 Guyomard, Hervd and Louis Pascal Mahe, "Producer Behaviour Under Strict Rationing and QuasiFixed Factors." September 1993. 94-1 Tsur, Yacov and Amos Zemel, "Endangered Species and Natural Resource Exploitation: Extinction Vs. Coexistence." May 1994. 94-2 Smale, Melinda and Vernon W. Ruttan, "Cultural Endowments, Institutional Renovation and Technical Innovation: The Groupements Naam of Yatenga, Burkina Faso." July 1994 94-3 Roumasset, James, "Explaining Diversity in Agricultural Organization: An Agency Perspective." August 1994. 95-1 Elbasha, Elamin H. and Terry L. Roe, "On Endogenous Growth: The Implications of Environmental Externalities." February 1995. 95-2 Roe, Terry and Xinshen Diao, "The Strategic Interdependence of a Shared Water Aquifer: A General Equilibrium Analysis." March 1995