Turkey : recent economic performance and medium-term prospects, 1978-1990

FILE COPY SWP6O2 Turkey Recent Economic Performance and Medium-term Prospects, 1978-1990 Jeffrey D. Lewis Shujiro Urata WORLD BANK STAFF WORKING PAPERS Number 602   WORLD BANK STAFF WORKING PAPERS Number 602 t7o2;o c~. Turkey Recent Economic Performance and Medium-term Prospects, 1978-1990 Jeffrey D. Lewis Shujiro Urata T,RNAInONAL MONETARY FUND JOINT LWREAY MAR 2 7 1984 UITEWATIONAL BANK FOR RECONSTRUCTION AND DEVELOPMENT WASHINGTON, D.C. 20431 The World Bank Washington, D.C., U.S.A.  Copyright g 1983 The International Bank for Reconstruction and Development / THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. First printing August 1983 All rights reserved Manufactured in the United States of America This is a working document published informally by the World Bank. To present the results of research with the least possible delay, the typescript has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. The publication is supplied at a token charge to defray part of the cost of manufacture and distribution. The views and interpretations in this document are those of the author(s) and should not be attributed to the World Bank, to its affiliated organizations, or to any individual acting on their behalf. Any maps used have been prepared solely for the convenience of the readers; the denominations used and the boundaries shown do not imply, on the part of the World Bank and its affiliates, any judgment on the legal status of any territory or any endorsement or acceptance of such boundaries. The full range of World Bank publications is described in the Catalog of World Bank Publications; the continuing research program of the Bank is outlined in World Bank Research Program: Abstracts of Current Studies. Both booklets are updated annually; the most recent edition of each is available without charge from the Publications Distribution Unit of the Bank in Washington or from the European Office of the Bank, 66, avenue d'Iena, 75116 Paris, France. Jeffrey D. Lewis is a research assistant and Shujiro Urata an economist in the Development Strategy Division of the World Bank's Development Research Department. Library of Congress Cataloging in Publication Data Lewis, Jeffrey D., 1954- Turkey','recent economic performance and medium-term prospects, 1978-1990. (World Bank staff working papers ; no. 602) Bibliography: p.' 1. Turkey-Economic policy-Mathematical models. 2. Foreign exchange problem-Turkey--Mathematical models. I. Urata, Shuj'iro, 1950- . II. Title. III. Series. HC492.L48 1983 338.5'443'0724 83-10263 ISBN 0-8213-0210-8  Abs tract This paper analyzes the 1978-81 foreign exchange crisis and 1982-90 medium- term prospects for Turkey using a computable general equilibrium (CGE) model of the Turkish economy. Through a series of historical simulations, the model attempts to capture the basic features of the 1978-81 foreign exchange disequilibrium in Turkey and evaluate their role in the crisis. The model is then used to undertake a series of forward runs for the 1982-90 period which reflect alternative scenarios as to the success and viability of the ongoing shift in Turkish development policy from inward-looking import substitution to outward-focused export expansion.  Acknowledgments The modeling effort that this paper represents was carried out in close collaboration with the Turkey Division of the World Bank as part of the work of the economic mission sent to Turkey in June, 1982. Seok Hyun Hong contributed substantially to the model construction and data collection efforts; the assistance and suggestions of Firouz Vakil and Fan-Fan Walker are also acknowledged. The authors wish to thank Sherman Robinson for his suggestions and guidance throughout the exercise, and Bela Belassa and Jayanta Roy for their comments on an earlier draft of this paper. Merih Celasum, Sibel Guven, and Tevfik Yaprak of Middle East Technical University in Ankara provided data and helpful discussion during our stay in Turkey. Finally, heartfelt appreciation is expressed to Kim Tran for her typing and retyping of various drafts, and to Isabelle Kim for her work on earlier versions. We of course accept full responsibility for all remaining errors.  Table of Contents Pe 1. Introduction ..................l 2. An Overview of the CGE Model................................ 4 3. The Turkish Economy: A Historical Analysis, 1978-81 7..........7 A. Turkish Economic Performance, 1973-77: A Perspective. ......7 B. Comparison with Other Middle Income Countries..........9.12 C. Recent Economic Events and Policy Reforms, 1978-81 ........17 D. Validation of the CGE Model, 1978-81 ... .......24 E. Analysis of CGE Model Results.........e..............o. .29 F. Sources Decomposition of Output Growth, 1958-81 ...........32 G. Analysis of the Foreign Exchange Crisis: Counterfactual Simulations .................................. *36 4. Analysis of Turkish Medium-Term Prospects, 1982-90 ........... 44 A. The Medium-Term Outlook: Basic Assumptions..e.e...e.e.e.9 45 B. The Base Run: Successful Export Expansion..*.*9.**..**e*48 C. Alternative Medium-Term Scenarios........ ............54 D. Macroeconomic Projections. .. .. .. ... .. ..... .. .e.e.58 E. Sectoral Implications of Alternative Trade Scenarios.eeteo*65 F. Sources Decomposition of Output Growth, 1981-90.... .s...77 References ..*.... ,8 Appendix I: The Turkey CGEModel..................................A-1 A. The OGE Model Formulation .. . . ..............e ........... . . .A-1 B. Modifications in the CGE Model ..........................A-10 C. Mathematical Presentation of the CGE Model...............A-20 Appendix II: The Data in the Turkey CGE Model .................... A-30   1. Introduction By 1978, Turkey was in the midst of the most serious foreign exchange and economic crisis it had experienced in three decades. Although the Turkish economy appeared to weather the first oil price increase and related international disruptions of the early 1970s with relative ease, the high (over 7 percent) GDP growth rate sustained over the 1973-77 period was largely attributable to increased foreign borrowing and favorable trends in workers' remittances. With the onset of the second oil price increase in 1978 and 1979, the persistence of stagflation in Turkey's major trading partners, and the growing strains in the Turkish economy (characterized by exchange rate overvaluation and accelerating domestic inflation), the growth momentum of the mid-1970s collapsed and Turkey entered a period of negative GDP growth and severe foreign exchange rationing. This stagnation continued until 1981 when the remarkable 62 percent growth of exports (in dollars) alleviated the foreign exchange shortage and permitted GDP growth of around 4 percent, which has continued into 1982. While the export-led economic recovery is attributable in part to the depressed level of domestic demand, another major factor is the government's reorientation towards an outward-oriented development strategy, as embodied in the exchange rate reforms, import liberalization and export subsidy programs implemented since early 1980. This paper presents an analysis of the 1978-80 crisis period in Turkey and an exploration of the feasibility and sustainability of the ongoing export-led recovery. We make use of the results derived from the latest variant of computable general equilibrium (CGE) models that have increasingly been applied to issues of policy choice and development strategy in developing countries, including earlier applications to Turkey (see Dervis and Robinson (1978) and (1982)) and Yugoslavia.  -2- Section 2 of the paper provides a brief introduction to the features of the CGE model, which is supplemented by a more detailed description and algebraic presentation in Appendix I. The current version of the CGE model incorporates a number of features that facilitate an historical examination of the turbulent foreign exchange disequilibria of the 1978-80 period. Turkish authorities responded to the severe foreign exchange shortage by rationing imports, resorting to a varying mix of quantitative controls and premia schemes to restrict imports. The CGE model simulates the simultaneous operation of both types of schemes, as well as producing measures of the total "rents" generated by such rationing mechanisms and their impact on productivity and capacity utilization. Section 3 presents an analysis of the 1978-81 historical period. The historical antecedents to the recent crisis are described and the macroeconomic characteristics of Turkey are contrasted with those of a sample of twelve other middle-income countries (MICs). Turkey is found to differ substantially from the sample average with regards both to the low importance of exports and imports in GDP (i.e. the degree of "openness") and to the sizeable share of agriculture in both GDP and employment. The economic developments of the 1978-81 period are described in some detail, and the features of the shift in development strategy that was first articulated in the 1980 trade reforms are outlined. Comparison of the sources of output growth for 1978-81 with earlier periods reveals strong discontinuities in the import policies adopted by the government, with periods of import liberalization alternating with foreign exchange shortage and import restriction. The quantitative analysis is based on the fitting of the CGE model to the 1978-81 period (or "validation"), and the subsequent simulation of a number of counterfactual "alternative" scenarios designed to disentangle  -3- the role of exogenous shocks and domestic policies in the economic performance of the period. Model simulations reveal that the "cost" to the economy of delay in the adoption of the new outward-looking strategy was substantial, and that the role of the oil price increase, while significant, was less important than changes in the real exchange rate due to high domestic inflation in bringing about the foreign exchange crisis. Turning from historical analysis to medium-term projections, Section 4 of the paper presents a number of alternative scenarios for growth in the Turkish economy during the 1982-90 period. Based on the phenomenal growth of exports during 1981 and 1982, and the high net inflow of foreign capital, Turkey enters 1983 in a better situation than at any time since the early 1970s. Using the CGE model,. this section develops a Forward Base Run that assumes that the government adopts trade and domestic management policies that will permit the ongoing shift to export-led growth to be successful, with exports reaching nearly 20 percent of GDP by 1990, and that facilitate the rapid growth of private investment, resulting not only in a drastic reduction in the role of public investment but an overall increase in the share of investment in GDP as well. This Base Run is then contrasted with a number of alternative forward scenarios, with special focus placed on the role of export performance, productivity growth, exchange rate policy, and private investment as critical determinants of economic performance in the coming decade. The macroeconomic implications of the various scenarios are considered, as well as their impact on sectoral structure. Finally, the feasibility of the Turkish export orientation is considered by using the "demand sources of growth" methodology to contrast the Turkish projections with the experience of other countries pursuing export-led strategies. Appendix I contains a detailed description and' algebraic summary of the model, while Appendix II outlines the data sources used for the study.  - 4- 2. An Overview of the CGE Model I/ The current CGE model of Turkey is the latest in a long tradition of non-linear, computable general equilibrium models that stem from the early work of Johansen (1960). In recent years, this class of models has been used as instruments of planning and policy analysis in conjunction with World Bank research and evaluation of the performance and prospects of developing nations, including Bank Missions to Turkey in 1979 and 1982 and the Yugoslavia Plan Review Mission in 1981-2/ The chief characteristic of a CGE model is that it simulates the working of a market system in which prices vary in various markets (for factors, products, and foreign exchange) in response to changes in supply and demand conditions. In contrast to the simpler, linear input-output models that assume no change in relative prices, a CGE model permits the simulation of the results of policy packages which operate through the price mechanism, such as exchange rates, tariffs and export subsidies. Economic performance is seen as the outcome of decentralized optimizing decisions made by producers and consumers in response to market conditions, which in turn are influenced by government policies, international economic conditions, and other exogenous forces. It is important to stress that a CGE model, like the simpler input- output model, is primarily a simulation model to be used for planning purposes, not a forecasting model. It is intended to be applied to periods long enough that market mechanisms can be expected to respond to changes in 1/ A more complete presentation of the model is contained in Appendix I. 2/ See, for example, World Bank (1980a).  -5- market conditions. Within each period, the model reaches, a static equilibrium solution in which all markets are modeled as clearing subject to specified behavioral rules and parameter values. This approach is in stark contrast to standard macroeconomic forecasting models, which derive their results from complex specifications as to the speed at which markets respond to changing economic conditions, and the different structure of these adjustment lags in various sectors of the economy. The CGE model is thus not intended for macroeconomic forecasting, but instead concentrates on the determination of relative prices among sectors and on the structure of real variables in the economy. The model is Walrasian and hence only determines relative prices. The selection of a numeraire or price normalization rule exogenously fixes the absolute price level, so that a solution to the model yields sectoral prices relative to the specified aggregate price level. The domestic rate of inflation is also specified exogenously, and the real focus of the model is further reflected in the complete absence of financial markets or monetary variables. Another consequence of the real focus of the model relates to the exchange rate. Just as the prices of outputs in all sectors are relative prices, so too is the exchange rate treated as a relative price within the model. What matters to the model is not the nominal level of the exchange rate over time, but rather whether the exchange rate movements are greater or less than the differential between domestic and world inflation. If the change in the nominal exchange rate from period to period just offsets differential world and domestic inflation rates, this corresponds to a constant real (or price level deflated) exchange rate. Only if the change in the exchange rate exceeds (falls short of) this differential will there be a depreciation (appreciation) in the real exchange rate. If one were to raise  -6- the estimate of domestic inflation for a certain period, and raise the nominal exchange rate by the same proportion, this would have no impact on any of the real variables in the model. It is worth emphasizing at this point that the idea of markets responding to changing economic conditions does not imply that markets are "perfect" in the neoclassical sense. Instead, the model explicitly takes account of rigidities and distortions in the economy, and attempts to model them formally. For example, capital already in place in each sector is fixed, with the capital stock modified only through the allocation of new investment, and sectoral wages for each type of labor are required to conform to an exogenously specified pattern embodying wage "distortions" existing in Turkish labor markets. When the model is operated intertemporally, it is necessary to specify how the exogenous variables change over time. When the model is used to examine future economic prospects, this implies forecasting the movement of these exogenous variables. In some cases (e.g. employment growth), these forecasts can be derived from extrapolation of historical trends; in other cases, (foreign capital inflows, investment allocation and levels), greater judgement is required since the magnitudes are presumably quite dependent on the performance of the economy itself, which in turn is responsive to the full range of government policy instruments (the exchange rate, tariffs and subsidies, tax policy). This last point deserves further emphasis. All projections generated by the model are conditional on the specific policy packages simulated and the projected levels of exogenous variables. Government policy plays a dominant role in the results, so that the choice among alternative scenarios is contingent on the government undertaking all of the policies simulated in the model.  - 7 - 3. The Turkish Economy: A Historical Analysis, 1978-81 A. Turkish Economic Performance, 1973-77: A Perspective The 1973-77 period in Turkey can be aptly characterized by the dramatic swing that occurred from a positive to a negative foreign exchange balance. Although foreign exchange crises had been associated with economic cycles in the past, the crisis which culminated in 1977 was much more serious than the previous ones, not only because of its magnitude, but also because of the severe domestic and foreign economic disruptions that accompanied it. Real GDP increased at an annual rate of 7.3 percent between 1970 and 1977 2. This growth rate compares favorably with 6.0. percent growth rate during the 1960s. Moreover, a high GDP growth rate of 7.2 percent was maintained during the Third Plan Period (1973-77), despite the first oil shock and its aftermath.21 Although this high growth rate of real GDP was partly attributable to government policies that insulated the Turkish economy and postponed adjustment to changing world conditions, dramatic increases in production in the industrial (mining, manufacturing and electricity, gas, water) and services sectors did occur. On the expenditure side, a rise in fixed investment contributed the most to the growth of GDP. The accelerated X24 Estimates of real GDP are calculated in terms of domestic prices. This tends to overestimate the growth rate because the output of the manufacturing sector, which was growing much faster than other sectors, was overvalued in Turkey compared to the rest of the world during the period concerned. See Balassa (1981), Appendix Table 1, who finds that the GDP growth rate for the 1967-77 period is thus overestimated by 0.5 percentage points. 2/ However, the quality of economic growth might not have been improved because of the following two reasons. First, the efficiency of investment declined as reflected in an increase in incremental capital output ratios from 2.6 in 1968-72 to 3.8 in 1973-77 (see Balassa (1981), p. 21). Second, there was considerable accumulation of inventories as indicated by an increase in their share in GDP from 0.9 per cent in 1973 to 2.5 percent in 1977 (Table 3.1).  - 8 - growth rate of real GDP in the 1970-77 period was accompanied by a higher domestic inflation rate, which accelerated steadily during the Third Plan period to reach an average of 20 percent annually, in contrast to the relative stability of the price level in the 1960s. However, this increase in domestic inflation was not excessive in comparison with the experience of other developing countries. The growth rate of agricultural production lagged far behind the rates in industry and services during the three plan periods, resulting in a shift in the structure of production away from agriculture. In 1962, at the beginning of the first Five Year Plan, the shares of agriculture, industry and services in GDP were 39.6, 21.4, and 39.0 percent respectively. However, by 1977 the agricultural share had declined to 28.1 percent whereas the shares of industry and services had increased to 26.1 and 45.8 percent. Despite this decline in the agriculture share of GDP, agricultural growth rates themselves were quite respectable by international standards. Although the change in the sectoral composition of employment shows a trend similar to that of output production, the share of workers employed in the agricultural sector remained high at 63.1 percent in 1977, while the employment shares of industry and services were 11.9 and 25.0 percent. The rapid growth of real GDP in the 1973-77 period stemmed largely from the growth of investment, as Table 3.1 illustrates. During the period the annual growth rate of real gross investment more than doubled from the 1968-72 period to reach 15.0 percent. Public investment leaped from a 5.4 percent annual growth rate in 1968-72 to 17.9 percent in 1973-77. During the 1973-77 period total consumption grew at 6.6 percent annually, while exports actually declined by 3.7 percent per year. As a result, the shares of gross fixed investment, consumption and exports in real GDP were 24.7, 78.4 and  -9- Table 3.1 Structure and Growth of Real GDP, 1973-77 (Percent) Structure Annual Growth 1973 1977 1973-77 Total Consumption 81.4 78.4 6.6 Private Consumption 70.0 66.8 6.3 Government Consumption 11.3 11.6 8.2 Total Investment 20.9 27.2 15.0 Private Fixed Investment 10.5 13.6 9.0 Government Fixed Investment 9.4 11.1 17.9 Stock Changes 0.9 2.5 37.7 Exports 7.8 5.0 -3.7 Imports 10.0 10.6 9.0 GDP 100.0 100.0 7.6 Source: World Bank (1982), Table 2.4, p. 410. 5.0 percent in 1977. The share of imports in GDP increased slightly from 10.0 percent in 1973 to 10.6 percent in 1977. The overall composition of the allocation of investment, which emphasized the industrial sector and related infrastructure because of the industrialization policy, stayed largely unchanged. In the Third Plan Period the transport and communications sectors received a larger share of investment whereas the shares of energy, health, education and housing declined. Such a trend is observed for both public and private investment. Although gross domestic savings grew significantly, at 8.8 percent annually in the 1973-77 period, it was not enough to finance gross investment, which was increasing at an even higher rate. Consequently, the investment-savings gap increased to 9.4 percent of GDP in 1977 from -2.5 percent in 1972. This potentially disastrous resource gap was filled by increased foreign borrowing.  - 10 - The balance of payments situation reversed during the 1973-1977 period. In 1973, although the balance of trade (goods and services) was in deficit at $592 million, this was more than offBet by the inflow of workers' remittances and net factor service income which resulted in a $515 million current account surplus (see Figure 3.3 below). However, the subsequent rapid increase in imports coupled with a reduction in exports increased the trade deficit to $3.88 billion U.S. dollars in 1977. One of the important factors contributing to the deterioration of the trade balance seems to be the appreciation of the real exchange rate during the period.'/ In addition to the unfavorable balance of trade situation, a substantial decline in workers' remittances and net factor service income contributed to a record high current account deficit of $3.572 billion. In order to finance the deficit, the Turkish authorities resorted to heavy reliance on short-term international capital markets. In 1977, Turkish exports consisted largely of food and livestock and raw materials, whereas its imports were mostly manufactured goods and mineral fuels. The export structure remained fairly stable between 1970 and 1977, with the share of manufactured goods increasing slightly at the expense of raw materials. A significant increase in the share of mineral fuels (petroleum) was observed in the import structure. In analyzing the foreign exchange crisis that began in 1977, one has to bear in mind the unprecedented occurrences in the international environment during the 1973-77 period. One was the "oil shock," which quadrupled the price of oil in 1974. Another important development was the existence of stagflation in the industrial countries, caused primarily by the oil shock and subsequent contractionary policies pursued by most of the countries. Against 24 According to Balassa (1981), the real exchange rate with Turkey's trading partners appreciated by 13.2 percent between 1973 and 1977 (Table 3).  this background, the Turkish economy was growing relatively rapidly. Naturally, the balance of trade situation worsened because of sluggish export demand abroad, the increased cost of imports due to higher oil prices, and higher import demand arising from sustained economic expansion. Moreover, the current account balance also deteriorated because of reduced workers' remittances, which could be attributed both to the recession in European countries and to the increasing overvaluation of the exchange rate. Dervis and Robinson have analyzed the important factors which led to the foreign exchange crisis in 1977 utilizing an earlier version of the CGE modelA.' Specifically, the impact of several major macroeconomic events on the exchange rate was examined. Five different experiments were performed starting from a base run which replicated historical developments: Exp.1 = base run and a flexible exchange rate and more moderate borrowing, Exp.2 = Exp.1 and moderate inflation, Exp.3 = Exp.2 and no oil price increase, Exp.4 = Exp.3 and no investment rate increase, Exp.5 = Exp.4 and higher remittances. The first experiment provides the "equilibrium" exchange rate, and the exchange rates obtained from the other experiments are compared to this equilibrium exchange rate to examine the influence of particular economic events. Their estimates show that the contribution of various factors to the depreciation of the equilibrium exchange rate from 1973-77 are as follows: differential domestic and world inflation rates, 32 percent; the oil price increase, 12 percent; the increase in the investment rate, 5 percent; the decrease in worker's remittances, 21 percent; and other factors, 29 percent. 1/ See Dervis and Robinson (1978) and Dervis and Robinson (1982).  - 12 - The higher inflation rate in Turkey relative to that in other countries is thus shown to have contributed the most to the devaluation of the equilibrium exchange rate. However, this effect only captures about one-third of the total equilibrium devaluation, so that simple calculations of the required devaluation based on comparative inflation rates would have been quite wrong. Dervis and Robinson concluded from the experiments that even if there had been no unfavorable external developments (i.e., no oil price increase and no decline in workers' remittances), the foreign exchange crisis would have appeared anyway, albeit at a later date, because of growing strains in the domestic economy due to the economic policies being followed. B. Comparison with Other Middle Income Countries In order to provide a benchmark for analyzing economic projections, it is important to examine how the Turkish economy has been performing in the last two decades compared to other middle income countries (MICs). Table 3.2 presents various economic indicators for Turkey and for the simple average of 12 other middle income countries for the years 1960, 1970 and 1977. 1977 is the latest year for which data for all the countries are available from the World Tables II. The comparison of Turkey with the average results of the other MICs highlights Turkey's performance relative to a diverse group of countries; individual comparisons can be made using the data in Appendix Table II-9. 2/ The population of Turkey has increased from 27.5 million to 41.9 million between 1960 and 1977, representing a 2.5 percent annual growth 1/ World Bank (1980b). 2/ The other countries in the MIC sample include (in increasing order of GNP): Egypt, Thailand, Philippines, Korea, Taiwan, Colombia, Peru, Brazil, Mexico, Yugoslavia, Argentina, and Spain.  - 13 - Table 3.2 Comparative Economic Indicators Population Per Capita GNP Average _- Average _ Annual Annual Number Growth Level Growth (million) Rate (x) (1977 US$) Rate (x) Turkey 1960 27.5 - 613 1970 35.3 2.5 866 3.3 1977 41.9 2.5 1110 4.0 MIC Sample Average 2/ 1960 26.6 627 1970 34.1 2.5 934 4.2 1977 40.2 2.5 1223 3.9 Ratios to Current Price GDP (Percent) 3/ Gross Agriculture Manufacturing Fixed Value Added Value Added Exports Imports Investment Turkey 1960 41.4 12.0 4.2 6.5 15.3 1970 39.0 15.9 5.8 8.0 17.1 1977 29.0 17.5 5.6 12.3 22.4 MIC Sample Average 2/ 1q60 27.3 21.6 11.6 13.2 19.5 1970 22.6 22.8 14.0 16.1 21.7 1977 19.8 26.1 18.5 21.5 23.9 Notes: 1/ Growth rates are for the 1960-70 and 1970-77 periods. 2/ Average figures derived as unweighted averages of values for all middle income countries in the sample, excluding Turkey. See Appendix Table II.9 for the individual country figures. The countries are: Egypt, Thailand, Philippines, Korea, Taiwan, Colombia, Peru, Brazil, Mexico, Yugoslavia, Argentina, and Spain. 3/ Agriculture and manufacturing are expressed as shares of GDP at factor cost; export, imports, and investment are shares of GDP at market prices.  - 14 - rate. These figures are very close to the averages for the other countries. Per capita GNP for Turkey was $613 and $1,110 (1977 prices) in 1960 and 1977, with Turkey ranking eighth and seventh among the 13 countries in the respective years. The annual growth rate of per capita GNP for Turkey increased from a below average rate of 3.3 percent in the 1960-70 period to above average at 4.0 percent in the 1970-77 period. In spite of its higher growth rate in the 1970-77 period, GNP per capita in Turkey was only about 90 percent of the sample average for 1977, and Turkey moved up only one place in the ranking. In the next five columns of Table 3.2 and in Figures 3.1 and 3.2, the shares of various economic aggregates in GDP are shown. The first two (shares of agriculture and manufacturing value added) are components of GDP in production whereas the last three (shares of exports, imports and gross domestic investment) are those of GDP in expenditure. As Figure 3.1 shows, the share of agricultural value added in GDP in Turkey declined substantially, from 41.4 percent in 1960 to 29.0 percent in 1977. However, these shares are significantly higher than the sample average shares of 27.3 percent and 19.8 percent in those respective years. In 1960, only Korea had a higher share than Turkey while Egypt, Thailand and Colombia had higher shares than Turkey in 1977. Conversely, Turkey had the second lowest manufacturing share (12.0 percent) in 1960, only slightly higher than Korea (11.5 percent). However, with the Korean industrialization drive of the 1960s and 1970s raising its share to 23.2 percent in 1977, Turkey represented the least industrialized of the 13 nations in 1977, with only 17.5 percent of value added in manufacturing. While for the sample of MICs, the manufacturing share of GDP had surpassed that of agriculture by 1970, in Turkey this reversal was not reached even in 1977 (see Figure 3.1).  - 15- -