Growth policies and the exchange rate in Turkey

World Bank Reprint Series: Number 181 Bela Balassa Growth Policies and the Exchange Rate in Turkey Reprinted with permission from The Role of Exchlange Rate Policy in Achieving the Outward Orientation of the Turkishi Econiomy, papers from a conference in Istanbul, July 20-21, 1979 (Istanbul: Meban Securities Brokerage and Finance Corporation, 1981), pp. 15-59.  GROWTH POLICIES AND THE EXCHANGE RATE IN TURKEY Bela Balassa* July 21, 1979 * The author is Professor of Political Economy at the Johns Hopkins University and Consultant to the World Bank. He is indebted to Turkish government officials, businessmen and academicians, and to members of the Bank mission to Turkey for helpful discussions on the subject areas covered by the paper, written following a visit to Istanbul and Ankara on April 23-28, 1979. Thanks are also due to Jacques de Groote and Merih Celasun for their valuable comments on an earlier version of the paper and to participants at the Seminar on the Role of the Exchange Rate in Shaping an Outward Oriented Economy, held in Istanbul on July 20-21, whose comments were taken into account in the final revision. None of the above should be responsible, however, for the views expressed in the paper, nor should the paper be interpreted as reflecting the views of the World Bank. The author was assisted by Michel Noel in the preparation of statis- ticol material for the woaer. 15  INTRODUCTIC'N This paper will review Turkey's growth performance during the sixties and the seventies and make recommendations .for policies for stable eco- nomic growth by taking the policy reforms instituted in the Spring of 1979 as the point of departure. The adjective 'stable' refers to the goal of mitigating the pronounced cyclical pattern that has characterized Tur- kish economic growth in the past. Section I will provide the background for the policy recommendations. It will examine Turkey's growth performance in relation to other major semi-industrial countries, analyse the effects of special factors on eco- nomic growth in Turkey, and describe the Turkish economic situation prior to the policy changes introduced in the Spring of 1979. Issues relat- ing to the measurement of economic growth in Turkey are considered in the Appendix. Section II will make comparisons with the situation existing in Brazil in the mid-sixties and examine the economic effects of the policy reform carried out at the time in Brazil. This will be followed by an evaluation of the policy measures adopted in Turkey between March and June 1979. In Section III, policy recommendations will be made for adopting an export-oriented strategy that would make increased use of market signals. The policy measures under consideration pertain to exchange rates, the system of protection, the allocation of investment funds, the treatment of foreign investment, interest rates and capital markets, the reform of administrative procedures, and reforms in the state enterprise sector: 16  1. TURKEY'S GROWTH PERFORMANCE IN AN INTERNATIONAL PERSPECTIVE ECONOMIC GROWTH IN MAJOR SEMI-INDUSTRIAL COUNTRIES, 1960-1978 Table 1 shows average annual growth rates of GNP, population, and per capita GNP in the periods 1960-66, 1966.73, 1973-75, 1975-78, and 1973-78, together with absolute figures for 1973, for selected semi-indust- rial countries that may be considered representative of alternative deve- lopment strategies. They include five Southern European countries: Greece, Portugal, Spain, Turkey and Yugoslavia; two Far Eastern count- ries: Korea and Taiwan; and a Latin American country: Brazil. In the following, changes in per capita incomes will be used to evaluate the growth performance of these countries. Among non-European countries, economic growth accelerated' in the 1960-66 period in Korea and Taiwan that adopted export-oriented policies in the early sixties (Table 1). In turn, continued reliance on im- port substitution behind high protective barriers encountered limitations in Brazil, contributing to the decline in the rate of economic growth ex- perienced during this period. With subsequent policy changes in the di- rection of export-orientation, described in Section II below, economic growth accelerated in Brazil between 1966 and 1973.(1) In the same period, an acceleration of growth occurred also in the two Far Eastern countries that have provided practically free trade treatment to exports. In Southern Europe, per capita income growth rates were consist- ently higher in countries with a greater export-orientation (Greece, Por- tugal, and Spain) than in Turkey, which - with the partial exception of the devaluation episode of 1970 - continued to follow a policy oriented towards import substitution. This conclusion also applies, albeit differen- ces in growth performance are smaller, if Turkey is compared with Yu- goslavia, where a partial reversal of export-oriented policies occurred during the sixties (ibid., p. 32). For the 1960-73 period, taken as a whole, (') Other Latin American countries with a similar experience include Aigentina and Colurnbia; in turn,- Chile and India provide examples of countries where inward-looking policies continued after 1966, with ad- verse effects on economic growth. For a detailed discussion, see my ((Export lncentives and Export Performance in Developing Countries: A Comparative Analysis)) (1978) that examines the experience of eleven semi-industrial developing countries, classified according to the policies followed, in the 1953-73 period. ?7  Table 1 GNP, Population and Per Capita GNP Growth Rates in Major Semi-Industrial Countries 1q73 (absolute values) 1960.66 1966-73 1960-73 1973-75 1975-78 1873-78 Greece Y 16.7 7.5 7.6 7.4 0.8 5.2 4.0 P 8.9 0.5 0.5 0.5 0.7 1.0 1G Y/P 1870 7.0 7,2 6.8 0.2 4.2 3.1 Portugal, Y 12.7 6.3 7.9 7.2 -2.8 5.4 2.3 P 2.0 0.1 0.8 0.3 3.5 0.7 1.8 Y/P 1410 6.2 7.1 6.8 -6.4 4.7 0.5 Spiln Y 59.4 8.0 6.3 6.7 3.0 2.3 2.2 P 34.7 1.0 1,1 1.1 1.1 1.1 1.1 Y/P 1710 7.0 5.2 5.7 1.9 1.2 1.2 Turkey Y 22.6 5.9 6.7 6.3 6.9 5.8 6.4 P 37.9 2.5 2.5 2.5 2.4 2.5 2.4 YIP 600 3.4 4.2 3.9 4.5 3.4 3.9 Yugoslavia Y 21.2 6.1 6.9 6.2 7.9 6.5 6.7 P 21.0 1.1 0.9 1.0 1.0 0.9 0.9 Y/P 1010 5.0 6.0 5.2 7.0 5.6 5.8 Korea y 13.3 6,9 10.0 9.0 8.2 11.6 10.6 P 32.9 2.7 2.1 2.4 1.7 1.6 1.6 Y/P 400 4.2 7.9 6.6 6.5 10.0 8.9 Taiwan Y 10.3 8,6 10,0 9.3 1.5 10.0 6.9 P 15.4 3.1 2.8 2.9 1.8 1.8 1.8 Y/P 660 5.4 7.2 6.3 0.3 8.2 5.1 Brazil Y 77.0 3.8 10.1 6.8 7.2 6.2 6.6 P 101.1 3.0 2.8 2.9 2.7 2.8 2.8 Y/P 760 0.9 7,3 3.9 4.5 3.4 3.8 Source: Data bank of the World Bank. Note: Y = GNP (in U.S. $ billion) P = Population (in million) Y/P = Per capita GNP in U.S. $ Growth rates of GNP and per capita GNP have been es- timated by regressing annual data, expressed in constant domestic prices, on time. Population growth rates have also been derived by regression analysis. 18  per capita incomes rose at an annual average rate of 6.8 per cent in Greece and Portugal, 5.7 per cent in Spain, 3.9 percent in Turkey, and 5.2 per cent in Yugoslavia.(') The import substitution-orientation of the Turkish economy is indi- cated by the relatively low shares of merchandise exports and imports in GNP that averaged 7.8 per cent in 1973 (Table 2). In the same year, the relevant figures for the other Southern European countrHes were: Greece, 14.7 per cent; Portugal, 21.2 per cent; Spain, 10.5 per cent; and Yugoslavia, 19.9 per cent. Finally, average export and import shares were 28.6 per cent in Korea and 40.2 per cent in Taiwan and reached 8.2 per cent even in Brazil, the largest semi-industrial developing country in terms of. GNP, having risen from 5.6 per cent in 1966 following the policy reform of the mid-sixties.(2) Table 2 Export Shares, Export and Import Growth Rates, and Capital Inflow Shares in Major Semi-Industrial Cou.ntries Merchandise Export and Import 1973-78 Average Annual Growth Rates Shares In GNP. 1973 of the Volume of Merchandise Trade Share of Not Capitol Inflow In GNP Exports Imports E&ports Imports 1973 1974 1975 1976 1977 1971 (1) (2) (3) (4) (5) (l) (7) (a) (9) (10) Greece 8.6 20.7 8.4 1.9 7.1 6.4 4.7 47 4.7 3,9 Portugal 18.0 26.4 06 1 1 -3.0 6.2 5.1 8.1t 93 44 Spain 7.3 13,8 11.9 1 5 -8.3 37 3.4 4.1 2.1 -1.6 Turkey 5.8 9.7 2.5 3.1 -2.8 2.1 5 0 4.7 6.9 2.8 Yugoslavia 15.4 24.3 5.3 2.4 -2.6 4.9 37 -0.5 4.0 2.1 Korea 24.7 32.5 197 14.8 2.3 11.3 9.5 1.1 00 2.4 Taiwan 43.1 37.3 14 5 91 -5.8 8.1 40 -19 -50 -7.2 Brazil 7.7 8.7 48 24 2.7 72 57 46 31 3.8 Sources: IMF: International Financial Statistics, July 1979. Spain: GNP in 1978, IMF estimate. Portugal: 1977-78 data: Banco de Portugal estimate. Yugoslavia: GNP in 1977 and 1978: IMF estimates; 1978^ net capital inflow: IMF estimate. Note: Net capital inflow has been equated to the current account deficit, it thus includes changes in foreign exchange reserves. Negative figures refer to a net capital outflow. (') Here and thereafter,%growth rates have been derived by regress- ing annual data on time. (2) Udiless otherwise noted, the data originate in IMF, International Financial Statistics. In the following, the expressions ((exports)) and ((im- ports) will be used to refer to merchandise exports and imports. 19  All the countries under consideration were adversely affected by the quadrupling of oil prices in late 1973 and suffered the effects of the 1974-75 world recession. Policy reactions to thesm changes, however, varied from country to country. In the Far East, Taiwan occeoted a virtual economic standstill in the years 1974-75, followed by the resumption of rapid economic growth based on the continued pursuit of an outward - ori- ented strategy. Similar conclusions apply to Korea, except that large foreign borrow- ing made it possible to limit the decline of growth rates in the years 1974- 75 (Tables 1 and 2). The policies followed by the two Far Eastern countries permitted them to surmount the adverse effects of the world recessign on exports, so that between 1973 and 1978 the volume of exports grew at an average annual rate of 14.5 per cent in Taiwan and 19.7 per cent in Korea. In Southern Europe, the continuation of an export-oriented stance contri- buted to the malntenance of high export growth rates in Greece (8.4 per cent) and Spain (11.9 per cent). And while Greece and, to a lesser extent, Spain, supplemented their export earnings by foreign borrowing, they accepted an increase in import volume much smaller (1.9 and 1.5 per cent, respectively) than that of exports. Export growth rates exceeded the growth of impo'rts also in Brazil (4,8 per cent and 2.4 per cent) and Yugoslavia (5.3 per cent and 2.4 per cent). At the same time, Brazil and, to a lesser extent, Yugoslavia, relied on foreign borrowing in order to reach high rates of economic growth. Finally, Portugal extensively used foreign loans to offset the economic dis- locations attendant upon the April 1974 Revolution. With policy improve- ments following the stabilization of the political situation, however, the volume of Portuguese exports surpassed the 1973 level in 1978 and the share of foreign borrowing in GNP declined to a considerable extent. In Turkey, efforts made to maintain high rates of economic expansi- on through increased foreign borrowing led to an increase in the volume of imports by one-half between 1973 and 1977. Correspondingly, whereas ir? the other semi-industrial countries under consideration the share of foreign borrowing in GNP tended to decline over time, it rose in Turkey, reaching 6.9 per cent in 1977. With increased difficulties encountered in borrowing abroad, this share decreased in 1978, lecding to a substantial fall of imports and a slowdown of economic growth. Notwithstanding the decline experienced in 1978, the annual average rate of increase of imports in Turkey exceeded that of exports for the 20  period taken as a whole. Exports were adversely affected by the appre- ciation of the real exchange rate,(') and their share in GNP fell from 5.8 per cent in 1973 to 4.5 per cent in 1978, despite the operation of special factors in that year (p. 25). ECONOMIC GROWTH IN TURKEY: AN EVALUATION While the more export-oriented semi-industrial countries experienced higher rates of economic growth, the 3.& per cent average annual in- crease of incomes per head in Turkey did represent a doubling of per capita incomes between 1960 and 1978. In this connection, the question arises what factors explain Turkey's growth perforrtance under the po- licies followed. As shown in the Appendix, the movement of labor from agriculture to industry and services importantly contributed to economic growth in Turkey during the period under consideration, It is also apparent, however, that the protection of industry and the disprotection of agriculture gave rise to an overestimation of the rate of producivity growth on the nationol economic level. The intersectoral movement of labor and increases in total employ- ment were supported by the high level of investment activity in Turkey. But, new investment brought successively smaller increments of output and created fewer jobs per capital invested. According to estimates based on official Turkish data, the overall incremental capital-output ratio in- creased from 2.3 in 1963-67 to 2.6 in 1968-72 and to 3.8 in 1973-77 while the amount of investment per job created, expressed in 1976 prices, rose from TL 373 thousand to TL 522 thousand and, again, to TL 722 thousand. The observed changes on the national economy level are explained in large part by the increasing capital-intensity of the manufacturing sector that was associated with continuing import-substitution. Thus, the incremental capital-output ratio in manufacturing rose from 1.6 in 1963- 67 to 2.4 in 1968-72 and to 4.7 in 1973-77. Parallel with these changes, the amount of investment per job created, again expressed in 1976 prices, increased from TL 267 thousand to TL 363 thousand and, finally, to TL 572 thousand, (i) The nominal exchange rate adjusted for changes in wholesale prices at home and abroad. See further p. 25 below. 21  The implications of alternatives to import substitution for the growth of the Turkish manufacturing sector have been examined by Anne Krueger who estimated the hypothetical growth rate of value added in the manu- facturing sector for the 1967-72 period on the assumption that a balanced export promotion and import substitution policy was followed. Measured in world market prices, a growth rate of 16.5 per cent is obtained under this alternative as compared to 10.3 per cent under the plan allocation (1974, p. 259). The differences are explained by the fact that, due to the capital-intensive nature of import substitution, incremental capital-output ratios under the plan were 48 per cent higher than under a balanced policy, which has been defined as one that would have allocated new in- vestment in proportion to each industry's value added in the base year, 1967 (p. 260). At the same time, the method applied underestimates the extent of inefficiencies under the import substitution strategy followed by Turkey, since there were high-cost import substituting industries already in the base year. It appears, then, that the continuation of the strategy of import substi- tution gave rise to inefficiencies in Turkey. This conclusion is supported by the fact that while the contribution of import substitution to the growth of the manufacturing sector was positive in the period 1963-68, it turned negative in the period 1968-73 (Celasun, Tables 1II. 8 and ll. 9),(') reflect- ing in part the decline in net foreign exchange savings from import substi- tution. At the same time, foreign exchange savings in import-substituting industries were obtained at a high cost to the domestic economy. In 1965, for example, the average ratio of domestic resource costs, estimated at the stiadow prices of the factors of production, to net foreign exchange savings (earnings) was 3.4 times higher in ten import-substituting induitri- es than in five export industries (Krueger 1966, p. 473). Notwithstanding the inefficiencies associated with the import substi- tution strategy, Turkey was able to avoid a decline in the rate of eco- nomic growth until 1977 because of increases in the share of gross fixed investment in GDP. According to official statistics, this share increased from 16.0 per cent in 1963-67 to 18.0 per cent in 1968-72 and, again, to 22.9 per cent in 1973-77. With incremental capital-labor ratios rising more (1) Similar results were reached for the same time periods by Ciller (Table 3.5) and for the periods 1963-67 and 1967-71 by Olgun (Tables 3.5 and 3 6) 22  rapidly than incremental capital-output ratios, the rise in the rate of in- vestment did not however suffice to fully absorb increases in the labor force, and unemployment rates as well as the share of ((discouraged)) workers in the potential labor force showed continuing increases.(') The growth of investment, in turn, was favorably affected by special factors, viz. i.ncreases in workers' remittances(2) and foreign borrowing. Recorded workers' remittances rose from negligible amounts in 1966 to 5.6 per cent of GNP in 1973. In the latter year, workers' remittances reached 94 per cent of the value of merchandise exports and surpassed this by a substantial margin if adjustment is made for the imported inputs used in export production. Recorded workers' remittances, expressed in terms of U.S. dollars, declined from the peak of $1.7 billion reached in 1974 and did not reach $12 billion in recent years. The resulting decline in the ratio of recorded remittances to merchandise exports and GNP appears to have been off- set, however by remittances channeled through the parallel foreign ex- change market. By 1977, these ,'unofficial' remittances reportedly exceed- ed recorded remittances by a conside.rable margin (Qiller, 1978, p. 16), so that the recorded and unrecorded remittances combined may have amounted to 6 per cent of GNP. Moreover, after the oil crisjs and the 1974-75 world recession, lur- key borrowed substantial amounts abroad. By 1977, the net capital inflow was nearly double of export value and approached 7 per cent of GNP, as compared to practically nil in the early seventies. As noted above, the subsequent decline in foreign borrowing to 3 per cent of GNP is explained by the limited availability of foreign funds. Apart from contributing to increased investments, workers' remittan- ces and the inflow of foreign capital have had multiplicative effects on Turkey's GDP, due to the existence of foreign exchange stringency that has aggravated to a considerable extent in recent years. According to one study, in the period 1973-77 the growth rate of GDP would have been 1.2 percentage points lower if net foreign borrowing did not exceed its (') This may be represented by the decline in labor force participation rates, shown by official data, from 80.0 per cent in 1962 to 75.0 per cent in 1967, 70.6 per cent in 1972, and 67.7 per cent in 1977 (2) The increase in the number of workers abroad also reduced un- employment rates. 23  assumed <(normal)) level. The corresponding figures are 3.7 percentage points for that of investment (Dervis-Robinson, 1979 Table 6.8).(l) With lower foreign borrowing, then, the rate of growth of domestic investment and consumption would have declined much more than that of GDP as borrowing permitted raising domestic expenditure much above domestic production. At the same time, the estimated figures represent the lower limit of possible values. To begin with, the norm of foreign borrowing chosen much exceeds the average for the period preceding Lhe oil crisis.(2) Moreover, the method applied underestimates the adverse effects of for-eign exchange stringency.(') Finally, the calculations assume that a flexible exchange rate policy would have been followed, with 'he exchange rate maintained at equilibrium levels throughout the period. In fact, Turkey has not applied such a policy during the period under con- sideration, and the shortfall in GDP and in domestic expenditure would have been larger under nonoptimal (fixed) exchange rate policies. Similar considerations apply to workers' remittances. At the some time, there is an important difference between the two, in as much as the former but not the latter gives rise to foreign indebtedness. In fact, with much of Turkey's debt being short-term, recent estimates show that the servicing of the debt surpassed $1.0 billion in 1978, amounting to 47 per cent of the exports of goods and services. This ratio may reach 55 per cent in 1979, TfHE TURKISH ECONOMIC SITUATION IN EARLY 1979 Debt-service ratios, calculated for a particular year, abstract from the dynamics of economic development. Thus, historical experience shows that large borrowing on a temporary basis will not adversely affect (') Under the definition used by the authors, onormalx foreign borrow- ing would have amounted to $940 million in 1977, i.e. approximately 2 per cent of GNP (ch. 6, p. 19). (2) In the years 1956-73, net foreign borrowing averaged .3 per cent of GNP. The average figure is .5 per cent if we exclude the years 1972 and 1973 that are characterized by an accumulation of foreign exchange reserves, which is considered as a capital outflow. (3) This is because of the use of a time trend of total factor produc- tivity (technical change) unaffected by changes in the availability of fo- reign exchange in the calculations and because of the assumption of continuous substitution between capital and labor (i.e. lack of disconti- nuity in the form of bulky investments). 24  a country's credit-worthiness, provided that it is followed by rapid export expansion that reduces the debt-service ratio over time. In the post-1973 situation, Korea provides a par excellc.,c.e case of such a development. After foreign borrowing reached 11.3 per cent of GNP in 1974 and 9.5 per cent in 1975, exports increased rcpidly, reducing the ratio of debt service obligations to previously-contracted debt and eliminating the need for further foreign borrowing. Similar developments occurred in Taiwan, except that the extent of foreign borrowing in 1974 and 1975 had been smalier and this has subsequently given rise to a net inflow of capital. In Brazil, too, debt-service ratios and the inflow of foreign capital, expressed as a percentage of GNP, declined as export growth accelerated, although in 1978 exports were adversely affected by weather conditions. Similar developments did not occur in Turkey, wiere the debt-service ratio increased in 1978 notwithstanding the fact that the drawing-down of inventories accumulated previously gave a boost to exports in that year. Also, as noted above, in the years following the oil crisis, imports in- creased to a considerable extent, surpassing the growth of exports in volume terms. Yet, in view of the deterioration of Turkey's terms of trade due to "Lhe quadrupling of oil prices, adding aboLit $0.5 billion (approxi- mitely 2 per cent of GNP) to Turkey's import bill in 1974, balance of pay- ments equilibrium would have required reducing imports and increasing exports, Insufficient export increases and the excessive rise of imports may be explained, in large part, by the direct and indirect effects of foreign borrowing. For one thing, the in;rease in incomes made possible by fo- reign borrowing syphoned off part of the exportable s!'rplus and led to Ihigher imports. For another thina, the inflow of foreign capital permitted an appreciation of the exchange rate in real terms, with adverse effects on exports and imports, Compared to the situation existing in 1973, by the first quarter of 1979 the real exchange rate appreciated by 26 or 31 per cent vis-a-vis the U.S. dollar and by 18 or 24 per cent vis-a-vis the cur- rencies of Turkey's major trading partners, on1 the average, depending on whether use is made of the wholesale price index compiled by the B3u- siness Research and Publications Department of the Ministry of Commer- ce or by the Istanbul Chamber of Commerce (Table 3). At the same time, the exchange rate applicable in 1973 already rep- resented a considerable appreciation as compared to the situation ex- isting after the large devaluation irn 1970, Moreover, the calculations ab- 25  stract from the fact that Turkey should have devalued more than purchas- ing power parity relationships, expressed in the real exchange rate cal- culations, would indicate. According to Dervis arid Robinson, the qua- drupling of oil prices would have required a devaluation 22 per cent greater than that indicated by purchasing power parity relationships. And, according to these authors, increases in the export prices of the deve- loped countries would have required an additional 12 per cent devalua- tion in order to offset the resulting deterioration of Turkey's terms of trade (1979, Figure 6.3). As the ratio of export to import prices in Turkey declined by 22 per cent between 1973 and 1977, these estimates imply that an additional 1.5 per cent devaluation would be needed to offset each one per cent deterioration of the terms of trade. The adverse effects of the appreciation of the real exchange rate on exports was offset only in small part by increases in export rebate rates. (1) At the same time, available evidence points to the existence of a posi- tive correlation between real exchange rates and export performance. Thus, it has been shown that the ratio of exports to value added in agri- culture and industry follows changes in real exchange rate with some time lag (TUSIAD, April 1978, pp. 32-33). In particular exports reacted strongly to the 1970 devaluation. While exporters received the official exchange rate augmented by rebates,(2) ((implicit)) exchange rates on imports were much higher and varied greatly from product to product. Thus, tariffs are high and show considerable dispersion; advance deposit requirements are in effect on most imports; and, furthermore, the scarcity value of imports varies de- pending on access to import licenses and to foreign exchange allocation. The situation was aggravated by reason of the fact that, in the past two years, foreign exchange allocations for private imports were practically (I) These rates were increased more or less steadily between 1973 and 1977; they were reduced on the occasion of the March 1978 devalua- tion, and raised again subsequently, averaging about 15 per cent on the f-b value of manufactured exports, including processed foods, in early 1979. Exports of fruits and vegetables too, received rebates. Note further that the rebates were supposed to provide compensation for the payment of indirect taxes but there were numerous cases of over and under-com- pensation. (2) However, be.ween July 1978 and April 1979 exporters could retain 25 per cent of their foreign exchange earnings to finance the importation of inputs and this ratio was raised to 50 per cent in April 1979 and the uses of foreign exchange extended to the domestic suppliers of the ex- porters. 26  Table 3 REAL EXCHANGE RATES IN TURKEY, 1967-79 INDEX OF INDEX OF THE REAL EXCHANGE RATE-VIS-A.VIS INDEX OF RELATIVE PRICES RELATIVE PRICES THE CURRENCIES OF TURKEY'S EXCHANGE RATE THE EXCHANGE VIS.A-VIS Vis-MAVIS TURKEY'S PERIOD LIRA/DOLLAR RATE THE UNITED STATES TRADING PARTNERS THE U.S. DOLLAR TRADING PARTNERS A B A B A B A B 1967 9.000 63.6 69.2 68.2 . 81.4 80.3 91.9 93.3 78.1 79.2 1968 9.000 63.6 69.6 66,8 85.0 81.3 91.4 95.5' 74 8 78.2 1969 9.000 63 6 71.9 69.4 88.2 85.1 88.5 91.6 72 1 74.7 1970 11.500 81.3 74.1 73.9 87.9 87.7 109.7 1100 925 927 1971 14.917 105.4 83.1 83.2 953 954 1268 126.7 110.6 110.5 1972 14.150 10.0 93.8 93.9 100.9 101.3 106.6 106.5 99.1 98.7 1973 14.150 100.0 1000 100.0 100 0 100 0 100 0 100,0 100.0 100.0 1974 13.927 98.4 109 2 106.7 109.2 106.6 90.1 92.2 90.1 92.3 1975 14.442 1021 110.0 108.5 108.7 107.1 92.8 94.1 939 95,3 1976 16.053 113.4 121.6 121.4 125.6 125.3 933 934 903 90.5 1977 18.002 127.2 142.1 146.5 142.1 146.5 895 e.8s 89a5 86 8 01 17.002 120.2 131.2 133.8 133.2 135.8 91.6 89.8 90.2 88.5 02 17.675 124.9 135.8 136.4 137.6 138.4 92.0 91.6 90.8 90.2 03 17.889 126.4 142.5 146.0 141.9 145.5 88.7 88.6 89.1 88.9 04 19.443 137.4 160.9 165.4 156,9 161.3 85.4 83.1 87.6 85.2 1978 24.282 171.6 201.1 209.3 186.9 194.6 85.3 82.0 91.8 88.1 01 21.379 151.1 176.8 180.5 165,7 169 1 85.5 83.7 91.2 89.4 02 25,250 178.4 190.5 197.9 181.9 189.0 93.6 90.1 98 1 94.4 03 25 250 178.4 206.5 214.8 190.7 198.4 88.4 83.1 93 6 89.9 04 25.250 178.4 219.9 230.0 197.9 214 2 81 1 75.0 90 1 83.3 *1979 01 25.250 1784 239.7 259.5 216.5 234.4 74.4 68.7 82.4 76.1 02 31.020 219.2 279.5 3053 258.8 2825 784 718 84.8 77.6 As of June II 47.100 3329 293.2 320.2 272.2 297.- 113.5 104.0 122.3 112.0 September 1 47.100 332.9 321.0 350.6 300.8 328.5 103.7 95.0 110.7 101.3 II 47,100 332 9 , 330.3 364.3 309.5 341.3 100 3 91.4 107.6 97.5 Decemberl 47,100 3329 351.4 3838 332.4 3631 947 88.7 100.2 91.7 1I 47.100 3329 372.1 410.4 352.0 3882 895 81 1 946 858 SOURCES: IMF, Direction of Trade, Annual 1971-77. IMF, International Financial Statistics, various issues. Turkey June 1979 exchange rate: Official Gazette, June 11, 1979 NOTES: The index of the real exchange rate has been calculated by ad- justing an index ofl the nominal exchange rate for changes in wholesale prices at home and abroad. Calculations for Turkey's principal trading partners, covering 63.8 per cent of Turkish ex- ports and 67.6 per cent of Turkish imports in 1973, (the United States, Belgium, France, Germany, Italy, Netherlands, Switzer- land, and United Kingdom) have been made by weighting with the sum of exports and imports combined in the year 1973. The sources for Turkish wholesale price indices are: A: Business Research and Publicatipns Department of the De- partment of Commerce. B: Chamber of Commerce, Istanbul. 27  Starting in June 1979, the real exchange rate has been calculated under the assumption that the premium inclusive rate of 47.10 continues to apply in Turkey and that dollar exchange rates for other countries also remain the same. Price.indices for Turkey's trading partners have been extrapolated by using the average monthly inflation rate for the preceqing twelve months. In the case of Turkey, it has been assumed that the reported price in- crease of 7.8 per cent in April has been followed by increases of 6.0 per cent in May and in June. The projections for Sep- tember and December 1979 have been made under alternative assumptions as to the rate of inflation in Turkey. Variant I assumes increases in wholesale prices of 4 per cent, Variant II 5 per cent, per month. unavailable. Private importers thus had to obtain foreign exchange on the parallel market while, at the same time, they were obligated to make advance import deposits.(') The increased scarcity of foreign exchange is indicated by changes in the ratio of the parallel market rate to the official rate. Taking the price of gold ingots in Turkey as compared to that in the London market to be representative of this difference, the percentage ex- cess of the parallel market rate over the official exchange rate rose from 4 per cent in January 1975 to 18 per cent in December 1975, fell to 14 per cent in December 1976, and increased again to 32 per cent in De- ceSriber 1977 (TUSIAD, July 1978, p. 204). By March 1979, the difference between the two rates reached 100 per cent. Estimates on the importance of the parallel foreign exchange market vary. Transaction values in 1976 were reportedly estimated at $1.6 billion by the Istanbul Chamber of Commerce and $2.1 billion by the Turkish In- dustrialists and Businessmen's Association (Ciller, 1978, p. 16). And, with the increased unavailability of foreign exchange, the importance of the parallel exchange market has greatly increased in subsequent years. Apart from the level of the reol exchange rate and of the rates of protection exports and imports are affected by fluctuations in these rates. Quarterly data for the years 1977 and 1978 shown in Table 3 exhibit considerable variations in real exchange rates. The variations are even greater if we extend the time series backward to 1970 and forward to June 1979 (see Section II below). (') However, between July 1978 and April 1979 exporters could retain 25 per cent of their foreign exchange earnings to finance the importa- tion of inputs and this ratio was raised to 50 per cent in April 1979 and the uses of foreign exchange extended to the domestic suppliers of the exporters. 28  Fluctuations in real exchange rates associated with intermittent and unforeseen devaluations, as well as changes in rebate rates, create un- certainty for exporters as they cannot foresee future developments in the domestic currency equivalent of export receipts and in the relation- ship between their revenues and costs. Importers and import-substitut- ing industries protected by tariffs, too, are subject to uncertainty.(') And, uncertainty is created for the government itself as shifts take place bet- ween the official and the parallel foreign exchange markets in response to changes in the real exchange rate. Negative real interest rates, too, have adverse effects on economic activity.(2) Thus, there is evidence that the savings rate is positively cor- related with the real rate of interest in Turkey (Fry, 1978). Also, in pro- viding inducements to channel savings into real estate, gold, and foreign assets, highly negative real interest rates reduced the availability of sav- ings for productive uses and contributed to Turkey's balance of pay- ments deficit. Finally, the allocation of savings available for productive uses is subject to inefficiencies, since interest rates cannot fulfill their role as a rationing device. The described factors adversely affected economic activity in Tur- key, leading to clisruptions in work and underutilized capacity. According to a recent w-rviy pf the Jstanbul ChamtWer of Industry, Overage capacity uOiiizatioi in nanufacturing vyas 5S.8 por cent in 1978, with the rqtios be- ing much higher than the average in industries based on domestic inputs, such as mining (80.0 per cent), forestry products (71.9 per cent) and food prQcessing (76.3 per cent), and the ceramics industry (65,2 per cent) and below the average in most industries using imported inputs (1979, p. 5). At the same time, these figures tend to understate the extent of the un- derutilization of capacity as they take one-shift operations as the norm, Excluding industries based on domestic inputs, over 50 per cent of respondents stated that the shortage of foreign exchange was the prin- cipal factor resulting in low capacity utilization. Another one-fifth of res- pondents indicated insufficient demand, mostly in regard to intermediate products, which is largely related to foreign exchange shortages in user (') A case in points is that of an importer who contracted to sell tractors at a given price only to find that the June 1979 devaluation raised its cost by nearly three-fourths. (2) Between April 1978 and May 1979, interest rates in Turkey were in the 8 to 20.per c@nt range for savings deposits and 16 per cent on non- preferential loans as compared to increases in consumer and wholesale prices of 50-60 per cent a year. At the same time, the transciction tax and various bank charges may have raised the lending rate to 23-25 per cent. 29  industries (pp. 9-10). Finally, one-tenth of the respondents named local material shortages and another one-tenth power outage and shortages as the principal cause of the underutilization of capacity. In the industries in question, shortages of domestic materials were in large part related to the foreign exchange difficulties. This is the case even more in regard to energy outages and shortages (pp. 12-13). While comparable data for earlier periods are not available, it ap- pears that capacity utilization levels in 1978 were substantially lower than on the occasion of previous foreign exchange crises. This fact, together with the existence of higher open and disguised unemployment and the much larger and more enduring current account balance deficits,(') in- dicate that the present c,isis is much more serious than the previous ones. 2) BRAZIL AND TURKEY: COMPARISONS AND POLICY IMPLICA- TIONS - THE CONSEQUENCES OF IMPORT SUBSTITUTION POLICIES(2) Turkey offers a number of similarities with the economic situation in Brazil prior to the reforms of the mid-sixties. Both countries moved be- yond the first (easy)) stage of import substitution, involving the replace- ment of the imports of nondurable consumer goods and their inputs by domestic production, and extended this to intermediate products, durable consumer goods, and machinery that tend to be more capital-intensive, require large-scale operations as well as the availability of parts, com- ponents and accessories at a low cost. Import substitution was carried out behind high protection, entailing considerable discrimination against primary activities and against manu- factured exports. The two countries also maintained overvalued exchange rates and adjusted those rates only intermittently, thereby creating an additional burden for exporters. At the same time, the use of tariffs and quantitative controls gave rise to considerable differences in incentives among import-substituting activities as well as among export activities that used protected domestic inputs. Finally, both countries had negative real interest rates and fragmented credit markets, with preferred borro- wers (usually in import-substitution industries) benefiting at the expense of others. (') The current account deficit equals the net inflow of capital, in- cluding changes in foreign exchange reserves. During the eighteen year period preceding the oil crisis, for which data are available, it exceeded 2.0 per cent of GNP only in 1963. (2) The discussion of Brazil in this section draws on the author's aln- centive Policies in Brazil)) (1979 b). 30  The effects of the policies followed in the two countries were also rather similar. After rapid expansion, they experienced a decline in net foreign exchange savings in import substituting activities, a rise in ca- pital-output ratios, and a decrease in rates of export expansion, with adverse repercussions for economic growth. And while in Turkey the ef- fects of workers' remittances and foreign borrowing delayed the decline of the rate of GNP growth until 1977, as noted above growth would have been more rapid if different policies were followed. Certain differences in the economic situation prevailing in the two countries in the mid-sixties and in the late seventies, respectively, should however be noted. Turkey enjoys the benefits of proximity to European and Middle Eastern markets and preferential access to the EEC whereas participation in LAFTA provides limited benefits in Brazil. Furthermore, Turkey has gone less far in import substitution than Brazil had by the mid-sixties, although the possibilities for efficient import substitution, too, were greater in the larger Brazilian market. At the same time, the share of exports in GNP was by one-third lower in Turkey in 1978 than it was in 1966 in Brazil. Last but not least, the present economic situation in Turkey may be judged more serious than that of Brazil in the mid-sixties. The foreign ex- change shortage, and the resulting distortions in relative prices, are con- siderably greater as is the balance of payments deficit and foreign in- debtedness, in particular short-term debt. And, unlike Brazil, Turkey has a large state enterprise sector which, after earlier contributions to indust- rial development, suffers from low efficiency levels and an antiquated system of management. POLICY REFORM AND ITS EFFECTS IN BRAZIL Notwithstanding these differences, the similarities in the policies followed and in their economic effects lend interest to the policy reform instituted in Brazil in the mid-sixties, and the impact of these measures on the Brazilian economy. The policy measures in question included: a) A substantial devaluation, accompanied by subsidies to exports and reductions in the level of import protection, both directly through the lowering of tariffs and the liberalization of imports, and indirectly through a slower rate of devaluation than what would have been necessary in the absence of export expansion; b) Incentives to the development of new agricultural exports; 31  c) The adoption of a ((sliding peg)>, with monthly devaluations under- taken in line with domestic inflation; d) The establishment of positive real interest rates and improve- ments in credit markets; e) Reductions in the rate of inflation through reductions in the rate of growth of government expenditures and the money supply; and f) Increased reliance on foreign direct and portfolio investment. The results were little short of spectacular. With the dollar value of manufactured exports rising at an average annual rate of 38.5 per cent between 1966 and 1973, the share of these exports in manufactured out- put increased from 1.3 to 4.4 per cent. In the same period, traditional primary exports grew 7.6 per cent a year while non-traditional primary exports increased 26.8 per cent, leading to a 19.9 per cent rise in total exports. With the increased inflow of forei'n capital, chiefly in the form of direct investment, the dollar value of imports increased even faster (24.5 per cent a year). As a result of these changes, the share of exports in GNP rose from 6.0 per cent in 1966 to 7.7 per cent in 1973 while that of imports increased from 5.2 per cent to 8.7 per cent. With the improved allocation of economic resources and increased capacity utilization, the incremental capital-output ratio declined from 3.8 in 1960-66 to 2.1 in 1966-73 in Brazil. Also, the share of domestic savings in GDP increased frcm 22.2 per cent in 1966 to 25.4 per cent in 1973 while that of foreign savings rose from 0.1 to 2.1 per cent. Rapid export expansion, increases in domestic savings, and the in- flow of foreign capital all contributed to the acceleration of economic growth in Brazil. Following the decline in the average annual rate of growth cf GNP from 6.3 per cent in 1953-60 to 4.1 per cent in 1960-66, the growth rate reached 10.1 per cent in 1966-73. During the same period, agricultural production rose by 5.3 per cent, and manufacturing produc- tioin 11.7 per cent, a year. The success of the policy reform in Brazil provides lessons for Tur- key. At the same time, given the differences in the economic situation, in the institutional framework, and in attitudes towards particular policy measures, it cannot be suggested that Turkey applies the same policies as Brazil did in the mid-sixties. These considerations are reflected in the policy recommendations made in Section III of the paper following an evaluation of recent policy changes in Turkey. 32  POLICY CHANGES IN TURKEY, MARCH-JUNE 1979 A variety.of measures were taken by the government between March and June 1979 in order to improve the economic situation in Turkey. The measures pertain to exchange rates, prices, the state economic enter- prises, the budget deficit, and interest rates, They will be considered in the following. The Turkish lira was devalued by 6.0 per cent on April 5,1979, bring- ing the rate to TL 26.50 per U.S. dollar, with an additional premium pro- vided for tourist receipts and workers' remittances on a degressive scale. Furthermore, exporters of manufactured goods were permitted to retain one-half of their export receipts for their own use or for transfer to their direct and indirect suppliers who are in the possession of import licenses. In turn, on June 11, an exchange rate was set of TL 35.00 to the dollar, with a premium of TL 12.10 on all sales of foreign currency except for traditional agricultural exports that are subject to support prices de- termined by the government,(') and on all purchases of foreign currency except for the importation of crude oil and its derivatives and the raw ma- terials used in producing fertilizer. At the same time, however, export rebate rates for manufactured exports were reduced by 5 to 15 percentage points and rebates on fruit and vegetable exports were abolished. Exceptions made for agricultural export appear to reflect the desire to avoid creating large rents to producers that would otherwise result as the expansion of agricultural output takes time, while setting a lower ex- change rate for imported inputs in fertilizer production so as to prevent increasing production costs in agriculture. However, one may query the decision of applying the TL 35,00 rate to the importation of petroleum and its derivatives as the adoption of a higher rate would be conducive to energy savings. In turn, the application of the TL 47.10 rate to all other .oreign ex- change transactions is to be welcomed. The adoption of this rate has given a boost to manufactured exports by increasing their profitability, when the large -c,,cess capacity further augments the possibilities for ex- port expar,,ion. Furthermore, import savings may be realized as the privi- leged rcElpients of import licenses respond to the higher cost of imports. (') Seedless raisins, hazelnuts, pistachio, dried figs, cotton, rice in husk, lentil, mohair, livestock, oil cake, and molasses. 33  The experience of May 1979 also indicates that immigrants' remit- tances respond positively to the exchange rate. Finally, the higher ex- change rate will permit the tourist sector to reduce prices in terms of foreign currency, thereby contributing to increased demand on the part of foreign tourists, given the high substitution elasticity in tourism as between different countries.(') Table 3 provides information on real exchange rates in Turkey follow- ing the June 11 devaluation, by taking TL 47.10 as the relevant nominal exchange rate. Compared to the situation existing in 1973, the results show a depreciation of the real exchange rate by 4 to 14 per cent vis-a-vis the U.S. dollar and by 12 to 22 per cent vis-a-vis the currencies of Turkey's main trading partners. These comparisons do not, however, take account of the depreciation necessary to offset the deterioration of the terms of trade by 33 per cent between 1973 and 1978 or that due to the rise of oil prices in 1979. And, an additional depreciation would be warranted if one chose as benchmark the situation existing in 1971 when the exchange rate was particularly favorable to exports. Furthermore, if recent trends continue, inflationary developments will soon undo the beneficial effects of the devaluation. This could be avoided if the government adopted a sliding peg on the example of Brazil and other Latin American countries, with further devaluations occurring par! passu with changes in relative prices. However, decisions have ap- parently been taken to Maintain the adjustable peg system. The last four rows of Table 3 show real exchange rates calculated on the assumption that differential inflationary trends in Turkey and in its major trading partners will continue and that exchange rates are maintained at existing levels. It appears that, even under favorqble as- sumptions as regards the rate of inflation in Turkey, the real exchange rate would return to nearly its 1973 level by September 1979 and it would fall below this level by December 1979. We will return to the policy implica- tions of these results in Section 'II below. The calculations shown under Variants I and 11 are based on the assumptions that wholesale prices in Turkey would rise by 4 and 5 per (') Estimates of this elasticity for developing countries range bet- ween 3 and 7; the latter figure has been obtained for the devaluation of 1959 in Spain that was at the time in a similar situation as Turkey is at present as far as tourism is concerned (Gerakis, 1965). 34  cent a month, respectively, between June and December 1979, represent- ing annual rates of 60 and 80 per cent under the two variants. By com- parison, wholesale prices rose at an average rate of 4.9 per cent per month in the first quarter of 1979 and approximately 6.5 per cent per month in the second quarter, when the prices of a variety of commodi- ties and services produced by state economic enterprises (SEEs) were substantially increased. At the same time, keepinig inflation below 5 per cent a month would require reducing the deficit of the public sector and slowing down increases in wages. Steps to reduce the public sector deficit were taken in the Spring of 1979, when the prices of SEE products were raised and ceilings were imposed on the expansion of employment in the public sector, on the deficit of the SEEs, on Central Bank credits to the public sector, and on the net domestic assets of the Central Bank. The imposition of these ceilings should have favorable effects on inflation by reducing the rate of money creation that has been associated with the public sector deficit in the past. Increases in minimum wages in the nonagricultural sector, from TL 110 to TL 180 per day in May 1979, will however add to cost- inflationary pressures, even though practically the entire increase rep- resents compensation for the rise of consumer prices since January- February 1978 when minimum wages were last determined. Also, recent wage settlements in the private sector represent annual increases of about 100 per cent and in June 1978 minimum wages for agricultural and forestry workers were raised by 78 per cent. Finally, interest rates on savings deposits were raised in May 1979, with the new rates ranging from 8 per cent on savings deposits between 3 and 6 months to 24 per cent on deposits between 3 and 4 years, and a premium of 10 percentage points provided on the repatriated savings of migrant workers. At the same time, interest rates on nonpreferential loans were raised to 20 per cent. Notwithstanding increases in nominal interest rates, real rates of interest remain negative by a substantial margin. Even assuming a rate oTf inflation of 4 per cent a month, real interest rates are between -36 and -52 per cent on deposits by domestic savers, -26 and -42 per cent on deposits of migrant workers, and -40 per cent on nonpreferen- tial loans, Apart from the situation existing prior to the policy reform in Brazil, there are few examples of such high levels of negative real in- terest rates. 35  3. POLICIES FOR THE FUTURE Towards Greater Export Orienitation We have seen that the continuation of a strategy of import substitu- tion entailed increasing economic costs and only through a combination of workers' remittances and foreign borrowing could Turkey avoid a decline in the rate of economic growth until 1977. Foreign borrowing also made it possible to postpone effecting the (transfer)) implicit in the deterioration of the terms of trade that was largely due to the quadrupl- ing of oil prices in late 1973. Growing foreign indebtedness, however, gave rise to a substantial debt service burden while limiting the possibilities for additional borrow- ing. As a result, economic growth has decelerated, with per capita in- comes remaining practically constant in 1978, and a decline expected in 1979, as compared to average increases of 4.8 per cent in the preceding three years. Turkey thus faces the short-term problem of servicirng its foreign debt and effecting the transfer implicit in the deterioration of the terms of trade through increased exports and reduced imports as well as the long-term problem of adopting a development strategy for sustained economic growth. At Turkey's present stage of economic development, the growth ob- jective would be served by a strategy of increased export orientation that permits the efficient use of economic resources. This strategy would also serve the short-term objective of improving the balance of payments, so that the use of appropriate policy measures would permit avoiding a conflict between short-term and long-term objectives. Increased export orientation, in turn, necessitates placing greater reliance on market forces in both the private and the public sectors. While central decisions could be appropriately taken in regard to a few large import-substituting investments, exporting requires decisions by individual firms that have to continuously adapt to the needs of foreign markets. At the same time, in order to ensure that the decisions taken by firms correspond to national economic interests, appropriate incen- tives need to be provided. Pursuing these objectives calls for adopting a variety of policy measures that support each other. The measures in question include the provision of appropriate market signals by adopting and maintaining realistic exchange rates and interest rates and by reforming the system of protection, as well as changes' in institutional arrangements by re-allb- 36  cating investment funds towards export industries, increasing the effici- ency of financial markets, simplifying administrative procedures, and reforming the state enterprise sector. These policy measures will be considered in the following. Exchange Rate Policy The devaluation of June 1979 has improved the competitive position of Turkish enterprises to a considerable extent. At the same time, as shown above, competitiveness would decline again over the next few months if the new exchange rate was maintained unchanged in the face of differential inflationary trends at home and abroad. Furthermore, the objectives of reducing uncertainty for exporters and importers and of avoiding economic instability would call for changes in exchange rates to take place in small steps, so as to maintain real exchange rates constant. While the sliding peg found easy acceptance in Brazil, the exchange rate has become a political issue in Turkey, with every devaluation seen as requiring political courage. It may be difficult, therefore, to envisage the adoption of a sliding peg for the present and, even if it were adopted, there would be a danger that the government may not carry out the devaluation that would be warranted by relative price changes. In this situation, there is a virtue in automaticity. Correspondingly, it is suggested here that reliance be based on the parallel exchange market for all transactions to which the TL 47.10 rate applies at present. This may be accomplished by raising the rate of foreign exchange reten- tion for manufactured exports to 100 per cent, allowing for the full trans- ferability of these receipts between exporters and importers, and trans- ferring to the parallel market tourist receipts and expenditures, immig- rants' remittances, and all nongovernmental imports that are presently subject to the TL 47.10 rate.(') The proposed transfer of transactions to the parallel foreign exchange market would avoid the re-emergence of the overvaluation of the Turkish lira and fluctuations in its real value, thereby reducing uncertainty for exporters and importers and contributing to stable economic growth. (') Government imports, chiefly petroleum and military equipment, could be financed from the proceeds of agricultural exports and foreign loans. However, the objective of energy conservation would necessitate imposing additional taxes on fuel. 37  Furthermore, in channeling import transactions by private and by public firms to the parallel exchange market, one would capture the scarcity rent on foreign exchange that presently accrues to the recipients of im- port licenses cind increase the integration of state enterprises in the market economy. This is highly desirable since, in the absence of app- ropriate price signils, the recent imposition of export targets on state enterprises may be met at a high cost to the national economy. At the same time, the inflationary effects of the proposed measures should not be exaggerated. With the parallel market rate reportedly used by private and public importers in pricing goods for sale in the domestic market, capturing the scarcity rent in these transactions would not ge- nerally result in higher prices. And while the prices of manufacturing and nontraditional agricultural exports would rise, exporters may not adjust their prices to the full extent of the change in the exchange rate in order to improve their competitive position in foreign markets. But, at any rate, increases in export prices would be necessary in order to provide incen- tives to expand exports. The above recommendations are given urgency by the rise of oil prices in the first half of 1979. Given the expected geographical com- position of Turkish oil imports,(') increases in oil prices may average about 60 per cent for Turkey. At 1978 import levels, the additional cost would be about three-quarter billion dollars, equal to over one-fourth of the value of merchandise exports and 2 per cent of GNP.(2) Thus, while Turkey has relied on foreign loans to postpone effecting the ((transfer)) implicit in the deterioration of the terms of trade after 1973, it now has another ((transfer)) to cope with. At the same time, foreign borrowing has given rise to substantial financial obligations for the repayment of short- term loans and the serVicing of the medium-and long-term debt.(3) Correspondingly, while an immediate extension of the parallel ex- change market to traditional agricultural exports does not appear desi- (') While traditionally the principal sources of Turkey's oil imports were Iraq, Libya, and Iran, it is planned to increase imports from Saudi Arabia in the future. (2) The calcUlations do not take account of the possibility that oil- producing countries would limit the price increases applicable to Turkey. They further assume unchanged transportation costs, (3) According to estimates of the Ministry of Finance, about $ 10 bil- lion of repayment, amortization, and interest are due during the period of the Fourth Five-Year Plan (1979-83), of which nearly one-half represents the repayment of short-term debt. 38  rable in order to avoid prov;<ding rents to producers, provisions would need to be made for the subsequent integration of the two exchange mGrkets. with commensurate changbj in support prices, so as to provide adequate incentives for the expansion of agricultural exports.(') This conclusion is strengthened'if we consider that large differences between the parallel market rate and the official exchange rate would create ex- cessive incentives for exporting agricultural products in proces-ed rather than in crude form. The proposed measures aim at effecting the trans'er implicit in the deterioration of the terms of trade through increased production for ex- port. Given the magnitude of the task involved, it would further be ne- cessary to limit the rise of domestic consumption by reducing the public sector deficit and moderating wage increases. As a result of these ac- tions, inflationary pressures too would mocieratej. Reforming the System f Protection The propc,:d extention of the use of the parallel exchange market would represent a first step in the reform of the system of protection. In order to irr,,fove the efficiency of resource allocation for sustained economic grow i, it WOL,Jd need to be followed by furthier steps. These should aim , ,aucing dciscrimination against primary and manufactured exports and rationalizing tariffs and subsidies. Except for hazelnujts in which Turkey has a quasi-monopoly position, it would be desirable to extend the application of the parallel market rate to all primary exports, with corresponding increases in support prices. Increases in agriculturcl product prices, in turn, should be accompanied by the elimination of various input and credit subsidies to agriculture. It would further be desirable to increase agricultural taxation, preferably in the form of a land tax. Such a tax may be introduced as a quid pro quo for raising the exchange rate and agricultuial support prices. Discrimination against the exports of manufactured goods vis-a-vis import substitution is associated with the protection of domestic sales in the form of tariffs, quota restrictions, and import licensing. Tariffs (cus- toms duties) on manufactured goods were generally in the 30-60 per cent 1) For the conditions of this integration, see the following diiscussion of the system of protection. 39  range in 1973 (Ciller, 1968, Table 2.2), to which a customs surcharge (15 pier cent of customs duty), stamp duty (9.0 - 9.5 per cent of cif value) and pier duty (5 per cent of the cif value of imports, customs duty, surcharge, and clearance expenses, combined) should be added (TUSiAD, July 1978, p. 133) At the same tisne, effective rates of protec- tion, i.e. the protection of value added in the production process, are substantially higher than nominal rates, given that various inputs, e.g. iron and steel, enter duty free (Table 2.6). Customs duties and other charges changed little since 1973. However, with increased foreign exchange stringency, import restrictions have come to be increasingly applied, thus raising the protection of domestic industry and increasing discrimination against exports. Changes in this situation would be necessary, lest exports do not receive adequate incentives. The bias against manufactured exports may be reduced by offsetting the taxes and tariffs that burden exports and by providing direct or in- direct exports subsidies. Firstly, on the example of other European count- ries, one would have to ensure that export rebates provide full compensa- tion for indirect taxes paid at the last and at all previous stages of the manufacturing process. Exporters should also receive rebates for duties paid on direct as well as indirect imported inputs that is not considered a subsidy under GATT rules, Furthermore, the expansion of exports would require the extension of the existing export credit scheme, the implemen- tation of the proposed export insurance scheme and, as discussed below, the increased availability of credits for production and investment in ex- port industries. Additional incentives to exports may be provided by way of a direct subsidy or indirectly through tariff reductions and the liberalization of im- ports that reduces incentives to import substitution. A combination of the two alternatives would be appropriate in the case of Turkey. For one thing, subsidies to manufactured exports would provide advantages to manufac- turing industries that is desirable at Turkey's present stage of economic development. For another thing, reductions in import prc-Iction would avoid excessive discrimination against primary activities and may be linked to a reform of the system of tariffs and import restrictions with a view to lessining existing disparities in rates of import protection. Reducing discrimination against exports would have beneficial effects on the national economy as it would permit earning foreign exchange at a lower domestic cost than foreign exchange is saved through import 40  substitution and it would tend to increase employment. Nevertheless, the proposed changes in the system of protection may be postponed until the effects of the extension of the parallel exchange market and the full rebating of indirect taxes and tariffs on export expansion become appa- rent. The Allocation of Investment Funds Providing incentives to exports is a necessary but not a sufficient condition of export expansion. Channeling resources into export activi- ties would also require the availability of credits for export production as well as for investment in export activities. This may be accomplished by modifying the allocation of investment funds, increasing their volume, and improving the efficiency of financial markets. Increasing the availability of investment funds for manufactured ex- ports would necessitate modifying the investment allocation for the ma- nufacturing sector in the Fourtlh Five Year Plan (1978-83), which calls for a 71.5 per cent investment share for intermediate products as against a 15.8 per cent share for investment goods and durable consumer goods and a 12.7 per cent share for nondurable consumer goods. Within inter- mediate products, petrochemicals and petroleum products, with a planned investment allocation of 13.6 per cent and primary metals, with an alloca- tion of 32.3 per cent (iron and steel, 26.0 per cent and nonferrous metals, 6.3 per cent), would receive the bulk of the total.(') In turn, among non- durable consumer goods, food processing would receive an allocation of 4.9 per cent, and textiles and clothing 5.7 per cent (Turkey's Fourth Five Year Plan, 1978, Table 142). This allocation appears to cQnflict with the projected annual average rate of growth of 25 per cent for manufactured exports (Table 141). Also, investments in intermediate products have a long gestation period while the existing foreign -exchange stringency in Turkey puts a premium on quick-yielding investments, And, in a situation of capital scarcity observed in Turkey, industries producing intermediate products, on the average, have capital-output and capital-labor ratios several times higher than either nondurable consumer goods or investment goods and durable con- sumer goods. (l) Other commidities in the intermediate products category include forest products, paper and printing, leather, rubber, plastics, chemicals, fertilizers, cement and its products, glass, and ceramics. 41  obtained for a particular year that relate installed capital to actual em- According to the data of the Fourth Five Year Plan, incremental ca- pital-output ratios for the 1978-83 period are 54 per cent higher in indust- i.es producing intermediate products than the overall average for the manufacturing sector. The difference is especially large for iron and steel (199 per cent) and for nonferrous metals (55 per cent); it is lower for petro-chemicals and petroleum products (10 per cent) by reason of the fact that investments undertaken during the previous period will come on stream between 1978 and 19S3. By contrast, the incremental capital- output ratio is only 43 per cent of the overall average in the case of nondurable consumer goods (of which, 34 per cent for food processing and 51 per cent for textiles and clothing) and 65 per cent for investment goods and durable consumer goods (Tables 139 and 142), Data on incremental capital-output ratios are affected by the length of the period of investment. Such is not the case for capital-labor ratios ployment. Data obtained for the United States indicate that capital re- quirements per job in petroleum refining and products are 6.1 times the overall average for the manufacturing sector while the corresponding ratio for primary metals is 1.6. By contrast, capital requirements per job are 46 per cent of the overall average for textiles, 10 per cent for cloth- ing, 49 per cent on nonelectrical machinery, 35 per cent on electrical machinery, and 57 per cent for transportation equipment; comparable data for food processing are not available. (Balassa, 1979a, Table 6). High capital-output and capital-labor ratios in petroleum products and primary metals, together with the expected oversupply of petroleum products in the world market and the poor quality of iron ore and cok- ing coal in Turkey, make a prima facie case against the further expan- sion of these industries. This conclusion does not apply to forest pro- ducts, leather, and ceramics that have relatively low capital-output and capital-labor ratios and make use of domestic materials. Nondurable consumer goods industries rely largely on domestic materials and utilize Turkey's abundant and relatively cheap manpower. Export possibilities are especially good in nearby Arab markets that presently account for less than 10 per cent of Turkish exports. And while the exports of textiles and clothing and of some simple processed food are subject to limitations in the European Common Market, Turkey be- nefits from its preferential access to the EEC in these commodities and does not encounter barriers for most food preparations. 42  The availability of low-cost skilled and semi-skilled labor also bes- tows advantages on Turkey in the production of investment goods and durable consumer goods that are relatively labor-intensive. Also, these goods are rarely subject to quantitative restrictions and bear low tariffs in the Common Market and in other developed countries. The expansion of industries producing investment goods and durable consumer goods and their inputs may be regarded as the next step in Turkey's industrial development. At the same time, in view of the import- ance of economies of scale in these industries, the parallel expansion of production for domestic and for export markets would be desirable. The above considerations point to the need for increasing the amount of investment funds available tQ consumer goods and investment goods industries an-d reducing the investment allocation of industries produc- ing intermediate products during the Fourth Plan period. The time thereby gained may be used for the economic project evaluation of the proposed investments in industries producing intermediate goods. Such evaluations should be made in terms of world market prices that represent the alter- natives available to Turkey. Furthermore, the planned 1.2 per cent share of tourism in total in- vestment under the Fourth Five Year Plan (Table 123) is not commen- surate with Turkey's possibilities for expanding tourism. Given its favo- rable climate, the natural beauty of the country, and the presence of a large number of archeological sites, tourism has considerable potential in TuJkey and could earn foreign exchange at a relatively low cost in terms of domestic resources while contributing to regional development. The implementation of these recommendations would entail increas- ing the investment share of the private, as compared to the public, sector. Such a shift would occur as lowering investment allocations to state en- terprises would increase the availability of investment funds to the pri- vate sector through reduced claims on private savings and a lower ((inf- lation tax)) that has adversely affected the real value of these savings in the past.(') (') According to estimates of the Ministry of Finance, in 1978 the fi- nancing requirements of state economic enterprises amounted to TL 90 billion, of which TL 57 billion represented investment expenditures and TL 33 billion the operating deficit. In the same year, the total financing requirements of the public sector were TL 106 billion, i.e. over 8 per cent of GNP. About one-sixth of the total was financed through the issue of debt obligations, reportedly exceeding private corporate issues ten times (Statement by Mehmet Gun Calika in The Wall Street Journal, May 31, 1979), one-sixth by foreign borrowing, three-fifths by borrowing from the Central Bank, and the remainder from other domestic sources. 43  Increased investments by the private sector would also be desirable for the pursuit of an export-orie-nted strategy, since private firms may better respond to the needs of foreign markets and take the risks neces- sary for successful exporting. In turn, a cutback in the public investment program in manufacturing would permit limiting the inflationary effects of deficit financing that would otherwise likely to continue, given the over- ambitious targets for public savings in the Fourth Five Year Plan. Considering further the substantial delays experienced by state en- terprises in implementing investment projects, a moratorium on new public investment projects in manufacturing may seem desirable. This would permit concentrating attention on the completion of ongoing pro- jects and undertaking an economic evaluation of the proposed new in- vestment projects as suggested above. But, ongoing investment projects in the public sectors would also need to be re-evaluated as the invest- ment expenditures necessary for completing them (TL 924 million) ex- ceed the entire allocation of the Fourth Five Year Plan to public invest- ments (TL 898 million). The differences are even larger for public in- vestments in the manufacturing sector that equal TL 199 million in the Plan as compared to TL 257 million of the investment needs of ongoing public projects in this sector (Tables 25, 122, and 123). At the same time, a number of these investments are at an early stage that would permit their discontinuation or postponement. The Treatment of Foreign Investment We have noted the need for the parallel expansion of the domestic and export sales of industries producing capital goods and durable qon- sumer goods. This is because, in the presence of substantial economies of scale, production for domestic markets alone would lead to small-scale manufacture at a high cost. Such an outcome is in fact observed in the automobile industry, where the average number of cars produced per plant in Turkey is only three per cent of that in the European Common Market (Manisali, 1979, p. 105). The above considerations also apply to the production of parts, com- ponents, and accesories. Government regulations have required firms producing automobiles, buses, and trucks to raise the ((national content)) of their production to 90 per cent, 80 per cent, and 75 per cent, respec tively, by 1978. While these targets have apparently been reached onlr in the case of bus monujfacture, national content is not likely to havo 44  surpassed 70 per cent in automobiles and 60 per cent in trucks (Ibid, p. 96). This, in turn, has entailed the small-scale production of parts, com- ponents, and accessories, with consequent high costs. A more appropriate solution is to concentrate on the production of of selected parts, components, and accessories for domestic use and for export and to import others. In this way, the same amount of foreign ex- change may be earned at much lower domestic resource costs since eco- nomies of scale can be exploited. Such an alternative also commends itself if the provisions of Turkey's association agreement with the Euro- pean Common Market are to be implemented. In the transport equipment industry, aircraft provides another ex- ample of the need for Turkey's participation in the international division of the production process. Rather than the local production of aircraft, envisaged in the Annual Programme for 1979, Turkey should carve out for itself a share in the huge world market for aircraft parts and com- ponents where duty free entry will be ensured under the Tokyo round of negotiations. Participation in the international division of the production process in transport equipment would permit Turkey to concentrate on the pro- duction of relatively labor-intensive parts, components, and accessories, where it has a comparative advantage. Similar considerations apply to the metal working, machine tool, and electrical and non electrical mac- hinery industries. At the same time, subcontracting arrangements in these industries may require the participation of foreign capital. Foreign par- ticipation may also contribute to the export-oriented production of ca- pital goods that are subject to rapid changes in technology. In this con- nection, reference may be made to the example of Brazil where foreign firms have spearheaded the expansion of the exports of investment goods. Foreign capital accounts for a small proportion of total industrial in- vestment in Turkey. On December 31, 1977, the total amount of foreign in- vestment was TL 2,2 billion as compared to paid-up capital of TL 28.4 billion in the hundred largest industrial corporations in 1976 (TUSiAD, July 1978, Tables 21 and 22). The differences would be substantially larger if comparison was made with total capital stock in the manufucturing sector, for which data are not available. In fact, it has been reported that foreign capital accounts for only 3 per cent of value added in manufac- turing. 45  At the some time, the slhare of foreign capitol in manufacturing in- vestment in Turkey is on the decline. Thus, athere are no recent examples of foreign companies receiving incentives, and there has been little fore- ign investment activity in Turkey in the past few years)) (Ibid. p. 26). These results reflect the negative attitude towards foreign investment and the bureaucratic difficulties experienced in the course of the approval pro- cess. In 1976, for example, while foreign investors made requests for the expansion of raooacity in six cases and for new investment in four cases, altogether eight approvals were issued and there were eight can- A more positive attitude towards foreign investment has been ex- cellations in the same year (ibid., p. 47). pressed by the Ecevit govertniment. However, under the 1978 decree the foreign investor has to take an obligation to increase the national con- tent of its production and to establish ((research and development cen- tres>) for the advancement of technology. Also, with the exception of in- vestments undertaken exclusively for export, the share of foreign capi- tal may not exceed 50 per cent. These requirements appear overly restrictive. As noted above, in- creasing national content leads to higher production costs, thus reduc- ing the competitiveness of Turkish industry. Also, greater competition at home and abroad is likely to give more inducement to technological im- provements than the formal requirement for establishing ((research and development centres)). Finally, making exceptions to the 50 per cent rule is of particular importance in regard to the investments necessary for the exploitation of Turkey's potential in food processing where toreign marketing is of crucial importance; for the establishment of plants to cater to Middle Eastern markets where delays may lead to the loss of these markets; and for the expansion of tourism where foreign expertise in construction, management, and marketing is needed. These conclusions are strengthened if we consider that, apart from the apport of technology and marketing know-how, foreign direct invest- ment would add the availability of foreign exchange in Turkey that is necessary to provide a ((cushion)) during the transitional period as noted below. It would further add to the availability of investible funds that would permit increased output and employment in Turkey. INTEREST RATES AND CAPITAL MARKETS While foreign investment will add to the availability of investible funds, further efforts would need to be made to increase public and pri- 46  vate savings in Turkey. Public savings would rise as a result of measures taken to reduce the deficit of the public sector. In turn, the generation of private savings would require changes in interest rate policy. Recent increases in interest rates are welcome but they have not fully ccmpensated for the acceleration of inflation. Correspondingly, in- terest rates are not sufficiently attractive to savers, in particular for longer maturities, severely limiting the availability of long-term finance to private firms. Also, the maximum interest rate on long-term loans by TSKB, the Industrial Development Bank of Turkey, is above the rate it has to pay for funds of similar maturity. Correspondingly, credit markets and institutions cannot appropriately fulfil their function of generating funds for new investments. In fact, in actual practice, private'borrowers are limited to loans of two year matu- rity. Furthermore, the increase of the withholding tax on dividends has led to a practical halt of stock market transactions since April 1978. In order to increase savings, to channel savings to productive in- vestments, and to ensure stability in the growth process, it would be necessary to raise real interest rates in Turkey and to avoid fluctuations in these rates. In this connection, reference may be made to Fry's esti- mates cited earlier that show a positive correlation between real interest rates and private savings' in Turkey. Similar results are observed in recent years when the decline in real interest rates has been associated with a fall in the marginal saving ratio that turned negative as inflation accelerated in 1978 (TOSiAD, July 1979, Page 104). Also, dramati6 in- creases in private savings occured in several Asian countries that have established positive real interest rates (Abe et al. 1977). Set'ing hIgher real interest rates also improves the allocation of investment funds as the interest rate can again serie as a rationing de- vice. As McKinnon has noted (1973) negative real interest rates lead to the misallocation of investment funds, in part because producers prefer low-return self-investment to negative returns on savings and in part because arbitrariness is introduced as the government and banks make choices among would-be borrowers, whose claims exceed the amount of available investment funds. Last but not least, higher real interest rates would have beneficial effects on Turkey's balance of payments by reduc- ing the desirability of gold purchases, which reportedly amounted to $600 million in 1978, diminishing the profitability of investment abroad. 47  As a first step, one may abolish interest rate ceilings on medium-and long-term obligations and loans. Consideration may also be given to the indexing of financial obligations that have been used to good effect in Brazil. Furthermore, there would be need for taking measures to encou- rage the development of bond and stock markets in Turkey. REFORMING ADMINISTRATIVE PROCEDURES Administrative procedures are notoriously inefficient in Turkey and would need to be simplified in order to ensure the success of an export- oriented strategy. Increasing foreign exchange retention to 100 per cent in regard to manufactured exports would serve this objective as one would avoid the need for fore;gn exchange allocation to purchase im- ported inputs, which is a time-consuming and uncertaih process. Thus, in the first quarter of 1979, exporters of manufactured goods requested foreign exchange allocation of $14.4 million for projecte,d exports of $176 million and received $9.8 million. It would further be desirable to abolish export licensing that has pur- portedly been used to provide for the needs of domestic users (cement) or to encourage exportation in processed form (synthetic fibers). Such interventions create uncertainty for the exporter, may lead to the loss of export markets, and unduly interfere with the operation of domestic markets. Also, there would be need for streamlining the system of investment incentives, which take the form of tax exemptions, accierated deprecia- tion, exemption from or reductions in customs duties, interest rate reba- tes, and permission to use external credits. While since 1973 the invest- ment incentives avoilable to sectors that meet certain criteria have been specified in the so-called promotion tables, discretionary decision-mak- ing continues in the selection of the recipients and in the choice and the extent of the incentive measures. At the same time, the scope of activities listed in the promotion tables varies, thereby creating uncertainty in firm decision-making.(') Also, (1) Seventy-three production activities from tomato paste to motor- cycle and bicycle tires were deleted from the list in 1979 while a number of activities from mushrooms and asparagus for export, to various motor vehicle parts and components were added (TUSiAD, July 1979, p. 75). 48  the inclusion of activities in the list and changes over time do not appear to have a clear economic rationale and, with it being a ((positive list)), it may exclude activities that are socially profitable. Correspondingly, it would be desirable to replace the ((positive list)) by a ((negative list)) that would specify activities which are not eligible for investment incentives. The exlusions should be limited in number, com- prising activities where foreign markets are restricted or domestic over- capacity exists. All other activities should receive investment incentives automatically. The handling of applications by potential foreign investors would also need to be streamlined by simplifying existing procedures and concen- trating decision-making in a single agency, so as to ensure that decisions are taken expeditiously. Apart from reducing bureaucratic obstacles, a positive policy of attracting foreign direct investment would be needed. In this regard, Turkey may learn from the experience of Ireland that has attracted foreign investment into export industries. There is further need to reform the system of import allocation that has increasingly become subject to discretionary decision-making as foreign exchange scarcity intensified. In creating uncertainty for the do- mestic producer, discretionary decision-making interferes with rational operations and it may lead to the interruption of production and low capacity utilization. Following improvements in the balace of payments, it would be desirable to gradually phase out quotas and import licensing, with exceptions made for luxury goods that are not producted domesti- with exceptions made for luxury goods that are not produced domesti- REFORMS IN THE STATE ENTERPRISE SECTOR State economic enterprises (SEEs) dominate in infrastructure, in the production of intermediate products, and play an important role also, in textiles, machinery, and electriaal equipment in Turkey. While the SEEs had made a substantial contribution to Turkey's industrial development, in recent years, they have been plagued by increasing losses. These los- ses have contributed to money creation and, in conjunction with the in- vestment requirements of the SEEs, have reduced the availability of funds for private investment (cf. pp. 35-36 above). Part of the explanation for the large losses of the SEEs lies in the lack of price adjustments for cost increases until the Spring of 1979. An 49  additional factor has been the pressure exerted on these enterprises to increase employment that doubled between 1970 and 1978 as a result. Pressure to increase employment, in turn, represents a manifestation of the prominence given to noneconomic factors in the management of the SEEs..The quality of management has further suffered by reason of the prevalence of political factors in choosing managers and in constituting the Boards of SEEs, the frequent interventions on the part of the minis- tries, and the low level of managerial compensation.(') As regards the comparative performance of public and private firms in the manufacturing sector, a technical study concludes that ((the ave- rage production efficiencies of the (publicly-owned) Sumerbank mills... appear to be-well below those of the more efficient private enterprises)) and that <(yarn quality in the private sector was markedly superior to that of the Sumerbank mills...)) (Ibid. pp. 2-36 and 3-9D). While similarly detailed comparisons have not been made for other industries, available data show that the state manufacturing enterprises have fallen behind their private counterparts in regard to both the productivity of labor and that of capital. For one thing, in practically every manufacturing industry, labor pro- ductivity is higher in the private than in the public sector, with the un- weighted average of the ratio of labor productivities in the two sectors being 1.68 in a 23 industry breakdown (TUSiAD, July 1978, p. 38). For another thing, increases in the share of the public sector in manufactur- ing investments from 32.7 per cent in 1963-67 to 43.6 per cent in 1973-77 were accompanied by a decline in their share in mabufacturing value added from 35.8 to 31.4 per cent (Ibid, pp. 39-41). High-cost production in the state manufacturing enterprises, in turn, has been sustained by protection and by the budgetary financing of losses. A reform of the state enterprise structure should aim at increasing the contribution of the SEEs to the national economy. In infrastructure, this would require reducing costs while practicing full-cost pricing. In the manufacturing sector, improvements in the international competiti- veness of the state enterprises should be the principal objective. Pursu- ing this objective would, in turn, necessitate increasingly integrating (1) According to a recent study of state manufacturing enterprises in Turkey, ((the institutional framework imposes severe limitations: poli- tical appointments, political interference in management, low executive pay, poorly defined performance and poor performance auditing)) (Walstedt, 1978, p. 1-64). 50-  these enterprises in the market economy and providing appropriate in- centives to management. On the example of Western European countries and, among social- ist countries, Hungary, the state enterprises should become self-managed and self-supporting units. Their managers should be chosen by, and be made responsible to, a Board independent from the Government, with the remuneration of management being made dependent on the firm's profits. The latter recommendation assumes the application of market prices to the output and the inputs of state enterprises, and the rationalization of these prices through the adoption of realistic exchange rates and interest rates and, subsequently, the reform of the system of protection. CONCLUDING REMARKS Following an analysis of the Turkish growth experience in an inter- national context and a review of Brazilian policy measures taken in a similar economic situation, in this paper recommendations have been made for policy changes in Turkey subsequent to the April-June 1979 measures. The policy recommendations aim at attaining stable economic growth in the frame-work of an export-oriented strategy while contribut- ing to the short-term objective of improving the balance of payments. The recommendations call for extending the scope of the parallel ex- change market, with a view to avoiding the re-emergence of the over- valuation of the Turkish lira and fluctuations in its value. This would in- itially exclude traditional agricultural exports subject to support prices that now receive the TL 35.00 rate. The stated objectives would also be served by reducing discrimind- tion against exports, channeling savings towards export activities, in- creasing public and private savings, and improving the operation of fi- nancial markets. Increases in private savings and the improved operation of financial markets would, in turn, necessitate raising real interest rates. Furthermore, there is need to improve the treatment of foreign direct in- vestMent that is preferable to borrowing abroad as it brings technolo- gical know-how and contributes to exports. The proposed measures would represent a shift from direct interven- tions to greater reliance on markets. Such a shift is overdue, given the increased sophistication of the Turkish economy and the need for in- creased export orientation that, in turn, requires incentives rather than quantitative targets. At the same time, apart from contributing to econo- 51  mic growth, export orientation would have beneficial effects on employ- ment and hence on the distribution of incomes. By contrast, import substitution in intermediate products tends to be capital intensive, thus limiting the growth of employment. Also, under present conditions, substantial profits may be made by the recipients of import licenses and by dealers in the parallel foreign exchange market. Finally, large firms associated with banks are at an advantage vis-a-vis medium-size and small firms in obtaining foreign exchange on the parallel matket as well as scarce credits, The proposed policW measures support-each other and should be considered a package. Thus changing exchange rates will have limited effects unless funds are provided for the expansion of export industries. Nor will it suffice to raise real interest rates as long as a substantial part of investment is not subject to economic project evaluation. In fact, for the proposed measures to have their full impact, the state economic en- terprises need to be brought more fully into the market economy. Note further that, apart from providing incentives to export, the des- cribed policy measures would contribute to efficient import substitution as well. As we have seen, in the machinery and equipment industries ex- porting and efficient import substitution go hand in hand. Equalizing rates of import protection will also contribute to efficient import substitution. Finally, this objective would be served by economic project evaluation in state enterprises. It should oe understood that it will take time until the full effects of the proposed measures will be felt. Correspondingly, the government needs a ((cushion)) to ease the transition from a large balance of payments deficit to an equilibrium position and from an iniport substitution strategy to an export orientation. Attracting foreign direct investment would help in this regard, but it would further be necessary for the developed count- ries to provide substantial financial assistance once decisions are re- ached by the Turkish government on the implementation of the necessary measures. 52  Appendix ECONOMIC GROWTH IN TURKEY: A STATIST'ICAL RE-APPRAISAL In examining Turkey's growth performance, one needs to separate the effects of sectoral growth rates of labor productivity (for short, sec- toral productivity growth) and those of the intersectoral movement of the labor force. This has been done in Appendix Table 1 where hypothe- tical rates of overall productivity growth have been calculated on the assumption that the sectoral composition of the labor force remained unchanged between 1967 and 1977. The difference between the reported rate of productivity grcwth, 4.9 per cent a year (Variant A), and that cal- culated on the assumption of the unchanged sectoral composition of thie labor force, 2.7 per cent a year (Variant B)('), indicates the impact of tile intersectoral movement of the labor force on overall productivity growth. The movement of labor from low-productivity agriculture to high- productivity industry and services is an integral part of the growth pro- cess in a developing economy. There is evidence, however, that inter- sectoral productivity differences have been overstated in Turkey, thereby leading to an overestimation of overall productivity growth. This ques- tion will be considered in the following utilizing international comparisons as well as data on protection in Turkey. of ) The reader will note that the latter figure represents an average of sectoral productivity growth rates (agriculture, 3.7 per cent; industry, 3.7 per cent; and services, 1.0 per cent), the weights being the sectoral composition of the labor force in the base year, 1967. 53  APPENDIX TABLE I Gross Domestic Product, Employment, and Productivity in Turkey, 1967-77(l) 1967 1967C 1967D I 1977A 1977B 1977C 1977D 1967-77A 1967-77B 1967-77C 1967-77D Value Added (TL million (absolute figures) (growth rates) 1968 prices) 93737 93737 93737 183321 148491 169952 174984 6.9 4.7 6.1 6.4 Agriculture 30506 46577 40107 42508 53471 64914 55893 3.4 5.8 3.4 3.4 Industry 24452 18613 14851 55314 42817 42102 33592 8.5 5.8 8.5 8.5 $ervices 38779 28547 38779 85499 52203 62936 85499 8.2 3.0 8.2 8.2 Employment (thousands) 11621 11621 11621 14151 14151 14151 14151 2.0 2.0 2.0 2.0 Agriculture 8160 8160 8160 7900 9937 1900 7900 -0.3 2.0 -0.3 -0.3 Industry 1545 1545 1545 -2430 1881 2430 2430 4.6 2.0 4.6 4.6 Services 1916 1916 1916 3821 2333 3821 3821 7.1 2.0 7.1 7.1 Labor Productivity (TL) 8066 8066 8066 12955 10494 12010 12365 4.9 2.7 4.1 4.4 Agriculture 3738 5708 4915 5381 5381 8217 7075 3.7 3.7 3.7 3.7 Industry , 15827 12047 9612 22763 22763 17326 13824 3.7 3.7 3.7 3.7 Services 20240 14899 20240 22376 22376 16471 22376 1.0 1.0 1.0 1.0 Source: Organization for Economic Cooperation and Development, Turkey, OECD Economic Surveys, No- vember 1978. Some of the inconsistent productivity data have been corrected by accepting data on sectoral value added and employment as accurate. Note : Variant A: Actual figures. Variant B: Hypothetical figures calculated on the assumption of the unchanged sectoral com- position of the labor force between 1967 and 1977. Variant C: Sectoral productivity figures for 1967 calculated on the assumption that produc- tivity ratios in industry-agriculture and services-agriculture comparisons in Turkey were the same in 1967 as in countries with per capita incomes of $200 to $574 in 1958 (Simon Kuznets, Modern Economic Growth-Rate Structure tind Spread, New Haven., Yale University Press, 1966, pp. 402-3) while taking overall productivity levels in 1967 in Turkey as given and calculating with observed sectoral employment and productivity growth rates. Variant D: Sectoral output and productivity figures re-estimated in terms of world market prices by utilizing ratios between domestic and world market value added of 1.70 for agriculture and 3.68 for industry while taking overall productivity levels in 1977 as given and calculating with observed sectoral employment and productivity growth rates.  In 1967, the base year of the calculations, the ratio of labor pro- ductivity in industry to that in agriculture was estimated as 4.23 in TurkeyA while the corresponding ratio for services as compared to agriculture was 5.41. These figures much exceed the average ratios estimated by Simon Kuznets (1966, pp. 402-3) for countries with per capita incomes in the $200-574 range in 1958, the group to which Turkey belonged in the base year (1967) of the calculations. The relevant ratios are 2.11 In the industry-agriculture, and 2.61 in the services-agriculture, comparison. The ratios are even smaller for' the income group below $200 a year (1.38 and 1.78) and over $574 a year (1.14 and 1.12)(.); Applying average productivity ratios estimated for the $200-574 in- come group to Turkey in the year 1967, and taking overall productivity levels in that year as given, results in the labor productivity figures shown in column 1967C of Appendix Table 1. Applying further actual rates of sectoral productivity growth between 1967 and 1977 to these figures, and calculating with actual employment data in the 1977, we have derived the hypothetical labor productivity figures reproduced in column 1977C of Appendix Table 1. The resulting estimates show a 4.1 per cent annual average rate of growth of overall productivity between 1967 and 1977 as compared to the reported rate of productivity growth of 4.9 per cent. The application of average sectoral productivity ratios for the $200- 574 income group to Turkey may be interpreted as a way to correct for distortions in relative prices resulting from the Turkish system of protec- tion. While intercountry differences in natural resources and in capital- labor ratios also affect the outcome, it does not appear that these dif- ferences would have introduced a bias in the comparisons. One may also address the issue of relative prices more directly by making adjustments for the price distortions actually observed in Turkey. This has been done on the assumption that, i,i the absence of price distortions, world market price relations would apply to Turkey. This assumption may be considered realistic as, with the exception of hazel- nuts (accounting for 14 per cent of exports in recent years) where it has a quasi-monopoly position, Turkey can trade at world market prices. Thus, world market prices will represent the opportunities available to Turkey and may be used to evaluate its rate of economic growth. (I) This comparison was first made in Hatiboglu, 1978, p. 193. 55  According to estimates made by Hasan Olgun in the framework of a 26-sector input-output table, in 1968 the ratios of domestic to world market value added in manufacturing and in agricultLure, respectively, were 4.15 and 1.70 at the actual exchange rate (1977, Table 6.16)('). These estimates have been used to derive the sectoral composition of Turkish GNP in 1967 in terms of world market prices, by taking total GDP as given and assuming that value added in the service sector is unaffected by protection. (2) The resulting estimates, shown in Column 1967D of Appendix Table 1, have in turn been utilized to re-estimate rates of overall productivity growth in terms of world market prices. This esti- mate is 4.4 per cent a year for the period 1967-77.(3) These results will overstate, however-, the rate of growth of overall labor productivity in Turkey, in part because differences within the manu- facturing sector are neglected and in part because no adjustment is made for the overpricing of services. As to the former, Anne Krueger has esti- mated that in the 1967-72 period the use of domestic prices has entailed an overestimation of the average annual rate of growth of the manufac- turing sector under the plan allocation by 0.8 per cent as compared to estimation in world market prices (1974, p. 259). Correspondingly, the above estimates of overall productivity growth in terms of world market prices need to be adjusted downwards.. (') Even higher protection rates for manufacturing industries were obtained by Tansu Qiller (1978, Table 2.4), whose results could not, how- ever, be used for present purposes because of the lack of aggregation for the manufacturing sector. (2) The same assumption has been made in regard to gas, electricity, and water, which are combined with manufacturing in the table. The in- dustry sector further includes mining, for which Olgun estimated the ratio of domestic to world market values added at 1.20 (1977, Table 6.16). With mining accounting for 7.9 per cent, and electricity, gas and water for 5.6 per cent, of value added in the manufacturing sector (OECD, National Accounts of the OECD Countries, 1975, Vol. II, Paris 1977), the ratio of domestic to world market value added is estimated at 3.68 for industry as a whole. (3) The 0.5 percentage points difference between estimates made at domestic and at world market prices exceeds similar estimates made by the author for other developing countries, Pakistan excepted, by a considerable margin. The estimated differences in terms of percentage points are: Brazil, 0.1; Mexico, 0.1; Pakistan, 0.5; and the Phillipines, 0.1. No differences are shown for Chile and Malaysia while for Norway there is a 0.1 percentage points difference in the opposite direction (Balassa and Associates, 1971, p. 34). 56  As regards the service sector, the data appear to overstate the contribution of the large and growing government bureaucrac)(') to eco- nomic welfare. This is because, under national accounts conventions, the increase of labor employed in public administration is considered as an additon to GNP, irrespective of its productivity. (1) In this connection, note that public consumption increased at an average annual rate of 11 per cent between 1967 and 1977 while private consumption grew by only 5,5 per cent (TOSiAD, July 1979, Page 101). 57  REFERENCES Abe, Shigeyuki; Fry, Maxwell J., Byoung Kyun Min, Pairoj Vongvipanond and Teh-Pei Yu ((Financial Liberalisation and Don'iestic Saving in Eco- nomic Development: An Empirical Test for Six Countries)), Pakistan Development Review, September, 1977, pp. 298-308. Balassa, Bela, ((Export Incentives and Export Performance in Developing Countries: A Comparative Analysis)), Weltwirtschaftliches Archiv, 1978, Vol. 114, pp. 24-61. (, aA 'Stages' Approach to Comparative Advantage)) in Economic Growth and Resources (Edmond Malinvaud, ed.). Proceedings of the 5th World Congress of the International Economic Association held in Tokyo in September, 1977, London, MacMillan, 1979 (Cited at 1979a). , ((Incentive Policies in Brazil)), World Development, 1979, forthcom- ing (Cited as 1979b). , and Associates, The Structure of Protection in Developing Count- ries, Baltimore, Md., Johns Hopkins University Press, 1971. Qiller, Tansu, ((Industrial Incentives in Turkey)), Appendix B. Import Re- gime in Turkey, Istanbul, 1978 (mimeo). Celasun, Merih, ((Sources of Industrial Growth and Structural Change: The Case of Turkey (1953-1973))), Ankara, October 1978 (mimeo). Dervis, Kemal and Sherman Robinson, Planning Models in Development Policy, Cambridge,. Cambridge University Press, 1980 (forthcoming). Fry, Maxwell J., ((Alternative Stabilisation EStrategies from a Model of Short Run Price and Output Fluctuations in Turkey)) Studies in De- velopment, Winter, 1978, p. 21-33. Gerakis, Andrias S., <(Effects of Exchange-Rate Devaluations and Reva- luations on Receipts of Tourism)), International Monetary Fund Staff Papers, November 1965, pp. 365-83. Hatiboglu, Zeyyat, An Unconventional Analysis of Turkish Economy, Istan- bul. Aktif Buro Basim Organizasyon, 1978. 58  Istanbul Chamber of Industry, Foreign Currency Shortage in 1978 and Effects on Manufacturing Industry, Istanbul, 1979. Kuznets, Simon, Modern Economic Growth: Rate, Structure and Spread, New Haven, Yale University Press, 1966. Krueger, Anne O., ((Some Economic Costs of Exchange Control: The Turkish Case)), Journal of Political Economy, October 1966, pp. 466- 80. --, Foreign Trade Regimes and Economic Development: Turkey, New York, National Bureau of Economic Research, 1974. Manisali, Erol, Foreign Economic Relations of Turkey, Istanbul, Guryay Matbaasi, 1979. McKinnon, Ronald, Money and Capital in Economic Development, The Brookings Institution, Washington, D.C., 1973. Olgun, Hasan, ((The Structure of Protection and the Policy of Industria- lization in the Turkish Manufacturing Industries, 1963-1971)), unpub- lished Ph.D. dissertation, Baltimore, Md., Johns Hopkins University, 1973. State Planning Organization, Turkey's Fourth Five Year Plan (1979-83), preliminary version, Istanbul, 1978. TUSiAD, Turkey's Industrial Sector in Foreign Trade with Special Refe- rence to EEC Relations, Turkish Industrialists and Businessmen's Association, Istanbul, April 1978. , The Turkish Economy, 1979, Prospects for Growth Within Stabi- lity, Turkish Industrialists and Businessmen's Association, Istanbul, July 1978. -, The Turkish Economy, 1979, Turkish Industrialists and Business- men's Association, Istanbul, July 1979. Walstedt, Bertil, ((State Manufacturing Enterprise in a Mixed Economy (The Turkish Case))), Wastington, D.C. 1978 (preliminary version). 59  THE WORLD BANK Headquarters: 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. European Office: 66, avenue d'Iena 75116 Paris, France Tokyo Office: Kokusai Building, 1-1 Marunouchi 3-chome Chiyoda-ku, Tokyo 100, Japan The full range of World Bank publications, both. free and for sale, is described in the World Bank Catalog of Publications, and of the continuing research program of the World Bank, in World Bank Research Program: Ab- stracts of Current Studies. The most recent edition of each is available with- out charge from: PUBLICATIONS UNIT THE V%ORLD BANK 1818 H STREET, N.W. WASHINGTON, D.C. 20433 U.S.A.