Richard Barth, William Hemphill
With contributiona from Irina Aganina, Susan George, Joshua Greene, Caryl McNeilly, Jukka Paljarvi

1. Trade Policies and Development Strategies Before 1980

2. The Reform Package of the Early 1980s

3. Developments Since the Reform Period
   
   1. The Period to 1988
      
   2. 1988–91
      
   3. 1992–93
      
   4. 1994–95
      
   5. Fiscal Developments in 1994–95
      
   6. Monetary Developments 1994–95
      
   7. Political Factors at the End of 1995 and Prospects for 1996

Part II.  Workshops on the Macroeconomic Accounts

1. Measures of Output, Employment, and the Price Level
   
   1. Introduction
      
   2. The System of National Accounts
      
   3. Measuring inflation
      
   4. Wage income and employment
      
   5. The labor market in Turkey
      
   6. Exercises and issues for discussion

2. Fiscal Accounts
   
   1. Some basic concepts
      
   2. Fiscal accounting
      
   3. Fiscal analysis
      
   4. The structure of the public sector in Turkey
      
   5. Exercises and issues for discussion
      
      Appendices:
      
      1. Seigniorage
         
      2. The Accounting Approach to Fiscal Sustainability

3. Balance of Payments Accounts
   
   1. Conceptual framework of the balance of payments
      
   2. Analyzing the external position
      
   3. Foreign liabilities
      
   4. Reserves and financing
      
   5. Shuttle trade in Turkey's balance of payments
      
   6. Exercises and issues for discussion

4. Monetary Accounting and Analysis
   
   1. Some basic concepts
      
   2. Monetary accounting
      
   3. Monetary analysis
      
   4. The structure of the monetary sector in Turkey
      
   5. Exercises and issues for discussion

5. Interrelations Among Macroeconomic Accounts:
   The Flow of Funds
   
   1. Some basic concepts
      
   2. Basic interrelationship between the saving-investment
      
   3. balance and the external current account
      
   4. Flow of funds accounting
      
   5. Exercises

Part III.  Workshops on Forecasting and Financial Programming

6. Forecasting Output, Expenditure, and the Price Level
   
   1. Some basic concepts
      
   2. Forecasting output and expenditure
      
   3. Forecasting output and expenditure in Turkey
      
   4. Determinants of the average level of prices
      
   5. Forecasting the price level in Turkey
      
   6. Forecasting a consistent set of national accounts deflators
      
   7. Exercises and issues for discussion
      
      Appendices:
      
      1. A Consumption Function for Turkey with Interest as an Exogenous Variable
         
      2. Estimating the Average Period Rate of Change of The Price Level
   
   
7. Government Sector
   
   1. Some basic concepts
      
   2. Forecasting revenue
      
   3. Forecasting expenditure
      
   4. Guidelines for forecasting the fiscal accounts in Turkey
      
   5. Exercises and issues for discussion
      
      Appendices:
      
      1. Some Tax Revenue Sources and Their Proxy Bases
         
      2. Proportional Adjustment Method—The General Case
         
      3. Turkey: Tax Summary as of May 1996
         
      4. Buoyancy and Elasticity with High Inflation
         
      5. Technical Aspects of Estimates of the Revenue Forecasting Equations

8. External Sector
   
   1. Some basic concepts
      
   2. Forecasting major items
      
   3. Regression equations for goods, services, and income
      
   4. Exercises and issues for discussion
   
   
9. Monetary Sector
   
   1. Some basic concepts
      
   2. Forecasting the demand for money
      
   3. Forecasting the rest of the monetary survey
      
   4. Forecasting the accounts of the monetary authorities
      
   5. Exercises and issues for discussion
   
   
10. Medium-Term Forecasting
    
    1. Some basic concepts
       
    2. The relationship between medium-term and short-term policy objectives
       
    3. Forecasting the medium-term implications of baseline projections
       
       Appendix: A User's Manual

11. Financial Programming
    
    1. Some basic concepts
       
    2. The framework for financial programming
       
    3. Policy and financing options
       
    4. Steps in economic forecasting
       
    5. Issues for discussion

Selected Data Series

1. Trade Policies and Development Strategies Before 1980¹

The role of the state fluctuated somewhat with domestic political changes and exogenous events (depression, war-time controls and the reaction to them, and periods of low or high agricultural prices), but changed little until the period of reform in the early 1980s. The party in power in the 1950s favored a reduced role for the state compared to prewar decades, but it took few concrete steps to achieve this objective. In fact, state economic enterprises (SEEs) were given access to central bank credit during this period. While the fiscal accounts of general government were roughly in balance, there were large public sector deficits because of SEE borrowing, especially to finance agricultural price supports. The deficits led to monetary expansion and inflation. With an exchange rate pegged to the U.S. dollar, the inflation led to an increasing overvaluation of the lira and falling exports. Import demand was limited by restrictions and licensing requirements.

Following the adoption of a stabilization program in 1959, including a devaluation, and a change in government,² the state's role in the economy was re-asserted in the 1960s with a constitutional amendment requiring a state planning organization and the official adoption of an import-substituting policy.³ Industry expanded by 10 percent per year in real terms from the middle 1960s, a time when per capita income in Turkey was equal to that in Korea,⁴ and the expansion continued and accelerated through the mid-1970s. Not only was growth strong, but inflation was low. As part of the stabilization program, SEEs were denied unlimited access to central bank credit to cover operating deficits, and limits were placed on commercial bank borrowing from the central bank. The modestly growing money supply was absorbed partly by falling velocity throughout the decade as monetization of the economy proceeded. Credit from banking system went more to finance private sector investment than government deficits. The general government deficit in relation to GDP hovered around 5 percent during this period, substantially financed by inflows of foreign aid.⁵ Growth in the supply of labor outpaced demand two to one—population growth in Turkey has averaged around two and a half percent per year in recent decades—but the availability of jobs in neighboring countries in western Europe provided an outlet for some of the excess workers.⁶

Under a policy of import substitution, the economic objective is to replace reliance on imports with an increased supply of goods made at home, especially manufactures. This strategy involves an innovative role for government in establishing new producing units with access to subsidized credit, and requires the erection of high tariff barriers to protect "infant" domestic producers against foreign competition. While higher-cost domestic producers can be effectively sheltered by this strategy, the high tariffs, besides discouraging imports, result in a more appreciated currency than would otherwise be the case, which penalizes exporters. In Turkey, a secondary objective of the inward-looking strategy was to export the manufactures produced in the government-supported sectors of industry. The two goals can in principle be pursued simultaneously (by taxing imports and subsidizing exports, or by fixing the exchange rate at a more depreciated real rate). However, Turkey's exports grew by only 1½ percent on average in the 1960s and ¾ of 1 percent in the 1970s.

In the early 1960s Turkey was successful in preserving balance in the external sector. The current account retained strength from the large devaluation at the end of the 1950s, and foreign aid flows financed modest deficits for a number of years. In line with import substitution policy, high import tariffs were supplemented by other controls and restrictions: import and export licensing, quotas, numerous surcharges, and advance import deposit requirements. As partial relief, rebates were granted on selected exports, and a structure of multiple exchange rates was used to favor importation of the producer goods needed to equip chosen manufacturing ventures.⁷ Nevertheless, by the end of the 1960s, pressure on the external sector had mounted as aid flows declined and foreign borrowing increased so that restrictions on imports were increasingly tightened. With growing pressure on the lira, a major devaluation took place in 1970, which strengthened the balance on goods and services temporarily and fostered a large inflow of emigrant workers' savings. Credit from foreign financial centers again became available. However, the devaluation was not accompanied by any other stabilization measures.

Although growth remained strong through the first half of the 1970s, it was a decade of mounting imbalances. The central government budget deteriorated due to large public sector wage increases and higher agricultural price subsidies at the time of the devaluation, subsidies to increase SEE investments plus heavy SEE borrowing from the central bank and from other state banks, and a relaxation of the limits on borrowing by the central government from the central bank. With a resumption of central bank financing and more rapid money growth, the pressures on the exchange rate increased continuously despite the administrative restrictions that were used to limit demand for foreign exchange. Two other external-sector developments accelerated the drift of macroeconomic performance. First, when world oil prices were raised in 1974, and again in 1979-80, Turkey resisted a pass-through of the higher energy costs to the domestic economy. Development of energy-using sectors proceeded according to the numbers in the five-year plans without regard for the need to economize and let demand and supply respond to the new relative price alignment.⁸ The conspicuously high costs of some public and private enterprises and the bloated petroleum import bill were covered ultimately by foreign borrowing. Second, in the aftermath of the first oil-price crisis, there was a decline in the opportunities for Turkish laborers to find employment in neighboring countries as guest workers; the growth of remittances and transfers from emigrants decelerated in step with the availability of foreign jobs.

One of the techniques devised to cope with the deteriorating external imbalance was a form of foreign borrowing known as the "convertible Turkish Lira deposit" scheme or the Dresdner Bank scheme. The program, dating from the late 1960s, was designed to attract the savings of Turkish nationals working in foreign countries and also the cash deposits that might have been earned in black-market trade, smuggling, or the mis-invoicing of imports and exports. According to the scheme, the Central Bank of Turkey offered interest rates on foreign exchange deposited in Turkish commercial banks 1¾ points above the Euromarket rate while also guaranteeing the foreign exchange value of both principle and interest. Beginning in 1975, the program was broadened to allow nonresidents in general, and not only Turkish nationals working abroad, to hold these deposits. Foreign exchange receipts from this source were transferred from commercial banks to the Turkish central bank and on-lent to government and SEEs, with expansionary effects on the money supply.⁹ Inflation accelerated markedly (still with a fixed exchange rate), worsening the underlying disequilibrium in the external sector. The Dresdner deposits constituted short term foreign loans, and therefore the maturity of Turkey's external indebtedness became increasingly short term as the decade progressed, despite earlier reschedulings intended to spread out debt servicing over time.

In 1977 Turkey went into arrears on its commercial debt, and capital inflows stopped. The need for a serious stabilization program was apparent but agreement on such a program could not be reached. There was serious political weakness in Turkey during the 1970s—a succession of five elections and five governments, each of the latter achieving less than a popular majority and being forced to form a fragile coalition with incompatible splinter parties as partners. The governments in this succession took no strong measures. Inflation accelerated markedly after 1977. There was increasing civil unrest. The second oil-price shock occurred in 1979-80, and GDP growth turned negative in both of those years. In the midst of a political vacuum and a deepening economic crisis, a military government took power in 1980 vowing to implement and enforce a stabilization program. The military government elevated the author of the plan, Turgut Özal, to the position of deputy prime minister.¹⁰

2. The Reform Package of the Early 1980s

The reforms that took effect in the early 1980s were limited largely to freeing some prices (including the price of foreign exchange and interest rates) and did not extend to the production sector. As a corollary to price liberalization, trade policy shifted from an inward orientation to outward orientation at this time.

Commodity prices were liberalized through a relaxation of direct government controls on the prices of some commodities (the Price Control Board of 1978 was abolished) and through allowing SEEs increased autonomy to set prices of their goods and services in accordance with cost and exchange rate developments. Goods whose prices were freed included petroleum, electricity, natural gas, petrochemicals, paper, cooking oil, auto tires, steel, telecommunications, textiles, and cement. The government retained direct control over the prices of goods produced under government monopoly (tobacco, tea, and other beverages) and also of certain basic commodities (bread, sugar, coal, fertilizer, the state railways, maritime transport, and electricity used in making iron, steel, and aluminum products). The reform plan reduced subsidies on inputs used by farmers except those for fertilizer and diesel fuel, although it retained direct control of support prices on agricultural goods. (Prices of some agricultural export goods were adjusted to world levels (cotton, raisins), and the support price for a single product—hazelnuts—was abolished.)

The Turkish lira was devalued and a policy of exchange rate flexibility was introduced whereby the nominal rate was to be changed frequently to offset the difference between the inflation rate in Turkey and major industrial trading partners. Officially from early 1981, the authorities embraced a policy of broad maintenance of the rate in real terms through frequent nominal rate changes.¹¹ As it turned out, the real rate depreciated by 30 percent against trading partners comparing the average level in 1980 with 1979 and continued to edge downwards—cumulating to an additional 20 percent over the next five years and 40 percent by 1987.¹²

Multiple currency practices were eliminated gradually during the first two years of the reform program, and payments arrears disappeared. Surrender requirements on exporters were reduced in several steps over the first half of the decade. Starting in 1984, banks were allowed to undertake foreign exchange operations on their own account subject to size restrictions, and foreign exchange deposits could be opened by residents and nonresidents. Banks were allowed to buy and sell foreign exchange at rates determined by market developments, without regard to the central bank rate, from 1985. During 1986-88, new foreign exchange pressures led to the re-imposition of a number of restrictions and limitations—including foreign exchange retentions, limits on operations by banks, increases in advance import deposits, and restrictions on trading currency with the central bank rate. In 1989 this reimposition of restrictions was reversed and additional liberalizing steps were taken; nonresidents were allowed to make purchases on the Turkish stock market and Turkish residents could purchase foreign securities. By March 1990, capital flows became virtually free.¹³

Before reform, tariffs and similar charges on imports averaged around 50 percent in manufacturing (an estimated 68 percent in effective terms).¹⁴ However, ultimate determination of import quantities rested on the use of quotas and licensing lists, and the high tariffs were redundant except as a source of revenue. The quota list was abolished in 1981, and some items were freed from licensing in the first two years of reform. Liberalization of trade slowed during 1982-83 but in the subsequent two years there was further reduction in the number of categories of goods for which licenses were required, so that during 1984 about half of imports were imported freely. In 1985, the number of items remaining on the licensing list was reduced further, by three quarters. In ensuing years, there were continuing changes, in both directions, in customs duties, luxury surcharges, and stamp duties, until Turkey began to take steps, in 1992-96, toward formation of a single tariff system prior to planned accession to the European Union.

Restrictions on, and preferential treatment of, exports did not end at the beginning of the reform period. Rather, subsidies in the form of tax rebates increased from an estimated 9 percent to 23 percent of value on average.¹⁵ Foreign exchange retention allowances were increased in the early 1980s; preferential credits were continued. However, in 1984 export licensing and price controls were eliminated, and in the following year the phase-out of incentive schemes was begun (along with the institution of a value added tax (VAT) in Turkey), which was finally completed in 1992 (though replaced at that date by a new credit scheme). The tax rebate system was ended in l988, and licenses were abolished in 1989. Surrender requirements for exporters were raised temporarily back up to 100 percent in 1988 during the balance of payments difficulties of that year, when multiple exchange rate practices for exports also reappeared briefly. Subsidized export credits, much of which leaked into domestic operations, grew until 1984, then were abolished, replaced by a 4 percent direct subsidy. Subsidized credits for exports were reintroduced into the payments system in 1988, 1989, and 1992, the last to replace the direct subsidy of 4 percent.

Interest rates, the third part of the reform package, had been set subject to ceilings in the low two-digit range in the period up to reform, while inflation had soared to triple digits during 1980. As part of the reform package, interest rate ceilings were raised though not abolished; over the period 1980-83, rates on time deposits became gradually positive in real terms, the maximum nominal rates still being controlled administratively (at least for larger banks). In 1980 the ceiling on bank lending rates was raised substantially, though not removed, to 70 percent initially in nominal terms. However, the number of preferential lending schemes proliferated. Until this time, not only had interest rates been controlled but credit had been rationed, and there were no institutional safeguards against unsound banking practices. The liberalization of deposit and lending rates precipitated a banking crisis due to a high demand for credit in the presence of the tight monetary policy that was followed during the early reform years. As part of the response to the 1982 banking crisis, the monetary authorities in 1983 simplified the structure of commercial banks' required liquidity¹⁶ and reserve ratios. (Lower ratios on preferential credits, such as to exporters, were ended.) Additional rates reappeared in 1984 but the structure was again unified in 1985. The practice of paying interest on commercial banks' reserve deposits at the central bank was ended (except for reserves in foreign currency on foreign exchange deposits). Tax rates on financial transactions were reduced though not abolished.

The reforms in the early 1980s were concentrated on the external sector. Relaxation of restrictions on trade and payments and setting the exchange rate at a more competitive levels had the effect of shifting the orientation of production from the home economy to a balance of home and export markets that depended on relative profitability. Comparative advantage succeeded self-sufficiency as the underlying rationale for external sector policy. The effect of this change on exports was dramatic. In the first two years following the 1980 devaluation, export values in dollar terms nearly doubled; the average annual increase in the volume of exports of manufactured goods was close to 40 percent from 1980 to 1985.¹⁷

In important respects, however, the reform program was partial rather than complete. It did not essentially affect the size and ownership of enterprises in manufacturing; this sector continued to be dominated by large, monopolistic, government-owned enterprises, many with significant non-economic objectives. While the changes in the external sector were significant, the reforms were not comprehensive. Numerous restrictions, controls, and subsidies remained, which in subsequent years were sometimes further relaxed and on other occasions tightened (for example, in 1986-88).

3. Developments Since the Reform Period

For two years following the beginning of reforms, macroeconomic balances in Turkey improved. The current account deficit in the balance of payments was reduced by about two-thirds, supported by rapid export growth, and the overall balance moved into surplus. Between 1980 and 1982, the central government's budget deficit was halved, making it possible to cut the public sector borrowing requirement (PSBR) to 6 percent of GDP in the latter year. The rate of monetary expansion was reduced, and the velocity of money declined rapidly as interest rates on time deposits became positive in real terms. Inflation decelerated from over 100 percent in the course of 1980 to 25 percent in 1982. Real GDP, which grew at a negative rate in 1980, increased at the rate of 5 percent in 1982.

In 1983, a year of significant further steps in trade liberalization, the direction of macroeconomic policy was reversed. Unemployment had grown during the "debt shock" years, putting pressure on Government for an expansionary stance, and an election was scheduled for November, marking a return to civilian government. Although imports grew rapidly in this period because of trade liberalization, GDP growth accelerated to 6 percent in 1984; the PSBR expanded to 8 percent of GDP; catch-up SEE price increases after the election, followed by a sharp nominal devaluation of the lira, pushed the inflation rate to 50 percent.

The period 1983-97 was one of increasingly expansionary monetary and fiscal policies. The real growth rate of GDP climbed from 5 percent to 8 percent in response. Nevertheless, inflation remained in the range of 30-50 percent (measured by the GDP deflator). And while imports grew faster than exports, the current account balance remained financeable. Several events occurred in this period to keep the economy from becoming more inflationary. The global oil price fell sharply in the mid-1980s (Turkey's terms of trade improved about 10 percent in 1986), so that the increase in import values was held down. The real exchange rate continued to decline slightly year by year, and exports continued to grow (though, except for manufacturing, not robustly). Real wages fell slowly in this period; collective bargaining agreements still required the approval of the High Arbitration Council, established by the predecessor military government. Moreover, cost-of-living adjustments (for unionized workers, generally those in large establishments) were influenced by the expectation that inflation would decline further in the aftermath of reforms and the opening of the external sector. In 1985 a value added tax (VAT) was initiated, with a very strong one-time effect on revenues beginning in 1986—an increase amounting to 5 percent of GDP (expenditure increased equally, however). Renewed foreign borrowing in the form of repatriation of bank accounts of Turkish nationals working abroad pushed the short-term share of external debt up to half again but provided a supply of foreign finance.

But the expansionary policies ultimately led to trouble. By 1987, the PSBR reached 9 percent (an election year), and civil servants insisted on two cost-of-living adjustments per year instead of one, the second to include compensation for inflation exceeding the official expected rate at the time of the signing of the labor agreement. Interest rates ceilings were reduced despite signs of a resurgence of inflation, and depositors shifted a significant share of bank funds to foreign-currency denominated accounts (which had became permissible in 1984). In December, the month following the election, catch-up increases in SEE prices led to a ten-percentage-point jump in the price level. By the end of 1987, the Turkish lira was discounted 15 percent in the foreign exchange market.

The year 1988 began with a foreign exchange crisis, downward pressure on the lira and an outflow of foreign capital. Mild monetary measures (higher interest rate ceilings and higher reserve and liquidity ratios) were implemented in February and eased somewhat in mid-year. A second foreign exchange crisis followed in October, and monetary policy was tightened again. On the fiscal side, steps were taken to reduce the borrowing needs of SEEs (higher prices, wage restraint, and a postponement of investment spending), and the PSBR fell from about 9 percent to 7 percent of GDP. The measures tended to slow the growth of output, which fell to 3½ in 1988 and to 2 in 1989; however, inflation accelerated to 66 percent. Inflation was to remain stuck in the range of 60-70 percent for five years, following a period of about equal length in the range of 30-50 percent.

In 1989 policies were loosened. A reflow of foreign capital after the 1988 crises added to money growth. Employees of state economic enterprises received pay increases amounting to an average of 142 percent (with inflation at 66 percent), and civil servants and the private sector managed to do almost as well. In the shorter run, the large wage increases were not reflected in an acceleration of inflation but in reduced profit margins (see the discussion in Part II, Chapter 1, Section e). To some extent the wage adjustment, while surprisingly large, was anticipated, although the correction was more than enough to make up for falling real wages in the years since 1980. The restrictive macro policies of the preceding year, taken in response to foreign exchange market turbulence, together with a drought, had a strong depressing effect on real growth in 1989, which was, at 2 percent, about equal to the increase in population. It was not a natural time to pass on cost increases in the form of fast-rising prices.

The strong wage increases in 1989 led to a rebound of domestic demand and growth in 1990. The current account, which had been in surplus the year before when domestic demand was weak, slipped back into deficit. Then, in 1991, the economy was beset by an external shock, the Persian Gulf War. Turkey was to lose a major export market (Iraq) as a result, and generous oil pipeline fees, but initially the effect of the shock was to bring downward pressure on the lira. The central bank defended the currency and absorbed large trading losses as a result. The change in trade flows and in expectations generally due to the Gulf War brought an end to the economic expansion, with GDP growing 1.5 percent. The current account improved, as domestic demand stagnated, but no substantial change in the inflation rate occurred. The year 1991 was an election year, and in 1990 and 1991 there were further large wage increases. Partly in response to the sharply higher wages, private-sector firms let as many workers go as possible; they sought to reduce labor costs through sub-contracting, investment in labor-saving technologies, and reductions in seniority (workers with longer times in a particular job generally received somewhat higher pay). Labor turnover approached 30 percent in firms paying the large increases. As before, SEEs suppressed price increases before the election, and absorbed wage increases in the form of operating losses. Monetary and fiscal policy were lax in the period leading up to the election—in the case of fiscal policy, partly on account of increases in agricultural price supports.

In 1992 there was a rebound of output, with real growth reaching 6 percent. The rebound would be easy to explain solely in terms of the very large wage increases of the preceding years and the end of the Gulf War. In addition, the PSBR increased from 11 percent to 14 percent, a strongly expansionary change, in part reflecting the wage increases in the public sector. In 1993 the expansion intensified, and imbalances became unmistakable toward the end of the year. The public sector deficit increased one more percentage point despite rising incomes, to 15 percent, an historical high. Real GDP growth accelerated to 8 percent, surprising in view of weakness in Turkey's export markets (domestic demand expanded by 13½ percent). The volume of imports of merchandise increased 39 percent, and the current account deficit widened from 1 to 3½ percent of GDP. The debt-GDP ratio increased moderately from 35 percent to 37 percent at the end of 1993; at this point, one quarter of the debt was short term. In January 1994, two rating agencies downgraded Turkey's credit rating. The loss of confidence on the part of creditors in the country's ability to meet its external obligations was met by a tentative tightening of monetary policy and increases in interest rates. However, the limited measures only served to feed expectations of stronger action to come, and foreign lenders reduced their short-term exposure.

The central bank seemed intent on keeping monetary policy tight in this period to offset lax fiscal policy. It shifted financing of the central government from advances to securities, which had the effect of keeping interest rates high and attracting plenty of foreign capital to finance the current account deficit. The exchange rate tended to appreciate over the years 1990-93 as debt grew. Ultimately the confidence of foreign creditors was undermined by these two developments.

The central bank of Turkey lost 40 percent of its reserve assets in the early weeks of 1994 and eventually was forced to let the rate go. The lira depreciated 60 percent by early April. The crisis prompted the authorities to put together a stabilization program, which was announced in early April, and involved deep cuts in discretionary government spending, immediate price increases for SEEs(followed by a six-month freeze), and a tightening of monetary conditions.¹⁹ During the second quarter of the year, interest rates on three-month treasury securities averaged 275 percent (annual rate), moving for short periods as high as 400 percent.

The loss of access to international capital markets together with a tighter policy stance, and the effects of the exchange rate change and higher SEE prices on real wages, resulted in a dramatic fall in domestic demand, by over 12 percent in real terms in 1994. Real wages fell 15-20 percent for 1994 on average.²⁰ The sharp fall in domestic demand was partly due to de-stocking; after unwanted inventories had been worked off, output in the industrial sector turned up again. Exports responded quickly to the exchange rate change and to slack domestic demand and rose by 35 percent in volume for the year as a whole; import volume declined 27 percent. Nevertheless, real GDP fell by 5.5 percent in 1994.

The large devaluation early in 1994 and the prices increases of SEEs (of 70-100 percent) initially pushed up the inflation rate. April over April, the rate was 125 percent, with perhaps a quarter of the acceleration due to the change in import prices in liras and half to the change in SEE prices.²¹ By mid-1994 the monthly rate of increase had slowed to only 2 percent. (At the time of harvest and of lower food prices, summer inflation rates are typically low.) However, with interest rates at very high levels and the expectation of a much reduced rate of nominal depreciation of the lira, short-term capital began to flow back into the economy by mid-year, enough that international reserves of the central bank were rebuilt and monetary expansion began to pick up. Besides the goal of continuing disinflation, and holding onto the competitiveness gains occasioned by the sharp devaluation early in the year, the central bank also needed to ease monetary tightness in order to reduce interest rates and the interest payments required on government debt. The tightness of stabilization policies was relaxed by autumn. The perceived resulting change in policy stance was interpreted as eroding public confidence in the sustainability of the anti-inflation program. Inflation re-accelerated, reaching 106 percent for the year on average. The real depreciation of the exchange rate for the year on average was 21 percent.

The economy rebounded in 1995, with GDP growing by 7 percent.²² Imports increased 33 percent in volume terms, throwing the current account again into modest deficit. Strong financial inflows were allowed to lift the exchange rate, by 13 percent in real terms for the year as a whole, to limit monetary expansion. Nevertheless, while inflation slowed early in the year, it accelerated in the second half. For the year on average the increase in the price level was 82 percent, 88 percent as measured by the consumer price index. These rates benefitted from the usual postponement of increases in prices of goods and services produced by the SEEs until after the election called for late December (see the final section, below), and from the real appreciation.

The stabilization program implemented in April 1994 was built around a large, front-loaded, fiscal correction. Under the program, the PSBR declined by 6 percentage points of GDP in 1994 and an additional three points in 1995. One percentage point of the decline in the PSBR in the first year came from the finances of SEEs. There were large increases in their prices, already mentioned, and a reduction in the stocks of agricultural goods held by SEEs involved in the price support program (the Soil Products Board). The larger improvement in SEE finances the following year included the effects of favorable crop prices (a decrease in the cost of price supports) and further postponement of investment projects. As it turned out, personnel costs for SEEs also declined because pay adjustments were delayed until the end of a strike by employees in October.

Tax measures yielded a decrease in the deficit amounting to 1½ points of GDP. They included a 10 percent surcharge on personal income and corporate profits tax liabilities. A ceiling was placed at the existing level on the number of employees in central government ministries and agencies for five years; new vacancies could be filled but there could not be an increase in the number of workers. Indexation of civil service salaries was ended. No unbudgeted wage increases were to be allowed in 1994; given the acceleration of inflation, this implied a fall in real wages. Personnel expenditure of government declined in relation to GDP from 11 percent in l993 to 8 percent in 1995.

These fiscal savings taken together add up to more than the changes actually achieved in the PSBR because the savings were partly offset by increased interest payments. In a stabilization period, monetary policy is tightened and the cost to government of borrowing in the form of treasury bills and bonds increases, often markedly.²³ If, as part of the reforms, a government intends to decrease monetary financing (advances by the central bank, which in general lead to money stock increases) by increasing the sale of securities, this will also lead to higher interest payments. The higher government interest expenditure frustrates the aim of reducing the PSBR relative to GDP.

Throughout the period 1991-93, the PSBR had been financed almost entirely through domestic borrowing. New external financing averaged 1 percent of GDP in 1992-93. But it turned negative in 1994-95 when Turkey's access to international loans was interrupted. Net sales of government securities, which accounted for about 50 percent of domestic financing in 1991, rose to 82 percent by 1995 as financing from central bank advances was cut back. In its 1995 financing program, the Treasury sought to extend maturities on government paper over the course of the year and to resume external borrowing during the second half of the year. Substantial success was achieved in pursuing these objectives, with the Treasury selling unexpectedly large volumes of nine-month and one-year paper in the first half of the year, thereby shifting a portion of the interest burden into 1996 and reducing its borrowing need during the second half of 1995. However, with the heightened uncertainties created by the unexpected fall of the Government in September (see Section g), borrowing costs rose sharply during the last quarter while the maturities on new sales of government paper shortened considerably.

The political developments connected with the election at the end of 1995 also caused the wage increase for civil servants that would normally have taken place in early 1996 to be brought forward to November 1995. During the course of the year, wages of civil servants were raised on three occasions, by 19.6 percent in January, 36.2 percent in April, and 57.6 percent in November. An additional increase of 50 percent was agreed for mid-1996.

The monetary objective of the program of April 1994 was to stabilize the value of the lira as a basis for bringing down the inflation rate over time. Higher interest rates, and a tighter stance of monetary policy generally, were expected to contribute to this result. However, there was considerable uncertainty at the time on the part of lira holders. For example, there was apparent doubt that the liberal policy regime that permitted deposits in foreign currencies would survive the crisis, and large-scale withdrawals of foreign exchange occurred. The central bank therefore moved rapidly to re-establish confidence in the banking system, extending the system of deposit insurance and opening new credit facilities for commercial banks experiencing liquidity problems. There was also a cessation of central bank advances to the central government, which was therefore forced to borrow completely at market rates. Interest costs were extremely high, as mentioned above.

The program to shore up the banking system was broadly successful. Only three small banks failed. It was supported eventually by the initial effects of fiscal measures to reduce the need for government borrowing from the banking system and by a strengthening of the lira (in real terms) in foreign exchange markets. However, the increase in confidence in the banking system was accompanied by an easing of monetary policy. In addition, capital flight was reversed by the later months of 1994, which added to upward pressure on the rate of growth of broad money. Unfortunately the growth of liquidity came at a time of renewed inflationary pressures and other factors undermining confidence of the public in the disinflation program (a constitutional court invalidated plans to hasten privatization of SEEs). Confidence of lira holders again ebbed. The monetary authorities, at the same time, declined to defend the lira with higher interest rates, seeking instead to lower the government's financing costs. Inflation speeded up substantially toward the end of 1994.

Despite the easing in the stance of monetary policy toward the end of 1994, there was a restoration of confidence the following year, based especially on a strong recovery of economic activity. Foreign currency substitution was reversed in Turkish deposit money banks, international reserves in the central bank increased, and the recovery in economic activity brought a rise in the demand for credit from the private sector. The central bank bought foreign exchange to limit lira appreciation, with the result that the rate of growth of the money supply early in the year was substantially above the inflation rate (around 130 percent per year, not seasonally adjusted). By the middle of the year the central bank eased credit policy further; it undertook large foreign exchange purchases without sterilization to ease the liquidity shortage, and rapid money growth continued. After the political uncertainties at the end of the third quarter, the increase in the money supply accelerated significantly.

In September 1995 the governing coalition in Turkey dissolved when one major party withdrew from participation. The prime minister, Ms. Çiller, refused to meet the demands of her former partner for public sector pay raises fully commensurate with inflation in the year. She resigned, and subsequently was asked by the president, Mr. Demirel, to form a new government. After talks lasting for several weeks, Ms. Çiller announced the formation of a coalition with partners from the far right and the center left. The new coalition failed a vote of confidence in parliament on October 15, and it was announced subsequently that national elections would take place on December 24. In that election, the voters gave no single party sufficient support to constitute a government on its own, so that either there will again be an attempt to form a coalition, perhaps with little more success than previously, or else another general election will be called. A long strike by employees in State Economic Enterprises occurred in October, and the government, in the pre-election atmosphere, ultimately gave in to generous wage increases; the wage increase for civil servants scheduled for January of 1996 was moved forward to late 1995, as already mentioned, and there were unscheduled cost-of-living increases to pensioners in mid-November. Financing for these expenditures involved the central bank, and the money supply grew rapidly during the final months of the year (in excess of 150 percent at an annual rate, private credit at a 280 percent annual rate), stimulating capital outflows and a drop in central bank reserves as the bank sought to avoid a rapid decline in the exchange rate of the lira without raising short-term interest rates. An extended period of political uncertainty was expected, while with the new year the government would have increased limits for borrowing from the central bank. The lira began to drop more rapidly on the foreign exchange market as the year closed despite central bank purchases. The annualized interest rate on treasury paper approached 240 percent.

It appears doubtful that the tightened stance of fiscal policy can avoid a fairly sharp increase in the PSBR relative to GDP in 1996. Measures taken to reduce SEE cost overruns—postponing wage increases, hiring, price-support increases, and investment projects—tend to depress spending and boost surpluses artificially. A return to a deficit in the neighborhood of the historical ratio of 3 percent of GDP for SEEs is to be expected following the small surplus in 1995 attributable to the emergency measures in 1994. In addition, the deficit in the social security budget is expected to increase from 1 percent of GDP in 1995 to 2 percent in 1996. Government investment was reduced to 4 percent of GDP in the crisis but had been 8 percent in 1991, and can be expected to recover significantly in 1996 and subsequent years. Tariff reductions scheduled for January 1 will reduce revenues by 0.3 percent of GDP. While personnel costs could grow more slowly in real terms following the freeze on the number of civil service positions imposed in 1994, similar policies have proved difficult to sustain in the past as the economy edged away from financial crises. There are a number of reasons to expect that interest payments will increase.