Report No. 26301-TU Turkey Country Economic Memorandum Towards Macroeconomic Stability and Sustained Growth (In Three Volumes) Volume II: Main Report July 28, 2003 Poverty Reduction and Economic Management Unit Europe and Central Asia Region Document of the World Bank Turkey: Country Economic Memorandum Currency Equivalents (ExchangeRate Effective July 28,2003) Currency Unit = Turkish Lira US$1 = TL1,410,000 Government Fiscal Year January 1 -December 31 Weights and Measures Metric System Abbreviationsand Acronyms AMC - Asset ManagementCompany PA - PrivatizationAdministration BK - Bag-Kur PFMC - Public FinancialManagementand Control BO - Build Operate PHC - Privatization High Council BOO -- Build Operate Own REER - Real EffectiveExchangeRate BOT Build OperateTransfer SDlF - Saving Deposit Insurance Fund BRSA - Banking Regulatoryand Supervisory SIS - State Instituteof Statistics Authority CAR CapitalAdequacy Ratio SOE - State Owned Enterprises CEM - Country EconomicMemorandum SPO - State PlanningOrganization CPI - Consumer Price Index SRS - Sustained ReformScenario EBITDA - Earnings Before Interest,Tax, SSK - Social Security Organization Depreciationand Amortization ES - EmekliSandigi TEAS - ElectricityGeneration and Transmission Company EU - EuropeanUnion TEDAS - Turkish ElectricityDistributionCompany EUAS - Turkish ElectricGeneration Company TEIAS - Turkish ElectricityTransmission Company FlAS Foreign InvestmentAdvisory Service TEKEL - Turkish Tobacco Monopoly FDI Foreign Direct Investment GDP - Gross Domestic Product TETAS - Turkish ElectricityTrading Company GDR - General Directorateof Revenues TFP - Total Factor Productivity GNP - Gross National Product TOBB - Union of Chambers of Commerce and Industry IA - IstanbulApproach TOOR --Transfer of Operating Rights ISE - IstanbulStock Exchange TSFAS Turkish Sugar Factories ISKUR - EmploymentOrganization TUSIAD - Turkish Industrialists'and Businessmen Association MOF - Ministryof Finance YOIKK - CoordinationCouncil for InvestmentClimate NPL - Non-PerformingLoans Vice President: Johannes Linn (ECAVP) Country Director: Ajay Chhibber (ECCU6) Sector Director: Cheryl Gray (ECSPE) Sector Leaders: Samuel Otoo (ECSPE) Team Members: Ismail Arslan (Team Leader, ECSPE), Mediha Agar, Pinar Baydar, Craig Burnside, Robert J. Cull ,Annette De Kleine, Asli Demirguc-Kunt, Ira Lieberman, Kamer Ozdemir, James Parks, Philip Suttle, Matthew Verghis (World Bank) Refik Erzan, Alpay Filiztekin, Nissan Liviatan, Fatma Taskin, Erol Taymaz, Ercan Uygur (Consultant) TURKEY: COUNTRY ECONOMIC MEMORANDUM TOWARDS MACROECONOMIC STABILITY and SUSTAINEDGROWTH TABLE OF CONTENTS CHAPTER MACROECONOMIC 1: FRAMEWORK 2 Financial Crisis and Response................................................................................. 1 Introduction............................................................................................................. 1 2 4 Achieving Macroeconomic Stability ....................................................................... 3 Sustaining the Recovery .......................................................................................... 6 9 5 6 The Quasi-Fiscal Dimensionof the Public Debt................................................... DebtandFiscal Sustainability ............................................................................... 13 17 7 EnsuringFiscal Sustainability ............................................................................... 21 8 The Medium-Term Agenda For Growth............................................................... 26 CHAPTER PUBLICSECTORREFORMS 2: 2 Structural Fiscal Reforms...................................................................................... 1 Introduction........................................................................................................... 28 30 4 Public Sector Governance ..................................................................................... 3 Structural and InstitutionalReforms to Upgrade PEM ......................................... 39 46 CHAPTER3: BANKINGSECTOR 48 2 Private Bank Restructuring and Recapitalization .................................................. 1 Introduction........................................................................................................... 3 State Bank Restructuring and Privatization........................................................... 52 4 Credit Crunch ........................................................................................................ 59 5 Changes inthe Incentive Environment and Intermediation inthe Long Term .....62 64 CHAPTER CORPORATE 4: SECTOR 68 2 Analysis of Corporate Distress .............................................................................. 1 Introduction........................................................................................................... 69 3 71 4 MediumTerm Structural Reforms ........................................................................ Short-Term Resolution Strategies ......................................................................... 80 CHAPTER5: SOCIAL POLICIES THEIMPACT OFTHE CRISISAND LABOR - MARKET MEASURES 1 Introduction........................................................................................................... TOWARDS THE GROWTH 86 CHAPTER TURKEY'S 6: EXTERNAL PROSPECTSA GLOBALIZINGWORLDECONOMY IN 2 Background - RobustPerformance InExternal Balance..................................... 1 Introduction......................................................................................................... 102 103 3 107 4 ProspectsFor Turkey's External Revenues,2001-2005 .................................... Global Outlook .................................................................................................... 109 113 6 The Role of ForeignDirect Investment............................................................... 5 External Vulnerability and Risks......................................................................... 118 REFERENCES ................................................................................................................... 123 TECHNICALANNEX ......................................................................................................... 128 1 Introduction......................................................................................................... 128 TABLES CHAPTER MACROECONOMIC 1: FRAMEWORK TABLE1,l: Key Economic Indicators, 1999-2003 TABLE1.2: Summary Business Cycle Statistics, 1988-2002 TABLE1.3: Dynamic Correlations with GNP, 1987-2002 TABLE1.4: The Stock ofPublic Debt inTurkey, 1994-2002 TABLE1.5: Public Debt Dynamics inTurkey, 1995-200 1 TABLE1.6: Public Debt Primary Balances inTurkey, 1995-2002 TABLE1.7: Alternative Fiscal Adjustment Scenarios TABLE1.8: ProjectedPublic Debt DynamicsinTurkey, 2003-2009 TABLE1.9: Fiscal Adjustment, 1999-2002 CHAPTER PUBLICSECTORREFORM 2: TABLE2.1: Size o f General Government inthe OECD TABLE2.2: Organization ofNationalTax Administrations inOECD Countries TABLE2.3: Public Sector Employment TABLE2.4: Public Sector Investment, 1996-2003 CHAPTER BANKINGSECTOR 3: TABLE3.1: Turkey Banking System- Selected Indicators TABLE3.2: SDIFBanks andTheir Resolution Process TABLE3.3: Fiscal Cost of Bank Restructuring TABLE3.4: Evolution ofNon-Performing Loans andReserves TABLE3.5: Capital Structure inPrivate Commercial Banks TABLE3.6: Turkey - Financial Development CHAPTER 4: CORPORATE SECTOR TABLE4.1: Net Profitability by Sector, 2001-2002 CHAPTER 5: SOCIAL POLICIES THEIMPACT OF THE CRISIS AND LABOR - MARKET TABLE5.1: Employment Indicators, 1998-2001 TABLE5.2: Summary Indicators of the Strictness of Employment Protection Legislation TABLE5.3: Public and Private Employment, 1999-2002 TABLE5.4: Productivity and Wages inSelected Countries TABLE5.5: Overview of Active Labor Market Programs CHAPTER6: TURKEY'S EXTERNAL PROSPECTS INA GLOBALIZINGWORLD ECONOMY TABLE6.1: Summary: Turkey's Links with the Global Economy, 1980-2000 TABLE6.2: Turkey: Current Account Balance TABLE6.3: GlobalConditions Affecting GrowthinDeveloping Countries, 2001-2005 TABLE6.4: Turkey: Indicative Projections ofForeignExchangeRevenues, 2001-2005 TABLE6.5: Turkey: External Financing Requirementsand Sources, 2000-2006 TABLE6.6: Mexico - External Financing TECHNICALANNEX TABLETA 1: Switching RegressionResults TABLETA 2: Summary Statistics TABLETA 3: Correlations TABLETA 4: Credit to Households and Government TABLETA 5: Variable Construction andData Sources BOXES TABLEB 1.1: Real GDP Growth inTurkey, 1967-2002 TABLEB 1.2: Volatility and Growth inEmerging Economies, Real GDP growth TABLEB 2.1: Central Government Tax Revenues FIGURES CHAPTER: MACROECONOMIC 1 FRAMEWORK FIGURE1.1: Macroeconomic andFinancial Indicators FIGURE1.2: Real Expenditure Growth and Real Interest Rates inTurkey FIGURE1.3: The Cost & Maturity of Domestic Borrowing FIGURE1.4: Inflation andDepreciation FIGURE1.5: Public Sector FixedInvestment FIGURE1.6: Public Sector Wage Bill FIGURE1.7: Deficit o f Social Security System CHAPTER PUBLICSECTORREFORM 2: FIGURE2.1: Turkey: Tax Performance FIGURE2.2: Turkey: Breakdown of Tax Revenues incl. SSI CHAPTER BANKINGSECTOR 3: CHAPTER 4: CORPORATE SECTOR FIGURE4.1: Export Orientation FIGURE4.2: Financial ExpenseBurden FIGURE4.3: Profitability by Size FIGURE4.4: Distressed Companies, 1998-2001 Q1 FIGURE4.5: Financial Expenses FIGURE4.6: Turkey: Privatization Under PA Portfolio and GSMLicenses CHAPTER SOCIAL POLICIES-THEIMPACT OFTHE CRISIS AND LABOR 5: MARKET MEASURES TOWARDS ECONOMY GROWTH THE AND FIGURE5.1: Real Wages, 1990-2001 FIGURE5.2: Adult Population, Labor Force andEmployment, 1923-2002 FIGURE 5.3: Indices of Value Added inManufacturing andReal Labor Costs CHAPTER TURKEY'S 6: EXTERNAL PROSPECTS INA GLOBALIZING WORLD ECONOMY FIGURE6.1: Trade to GDP has risen by more than 50 percentagepoints FIGURE6.2: Change inTurkey's openness strong among comparators FIGURE6.3: "Speed of integration" bestedonly by Mexico over 1996-2000 FIGURE6.4: Growth of Turkey's exports exceedthe growth of the export market FIGURE6.5: Turkey's revenues:expansion strongest intourismandautos FIGURE6.6: Turkishadvantage inproductivity levels inTAC is fading FIGURE6.7: A Sequence of crises duringthe 1990s FIGURE6.8: Volatility of REERand GDP Growth FIGURE6.9: Slackening pace of growth inper-capita import purchasing power FIGURE6.10: Turkey's exposureto international risks:well above comparators FIGURE6.11: FDIInflows: After Crisis 1& 2 years after relative to 1year prior to crisis FIGURE6.12: Total WorldFDIflows to Developing FIGURE6.13: FDIinflows rankedby EBRDInstitutionalPerformance Indicator FIGURE6.14: Annual FDIInflows to Turkey TECHNICAL ANNEX FIGURETA 1: Real Credit to the Private Sector FIGURETA 2: Credit Demand and Supply (Model 2) FIGURETA 3: Capital andReserves at Central Bank FIGURETA 4: Credit to Households and SMEs FIGURETA 5: Credit to Government BOXES FIGUREB 1.1: Fiscal Costs ofBanking Crises Since 1990 BOXES CHAPTER1: MACROECONOMIC FRAMEWORK Box 1.1: Volatility and Growth Box 1.2: Currency Crises andthe Fiscal Costs of Banking Crises Box 1.3: Contingent Liabilities inthe Energy Sector CHAPTER 2: PUBLICSECTORREFORM Box 2.1: International Perspectiveson Tax Effort inTurkey Box 2.2: Tax Strategy Priorities Box 2.3: The Management of Fiscal Risks inTurkey Box 2.4: Priorities undertheNationalStrategy to Enhance Transparency and Good Governance inTurkey's Public Sector CHAPTER 3: BANKING SECTOR Box 3.1: Bank Recapitalization: International Experience Box 3.2: Coordinating Financial and CorporateRestructuring: International Experience Box 3.3: State Bank Restructuring and Privatization: International Experience CHAPTER 4: CORPORATE SECTOR Box 4.1A: Institutional Structure for IstanbulApproach Box 4.1B: IstanbulApproach Process Box 4.2: Key Recommendationsfor CorporateResolution Box 4.3: Legal Framework for Creditor Rights andEnforcement inTurkey TURKEY: COUNTRY ECONOMICMEMORANDUM TOWARDS MACROECONOMIC STABILITY and SUSTAINEDGROWTH CHAPTER 1:MACROECONOMIC FRAMEWORK 1 INTRODUCTION 1. This report addresses key questions facing Turkish policymakers: how to sustain the economic recovery that began in 2002 following the deep crisis o f 2001, how to ensure disinflation and public debt sustainability, and how to foster broad-based and equitably distributed growth in the future. After a brief review o f the 2001 crisis and the Government response, the report analyzes the economic opportunities and challenges facing Turkey, and identifies policies to build on the economic recovery which began in2002. The C E M develops a comprehensive four-point agenda for sustainable and more equitably distributed growth. The agenda encompasses: (i)macroeconomic stability, (ii)effective government, (iii)improved business environment, and (iv) stronger social policies. The report includes an analysis o f the external environment for Turkey which highlights the importance o f expanded foreign direct investment. 2. This chapter summarizes recent economic developments and analyzes the key policies needed to achieve macroeconomic stability and sustained growth over the medium term. The macroeconomic analysis in this chapter argues that, given the state o f the economy and the amount o f public sector debt that has now been accumulated, macroeconomic stability and growth can only be sustained by a deep fiscal adjustment that would not only achieve ambitious targets for the primary balance, but would also address the key sources o f Turkey's prior fiscal problems. It i s estimated that Turkey will need to maintain its primary surplus at the current target o f 6.5 percent o f GNP over the medium term to bring its debt level back to manageable proportions. International experience has shown that fiscal adjustments are usually only sustained when they are o f high quality. Since Turkey's debt i s only manageable with a substantial and sustained fiscal adjustment, the quality o f the adjustment will be key. Quality i s reflected inthe use o f expenditure cuts, rather than revenue increases to achieve the adjustment. It is also reflected in addressing structural fiscal problems, such as extensive contingent liabilities, inefficient public sector enterprises, and distortionary subsidyprograms. 3. With the Government now needing to focus on fiscal sustainability and policy coordination, the public sector will naturally end up playing a smaller direct role in the Turkish economy. Loss making state-owned enterprises, un-funded subsidies to the agricultural and small business sectors through the state banks, and indirect subsidies to financial intermediation, are no longer feasible. Thus, the private sector must take responsibility for longer-term growth. The Government's main role should be to provide the right environment for private sector activity. Many o f the conditions for sustained, private sector-led growth are inplace in Turkey. Export competitiveness i s strong. The very rapid pace of opening and integration o f Turkey's economy over the past twenty years has set the stage for even stronger export performance inthe future. There is no capacity constraint holding back increased consumer and private investment spending in the near term. Utilization rates are low, and the economy has the productive capacity with which to meet increased demand without fueling inflation. Growing demand will, inturn, stimulate private investment. Onthe other hand, private sector access to financing may 1 become a constraint on the recovery over time. Given the government's external debt position, the private economy has limitedaccessto external finance and domestic financing is constrained by the ongoing restructuring in the Turkish banking sector. Thus, strictly limiting its own financing needs and ensuring the recovery o f the banking sector are key objectives that the Government must focus on to sustain the recovery. 2 FINANCIALCRISISAND RESPONSE 4. In2001,Turkey suffered through one of the worst economic crises it has faced inrecent memory. Two overriding factors played key roles in Turkey's economic crisis. These were the build up o f huge fiscal imbalances over the 1990s and the lack of attention to the accumulation o f systemic banking sector risks. The poor initial condition o f public finances when the original disinflation program was launched in late 1999 was combined with a weak banking sector and a slowdown infiscal and structural reforms inmid-2000. As a result, an initial consumption boom was soon dissipated in a loss of confidence inthe crawling peg exchange rate regime used as the principle nominal anchor for disinflation. A pull back by foreign creditors in the face of a widening current account deficit led to large-scale financial turmoil in November 2000. The Government responded by taking over a major private bank and extending a blanket guarantee to bank creditors. However, this exacerbated concerns about fiscal sustainability and the problems came to a head in a full-fledged currency crisis inFebruary 2001 that forced the Government to float the Lira and bail out the banking system. 5. The government announced a crisis response program in May 2001 following the collapse o f crawling peg and subsequent devaluation. The key structural and social elements o f the program were (i)macro-framework designedto restore financial stability and ensure public a debt sustainability-principally through a further tightening o f fiscal policy with a primary surplus target o f 5.5 percent o f GNP in 2001 increasing to 6.5 percent thereafter; (ii)rapid restructuring o f the banking sector-especially o f state banks and insolvent private banks intervened by the regulatory authority (BRSA)-based on large resource transfers from the budget; (iii)a much more ambitious program of public sector reforms centered on deeper structural fiscal reforms and institutional reforms to improve public expenditure management and public governance; (iv) a renewed privatization drive in combination with further liberalization measures (particularly in energy, telecommunications and agriculture) and strengthening o f the role o f independent regulatory bodies to improve the climate for private investment; and (v) strengthening of social assistance to help low income groups adversely affected by the crisis. The Government secured US$lO billion in additional exceptional financing from the IFIs to helprestore investor confidence and close the financing gap opened by the crisis-which caused large private capital outflows-and the bank restructuringprogram. 6. Macroeconomic outcomes under the crisis response program were mixed. The financial turmoil was contained, but financial markets remained fragile. The initial depreciation o f the Lira was quickly absorbed into consumer prices, but inflationary pressures continued well into the year. Following the decision to abandon the peg, exchange rate related uncertainty persisted as the government was slow to reaffirm its commitment to the float and the central bank repeatedlyintervenedinthe foreign exchange market. The exchange rate stabilized by late 2001, albeit at a more depreciated level than originally projected. Interest rates were brought down from their crisis peaks, but remained well above the original program path throughout 2001, 2 mainly due to higher-than-expected inflation and fears about the sustainability of the public sector debt. The primary surplustarget was met, but there was continued doubt about the 1999 2ooo 2o01 2o02 2003 recession was much est. deeper than projected and GNP Growth -6.1 6.3 -9.5 7.8 5.0 recovery was delayed until CPIInflation(Dec-Dec) 68.8 39.0 68.5 29.7 20.0 2002. A serious drought NominalInterestRate 106.2 38.0 99.1 63.8 45.0 exacerbated the recession UnemploymentRate 7*1 6.6 8*5 12.3" Primary Balances(%GNP) " -2.0 2.7 5.5 3.9 6*5 with agricultural output NetPublic Debt (YoGNP) 61.0 57.7 95.0 79.8 77.2 contracting by some 6.1 Privatisation($bn) 0.1 3.3 2.8 0.3 1.5 percent. Overall, GDP Current account balance(YoGNP) -0.7 -4.9 2.4 -1.O -3.2 shrank by an estimated 7.5 REER(1995=100) 123.1 136.5 112.5 125.3 138.0 percent in 2001, GNP by 8. The economy recovered strongly in 2002 and macroeconomic targets were generally exceeded. GNP is estimated to have increased by 7.8 percent, more than double the original growth rate target of 3 percent. Of course, the strength of the recovery in 2002 largely reflected the rebound expected from a deep recession. It does not, therefore, indicate a similar rate of growth in the longer-term. The recovery was underpinned by good news on exports, which responded quickly to the real exchange rate adjustment following the February 2001 crisis. In real terms, exports rose by 7.4 percent in 2001 and a further 11 percent in 2002. Private consumption also started to recover, growing 2 percent in 2002 after a 9.2 percent decline in 2001. Private investment, which fell by 31.5 percent in 2001, stopped declining in 2002. It was down 0.8 percent for the year, but was up 22 percent on a Q4 to Q4 basis. CPI inflation was 45 percent based on annual average levels. However, on an end-year basis, CPI inflation declined to just under 30 percent. There was substantial real appreciation in 2002, though the real exchange rate remained substantially below its pre-crisis levels. The balance of payments remained strong, supported by strong export performance and tourism receipts, despite the appreciation of the CPI-basedreal exchange rate. This may well have been due to the sharp fall inprivate sector real wages after the crisis. With the economic recovery, imports increased 16 percent inreal terms, and the current account balance moved from a surplus o f about 2.4 percent o f GNP in2001 to a deficit o f about 1 percent o f GNP in2002. Figure 1.1:Macroeconomic and FinancialIndicators :I (a)Annual CPI Inflation (b) Real Effective Exchange Rate 80 I 160 70 140 - 60 0 I c 0 r $ 50 120 - a, Q 40 m m r 100 - 30 2 0 4 , , , , , , I I , , , I 80 , , , , , , , , , , , , ~ (c) Nominal T-bill rate (d) Manufacturing Sector Real Wage Index (1997=100) 250 110, I 200 12 month nominal interest rate on T-bills . 105 E 150- 100 8 100- 95 50 - 90 85 80 (e) Unemployment Rate (9GNP growth 14, -15 J 1 (9) Real Private Credit 83 , (h) Capacity Utilization in Manufacturing I 654 , , , , , , , , , , , , , I Source: Government, IMF and World Bank 4 The massive capital outflow o f 2001 was halted, and, infact, the capital account balance was in surplus to the tune of 1.1 percent o f GNP in 2002. Thus, overall, the balance o f payments was in balance, an improvement over estimates under the original program. The Central Bank's gross foreign exchange reserves ended the year comfortably at some US$28 billion. 9. Fiscal performance fell well short o f the program targets for 2002 as fiscal discipline was relaxed during the run up to the November elections. The consolidated public sector primary surplus for 2002 is estimated to have been 3.9 percent o f GNP, well short o f the target 6.5 percent. The main reasons for the emerging fiscal gap included: (i) unplanned civil service an wage increase; (ii) overruns in social security expenditure; (iii) in adjustingpublic goods delays and services prices; and (iv) a drop intax revenues linked to expectations o f a post-election tax amnesty. While higher than programmed public expenditures contributed to the recovery in 2002, the slippages will have serious implications for meetingthe 2003 fiscal targets. 10. Financial markets fluctuated sharply over the course o f 2002, in line with political developments. After falling below 60 percent in early 2002, domestic interest rates on benchmark government securities increased once more in May following a deterioration in the former Prime Minister's health which triggered renewed political uncertainty and the announcement o f early elections scheduled for November 3. Interest rates remained fairly steady inthe 70-75 percent range through the Summer and Fall o f 2002. Financial markets rallied on the election results, regaining much of the ground lost to political uncertainty over the summer. Treasury borrowing rates dropped once more to the 50 percent range, considerably easing the debt burdenand improving the prospects of sustaining the recovery. However, secondary market rates started to increase again inmid December as markets became increasingly concerned about the new Government's commitment to the economic program. For 2002 as a whole, domestic interest rates averaged 63.5 percent, somewhat lower than the original program projection o f 69.1 percent but in line with the lower than expected inflation realization. While real interest rates (based on backward-looking annual inflation) have fallen sharply from their post-crisis peaks, they have remained in the 20 percent range since mid-2002. Net public sector debt fell from an estimated 95 percent o f GNP at the end o f 2001 to 80 percent by end-2002. The main factor behindthe fall inthe debt to GNP ratio was the real exchange rate appreciation duringthe course o f 2002, together with the economic recovery. 11. The economic program for 2003 aims to sustain the recovery with the projected growth o f 5 percent. Continued strong export performance and a return of private consumption and investment demand are expected to lead the recovery. Growth in 2003 will also be helped by strong carry over effects from the second half o f 2002. The recovery continued into the first quarter o f 2003, with GNP 7.4 percent higher than in the first quarter o f 2002. Importantly, private consumption and investment led the way for the first time since the crisis, recording increases o f 6.5 percent and 20.4 percent respectively. Industrial production and export data suggest that the recovery continued in the second quarter, albeit probably at a slowing rate. Building on the positive outcome last year, the program targets a further reduction in CPI inflation to 20 percent by the end o f the year. The authorities are counting on tight fiscal and monetary policies, a positive output gap and a stable exchange rate to bring the inflation rate down to the targeted level. The program aims to ensure sustainability o f the public debt through sustained fiscal adjustment. The primary surplus i s targeted to returnto the program level o f 6.5 percent of GNP in2003 based on a package o f fiscal measures equivalent to about 4.9 percent o f 5 GNP introducedby the Governmentto cover boththe slippagerecordedin2002 anda number of temporary measureswhich expired. Together with the continuedrecovery, the tight fiscal stance i s expectedto underpin a further modest decline inthe public debt to GNP ratio. On the external side, the current account is projected to weaken substantially, reflecting increasing imports, higher oil pricesand flat tourismrevenuesfollowingrecordreceiptslast year. 3 SUSTAININGTHE RECOVERY 12. In the short-term the keys to overall economic recovery are improved consumer and business confidence, because, as in previous recessions, the sharpest declines in aggregate spendingwere inprivate investmentandinconsumer durables spending. Inthe recent recession, like others before it, consumer and business confidence was shattered. In particular, in 2001 consumer spendingon durables andbusinessfixed investmentbothplummeted,by 30.4 and34.9 percent, respectively.In2002, spendingon consumer durablesrose slightlyby 2.1 percent inreal terms. But private sector fixed investment fell by a further 7.2 percent. Only in the fourth quarter of 2002 did signs of a recovery in investment spending materialize. Private fixed investment in 2002Q4 was up by 21.8 percent compared to 2001Q4. But both categories of spendingremainwell downfrom their pre-crisislevels (Table 1.2). Table 1.2: Summary BusinessCycle Statistics, 1988-2002 Std. Deviation Correlation Change Change with GNP (2000-01) (2001 02) - Percent Relative percent Shareo f percent Shareof to GNP A in A in GNP GNP GNP 6.1 1 1 -9.5 1 7.8 1 Privateconsumption 5.9 0.97 0.91 -9.2 0.65 2.0 0.18 Nondurables & services 3.0 0.49 0.80 -2.9 0.13 1.8 0.11 Semi-durables 7.9 1.30 0.85 -9.0 0.10 3.1 0.04 Durable goods 22.6 3.73 0.80 -30.4 0.42 2.1 0.03 Private fixed investment 17.3 2.85 0.80 -34.9 0.76 -7.2 -0.14 Public consumption 4.8 0.80 0.73 -8.5 0.08 5.4 0.06 Publicfixed investment 18.1 2.99 0.72 -22.0 0.17 14.5 0.12 Change in inventories 1.9* 0.50' -1.4* 0.42 4.7* 0.92 IExports 8.1 1.34 0.34 7.4 -0.26 11.0 0.55 Imports 19.1 3.13 0.92 -24.8 -1.04 15.7 -0.67 Source: Annual nationalaccountsdata for 1987-2002 were obtainedfromthe SIS. Note: Residualcomponentsofthe change in GNP not reportedinthe table are other net incomeand the statistical discrepancy. *Where notedwith an asterisk, the statistics for the change in inventories are for the change in inventories as a percentageof GNP, notthe growth rateofthe change in inventories. 13. To gain some sense of the importanceof the various categoriesof demandinthe Turkish business cycle, consider the summary statistics presented in Table 1.2. These describe the volatility of different components of the national accounts (on the expenditure side), as well as their correlation with GNP. As Table 1.2 indicates, consumer spending on durables and semi- durables is much more volatile than GNP, and much more volatile than consumer spending on 6 nondurables and services.' Private fixed investment spending is about 3 times as volatile as GNP. Another indication of the importance of spending on durables and investment goods is found by computing the share of the decline in GNP in 2001 due to different spending components. Together, consumer semi-durables and durables, along with private fixed investment,account for 128 percent of the decline inGNP. By comparison, consumer spending on nondurables, along with changes in public sector demand account for only 38 percent of the decline inGNP. 14. On the basis of data for 2002, Turkey appears to be recovering strongly from the crisis- driven recession of 2001, but there are lingering doubts about whether the recovery will be sustained and built upon inthe coming year. After the economy bottomed out inthe first quarter of 2002, the performance of the economy inthe second, third and fourth quarters was strong, and real GNP grew 7.8 percent for the year. As Table 1.1 indicates, the recovery has mostly beendue to continued good export performance (exports were up 11percent inreal terrrs for the year) and a sharp increase in business inventories. There were some other signs of recovery in the real economy in 2002. Production in the industrial sector was up 9.4 percent in 2002. Agricultural production was up 7.1 percent. Wholesale and retail trade also recovered strongly (up 12.8 percent) after a dismal year in 2001. Banking, not surprisingly, remains a weak sector, with the services of financial institutions slipping a further 7.1 percent in2002. 15. There are signs that confidence is improving and that the most likely outcome for 2003 is continued growth, albeit at a more moderate pace. The modest increases in consumer durables spending, andthe continued decline of private fixed investmentspending indicate that consumers and firms remained wary about the state of the economy in 2002. Fixed investment spending bottomed out in the third quarter of 2002, and rose on a 01Q4-02Q4 basis by 22 percent. While construction spending remains depressed, investment in machinery has risen substantially. Still consumer durables and fixed investment spending have a long way to go to reach their pre- recession levels. It is unlikely that export performance will continue to improve at the dramatic rate observedin2002, particularly as real appreciation of the Lira is likely to slow export growth. The increase in business inventories i s a positive sign, in that it presumably reflects firms' anticipation of a continued recovery. However, the fact that most of the increase in GNP i s accounted for by the inventory buildup highlights the vulnerability of the recovery. If expectations are not met, and private demand does not increase sufficiently, the recovery could be infor ahard landing. 16. It is difficult to predict movements in GNP in Turkey based on national accounts data, giventhat there are no obvious leading indicators to look at. Most expenditurecomponents seem to have relative symmetric dynamic correlation patterns with GNP at the quarterly frequency.2 This i s indicated in Table 1.3. If anything, it would appear that expenditure on imports is Overallconsumer spending is slightly less volatile than GNP, despitethe fact that spendingon semi-durables and durables is more volatile than GNP. This reflectsthe fact that nondurables spending is considerably less volatile than, though positivelycorrelated with, GNP, and is by far the largest component of private consumption. A leading indicatorhas a skewed pattern of dynamic correlationwith GNP. For example, importsappear to be a leadingindicatorbecausethe correlationbetweenimportsand GNP one quarter ahead is stronger than the correlationbetween importsand GNP one quarter lagged. 7 somewhat o f a leading indicator, while private fixed investment and public sector demand are somewhat lagging indicators o f the business cycle. But the patterns revealed in the correlations are not particularly strong. Table 1.3: Dynamic Correlations with GNP (quarterly, seasonally-adjusted data, 1987-2002) Lead time 01 -4 -3 -2 -1 0 1 2 3 4 X variable Private consumption 0.38 0.54 0.70 0.83 0.94 0.80 0.62 0.45 0.26 Nondurables& services 0.40 0.51 0.63 0.74 0.82 0.74 0.65 0.58 0.49 Semi-durables 0.25 0.43 0.61 0.74 0.86 0.72 0.59 0.42 0.30 Durables 0.26 0.38 0.53 0.66 0.76 0.60 0.38 0.15 -0.08 Privatefixed investment 0.40 0.58 0.72 0.80 0.84 0.75 0.62 0.52 0.43 Public consumption 0.27 0.22 0.34 0.40 0.38 0.24 0.04 -0.17 -0.28 Public investment 0.05 0.07 0.08 0.12 0.16 0.01 -0.10 -0.30 -0.42 Exports 0.11 0.13 0.14 0.13 0.15 0.10 0.06 -0.03 -0.07 Imports 0.16 0.31 0.53 0.74 0.89 0.78 0.57 0.33 0.09 Source: SIS Note: Correlation is betweenX, and GNP, The datawere seasonally adjustedusingthe X11 procedurein Eviews. The datawere then expressedin logarithms and linearlydetrendedbeforecomputingthe correlations. 17. Can the economy accommodate increased consumer and private investment spending? On the production side the answer i s unequivocally yes3. Capacity utilization rates, which rose sharply towards the end o f 2002, have returned to relatively low levels-the economy has the productive capacity with which to meet increased demand, without inflationary pressure arising. The Government has relatively limited means at its disposal to spur private sector confidence. Given its budget constraint, the best the government can do is to maintain fiscal restraint-to help lower interest rates and build confidence in the overall economic program. On the other hand, consumer and business access to financing for major purchases and investment is an important problem. Given the Government's external debt position, the private economy has limited access to external finance. Domestic financing is also limited as a result of the ongoing restructuring inthe domestic financial sector. Hence, a stronger fiscal position within the public sector i s crucial in order for there to be sufficient funds for private borrowing to finance investmentand durables purchases. Hopefully, the Government can pursuethe fiscal adjustment and structural reforms needed to spur the improvements in confidence required sustain the recovery in2003. See Taskin (2002), a backgroundpaper for this Report, for the automotive industry. 8 4 ACHIEVING MACROECONOMIC STABILITY 18. Figure 1.2 illustrates the importance of economic stability and confidence for the Turkish economy. It plots GNP growth, the growth o f private fixed investment, and the growth o f Figure 1.2: RealExpenditureGrowth and Real InterestRates inTurkey consumer durables and semi- 35 durables spending against the real Growthof Pnvate Growthof Durables money market interest rate.4 It i s 25 arguable that real interest rates are 15 a function o f private confidence in the economy, and the 5 government's policy framework. Whether this i s the case, or not, -5 what i s clear, from Figure 1.2, i s -15 that real expenditure i s strongly inversely correlated with the real -25 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2WO 2001 2002, the correlationsbetween interest rate. Between 1990 and Source: Central BankofTurkey Data Delivexy System for annual national accounts data IMIntemational GNP growth, private fixed Financial Statistics for the moneymarket interest rate. investment growth, and consumer durables growth and the real interest rate were, respectively, -0.92, -0.83 and-0.75. 19. Macroeconomic stability i s important inthat (i) Turkey can illafford another fiscal crisis, and (ii) the recovery will only be sustained if confidence in the economic program increases. But macroeconomic stability is also crucial for medium-term growth. Turkey has experienced unusually volatile growth since the later 1980s which has hampered its development. In particular, it has diverted vast amounts o f managerial expertise into financial management and away from decisions about technology, investments and human resources. Consider, for a moment, the decision making process for a firm considering investment in machinery or equipment. The firm must compare the merits o f alternative choices for this new capital. The firm must also compare the returnto the capital investmentto the returnon alternative financial investmentsand the cost o f borrowing'. InTurkey each of these items is incredibly volatile due 4The realmoney market interestratefor year t is measuredas rl = 100 [(lRJ(1 + zq) - 11, where r,is the + 'inflationrateofthe GDP deflator betweenyear t-1 andyear t. For example, we couldwrite the gross nominalreturnto physical investmentas: Here PK represents the cost of a unit of capital, Py is the price o f output, delta is the depreciation rate, and MS representsthe marginal sales the firm generates by making the investment in new capital. The firm will compare its forecast of this return to the gross nominal interest rate (in TL), 1 + RI, as well as to its forecast of the gross dollar cost ofborrowing: 9 to the economy's volatility. Historically, the exchange rate has been extremely volatile, as has inflation. Furthermore, a firm making capital investments faces extreme uncertainty about its future sales due to the volatility of the real sector. As a result, Turkish firms use an enormous portion o f their managerial and technical services in financial management. They spend proportionally less assessing the relative merits o f alternative physical investments. It also leads to conservative decision-making: business decisions often come down to choosing the least risky strategy, and this is often to hold FX-denominated financial assets. Firms also spend less time trying to improve productivity and more time on cash management (Box 1.1). 20. Supporting evidence for the view that uncertainty i s a major concern for firms comes from a background study for this report6 Firms surveyedin 1996 cited devaluation, contractions in domestic demand, and rises in interest rates as by far their most significant concerns during the 1994 crisis. They also stated that they responded to the crisis most often by decreasing their borrowing requirement. Though not many cited postponement o f investment as a major response, presumably this would be an important way o f reducing borrowing needs. Of course, investmentcertainly declined sharply duringthe 2001 crisis. 21. There were some signs that a degree o f stability and confidence was Figure 1.3: The Cost & Maturity ofDomestic restored in 2002. Through most o f 2002, Borrowing treasury bill rates were on the decline, 240 24 while average maturities, for the most part, 200 20 lengthened (Figure 1.3). This suggests a 3 -a 160 16 declining risk premium on domestic -8 (z currency denominated debt. Until July f 120 E 12-2 al 2002, inflation declined faster than the a t 80 a m f nominal interest rate. If inflation 40 4 s expectations moved in line with observed 0 0 inflation then this would imply that real 1999 2000 2001 2002 2003 interest rates remained high. And this I A v s r a g e Maturity -A=. T-bill Rate would help explain the continued slump in Source: TreasuryandWorldBank investment demand. In the latter part of 2002, interest rates continued to decline while there was a renewed bout o f inflation associated with the depreciation of the lira that occurred between April and July. Thus, perceived real borrowing costs may have declined. It i s as yet unclear which direction real borrowing costs are moving in 2003. While T-bill rates have settled at about 50 percent range, during the first six 6 months, inflation ran at annual rates near 40 percent (Figure 1.4), suggestingthat real borrowing costs have fallen. On the other hand, in the past 6 months the value of the Lira has increased 1+RP =(l+Rj)-, St+l c where R' is the dollar interestrate, andS is the TL/$exchange rate. For a firm locatedin Turkey forecastingthe TL- returns in(1) and (2) is extremely difficult. %ygur (2002), "The Two-TierCrisis and Investment inTurkey". 10 suggesting that inflation may soon drop further substantially. Thus, it may be the case that impliedreal borrowingcosts remainquite high. Box 1.1: Volatility and Growth annum, a levelo fvolatility similar to that experiencedby Canada inthe 1962-85 period.' Furthermore, 1979 was the only recession Geometric Standard year inthis period. Inthe 1980s, growthslowedto 4 percentwhile Average Average Deviation volatility rose slightlyto 3.5 percent. Inthe 1990s, Turkey 1967-79 4.8 4.8 3.O suffered a hrther decline inaverage growth and a dramatic 1980-89 4.1 4.0 3.5 increaseinvolatility; growth slowedto 3.8 percentwhile volatility 1990-99 3.9 3.8 5.3 rose to 5.3 percent. For 1990-2002, the average growth ratewas 1990-2002 3.6 3.4 5.8 As Table B 1.2 shows, Turkey's increased Table B 1.2: Volatility and Growth in Emerging Economies, volatility is not a worldwide phenomenon. Real GDP growth (percent) Somethingspecific has madeTurkey movefrom beinga country with relatively low volatility and 1980-89 1990-01 slightly above average growthamong emerging Average Std.Dev. Average Std. Dev. market countries,to havingslightlybelow 2hile 3.5 7.1 Thailand 4.8 6.0 average growth and significantly morethan 'hilippines 1.9 5.1 Turkey 3.1 5.9 average volatility. kgentina -1.O 4.9 Argentina 2.9 5.4 3razil 3.0 4.6 Malaysia 6.6 5.2 It is oftenarguedthat as countries developthey Corea 7.5 3.9 Korea 6.1 4.7 become less volatile as agriculture shrinks in importance. However,the 2000 CEMpointed rurkey 4.0 3.5 Chile 5.8 3.6 out that agriculturalproductionand servicesare Vlalaysia 5.8 3.4 Philippines 2.9 2.2 both less volatilethan industrialoutput in rhailand 7.2 3.2 Brazil 2.6 2.2 Turkey. Onthe expenditureside, private ndia 5.9 1.9 India 5.5 1.9 investment, consumerdurablesandthe current Source: CEM 2002 accountare the most volatile components. Why has growth become more volatile despiteTurkey's development? One relevantfactor is that Turkey's increased openness, especially inthe capital account, has enhancedthe impact o f externalshocks giventhe underdeveloped financialsystem. Another importantfactor is that the effects o fexternal and internal shocks have been exacerbated by highlypro-cyclicalfiscal policy. ' See Ramey, Garey andValerie Ramey (1995) "Cross-country Evidence on the LinkBetweenVolatility and Growth,'! American Economic Review, 85, 1138-51. 11 22. The question for the Government i s how to enhance the Figure1.4 MationandLkpmiation degree o f confidence the public (backwrd-lwking6mnthm caverages,annualrates) and investors have in its economic loo f l program. In this regard, fiscal 80 sustainability i s a key issue. The 60 Government needs to stay the - 4 0 course with the size of its fiscal 8g 20 adjustment-otherwise the stated o dual goals o f reducing inflation and -20 restructuring the financial and corporate sectors will not have 199 2000 2001 m 2003 credibility. Absent a fiscal adjustment of the magnitude I+mation+m a t i o nI targeted under the program, it Source: Central BankandWorldBank simply would not be credible that ~otes:maiioniscm-quta~ theCPI.~epreclation for is forthen/$ acchz~lgerate. the cost associated with restructuring could be absorbed without resort to inflationary finance. Another reason the size o f the fiscal adjustmenti s crucial is to create room for renewedlendingto the private sector. While growth in credit to the private sector i s unlikely to become a leading indicator, expanded credit access will be neededto sustain the private sector-led recovery over the medium term. Over the past decade the public sector has absorbed an increasing fraction o f the available domestic financing. The financial crisis in2001 only exacerbated this situation. Inaddition, the crisis has reduced the willingness o f external creditors to lendto the private sector. This has beenfactored into projections for Turkey's current account balance in the coming years: it i s anticipated that only small current account deficits will be possible.' Inthis environment, it is a mathematical necessity that there be a sharp increase in domestic savings to finance durables and investment spending. A sharp increase in public savings will play an important role, especially at a time whenprivate incomes are depressed. The Role of FiscalPolicy 23. As Turkey emerges from yet another period of recession, it i s important to learn from the policy mistakes made inthe past. After each o f the past crisis episodes Turkey has experienced, memories have beenshort. As part fiscal crises have moderated and tax revenues improved with renewed economic expansion, control over public finances has tended to relax. This is exemplified inthe fact that public sector consumption, especially non-wage spendingand public sector fixed investment, are highly pro-cyclical (Table 1.2). As a consequence, fiscal policy has tendedto accentuate demand shocks ratherthan offsetting them. Some examples are informative: 0 Inthe 1994-95 recession public sector investmentfell by 43 percent, then it rose 80 percent through 1998. During the 1999 recession, public sector investment fell by 9 percent. In Projections for the current account deficit from 2003-07 are in the range of-0.8 to -3.7 percent o f GNP. 12 2000, it rose again by close to 20 percent before falling by 22 percent during the 2001 recession. In 1994, non-wage public consumption fell by 10 percent. By 2000 it had risen by an astonishing 107 percent before falling substantially inthe 2001 recession. 24. The silver lining o f the current precarious fiscal position is that there i s very little room for fiscal policy to become pro-cyclical during this recovery. However, as the economic situation improves further and the constraints on fiscal policy slacken even a little, it i s important that policy makers display more discipline than inthe past. The cyclical pattern o f government spending should not be dominated by market-leading and generous wage settlements in the public sector, or pro-cyclical hiring policies and fixed investment spending that mops up any spare public funds which should go to reducing the debt. Infact, as presentedinthe next section, the debt dynamics would strongly argue for an over-performance o f the fiscal targets inthe near term. Over the longer term, government spending should shift to an acyclical stance, while social spending should gradually move to a counter-cyclical stance as the institutional basis for target social assistance programs i s developed further. The public expenditure and governance reforms, and strengthening of social assistance discussed elsewhere in this report, are key reforms to support a more optimal fiscal policy mix. 25. Tight fiscal policy will provide additional credibility to monetary policy-already anchored by central bank independence under legislation passed in 2001-and open the way for continued disinflation. As part o f disinflation efforts Central Bank i s planning to introduce inflation targeting and has made necessary technical preparation. Other complementary macro policies include: (i)continued exchange rate flexibility under the floating foreign exchange regime; and (ii) policies to ensure competitiveness including efforts to attract FDI. There active i s an extensive body of research highlighting the importance o f policy credibility to successful stabilization outcomes. The experience o f the Central and EasternEuropeanAccession countries (CEE10) in the second half of the nineties i s instructive. In these countries, the disinflation process enjoyed considerable credibility because it was part o f a broader change in economic regime geared towards EUaccession. As a result, they were able to maintain positive growth on average throughout this transition period. The lesson for Turkey i s that if it can move credibly towards EU standards, there i s a good chance of minimizing trade-offs between recession and disinflation, at least inthe mediumterm. 5 DEBT AND FISCAL SUSTAINABILITY Debt Dynamics 26. Betweenthe end of 1995 and the end o f 2001, Turkey's debt stock more than doubled in terms of the debt to GNP ratio. As Table 1.4 shows, at the beginning o f the period, debt was about 41 percent o f GNP, with over two-thirds o f the stock being external debt. The fiscal costs o f cleaning up the banking sector after the 2001 crisis have weighed heavily on the public debt burden. By the end of 2001, net public debt amountedto 95 percent o f GNP. In2002, the debt stock shrank somewhat to 80 percent o f GNP, but it remained at almost twice its level in 1995, and external debt represented just 40 percent o f the stock. As Table 1.4 indicates, a sizeable 13 fraction of the domestic debt stock is denominated in foreign currency, but this is a relatively new phenomenon. Until2001, the vast majority of Turkey's domestic debt was denominated in local currency and was issued at fixed interest rates. The currency composition of the domestic debt began to change following implementation of a major debt swap as part of the bank restructuring in mid-2001. Now, a substantial fraction of Turkey's domestic debt i s either indexed to foreign currency or denominated in foreign currency. Much of the domestic debt i s now issuedat a floating interest rate. Table 1.4: The Stock of Public Debtin Turkey, 1994-2002 (percent of GNP) 1994 1995 1996 1997 1998 1999 2000 2001 2002 Domesticdebt 14.0 12.2 20.5 20.4 24.4 40.9 39.1 57.2 47.7 FX-denominatedhdexed .. 2.7 20.4 15.3 Floatingrate .. 28.6 20.5 Externaldebt 30.7 29.1 26.0 22.5 19.3 20.1 18.3 37.7 32.1 I TOTALdebt External+ FX-denhdexed .. 21.0 58.1 47.4 44.7 41.3 46.5 42.9 43.7 61.0 57.4 95.0 79.8 Source: TreasuryandWorld Bank Note: `..,indicatesdataeither notavailable or not calculated. 27. One approach to understanding the evolution of public debt in Turkey is to decompose the annual change in debt into various components implied by a standard model of the government budget constraint (Table 1.4)8. Turkey's public debt stock was relatively stable from the end of 1995 through the end of 1998. Essentially, it increased in two steps, from 43.7 to 61.0 percent of GNP in 1999, and from 57.4 to 95.0 percent of GNP in 2001. Why was Turkey's debt stable inthe earlier period? And why did indebtednessrise so quickly in 1999 and 2001? 28. There is a sense in which during the 1995-98 period loose fiscal policy was being accommodatedwith loose monetary policy. On average during the period, the public sector ran Inparticular, we can write the public sector's budget constraint in local currency as: (3) netfinancing = interestpayments-primary balance- seigniorage The change indebt, of course, is the same as net financing.Normally,debt is expressedeither as a percentageo f GNP or GDP for analyticalpurposes.As the appendix shows, the governmentbudgetconstraintcanbe usedto solve for the change inthe debtto GNP ratio. (4) A in(debt/GNP) = interestpayments- primarybalance- seigniorage -growtheffect-inflationeffect-revaluationeffect. Equation (4) captures the ideathat part o fthe government's effective financing comes implicitly from either a (i) growth effect, (ii) an inflationeffect or (iii) a revaluationeffect. These effects canmost easilybe understoodby consideringthe case where net financing is zero, so that no new debt is issuedinthe year. Despitethe fact that the governmentissues no new debt, the debt-to-GNPratio can change for three reasons: (i) ifthere is realgrowth, GNP will rise so the ratio of debt to GNP will fall; (ii) ifsome portionofthe domestic debt is not indexedto the price levelthen nominaldebt will remainconstant, while nominal GNP will rise; (iii)ifthe exchange rate depreciatesat a slower ratethan inflation,the local currency value o fexternalor FX-denominateddomestic debt will rise slower thanthe price level. 14 a very small primary surplus o f about 0.1 percent o f GNP although the overall fiscal balance was very negative, averaging a deficit o f 11.1 percent o f GNP. A substantial fraction o f the public interest burden was eased by the fact that the face value o f debt was being eroded through inflation and real appreciation as captured in the inflation and revaluation effects in Table 4. Turkey printedmoney at a relatively rapid pace inthe mid-1990s. This implied seigniorage that averaged 2.7 percent o f GNP. Turkey was also fortunate to have positive growth throughout the period at an average rate of 6.8 percent per annum. This contributed 1.5 percent o f GNP in implicit financing. These factors combined were sufficient to keep Turkey's debt relatively stable. Table 1.5: Public Debt Dynamics in Turkey, 1995-2001 (percent of GNP) 1995 1996 1997 1998 1999 2000 2001 2002 Change in debt -3.4 5.2 -3.6 0.8 17.3 -3.6 37.6 -15.2 Debt creatingitems A. Interest payments 7.3 10.0 11.0 16.2 22.1 21.9 23.5 16.3 Debt reducing items B. Primary balance 2.7 -1.2 -2.1 0.9 -2.0 2.7 5.5 3.9 Growth effect 1.7 1.5 2.0 0.9 -1.8 2.4 -3.9 4.8 Inflation effect 6.5 5.3 9.2 8.8 8.7 13.8 13.0 11.2 Revaluation effect 4.4 1.9 1.6 2.5 -1.2 3.8 -13.2 10.1 Seigniorage 3.0 2.4 2.9 2.4 3.2 1.8 1.4 1.5 Other (sum o f below) 0.0 0.0 0.1 0.5 0.1 1.6 -18.1 -1.8 Privatization 0.0 0.0 0.1 0.5 0.1 1.6 1.9 0.1 Cost o f financial sector bailout 0.0 0.0 0.0 0.0 0.0 0.0 -20.0 -1.9 Memo items: Official balance (B-A) -4.6 -11.2 -13.1 -15.3 -24.1 -19.2 -18.0 -12.4 I Effective service cost 4.0 7.9 -0.8 5.5 16.8 4.9 22.5 -6.8 Errors and discrepancy 7.6 5.2 -1.0 0.6 2.2 0.6 -1.2 -1.8 Source: IMFreports and team calculations. Notes: Theterm `Other' reflects the fact that there are off-budget items that require or create financing and are not included in the standard measure o f the primary balance. Effective service cost is interestpayments net ofthe inflation and revaluation effects plus the errors and discrepancy. `Errors and discrepancy' equals the debt reducing items minus interest payments and the change indebt. The remaining fiscal costs ofbank restructuring are reflected in separate entries in Table 1.4. 29. In 1999, the various factors that had helpedto keep Turkey's debt incheck inthe 1995- 98 period were reversed. The primary balance deteriorated to a deficit o f 2.0 percent o f GNP. This was largely due to two factors: the economy slipping into recession and the fiscal easing that occurred after the Marmara earthquake in August. The real exchange rate depreciated, implying a revaluation effect of -1.2 percent of GNP, as opposed to an average effect of 2.6 percent o f GNP in the 1995-98 period. These two factors raised the financing burden by 5.5 percent o f GNP compared to the 1995-98 period. Nominal interest rates remained highin 1999, despite a decline in inflation. One interpretation is that this was the result o f a monetary policy stance geared at consolidating the disinflation that began in 1998. Another i s that real interest rate movements were the result o f a loss o f confidence in the domestic debt market. Regardless o f the explanation, the net impact on the real domestic interest burden was substantial-on the 15 order o f 9.7 percent o f GNP. The policy mix and other economic events thus created a financing nightmare: the overall balance deteriorated sharply, but none o f the implicit sources o f financing (inflation, revaluation and growth) helped offset this increase.' With seigniorage revenue already near its natural limit an enormous increase in net debt was inevitable." Almost all o f this newdebt was raisedinthe domestic debt market, or through the state banks. 30. Although there was a slight improvement inthe debt situation in2000 that coincided with a rebound inthe real exchange rate and resumption o f growth, the economy fell into a full-scale financial and economic crisis in February 2001. Much as in 1999, this crisis played havoc with public finances. In 2001, the debt stock rose by 37.6 percent o f GNP. One reason for this increase was that the Government issued new bonds in order to recapitalize failing banks. A rough estimate o f the quantity o f bonds issued for this purpose i s around 20 percent o f GNP." So the direct impact o f the financial crisis only explains about half the increase in the ratio o f debt to GNP. Inother respects 2001 was similar to 1999, but with some important differences. Unlike 1999, the primary fiscal balance did not deteriorate relative to the stance o f the mid 1990s. Indeed, the government successfully implemented a sharp fiscal contraction despite the recession, with the primary fiscal balance improving to 5.5 percent o f GNP. In2001, there was a sharp depreciation o f the Lirabrought on by the financial and fiscal crisis. Inflation didnot bring the price level in line with the exchange rate by the end of the year, so there was a sizeable real depreciation and a revaluation effect on the order o f-13.2 percent o f GNP (compared to 2.6% o f GNP in 1995-98). The revaluation effect was so large that it completely offset the inflation effect on domestic debt, leaving Turkey with an effective debt service burden of 23 percent o f GNP. As in 1999 the government faced an enormous financing challenge. With seigniorage kept to just 1.4 percent o f GNP, and with sharply negative growth, the Treasury was forced to issue a great deal o f new debt. Unlike 1999, when the domestic market absorbed most o f the financing requirement, only about half the new debt was issued domestically as the IFIs steppedup their support. Almost all net financing obtained from the domestic market was raised by issuing debt denominated inor indexedto foreign currency. 31. In 2002, the debt picture improved, mainly due to the economic rebound. Growth returned, and was unusually strong giventhe depth o f the preceding recession. The exchange rate appreciated substantially in real terms, after the overshooting o f 2001. Interest rates moderated somewhat. Together. these factors implied a sharp reduction in Turkey's public debt from 95 to about 80 percent o f GNP. At the sametime as the overallfiscal balancedeterioratedby 8.8 percentagepoints of GNP, the growth effect becamenegativeand deterioratedby 2.7 percento fGNP, the inflationeffect shrank by 0.1 percento f GNP, andthe revaluationeffect becamenegativeand deterioratedby 3.7 percento fGNP. l oItwould have beenvery difficult for Turkey to increaseits seigniorage revenueover the average levelinthe 1995- 98 period o f 2.7 percent of GNP. Estimateso f the seigniorageLaffer curve for Turkey suggestthat with stable money demand, this is very close to the maximal leveloffeasible seigniorage. See the 2000 CEM. The other itemsentry in Table 4 is -1 8.1 percento f GNP. This reflects 1.9percentof GNP inprivatization revenue, minus 20 percento f GNP inrecapitalizationcosts. 16 6 THEQUASI-FISCAL DIMENSIONOFTHE PUBLICDEBT 32. The previous section took an accounting approach to explaining the accumulation o f public debt in Turkey over the period 1994-2002. This section explores the deeper root causes o f the fiscal problem. It suggests that the following factors have been the key contributors to what has become a very serious debt problem (Box 1.2 and Box 1.3). Combined, these three factors have contributed a flow-equivalent fiscal cost o f about 6.5 percent o f GNP per year. Off-budget subsidies to farmers and small businesses that were channeled through the state banking system. Indirect subsidiesto the banking systeminthe form o f implicit and explicit guarantees to bank creditors, which created contingent liabilities that were converted to actual liabilities inthe February2001 crisis. Other losses incurred outside the central government budget. Subsidies throughStateBanks 33. For many years the Government conducted quasi-fiscal operations through the two largest state banks, Ziraat and Halk. These banks provided subsidized loans to the private sector (Ziraat to farmers, and Halk to small and medium-sized businesses), the cost o f which they have not been able to cover from their operations. For example, from 1994 through 1999, Ziraat's cost o f funds was nearly twice its subsidized lending rate on average. The state banks covered their losses from subsidized lending by accumulating claims on the government referred to as unpaid duty losses. Essentially, these non-cash assets preserved the balance sheets of the state banks without providing them with any cash flow. Until 2001, the Government exercised little control over this quasi-fiscal operation. Losses continued to accumulate, as did interest on the existing stock o f losses. As the Government was in principle liable for the losses that were reported, the state banks had little incentive to improve efficiency or seek out more profitable lending opportunities. By the end of 2000, the stock o f non-cash government debt held by the state banks had reached 14.4 percent o f GNP. By the end o f 2001, because inflation eroded the value of some o f these claims and because some claims were replaced by cash debt, the stock of non-cash government debt held by the state banks had fallen to 12.7 percent o f GNP. The fiscalization o f these unpaid duty losses after the 2001 crisis was a major shock which prompted passage o f legislation to forbid the accumulation of further duty losses inthe state banks, at least inprinciple. 17 Box 1.2: Currency Crisesand the Fiscal Costs of Banking Crises In2001,Turkey experienceda twin bankingand currencycrisis: the Turkishlira depreciatedby almost 54 percent, while a substantialportionof the bankingsector was threatenedby insolvency. Inthis regard, Turkey sharedthe experienceof a numberof countriesthat haverecently sufferedthroughsimilar crises. An important aspectof bankingcrises is that they can imposelargefiscal costs on a government (Figure B 1.1.). These costs arise when the government steps into liquidate banks and payoff depositors, or when it helpsto recapitalize banksthat are on the brink of failure. The link betweenbankingcrises and currency crises is establishedthroughfacts andtheory. First, it is frequently the case that banks in emergingmarketshave either openpositionsthat expose them to foreign exchange risk, or they face foreign exchangerisk indirectlythroughcredit risk stemmingfrom their customers' exposure. This means that severe currencycrisestend to magnify existingproblemsinthe bankingsector and canraise the fiscal cost bornebythe governmentto solvethese problems. Second,theory suggeststhat the larger the fiscal costs the governmentmustbear, andthe lower itswillingness to financethese costs explicitlythroughtaxationor expenditure cuts, the greater the degree o f currency depreciationwill be.' Fromthis perspectivethe effects ofTurkey's bankingcrisis onthe economyare relativelyeasy to understand. By the end of2000, state bankshadaccumulated losses amountingto about 16.4percentof GNP. To this, the financial crisis itselfadded an additional 14.4 percentof GNP in2001and a hrther 1.9 percentof GDP in2002 to the government's restructuringburden. Thus, the totalfiscal cost o fTurkey's bankingsector woes amounts to some 33 percentof GNP. Incomparisonwith the crises includedin FigureB1, this puts Turkey inselect company. Figure B1.1: Fiscal Costs of Banking Crises Since 1990 70 I 60 50 40 30 20 10 0 Sources: S&P SovereignRatings Service, Caprio and Klingebiel(1996),Dziobek and Pazarbasioglu(1997), Polackova(1999), and WorldBank estimates. 'See, for example, Craig Bumside, Martin Eichenbaum and Sergio Rebelo "On the Fiscal Implications o f Twin Crises," in Michael P. Dooley and Jeffrey A. Frankel, eds. Managing Currency Crises in EmergingMarkets. Chicago: University o f Chicago Press, 2003. 18 Box 1.3: Contingent Liabilitiesinthe Energy Sector It is importantthat, in additionto publicdebt, the Governmentshouldcarefullymanage contingent liabilitiesinkey sectors. This was not done adequately inthe past andas a result a large stock ofactual and potentialcontingent liabilitieshave built up includinginthe electricityand gas sectors. Electricity. The Treasuryhas guaranteedthe power purchasecontractsbetweenTETAS (the state owned electricitytradingcompany) and privatelyownedpower plants. These contractsspecify the minimum amount ofpower TETAS musttake andthe price. IfTETAS is unableto usethe power or unableto pay for it, the paymentliability would haveto be bornebythe Treasury. The contractsare oftwo types:those with private power plantsbuilt under build-own-transfer arrangements(BOTs) andthose built under build- own-operate arrangements(BOOS). The face value o fall these guaranteedpowerpurchasecontractsis aboutUS$3.0-3.5billionyearly. IfTETAS cannotuse this powerthe Governmentcould intheory be requiredto paythis amount to privatepower producers.AlthoughTETAS shouldbe ableto sell mostofthe power fromthe privatepower producersto meet demand, there remainsa risk that TETAS could fall short onthe revenuerequiredto pay for contractedpower ifretailprices are set too lowor there is no associated obligationonthe electricityretailcompaniesto purchasepower from TETAS at a price adequateto meet its costs. This is particularlyimportant as the contractedpricefor muchofthe privatepower is quite high, especially for the BOTs. IfTETAS incurs a loss, the contingent liability turns into a realcost for the Treasury. The situationrequiresclose monitoringandthe Governmentshouldavoid guaranteeingfuture powerpurchasecontracts. -BOTAS Gas. (the state ownedgas company) has signedcontracts for large volumes ofnaturalgas imports on a take-or-paybasis. Underthese contracts,BOTAS has agreedto take specifiedvolumes o f gas at specified prices andto payfor the gas that it cannottake. The face value o fthese gas contracts is about US$3 billionper year and risingover time. Projectionsindicatethat there is unlikelyto be sufficient demand for naturalgas inTurkeyto absorb the volumesthat BOTAShas agreedto take, even assuming that gas under certain contracts is never delivered. As a result, BOTAS maybe obligedto take natural gas for whichthere is no demand. Whether an actual loss will materializedepends on a number of factors beyondthe scope ofthis report. However,the Governmentneedsto monitorcarefully this potential contingent liability and instructBOTAS to not sign new gas importcontractswhich would aggravatethe . . GovernmentGuaranteesto BankCreditors 34. After the financial turmoil of November 2000, the Government issued a blanket guarantee to the creditors, both foreign and domestic, o f Turkishbanks. While the central bank typically fills a lender o f last resort role inthe banking system, the role o f deposit insurance, and more generally, government guarantees, in ensuring financial system stability has been questioned. International evidence suggests that deposit insurance can, in fact, lead to greater instability.l2 Theoretical arguments suggest that government guarantees can even open an otherwise healthy financial system to self-fulfilling speculative attacks.I3 Regardless o f the advisability o f providing guarantees to bank creditors, doing so creates a contingent liability to the government. In Turkey's case, this contingent liability was realized, and resulted in an estimated fiscal cost equivalent to about 18.2 percent o f GNP. l2 See, inparticular, EdwardJ. KaneandAsli DemirgupKunt"Deposit InsuranceAround the Globe:Where Does it Work?" NBERWorkingPaper W8493, September2001. l3 CraigBurnside, MartinEichenbaumand SergioRebelo"Government Guarantees and Self-Fulfilling Speculative Attacks,'' manuscript, NorthwesternUniversity,September 2001.Forthcoming,Journal of Economic Theoly. 19 35. Like subsidized lending, government guarantees to bank creditors act as a subsidy to private activity. Together, these two distortions to financial intermediation contributed about 57.8 uadrillion TL (32.8 percent o f GNP) to the public sector's indebtedness by the end o f 2002!4 Inother words, they were directly responsible for over one thirdof the stock ofdebt. It i s furthermore arguable that the exchange rate overshootingthat was observed in 2001 would not have taken place inthe absence of the government's contingent liabilities to bank creditors. The fiscal shock associated with the banking s stem bailout may well have beenthe underlyingcause o fthe Lira's sharp nominal depreciation.lY Other Quasi-fiscal Losses 36. Table 1.6 decomposes the public sector primary balance into components. This decomposition indicates that, from 1995 through 2002, the central government ran a primary surplus with the exception of 1997. On the other hand, other entities within the public sector generated net primary losses throughout most of this period. Apart from unpaid duty losses and miscellaneous net items, the remainder o f the public sector ran primary losses averaging 1.2 percent of GNP per year inthe period 1995-2000. Given the real interest rates being paidby the public sector inthis period, it is arguable that these losses contributed about 13 percent of GNP to the stock of debt at the end o f 2001. Judging by the budget outcomes in 2001 and 2002 it would appear that the Government has now obtained better control over entities outside its direct budgetary control, namely the state-owned enterprises and state banks. However, it remains to be seen whether the Government will be able to maintain and consolidate this control over the longer term. Table 1.6: Public Debt Primary Balancesin Turkey, 1995-2002 (percent of GNP) 1995 1996 1997 1998 1999 2000 2001 2002 Overallprimarybalance 2.7 -1.2 -2.1 0.9 -2.0 2.7 5.5 3.9 Centralgovernment 3.4 1.2 -0.2 3.8 1.5 4.3 4.8 2.4 Rest o fpublicsec. (sum of 1) -0.7 -2.4 -1.9 -2.9 -3.5 -1.6 0.7 1.5 EBFs -0.6 -0.2 0.1 0.0 -0.5 -0.2 0.1 -0.1 SEES 1.3 -0.1 -0.4 -1.1 -1.2 -1.5 0.1 1.1 Local government 0.0 -0.1 -0.1 -0.4 -0.8 -0.2 0.0 0.0 1 SOC.Sec., U.I.,etc. -0.6 -0.2 0.0 -0.4 -0.1 0.3 0.5 0.4 Other -0.8 -1.8 -1.5 -1.0 -0.9 0.0 0.0 0.1 Source: IMFand WorldBank Note: `Other' includesprimary unpaidduty losses, whichwere discussedabove. l4 This estimatecombinesthe duty losses o fthe state banks andthe additionalcost of financial sector restructuring incurredin2001. l5 This argumentis consistentwith the fiscal theory ofthe price level, andrecentwork on currency crises. See, for example, Craig Burnside,MartinEichenbaumand Sergio Rebelo"Prospective Deficitsandthe Asian Currency Crisis," Journal of Political Economy, 109(6), December2001, pp. 1155-97. 20 SummingUp the ContributingFactorsto the Debt Problem 37. Totaling up the factors that have been presented above gives a total contribution to the stock o f public debt equaling 59 percent o f GNP. On other words, in the absence of these factors, the stock o f public debt would have been about 36 percent of GNP at the end o f 2001, rather than the actual value o f 95 percent o f GNP. The striking thing about this calculation is that almost none o f the contributing factors to the public debt buildup are items within a standard government budget. They all represent losses due to quasi-fiscal activities over which successive Turkish governments did not exercise sufficient control. This analysis points to a very basic diagnosis for Turkey's public debt problem: quasi-fiscal operations must be brought under control and this control must be consolidated over the medium term. Since the 2001 crisis, the Government has taken considerable action to reign in quasi-fiscal losses as shown in Table 1.6, butthe longer term effectiveness ofthese measuresremains to be seen. To ensure a highquality fiscal adjustment, other structural fiscal reforms will also be needed to underpinthe adjustment. These reforms are discussed in more detail in the following section and in the public sector reform chapter. 7 ENSURINGFISCALSUSTAINABILITY 38. Moving forward, the main question the Government faces i s how to design and maintain i t s fiscal retrenchment. Clearly, fiscal conservatism will be necessary through the short and medium term, in order to help finance the cost of recapitalizing and restructuring the banking system and lower the public debt burden. But important design questions need to be considered: how big does the primary balance needto be, how has the strong primary balance beenachieved inthe last two years, andhow should fiscal policy bedesignedto maintain it inthe future? The SustainablePrimary Balance 39. Although there i s no one standard definition o f sustainability, all definitions are built upon the notion that the public sector must satisfy its lifetime budget constraint, which, stated succinctly, is: (5) initial debt =presentvalue o f future primary balances plus seigniorage. Table 1.7 presents estimates o f the primary surplus needed to maintain fiscal sustainability in Turkey based on a special case o f (5) where the debt to GNP ratio is assumed to remain constant16.To calibrate, Table 1.7 uses end-2002 value o fthe ratio o f debt to GNP, b = 0.798, n + g + r g 16(6) x=- r - g b - l + g l+n+g+Zg m where x, b and m representthe primarysurplus, debt stock andmonetary base, respectively,relativeto GNP, andr, g and n are the real interestrate, realgrowthrate, and inflationrate, which are all assumedto be constant.Equation(6) 21 and sets the monetary base, m = 0.04. Rather Table: 1.7: Alternative FiscalAdjustment Scenarios than using historical inflation rates which are (Required Primary Surplus, YOof GNP) much higher than Turkey i s targeting, the table I I Real assumes a long-run inflation rate o f 5 percent GNP Real Interest Rates per year (z= 0.05). The table presents varying Growth 7 9 11 13 15 17 assumptions regarding real growth and the real 3 2.8 4.3 5.6 7.3 9.0 10.5 interest rate that are motivated by historical 4 2.0 3.5 5.0 6.6 8.1 9.6 experience. Turkey's real growth averaged just 5 1.1 2.7 4.2 5.7 7.2 8.7 under 3 percent between 1988 and 2002 (the &"N: Staff calculations mediangrowth rate was higher)butthe medium-termtarget growth rate is 5 percent, so the table uses a growth rate o f between 3 and 5 percent. From 1995 to 2002, the average effective real cost o f public debt service was about 11.5 percent, but real interest rates have fluctuated considerably. Table 1.7 presents results for a range o f real interest rates: 7 to 17 percent. 40. The analysis presented in Table 1.7 confirms the view that a very conservative fiscal position i s requiredto maintain fiscal sustainability inTurkey. With a 5 percent real growth rate, and a 13 percent real interest rate, a primary surplus o f 5.7 percent o f GNP would be requiredto keep the stock o f debt stable at 79.8 percent o f GNP.17 Notice, however, that if growth were only 3 percent per annum, the primary surplus would needto be 7.3 percent o f GNP. Under the program projections o f 5 percent growth and a primary surplus of 6.5 percent o f GNP, the stock o f public debt to GNP i s projectedto decline only slowly, thereby indicating the needto maintain the primary surplus at this level over the medium term and beyond. To the extent that actual fiscal outcomes exceed the primary surplus target, the needto maintain such large surpluses will diminish more quickly as will the risks inherent in the program. Therefore, fiscal over- performance inthe near term would accelerate progress towards macroeconomic stability. Medium-termDebt Dynamics 41, Table 1.8 indicates the medium-termbaseline projections for Turkey's public debt stock, which imply a decline indebt to 71.7 percent o f GNP by 2008. As the inflation and depreciation rates converge towards program targets and existing debt i s retired, interest flows and the inflation and revaluation effects will deviate from their long-run values, but these deviations do not account for the decline in the debt stock because, under the baseline scenario, depreciation and inflation converge fairly quickly to their assumed long-run values. The modest decline in Turkey's public debt stock inthe projections results from the programmed 6.5 percent of GNP primary surplus exceeding the "sustainable" primary surplus, described above, by about 1 percent o f GNP'*. i s a special case o f (5) in whichthe ratio o f debt-to-GNP is assumedto remainconstantat the level b. This is achievedby runninga primarysurplus o fexactlythe magnitudex. " IMFprojectionsas o fthe 4" rev of Turkey's debt dynamics assume an effective real interest cost of roughly 13 percento fthe debt stock over the period2003-08. '*Basedonthe TL appreciationsince April, inthe 5' programreview scheduledto be completedinAugust 2003, -----.", --,--- - .-. - .- --- .- r------- . 4 .. . 1 . .I._ r'vJ y..""- ---.."--.* .....-*".-------- -,-r 1- -I 11. 1 I. -1 compared to 77 percent inthe 4" reviewbaseline. Ifrealized, this downwardadjustment would improvethe medium-termdebt pathprovidedthe exchangerate adjustmentis maintained. 22 Table 1.8: ProjectedPublic Debt Dynamics in Turkey, 2003-08 (percent of GNP) 2003 2004 2005 2006 2007 2008 Debt stock 77.2 76.8 74.6 73.4 72.2 71.7 Change in debt -2.6 -0.4 -2.2 -1.2 -1.2 -0.5 Debt creatingitems A. Interest payments 16.9 15.0 12.5 12.4 11.7 11.7 Debt reducingitems B.Primary balance 6.5 6.5 6.5 6.5 6.5 6.5 Growth effect 3.2 3.3 3.4 3.4 3.3 3.3 Inflation effect 5.4 3.4 2.4 1.6 1.5 1.5 Revaluation effect 1.1 0.2 0.1 0.1 0.1 0.1 Seigniorage 1.1 0.7 0.6 0.5 0.5 0.4 Other (privatization) 0.8 1.o 0.5 0.5 0.5 0.5 Memo items: Official balance (B-A) -10.4 -8.5 -6.0 -5.9 -5.2 -5.2 Effective service cost 9.0 11.1 8.8 9.7 9.6 10.2 (as a percentage o f debt) 11.6 14.4 11.8 13.2 13.3 14.2 Real GNP growth 5 .O 5.0 5.0 5 .O 5.O 5 .O IDepreciation o fthe TL 17.8 12.0 8.0 5.0 5.0 5.0 Infiation rate 20.0 12.0 8.0 5.0 5 .O 5.0 Source: Staff calculations Note: All values are calculated based on IMFprojections ofthe debt stock for the 4' program review in April 2003, including the exchange rate, inflation and real GNP. These assumptions are indicated inthe table. Improvingthe Quality of FiscalAdjustment expenditure are frequently transitory. In2000-01, Turkey 1999 2000 2001 2002 1999- 2o02 implemented a sharp fiscal Total expenditures 40.5 41.5 48.7 43.7 adjustment that brought the Non-interest expenditures 20.0 20.8 22.2 22.8 2.8 public sector primary balance Total revenues 26.2 28.6 32.3 31.5 5.3 to 5.5 percent o f GNP in 2001 from a low of -2.0 percent o f PSBR 14.3 12.9 16.4 12.2 GNP in 1999. However, in Quasi-fiscaldeficit 10.4 6.3 1.5 0.4 2o02,. the primary AdjustedPSBR 24.7 19.2 18.0 12.6 shrank back somewhat to 3.9 Net interest payments 22.0 21.9 23.5 16.5 percent ofGDP- This suggests Primarybalance -2.0 2.7 5.5 3.9 5.9 Inflationary comp. interest paym. 9.7 9.6 11.1 9.9 that a close examination o f how Operational balance -12.4 -6.9 -5.6 -4.9 23 balances. In fact, non-interest expenditure o f the central government actually have increased significantly since 1999. 43. I s Turkey's fiscal adjustment o f high quality? Of course, a sizeable adjustment was necessary, and it hadto be undertaken inthe short-run. As a result, it i s not surprising that much o f the fiscal adjustment has come from the revenue side o f the budget. Nonetheless, there i s reason to be concemed over the longer-term. As time passes, the Government should try to move away from one-off revenue Figure 1.5: Public Sector Fixed Investment (% ofGNP) measures to more broadly based tax reforms. Furthermore, the Government should make efforts to improve the quality o f the fiscal adjustment by generating expenditure savings through structural measures. Within the budget, there are three structural priorities: (i) the public wage bill, (ii) public investment, and (iii)further reforms to the social security system in 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 order to reduce the extent o f transfers from the budget. Under the public Source: SPO and World Bank. sector reform program presented in chapter 2, a major rationalization o f the public investment program has been undertaken easing the real impact of the reduction inthe investmentbudget following the crisis (Figure 1.5). Initial steps have been taken to rationalize public employment, including Figure 1.6: Public Sector Wage Bill (%of GNP) establishment of a monitoring system I and reduction o f redundancies in state enterprises, which has 12 facilitated a reduction in the public wage bill (Figure 1.6). Progress i s 8 also being made to tackle problems 4 with the tax system under the strategy adopted in late 2001. In 0 many cases, these reforms will not 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 only underpin the fiscal adjustment effort, they will also promote development o f a stronger, more ource: SPO and World Bank. resilient private sector. 44. Looking ahead, further structural reforms are needed to support the fiscal adjustment. Continuation o f the public investment and public employment rationalization programs can support further fiscal adjustment inthese areas. However, the most pressingtask i s to tackle re- emergingproblems inthe social security system. While parametric reforms to the public pension system adopted in 1999 raised the pension age and linked benefits to lifetime contributions, the fiscal sustainability o f the social security system has not yet been secured. After initial improvement in 2000, the fiscal position o f social security system worsened considerably in 2001-02 (Figure 1.7). With the ad-hoc pension increase introduced in January, the deficit could 24 well approach 5 percent o f GNP in 2003. The first reform priority i s to Figure 1.7: Deficit of Social Security System enact pending legislation to underpin (%of GNP) institutional and administrative 3.5% n n I I n reforms designed to improve 3.0% I I collections and expenditure management, contain health 2.5% expenditure, consolidate the legal 2.0% basis for the unemployment insurance 1.5% fund, and facilitate implementation of 1.0% the 1999 reforms. A second priority is 0.5% to introduce further institutional 0.0% reforms to reorganize the social 1995 1996 1997 1998 1999 2000 2001 2002 protection system by function 1 1 H SSK W BAG-KUR 0 ES Est. (pensions, health insurance, social Source: SPO and World Bank. assistance, and employment services), together with a new phase o f parametric reforms notably in Emekli Sandigi-the pension fund for civil servants. A third priority i s to improve management o f the system, particularly financial management including credible efforts to address the massive buildup o f contribution arrears. While the prospects of generating additional short-term fiscal savings relative to the program baseline may be limited, decisive action in these key structural areas (public investment and wage bill, and social security reform) can have a major impact on the quality o f fiscal adjustment and hence on real interest rates. Risksto Fiscal Sustainability 45. Improvements inthe central government's primary surplus are necessary now that debt i s about 80 percent o f GNP, but are they sufJicient to ensure fiscal sustainability? Considering the fact that it was not the primary fiscal position that was responsible for the huge increase inpublic sector debt between 1994 and 2002, it i s clear that a strong primary fiscal position will not be sufficient. The Government must also control off-budget losses by eliminating the system o f duty losses. It must also control contingent liabilities. Probably the four most important contingent liabilities the Government now faces are (i) potential further liabilities from troubled banks and corporations, (ii)future liabilities o f the social security system, (iii)contingent liabilities generated duringthe privatization o f utilities, and (iv) FX risk inthe government's debt portfolio. The first source o f contingent liabilities i s discussed ingreater detail elsewhere inthis report. Clearly the banking system i s an important source o f risk to the government's finances given the events o f 2001. The issues surrounding further reforms to the social security system have been discussed above. The sources o f contingent liabilities arising from the utilities sector were discussed indetail inthe 2000 CEM on Turkey. 46. The exchange rate is an important source o f fiscal risk. As described inan earlier section, the Government has shifted a great deal o f its domestic debt portfolio to foreign currency denominated securities. At the end o f 2002, about 59 percent o f the public debt was either external or FX-denominated domestic debt. This lowers debt service cost relative to borrowing in domestic currency to the extent that, ex-post, the premium on domestic debt turns out to be higher than the rate o f depreciation o f the Lira. Given the Government's commitment to 25 disinflation and fiscal reform, this makes borrowing in foreign currency attractive, but it obviously means that the Government bears an FX risk that it would not have if it borrowed in domestic currency. To quantify this risk, consider the fact that at the end of 2000 only 37 percent o f Turkey's debt stock was external or FX-denominated. With this level o f FX exposure, Turkey's debt stock rose by 13.2 percent of GDP in 2001 due to the revaluation effect. If Turkey's exposure had beenhigher, say at the 2002 level, with 59 percent o f its debt inforeign currency, the revaluation effect would have increased the debt stock by 15.9 percent o f GNP. Another way o f quantifying currency risk i s to ask what would happen to the debt dynamics in Table 1.6, if the Lira depreciated faster by 10 percentage points in any given year. Absent a similar acceleration o f inflation, the net effect would be to raise the stock o f debt by about 3.3 percentage points o f GNP. Given the volatility in Turkey's economy, a depreciation o f this magnitude i s not out o fthe question. 47. Turkey's debt dynamics are fragile and vulnerable to important risk factors including contingent liabilities, FX risk and o f course interest rate risk. The highdegree o f exposure to FX and interest rate risk will be a constraining factor on monetary and exchange rate policy over the medium term. The risk analysis emphasizes the importance of staying the course with fiscal adjustment. Credible fiscal policy will reduce fiscal dominance over time and increase the effectiveness o f monetary and exchange rate policy instrumentsas well. 8 THEMEDIUM-TERM AGENDAFOR GROWTH 48. Turkishpolicy makers face a challenging agendafor sustaininggrowth. The Government must address the structural roots o fmacroeconomic imbalance inorder to preventa recurrence of the "boom-bust" cycle that has characterized Turkey's economic performance over the past decade. Positive public debt and external financing dynamics hinge on rebuilding investor confidence. Over the longer-term the public sector will play a smaller role in the Turkish economy. There i s little room in the budget for a state owned enterprise sector that generates losses to be covered by the budget or for direct or indirect subsidies to the private sector. At the same time, the government can, and must, play a positive role in making it possible for the private sector to achieve high levels o f growth. The Government must also use its limited resources more effectively to meet Turkey's social needs. The medium-term agenda for sustained and equitable growth can be summarized as follows: The path to macroeconomic stability lies in sustaining a high level o f fiscal adjustment for the foreseeable future. Fiscal adjustment must be underpinned by structural reforms and complemented by credible macro policies. Moving credibly towards international-notably EU-standards, will help minimizingtrade-offs betweenrecession and disinflation. Determinedaction to modernize public sector institutions and to improve public governance i s critical to achieving quality fiscal adjustment and durably improving the business climate. Effective government i s also key to realizing Turkey's social and human development objectives through the efficient delivery o f public services. Sustaining and expanding the growth process will depend critically on improved performance of the financial and real sectors. Wide-ranging reforms must be deepened to 26 bolster the financial system, promote corporate restructuring, complete regulatory reforms, and accelerate privatization and market liberalization. Social policies needto be strengthenedto ensure the robust social contract among its citizens that Turkey needs to meet the challenges o f the new century and secure the social development o f the country in light o f the millennium development goals. Priorities for social policy include mitigating adverse consequences o f the ongoing structural adjustments, raising the level o f human capital, and ensuringeffective operation o fthe labor market. Sustained growth will hinge on Turkey's further integration with the global economy. Structural reform i s essential to maintain strong export performance, attract foreign direct investment and bolster confidence. Given continuing vulnerabilities to exogenous shocks, maintaining a policy o f exchange rate flexibility will be important. Sustained progress towards EUaccession would provide a strategic external anchor for Turkey's reform efforts. 49. Deeper structural reforms will support both fiscal adjustment and the private sector response. The structural reforms will underpin fiscal adjustment by rationalizing public expenditure, improving the efficiency o f revenue collection and containing contingent liabilities. The reforms will promote private sector led growth and raise total factor productivity by improving the incentive framework, stimulating competition, and encouraging productive investment. Core structural reforms inthe public sector, banking and finance, corporate sector, and labor market are analyzed in detail in this CEM. Other key elements o f the medium-term structural agenda include completing the agriculture, energy and telecommunications reforms launched since 1999. These reforms were discussed in the 2000 C E M and are therefore not covered further here. Reemergingdeficits inthe social security system have focused attention on the urgent need for further structural reforms inthis area to more effectively implementthe 1999 reforms and introduce deeper changes to the system. The issues surrounding social security reform were also analyzed in depth in the 2000 C E M and the most recent developments are summarized inearlier inthis chapter (para. 44). 27 CHAPTER 2: PUBLIC SECTORREFORMS' 1 INTRODUCTION 50. Turkey's public sector is large relativeto OECD comparatorssuch as Spainand Portugal (Table 2.1), andpublic spendinghas greatly exceededavailable resources, while service delivery has lagged. The underlyingproblems with the structure and management of the public sector in Turkey have been analyzed in detail in the 2000 CEM on Structural Reforms for Sustainable Growthand the Public Expenditureand InstitutionalReview. As shown inthese reports, central to the problem are: (i): weak fiscal fundamentals driven in part by extensive government intervention throughout the economy, and (ii) an under performing system of public sector Table 2.1: Size of GeneralGovernmentin the OECD (as YOof GDP) I 1996 I 1999 Revenue '/ Expenditure */ Deficithrplus I Revenue '/ Expenditure '` Deficit/Surplus Australia 3i 33.1 36.3 -3.2 33.3 34.4 -1.1 Austria 48.2 54.1 -5.9 47.3 50.9 -3.6 Belgium 47.7 53.2 -5.5 48.2 50.5 -2.3 Canada 3` 42.8 46.7 -3.9 43.4 44.4 -1 Denmark 55.3 58.7 -3.4 55.1 54.4 0.7 Finland 51 56.6 -5.6 48.7 49.2 -0.5 France 47.5 53.8 -6.3 48.1 52.1 -4 Germany 43.9 49.1 -5.2 44.5 47.6 -3.1 Greece 44.9 52.5 -7.6 50.2 52.0 -1.8 Italy 44.3 52.6 -8.3 45.2 48.4 -3.2 Japan31 31.7 36.5 -4.8 31.6 42.7 -11.1 Netherlands 43.8 48.3 -4.5 44.2 45.8 -1.6 Portugal 37.4 43.2 -5.8 38.9 42.8 -3.9 Spain 39 44.6 -5.6 n.a. n.a. n.a. Sweden 56.8 62.5 -5.7 57.2 57.7 -0.5 Switzerland 3i 33.6 37.3 -3.7 34.4 37.7 -3.3 TURKEY 4i 28.5 33.9 -5.4 36.0 46.5 -10.5 UnitedKingdom 37.3 43.0 -5.7 39.1 39.1 0 United States " 32.4 35.2 -2.8 32.8 34.3 -1.5 Source: OECD Historical Statistics, 1970-1999 OECD definition of total revenues is taxes on productionand imports, property income receivable, currenttaxes on income and wealth receivable, social contributions and other current transfers. Z iOECD definition of total expendituresis the current disbursements plus gross capital formation, acquisitions less disposals of non-producednon-financial assets and net capital transfers. 31 The figures inthe second part of the table are for 1998 4/ Datafor Turkey from PEIR esimates "Thefiguresinthesecondpartofthetablearefor 1997. management. The impact of many years of weak fiscal discipline and poor expenditure management is also reflected in the composition of public expenditure. The share of public spending on interest payments and wages and salaries in Turkey is much greater than in almost all of its OECD counterparts and priority areas such as education are under-funded. This difference is inconsistent with the objectives of sustainedeconomic growthand equity. 'This chapter drawsonthe PFPSAL-I1andPFPSAL-I11documentsand from inputs o fthe Bank teamworkingon public sector reforminTurkey. 28 51. Turkey's public sector reform strategy has three prongs: The first prong encompasses structural fiscal reforms to underpin fiscal adjustment. To date, the program has emphasized (i) tax reform to improve equity, reduce distortions and broaden the tax base and (ii)public employment policies to address over-employment. These efforts complement broader structural reforms in improving the quality o f fiscal adjustment including financial sector reform, overhaul o f agriculture support policies, introduction o f markets for electricity and gas (aimed inter alia at eliminating the need for government guarantees), and privatization. Given the re-emergence o f large deficits in the social security system, the structural fiscal reforms under public sector reform program will need to be expanded to include a new round o f social security reforms as presented in Chapter One. The second prong o f the public sector reform program involves structural and institutional reforms to modernize public expenditure management (PEM) through actions to improve budget preparation and execution, enhance the transparency o f financial management, and ensure prudent public liability management including containment o f contingent liabilities. These P E M reforms are laid out in the Strategic Frameworkfor Public Expenditure Management Reform prepared in July 2001 as part o f the PEIR and PFPSAL program. The third prong o f the public sector reform program focuses on institutional reform to improve public sector governance. Inthis context, a national anti corruption strategy under the slogan A Transparent and Clean Turkey: Together Hand in Hand was adopted by the Government and published in March 2002. The third prong includes a civil service reform component. 52. There has been significant progress in the enactment o f the legal framework neededto underpinthe public sector reforms. A Public Procurement Law based on the UNICTRAL model was enacted inJanuary 2002 moving Turkey towards compliance with EUrequirements, and the independentPublic Procurement Agency established under the law i s now fully operational. The Public Finance and Debt Management Law was enacted in March 2002 creating the foundation for a comprehensive risk managementframework. Implementation o f medium-termtax strategy i s moving forward with enactment o f a unified special consumption tax in mid-2002 and enactment o f a first legislative package for the direct tax reform in March 2003. A Public Financial Management and Control (PFMC) Law was submitted to Parliament in August 2002 which seeks to provide for a comprehensive presentation of the budget, modernize budget preparation and execution, allow decentralization o f financial control to spending agencies and unifyexternal audit underthe TCA. 53. While there has beengood progress inenacting the legal framework, implementationhas been slower than expected in several core areas. Areas where concrete results have been achieved include tax policy reform, public employment, rationalization o f the public investment program and improvement o f public liability management. What i s most critical inthe eventual success o f the public sector reform program i s a political commitment to and coordination o f the reform process. None o f the ministerial or high-level committees set up under the original public sector reform program was operational during the previous government in 2002. The new Government has taken a somewhat different approach in incorporating the core elements o f the public sector reform into its Urgent Action Plan and assigning clear responsibility for each element to specific agencies. A ministerial committee for enhancing transparency and improving governance was established in March 2003, which will steer relevant actions o f the both the urgent action plan and national strategy in this reform area. More detailed information on the status of the public sector reform program i s presented in the remainder o f this chapter. 29 2 STRUCTURAL FISCALREFORMS Tax Reform 54. The rationalization of Turkey's tax system is a key building block in moving towards a sustainable fiscal stance. The complexity and lack o f transparency o f the Turkish tax system, which i s exacerbated by tax policy instability coupled with high and unstable inflation rates have been highlighted in several detailed studies '. Much o f the complexity and instability arises from the addition o f special and additional taxes and surcharges inresponse to the ongoing effort to contain the budget deficit. In addition, in the high inflation environment, partial inflation indexation in combination with differential nominal tax rates and investment incentives across financial instruments have created highly distorted real effective tax rates across financial instruments and real business investments. Small businesses and low income employees face a hightax burden. Finally, investors have identifiedconcerns with bureaucratic redtape, delays in processing applications for licenses or investment allowances and VAT refund claims, and corruption as negative factors for investment and business operations inTurkey. Tax Policy 55. Over the past decade Turkey has steadily increased its revenue effort. Total public revenues (tax and non-tax revenues) collected by the central government, including social security contributions, have increased from I I 21.3 percent o f GNP in 1990 to 37.6 percent Figure2.1: Turkey:TaxPerformance o f GNP in 2002. Total tax revenues (including local administrations and social 35 security contributions) rose from 18.9 30 percent o f GNP to 31.7 percent over the 25 same period (Figure 2.1). The tax revenue available to the central government 2 20 g (consolidated budget) can be divided into 15 three groups: taxes on income, taxes on I consumption and other taxes. The taxes on income are the personal income tax (PIT), \8 99% 99." +qb ,qO %p .rp" the corporate income tax (CIT), and the 1 ITaxRevenues 0 TaxRevenuesInc.SSII payro1'tax that funds security* taxes on income are the primary source o f LSource:Ministry of I I FinanceandWorld Bank tax revenues, rising from about 10 percent o f GNP in 1990 to 14.5 percent o f GNP by 2001 (Figure 2.2). How does the tax performance o f Turkey compare with other groups o f countries over 1990 through 2000? Box 2.1 shows a comparison o f central government tax revenues in LiamEmbrill, Victoria Summers et al, Turkey: TheAgendafor Tax Reform - TheNext Steps, IMF, FiscalAffairs Department(September2000), and SheetalK.Chand, DavidRobinsonand Jose da Silva-Lopes,Turkey: Reform of the Financial Sector Taxation, IMF (March 1990). IsmailArslanand Sweder vanWijnbergen, TheInvestment and Export Incentive System in Turkey,WorldBank, December 1995. ForeignInvestmentAdvisory Service, Turkey:A Diagnostic Study of the ForeignDirect Investment Environment, InternationalFinanceCorporationand World Bank, February2001, and FIAS, Turkey: Administrative Barriers to Investment, (June 2001). 30 Turkey and other countries. This analysis confirms that the revenue effort in Turkey has grown significantly over the last decade such that Turkey has relatively high revenue effort in its per capita income group. Turkey is currently at the lower end o f the upper middle-income countries interms ofper capita income level. 56. The tax effort in Turkey has been raised through exploiting short-term opportunities rather than sound structural changes that ensure revenue growth in an equitable and Figure2.2: Turkey:BreakdowmofTaxRevenues efficient manner over the long-term. incl.SSI Increases in revenue effort in recent years have largely been achieved through ad hoc special taxes (such as a high excise on mobile telephone services, high withholding taxes on the inflationary component o f interest income) and non-tax revenues, rather than through major improvements in the structure and collection ofthe core taxes. A large proportion of the increase inincome tax revenues was achieved through taxation 1 IncomeTaxes E! ConsumptionTaxes of thehighnominal interest (and high real) Source:Ministry o f Finance and World Bank. revenues on government securities and large increases in social security contributions. The increase in tax revenues on interest income i s transitory and will decline as inflation rates and nominal interest rates decline; and the increase in social security tax rates i s self limiting. Outside o f the special tax (banking and insurance tax) on the banking sector, broader interest income has been subject to limited taxation or exempted, while interest deductions by corporate sector have allowed the deduction o f the inflation component with some limited exceptions. In the last few years, the authorities have increased excise taxes, notably those on petroleum products, in some cases with significant negative externalities and distortions. Distributional effects o f consumption taxes are negative on low income groups. Due to pressing fiscal requirements, consumption taxes have been increased significantly and their share in GNP increased from 3.7 percent in 1990 to 11percent in2001. 31 Box 2.1. International Perspectiveson Tax Effort in Turkey Turkey is classifiedas an upper middle incomecountry with its per capita income falling at the lower end ofthis group. The revenueeffort ofTurkey improvedsignificantlythroughthe 1990s. Currentrevenuesofthe central governmentrose to 26.7 percentof GDP by 2000, up from an average o f 19.3 percentover the 1990-2000period. Whencomparedto lower middle-incomecountries,this revenueeffort is markedlyabovethe currentrevenue average ofthe lower group of 13.6 percent,o fGDP. By2000, the current revenuesinTurkey hadrisen significantlyabove the average for all upper middle-incomecountries of21.3 percent. The currentrevenuesof the central government inTurkey are comparableto those inhigh-incomeOECD countries and inthe Europeand CentralAsia region, ingeneral, but they still fall belowthe average current revenuesof 34.2 percentof GDP inthe member countrieso f the EuropeanMonetaryUnion. Whencomparedwith a rangeo f other relativelylarge countrieswithin the upper middle-incomegroup, the revenueeffort o f Turkey is similarto Brazil, Chile, Malaysia and SouthAfrica, but significantlyhigher thanthat inMexico. Whenthe core taxes are analyzedacross the various groups of countries, the tax effort ofTurkey at 10.3 percentof GDP with indirecttaxes on goods and services is higher than inmost countries and comparableto the countries in the EuropeanMonetaryUnion. Directtaxes are composed ofbothtaxes on income, gains and profitsandtaxes and social securitytaxes (typicallypayrolltaxes). The combineddirect tax effort ofTurkey in2000 o f 13.9percent is significantlyabovethe average inthe upper middle-income groupof7.2 percent and similarto the averageeffort in Europeand Central Asia o f 11.6 percent,but not as highas inthe high-incomeOECDcountries and especiallythe EuropeanMonetaryUnioncountrieswhere the direct tax effortto 21.7 percent of GDP. Ofthe selectedupper middle-incomecountries, only Brazil has a higher social securitytax effort, but SouthAfrica has a higher combined direct tax effort of 13.9 percent. Ingeneral, these tax effortdata show that Turkey hasthe potentialto collectsignificantlyhigher shareso fGDP in combineddirect taxes as per capita income levelsrise inthe future, particularlythroughgrowth informal sector employment and realwages as has occurredinhigh-incomecountries, ifthe tax structures are appropriateto achieve this. Table B 2.1: Central GovernmentTax Revenues(Share of GDP, %) Taxes on Taxes on income, Social goods Current gains and security and revenue profits taxes services Turkey (2000) 26.7 8.3 5.6 10.3 Lower middleincomecountries 13.6 2.7 0.4 4.2 Upper middleincomecountries 21.3 4.1 3.1 6.8 Europeand Central Asia 26.1 4.1 7.5 9.8 High incomeOECD countries 26.2 8.0 7.0 7.0 EuropeanMonetaryUnion 34.2 9.2 12.5 9.1 Selectedupper middleincome countries: Brazil 24.1 4.1 7.6 5.2 Chile 22.6 4.0 1.5 10.3 Mexico 14.8 4.8 1.9 8.0 Malaysia 26.0 8.9 0.3 5.9 SouthAfrica 25.8 13.4 0.5 8.9 Source: Ministrv o fFinance.Treasurv and WorldBank 32 57. The prospects for sustainable increases inthe revenue effort from the current already high levels are limited. Turkey has a difficult tax base out o f which to collect taxes relative to the typical OECD country. It has a high share o f employment in the agricultural sector (about 45 percent), a large share o f employment i s at the minimum wage (about 40 percent) and it has relatively large employment inthe informal sector. Low-income taxpayers are already bearing a significant burdenthrough indirect taxes, social security payments and the income tax. Tax rates are already fairly high and such that further rate increases will not be very productive. For example, raising the VAT rate from 18 percent to 20 percent will only raise about 0.7 percent o f GNP, assuming the same level o f administrative effectiveness as currently prevailing. Giventhe 2001 recession and Turkey's generous tax incentives, a large share o f corporations are experiencing tax losses. Loss carry forwards are dampening recovery in corporate tax revenues even though the economy has started to rebound. In the short-term, these loss carry forwards will also dampen the effects of reducing the investmentallowances and bringing a truer measure o f economic income into the tax base. 58. In light of the problems outlined above, core objectives for tax policy reform emerge. Tax structures should be amended to: (i) reduce tax distortions, complexities and uncertainties, particularly in tax treatment o f capital income, both in the context o f the current inflationary environment and in a medium-term environment o f stable prices; (ii)increase tax policy transparency and stability in order to enhance investor confidence and lower the perceived costs o f investment; (iii)enhance equity-shift the tax burden away from low wage employees, and (iv) move towards an internationally competitive tax environment in the context o f EU accession. If the tax structure, particularly of the income tax is amended such that broad measures o f income are made subject to tax (along with a reduction inthe burdenon low income wage earners), then there i s still significant potential for further improvements in income tax revenues as the economy grows which can provide scope for rebalancing the tax structure. Comparison o f the direct tax yield from social security payroll taxes, individual and company taxes show that these revenues can grow to yield in excess o f 20 percent o f GDP as per capita income levels grow to higher levels. This emphasizes the importance o f structuring the core taxes on income and consumption to capture the future growth inthese tax bases. Restructuring the taxation o f corporate and capital income towards taxing a broader and truer measure of profits and gains would come about if investment tax incentives are moderated and more investment income i s brought into the tax base. Tax Administration 59. Reforms are urgently needed to improve the efficiency and effectiveness o f Turkey's tax administration in order to broaden the tax base and increase collections at current or lower tax rates. The tax administration i s one o f the larger administrative services in Turkey with more than 1,000 offices throughout the country and a staffing level o f more than 40,000. It i s responsible for collecting the bulk o f government revenues. However, the General Directorate o f Revenues (GDR) i s placed at a relatively low level within in the organizational structure o f the Ministry of Finance as one of six general directorates. The importance is not sufficiently recognized in the organizational structure o f the Ministry and the tax administration lacks a sufficient number o f senior management layers. The organizational structure for tax administration should be upgraded inline with the practice inother OECD countries (Table 2.2). The complex organizational structure of tax administration headquarters, with a large number of 33 individual units structured around different types o f tax with some elements o f a functional organization, has proven to be an impediment to building managerial capacity in the key operational functions o f tax administration. Moving to a more streamlined and functional organization at the headquarters level i s the first condition for strengtheningoperational capacity. The restructuring andconsolidation of the regional and local office network will then be required to properly implement the modernized compliance management strategies. These institutional reforms will facilitate the GDR's efforts to develop appropriate strategies, performance benchmarks, and monitoring instruments to improve its enforcement and service capacity. Source: IMF and World Bank. 60. Strengtheningtax audit capacity stands out as a crucial priority. Although tax audit is a key element o f compliance management and thus an important tax administration responsibility, the audit function inTurkey is largely outside the control of the GDR. Efforts must be made to integrate and better coordinate the tax audit function inthe tax administration. This will require substantially increasing the capacity o fthe GDR to determine criteria for audit selection based on risk assessment and the development o f a more comprehensive large taxpayer audit approach. Options should be explored to fully integrate large taxpayer audits into the tax administration. 34 The audit o f large taxpayers requires substantial experience and advanced knowledge o f audit techniques and large taxpayer evasion and avoidance techniques. Building o f sufficient audit capacity will take time to achieve and efforts towards this objective should start now. 61. A number o f other reform challenges will needto be addressed inthe design o f the tax administration reform strategy. The GDR needs to develop a proper internal audit function. Internal audit still plays only a marginal role in the structure o f GDR and a clear separation between internal and external audit responsibilities is lacking. This is a major issue o f concern as according to international experience the risk o f corruption tends to be particularly high in revenue administration. Risk analysis in revenue administrations in other countries highlight the audit function as a particular corruption risk area due to the frequent direct contact between taxpayers and tax officials. Controlling this risk requires clear separation o f external from internal audit responsibilities and creation of a special internal audit unit reporting directly to the tax administration director general. Inthe taxpayer service area, the GDR has beenpreoccupied by issuing rulings on request o f individual taxpayers and a proactive approach to taxpayer service delivery has not been developed. However, the implementation o f self-assessment and promotion o f voluntary compliance requires a well targeted and diversified taxpayer service program. Modern tax administrations also have developed service pledges and clear and ambitious service standards which can be easily monitored. To reach a level o ftaxpayer services which corresponds to good practice among OECD countries, a dedicated taxpayer service function will be required in the GDR. Finally, the problem o f arrears collection from state- owned and municipal enterprises has not been sufficiently addressed. This i s more than a pure tax administration reform issue. It requires cooperation with ministries and other institutions responsible for state owned and municipal enterprises, and a clear government decision to eliminate the protection of public enterprises from tax collection enforcement. 62. As part of the tax reform agenda, the institutional arrangements for tax policy making in Turkey mustbe revisited. Incontrast withthe vast majority o f OECD countries (the only notable exceptions being Belgium and the United Kingdom), the Ministry o f Finance has no organizational unit responsible for tax policy formulation. Responsibility for tax policy i s dispersed among numerous departments in GDR and overlaps with administrative responsibilities for implementation. The quality of tax policy making also i s negatively affected by the fact that the existing capacity for tax modeling and revenue forecasting is inadequate. A reorganization o f the tax policy function i s required. The M O F does not intendto separate the responsibility for tax policy from implementation or to create a tax policy function outside the GDR. Therefore, the organizational reform ofthe GDR must include consolidationo foverall tax policy responsibility inthe tax policy unit and strengtheningo f this unit as part o f the functional reorganization. The tax policy unit needs to develop the capacity to analyze and forecast tax revenues based on samples o f actual tax payer information. This i s o f critical importance to reform o f the complex tax incentives offered business sector and tax treatment o f financial instrumentsand institutions. Medium-termTax Strategy 63. The Governmentadopted a medium-termstrategy for improving the tax systeminTurkey in January 2002. The overarching objective o f the strategy is to improve the stability, transparency and equity o f the tax system through measures to minimize tax distortions, broaden 35 the tax base and improve the efficiency of tax administration. This will allow an equivalent amount o f revenue to be raised at lower marginal tax rates thereby promoting economic growth. The strategy builds on the 1998 tax reform and draws on earlier analysis o f the tax system by the Bank, IMF and Foreign InvestmentAdvisory Service (FIAS). The tax strategy lays out a series of coordinated actions over the course o f the 2002-04 period to address certain serious tax policy and administration issues facing Turkey. The priority objectives are summarized inBox 2.2. Box 2.2: Tax Strategy Priorities Tax Policy Simplify and consolidatethe indirecttax structurewith a VAT with a standardrateand two low rates (1 percent for agriculture and 8 percentfor other basic consumptiongoods). Introducea unifiedSpecialConsumptionTax that will consolidatethe current range o f excise and specific taxes into a single tax charged on a limitedrange o f luxury goods. A threshold will be establishedfor obligatory filing of VAT returns. Earmarkingo f revenues will be abolishedand all revenuesfrom earmarkedtaxes will be brought intothe budget. Simplify and consolidate the direct tax regime in line with OECD standards and international best practice. Convert tax rebate for wage earners into a tax credit to simply administration and reduce tax burden on low income wage earners. Move to incometax base to a more globalmeasure of incomeand away from a reliance on final withholdingtaxes. Minimizetax exemptionsand exceptions. Harmonizerealeffective tax rates across all types of financialinstrumentsand maturities. Phase in the full indexation of the corporate and business income taxes (real interest deduction to replace nominal interest deduction). Reduce the scope o f investment incentives, consolidate into a single investment allowance rate across sectors and geographic regions, and gradually eliminate withholding tax on these allowances. Makethe incentivesystemmoretransparentand automatic. Overhaulthe tax regime for free trade zones to removeunintendeddirect tax losses and limit indirect tax losses through transfer pricing. Terminate the corporate tax and payrolltax exemptions. Provide reduced corporate tax rate only on conditionthat separate subsidiaries are establishedand inwardtrading is below 15 percent of sales. Adjust taxation of financial leasingin line with internationally generally acceptedaccountingpractices. Tax Administration Reorganizethe GDRto introduce functionalstructure in line with OECD practice. Revise payment scheme for managers and key technical staff and design a bonus and incentive system which allows more competitivetotal remunerationfor these positions. Upgradethe tax policy departmentwithin the GDR. Strengthen tax audit management capacity within the MOF by establishing an audit coordination unit and upgrade the role o f the audit department in the GDR. Prepare a coordinated annual auditingplan. Develop databasesfor tax audit planningandthe use of informationfrom the tax identificationnumber (TIN) systemfor effective auditing. Design a more comprehensive large taxpayer audit strategy and improve knowledge and databaseon largetaxpayers andtheir businesstransactions. Develop a strategyfor taxpayer services and improvementsin relations betweentaxpayersandthe GDR. Upgradeinternalaudit functionfor the tax administrationreportingdirectlyto the heado f GDR. Designand implementprogramto reducetax arrears. 0 Upgradethe management informationsystemsthroughoutthe GDR. 64. Implementation o f the medium-term tax strategy i s progressing well. O n the tax policy side, a unified Special Consumption Tax (SCT) to consolidate a range o f excise and specific taxes into a single consumption tax charged on a limited consumption and luxury goods was enacted in June 2002. The first legislative package to address issues in the direct taxation was enacted in April 2003. This legislation: (i) harmonizes taxes on investment income at the declaration stage; (ii) rationalizes the system o f investment incentives; (iii) reforms the system o f 36 personal income tax credits; and (iv) introduces a partial dividend deduction to integrate and lower the double taxation of distributed corporate earnings. This legislation brings Turkey's personal and corporate income tax regimes closer to OECD standards and international best practice. However, in a number o f important areas, further steps will be needed in the future including introduction of a single tax rate schedule for wage and non-wage incomes and global declaration by all individual taxpayers. A second package o f direct tax reform legislation i s under preparation which aims to minimize geographical, sectoral, and other investment incentives-including rationalizing benefits in Free Trade Zones. The authorities are also working to develop proposals to harmonize taxation o f financial income at the withholding stage to address significant distortions betweenthe real withholding tax rates on different instruments. For example, deposits in Turkish lira are more heavily taxed than deposits in foreign currency. Legislation on inflation accounting which will reform ad hoc inflation adjustments in the tax systemhas beenprepared. 65. The strategy also tackles tax administration issues encompassing institutional improvements, automation, transparency, compliance, taxpayer services and tax audit. It i s designed to align Turkey progressively with best practice in other OECD countries. Implementation of the tax administration reforms has proceeded somewhat more slowly than originally envisaged, in large part due to the elections. In addition, the general tax amnesty announced by the new Government in early 2003 poses an important challenge for the tax administration interms o f ensuringthe credibility o f future tax collection, despite the authorities' public commitment against future amnesties. In a positive step, the audit coordination committee has been established and a coordinated audit plan for 2003 has been prepared and approved. Resources have been includedinthe 2003 budget to enhance the audit capacity o f the GDR. Effective implementation of the audit plan will help discourage future tax evasion, but broader efforts and organizational changes will be required to bring tax administration up to international norms. The new functional structure within the GDR was largely agreed in mid- 2002 and new management positions were established. However, full implementation o f the functional reorganization has been delayed and i s now scheduled to be completed by end-2003, at which point the GDR will start to extend this structure to the local level as well. The functional reorganization will reorient the GDR from a structure based on administration o f individual taxes to one based on the core functions of tax administration including taxpayer registration, taxpayer services, collections, audit, legal, information technology, andHR. As part o f the functional structure, the tax policy unit o f GDR will be upgraded. PublicEmploymentProgram 66. The comprehensive public employment program is part o f the longer-term strategy to modernize the public sector and ensure medium-termfiscal sustainability. The retrenchment in SEE employment supports the policy o f a non-inflationary, sustained shift in the overall SEE balance from deficit to surplus. While reducing SEE employment i s always difficult and entails social costs, the fiscal adjustment it supports will reduce crowding out and promote the strategically important objective of private sector investment and job creation. Under the program, a system o f quarterly monitoring has beenestablished which provides a clear picture of public employment across the general government. The stricter policies on public employment introduced in 2002 are being maintained and strengthened in 2003. These include: (i)a cap on the total number of civil servants inthe central and local government-the ceiling on new hiring 37 inthe consolidated budget in2003 has beenreducedto 35,000; and (ii) limit on replacement a hiring inthe SEEs equal to 10 percent of attrition and no replacement hiring for SEEs requiring finance from the budget. A system o f quarterly monitoring has beenestablished which provides a clear picture of public employment across the general government (Table 2.3). Using company-specific information, an assessment of 45,800 SEE redundancies was completed in April 2002. A circular on implementation of the redundancy program was issued inMay 2002 after consultations with the trade unions. As of end-June 2003, the identified redundancies reduced by 28,102, essentially through voluntary retirements. The Government now expects the remaining identified redundancies to be phased out by December 2003. A new government regulation lifting all existing restrictions on retirementof public sector workers was publishedin March 2003. For this well-targeted attrition, no replacement hiring is being allowed and the resulting unfilled positions are being eliminated. The Government i s also undertaking coordinated efforts to ensure that adequate funding i s in place for the lump-sum severance payments due to retiring and laid-off worker^.^ This is particular1 important for loss-making SEEs which cannot cover these payments from their own cashflow.' Implementation ofthe SEE redundancy program i s beingtracked under the quarterly monitoring system. Table 1/ 2002 Number % o fTotal Number % of Total Consolidated Budget 1,903,473 71.6 1,897,067 73.0 Civil Servants 1,700,000 64.0 1,691,458 65.1 Workers 203,473 7.7 205,609 7.9 SEEs 415,332 15.6 383,064 14.7 Civil Servants 10,079 0.4 9,715 0.4 Workers 405,253 15.2 373,349 14.4 State Banks 47,979 1.8 32,558 1.3 Civil Servants 3,250 0.1 307 0.0 Workers 44,729 1.7 32,25 1 1.2 Social Security Institutions 77,854 2.9 75,536 2.9 Civil Servants 70,34 1 2.6 68,O 16 2.6 Workers 7,5 13 0.3 7,520 0.3 Local Administrations 213,428 8.a 211,702 8.1 Civil Servants 104,587 3.9 102,941 4.0 Workers 108,841 4.1 108,761 4.2 TOTAL 2,658,066 1oo.a 2,599,927 1oo.a Civil Servants 1,888,257 71.C 1,872,437 72.0 Workers 769,809 29.C 727,4901 28.0 Source: Ministry o f FinancechairedPul: dBank. 1/ "Workers" category include contractedpersonnel. Withthe introductiono fthe new unemploymentinsurancescheme,the Bank has recommendedto the authorities to consider a fundamentalrestructuringo fthe severancepaymentsystem over the mediumterm. 5Experienceto date under the Bank's PSSP project providesa roughupper boundfor severance paymentcosts o fup to US$20,000 per worker. 38 67. Looking ahead, a key issue for the Government under the public employment program will be to set overall objectives for the appropriate level and structure o f public employment in Turkey. The policies inplace so far are a useful start but generally have a short-term focus. The public employment program will soon need to be put in a broader strategic perspective. Important input into this strategy formulation will be provided by the functional review o f government currently underway (para. 82). Inthe SEE sector, following implementation o f the on-going redundancy and privatization programs, a determination will be needed as to which enterprises should remain in the public sector (and why) and what employment policy these firms should follow. 3 STRUCTURAL AND INSTITUTIONALREFORMS UPGRADE PUBLIC EXPENDITURE TO MANAGEMENT 68. The key elements of the Government's multi-year strategy for public expenditure management (PEM) reform are laid out in its Strategic Framework for Public Expenditure Management Reform. The framework i s structured into three priority areas in support o f three core objectives. The objectives, commonly referredto ininternational practice as levels 1-3, are: strengthening aggregate discipline (level l), capacity for policy formulation (level 2), building and improving theperformance ofpublic agencies (level 3). The three priorities for PEM reform are to: Reform the processes for budget preparation and execution. This involves steps to: (i) improve the transparency and comprehensiveness o f the budget in line with international standards, (ii)strengthenthe credibility o fthe budget preparation process, (iii) capacity build for policy formulation at all levels o f government, and (iv) realize concrete improvements in the operational performance o f line ministries and agencies through a progressive shift to performance-orientedbudgeting. Upgrade public accounting, procurement and audit functions to ensure adequate financial accountability. This encompasses legal changes to introduce international fiduciary standards, as well as institutional changes to build capacity to implement the new standards and shift from formalistic ex ante control to effective ex post monitoring. Ensureprudent public liability management. This involves legal measures to establish clear lines o f borrowing authority and transparent reporting o f public liabilities, as well as institutional measures to buildup the capacity for modern fiscal risk management. BudgetReform 69. The Public Financial Management and Control (PFMC) Law scheduled to be enacted by end-October 2003 will form the cornerstone o f the legal framework for modern public expenditure management in Turkey. By establishing clear linkages betweenpolicies, programs and (comprehensive) budgets within a consistent medium-term macro-framework, the PFMC law and related reforms will provide the framework for institutionalising the fiscal adjustment process in Turkey. In this sense, the PEM reforms directly complement the structural fiscal 39 reforms which provide the actual instruments o f fiscal adjustment. As a modern public finance act, the PFMC law will define and enable: A well defined framework of responsibility and accountability for all relevant aspects of public finance management Aggregate fiscal discipline through political commitment to a comprehensive and transparent mediumterm fiscal framework consistent with overall macroeconomic stability Strategic decision making on policies and projects within the constraint imposed by the mediumterm fiscal framework The introduction of strategic planning principles within public administration units to ensure that their resource requests are aligned with policy objectives and key performance goals o f such units The gradual shift from a system of central (external) financial control to a system o ffinancial control that i s internal to public administration units Improved operational efficiency o f public administration units through effective internal financial control and audit, greater flexibility in managing resources and a greater performance orientation inresource allocation. 70. The transparency and credibility of Turkey's government budget is improving although the process has proved to be somewhat more difficult and complex than initially envisaged. Progress can be summarized as follows: Budgetary and extra-budgetary funds. All budgetary funds and extra-budgetary funds (EBFs) were closed in the 2002 budget with the exception o f the Support Price Stabilization Fund (DFIF), linked to the restructuring o f the agriculture sales cooperative unions (ASCUS), and five EBFs. Off-budget earmarking has been substantially scaled back in the 2003 budget. However, a more permanent solution i s needed to terminate the system o f earmarked expenditures linked to the closed funds and incorporate their revenues into the general revenue base. This will be accomplished through legislation which i s expected to be enacted by end-June 2003. For the remaining EBFs (excluding the SDIF which i s covered under the financial sector reform program), the following measures have been included inthe PFMC law: (i) submission o f EBF budgets to Parliament for approval, (ii)audit o f EBF accounts by the TCA, and (iii) monthly reporting o f EBF operations on a consolidatedbasis with the central government. These measures will go into effect following enactment o fthe law. 71. Budget classification. Introduction o f the new Government Finance Statistics (GFS) compatible budget classification will improve budget transparency by allowing for a comprehensive and consistent presentation o f the budget across all general government agencies with a complete functional breakdown of expenditure. The new GFS budgetclassification being rolled out in the consolidated budget agencies for the 2004 budget cycle and will become fully operational for the entire general government in the 2005 budget. In order to ensure that this timetable i s adhered to, the MOF issued in July 2003 an action plan and circular showing objectives, actions, and dates from the start o f 2003 to the completion o fthe GFS reform. 72. Budget preparation. Important steps have been taken to improve the credibility o f the budget preparation process. The framework for these actions was established by a High 40 Planning Council (HPC) decision issuedinJune 2001 to accompany the Prime Minister's Budget Call. The HPC decision provided a macro-fiscal framework for 2002 budget preparation consistent with the primary surplus target and established indicative ceilings for both the recurrent and investmentbudgets for ministriesand line agencies based on an indexation formula applied to the actual budget allocations that each ministry and line agency received in2001. As a further step, the Prime Minister's Budget Call for the 2003 budget included individually specified indicative ceilings for each ministry and line agency. 73. Public Investment Program. A Public Investment Review was carried out on an urgent basis by SPO in October 2001 with support from the Bank. The review confirmed that the average completion time i s excessively long, reflecting a lack o f prioritization among projects. In 2001, the average completion time was 12.5 years (more than 30 years in the agriculture sector), a sharp increase from 9.2 years in 2000 given the impact o f the Lira devaluation and higher than expected inflation. Scarce funds are distributed among an excessive number o f projects which leads to inefficiencies in project implementation. The PIP has traditionally included projects that will not be implemented immediately (the so called "trace allocations"), which hampers the credibility o f the program. The 2002 PIP rationalization cut the average completion time to 8.5 years, a 32 percent reduction, which significantly exceeded the 20 percent benchmark set by the HPC. The 2002 PIP: (i) phased large projects with the view to focusing on priority subprojects; (ii) excluded from the investment program numerous projects with "trace allocation"; (iii) terminated a large number o f uneconomic projects; and (iv) reduced by more than half the number o f new multi-year projects included inthe 2002 investment program. Most o f the new multi-year projects were emergency projects (e.g. flood prevention) or projects o f a social nature. Overall, 353 individual projects and 649 subprojects were dropped from the PIP andthe rationalizationhas reduced the total cost o fthe PIP by TL 36.9 quadrillion. Number Total Total Expenditure by Initial Actual Investment AverageTime New Projects of Value of the end of previous 3/ (in percent of o f Completion Projects Projects year Allocation Investment21 GNP) 41 Total Multiyear 1996 5221 8,155 2,629 467 545 3 6 10 8 1449 459 1997 5332 16,597 4,788 1,008 1,298 4 4 I O 7 1428 577 1998 5556 34,084 9,657 2,135 2,378 4 4 10 4 1,574 714 1999 5458 57,126 17,380 3,560 3,649 4 7 I O 2 1,186 305 2000 5321 86,219 26,125 5,905 6,183 4 9 9 2 1,176 249 2001 5/61 5047 142,919 45,885 7,167 7,570 4 3 12 5 1,234 286 2002 71 4414 166,797 66,021 10,590 1 1,047 4 0 8 5 1,066 128 2003 3851 187,110 80,372 12,464 3 5 7 6 1,032 134 ~ 41 outlines steps to strengthen the role o f formal costhenefit analysis in the project approval process and establish a framework for project monitoring and evaluation. An assessment prepared by SPO confirms the reduction inthe average completion time to an estimated 7.6 years inthe 2003 PIP alongwitha continuedreduction inthe numberofprojects (Table 2.4). 75. Budget reforms are also underway to build capacity for policy formulation and improve the performance o f public agencies. The SPO has issued guidelines for strategic planning by key line ministries and departments including guidelines for costing policies. The Government will also initiate policy reviews by selected line ministries and agencies to derive clear statement o f objectives and expenditure priorities inpreparation o f 2004 budget. This will be complemented by action to revive the role o f ministerial departments responsible for policy reviews with support from SPO. The next step i s to finalize these guidelines and implementthem in a phased manner across central government in order to effectively link planning, policy and budgets. The Governmenthas been carrying out a pilot program since 2002 for a group of line ministries and agencies willing to volunteer for reforms that will provide them with greater discretionary control over inputs linked to adoption o f effective internal controls and commitment to improved performance. The pilot agencies have been provided access to resources consistent with their performance plans inthe 2003 budget. Based on the lessons from the pilot, a plan to expand the program will be prepared by October 2003 Financial Accountability 76. Modernization o f Turkey's system o f public financial accountability in line with international standards i s an urgent PEM priority. As detailed inthe August 2001 CFAA, some of the basic building blocks are in place: an established legal framework; reliable, albeit fragmented, accounting processes; regular reporting o f financial results; an apparently independent Supreme Audit Institution, the Turkish Court o f Accounts, which reports to Parliament; a relatively open system o f accountability and plenty o f skilledpersonnel. However, deeper analysis reveals significant problems and lack o f systemic coherence. Complex institutional relationships, multiple sources of public funds, heavy emphasis on ex ante controls and inadequate reporting to the legislature reduce transparency and weaken financial management. Accounting and procurement legislation have not kept up with evolving international standards. Similar to the budget, the audit system i s fragmentedwith many bodies besidesthe TCA carrying out activities analogous to external inspection and audit. 77. Accounting. Efforts to upgrade the public accounting system are steadily gaining momentum under the PFPSAL program after many years o f limited progress. The suy2000i automated accounting system i s operational in virtually all o f the approximately 1,500 sites o f the MOF's General Directorate o f Public Accounting across Turkey. Under this internet based system, the MOF's central accounting data base contains data on all individual transactions wherever they occur in the system. The initiative to introduce modified accrual accounting in compliance with GFS requirements i s making progress. The PFMC law will replace the existing public accounting law (No. 1050) and provide MOF with permanent authority to issue budget coding and accounting standards for all general government agencies. The MOF has prepared a regulation referring to accounting policies and principles, and containing framework accounting standards and a framework chart o f accounts consistent with GFS. This framework chart of accounts will be adapted to meet the specific requirements of individual agencies while 42 maintaining full compliance with GFS standardse6 The new chart o f accounts i s being piloted within consolidated budget agencies including several joint pilots with the GFS budget classification. Following approval by TCA, the new accounting regulation will be published in the official gazette by mid-2003.7 Based inpart on the experiences from the pilots, the timetable for the accounting reform has beenrevised and i s now as follows: (i) introduce modified accrual accounting in consolidated budget entities in 2004, (ii)begin introduction o f the full accrual basis in entities outside the consolidated budget in 2004, and (iii)introduce full accrual accounting for consolidated budget entities by 2007. The government accounting standards board to be established through the PFMC law will be responsible for transforming the framework standards included inthe accounting regulationinto full-fledged accounting standards over time. 78. Procurement. Turkey has moved decisively over the past year to upgrade its public procurement legislation and practices in line with international standards. The new public procurement law was enacted inJanuary 2002. The new law i s based on the UNCITRAL model and moves Turkey in the direction o f compliance with EU standards. The independentPublic Procurement Agency (PPA) established under the law to oversee public procurement and ensure enforcement o f the new procurement standards i s fully operational. Although the new Government initially considered delaying implementation o f the law or amending key provisions, in the end, the new procurement standards came into force on January 1, 2003 as scheduled. The Government i s now working on a set o f technical improvements to the law as well as new procurement legislation for SEESin the public utilities sectors consistent with the relevant EUdirective. 79. Auditing. Effective financial accountability requires extensive modernization o f Turkey's public audit system. The objectives are two fold: (i)clarify institutional responsibilities, promote improvements in audit quality in line with international standards and support the shift from ex-ante controls to ex-post monitoring in harmony with the efforts to improve operational performance; and (ii) expand the scope o f TCA audits to cover the entire general government including local administrations, autonomous agencies, SSIs, remaining EBFs and revolving funds, with the overall objective of transforming TCA into an effective SAI. These objectives will be facilitated by enactment o f the PFMC law. With regard to audit quality, the law will clarify the roles and responsibilities of the line agencies, the MOF and the TCA. The law will establish a framework for decentralizing internal audit to the line agencies in line with EU requirements. It will allow the TCA to discontinue involvement in budget execution and focus on ex-post audits, including performance audits. With regard to the transformation o f TCA, the law will: (a) include all consolidated budget agencies under TCA's annual audit; (b) subject TCA's own accounts to external audit with reports submittedto the Parliament; and (c) expand the scope o f TCA audits to cover the entire general government. As with any major institutional reform, the transformation o f TCA into a modern SA1must be carefully designedto build consensus with the public administration. The TCA is committed to undertake internal reforms to align its institutional structure with international standards for SAIs, to upgrade its 6 The new chartof accounts is basedon an economic classification. The coding for the functional, institutionaland 'financialclassificationsarecharto beingbuilt intothe accounting softwarerunningunder say2000i. This will be a framework f accountswhich will be adaptedby MOF to meetthe specific requirementso f individualagencies while maintaining overallconsistency with GFS standards. 43 audit capabilities, and to reach consensus with the other government audit bodies on implementation o f the reform. A draft action plan for the TCA's internal reform has been prepared and will be improved, including through a peer reviewby auditors from other European SAIs. The action plani s expectedto be adopted by the Government and TCA by October 2003. PublicLiability Management 80. The new law on Public Finance and Debt Management, enacted in March 2002, has established the foundation for a comprehensive risk management framework in Turkey. The new law: (i)confirms that the Treasury is the single borrowing authority for the central government; (ii)creates a more transparent framework for the issuance o f government guarantees; and (iii) establishes a risk account in the budget, starting with the 2003 budget, to cover the expected fiscal cost o f called guarantees and unexpected losses from other fiscal risks. Inaccordance withthe law, more stringentregulations for the issuance ofgovernment guarantees and for borrowing by SEESwere issued in April 2002. The Treasury has established a new office (the "middle office") for public liability and risk management. The circular on establishment o f the middle office and a Public Debt Committee to set policy within the Treasury, originally published in September 2002, was reissued in January 2003 with a strong emphasis on the risk control function of the middle office. The latter has launched monthly fiscal risk monitoring notes for the Treasury Management, and i s currently implementing systems to formulate strategic risk benchmarks and medium-termborrowing scenarios for the portfolio of direct government debt. The middle office has established a valuation model for contingent liabilities and a tool for the formulation o f strategic portfolio benchmarks and medium term borrowing scenarios. The Treasury publishedin late April 2003 its first Quarterly Fiscal Risk Report covering risks coming from both its direct debts and contingent liabilities, This new report i s the main vehicle for the dissemination o f the Treasury's liability management strategy to other Government agencies, the legislature, investors, and the general public. A review o f the government guarantee portfolio was completed in May, providing a preliminary valuation. The new structure for public liability management i s expected to be fully operational by the end o f 2003. By that time, the Treasury will have produced an operational valuation o f Turkey's portfolio of explicit contingent liabilities and an updated list o f its implicit contingent liabilities. On the basis o f these assessments, the Government will be in a position to issue its public liability strategy as an annex to the planned medium-term expenditure framework (Box 2.3)8. 8 Liability management related issues havebeen explored in detail inthe World Bank Publication edited by PolackovaBrixi and Schick (2002). 44 Box 2.3: The Managementof FiscalRisks inTurkey Over the past 20 years mostOECD countriesand manyemergingeconomies have introducedwide-rangingreforms inthe area ofpublic liability management.Firstand foremost,these changes affectedthe managementof direct governmentdebt. Public financial managers havebeengiven greater operationalflexibility and portfolio managementpracticespioneeredbythe private sector have beenadopted. With regardto managementof off-budget liabilities, the bestpracticesare fewer and more recent. New Zealand ledthe way inthe mid-1980sby adopting accrual-basedaccountingand budgetingstandards practicedbythe privatesector; Canada(1986) andthe USA (1989) have introducedcomprehensiverisk controls for governmentguarantees. Most countries, includingmany emergingeconomies,have integratedthe managemento fguaranteeswith the managemento fdirect debt, however, only few ofthem (e.g., USA,Netherlands,Sweden) are applyingmark-to-marketvaluations for their guarantee portfolio.Also inthe USA and now inother OECD countries, actuarialvaluation methods have been introducedfor governmentinsuranceschemes. Inmost countries,other categoriesofcontingentliabilitiesare still tracked and reportedonly in nominalterms, andtheir issuanceand managementare generally not filly transparent. The 2001crisis revealedseriousdeficiencies in fiscal risk managementpracticesinTurkey. Inresponseto these challenges, the Governmenthas initiated wide-rangingreforms inpublicfinance management, municipal sector, financial system, energy, agricultureand other relevantsectors all o f which will improve fiscal risk management. Reformsfocus on two large classes ofrisk:those presentinthe direct debt portfolioandthose comingfrom off- budget liabilities.Public finance management is beingstrengthenedsignificantlybythe promotionof systemic legal acts suchas the Law on Public Debt andRisk Management(April 2002), andthe PublicFinanceManagementand Control Law (scheduledto be adopted by end-October2003). Importantinstitutionalandtechnical changes for betterpublic liability management(PLM) are also under way: Within the Treasury, a DebtManagement Committeehas beencreated,providingsenior managementwith a tool to formulate an integratedapproach to public liability management, and a DeputyGeneral Directoratein charge o fRisk Management ("Middle Office") has beenestablishedto conductpublicdebt analysisand monitor all fiscal risks.The MiddleOfficehas implementeda guaranteevaluation tool which compares favorably to the best internationalpractices. For direct debt, the MiddleOffice is building quantitative benchmarksto measure refinancing, currencyand interestrate risks more effectively The Treasuryhas strengthenedguaranteemanagementproceduresby introducinga guarantee ceilingand a risk account. The guarantee ceilingcovers mosttypes o f guarantees. ManyPLMreforms currently underwayinTurkey compare well to best internationalpractices,while insome other areas Turkey still followstraditional, less effectivepractices. Institutionalstrengtheningo f the Treasurywill help prioritizea more strategic approachto debt creation. Guarantee designnow aims to constrainoverall government exposure and apply morerigorous financial criteriato the local beneficiariesof suchguarantees. The Treasury is also addingto its arsenalo f state-of-the-arttechnicaltools, Itnow implementsthe same guarantee valuation model which was introducedin Sweden over 2000-2002-very few OECD countries operate valuationtools o f equal sophistication.Insum, progressintechnicalmodernizationofTurkey's liability managementhas been impressive. Remainingchallengesprimarilylie inthe consolidation and strengtheningof institutionalmandates for PLM in Turkey, as well as pursuingthe broaderpublic finance management agenda. Prioritiesinclude: Introducinga medium-termfiscal framework and accrual budgetingproceduresthat help revealthe true cost o f contingentliabilities; Matchingthe Treasury's improvedrisk managementcapacity at the budget execution stage with similar improvementsincapacityat the budget formulation stage (MOF) and at the audit stage (TCA). Strengtheningrisk managementpracticesofthe guarantee beneficiaries(municipalities, SEES,public banks); municipalsector reform, ifproperly designedand implemented, holds great promiseof improved PLMpracticeswithin the broaderpublicsector. 45 4 PUBLICSECTOR GOVERNANCE 81. A national strategy to enhance transparency and good governance in the public sector was published in March 2002 under the slogan "A Transparent and Clean Turkey: Together Hand in Hand". The objective o f the strategy is to provide a comprehensive framework for improving governance and reducing political influence over the economy which establishes clear priorities and benchmarks, and empowers and energizes public opinion to fight corruption. Priorities under the national anti-corruption strategy are summarized in Box 2.4. The basic structure and actions o f this strategy have been reflected in the UAP o f the new Government. The ministerial committee for enhancing transparency and improving good governance was established in March 2003. The committee will steer relevant actions o f both the UAP and national strategy in this reform area. The ministerial committee will prepare regular implementation reports for the national strategy within the context o f the UAP. Among a comprehensive set o f substantive actions i s the Law on "Freedom o f Information for Citizens", which has already been drafted and sent to agencies for comments. This law i s expected to be passed by October 2003. Legislation establishing a code o f conduct for civil servants and public administrations i s expected to be passed by end-July 2003. The Government i s counting on the active involvement of national non-governmental organizations (NGO) to improve public awareness about good governance. One o f the major national NGOs, Foundation for Economic and Social Studies (TESEV),has completed two diagnostic surveys co-sponsored by the Bank on corruption covering households and the business community respectively. Work on a third survey covering civil servants was delayed by the elections. Box 2.4: Prioritiesunderthe NationalStrategy to EnhanceTransparencyand Good GovernanceinTurkey's Public Sector Accelerating the completion of ongoing work, within the framework of public administration reform, for institutionalimprovementand financial management. Implementingamendments to the Law on the Contentso f a Declarationof Wealth, Briberyand Anti-Corruption (No. 3628) in order to make these declarations public and to require mandatory audits and public access to the declarationsfor all electedofficials. Enactingregulationsto ensure that inspections o fcampaignfinancing, incomeandthe expensesofpolitical parties and electioncandidates are made available to the public. Amendingthe existing Law on PoliticalPartiesto requirethat politicalpartiesand candidatesmake a public declarationo fcontributionsfrom individuals and legal personsabove a fixed amount. The HighElections Boardwill be responsiblefor publishing and auditingthis information. Implementingamendments to the Law on the Preventiono f Money Laundering Law (No. 4208) to expand the list o f criminal activities giving rise to illegitimate profits and to restructure MASAK in order to provide investigativepowersto MASAK experts in specific areas. Creation o f an Inspection and Audit Services Class and passage o f an Inspection Law to put into effect inspection standards, to allow for a complete restructuring of all inspection and auditing units, and to implement changes to appointmentproceduresto allow for fixed-term appointment of presidents. Amending the Civil Service Law(No. 657) to includea specific employment category for inspectionand auditing services. Establishingspecialisedcourts to facilitate quick resolutiono fcorruptioncases. Foundingspecialisedunits within the security forces and under the supervisionand inspectionofthe Chief State Prosecutorto investigatecorruption cases. Implementingethic agreements inTurkey. 46 82. Civil Service Reform. Sustained improvements in public governance will hinge in large part on effective reform o f the civil service. Indeed, civil service reform features prominently in the national anti-corruption strategy and features in the UAP. Civil service reform is vital to raising the quality o f public services and ensuring the quality o f the fiscal adjustment. The Government has already initiated work on a norm cadre system to establish benchmarks for staffing inpublic agencies which i s now operational within the Ministryo f Education. There is widespread acknowledgement that a more complete, medium-termapproach is neededin order to adapt the civil service to the changing role o f government in Turkey and to internalize the on- going reforms to PEM. In addition to questions about the size and composition o f the civil service in a changing Turkey, the reform will need to address issues such as distorted salary incentives, career development requirements and inadequate operational budgets in most agencies. InFebruary 2002, a ministerial committee was established to oversee the preparation o f a civil service reform strategy and carry out a full functional review o f government in preparation for further decisions about what functions to retain within government and what to devolve to the private sector. Preparations for the functional review were initiated inmid-2002, but then delayed by the run-up to elections. The functional review has now been re-launched underthe responsibility of SPO andthe report will be submittedto the Government inJuly 2003. Responsibilityfor preparation of the civil service reform strategy has now been confided inthe UAP to the State Personnel Presidency under the overall direction o f a Deputy Prime Minister. A timetable of end-2003 has beenset for the preparationand adoption o fthe civil service reform strategy by the Council o f Ministers. 47 CHAPTER 3: BANKING SECTOR 1 INTRODUCTION 83. Over the last decades, public policy contributed to the deterioration o f the Turkish banking system which led to the 2001 crisis. Problematic policies included large public sector borrowing requirements, weak monetary policy which fueled chronic high inflation, use o f tax advantages to favor government debt, deposit insurance that undermined market discipline, lax prudential regulation and supervision, and undercapitalized state banks. Generous deposit insurance and a lax regulatory and supervisory environment encouraged risk taking. Connected lending, high exposure concentrations, and large foreign exchange positions allowed banks to increase risk, while lenient prudential regulations allowed these risks to remain hidden. In response to the 2001 crisis, Turkey has taken a number o f serious financial sector reform actions. These encompass: (i) significant reform o f prudential regulations; (ii) strengtheningo f regulatory and supervisory authority; (iii) restructuringo f private and public banks; (iv) re-capitalization o f the banking system; and (v) debt re-structuring for the corporate sector. The government's financial sector reform program has been supported by the FSAL, and PFPSAL operations. The Bank team has also carried out three major banking reviews since late 1998 which contain a thorough review of the events that contributed to the crisis and the impact o f the crisis on the financial sector. 84. Today, the Turkish banking system i s undergoing rapid change as part o f the banking reform program. As of the end o f 2002, there were 54 banks o f which 40 were deposit taking institutions and 14 were investment and development banks. As given in Table 3.1, around 32 percent o f total banking assets are held by the three state banks, and another 3- percent by foreign and investment banks. With the June 2002 takeover o f a large private bank (Pamukbank) by the Savings Deposit Insurance Fund (SDIF), the share o f banking assets under SDIF control reached to 4.4 percent, leaving 56.2 percent in private banks. Audits o f the banking sector carried out in 2002 revealed that non-performing loans (NPLs) in the private banking system constituted 23.7 percent o f total loans at the end o f 2001, instead o f the 6.8 percent published previously. The audits also showed a dramatic drop in bank profitability. As a result, BRSA assessed that banks faced an overall capital shortfall of close to U S $ 1 billion, excluding Pamukbank, as o f the end of 2001. 85. This chapter addresses four main issues related to the financial sector reform program. These are (i)private bank restructuring and recapitalization, (ii)state bank restructuring and privatization, (iii)credit crunch issue inthe short run, and (iv) changes inthe incentive structure o f financial institutions and increasing the level and quality o f financial intermediation in the long term. The chapter summarizes recent developments in each o f these areas and analyzes the challenges facing the authorities incompleting the reform and restoring the banking systemto its key role of intermediating financial resources for growth. 48 Banking System Total assets (inTrillion TL) 40,988 79,763 117,649 182,265 212,681 CAR in % 12.4 9.1 17.3 16.8 25.4 NPL/total loans 6.7 9.7 9.2 12.6 17.6 Provisions/NPLs (YO) 44.2 61.9 59.8 65.7 64.1 ROA 1.9 -0.4 -0.8 -1.9 1.7 Share in assets (%) 100.0 100.0 100.0 100.0 100.0 Private Banks Total assets (inTrillion TL) 23,596 42,165 56,179 99,946 119,471 CAR in% 12.0 9.0 18.3 10.9 19.7 NPL/total loans 6.9 3.5 3.5 5.3 8.9 Provisions/NPLs (%) 41.2 62.2 63.O 46.1 53.0 ROA 2.8 3.8 2.3 1.2 2.2 Share inassets (YO) 57.6 52.9 47.8 54.8 56.2 State Banks Total assets (inTrillion TL) 14,150 27,104 40,655 57,676 67,831 CAR inYO 8.6 11.8 7.9 26.8 50.2 NPL/total loans 5.3 9.1 11.1 29.0 37.4 ProvisionsiNPLs (%) 30.2 35.1 30.3 66.0 73.9 ROA 0.5 1.o -0.4 0.0 2.4 Share in assets (%) 34.5 4.0 34.6 31.6 31.9 ITotal assets (inTrillion TL) I 667 5,480 12,912 11,736 9,3101 CAR in % 30.0 -7.5 NPL/total loans 49.7 61.9 41.4 64.9 69.4 Provisions/NPLs (%) 80.9 75.3 80.3 89.1 60.5 ROA -19.0 -46.5 -17.9 -43.6 -2.5 Share in assets (%) 1.6 6.9 11.0 6.4 4.4 Total assets (in Trillion TL) 2,576 5,014 7,903 12,908 16,068 CAR in% 21.0 20.3 29.4 29.5 45.6 NPL/total loans 2.1 2.1 1.8 6.6 4.3 Provisions/NPLs (%) 54.0 51.0 51.8 81.0 69.3 ROA 5.8 6.8 4.1 3.3 5.3 Share in assets (YO) 6.3 6.3 6.7 7.1 7.6 Memo items CPI 84.6 68.8 39.0 68.5 29.7 M2/GNP 51.3 44.6 58.8 38.1 Total Bank assets/ GNP 57.2 71.3 77.1 87.9 77.8 Private Credit/GDP 22.0 21.4 22.8 21.8 16.8 Real credit to the private sector, percantagechange -6.6 -14.6 23.8 -26.2 -9.1 Average nominal treasury bill interest rate 106.0 106.2 38.0 99.1 62.7 Av. Ex-ante interest rate on debt/ CPI 32.0 -9.5 32.8 61.7 Spreadon Turkish dollar Eurobonds(in Basis Points) 548.0 550.0 443.0 718.0 654.0 49 86. Regardingprivate bank restructuring and recapitalization, there has beengood progress in the development and implementation of an effective strategy to resolve bank insolvencies. As a result o f the audits, the bank balance sheets are now much more transparent. The recapitalization program for private banks has led the banks to increase their capitalization. Looking ahead, two concerns remain. First, the banking system i s still undercapitalized even after the recapitalization exercise. While regulatory capital appears high, the low level o f economic capital calls for cash injections. The concern arises from (i)the high level o f related-party loans, (ii) fixed assets which represent 39 percent o f total shareholder's equity, (iii) equity participations and affiliated assets which represent more than 46 percent o f shareholder equity, (iv) zero risk weighing for government securities which i s almost as large as the loan portfolio and (v) over 30 percent non- performing loans, o f which only 50 percent are reserved for. A calculation o f "free capital" that excludes affiliated investments, fixed assets, and the unreserved part o f non-performing loans indicates that free capital o f Turkish private banks i s much lower than the regulatory capital adequacy ratio. Second, the banks needto restructure their large non-performing loan portfolio. Part o f the solution i s the Istanbul approach (described inthe Corporate Sector chapter) which i s a cornerstone o f the Government's efforts to revitalize the private sector. It i s not clear if the Istanbul approach will prove adequate to deal with the problem completely. It may be helpfulto provide additional incentives to encourage quick restructuring. Possible measures include further tax relief to banks for provisioning or restructuring loans and accounting relief to corporations for recognizing foreign exchange losses. These should be temporary measures to support the ongoing economic recovery and growth. Another approach would be to create one or more asset management companies, which may be a more effective way o f removing non- performing loans from banks' portfolio. Efforts to create an A M C are on-going. Similarly, progress on asset management and recovery o f banks taken over by SDIF has been slow, partly because the SDIF i s responsible for asset management which i s not ideal. The process could be speeded up by transferring non-performing assets to a new private asset management company to be created. 87. There has been significant progress in state bank restructuring and privatization yet important challenges remain. UntilJuly 2001, nearly a third o f the banking system assets were accounted for by four state banks: Ziraat, Halk, Emlak and Vakif. Emlak has been liquidated, Ziraat and Halk were recapitalized and are being managed by a new independent board on commercial principles. Strategic studies o f Ziraat have shown that the bank cannot prosper by financing agriculture alone, and will need to become a full service rural bank to be profitable in the long run. While a similar study for Halk will be undertaken, there are serious questions whether there will be a separate market niche for Halk since it would face competition from Ziraat, including for rural SME clients. Therefore, consideration should be given to ensuring viable strategies for these two banks prior to privatization. Both banks need to go through pre- privatization financial restructuring since their very large holdings o f government securities (60 percent o f asset portfolio, compared to 17 percent in loans) makes them unattractive to potential buyers. Spinning off the portfolio o f government securities into a money market fund would shrink the size o f these institutions and speed up the privatization process. Vakif, which was initially thought to be in better condition, was put up for sale. However, the bank was not privatized because it has significant non-performing loans (35 percent), holds too many non- banking assets such as real estate and tourism-related assets, and lacks clear ownership and therefore has no legal freedom to sell assets or recapitalize. Since Vakif i s a large bank (7 percent o f banking system assets), it requires urgent restructuring to make its privatization 50 feasible. Privatization o f state banks i s very important in creating an efficient, growth-promoting financial sector. The critical issue i s to maintain the momentum that has been built up and continue with privatization efforts. An important step would be to reduce Government's reliance on these banks for financing. In the transition to better fiscal discipline, development o f a broader institutional investor base for government securities may help ease the burden on the state banks. 88. There has been a significant contraction o f bank credit after the crisis. An empirical analysis o f whether this was caused by supply (banks beingunwillingto lend) or demand (due to lower levels o f economic activity and high lending rates) was carried out. The results suggest that there was no credit crunch, Le., the demand for bank credit was lower than the supply, although capital regulations did reduce supply to some extent, and households and SMEs were relatively more likely to be credit constrained. Thus, it i s very important to avoid artificial measures to encourage banksto channel credit to undeservingfirms inan effort to put an end to a perceived credit crunch. If necessary, certain temporary measures directed to firms rather than banks could provide relief. These measures could include: (i) schemes that give firms some government backed financing inthe one or two years following the crisis, e.g., equal to a fraction o f taxes paid in the previous few years, (ii) automatic roll-over o f small SME loans, or (iii) a trade finance scheme. Perhaps more importantly, it i s often overlooked that SMEs finance themselves primarily not from banks but mostly by using trade credit. Thus, measures such as allowing receivables to be protected in bankruptcy and allowing borrowing against invoices and accounts receivables are likely to be much more important in relaxing the credit constraints o f SMEs than any measures directedat bank credit. 89. Further changes inthe incentive structure of financial institutions and increasingthe level and quality o f financial intermediation will be important to sustain growth over the medium term. Political commitment to the reform process i s key to sustaining the reforms already made and continuing the reform efforts. Although there are many sophisticated players inthe financial system, Turkey's financial system i s not deep relative to the size o f the economy. One would expect to see higher ratios o f money supply to GNP, greater credit to the private sector, and lower interest margins. One o f the most important reasons for the low level o f intermediation in Turkey is crowding out by public borrowing. Unless the state eases pressure on the banking system to finance its large budget deficits, a financial system that promotes growth cannot be expected to emerge. Reliance on the banking system for deficit financing i s also impeding the privatization o f public banks which i s necessary for an efficient financial sector. If the move towards fiscal discipline and a low inflation environment can be sustained, the banking system's role will take on a much greater importance. Inthe transition to such an environment, however, there are also important risks. Banks will have to start behaving commercially, expanding their credit portfolio and avoiding the potential risks that accompany such credit booms. At present, most o f banks' interest income comes from government securities. To offset the income losses due to lower inflation and the move away from government securities, banks will have to improve their profitability. This requires the remaining regulatory and excess taxation burdens on the banking sector to be phased out. Liquidity and reserve requirements-together with transactions taxes on foreign exchange, stock market and other individual transactions-need to be decreased. Such distortions lead banks to book their assets off-shore. Indeed, some o f the larger foreign banks keep only one fifth o f their Turkish assets and liabilities inTurkey. A credit bureau may also be helpful in allowing the banks to better evaluate credit risk. Currently, the 51 Central Bank only collects information on loan defaults. Finally, the current deposit insurance scheme o f blanket coverage distorts incentives for the banks. The move to a limited deposit insurance system has been delayed to allow for improvements in the financial health o f the banking sector. Designfeatures of the new insurance system, e.g., coverage limit, co-insurance, and risk-based premiums', will be important in reducing the moral hazard inherent in such systems. Table .2: SDIF Banks and Th r Resolution Pr :ess Banks that were taken over by Merged Banks Sales Liquidations Banks the SDIF (those listed below Remaining merged into the bank Under the that is underlined) SDIF Turk Ticaret Nov. 6, 1997 Sumerbank BankEkspres Turk Ticaret Bayindirbank" Bank Expres Dec 12, 1998 Egebank Sumerbank Imarbank Pamukbank Interbank Jan 7, 1999 Yurtbank Demirbank Imarbank Egebank Dec 22, 1999 Yasarbank Sitebank Yurtbank Dec 22,1999 Kapital MilliAydin B. Sumerbank Dec 22, 1999 ----- Bank UlusalBank Esbank Dec 22, 1999 Bayindirbank Yasarbank Dec 22, 1999 - Etibank Etibank Oct 27,2000 Bank Kapital Oct 27,2000 -- EGSBank -- Interbank Esbank Demirbank Dec 6,2000 UlusalBank Feb 28,2001 Toprakbank IktisatBankasi March15,2001 Sitebank July 9,2001 -- IktisatBank KentBank Bayindirbank July 9,2001 Kentbank July 9,2001 EGS Bank July 9,2001 Toprak Bank Nov.30,2001 Pamukbank June 19,2002 Imarbank July 9, 2003 Source: BRSA Note: "Bayindirbankisbeingrestructuredasabridgebankwhichwillperformassetmanagementfunction. 2 PRIVATE BANKRESTRUCTURING RECAPITALIZATION AND 90. Turkey has made very significant progress in bank restructuring efforts. These can be grouped under three headings (i) problem bank resolution, (ii) recapitalization o f private banks, and (iii) asset management and recovery. 91. Problem Bank Resolution. An important component o f the 1999 reform program was the development o f an effective and orderly strategy to deal with problem banks. Despite the financial volatility of the 1990s, problem bank resolution had been slow. In 1994, three small banks were closed and sent into liquidation. From 1997 to early 1999, ownership of three more banks was transferred to the SDIF. Reform efforts accelerated in 1999. Immediately after legal changes were made to the Banking Act, through Law No. 4491, the authorities intervenedin six deeply insolvent banks. This law authorizes the SDIF to participate in financial restructuring arrangements, exempts some o f its operations from taxes, duties and levies, allows it broad freedoms inrestructuring operations, and provides legal protection and rights for SDIF. 'BRSA issueda new regulationon May 15,2003 to introducerisk basedpremiumsystemfor deposit insurance. 52 92. As a result of the 2001 crisis, the financial condition o f private banks deteriorated sharply, necessitating further interventions by the SDIF. By July 2003, additional banks were transferred to the SDIF, bringing the number to 20 banks representing over 6 percent o f total assets o f the banking system (See Table 3.2.). Twelve o f the intervened banks were merged under three transition banks. One of these transition banks and four others were sold to the private sector and foreign investors. Turk Ticaret and Imar Bank have been under liquidation.. The remaining banks are still under SDIF ownership. As of December 2002, the initial cost of restructuring SDIF banks-before any asset recovery-was about 16 percent o f GNP (See Table 3.3 for a breakdown o f fiscal costs o f bank restructuring). This rapid bank resolution process was the result o f a serious commitment to reform and the reforms made to the Banking Act. banks, limiting their foreign exchange open TL trillion % of GNP positions2, and encouraging mergers. The bank Public banks 29,022 16.4 recapitalization process had three phases. Inthe first Ziraat 15,385 8.7 phase, bank balance sheets were audited (first by Halk 12,627 7.2 banks' contracted independent auditor, second by a Emlak 710 0.4 different set o f independent auditors for compliance Vakif 300 0.2 with principles o f independent audits, and finally by SDIF banks 28,784 16.3 BRSA teams). After these audits, the banks' NPLs Total 57,806 32.8 94. The second stage entailed bank recapitalization by owners up to 8 percent capital adequacy ratio (CAR). After the completion of the audit results, Pamukbank, owned by the Cukurova group and one of the top five banks, was found to be seriously insolvent. The authorities considered and then rejected an application by Yapi Kredi Bank, also owned by the Cukurova group, to take over Pamukbank. Excluding these two banks, the capital shortfall was determinedto be on the order o f US$1billion. Over 80 percent o fthis amount was provided by the bank owners themselves. Inthe final stage, those that could not meet the 8 percent CAR requirement were able to qualify for SDIF support if they met the necessary conditions. According to the rules, for banks with CARSthat were positive but less than 5 percent, SDIF was authorized to invest direct capital not more than the additional capital that will be put up by the shareholders to increase the CAR to 5 percent. This option was limitedto banks having an asset share o f over 1percent o f total banking assets inorder to encourage mergers. After the crisis, the incentives introducedby the floating exchange rate regime led to anarrowing o f foreign exchange short positions. Voluntary domestic debt swap operation also contributed to the reduction ofFX open positions inthe private banks. Infact, private banks' open FX positions declined to US$O.8 billion in2002, down from US$ 8.4 billion in2000. 53 Table 3.4: Evolution of Non-performingLoansand Reserves December 1999 Net NPL (TL Billion) 886,030 GrossNPL (TL Billion) 2,327,294 Net NPL / Total Loans (%) 4.4 GrossNPLITotal Loans (%) 11.7 December 2000 Net NPL (TL Billion) 1,459,563 GrossNPL (TL Billion) 3,959,774 NetNPL/ Total Loans (%) 4.6 GrossNPL/ Total Loans (YO) 12.4 December 2001 NetNPL (TL Billion) 6,530,473 GrossNPL (TL Billion) 12,798,377 Net NPL / Total Loans (YO) 17.2 GrossNPL/ Total Loans (%) 33.7 October 2002 Net NPL (TL Billion) 6,3 17,022 GrossNPL (TL Billion) 13,847,lO 1 Net NPL / Total Loans (YO) 14.4 GrossNPL/ Total Loans (YO) 31.7 December 2002 Net NPL (TL Billion) 3,738,769 GrossNPL (TL Billion) 10,429,680 Net NPL / Total Loans (%) 7.1 GrossNPL/ Total Loans (%) 17.6 Source: BRSA. 95. Since recapitalization efforts by the owners did not produce results, Pamukbank was transferred to the SDIF in June 2002. However, SDIF intervention in the bank was legally challenged, and the Administrative Appeals court stayed the execution o f the takeover in November 2002, instructing SDIF to returnthe bank to the Cukurova group on the grounds that the owners were not given sufficient time to recapitalize. However, the owners did not want to take back the insolvent institution and an agreement betweenSDIF and the Cukurova group was finalized inJanuary 2003. According to this agreement, legal proceedings are waived and SDIF will continue to exercise the ownership and management rights of Pamukbank. SDIF will also 54 exercise Pamukbank's shareholder rights (except dividend rights) in Yapi Kredi3. Since Cukurova i s no longer a fit and proper owner, its shares inthe latter bank will be sold within two years. In the meantime, the management board o f Yapi Kredi will have members appointed conditional on approval by the SDIF. A debt restructuring agreement betweenYapi Kredi and Cukurova group continues to be effective. Cukurova group debts to Pamukbank have been restructured according to a fifteen year payment plan and the bank will be available for sale after beingrestructuredunderthe SDIF. 96. The Pamukbank agreement is less than ideal, but given the current economic circumstances in Turkey, it had the merit o f preventing a drawn out legal battle that may have jeopardized even the legality o f past bank takeovers. While the deal removed insolvent Pamukbank from the banking system, it was criticized because it constituted a departure from the rules o f the recapitalization program. The Cukurova group did not have to put up any new capital. By allowing the group's loans to other institutions to be restructured under the Istanbul Approach, the authorities runthe risk o f not completely cleaning up the system. 97. The Turkish bank recapitalization program tried to incorporate many elements o f international best practice (Box 3.1). Developing a long-term approach begins with good diagnostic work, as the Turkish Government did with the aforementioned audits. Second, empowerment o f a single restructuring agency-in Turkey's case the SDIF-is an important step to avoid intra-governmental conflicts; and transparency with respect to that agency's actions has been important at every step to restore confidence. Third, the Turkish program has been comprehensive and credible, designed to resolve all troubled banks as promptly as possible. Fourth, the program relied on market forces to determine which banks should be recapitalized and those that failed the market test were taken over by the SDIF. International experience makes clear that the selection o f banks to be capitalized should be made by markets, not governments and that bank owners should be given incentives to recapitalize and rehabilitate their banks as effectively and as quickly as they can. 98. There are still concerns that the banking system may not be adequately capitalized even after the recapitalization exercise. First, although shareholder equity increased by roughly US $1.5 billion during the fourth quarter o f 2001, only 40 percent o f this came from an increase in paid-in capital. Much o f the rest came from increased asset valuation .4 Second, prior to the recent adoption o f the new regulation, loans extendedto shareholders holding 10 percent or more o f the capital and to entities with connected relationships with those shareholders were deducted from the capital base indeterminingcapital adeq~acy.~ This rule has beeneliminated giving the Yapi Kredi is 41.8 percentpubliclyheldwith all remainingshares belongingto Cukurovawhich held 13.1percent of the sharesthrough Pamukbank. Accordingto the BasleCommittee on BankingSupervision, revaluationsofrealestate holdings(or equity holdings)canbe includedas Tier 2 supplementarycapital, and since Tier 1capital mustcomprise at leastfifty percentoftotal capital, realestate revaluationsmay not exceedfifty percentoftotal capital. Also, the committee emphasizesthat suchreservescanbe includedwithin supplementarycapital, "provided that the assetsare considered by the authority to be prudently valued, fully reflectingthe possibilityo fprice fluctuations and forced sale." (Basle Committee,July 1988,p. 5). It was possibleto circumventthis regulationby separatingownership by an industrialgroup into individual companies, each with less thanthe 10 percentownership ceiling. Thus, loansto these entities did not have be deductedfrom capital. Credits madeto operatingcompaniesof shareholdersdid not have be deductedeither. Nevertheless,this lawwas conservative. banks with a high level of related-party lending greater flexibility incomplying with the capital adequacy requirementsand artificially inflatingtheir capitalratiosrelativeto previousyears. Box 3.1: BankRecapitalization:International Experience Acting promptly, being comprehensive and credible is critical in bank restructuring. Piecemeal solutions almost never work. For example, Japan's attempts to resolve financial sector difficulties through a policy of low interest rates and hidingthe real losses have clearly failed- after two rounds of recapitalizations in 1998 and 1999, market observers still consider the banks to be severely undercapitalized. Fixes such as increasing lending margins or an inflation tax to restore profitability never work and have distorting effects. In Korea, limits on deposit rates have helped develop a curb market, which later formalized into non-bank financial institutions, adding to financial instability. Again, forbearance on capital adequacy targets are often counter-productive. Throughout the 1980s, US regulators used forbearance techniques not to recognize the mounting losses in the S&L industry, which eventually led to a crisis and even greater losses. Clearly, not restructuring is not an option. In the past many countries, including Turkey, Thailand (1983-87), and Indonesia (1994) chose not to restructure, for which they eventuallypaid. Delaysor partialrecapitalizationefforts may backfire since undercapitalizedbanks in a context of broadgovernment guarantees have incentivesto increaserisktakingbehavior since gains accrue to the bank ownerswhereas losses end up with the taxpayers. One common trap in designing restructuring plans is to become pre-occupied with minimizingthe up-frontcash costs o frecapitalizationat the expense ofhigher longer-termfiscal costs andto neglect the incentives that are createdfor the restructuredbanks.Because it was partiallydone, Hungary, for examplehadto recapitalize its banks several times. Venezuela is a similar case with repeated recapitalizations. Even industrial countrieshave not been immune.France's Credit Lyonnaisinthe 1990sis such an example. Once the decision to adopt a comprehensive bank restructuring has been taken, a single agency should be responsible for its implementation, and that agency's actions must be transparent. In many countries, adequate safeguardswere not built in during the evaluationand decision making, andtransparencywas limited. For example, when stronger Korean banks acquired five weaker banks in 1998, the market reacted adversely since details on the government support for acquiringbankswere initially limited. The evidence also indicates that markets rather than governments should determine which banks merit recapitalization. One way o f doing this is to maximizeprivate contributions before injectingpublic capital, andthen safeguarding the government funds by imposing strict measures on recipients.No support should be given to non- viable banks, which should be closed promptly. Hungary's recapitalization attached no conditionality to disbursemento f funds and hadto go through three rounds. Finally, it was sale to foreign investors of 60 percent o f the banking systemthat resulted inadequately capitalizedbanksby late 1990s. Similarly, Japaneserecapitalizations includedonly weak conditionalityanddid not leadto serious improvements. Onthe other handNorway(87-93) and Poland (92-95) included strict conditionality (in the case o f Norway, funds were injected only after shareholder capital was written down, and in Poland government bonds were accompanied by strict penalties for non compliance)and are consideredto be successfulcases, resultinginsustainablefinancialstructures. Not all market-basedrecapitalizationshave been successful, however. Thailand's scheme encountereddifficulties because banks were reluctant to take advantage of it due to requirements that diluted ownership and changed management. The programoffered banks tier 1 capital support conditionalontheir adhering to stringent conditions and making managementchanges. While the approach ensured burden sharing between private and public sector, banks that needed fresh capital could delay their application and were allowedto recapitalize on a flow basis and comply with more stringent capital adequacy requirementsat a later point intime. Thus, banks resistedusingpublic funds and remainsubstantiallyundercapitalizedwith NPLs o f over 25 percentat the end of 2001, 56 99. Since Turkish banks have significant equity participation in affiliates and associated companies, free capital is probably a more relevant Table 3.5: Capital Structure in Private Commercial Banks concept in determining capital Total Private Banks September December December adequacy. Free capital i s (in millions of USD) 2001 2001 2002 defined as shareholders' equity Shareholders'equity 5,507 5,238 9,296 less investment in fixed assets, Less: equity participations, affiliates Equityparticipationsand affiliates 2,512 4,138 4,199 and associated companies. The Premises andEquipment 1,560 2,987 3,365 remaining equity represents the FreeCapital 1,435 -1,887 1,73 1 capital available to leverage a Less: bank's ordinary banking Additionalreservesto reach 100% business. To be even more coverage o fNPL 519 3,828 960 stringent, one could deduct the Adjusted FreeCapital 916 -5,715 772 unreserved amount o f non- TotalAssets 56,394 68,027 73,093 performing loans and/or Equity/Assets(%) 9.77 7.70 12.72 accrued interest on loans. FreeCapitaUAssets(%) 2.54 -2.77 2.37 Given that over 30 percent of Adjusted FreeCapital/Assets(%) 1.62 -8.40 1.06 RevaluationReservesEquity 22.00 11.13 1.45 loan portfolio i s non- performing, o f which only 50 percent i s reserved for, this i s a significant issue. Free capital for private banks as o f Sept. 2001 was substantially lower than the capital adequacy ratio, illustrating these issues (Table 3S). Another issue i s the assignment o f zero risk weighing to TurkishTreasury bills and bonds. Given the large spread on TurkishEurobonds (Table 3.l),may be prudent to attach some risk-weight it to these instruments. The fact that this i s not international practice i s because most BIS regulations are devised for countries with very high rated securities. Beingmore conservative in the Turkishcase would obviously diminishthe capital adequacy ratios o f most institutions since they tend to hold a high percentage o f government securities.6 However, not doing so also has potential consequences as international rating agencies and investors may make their own adjustments to the reported ratios. 100. The above issues underpinconcerns that the banking system is still undercapitalized even after the recapitalization exercise. While regulatory capital i s high, the low level o f economic capital-the inadequacy o f free capital to support traditional banking business-calls for an additional cash injection. International experience shows that partial recapitalization leaves banks fragile, distorts risk-taking incentives, wastes already injected resources and necessitates costlier repeat recapitalizations. 101. Asset Management and Recovery. Restructuring o f the large non-performing loan portfolio o f private banks and asset recovery o f SDIF banks are important concerns. Currently, the SDIF is responsible for managing and collecting the bad loans o f the intervenedbanks. Out of the US$ 10 billionNPL portfolio o f these banks, only approximately US$ 1.7 billion had been Indeed, the proposednew accordstatesthat riskweight on government securities will be determinedat national discretion and recommendshigher weights where needed, so that zero weight does not come as default anymore. 57 collected at the end o f March 2003.' Having asset management as a part o f SDIF has been counterproductive and lengthens the recovery process since public agents tend to be more restricted legally and are less flexible in making deals. Also, such public institutions lack incentives to finish the asset recovery quickly. To deal with the NPL portfolio o f the remaining private banks, the Government has put inplace a regulation on asset management companies to complement the Istanbul Approach. A private A M C with 20 percent ownership by the SDIF i s being considered and the SDIF has initiated feasibility studies. In case a private AMC is established, it may be possible to transfer to it at least some o f the NPL portfolio managed by the SDIF.The legal infrastructure for the establishment of anAMC has beencompleted. 102. It is also important to coordinate financial and corporate restructuring. Key issues are whether (i) to recapitalize banks before or after corporate restructuring, and (ii) restructure to non-performing loans through a centralized collection authority or through the banks themselves. Under an ex-ante recapitalization with appropriate burden sharing, as inthe case o f Turkey, the governmentrecapitalizes banks based on an assessment o f probable losses determined by outside reviews. Some loans may be transferred, at the time o f the recapitalization or later, to asset management companies (centralized approach). For remaining loans to individual firms, the main function o f the bank and other creditors i s to drive operational restructuring, reduce debt to manageable levels, and if necessary provide working capital (decentralized approach). This mixed approach has been used in many developing countries, with success often depending on the degree to which they also enhanced other elements o f the environment for private sector operations such as bankruptcy procedures (Box 3.2). 103. Since the Turkish recapitalization process i s not linkeddirectly to corporate restructuring, it may be necessary to provide additional incentives to encourage quick restructuring. These incentives can take the form o f tax, accounting and other measures. Banks may be given more tax relief for provisioning or restructuring loans. Corporations may be given more favorable accounting relief for recognizing foreign exchange losses. O f course such measures may speed recovery but not necessarily contribute to fundamental reform-as in the case of Korea, Indonesia and Mexico-and should not be made available permanently. Finally, the recapitalization scheme also includes the requirementthat 60 percent o f SDIF assistance for tier 1capital be lent out to non-group institutions by the end o f June 2003. Such pressure on banks may be premature and dangerous in that it may add to the N p L s and erode the newly injected capital. This issue i s addressed inthe credit crunch section. 'At least half o fthe US$ 10 billion were loans to bank owners, which are considered less likely to be collected. Also US$308 million was generated by the sales o f subsidiaries and real estate. 58 Box 3.2: CoordinatingFinancialand CorporateRestructuring: InternationalExperience Ex-ante recapitalization, such as the one Turkey has undertaken, can be implemented quick;y and signal to the market that problems are being resolved. Provided that it is carried out well, ex-ante recapitalization can be an investment which leads to lower costs. However, ex-ante recapitalization also carries big risks. If the recapitalization is not properly structured and bank operations and governance is not changed, the effort may be wasted. Banks will havebetter capacityto work out loans andtake losses whencapitalized, butthey may still delay restructuring and roll over non-performing loans since they are able to attract new funds without difficulty. Similarly, firms may have little incentive to undertake operational restructuring if they have access to new funds. Banks in Japan, for example, did not have incentives to undertake corporate restructuring, as long as they could carry non-performingloans at low cost. These risks are especially high indevelopingcountrieswhere banks are not as strong as the big family-controlled conglomeratesor large state-ownedenterprises. Most banks also do not have the necessary skills andtechnicalcapacity,whichtakes time to build. The alternative model requires tighter links between financial and corporate sector restructuring. Carried to the extreme, this is ex-post recapitalization, where banks receivepublic funds ifand when they provide financial relief to corporations. This model maintains pressure on the banks and corporations to do realistic restructuring, but it introduces significant uncertainty in the meantimesince depositors and creditors cannot be certain about the quality o f their claims. This is why there are no examples o f ex-post recapitizationin practice. One compromisewould be to provide some public funds immediately with further support conditional on progress in corporate debt restructuring. Another important issue is whether to use a central entity such as an asset management company to resolve nonperforming loans. Indonesia, Korea, Malaysia, and Thailand, all eventually used AMCs, and out-of-court systems for corporaterestructuring. All o f them also enhanced to varying degrees the basic legal framework for the private sector including bankruptcy and corporate governance frameworks, liberalization of foreign entry in the financialand corporatesectors. Success with implementationo fthese measureshas varied with MalaysiaandKorea leading, Indonesialaggingbehind, andThailand inthe middle (Claessens, Djankov andKlingebiel,2001). Empiricalevidence on the centralized versus decentralized approach is limitedbut tends to favor the decentralized approach. A study of seven countries showed that most asset management companies did not achieve their stated objectives with only the Swedish AMC successfully managing its portfolio. Rapid asset disposition vehicles in Spain and US also achieved their objectives. However AMCs in developing countries (Ghana, Mexico, and Philippines) were not successful mostly due to weak legal frameworks, underdevelopedcapital markets, lack of professionalmanagement, and politicalpressures (Klingebiel,2001). Dado and Klingebiel(2002) analyzebank-led restructuringin seven countries and find that the success of this approachdepends on the quality o fthe institutional framework, includingaccounting and legal rules, and on initial conditions includingthe capital positions o f banks and ownership links. Norway and Chile were success cases. Poland and Hungary lagged behind Chile, although Poland improvedits framework faster than Hungary. Thailand, Japan andArgentina made little progress. 3 STATE BANKRESTRUCTURINGPRIVATIZATION AND 104. The international evidence indicates that state ownership o f banks has negative consequences for financial development and economic growth (Box 3.3). Comprehensive restructuring and eventual privatization o f state banks i s an important pillar o f the Turkish financial sector reform program. Until July 2001,nearly a third o f banking system assets were accounted for by the four state banks, Ziraat, Halk, Emlak and Vakif (See Table 3.1). Ziraat traditionally focused on agricultural finance, Halk on lendingto small business and artisans, and Emlak on public housing construction. Vakif i s a retail and general corporate sector focused 59 bank.' Prior to the crisis, the state banks-principally Ziraat and Halk-suffered from large "duty losses," subsidized credits to agriculture and small businesses whose losses accumulated on the balance sheets o f the banks. The duty losses were only occasionally paid down from the budget. In 2000, these losses ballooned to 12 percent o f GNP. The continuous liquidity shortages o f the state banks ledto very highinterest rates inthe interbank market with damaging effects for the rest o f the banking sector. 105. As part o f the reform program, Ziraat and Halk were re-capitalized with government securities at market terms and all their duty losses were eliminated (Table 3.3). Emlak was liquidated. A new independent board was appointed to manage Ziraat and Halk in accordance with commercial principles, and to streamline and eventually privatize them. Vakif, because it was thought to have a much stronger client base and a diversified loan portfolio, was put up for privatization. 106. International experience suggests that pre-privatization injection o f funds may decrease the probability o f sale and can be easily squandered. Thus, while the 16percent o f GNP capital injection to the Turkish state banks recapitalizedthem, it i s also important to make sure that this does not slow down the privatization efforts by removing fiscal incentives to sell the banks. As for the private banks, the issue o f new bank lending i s discussed in the credit crunch section below. 107. International experience also suggests that preparing the banks for privatization while improving the regulatory environment i s very important. Turkish state banks are going through this process. Already, they have made significantprogress inreducingpersonnelandthe number o f branches. Diagnostic studies by international consulting firms that have identified necessary improvements in banking operations-such as human resources, technology, and back office operations-which are being undertaken. The state banks will now be run according to commercial principles and there will be no mandated loss-making operations that are not directly paid out o f the budget. These are significant steps. A recently completed pre-privatization study o f Ziraat by international consultants has shown that the bank cannot prosper by financing agriculture alone. Ziraat i s expected to become a broadly focused bank, including developing a presence in the SME market, as well as in consumer finance. This would be similar to the manner in which banks such as French Credit Agricole and Dutch Rabobank have developed. Traditionally, both had their roots in agriculture finance, but over time developed into broad based banks operating across all sectors o f the economy, with their exposure to agriculture falling to less than 20% o f their portfolios. This finding has implications for Halk which i s likely to face competition from Ziraat in its new role. While a similar study i s beingundertaken to assess the future role o f Halk, it i s doubtful that there will be an independentrole for Halk. Vakif has a different ownershipstructurethanthe other three state banksdating from Ottomantimes. Most of its capital is inthe form of shares historicallybelongingto foundationsnow representedby the government's General DirectorateofFoundations. 60 Box 3.3: State BankRestructuring and Privatization: International Experience Empiricalevidence usinginformation from a largenumber of countriesshows that greater state ownershipof banks i s associatedwith less financial sector development, lower growth and lower productivityandthat these effects are magnifiedin countrieswith lower levelso f income(La Porta,Lopez-de-Silanez,and Shleifer, 2000). Generally, ratherthan respondingto principlesofprofit maximization, governmentstendto be responsiveto various interests, andespeciallyto bolsteringpoliticalsupport. Bank privatizationcanyield realbenefits, especially indevelopingcountrieswhere state ownershipof banks is particularlydamagingto growth. However,privatization is a complicatedprocess and, ifnot well-designed, it can leadto crises. Chile (with bankprivatization in 1975 and a crisis in 1982) andMexico(1992 and 1994, respectively) are examples where privatization in a contexto f an underdevelopedregulatory and supervisory framework ledto crisis. The emphasis was on the speedof sale. Althoughthere were multiplecauses, a weak regulatoryenvironmentwas acknowledgedto have ledto fraud, lootingandthen crisis. Generally, it is importantto have a deliberate and credible phasingout of state ownership over some periodduringwhich the environment is being improvedas is beingdone in Turkey. The Argentine bankprivatizationexample supportsthis sequencing. Onthe eve o fprivatization, Argentina hada reasonablydevelopedregulatory and infrastructureenvironment. Indeed,by 1997,Argentina hadone of the strongestbank regulatoryenvironmentsamong emergingmarket economies(World Bank, 1998). Following privatization, the newlyprivatizedbanks' balance sheetsand incomestatements beganto resemble more those o f other banks, with lower administrative costs and greater share of credit extendedto the privatesector (as opposedto state enterprises). Another important lesson from Argentina's privatizationi s the fact that weaker banks were more likely to be sold becauseofthe pressurethey exert onthe budget. By removingthis pressure,at least inpart, recapitalizationo f state banksthrough injectionofpublicfunds can reducethe probabilityo fprivatization. Countriesthat can attract foreign entry from reputableforeign namesare ingeneral ina stronger position in privatizationoperations. Even ifthe regulatory environment is not up to par, strongforeign bankscan bringgood skills, products,and even a capacity to train local bankers, and would presumablybe motivatedto protecttheir reputationsto behave inlinewith the highestfiduciarystandards. Indeed,the share ofbankingassets controlledby foreign banks has soared inseveralcountries inrecentyears. For example, foreign banks have beenvery active in EasternEurope-including CzechRepublic,PolandandHungarywith more 50 percentforeign ownership-and LatinAmerica-including Venezuela, Chile andArgentina, againwith foreign ownership levels of 50 percent. Privatization o fbanks, especiallyintransitioncountries, and sales of failed bankshave providedopportunitiesfor banks from developedcountriesto acquire pre-existingbranchnetworksand enter retailbanking. Until recently, foreign banks hadcommittedvery little capitalto Turkey due to instability inthe macroeconomic environment and fears o f havingtheir investmenteroded by inflation. The prospectofa bettermacro-economic environment, the potential for growth in banking, improvedtransparency inthe system, the needof Turkish banks for fresh capital and good partnership opportunitiesmay be able to lure foreign banksto invest in Turkey. Two significant investmentshave beenthe takeover by HSBCofthe failedDemirbankandthejoint venture between Kocbank and Unicredito. 108. An important factor in allowing a faster pace o f privatization o f the state banks is to shrink their size. About 60 percent o f the asset portfolio o f these banks i s made up o f government assets (with loans making up only 17 percent), making them unattractive for privatization. Thus, privatization will require the government to reduce its reliance on these institutions to finance the budget. It may be possible to spin off this part o f the banks' operation into a money market mutual fund invested in government paper managed by the banks. An alternative approach would be to create a narrow bank invested in government securities whose size would shrink as the government's financing needs are gradually reduced. An additional step incommercializing these institutions would beto limittheir role inprovision of subsidies. First, 61 it is important to ensure that all subsidiesare transparently provided through the budget. Second, all commercial banks should be allowed to distribute these subsidies. In a transition period, some specific role in subsidy distribution could be reserved for the public banks since at present Ziraat and Halk have important information and expertise inthese sectors. However, eventually all banks should be able to compete with each other inthis area on a commercial basis. Inthis way, it would be much easier to privatize and attract foreign capital. Such concrete steps would prevent a very long and drawn-out privatization effort and minimize the chances that the restructuring/privatization o f the state banks runs into political interference. These steps will also preventthe subsidy operations from revertingback to the "duty loss" ~ystem.~ 109. Privatization o f Vakif Bank has not materialized as expected. Privatization efforts made it clear that the bank could not be sold as is for the following reasons: (i)very high non- performing loans (about 35% o f its loan portfolio), (ii)excessive non-banking assets on its balance sheet, such as real estate and tourism related assets, other non-financial subsidiaries, and (iii) ofaclearlydefinedownershipandcorporatestructure. Vakifisarelativelylargebank lack (about 7 percent o f banking system assets) which has an important impact on the financial system. Although it was recapitalized as a result o f the bank recapitalization program, the bank i s still in need o f urgent restructuring. Currently, there are plans for a buy-out by the employee pension fund, which already owns 25 percent o f the bank. While this can be a transitional arrangement, care needs to be taken so that losses are not born by the employees. Another option would be to merge the bank with Ziraat Bank, which already has excess capital after the capital injection by the Government. 110. To summarize, Turkish state banks need to be privatized. While hasty privatizations without proper preparation o f the banks and the regulatory environment can be destabilizing and costly, delaying bank privatization has higher economic costs. The Government has taken important steps in restructuring the state banks, merging and recapitalizing them, eliminating duty losses, establishing an independent board, reducing operating costs, and in general streamlining operations. The risk factors are the reduced fiscal incentives to privatize since the banks already received significant capital injections, potential revival o f the "duty loss" practices giventhe highlevel o f capitalization, and loss o fpolitical will for privatization. The requirement that the banks should make significant new loans may be difficult to meet in this economic environment. This requirement may increase the bad loan portfolio and squander newly injected capital. It i s crucial to continue the privatization efforts o f state banks. To do this, it i s important to prepare specific privatization plans for the banks, to reduce government reliance on these institutions to finance the budget deficit, and to allow distribution o f subsidies through all commercial banks. 4 CREDIT CRUNCH 111. After the crisis in Turkey, there has been a significant contraction o f bank credit (see Figure 1 from Annex). Whether this is due to reduction in demand due to the sharp decline in economic activity and higher lending rates that accompanied the crisis or supply side factors i s 9 To take advantage o fthe improved financial position ofthe banks and to ease the pressure on the public debt burden, some o fthe interest and principle payments due to Ziraat and Halk were capitalized rather than paid in cash. It is important that such capitalization does not affect the liquidity requirements o f the two banks. 62 not clear. A credit crunch can be definedas a reduction inlendingby financial intermediaries to an amount that i s less than the quantity that is desiredby the borrowers, even if they are willing to pay for the cost o f credit and are able to satisfy the non-price terms o f credit required by the lenders. Definedinthis way, a credit crunch represents a rationing o fthe supply o f bank credit. 112. A reduction inthe supply ofbank loans may be causedby an erosion inthe banks' capital position or deposit base, possibly accompanied by portfolio reallocation decisions by the banks. For example, tight monetary policy, by draining the banking system o f its reserves, reduces the amount o f funds available inthe interbank market. Unless this reduction i s offset by an increase indeposits or capital, it will leadto a reduction inloans. Ifborrowers are highly dependent on bank credit for financing their operations, then the cut back in lending will lead to reductions in real spending. The supply o f credit may also shift (or bend back) because as borrower's net worth deteriorates following a crisis and collateral values decline, lenders are more inclined to charge a higher premium on loans to compensate for higher risk and higher cost o f monitoring (financial accelerator view/credit rationing). Since ease o f monitoring and quality o f the borrower i s an important factor here, small, unlistedand more specialized firms are more likely to bear the brunt o f the crunch. Other factors that cause banks to cut back on credit may include an erosion o f their capital base, greater regulatory scrutiny o f bank portfolios and the imposition o f tougher regulatory standards. The stock o f credit may appear to be reduced when authorities require the banks to remove NPLs from their balance sheets as a result o f restructuring operations. Bank lending could also be severely reduced when losses begin to erode the banks' capital to the point where they are unable to comply with regulatory capital adequacy ratios. In order to meet regulatory requirements, banks are forced to reduce their loan assets which require 100percent risk weighing. This latter effect i s generally characterized as a capital crunch. 113. International evidence on credit crunch arising from crisis i s mixed as reviewed in the Technical Annex. The general message i s that, while periods of credit crunch can occur following crises, particularly for smaller enterprises, these episodes are likely to be short. Once macroeconomic activity resumes, bank lending also increases. A disequilibrium framework was estimated to investigate the credit crunch effect in Turkey. The methodology and the estimation results are given in the Annex. The results indicate that in general there has been no credit crunch after the most recent crisis (although there was one after the 1994 crisis) because the demand was much more suppressed compared to supply. However, a downturn in credit supply due to capital regulations can be observed. With regard to credit to households-which includes credit card debt-there i s some evidence o f a credit crunch in the last six months o f 1999. This can be thought o f as a proxy for small and medium enterprise (SME) credit since most SMEs around the world use credit card debt to finance their operations. 114. To alleviate reductions in credit supply-particularly for smaller firms-most countries have eventually followed mildly expansionary fiscal and monetary policies. This i s a luxury Turkey does not have. However, intrying to put an end to these perceived "credit crunches", the most important issue i s to prevent bad firms from getting too much financing. Since the period o f real credit crunch tends to be short, there i s reason to do less rather than more. Simple, transitory rules operating on the firm side may be helpful. These can take different forms: a tax break for corporations; a scheme which gives a firms some government backed financing in the current year equal to a fraction o f taxes paid during the previous one or two years (which i s a crude test whether the firms were making profits, and may thus be viable firms if not for the 63 crisis, and were obeying the tax code); and/or some accounting relief, e.g., on foreign exchange losses. For the SMEs, an automatic roll-over for very small loans, lower capital adequacy weights on SME loans, or some trade finance scheme may be helpful. Indeed, some o f these measures, particularly roll-overs and credit guarantees for SMEs, were successfully applied in Korea. Most SMEs around the around get the bulk o f their external finance not from the banks, but from trade credit. This is also true inTurkey. However, receivables financing inTurkey is hamperedbecause such claims are not protected inbankruptcy. Although it is possible to obtain credit against post-dated checks, it i s not possible to do this using invoices or accounts receivables. A change inthe legal framework to allow re-assignment o f intangibles would go a long way inrelaxing short-term financing constraints o f SMEs. 115. Trying to promote credit expension by taking measures on the banking side may be counter-productive. Generally, resumption o f bank credit i s not a leading indicator. To the contrary, it follows output recovery. Requiring the banks to lend too much too soon may backfire and lead to additional non-performing loans. The private bank recapitalization program requires 60 percent o f Tier 1 capital support provided by the SDIF to be extended as credits to non-group firms by the end o f June 2003. Given the uncertain growth prospects, such requirementsmay endup eroding the capital soon after the recapitalization. 5 CHANGES INTHE INCENTIVEENVIRONMENT AND FINANCIAL INTERMEDIATION INTHE LONG TERM 116. The Government has taken a number o f urgent reform actions after the crisis to improve the incentive environment. Prudential regulations and the supervisory framework for banking have been strengthened. Prudential regulations on loan loss provisioning, large/connected lendingpractices, and foreign exchange exposures were brought inline with BasleEU standards and the criteria associated with these regulations are being phased in. Tax advantages o f repurchase (repo) transactions have been removed and these transactions are now requiredto be reported on balance sheets as collateralized finance transactions. Legal and tax measures are underway to encourage bank mergers. Bank supervision has been made independent and strengthened by the establishment o f the BRSA. The deposit insurance agency (SDIF) has been transferred from the central bank to the BRSA, where it has responsibility for administering the deposit insurance scheme, resolving problem banks, and managing and recovering assets. As discussed above, a major restructuring effort has been started for the public banks with the end goal o f privatization. Twenty insolvent private institutions have been taken over by the SDIF and efforts have been made to strengthen the capital position o f the remaining private banks. The legal framework for creditor rights, bankruptcy and debt workouts is being strengthened. Further legal improvements to allow more effective insolvency procedures and better protected creditor rights are being prepared. 117. These are very impressive developments. Establishing stronger regulation and supervision, capitalizing the private banks and reducing government ownership o f banks are important steps inestablishing an effective, healthy banking systemthat would contribute to long term development o f the Turkish economy. In banking, incentives are key to making sure the regulation and supervision i s not circumvented and the newly injected capital i s not wasted. Any 64 attempt to introduce additional capital without adequate attention to the underlying incentive structure cannot be sustained. Partial recapitalizations are also very dangerous because undercapitalizedbanks have distortedrisk-taking incentives evenunder best circumstances. 118. Comparing Turkish banking indicators with those o f selected countries around the world shows that there i s still significant room for financial deepening. This includes higher ratios of money and bank assets to GNP,more credit to the private sector, higher efficiency reflected in lower interest margins, greater foreign bank share and lower state bank share in the long run (Table 3.6). InTurkey, there are still a number o f factors that distort the incentive environment. In the past, the banking system's main role has beento finance the ever-growing needs o f the public sector by mobilizing internal and external resources. The banks were given incentives such as very high yields on government paper, tax breaks, and deposit insurance to make this possible, and the shifts in their asset and liability portfolio reflected these incentives. Furthermore, state banks-which represent 30 percent o f the banking sector-suffered from liquidity and solvency problems which undermined discipline and promoted excessive risk- taking by the whole industry. International experience shows that such a banking system cannot contribute significantly to economic development. A healthy financial system contributes to economic growth by mobilizing savings and allocating these to the best possible uses, monitoring management decisions and managing risk. Ifthe government's demand for credit i s excessive, there i s not enough room for the banks to perform this resource allocation function. As discussed, empirical evidence shows that the greater the role o f state banks play, the lower the level o f financial development and economic growth, the more concentrated i s credit allocation, and the higher the likelihood and cost of banking crises. Moreover, government reliance on the banking system for financing also impedesprivatization o f state banks. Table 3.6: Turkey FinancialDevelopment - bank creditto Bank net gdplcap the private assets/ interest bank foreign bank state bank (95) m3/gdp sector/gdp gdp margin concentration share share Argentina 8001 0.26 0.22 0.35 0.07 48.00 49.00 30.00 Chile 4992 0.43 0.59 0.61 0.04 59.40 32.00 11.70 Germany 30794 0.76 1.11 1.40 0.02 12.00 4.20 42.00 Greece 11901 0.38 1.03 0.03 70.00 5.00 13.00 Japan 42391 1.90 1.16 1.51 0.02 31.10 5.90 1.15 Korea 11475 0.84 0.65 0.78 0.02 47.50 0.00 29.70 Mexico 3393 0.30 0.18 0.31 0.05 80.00 19.90 25.00 Malaysia 4536 1.24 0.94 1.33 0.03 30.00 18.00 0.00 Turkey 3007 0.39 0.21 0.50 0.10 50.00 5.00 35.00 USA 29253 0.59 0.45 0.52 0.04 20.80 4.70 0.00 Source: Beck Demirguc-Kunt and Levine (2000). Note: Values are 1995-99 averages. 119. Ifthe move towards fiscal discipline and a low-inflation environment can be sustained, the banking system's role will take on much greater importance in Turkey. In the transition, there are also important risks. Banks will have to start behaving as commercial banks, expanding their credit portfolio and avoiding the potential risks that accompany such credit booms. At present, most o f banks' interest income comes from securities. To offset the income 65 losses due to lower inflation and the move away from government securities, banks would have to improve their profitability. This requires the remaining regulatory and taxation burdens on the banking sector to be reduced. The latter i s particularly important since such burdenscan become more significant as inflationary income disappears (Eroglu, 2001). Liquidity and reserve requirements, and transaction taxes-on foreign exchange, stock market, and other individual transactions-need to be decreased. Some o f these taxes also lead the banks to book their assets and liabilities abroad. Foreign banks are in a better position to do so and thereby gain an unfair advantage. Some o f the larger foreign banks only keep approximately one fifth o f their Turkish assets and liabilities in Turkey. There has been limited progress in this area. The Central Bank i s now paying interest on reserve requirements and there are plans to phase out some o f the costly transaction taxes and replace them with other revenuemeasures. Another area that would facilitate better evaluation o f credit risk i s the establishment o f a credit bureau. Currently, the Central Bank collects information on only loan defaults which impedescredit analysis. 120. One o f the important factors that distorts the incentive structure o f the Turkish banking system is deposit insurance. Empirical evidence from the experience o f a large number o f countries around the world shows that poorly designed deposit insurance systems can do more harm than good by undermining market discipline and causing moral hazard, particularly in countries with weak supervision and regulation and undercapitalized banks (Demirguc-Kunt and Kane, 2002). InTurkey, deposit insurance scheme was established in 1983 and its coverage was made unlimited after the 1994 crisis. This allowed Turkish banks to increasingly engage in riskier operations-keeping large open foreign exchange positions for example-without much concern since their creditors and depositors were covered by the guarantee. This blanket coverage was extendedto cover all creditors after the recent crisis. 121. Inthe long term, it is not feasible to retain a blanket guarantee since it would make the job o f official regulators very difficult by underminingprivate market discipline. Indeed, there were plans to reform the deposit insurance scheme after the bank recapitalizationwas completed, but given the economic uncertainties these decisions have been delayed. One of the important decisions that will have to be made i s the new coverage limit. Currently, there are plans to reduce coverage to 50 billion TL, which corresponds to the EU scheme. However, this coverage i s over ten times the Turkish GNP per capita and i s therefore too high. International evidence shows that best designed schemes have coverage levels o f at most 2 times GNP per capita- like most European countries-which would be sufficient to protect the deposits o f the majority o f the depositors. Higher coverage levels would reduce market discipline, increase moral hazard and make banking crises more likely". Currently, Turkey has a risk-based premium structure although the differentiation betweenpremiums for high/low risk banks i s not significant. It may be a good idea to retain the risked-based premium structure inthe new limited deposit insurance scheme to replace the blanket guarantee. 122. In summary, the long-run institutional framework is key to make the banking reforms sustainable. Political commitment to the reform process i s particularly important in continuing the reform efforts. Important steps have been taken to improve the incentive structure of the Turkish banking system since the crisis. Strengthening prudential regulation and supervision, efforts to recapitalize the private banking sector, the progress made so far in restructuring the loDemirguc-KuntandHuizinga, 2001 andDemirguc-KuntandDetragiache,2002 66 state banks, efforts to resolve the NPL problem and to ease the restructuring o f the corporate sector are very important steps. However, in the long term, a healthy and effective banking system requires a stable macroeconomic environment, reduced public sector borrowing needs, lower state ownership, a lower regulatory and taxation burden, and a better designed deposit insurance scheme that does not undermine private incentives to monitor the banks. As the reform process moves forward it will be important to keep inmindthat public supervision tends to be much more effective when it i s free from political intervention and reinforced by private discipline, and partial bank recapitalizations without paying adequate attention to factors that undermine the effectiveness o f regulatory/supervisory environments and impediments to bank profitability are doomed to be repeated. 67 CHAPTER CORPORATE 4: SECTOR 1 INTRODUCTION 123. The cumulative effects of the 2001 crisis were felt throughout the economy-starting in the financial sector and causing increasing distress inthe real sector.' Net profitability of listed companies-on the decline between 1997 and 2000-turned into major losses in the first half of 2001 inevery manufacturing sector except cement and glass. Losses totaled US$l billion during the first semester of 2001,down from profits ofUS$1.6 billion inthe same periodin2000. From the first quarter of 1999 through the first quarter of 2001, the share of listed companies at medium risk rose from 63 to 69 percent-while those at high riskjumped from 2 to 11 percent. Together, these companies accounted for 58 percent of listed companies' sales, 56 percent of employment, and 89 percent of bank liabilities. The corporate sector rebounded strongly in 2002, largely due to export-led growth and a modest recovery in domestic demand. Net profit in the corporate sector for the first ninemonths of 2002 reachedUS$2.3 billion, slightly higher than earnings for the entire year 2000. The main factor was a decline in financial charges, including foreign exchangelosses. 124. To assist inrestoring growth, Turkey is inthe process of implementing a comprehensive corporate restructuring program known as the Istanbul Approach (IA). Once firms enter the IA, there i s a formal review process by a Creditors Committee. Once a workout i s agreed to, the creditors and the debtor (the company) sign an agreement stipulating the terms and conditions of the workout, reporting requirementsof the debtor and events which would trigger a review of the agreement. The creditors also sign an inter-creditor agreement specific to the specific workout case. The IA does not guarantee the survival of a company in distress. Regulatory authorities do not intervene and, because of its voluntary nature, the IA is only effective if it i s mutually supported by the banking and corporate communities. Progress to date in implementing the IA has beenreasonable. As of end-June 2003,299 firms, 208 large firms in32 groups and 91 small firms had entered the IA with US$4.8 billion in loans being restructured. These firms have approximately 46,335 employees, sales of about US$2.9 billion and exports of close to US$0.8 billion, and over US$ 7.5 billion in assets. Nevertheless, a number of issues have emerged which if left unresolved will impede progress. One issue is the lack of an adequate bankruptcy framework. Other issues involve technical concerns including the handling of so called category 2 loans in the balance sheets of banks and the participationof state banks. The Advisory Committee consisting of Treasury, the Central Bank, BRSA, the banks, TOBB and TUSIAD may be helpfulinpushingthe process forward. 125, Several medium-term reforms should complement the short-term corporate resolution process. These include measures to attract foreign direct investment and accelerate the privatization program. A careful review of the performance o f state enterprises i s required to assess their need for workouts and more fundamental restructuring. In addition, there is a need for accounting reform and improvedcorporate disclosure, as well as better corporate governance including strengthenedindependencefor boards of directors. 1This review ofthe impacto fthe crisis is basedupona recentWorld Bank Study, "Turkey: CorporateSector Impact Assessment Report", March2003. 68 2 ANALYSIS CORPORATE DISTRESS OF 126. The 2001 crisis hurt all sectors-industry, construction, trade, and services. Industrial production fell 6.9 percent inthe first half of 2001 and 9.9 percent for the year. Plant capacity use in the private sector fell to 62 percent in mid year, before recovery to 68 percent in the fourth quarter o f 2001. Lower domestic 1 Figure4.1: ExportOrientation 4G% I demand was the most important factor undermining capacity utilization. Increased 30% exports in a number o f sectors-such as 20?? automobiles and textiles-helped mitigate 1G?? corporate losses but even these sectors suffered significant losses in 2001 (Figure 4.1). The downturn forced most firms to reduce their workforces throughout 2001. Efforts have also been made to control wage costs, with shorter -+-Large +Medium +Small work weeks, strict control on overtime, Source: CSIA Report 2003, World Bank accelerated vacation and maintenance schedules, and temporary furloughs. Although Turkish firms have beenagile inresponding to the crisis, worker distress has been considerable. Over the 2000-02, period, the unemployment rate increased from 6.6 percent to 10.6 percent. 127. Data for non-financial companies (industry and services) listed on the Istanbul Figure 4.2: Financial Expense Burden Stock Exchange (1SE)-a sample 1 40% 1 representative o f Turkey's mid-cap and large 30% companies-show worrisome trends. The 20% series o f crises over the past decade (1994, 10% 1999, November 2000 and February 2001) have lowered company profits and increased financial fragility. N e t profitability declined over the 1997-2000 period, then turned into +Large -Medium -+-Small major losses in the first half o f 2001 in every manufacturing sector except cement and glass. Source: CSIA Report 2003, World Bank Losses totaled US$1 billion at the end of second quarter o f 2001, down from profits o f Figure 4.3: Profitability Size US$1.6 billion in the same period in 2000. 10% I - 1 Interest rate coverage on an earnings before interest, taxes, depreciation and amortization (EBITDA) basis dropped below one beginning -1 inthe first quarter o f 2001 and stayed that way until the fourth quarter. By contrast, the debt -20% 3 coverage ratio was consistently above one -30% ' 1 through 1999-2000 for most listed companies. In addition, financial expenses rose sharply as 1+Large -Medium +Small 1 a percentage o f revenue in the first quarter o f Source: CSIA Report 2003, World Bank 2001, hitting 27 percent (Figure 4.2). -For 2001 as a whole, sales o f listed companies were down 69 13 percent, earnings (EBITDA) were down 11percent, andnet earnings were down a staggering 80 percent. 128. The 2001 crisis put a larger share of listed companies at medium or highrisk o f default. Between the first quarters of -1999 and 2001 the share o f listed companies at Figure 4.4: Distressed ComDanies. 1998-2001 Q1 SE listed Companies) medium risk rose from 63 to 69 percent- while those at high risk jumped from 2 to 11 percent. Together these companies Sustainable accounted for 58 percent o f listed Operationally companies' sales, 56 percent o f Distressed employment, and 89 percent o f bank ~9 Financially Distressed liabilities (Figures 4.3 and 4.4). Distressed companies have limited opportunities for a 0 Technically insolent turnaround in the short term. Companies have cut production costs through import 1998 1999 2000 2001Q1 working capital, buta general lengthening of receivables and inventories Table 4.1: Net Profitability by Sector, 2001-2002 demonstrates the difficulty o f doing so ina (Millions of U S Dollars) systemic crisis. Companies have also cut 2001l12 2002112 Consumer Durables -35.9 168.9 investments, with private investments Electronics-Communication -530.7 139.6 down nearly a third in 2001. A few Textile -209.77 22.45 companies and banks have made rights Transportation 48.6 100.3 offerings to increase equity capital, but the ISE Total -1175.44 933 I depressed stock market discourages equity Source: TSKB Research, Weekly Bulletin, April 21-25, financing. 2003 129. A major turnaround inprofitability occurred inthe consumer durables sector in2002, in particular the white goods sector, and that has continued in the first quarter o f 2003. Figure 4.5: Financial Expenses (Total USDm) Other sectors with significant turn around Total Industry I in profitability in 2002 were electronics Textile and communications, food processing and Chemicals textiles (Table 4.1). The primary factor for the turnaround in profitability was a Food reduction o f financial charges. Financial Electronics-Communication expenses include both interest payments ConsumerDurables and foreign exchange losses under the Turkishaccountingprinciples (Figure 4.5). 0 2000 4000 6ooo 8000 loo00 Source:TSKB 70 3 SHORT-TERM RESOLUTION STRATEGIES 130. Implementing a comprehensive resolution strategy has been a continuing priority in dealing with the aftermath o f the 2001 crisis and its effects on Turkish companies. Lessons from the recent crises inEast Asia and Mexico indicate that the short-term resolution strategy should include: (i)segmenting problem companies by size so that companies o f different sizes and needs can be treated appropriately; (ii) implementinga voluntary workout program; (iii) adopting a new bankruptcy framework, including the option for pre-packaged bankruptcies that would recognize the workouts in the courts; (iv) easing tax, legal, and regulatory impedimentsto crisis resolution; and (v) development o f an asset management company or companies. Each o f the major crisis countries in recent years-e.g., Mexico, Thailand, Korea, Malaysia and Indonesia- has implemented a voluntary workout program that was important in restoring the creditworthiness o f the real sector. Getting the incentives right has proven difficult in other crisis countries including East Asia and Mexico2. Companies that have defaulted on loans could face seizure o f their collateral if they abstain from the workout process. However, Turkey's weak bankruptcy framework may well serve as an obstacle to the workout program as debtors seek shelter from their creditor through the bankruptcy process. A resolution program i s also vital to restore the soundness o f viable banks and may include an asset management agency or company to remove the most distressed assets or special classes o f assets such as real estate, cars and other saleable collateral, from the books o f these banks. Segmenting ProblemCompanies 131. Experience in other crisis countries suggest that thousands o f small and medium-size enterprises, hundreds o f mid-cap enterprises, and a few large groups may all be introuble at the same time. Countries have dealt with these issues in various ways, but all have eventually recognizedthe needto segment problem companies. The types o f segmentation discussed below are at the heart o f the resolution strategy that has been developed and i s known as the Istanbul Approach. 132. Companies can be broadly segmented into three groups. Small and medium-size enterprises-the first and most numerous casualties o f such crises, with little power to effect outcomes-are usually too numerous to include in a corporate workout program. Solutions in support o f small and medium-size enterprises needto be systemic, such as extendedrollovers o f credits including re-capitalization o f past-due interest and extended grace periods for principal repayments. Mechanisms to improve liquidity in small and medium-size enterprises can help keep them afloat. The financial costs o f stabilizing small and medium-size enterprises are small relative to the social and political costs o f allowing thousands o f them to sink, and should not take much time or effort to apply. At the other end o f the spectrum are the large corporate groups that dominate the economy, with complex structures that may include a mix o f industrial and financial companies. To the extent that these groups exhibit distress, they need to be monitored carefully and addressed on a case-by-case basis so that their problems do not provoke a secondary crisis and spill over into the banking sector, causing banks and non-bank financial institutions to collapse. The Daewoo collapse in Korea i s perhaps the most prominent recent Kawai, LibermanandMako (2000), "Financial StabilizationandInitialRestructuringofEast Asian Corporations: Approaches,Results,and Lessons". 71 example o f such a case. Finally, in the middle are mid-cap corporations. Depending on the economy, these firms have US$25-100 million in annual sales, normally require bank financing for working capital and investment, and may be the most adversely affected if a crisis causes a contraction in the stock o f credit, as has happened in Turkey (and happened in East Asia and Mexico). Mid-cap companies-usually found using several banks and other financial intermediaries such as leasing companies, as well as supplier credits-normally form the core o f a voluntary workout program. With adequate due diligence and a well structured workouts, many mid-sized companies can be restored to creditworthiness. The IstanbulApproach 133. The Turkish Bankers Association, working with Government and industry representatives, has taken the lead in developing a voluntary, non-judicial workout program based on the London approach. The program, informally known as the Istanbul Approach (IA), has requiredthe strong backing o fthe Treasury and the Ministryo f Finance because o f the policy and regulatory implications and the tax incentives normally required to make such a program operational. The IA i s supported by amendmentsmade by the law (No. 4743) dated January 31, 2002 to the articles o f Banking Law (No. 4743) and by a circular o f BRSA in April 2002. In addition, the BRSA has established clear provisioning rules in support o f workouts and in coordinating workouts with intervenedbanks. At the heart o f the IA i s a Framework Agreement (inter-creditor agreement) signed in June 2002 by 34 commercial banks and non-bank financial intermediaries. Intervened and state banks have also signed the agreement. Finally, the agreement has beenapprovedby the BRSA. 134. The IA incorporates key aspects derived from international experience with voluntary workouts. First, creditors must signal that they are willing to pursue a non-judicial resolution o f a company's financial difficulties rather than resort to a formal process o f seizing collateral or an insolvency procedure such as liquidation. Second, the out-of-court process under the IA i s case- by-case, Le., each company entering the process i s addressed by a separate Creditors' Committee and that committee i s directed by a lead bank holding the largest share o f the credits or i s appointed by the other creditors holding 75 percent of the credits by value. While the inter- creditor agreement has been signed by banks and non-bank financial institutions, other creditors such as trade creditors may join the workout process through the Creditors Committee. Third, the IA is time bound. That is, the Creditors Committee has a maximumof 90 days plus three 30 day extension periods (180 days) to reach a resolution and agree to a workout with the debtor. Fourth, the IA incorporates a due diligence process. The creditors commission an independent due diligence review of each distressed company's long-term viability, drawing on comprehensive information made available by and shared between all the likely parties to any workout. 135. A Coordination Secretariat, under the direction o f TSKB, is responsible for oversight of the process and to ensure that its logistics and timing o f the work (Box 4.1A). A lead bank is chosen to head the Creditors' Committee formed for each case. Ifthere i s no resolution, i.e., less than 55 percent o f creditors by value agree to a proposed workout plan, then the case i s dropped from consideration. A three-person Arbitration Committee, appointed by the board of the TurkishBankers Association, makes decisions on workouts ifthey pass the 55 percent threshold, but do not reach approval of 75 percent of creditors. The arbitration panel can also decide on 72 any individual matter which the Creditors Committee bringsto its attention. For the large groups or conglomerates, a group owned bank cannot take the lead for a group owned subsidiary, nor can the group bank vote in the creditors committee. A group bank i s also not able to provide newfinancing to a group company. Box 4.1A: InstitutionalStructure for Istanbul Approach I TSKB Arbitration Coordinationnntermediation Panel I f f f Committee Committee Committee 136. Because o f its voluntary nature, the IA i s only effective to the extent it i s mutually supported by the banking and corporate communities. Regulatory authorities do not intervene and the IA does not guarantee the survival o f a company in distress. During the period o f the review and negotiation the creditors must agree to support a standstill. That is, each of the creditors agrees to maintain its credit facilities and not to move against its collateral to the disadvantage o f the other creditors. This maintains the confidence o f the workers, suppliers and the customers by allowing the company to continue operating. At the end of the due diligence period, the creditors have to form an opinion with respect to the company's viability and make a decision whether or not to go forward with the workout. Until the workout i s agreed upon, creditors preserve their legal rights to pursue the debtor through the courts inthe event a workout agreement i s not reached. 137. Once a workout i s agreed to, the creditors and the debtor sign an agreement stipulating the terms and conditions o f the workout, reporting requirements o fthe debtor and events which would trigger a review o f the agreement. The creditors also sign an inter-creditor agreement specific to the specific workout case. The IA allows companies to supplement their borrowing incase of a liquidity shortfall or pressingmaintenance needs. Newmoney can be provided on a pro rata basis by all existing lenders, by specific lenders with priority arrangements, or by the release o f asset disposal proceeds subject to priority considerations. Banks can also benefit from the tax incentives that are a part ofthe IstanbulApproach. Banks are eligible for these incentives under the program for a period o f three years with the option for extendingthe program. Other principles underlyingthis critical period o f financial support include recognizing the seniority o f existing claims and sharing losses on an equal basis betweencreditors in the same category. If creditors agree that a company i s viable over the long term, they can consider a formal rescheduling-such as an interest holiday, re-capitalization o f interest in arrears, extension o f 73 loan maturities, lending o f new money, or conversion o f debt to equity. These longer-term financial changes need to be conditional on the implementation o f an agreed business plan that may well require restructuring o f the company over the term of the workout. Restructuringmay include management changes, injections of fresh equity, sales o f assets or divisions, mergers or even company takeovers by new owners. Box 4.1B diagrams the IA process. Box 4.1B: Istanbul Approach Process 138. Eligibility and sequencing are key issues for the IA. Eligible firms are primarily large and mid-cap manufacturing firms. To qualify as a large firm, companies must meet at least two of the following criteria as o f end-2001: having two or more banks as creditors, debt o f more than US$lO million, more than 100 permanent employees, annual exports o f US$15 million, total annual turnover o f at least TL 25 trillion, and assets on an audited balance sheet o f at least TL 15 trillion. There i s also a need to address sequencing-that is, the criteria used to decide which firms enter the process initially. Although small and medium-size enterprises are covered by a modified form of the process handled primarily by the lead bank, the number of firms that can be restructured inthe first phase may needto be limitedto preventthe process from bogging down. In Thailand, the workout process bogged down with thousands of SMEs involved in restructuring. Thus a more systemic approach needs to be adopted for small firms3. Excluded from the process are marketing and trading firms (that is, non-producing firms) and firms with more than 50 percent state ownership. ProgressinImplementingthe IstanbulApproach 139. As of end of June 2003,299 firms (208 large firms in32 groups and 91 small firms) were participating inthe IA with some US$ 4.8 billion in loans being restructured. These firms have approximately 46 thousand employments, sales o f US$2.9 billion and exports o f about US$ 761 million, and about US$ 7.5 billion in assets. Many o f the cases have completed the 90 days and are on to the first extension. Experience in most crisis countries i s that a systemic workout process may well take several years to complete. Moreover, it i s difficult to get the workout just Erzanand Filiztekin(2002), a backgroundpaper to this Report, analyzesgrowth, employmentandproductivityin Turkishmanufacturingindustry with respectto industryand size, and scrutinizesthe vulnerabilityof firms with respectto the economic environment. 74 right the first time, given the unstable economic conditions that normally exist when the workouts are concluded. Therefore, the process has to allow at a minimumfor modifications of some workouts a couple of years into the process once economic conditions have stabilized. Although the IA was formally introduced inJune 2002, implementation was realistically delayed until the effort to flush out NPLs and re-capitalize the banks was completed as outlined in chapter 3. At that point, newly capitalized banks had the proper incentive to have loans re- classified and to collect on provisioned loans. Following the recapitalization, improvements in loan classification and collection on provisioned loan are directly reflected in increased profits line and capital o f the banks. 140. There are several impediments to the IA that prevent it from becoming a more robust restructuring process. The first, and most important, problem i s with category two loans4. The banks are unwilling to bring companies with potentially viable, category two loans that need restructuring into the IA. Banks are also reluctant to provision for these loans, but recognize that they may well need restructuring and/or rescheduling and that the borrowing companies may require new working capital. There are three issues related to category two loans: (i) i s unclear it under the legislation governing the IA whether these loans would automatically require provisioning if they are restructured under the IA; (ii)a number of category two loans are denominated in foreign exchange and with 12-18 months maturity which may have already become ineligible for the IA; (iii) the IA requires a 90 percent vote o f creditors to approve workouts o f performing loans-including category two loans-which i s too restrictive. 141. The second problem involves differences between state and SDIF banks and commercial banks. Commercial banks claim that 3-4 large commercial banks are bearing most o f the burden o f the IA restructurings. That is, state and SDIF banks cannot participate in debt/equity swaps, cannot accept real estate collateral, and are reluctant to extendnew financing. This i s a material issue as the state and SDIF banks are major creditors in the workouts already underway. State banks claim they are ready to adhere to the IA rules as long as there is independent due diligence o f IA cases by TSKB or external investment banks/accountants. Also, to provide new money, the state banks argue that they must be given first priority in accord with IA rules and this new money needs to be backed by new or shared collateral. The third problem i s that banks have focused their efforts on rescheduling and notfundamental restructuring of problem companies. Most o f the companies that have entered the IA process have large non-performing loans. A number of these companies hadproblems before the recent crisis. The banks are opting for long- term rescheduling (6-14 years) to lower the debt amortization burdens and to increase free cash flow in these companies. A number of the IA companies are being restructured through debtlequity swaps, debvasset swaps and even change o f management, but more pressure for real restructuring i s needed. Most o fthe workouts will also require some fresh cash, but management o f the banks are reluctant to increase their exposure to these companies. This was the experience initially inKorea where the banks focused on rescheduling. Inthe end, 50 percent o f the Korean companies went through a second and third round of restructuring with a number o f companies needingto be liquidated. This i s likely to happen inTurkey as well. The banks will take larger write downs ifthey do not restructure the workout cases under the IA. Also, the corporate sector and the economy will be adversely affected. ~ ~~ 4 Loansthat needto be extendedbut for which delays in repayment, ifany, are less than 90 days and due to acceptablereasons-these loando not needprovisioningas NPLs inthe balancesheet 75 142. Looking ahead, experience from East Asia points to some other issues that are likely to come up over the course o f implementation o f the Istanbul Approach. One will be the disposition o f assets as loans are restructured. The SDIF i s operating a public asset management company, but public agencies have often found it hard to operate such companies because o f the commercial nature o f the work. Companies that go through restructuring may well have a need for funds to manage assets that are difficult to dispose o f during a crisis. Distressed companies and banks may also seek to divest land and other fixed assets from their portfolios. The lack o f an adequate bankruptcy law limits the ability o f the banks to press for restructuring. That is, they lack the "stick" in the system to require corporate restructuring. In countries that lacked an adequate bankruptcy system4.g. ,Mexico, Thailand and Indonesia-it proved very difficult to follow through on voluntary workouts. In Korea, on the other hand, the bankruptcy system complemented the voluntary workout program. Some 13 large chaebols went into bankruptcy duringthe crisis and this served as a signal to other firms that they neededto restructure5. 143. The voluntary nature o f the Istanbul approach i s critical to its success and it is important that Government not interfere in individual transactions. Nevertheless, there are several critical steps that Government can take to facilitate the process. Key recommendations are summarized inBox 4.2. Recommendation Timing of expectedimpact Short term Medium term Policiesto Support Corporate Restructuring Implementa voluntary workout programbasedonthe Londonapproach X Commercializeasset resolutionmechanisms - Sell distressedassetsthroughan asset resolution agency- X - preferablyone not createdbythe Government Sell landandrealestate through realestate investmenttrusts (REITs) Strengthenthe bankruptcy systemand allow for pre-packaged bankruptciesto recognizeworkouts throughthe courts X X Easetax, legal, and regulatory impedimentsto restructuring. X X Asset Resolution 144. Corporate workouts will leave banks with many assets that they are poorly equipped to manage and that might be difficult to dispose o f during the crisis. Moreover, these assets will not earn interest and will impair banks' reported financial performance. For example, a debt overhang will inevitably require debt-equity conversions or the swap o f straight debt at highreal interest rates for longer-term convertible debentures at lower interest rates. InKorea the interest rate on convertible debentures introduced in workouts was only 1 percent. In addition, many companies going through a workout will require additional long-term debt or equity injections to operate inthe future. 145. Virtually all crisis countries have established asset management agencies or companies. However, most countries have found it difficult to operate publicly created agencies due to the commercial nature o f asset disposal. Also, the lack o f a secondary market for these assets in most countries significantly slows the process o f asset disposal. As a result, the value of such 'Woo-Cumings (2001), "Miracle as Prologue:The State andthe Reformof the Corporate Sector inKorea". 76 assets quickly plummets toward zero. Turkey infact has already had a public asset management agency. Through the intervention o f some twenty commercial banks, the SDIF i s operating a public asset management agency. A BRSA circular issued in October 2002 established an enabling environment for the creation o f private asset management companies. Diverseforms o f asset management companies can be created by private entrepreneurs or major financial institutions. An alternative approach to the SDIF problem o f asset disposal, i s to consider profit participation deals. In East Asia and Eastern Europe, large financial groups have entered the market to purchase distressed assets-for example, in Korea they did so through a profit participation arrangement with the Korea Asset Management Company. International debt and equity funds could manage these assets for banks and potentially bring in institutional capital to provide some new money. Korea, for example, set up four funds with the participation o f the Korean Development Bank, commercial banks, and internationally recognized fund management groups to helpmid-cap companies seeking investors. The International Finance Corporation also made several direct investmentsin mid-cap companies in Korea. Such a role might be ideal for Turkey's private development bank, supported by commercial banks. A major impediment to the success o f a commercially viable asset management company in Turkey will be the willingness o f banks to dispose o f distressedassets at a fraction o ftheir value. Ifthe banks need liquidity following the provisioning of NPLs, then they may be disposed to sell these loans. However, all indications are that the Turkish banks prefer to wait until the economy improves so that they can realize a better return on these impaired assets. 146. Land and real estate sales. Distressed companies and banks may seek to divest excess fixed assets or sell and lease back primary real estate. Banks will also need to divest land and buildings acquired as collateral during the crisis. Duringits crisis Korea established a land bank to allow companies to divest excess property, with an option to reacquire it in the future. It did so to prevent a precipitous fall in values as occurred inJapan. Although Korea's land bank was owned by the state, there i s no reason Turkey could not attract institutional investors to establish such real estate investmenttrusts (REITs). The BankruptcyFramework 147. An important step i s the establishment o f adequate bankruptcy procedures. As noted above, it will be difficult to get firms to go through the needed restructuring (as opposed to a simple reschedulingo f loans) without a functioning bankruptcy framework as a fall-back option. Corporate workouts and restructurings generally take place inthe shadow o f the law. Voluntary resolution requires reliable fallback options through established legal mechanisms for individual enforcement and debt collection or through collective insolvency procedures. As the Mexican and East Asian experiences show, voluntary workout programs are often constrained by inadequate insolvency systems and by legal environments that discourage restructuring. Inmany crisis countries, weak bankruptcy systems have proven to be an impediment to the voluntary workout process. Turkey also has a weak bankruptcy systemthat can leave companies protected by the courts for years (Box 4.3). Currently, Turkishcourts can take up to two years to resolve enforcement proceedings. During a crisis, the bankruptcy system is not by itself a viable alternative for restructuring problem companies. The number o f distressed companies inTurkey would quickly overwhelm the court system were it to be the only alternative. However, the bankruptcy framework must be an effective alternative to voluntary workouts in order to provide 77 an important incentive (Le., "stick") for companies to participate in the workout process (the "carrot"). Box 4.3: Legal Framework for Creditor Rights and Enforcement in Turkey Turkey's legal framework for creditor rights and enforcement deals with procedures for enforcing the commercial rights of unsecured and secured creditors. For secured creditors it also covers laws and procedures for creating, registering, and enforcing collateral rights such as mortgages and pledges. Up to 90 percent o f bank lending in Turkey is done on a secured basis, and most is guaranteed by a major shareholder. Types of collateral include real estate and movable assets; an unconditional mortgage on real estate is the most reliable form o f security because it allows the creditor to bypass the courts and move directly to execution or foreclosure procedures. Inthe event of a default, creditors can pursue execution or apply for an insolvency proceeding. Nearly all cases- almost 500,000 a year-are filed under the execution procedure. Procedures for debt recovery and foreclosure are generally effective, lasting an average o f four to eight months over the past decade. Shorter proceedings tend to be uncontested, meaning they go directly to an executioner, which sends a collection order to the debtor. If creditors choose to go to court or a collection order is contested, it generally takes two to three years to reach a final judgment. In most cases, however, either the debtor does not contest the proceeding or a default judgment can be obtained. Courts are stretched to capacity and will likely become even more overburdened by efforts to collect on the non-Derforming loans o f failed banks. 148. Enforcement o f secured loans. Turkish banks typically secure their loans with real property worth far more than the loans. Ifa lender i s unable to resolve a dispute with a borrower when a loan is not being properly serviced, the bank may initiate an enforcement action in the courts. While initiation o f such lawsuits has traditionally been the trigger for banks to classify such loans as non-performing, new loan loss provision rules require classification based on performance and creditworthiness. The recent effort to pushbanks to flush out their NPLs could lead to a dramatic increase in the volume o f enforcement actions. Given that the courts are already overloaded, the increased caseload will likely slow the completion o f enforcement cases as well as all others. Because a contested enforcement proceeding can take more than two years to resolve, the prospect o f further delays i s worrisome. The banks are confident that court enforcement procedures are effective, if slow. The question i s whether the system will work as well when caseloads increase further. Just as important is whether an enforcement system that works for lenders-because they typically hold considerable excess security-is effective in protecting borrowers. Except where lenders work together on an informal or ad hoc basis to work out a restructuring plan, the answer may well be negative. 149. The Savings Deposit Insurance Fund (SDIF) has intervened in some twenty private banks. To expedite recovery, the loan portfolios o f these banks have been redefined as state receivables subject to collection under the Public Receivables Act (No. 6183). Because court judgments are not necessary for state receivables, this change could shorten average recovery periods to a little as six months. However, giving the loans o f SDIF banks the status o f state receivables raises several issues. First, accelerated collections enable the SDIF to recover inless time, thereby potentially giving it access to debtor assets more quickly than other creditors. Second, the Public Receivables Act applies to all assets o f a debtor-not just assets that have beenpledged-which may give the SDIF an effective priority and greater likelihood that it will be able to recover amounts owed to it relative to other creditors. Because the Act provides limited scope for rescheduling and almost no capacity to write off debt, it can be difficult to 78 develop a workout plan for a troubled debtor. Just as important, enforcement proceedings under the Public Receivables Act take place outside the commercial court system while other creditors must use the courts-making restructuring more cumbersome, The question of whether the SDIF acquires greater rights than the original holder o f the debt simply by virtue of subrogation or assignment i s certain to be challenged in the courts, resulting in delays. Indeed, this issue could have a widespread impact because other courts will likely take a wait-and-see approach. Depending on the final ruling, the result could potentially undermine the voluntary resolution process. 150. Legal framework for insolvency. Turkishinsolvency procedures are usually not relied on as a means o f recovering debt even incases where borrowers are clearly insolvent. The average number of cases, about 1,500 a year-less than 0.5 percent o f execution proceedings-reveals that insolvency has been marginalized. Unsecured creditors see the process as ineffective, fraught with extremely long delays, and yielding little or no benefit. Distributionsto unsecured creditors reportedly do not exceed 3-5 percent, and no consideration i s givento the time value o f payments. As a general rule, secured creditors enjoy a privileged status under the Execution and Bankruptcy Act. The declaration o f bankruptcy does not result in a stay on actions by secured creditors to foreclose on their collateral. Where foreclosure proceedings have not commenced, pledgedassets can be entered as part o f the bankruptcy proceeding, and the secured creditor has the option o f disposing o f the collateral on its own or allowing the administrator to do so. As a general rule it i s difficult to salvage viable enterprises when secured creditors are exempt from the process and can collect on collateralthat may be essential for the business. Weak insolvency procedures seriously impede the development o f viable workout plans for illiquid debtors. The fact that secured creditors can execute against a debtor's core assets limits the scope for developing a plan to rehabilitate the debtor. However, given the inordinate delays and marginal use o f the bankruptcy process, the current treatment o f secured creditors offers a more reliable framework for risk assessment and management by lenders. 151. Pre-packaged bankruptcies. Experience in East Asia and Mexico highlights the impediments to corporate restructuring that can arise from valid legal checks and balances. These legal concerns-such as protections for public shareholders and controls on tax deductions for credit write-offs-often serve broader public policy interests and cannot be simply eliminated. Public shareholder rights to oppose diluting equity restructuring can be particularly difficult or impossible to override in an out-of-court workout, as seen in recent cases involving Daewoo affiliates inKorea. To deal with valid legal concerns, it may be necessary to rely more on court-supervised processes (such as "prepackaged" reorganizations) to effect equity restructurings o f companies. Inaddition, the insolvency system can provide "front end" support to encourage debtors to cooperate with out-of-court workout efforts-as well as important "back end" support that makes it possible to override dissident creditors or shareholders in a corporate workout. 152. Turkey's insolvency law should be amended to restrain secured creditors from executing against their collateral as long as their property interests have not beenaffected. The insolvency law should provide a mechanism that preserves the value o f the security for secured creditors, but that allows the excess security value to be used to help restructure the debtor. The amended law should also include provisions for pre-packaged bankruptcies. There i s little point to these changes, however, unless it i s accompanied by wholesale changes that allow the procedure to be 79 expedited. Time destroys the chances of successful restructuring-a process that takes years, insteado f months, has little to offer. Expeditingmatters requires changes to judicial procedures, and ideally to the structure o f the courts. In light o f the comprehensive legal and institutional reforms needed, the Government undertook an overall assessment o f its legal and institutional framework for creditor rights and insolvency, usingthe World Bank's Principles and Guidelines for Effective Insolvency and Creditor Rights Systems (April 2001). This assessment has identified specific reforms required in Turkey's legal and institutional framework in core areas related to the enforcement o f creditor rights (including creation, registration and enforcement o f security), and formal and informal corporate insolvency. The Parliament passed a set of comprehensive amendments to the EBA on July 17, 2003 which address enforcement o f creditor rights, formalhnformal corporate insolvency and restructuring systems, credit risk management practices, and institutional and regulatory frameworks. In addition, The Government plans to enact a new section o f EBA on pre-packaged bankruptcy by end-2003. 153. Easing tax, legal, and regulatory impediments to restructuring:. The East Asian crisis highlightednumerousimpedimentsto corporate restructuring, mergers and acquisitions, and the entry o f foreign direct investment. Turkish firms face many o f the same obstacles. Significant impediments encountered in one or more East Asian countries included: (i) treatment of debtor gains from debt restructuring as taxable gains; (ii) inability of corporate creditors to deduct an losses from corporate debt restructuring and voluntary workouts to reduce taxes; (iii) immediate taxation o f non-cash corporate reorganizations such as mergers or spinoffs; (iv) little or no opportunity to transfer net operating tax loss carry-forwards to a corporate acquirer or new merged entity; (v) excessive fees and taxes on property transfers; (vi) constraints-such as the personal liability o f staff-on the ability of state-owned financial institutions to incur losses through corporate debt restructuring; and (viii) constraints on the ability o f banks to hold equity in debt-equity swaps or similar instruments such as convertible debentures. The Turkish Government has amended corporate tax laws and regulations to encourage mergers, acquisitions, asset spin-offs, and the sale o f real estate and participation shares by banks augmenting their capital. It has also eliminatedstamp, value added, and other duties on such transactions. These amendments address several o f the obstacles to restructuring and workouts encountered in East Asia. However, more impediments may surface that could requireGovernment action. 4 MEDIUM TERMSTRUCTURAL REFORMS 154. In addition to the corporate resolution process, other measures are neededto foster the environment for a more dynamic corporate sector. Some o f the most important o f these measures are discussed in other chapters of this report. Macroeconomic stability including a stable and competitive exchange rate and lower real interest rates are very important for corporate sector recovery. These are discussed inchapter 1. Similarly, a healthy financial sector i s critical to growth o f the real sector and the relevant financial sector policies and reforms are discussed inchapter 2. Here, three other medium-term structural reforms are briefly considered: privatization, improved financial reporting andaccounting standards, and corporate governance. 80 Accelerating Privatization. 155. Accelerating privatization i s a key supply-side element in the effort Figure 4.6: firkey: Privatization Under PA to sustain growth. While Turkey Portfolio and GSMLicenses Sales ($ million) began its privatization program in 5,500 5,000 1985-making it the first developing 4,500 country to do so-it i s now lagging far 4,000 3,500 behindcomparable economies inLatin 3,000 America (Argentina, Chile, Mexico) 2,500 2,000 and the main economies in Central and 1,500 Eastern Europe (Czech Republic, 1,000 500 Hungary, Poland). Privatization sales have consistently fallen short o f expectations, regularly falling short o f US$1 billion per year with the Source: PrivatizationAdministration and World Bank exception o f 2000 when sale o f a majority stake in POAS (petroleum distribution) and a third GSM license pushedsales above US$5 billion (Figure 4.6). The public sector still accounts for 15 percent o f non-agricultural production. In most emerging market economies and many OECD countries, privatization has been major driver o f foreign direct investment. Disappointing privatization performance in Turkey may have been a factor in its failure to attract significant levels o f FDI. 156. Accelerated privatization could generate significant external resources, improve the efficiency o f strategic sectors-such as telecommunications and energy-and send a strong signal to foreign investors and financial markets about Turkey's commitment to reform. But increased efficiency will not happen automatically: it will require implementing appropriate regulations, creating independent regulators, and allowing competition in these sectors. To achieve these tasks, the Government must make a long-term commitment to structural reform. In moving forward with privatization, the Government faces at least two important trade-offs. First, inthe energy and telecommunications sectors, there is a clear need for revenues to be raised, but at the same time, establishing as much competition as possible i s critical to securing Turkey's competitive position. Decisions favoring a more competitive structure could lower up-front revenues from privatization, but will maximize the long-run benefits. The second trade-off involves whether to prioritize sale o f large companies, which can make a significant contribution to the fiscal program, or to focus on the sale o f smaller firms which may not generate as much cash revenue but may be extremely important for stimulating the private sector, at least in the short run. 157. Despite the high competence and considerable experience o f Privatization Administration (PA) managers in executing privatization transactions, the program has suffered from many flaws, The requirement that the Privatization High Council approves all sales, regardless o f size, has forced even small divestituresto be recycled several times-significantly lowering their value and reducing transparency. A number o f loss-making companies have beentransferred to the PA, and they are absorbing its resources-both financial and human. The PA has effectively become a large state holding group-a function it i s not equipped to handle. The bulk o f the funds from privatization have been used to finance the unprofitable holdings rather than being 81 returnedto the Treasury to retire debt. Privatization has been viewed as an important way o f generating revenue to overcome fiscal problems, particularly during crises. The more fundamental objective o f privatization-making Turkish industry more efficient and more competitive-has beensubordinated. 158. Determined action to improve and de-politicize the privatization process will help strengthen the Government's reform credentials. As a first step, the PA's mandate should be broadened to include a "fast track" privatization process for non-strategic companies in its portfolio. Inparallel, the cross subsidizationmechanism among companies inthe PA's portfolio should be terminated to improve transparency and operational performance. The P A should be free to clear out its inventory of loss-making companies through mergers and liquidations. All privatization revenues and dividends from profitable companies in the PA's portfolio should be transferred to the Treasury and recorded inthe budget. The costs of subsidizing loss-makers and of severance payments and other payments to workers laid-off or retired as a result o f privatization should be provided transparently through the budget. Finally, future privatizations should not be carried out by line ministries but by the P A in order to take advantage o f its experience and technical abilities. Despite significant tariff and non-tariff protection and price adjustments, some o f the companies to be privatized experienced losses in recent years. The substantial accumulation o f arrears needs to be cleared as part o f the privatization process for these companies. 159. A new privatization program for 2003, which includes major companies of national importance, has been approved. The current privatization program targets several major holdings: those in PETKIM (petro-chemicals), TUPRAS (oil refining, partly divested in 2000), Turkish Airlines, ERDEMIR (steel), TEKEL (tobacco and spirits), and SEKER (sugar). The Privatization Administration should urgently adopt privatization plans for TEKEL and SEKER. The Privatization Administration should also continue to divest its smaller holdings. It should also amend the telecommunications law to authorize privatization o f Turk Telekom subsidiaries and adopt new Turk Telekom privatization plan. In addition, state banks will be privatized over the next several years, and intervened banks will be divested or liquidated by the end o f 2003. Moreover, the Government plans to sell other assets, such as some of its landholdings. 160. The 2003 privatization program has made reasonable progress. Based on sales contracts signed as o f mid-May 2003, cumulative cash proceeds were in line to reach the end-June target o fUS$90 million. Highlightso f the program are as follows: The PA has completed tenders for four TDI ports, Taksan (machinery), Gerkosan (iron and steel), TZDAfj (agriculture machinery), and six SEKA facilities in the framework o f the 2003 program. All o f these tenders were approved by the Privatization High Council except for those which did not receive sufficient bids (Gerkosan, two o f the SEKA facilities, and two TDI ports). Block sale tenders for 51 percent of the ET1 group companies (ET1 Copper, ET1 Silver and ET1 Electrometallurgy) have been announced. The PA has also finalized the merger o f TURBAN, TZDAS and TUMOSAN under the Siimer Holdingumbrella. 82 0 A block sale tender for the entire 88.86 percent state stake in PETKIM (petrochemicals) was finalized on June and the winning bid was approved by the Privatization High Council inJuly. 0 A block sale tender for the entire 65.76 percent state stake in TUPRAS (refineries) was launched inearly June and will close in September 2003. 0 The PA, working with international consultants, has formulated a privatization plan for TEKEL (alcohol, tobacco and salt), including strategies for dealing with the key issues o f over-employment, debt settlement, and excess inventory. The TEKEL privatization plan was approved by the Privatization High Council in March 2003. Tenders for 100 percent o f TEKEL's alcohol and tobacco operations were launched in early June and will close in September 2003. 0 A block sale/public offering for THY (national airlines) is scheduled for the third quarter of 2003. ImprovingFinancialReportingandAccountingStandards6 161. Turkey's accounting standards do not fully comply with IAS or generally accepted accounting principles. This setup makes it impossible to assess accurately the real financial situation o f companies because o f obstacles such as high inflation, use o f accelerated depreciation laws for taxation purposes, accounting treatment o f foreign exchange losses and profits, treatment o f finance charges as revenue in cases o f deferred payment sales, and lack o f consolidation requirements. Inthe absence o f clear financial data, banks often make corporate credit decisions based on collateral. As a result companies that would not be eligible for credits based on business and cash-flow viability are able to borrow, while others may not be able to even though they have a viable business-but not the required collateral. The main problem i s the lack o f IAS in the reporting and disclosure required for listed companies by the Capital Markets Board, and for unlisted holding groups that have listed subsidiaries or affiliates or that own banks. Turkey's recovery from its crisis involves, among other things, an expectation o f substantial foreign direct investment. While there are many reasons Turkey has not received as much foreign investment as expected, if it i s to become more attractive to foreign investors inthe future, Turkey mustprovide financial statements that meetinternational standards. 162. The solution is to apply IAS in order to determine the real profitability of companies which will benefit all stakeholders inthe long run. It i s important that accounting standards in Turkey be fully consistent with international standards. Maintaining separate Turkish standards will slow firms' access to international markets. While Turkish banks were required to adopt IAS for 2002, the Capital Markets Board should apply these standards to all listed companies, including unlisted subsidiaries and affiliates o f major groups. A number o f groups have on their own moved to IAS and to audits based on generally accepted accounting principles, primarily to raise funds in international capital markets. In addition, reporting should be required on a consolidated basis for all groups. Disclosure rules are needed for all contingent liabilities between a holding company and its subsidiaries as well as collateral (such as shares o f the subsidiary) given to banks for loans or guarantees provided by the parent for subsidiaries or This discussion on accounting standards is based inpart on a Deloitte and Touche report. The Bank's recent Study on Non-Bank Financial Institutions and Capital Markets in Turkey (April 2003) also discusses the accounting standards issue. 83 cross-guarantees from one subsidiary to another. During Korea's recent crisis, a web of guarantees left many chaebols in trouble. The chaebols were unable to detach good assets (subsidiaries and affiliates) from bad because o f the cross-guarantees from subsidiaries and affiliates to other subsidiaries and affiliates. ImplementingIAS 29 in Turkey will enable banks and capital markets to analyze the financial situation o f holding companies and their subsidiaries. An Accounting and Auditing Standards Board should be established to provide the leadership and incentives needed to rapidly expand the use o f these standards. IAS may not be suitable for all types o f companies. For example, small and medium-size unlisted companies may not need financial statements prepared according to these standards. However, companies interested in raising capital from foreign sources must prepare financial statements basedon IAS. 163. The Capital Markets Board has taken a major step with its new consolidation standards applicable starting from 2003. Previously, Turkish accounting standards did not require consolidation, except for a requirement that banks prepare, in addition to their main annual financial statements (which are unconsolidated), financial statements consolidating their financial sector participations. The lack of consolidation has been a serious weakness considering that large parts o f Turkish industry are controlled by holding companies. The financial statements o f these holding companies show only the dividend income from subsidiaries. Transactions among group companies often distort the reported profits and assets o f individual companies. Consolidated reporting will correct for such transactions. International Accounting Standard 28 requires the equity method o f accounting for associated companies (normally those with a shareholding o f 20-50 percent). Under the equity method, the reporting company includes in its income statement its share o f the profit or loss o f its associated companies and adjusts the carrying value o f its investmentin each associated company to reflect changes in each company's net assets. Untilthe new Capital MarketsBoard standards came into effect, equity accounting was not required by any Turkish standard except in the financial statements o f banks consolidating their financial sector participation. 164. The lack o f inflation accounting has traditionally been a problem in Turkey which distorts the financial statements o f businesses in fundamental ways. First, reported profitability may not be "real" for several reasons. Ifthe real value o f receivables erodes during the period up to collection, companies will charge higher prices to compensate for the delay in collection. High prices inflate reported gross margins, but losses from the erosion o f receivables during the credit period do not appear in financial statements. The amount can be significant when credit periods are long. Payables erode inthe same way, but receivables are often larger than payables. Reported profitability and gross margins are overstated because the goods sold were based on inputs purchased earlier, when prices were lower. Second, the real value o f fixed assets- especially land-becomes much higher than the book value. Similarly, the current value of stocks i s often higher than the purchase cost. Finally, if interest rates are very high but a large part o f the interest compensates for an erosion o f the principal, companies with borrowings may report profitability below real levels. Financial statements prepared in Turkey ignore all these problems except for the annual revaluation o f tangible fixed assets and conversion o f revaluation reserves to share capital. Compulsory standards for inflation accounting have beenissuedby the Capital Markets Board and they have become a requirement for the listed companies starting in 2003. Draft legislation on inflation accounting has also been prepared by GDR to address taxation related issues. At the same time, if disinflation progresses in line with the program targets, then the situation will improve automatically. 84 165. There are other obstacles to understanding the real position and financial performance of Turkishbusinesses. These include the lack o fpublicly available information, which can only be solved with changes in legislation, and a widespread belief that many transactions are not recorded infinancial statements for tax evasion purposes. Although there i s little research on the second problem, it i s thought to affect mainly small andmedium-size enterprises. EnhancingCorporateGovernance 166. Internationally, there is growing recognition that strong corporate governance standards are important in establishing an attractive environment for long-term institutional investors. The Commercial Code provides the legal framework for corporate governance inTurkey, focusing on the duties andobligations ofdirectors andprotection ofminority shareholders. Inaddition, listed companies must comply with Capital Markets Board requirements on disclosure, independent audits, and stronger minority shareholder rights. Apart from these laws, there i s no separate corporate governance code in Turkey. The Capital Markets Board should consider adopting a Corporate Governance Code that incorporates principles o f transparency and disclosure and that requires timely, accurate information along with enhanced protection o f minority shareholder rights. Inaddition, an institute should be established to train independentdirectors for seats on the boards o flisted companies. 167. Corporate governance should follow a few basic principles. Boards should operate independently from management with a clear fiduciary mandate to oversee company performance and make policy decisions on such crucial issues as reorganization, sales o f substantial assets, mergers with or acquisitions o f other companies, efforts to borrow or raise capital, and adoption o f corporate strategies. Increasingly, good practice principles o f corporate governance require that a substantial portion o f board members be independent-that is, not management. It i s important to foster responsibility for performance and protect the decision- making power o f top managers, as well as the fiduciary and oversight responsibilities of the board. To this end, the roles and decision making responsibilities for execution (management) and supervisiodpolicymaking (the board) should be separated. InTurkishbusinesses, board and management positions are usually filled by the same individuals making for less effective overall governance. All shareholders, including foreign and minority shareholders, should have equal voting rights and transparent, enforceable rights in the case o f significant corporate events such as dividends, mergers, acquisitions, liquidation, and bankruptcy. The Turkish Commercial Code does not provide sufficient voting equality and minority shareholder rights. Each share does not necessarily have only one vote; multiplevoting powers are allowed. Minority shareholder rights are infringed on inthree main areas: dividends policy, voting rights (that is, blocking rights) for key corporate events, and treatment in cases o f bankruptcy or liquidation. Finally, as noted above, corporate activities and performance should be reported and disclosed in a clear, understandable, consistent way based on IAS. Transparency should also be required for entire holdings. Hence requirements for consolidating and disclosing financial results should be imposed on listed holding companies as well as private ones that have listed subsidiaries. 85 CHAPTER SOCIAL POLICIES 5: - THEIMPACTOFTHE CRISISAND LABOR MARKET MEASURES TOWARDSTHE GROWTH' 1 INTRODUCTION 168. This chapter evaluates the evidence on the social costs o f crisis in Turkey and then addresses labor policy measures to sustain economic recovery and growth-the principle theme o f this CEM. The three year period between 1999 and 2001 was a tumultuous one for Turkey: There was a financial crisis, drought and the Marmara earthquake in 1999. Although the economy began to recover in 2000, a banking crisis hit towards the end o f the year followed by the full-fledged economic crisis o f 2001. The evidence suggests that the cumulative social impact o f the crisis i s likely to have been high. Unemployment has increased and real wages have fallen sharply. As might be expected, the impact on poverty rates has been marked. The urban poverty rate i s estimated to have almost tripled between1994 and 2001. Coping strategies by affected households have included migration, reduction o f expenditures on food, education and hea1th, 169. Turning the focus to labor market measures that might generate jobs, three issues that potentially affect the demand for labor are considered: labor laws; wages and productivity; and active labor market policies. The review o f labor laws and a comparison with OECD countries finds that Turkey's labor laws are the strictest inthe OECD interms o f employment protection. O f particular note are Turkey's severance pay requirements which are higher than any other country except Portugal, and the restrictions on use o f temporary workers. These laws are meant to protect workers from exploitation and business cycle fluctuations. For most workers, the laws may have the paradoxical effect o f increasingthe insecurity faced by workers as employers avoid paying severance all together and hire short term workers illegally. Relatively low real wages in manufacturing has meant that Turkish labor i s competitive with a selection o f other countries, as wages are low enough to compensate for the higher productivity. The review o f international experience with active labor market policies suggest that these can have limited impact on job creation but there are lessons for program design. The Unemployment Insurance Scheme introduced in August 1999 has had a promising start but now that it i s established, the redundancy with the mandated severance pay i s marked. An important prerequisite for more effective functioning o f the unemployment insurance scheme-and indeed any active labor market policy-is the establishment law for the Turkish Employment Organization (ISKUR) adopted by Parliament in July 2003 as part o f the broader legislative package to underpin the institutional and administrative reform o fthe social security system. 'Assessments inthis chapter are basedon the former Labor Law (No.1475).To improvethe flexibility o f labor markets,Parliamentenacted a new Labor Law (No. 4587)on May22, 2003. The new law aims to increasejob creationinthe economy, and it allows for differentiationo f labor contractsby introducingprovisions for part-time and temporaryemployment. The new lawalso improvesworkers' rights, includingthrough requiringemployers to provethat the terminationemployment is basedon a valid reasonas set forth inthe law.The new law will be fully analyzedas partof the on-goingLabor MarketStudy. 86 The Social Impactof the Crisis 170. The crisis conditions have led to higher unemployment and lower real wages2. Between 1998 and 2001,real per capita GNP fell by 14percent, with a particularly deep recessionin2001 before recovering in 2002. However, the unemployment rate actually increased in 2002 to 10.6 percent despite the growth in output. The pattern of employment has also changed. Between 1981 and 1987, employment inindustry remained flat, while construction and services generated growth (World Bank (2000)). Between 1998 and 2002, in contrast, employment in industry includingmanufacturing actually rose, while employment inconstruction and services fell (Table 5.1). This is consistent with the argument inUygur (2002) that Turkey experienceda real estate bubble,particularly a residential real estate bubble, starting inthe late 1980sthat beganto deflate inthe secondhalfofthe 1990sandburstafter the financial crisis of2000-2001. Also to benoted i s the sharp fall in agricultural employment. Agricultural employment peaked in 1999, but then fell. Since a large part of employment in agriculture i s on family-owned farms it i s not typical for there to be lay-offs in the same manner as in urban employment. So the fall in agricultural employment is likely to have come from increased migration to urban areas. Since the urban sectors that typically first absorb rural labor, construction and services, were themselves in a downturn, recent migrants are not likely to have found work easily. Table 5.1: Emplo; nent Indicators 1998-2001 iNP per capita(constant 1987, TL) $3 1,010 1,690,13 1 1,767,1871,572,820 1,668,421 lnemploymentrate 6.4 7.1 6.6 8.5 10.6 Employment,total(inthousands) 20,872 21,414 20,579 20,367 20,287 Employment inagriculture 8,461 8,872 7,103 7,217 6,745 Employment in industry 3,338 3,440 3,738 3,734 3,888 Employment inmanufacturing 3,088 3,386 3,570 3,548 3,675 Employment inconstruction 1,290 1,294 1,313 1,073 931 Employment inservices 7,783 7,808 8,425 8,343 8,723 Public Employment(in thousands) 2,952 2,884 3,002 3,061 3,084 Inflation,consumer prices(annual `YO) 84.6 64.9 54.9 54.4 45.0 This reviewo fthe impact ofthe crisis is basedupona recentWorldBank Study, "Turkey: Poverty and Coping After the Crisis", 2003. 87 171. While the unemployment rate has increased, perhaps even more widely felt has beenthe fall in real wages. For public sector workers real wages fell by 14 percent in 2001, and for civil servants real wages fell by 9 percent. Figure 5.1 shows how for the last decade real wages in Turkey have tracked economic crises in 1994-95 and again in 2000-01. Real wages were increasing quite sharply till the financial crisis of 1994-95 when they fell precipitously. Real wages then started -rising again as the economy began to recover, rising very sharply in 1998 and 1999. Following the financial crises o f 2000-2001, real wages fell in 2001, Real wages for public sector workers are still close to the all time peaks o f 1993 and 1994, although they are likely to fall again in 2002. Real wages for civil servants have tended to be less volatile. I1,000 500 i- Also worth noting are the levels: on 0 1 , I I , I I , , I I I I average, wages for public sector 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 employees have tended to be higher than wages for private sector employees which -Public Sector -+-Private Sector --t Civil Servant inturnarehigherthan for civil servants. Source: SIS 172. As might be expected, the impact of crisis on the poor has beenquite severe. The impact on poverty rates i s most directly observed through household survey data. Such data are available from a 1994 household income and expenditure survey carried out by the State Institute for Statistics and the 2001 household consumption and income survey conducted by Turkish researchers with World Bank support. The surveys are not easily comparable because o f differences in methodology and coverage but do they provide an indication of the change in living standards between 1994 and 2001. In 1994, for the country as whole (that is, urban and rural), 7.3 percent o f the population had expendituresbelow the poverty line, 8.5 percent inrural areas and 6.2 percent inurbanareas. In2001, the urban poverty rate hadjumped to 17.2 percent. It is difficult to assess the impact on rural poverty in 2001 because of difficulties in comparing the data, and the estimate for urban areas should be treated as an upper limit. Nevertheless, it seems quite clear that there was a sharp increase in poverty. Many households were forced to reduce food consumption during the crisis. More than three-fifths o f the urban households reported cutting down on food consumption according to the 2001 survey as did nearly three fourths o f rural households. Thirty five percent o f urban households and 44 percent o f rural households reported spending less on health and education. 173. Poverty inTurkey i s most closely linked to where one lives. The Southeast accounted for less than 7 percent o f the urbanpopulation but 25 percent o f the poverty. This region has been the major source o f migration to urban areas, but much of this migration i s to the Marmara region rather than to the urban areas within the Southeast. Households with children were also associated with urban poverty, particularly households having more than one child. Households with elderly dependents or a female head were not more likely to be poor than the general population. L o w levels of education and not having ajob were both associated with poverty but not by as much as living in the Southeast or having children in the household. Those with a university education are very unlikely to be poor. Urban households whose head had only a 88 primary education had a poverty o f 25.9 percent compared to the average o f 17.2 percent. The poverty rate for households whose heads reported themselves as unemployed was 30.6 percent, but only 7 percent of the urban population reported their heads as unemployed. While a comparison o f the correlates o f poverty in 1994 and 2001 was not possible for this study, it seems unlikely that these correlates have changed significantly. Traditionally, one o f the most visible means that poor households in Turkey use to improve their livelihood is to migrate to urban areas. The major cities, Istanbul, Ankara and Izmir have grown dramatically, although there was some anecdotal evidence o f reverse migration in2001. Labor Market Measures for Job Creation 174. Over the medium term, a private sector that generates well-paid jobs will be the key to improving livelihoods in Turkey. The rest o f this chapter focuses on the other side o f the equation - labor market measures to generate private sector jobs. Some very important measures for job creation that do involve the labor market are discussed elsewhere in this Memorandum, including restoration o f macroeconomic stability, a sound financial system and a well functioning corporate sector. The focus in the remainder o f this chapter is on possible measures in the labor market for job creation through labor laws, labor costs and active labor market policies. 175. Labor market characteristics. A striking feature o f the Turkish labor market i s the large and growing gap between the labor force (adult population who are either employed or looking for work) and adult population (Figure 5.2). In2001, o f the roughly 46 Figure 5.2: Adult Population, Labor Force and Employment million adults, only 22 million were in (1923-2002) 50,000 the labor force and o f these, 20 million 45,000 were employed. Half the adult 40,000 35,000 population, (conventionally defined as 30,0004 , /- the population above 15) was not 25,000 20,000 employed and not looking for work. 15,000 This is largely because women are not in 10,000 the labor force. What appears to be 5,000 0 happening (World Bank (2000)), i s that 1923 1930 1940 1950 1960 1970 1980 Oct. Oct Oct 2000 2001 2002 as Turkey urbanizes, women in rural 1988 1990 1995 areas are finding that there are not many &Population(lS years andover)('000) +Laborforce('000) +Employment ('OOC wage earning opportunities, particularly Source: SIS for those with low education. They are thus choosing to stay at home. Other possible explanations for the falling labor force participation include discouraged workers dropping out of the labor force andundercounting o fworkers ina growing informal sector. 89 176. An earlier World Bank report, "Turkey: Economic Reforms, Living Standards and Social Welfare Study" completed in 2000 included a chapter on labor market issues in Turkey. The report drew on several background papers3 and explored why the labor market seemed to under- perform in absorbing new entrants and generating employment. Two broad possibilities were investigated: (i) that the growth rate in the economy had been low and (ii) the pattern o f that growth has not been labor absorbing. The finding was that both o f these factors had contributed to the limited absorption o f labor which suggestedthat policy should be aimed at the barriers to growth as well as the biases inthe pattern o f growth and constraints to labor demand. While not a study on the constraints to growth per se, the report pointed to the low investment rate in tradeable sectors and low rates o f productivity growth as major constraints on growth. Value added per worker, a measure o f productivity, was also low relative to Spain and Korea during their growth spurts. Productivity in agriculture was negative for 1975-1990 which held back overall growth and was a contributor to the growing inequality betweenregions. 177. On the question o f why growth had not ledto more labor absorption, the report noted that while services (and in particular construction) and industry had both experienced growth, the growth had led to productivity gains in industry and to employment growth in services. The report documented relative wage movements as a possible explanation. A series o f wage hikes ledto a sharp increase inwages inmanufacturing starting from 1988 untilthe currency crisis in 1995, far outstripping value added per worker, and relative wages were increasing in tradeables (mining, manufacturing, energy and transport) relative to non-tradeables (including construction and trade). Econometric evidence from a panel o f manufacturing industries found that the elasticity o f employment with respect to wages i s indeednegative so that increases inwages will lower employment, but that the elasticity was low. However, even with low elasticity, the magnitude o f the increase in labor costs was suggestedas a factor inthe low labor absorption in industry. 178. Labor law. Work i s such an important part o f well-being that inevery economy that the labor market i s always more regulated than other markets. Regulation i s often classified into two broad types4. The first i s statutory regulation which refers to the laws and decrees that define the legal framework such as legislation governing the hiring and firing o f workers, occupational health regulations, minimum wage or non-wage benefits, or the right to unionize. The second type o f regulation, collective voice, results from negotiations between labor represented collectively and employers who are also sometimes represented collectively. All economies use both kinds o f regulations although countries with an Anglo-Saxon heritage o f common law tradition such as the UK and U S tend to rely less on statutory law than countries with a civil law tradition such as France and Spain. Turkey tends towards the latter group (Sural (2001)). What actually happens in practice betweenemployer and worker i s context specific. Labor laws may not be followed in all instances. Union density may be low and/or bargaining agreements may extendto non-unionized labor (Betcherman et a1(2001), Flanagan(1999)). The backgroundpapers on labor marketswere by Erol Taymaz(Trade Liberalizationand Employment Generation:the experienceo f Turkey inthe 1980s); HakanErcan(The Structure of Turkish Labor Markets: 1988- 2025); Alpay Filiztekin(ConvergenceAcross TurkishProvinces and SectoralDynamics) and MarniaLazreg(Rural to UrbanMigrantWomen's Participation inthe Labor Force) See Betcherman,Luinstraand Ogawa (2001) for a very usefulreview. 90 179. Labor law i s customarily divided into individual and collective labor law. The first law on unions was enacted in 1947, collective bargaining i s currently governed by the Unions Law passed in 1983. Individual labor law governs relations between employers and individual workers. In Turkey', this i s provided by the Labor Act o f 1971. It regulates the formation, form, types, content, and termination o f labor contracts, and the reciprocal rights and duties of the worker and employer, and the settlement of individual labor disputes. Two aspects o f labor market regulation might be considered especially important for labor market outcomes: employment protection and wage determination. Some o f the critical features o f Turkish labor law are presentedbelow: Temporary employment i s defined as employment that will be continue for at most thirty working days due to its nature, and i s subject to the Code o f Obligations rather than the Labor Act. Employment contracts necessitating employment for more than thirty days are regarded as permanent. The minimum age for employment i s 15 years except for heavy and hazardous work in which case it i s 18 years. The statutory working week i s 45 hours and overtime i s compensated at one and half times the normal hourly rate. Discrimination i s prohibited basedon the principle o f equal pay for equal work for wage determination. Employment protection i s provided through the provisions o f the Labor Act that deal with termination o f employment. An employer can dismiss a worker either by giving a term o f notice or for ajust cause. With the term o f notice, no reason needs to be given and the length o f the notice requireddepends on the lengtho f employment. Just causes are classified under reasons o f health; misconduct or improper behavior by the worker; or force majeure (if the worker simply does not report to work for a week). Unless the worker i s dismissed for reasons o f misconduct or improper behavior, severance compensation i s due to the dismissed worker who have been employed for at least a year. Severance compensation i s equal to thirty day's pay for each year's service although the thirty day period may be increased through labor contracts or collective bargaining. There is however a ceiling on severance compensation which i s equal to the amount o f the retirement bonus accruing to the highest ranking official subject to the Civil Servants Act. To protect against abusive or discriminative dismissal or for union activities, a worker who can prove that dismissal was for one o f these categories i s eligible for bad-faith compensation amounting to triple the notice compensation or for unionism compensation for which the minimumamount has beenset at the basic annual wage. 180. The rules can be expected to have a significant impact on labor market outcomes. For example, employment protection rules are meant to enhance job security by making dismissal costly to the employer. One result should be to stabilize employment levels, all else being equal, reducing layoffs indownturns but also reducinghiringas the economy recovers. There may also be the unintendedconsequence o f shifting labor demand to the informal sector or employment types. Betcherman et a1 (2001) suggest that over the last decade, OECD employment protection rules have remainedrather stable, while these have been scaled back in Latin America. The US i s regardedas having the least restrictive laws on employment protection. The description of Turkish labor law is basedon Sural(2001). 91 181. Although it i s intrinsically difficult to compare employment protection across countries, perhaps the most comprehensive effort has beenmade by OECD and presentedinOECD (1999). The strictness of employment protection is measured along three dimensions: regular employment, temporary employment and collective dismissals. These are then combined into overall measures o f the strictness o f employment protection legislation (see Table 5.2). Version 1 of the overall measure does not incorporate the indicators on collective dismissal to allow comparability with data from the 1980s. Version 2 does include the indicators on collective dismissal and i s thus the more comprehensive measure. Table 5.2: SummaryIndicatorsof the Strictnessof EmploymentProtectionLegislation Regular Temporary Coliectiw Overall EPL strictness ' employment employment dismissal Verrton 1 \'errlo" 2 Late 1980s Late 19903 Late 1980s Late 1990s ,:::~ Late 1980s L a t e 1990s Late 1990s entral and Western Europe Austria 2.6 2 6 1 8 1.8 3.3 2.2 (8) 2 2 2 3 Belgium 1.5 1.5 4.6 2 8 4 1 3 1 (13) 2.1 2.5 France 2.3 2 3 3 1 3.6 2 1 2 7 ( I O ) 3 0 2.8 Germany 2 7 2 8 3 8 2 3 3 1 3 2 (14) 2.5 2 6 Iceland 1.6 1 6 0 3 0.3 2.1 0 9 (4) 0 9 1 1 NetherIan dr 3.1 3 1 2 4 1.2 2.8 2.7 (11) 2 1 2.2 Sw,tzerlan d 1 2 1 2 0 9 0.9 3.9 1.o (6) 1 0 1 5 United Kingdom 0 8 0 8 0 3 0 3 2.9 0 5 (2) 0 5 0.9 ,"them Europe Greece 2.5 2.4 4 8 4 8 3.3 3.6 (16) 3 6 3.5 Italy 2 8 2.8 5.4 3.8 4.1 4 1 (18) 3.3 3.4 Portugal 4.8 4.3 3 4 3 0 3 6 4 1 (19) 3.7 3.7 Spat" 3 9 2 6 3 5 3.5 3 1 3 7 (17) 3.1 3.1 Turkey 2.6 4.9 2 4 3.8 3.5 ordie countries Denmark 1 6 1 6 2 6 0 9 3.1 2.1 (7) 1 2 1.5 Finland 2 7 2.1 1.9 1 9 2.4 2.3 (9) 2.0 2.1 Norway 2 4 2.4 3.5 2 8 2.8 3.0 (12) 2.6 2.6 Sweden 2.8 2 8 4.1 1 6 4.5 3.5 (15) 2.2 2.6 ransition :anomies Czech Republic 2.8 0.5 4.3 1.7 2.1 Hungary 2.1 0 6 3 4 1 4 1.7 Poland 2 2 1 0 3 9 1.6 2 0 o r t h America Canada 0.9 0,9 0.3 0.3 3.4 0 6 (3) 0.6 1 1 I\textco . . 2.3 3.8 United Stater 0 2 0.2 0 3 0 3 2.9 0.2 (1) 0.7 pia and O c e a n i a Australia 1.0 1.0 0.9 0 9 2.6 0.9 (5) 1.2 Japan 2.7 2 7 2.1 1.5 2.4 (17) 2.3 Korea 3 2 2.1 1 9 2 6 (20) 2.5 New Zealand I.7 0.4 0 4 1 0 (7) 0 9 )at& not available. a) Figures m brackea show country rankings All rankings increase wtrh the stiicfness ofemployment prorectmn Source OECD (1999) Employment Outlook 182. Looking at the country rankings, Turkey i s rated as having the strictest employment protection legislation o f the 26 countries for which data are available under version 1 and second only to Portugal by version 2. The regional pattern i s striking, the 5 countries listed in Southern Europe have the strictest employment protection legislation o f all the countries listed by both measures. An extension o f the OECD measures to Central and Eastern European countries (Riboud, Sanchez-Paramo and Silva-Jauregui (2002)) suggests that most o fthese countries fall in the middle of the OECD scale, implying that Turkey's legislation is stricter than in these countries as well. On examining the breakdown, Turkey's employment protection legislation i s ranked 19fhout o f 27 (only 8 countries were stricter) on employment protection for regular employment. Turkey is rated as the strictest country o f all on regulation o f temporary 92 employment, and has relaxed regulation o f collective dismissal (relative to individual dismissal) with only 5 countries have lighter regulationon collective dismissal. 183. What i s driving these overall results? O n regulation o f permanent workers Turkey's regulations are about average on procedural inconvenience (12 out o f 27) or difficulty o f dismissal (12 out o f 27). The Job Security Act passed in August 2002 i s likely to change this ranking and make Turkey's regulation on these issues more strict withinthe OECD. Interms of notice and severance pay required for no-fault individual dismissals, only Portugal i s rated stricter. Turkey's severance pay regulations described earlier imply that a typical worker with 20 years tenure i s entitledto 26 months severance pay. Infact, for public sector workers this may significantly understate the benefits received on termination. The World Bank has carried out detailed calculations o f the payments to be received by 40,000 workers in 23 state-owned enterprises (SOEs) in Treasury's portfolio and 22 SOEs in the Privatization Administration portfolio. These calculations which include job loss compensation and notification payments besides the other statutory payments indicate that average payment to worker i s equal to 36 months of net salary for workers in the Treasury portfolio and 38 months for workers in the Privatization Administration portfolio. This makes costs o f dismissing workers almost prohibitively high. 184. On regulation o f temporary workers, Turkey's regulations are rated very strict because fixed-term contracts are allowed only for objective reasons such as the task itself being o f temporary duration. It i s used mainly for seasonal and agricultural work. This i s a particularly strict rating as only Greece and Turkey o f the countries rated have such restrictive criteria for the use o f temporary workers. Temporary work agencies are not allowed at all inTurkey, again with the exception o f agricultural work. Here too, Greece and Turkey have the strictest regulation against temporary work agencies. Indeed, fixed term contracts and temporary work agencies are the areas where OECD countries have by and large moved to lighten the regulations. In the past, countries in Europe in particular allowed fixed term contracts only for objective or material reasons but most o f them have moved to lift this requirement. Similarly, the general trendhas beento liberalize the use o ftemporary work agencies. 185. For collective dismissals relative to individual dismissals, the OECD measured strictness by the number o f workers at which the dismissals are deemed collective, the additional notification requirements, additional days involved and other special costs to employers. On all these measures, Turkey's regulations are not among the strictest relative to the comparator countries. 186. Turning to the empirical results o f the international evidence on the impact of labor market regulation, the evidence on whether stricter regulation hampers job creation i s mixed. Influential studies inLatinAmerica (Heckman and Pages (2000)) and the OECD (Lazear (1990)) indicate that job security provisions do increase unemployment, although Blank and Freeman (1994) and Freeeman (1999) argue otherwise. The survey by Betcherman et a1 (2001) o f the empirical literature suggests that stricter employment protection for permanent employees leads to lower labor turnover (less firings but also less hiring); longer average unemployment durations, lower labor force participation rates and positive employment effects for skilled prime age males, but lower employment for women, young people and less-skilled workers. 93 187. Insum, it appears that to create a more dynamic labor market with more employment in the formal sector particularly for women and the young, Turkey should reexamine the regulations on temporary workers as well as the severance pay and notice required for dismissal of permanent workers. This will weakenjob protection for permanent workers but on both these dimensions, Turkey has some o f the strictest protection inthe OECD. The Labor Security Law approved by Parliament in August 2002 i s likely to have made Turkey's regulation o f regular workers (where Turkey ranked in the middle) more strict. As consideration i s given to labor market reforms, attention should be paid to the unintended consequences o f the existing regulatory framework. Labor regulation i s meant to protect workers from exploitation, and from undue business cycle fluctuations. The evidence suggests that for a substantial majority of workers it does neither. As will be seen in the next section, real wages in Turkey have shown considerable flexibility. For the roughly 10 percent o f the people who are employed in the public sector, job security i s high. For those employed in large firms who obey the law, job security may also be high because the very high severance payments makes it expensive to fire workers. 188. For the majority o f the work force, labor laws are reportedly disregarded through measures such as unregistered workers, firings o f workers after 11 months just before they are eligible for severance pay and then rehiring, and misreporting o f wages. Inthese cases the labor laws may actually increase insecurity by giving firms incentives to disregard them. Indeed, a survey carried out by the Foreign Investment Advisory Service (FIAS) finds that employers do not find labor laws to be a major constraint precisely because they often do not follow them. However, these laws may well be a factor hindering the entrance o f foreign investors who typically do not have this choice. 189. Jobs, productivityand labor costs. The impact of wages onjob creation is a complex interaction with multiple feedback loops. Two channels are analyzed here: (i)the impact o f wages on macroeconomic stability and (ii) the impact of wages on competitiveness of Turkish firms. There are a multitude o f studies which demonstrate the importance o f macroeconomic stability to sustained growth and hence job creation. Wages can affect macroeconomic stability inat least two ways. The first is the role ofpublic wages indeterminingthe government deficit. The second is the role o f wages ininflation. 94 190. Stabilizing the economy with reasonable inflation and interest rates requires a sustainable public deficit which in turn requires sustainable public expenditures. An analysis o f public expenditure patterns in the Public Expenditure and Institutional Review (World Bank (2001)) found that, relative to OECD countries, Turkey stood out on two categories: interest payments, and wages and salaries. Public sector wages and salaries in Turkey amounted to 24 percent o f central government spending in 1999. In the OECD only Portugal and Greece have higher shares allocated to wages, -while in the other countries the shares range from a low o f 2.9 percent inAustralia to a higho f 16 percent in France. The data on public employment in Table 5.3 suggest that, unsurprisingly, total employment is less flexible than wages. During the crisis period from 1999-2001, public employment was actually increasing. As the PEIR points out reducing the public wage bill to sustainable level should be addressed in the context o f a comprehensive public sector reform Private 17,920 18,530 17,577 17,306 17,203 process that strengthens institutions to Public 2,952 2,884 3,002 3,061 3,084 deliver services. Such an approach should Grand Total 20,872 21,414 20,579 20,367 20,287 involve a functional review o f government and be grounded in a broader civil service reform strategy, in addition to privatization and specific policies to address over-employment in the state enterprises. Turkey's public sector reform program, as presented in Chapter 2, i s consistent with this approach. As yet, however, the program has not set an explicit target for public sector salaries or for reductions inthe public wage bill. 191. The evidence is mixed on whether wages a major source o f inflationary pressures inthe Turkish economy. As can be seen in Figure 5.3, real wages, particularly for public sector workers, rose sharply immediately before the crises o f 1994-95 and in 2000-01, and then fell immediately afterwards. This pattern was partially the consequence of having backward indexation in an institutional setting where wage setting i s staggered, Le., determined once or twice a year (Taylor (1998) provides an overview of the large literature). When inflation declines rapidly, wages are slow to adjust since they are set once or twice a year, and real wages rise; conversely, as inflation rises real wages fall. In 2000, the macroeconomic stabilization program envisaged a shift to forward indexation to the inflation target (World Bank 2001) but true forward indexation in private sector has not materialized and incomes policy has remained relatively weak under the program. Nevertheless, empirical analysis by the IMF Research Department suggests that inflation inertia has significantly decreased inTurkey IMF (2002)). 95 Mexico 21 17 14 Poland 28 27 Portugal 27 23 Spain 28 28 31 36 96 192. The impact of wages on competitiveness can be measuredby looking at wages relative to productivity and also by comparing with other countries. Changes in labor costs should track changes in productivity over the long run. While long-run data on productivity are not readily available inTurkey, Figure 5.3 charts the evolution of labor costs and value added per worker in manufacturing, a proxy for productivity, between 1991 and 2001. Figure 5.3: Indices of Value Added in Manufacturing and Wages in both the private sector and I Real Labor Costs public sector increased faster than 1200 value added prior to 1994, thereby reducing competitiveness and presumably labor demand-at least in manufacturing. Wages subsequently fell relative to value added between 1994 and 1996 and then rose again, 0 ~ 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 particularly in the public sector. Despite falling in 2001, labor costs -tValueAddedperWorker(manufacturing) remained significantly above --ePublicSectorLaborCostIndex --tPrivateSectorLaborCostIndex productivity levels from the 1990base. Source: SIS 193. OECD data can be used to present international comparisons of labor costs (Table 5.4). These comparisons (in US Dollars) are indicative as the underlying data are collected by the national statistical agencies in slightly different manners and the data reflect exchange rate movements. Currently, labor productivity in Turkey is roughly comparable to Hungary and Poland, but well below the levels reached in Korea, Portugal or Spain. Labor costs are slightly below levels in Hungary, Poland and Mexico and well below Korea, Portugal and Spain. In terms of the ratio of value added to labor costs, the ratio in Turkey has declined from the very high levels seen in the 1980s, but labor i s still competitive in Turkey compared to the other countries. These results are confirmed by the detailed background sectoral studies (textiles and clothing, and automotive sectors) carried out for this report. 194. Active labor market policies. Countries in the OECD have increasingly been implementing so called Active Labor Market Policies (ALMP) to assist job seekers to secure jobs. The genesis of these programs is traced back to the 1970s. Rising unemployment in Europe led to increasing concerns about the costs of the social safety net for the unemployed along with concerns that the safety net was creating incentives againstjob seeking and work. It was felt that a "trampoline" that assisted the unemployed to actually secure jobs was more appropriate than a safety net. Current terminology usually refers to safety net type programs as passive labor market policy and trampoline type programs as active labor market policy. Active labor market policies originated inEurope and most experience is still inthe European countries. Typically active labor market programs are implemented side by side with existing safety net programs but ALMP are increasingly spreading through eastern Europe and Latin America where formal safety net programs are less prevalent. Active labor marketprograms6can include (i)job creation, (ii)training and(iii)employmentprogramsprogramsto matchworkers and employers. These are describedand design issues are briefly discussed below. Also described This reviewof active labor market policesdraws heavilyon Betcherman,Dar, Luinstraand Ogawa (2000) 97 i s the now-significant empirical evidence on how the various programs have worked in practice. A large number o f studies that evaluated labor market programs were reviewed by Dar and Tzannatos (1999) and summarized inBetcherman et a1(2000). 195. Job creation programs fall into three broad categories. Wage subsidies are meant to encourage workers to hire new workers or to keep employees who might otherwise have been laid off. These subsidies are targeted to a particular category o f worker or employer (examples include the long term unemployed, those coming from disadvantaged areas or youth) and can take the form o f a direct wage subsidy or social security offset. The level and duration o f the subsidy varies significantly betweenprograms and countries. Inthe US, under the Targeted Job Tax Credit, firms are paid 50 percent o f the individual's wages for a period o f two years, while the UK job subsidy program provides up to 100 percent o f wages for a period o f six months. Monitoring employer behavior to prevent abuse i s critical. A review o f 22 evaluations o f wage subsidy program in different countries found that these programs are susceptible to deadweight losses (program outcomes not being different that what would have happened without the program) and unintended effects such as subsidized workers replacing unsubsidized ones or employers hiring subsidized workers and laying them off once the subsidy period ends. While very few o f the studies carried out a full cost benefit analysis, wage subsidies are unlikely to have positive economic returns. 196. The second category o fjob creation programs are public works where government either createsjobs directly or contracts with business or non-profits to provide jobs. Public works are typically targeted to the long term unemployed or the poor. They can lead to the creation o f public infrastructure. The key design issue i s the setting o f the wage. The rule o f thumb i s that the wage should be equal to or less than the market wage for unskilled manual labor. This has the advantage o f allowing selftargeting. Only disadvantaged workers are likely to apply at these wages and restrictions on hiring can be avoided. Public works programs are among the most heavily funded in the OECD countries. Seventeen evaluations are available. These show that public works program can have desirable short term effects in raising employment, but they do not have a impact interms o f reducing long term unemployment. These are best usedas a short- runanti-poverty intervention. 197. The third category o f job creation is micro enterprise development which are seen as particularly useful in countries where banks are unable to conduct comprehensive risk assessments required to offer credit to unemployed workers who wish to start their own enterprises. Such programs are common in low income countries such as Bangladesh but have also been offered over the last twenty years inMassachusetts inthe US, Denmark, Hungary and Poland. Experience suggests that good design characteristics include offering small amounts o f initial credit with subsequent loans dependent on repayment record, use o f market interest rates and use o f groups lending rather than formal collateral. The 15 evaluations o f micro enterprise development programs show high deadweight losses and failure rates for the assisted business: typically one-third to one-half o f assisted businesses shut down in the first year o f their operation. However, participants inthe programs were more likely to be employed than control group participants. Little evidence i s available on the cost-benefit o f these programs. 198. Labor market trainingprograms are usually aimed at retraining long term unemployed, retraining displaced workers especially when there are mass lay offs, and training young school 98 drop outs. Key design issues for these programs focus around finding the appropriate role for the government and linking these programs with employers. In Chile, for example, representatives o f employers, employees and government sit on the governing board o f vocational training institutes which has strengthened the ability o f these institutes to respond to the needs o f employers. Training programs typically account for a significant portion o f expenditure on ALMP, typically between 40-60 percent, and it is not surprising that a large number of evaluations are available. 28 studies o f training programs for the long term unemployed found that while these can have a positive impact, it i s no more than the much cheaper job search assistance programs. They are more effective when the business cycle i s expansionary. Tightly targeted on-the-job training for disadvantaged groups offered the best returns. Retraining programs for those laid off en masse (for example through large scale lay offs) show very small and statistically insignificant increases in employment possibilities, and often negative wage effects. Retraining for youth are rarely found to have an effect on employment probabilities or wages. 199. Employment services' programs seek to improve the functioning o f the labor market by matching employers and employees. This assistance can include initial interviews at employment offices, in-depthcounseling during the unemployment spell, job clubs etc. These kind o f services are relatively inexpensive and can shorten unemployment spells. However, often the individuals who benefit most are the better-qualified, and thus are likely to have found jobs on their own. In many countries, private employment services have been established alongside public ones. These should be considered complementary rather than competitive. Private employment services charge fees and typically reach skilled segments o f the work force while public services can target the most needy. Empirical evaluation o f employment services programs found that these have often had positive effects, and i s usually cost-effective relative to other ALMPs. Studies inCanada, Hungary, Netherlands and Poland suggest that these are more effective during economic expansions. 200. The particular mix o f policies that is appropriate for each country is, o f course, context specific. Policy makers have to consider the needs and priorities, institutional capacity and fiscal capacity in guiding their choices. Inrecent publications (OECD 2001,OECD 2002), OECD has been recommending that countries use a mix o f all three kinds o f programs, while noting that these are not panacea. Factors identified as important for success include availability o f information on unemployed people and nationwide information on placements, effective management control mechanisms and supportive legislation. that are important for success. Turkey, after implementing the Unemployment Insurance Scheme described above, i s now preparing to implement active labor market programs with the Ministryo f Labor and the Turkish Employment Organization taking the lead. 99 Table 5.5: Overview of Ac re Labor Market Programs Program Appear to Help Comments 1. Job-search Adult unemployedgenerally Relativelymore cost-effectivethan other labor market assistance/ Emp. when economic conditions are interventions(e.g. training) - mainlydue to the lower Services(19 improving; women may benefit cost, youth do notbenefit usually. Difficulty lies in evaluations) more. decidingwho needshelp inorder to minimize deadweightloss. 2. Training of Women andother No moreeffectivethanjob-searchassistancein long-term disadvantagedgroups generally increasingre-employmentprobabilities and post- unemployed(28 when economy is improving. interventionearningsand are 2-4 times morecostly. evaluations) 3. Retrainingin Little positive impact mainly - No more effectivethanjob-search assistanceand the case of mass wheneconomy is doingbetter. significantly more expensive. Rate ofreturnonthese layoffs(12 programsusuallynegative.However,job search evaluations) assistance maynotbe a direct substituteas it may cater to a different groupso fthe unemployed. 4. Training for No positive impact. Employment'earnings prospectsnot improvedas a youth (7 result of goingthroughthe training. Taking costs into evaluations) account - the realrateo f return o f these programs is negative. 5. Employment' Long-termunemployedin Highdeadweightand substitutioneffects. Impact Wage subsidies providingan entry into the analysis shows treatment group does not do well as (22 evaluations) labor force. comparedto control. Sometimesusedby firms as a Dermanentsubsidvmoeram. 6. Public Works Severelydisadvantagedgroups Long-termemploymentprospectsnot helped:program Programs(17 in providingtemporary participantsare less likely to be employed in a normal evaluations) employmentand a safetynet. job and earn less than do individuals inthe control group.Not cost-effectiveifobjective is to get people into gainhl employment. 7. Micro- Relatively older groups,the Very lowtake-up rate amongunemployed. Significant enterprise moreeducated. failure rate o f smallbusinesses. High deadweightand Development displacement effects. High costs -cost-benefitanalysis Programs(15 rarelyconductedbut sometime show coststo UIbudget evaluations) higherthan for control group. Source: Betcherman,Dar, Luinstra and Ogawa (2000) 201. An important leg of Turkey's passive labor market policy is the recently introduced Unemployment Insurance Scheme. The Unemployment Insurance Scheme was established in June 2000 and has enrolled approximately 4 million workers. Enrollment i s compulsory for workers registered with SSK the social security system for private sector and state owned enterprises employees. Individual accounts are created to which workers transfer 1 percent o f their salary while employers contribute 2 percent and the state 1percent. Workers can draw on their account when they are separated from their jobs. Despite the increase in unemployment with the crisis, only about 39,000 workers were receiving payments as o f October 2002 because o f the requirement that contributions have to be made for 600 days before workers can start receiving payments. The Turkish scheme is a type o f compensation fund where the total unemployment benefit i s a function o f payments made by the worker [n.b this i s to be confirmed]. These funds are becoming increasingly common in Latin America and have been introduced in Chile, Panama, Peru and Venezuela. Compensation funds are attractive because 100 they do not change the incentives for work which is a possibility with severance payments or traditional unemploymentinsurance schemes (Domelundand Gill (2002)). With the successful establishment of the Unemployment Insurance Scheme in Turkey there is now a redundancy with the mandated severance payments and the way i s open to reduce or even phase out mandated severance payments. Given the large surplus building up in the Unemployment Insurance Fund, it i s essential that there i s a comprehensivereview of the future use of the fund with the intention of ensuring that it i s brought into balance with due regard to the existing employment benefit policies. The review should attempt to bringthe fund into balance with an excessof contributionincome over expenditureof notmorethan 15 percent. 101 CHAPTER6: TURKEY'SEXTERNAL PROSPECTSA GLOBALIZINGWORLDECONOMY' IN 1 INTRODUCTION 202. Over the last twenty years, Turkey has become increasingly integrated into the world economy with imports and exports growing substantially as a share o f GNP. There has been a diversification o f the export mix towards manufactures exports and some change in the composition o f export markets with the loss o f Iraq as a significant trading partner after the Gulf war. This has been compensated by increased exports to the European Union, which remains Turkey's most significant market, and some gain in emerging Europe and Central Asia. The liberalization o f trade policies that began inthe 1980s, followed by the 1996 customs union with the EU and more recent free trade agreements with the central European countries, has beenan important enabling factor inthis good performance. 203. A forecasting exercise suggests that prospects for continued export growth are good under a scenario of economic reform. For the period through 2005, Turkey's export growth is expected to be around 9 percent on the average under the assumptions o f Turkey's current strengths, recovery in the EU and more evenly balanced global growth. Goods exports are expected to be driven by growth in exports o f automobiles and parts underpinnedby increased FDI, a reorientation from existing plants to the export sector and aggressive capital expansion2. Electrical machinery and industrial equipment are also expected to grow strongly. Turkey's productivity advantage in textile apparel and clothing i s above competitors, but the advantage i s eroding. Competitive pressures are likely to restrain growth inthe sector3. Tourism revenues are anticipated to grow at 12 percent per year. Finally, factor income receipts will be restrained by lower returns on holdings o f overseas assets, but remittances should recovery strongly. The projected growth inexports i s inline with performance inthe second half o f the 1990s. 204. An analysis of the sensitivity o f the Turkisheconomy to external shocks emphasizes the importance o f macroeconomic stabilization to sustained export-led recovery. While many countries have experienced volatility in income or the real exchange rate, Turkey has suffered exceptional volatility inboth indicators. Volatility in GDP growth and the real exchange rate has beenmuch higher than countries that went through crises such as Malaysia, Korea, Thailand and Mexico, only Indonesia and Russia saw comparable levels o f volatility. The analysis suggests that the main sources o f vulnerability are exchange rate shocks, interest rate shocks, slowdown in export markets and oil shocks. Interestingly, exposure to oil shocks has diminished over the 1990s with increased efficiency in energy use in Turkey and reduced importance o f trade relations with Iraq. Slow recovery in the world economy-particularly delayed recovery in the EU,increasedrisk aversionof international investors, and the potential for post-war instability in Iraq are all significant external risk factors for the medium term. These risks highlight the importance o f maintaining exchange rate competitiveness, containing import demand and attracting more foreign direct investment. A financial crisis usually improves competitiveness as * Taskin 'This chapter draws fromthe backgroundpaper byRiordian, De Kleine, Suttle andTimmer (2002). (2002), a backgroundpaper for this Report, in"The Automotive Industryin Turkey: Overview o f the Sector and Opportunitiesfor Exports". Taymaz(2002), abackgroundpaper for this Report, in"Competitiveness o fthe Turkish Textile and Clothing Industries." 102 the nominal exchange rate collapses, but the tendency for the real rate to appreciate with the recovery should be watched closely. While a combination o f fiscal and monetary tightness i s required, the external analysis reinforces the importance o f fiscal adjustment in order to lower domestic real interest rates and thereby ease pressure on the exchange rate. 205. Turkey must attract more foreign direct investment to promote competitiveness, ensure more stable capital inflows, and repay the large official debt coming due over the medium term. The experience o f other countries such as Chile, Indonesia and Malaysia suggests that it is indeed possible attract FDI immediately after a financial crisis. Despite its large market size, skilled domestic labor, strategic location and competitive local firms as suppliers to multinationals Turkey has not been able to attract substantial FDI inflows. In addition to macroeconomic instability, factors inhibitingFDIhave included political uncertainties, excessive bureaucracy, inadequate regulatory framework, and an opaque and unstable legal environment. Following significant analytical work with inputs from the private sector and the Foreign Investment Advisory Service (FIAS) o f the World Bank, an FDI promotion strategy is being developed and implemented. A structured program to improve the investment environment was adopted by Government in December 2001. This program centers around a Coordination Committee (YOIKK) and technical committees to review and advise the Government on such areas as FDI regulation, company establishment, sectoral licenses, investment promotion and taxes and subsidies. A new FDI law was enacted in June 2003 to upgrade the legal framework and legislation has been preparedto establish an investmentpromotion agency. The Government can build on this momentum by: (a) activating the Investor Council consisting o f high-level representatives of the international business community, (b) initiating a full-scale review o f commercial law and competition policy inTurkey to identify and address important obstacles to FDI, and (c) developing a strategy to facilitate the flow of "greenfield" FDIto SMEs. 2 BACKGROUNDROBUST - PERFORMANCEINEXTERNAL BALANCE Historical Developments 206. Turkey's integration in global trade,andfinance increased substantially over the past two decades with the establishment o f a customs union with the European Union, diversification of the export product and market mix, and further opening o f the capital account. External revenues are now generated from a highly diversified set o f sources, notably with tourism, other emerging services exports and remittances accounting for almost one-half o f earnings. Rapid growth in goods and services exports (11 percent over the decade and 12.5 percent during the latter half), has been a positive development for the country, as exports have come to contribute more-than 50 percent to GDP growth performance during the 1990s. Turkey's substantial gains inexport marketpenetration inthe industrial-as well as developing countries, finds the country well placedto benefitfrom increases inmarket over the nextyears. Turkey's current positioning within the international environment is the result o f a long-term evolution o f the country's integration into global trade and capital markets (Table 6.1). Several points stand out when viewingthese developments. 103 Table 6.1: Summary: Turkey's Links with the Global Economy, 1980-2000 Percentagepoints 1981-1990 1991-2000 1991-1995 1996-2000 2000 Trade in goods and services Openness ratio /a 26.9 48.3 38.9 57.8 66.2 Exports / GDP 12.7 21.9 17.5 26.2 29.5 Imports / GDP 14.3 26.4 21.4 31.5 36.6 Speed of integration /b 11.0 7.3 5.5 9.2 15.2 Export growth 19.7 10.8 9.0 12.5 19.2 Import growth 13.6 10.2 7.6 12.8 25.4 Manufactures share of 63.O 75.2 71.8 79.4 83.8 exports Intra-region trade /%i /c 71.8 74.3 73.2 75.0 73.6 EU-15 39.3 50.3 48.5 52.0 51.7 Europe & Central Asia 5.4 11.8 10.3 12.8 13.4 MENA region 27.0 12.2 14.3 10.2 8.5 Exports per capita /d 290 645 475 810 935 Terms of trade ch% -0.3 -0.6 0.6 -2.0 -8.4 Finance and Debt Current balance /GDP (%) -1.6 -1.1 -0.8 -1.5 -4.9 Private capital inflow /GDP 2.3 3.6 2.2 5.3 8.2 (%I FDI / Fixed investment(YO) 0.12 0.19 0.20 0.18 0.22 EDT/ GDP (%) 29.9 46.5 42.1 49.1 57.7 Variable interest share 32.0 42.3 34.0 45.7 46.8 Debt service / exports (%) 28.7 28.2 29.5 27.4 36.1 Source: World Bank. 207. Openness-or the ratio of trade in goods and services to GDP-has risen Figure 6.1: Trade to GDP has risen by morethan - impressively from 13 percent in 1980 to 66 50 percentage points percent by 2000, reflecting a pick-up in export share o f 25 points and 27 point rise in 25 .................................................................... 70 import share o f GDP (Figure 6.1). The evolution o f Turkey's trade openness has been exceptionally favorable over the 1980s through 2000 (Figure 6.2). The change in share o f trade to GDP since 1980 (about 53 points) finds close similarities with the -5 1......Y.. .--!.....-.-..-......-.......-.......-......!............I 10 performance o f some o f the stronger 1980 1981-90 1991-95 1996-2000 2000 integrating countries, for example Mexico Source: WoYorldBonkdatu,EPPG colculalrons. Opennes: share in GDP ~Jerporrsplusrmpom oJgoodr andnon-jaclor svcs Speed of Integration: Trade growlh less GDP growth and Korea; has well outpaced that of the European Union as a group, as well as the aggregate of middle-income countries. Turkey's performance has narrowed the path to convergence in the level o f openness with the European Union.Turkey's openness ratio (66 percent in2000) contrasts with 74 percent for the EU, stands well above that o f Latin American comparators as a group (42 percent), yet well below that o f 104 East Asian comparators (e.g. 95 percent for the region, Korea: 88 percent; Thailand 101 percent). 208. The pace at which the economy has opened to trade (or speed of integration) has varied over the last decades, with a large portion o f the increase occurring during the 1980s, slower performance in the early 1990s, but a re-acceleration o f trade with respect to output growth during the second half of the 1990s-and markedly during the global boom o f 2000. Within the sample of comparators, Turkey's speed of integration during the second half o f the 1990s (9.2 percentage points) ranks second only to Mexico's rapid rate o f near 12 percent, spurred by NAFTA links (Figure 6.3). Among Latin American countries, Turkey's performance falls closer Figure 6.2: Change in Turkey's openness" Figure 6.3: "Speed of integration" bested strong among comparators only by Mexico over 1996-2000 (*excess of eade growth over GDPgrowth) 81 83 85 87 89 91 93 95 97 99 -21 MericoTurkey Korea EU Argma Mid-Y Brazil Thai Malay Chile lndonera Source. A`ol~onologenc~rr. WorldBank, Economic Policy ondPrmpecrr Group to that o f Chile in terms o f degree o f openness and to that o f Mexico regarding the speed c ~~~~~~ integration over the 1990s. Among East Asian countries, Turkey i s near par with Korea on both indices but still well belowthe extremely open economies o f Malaysia and Thailand. 209. The mix of export markets has also shifted substantially over the 1980s to today. Turkey i s well positioned geographically and culturally to trade intensively with three distinct markets- the EuropeanUnion, Central Europe and the CIS (ECA), and the Middle East and North Africa (MENA) regions. Table 6.1 shows the persistence of western Europe as Turkey's principal export market, rising from 40 percent of Turkishtrade during the 1980s to 50 percent during the 1990s. At the same time the importance o f the MENA region has fallen sharply, in part due to U.N.sanctions upon Iraq, formerly Turkey's largest export market after Germany ,and difficult economic conditions in the region tied to volatile oil markets and sporadic tensions. But a revitalization o f trade with Central and Eastern Europe and the countries o f the former Soviet Unionis inevidence, reflecting stabilization and growth inthis market following the initialyears of transition. On balance, a measure o f the strength o f performance in Turkey's trade-real exportsp e r capita-has risen three-fold from the average o f the 1980s to 2000. 210. Services exports expanded rapidly at 12.7percent per year during the early 1 9 9 0 ~ ~ fell but to half that pace o f advance during the second half o f the decade. Tourism revenues sustained healthy growth o f 9 percent across the decade, while growth inarrivals fell somewhat duringthe second half (realizing an increase in tourism expenditure per visitor). However, volatility and decline has characterized several components o f services revenues in recent years especially those related to construction. The earnings o f Turkish engineering firms abroad declined by 11 percent per annum during the second half o f the 1990s, due to reduced contract work inthe CIS 105 and Middle East regions. This served to slow the advance o f "other services" from a robust 15 percent pace in 1991-95 to 4.2 percent over 1996-2000. The growth o f factor income receipts and transfers ratcheted higher during the latter 1990s, partly offsetting the slackening growth in other foreign revenue sources. Strong earnings growth associated with Turkish investmentassets abroad-tied to the global boom in equity and fixed income markets o f the mid-to-late 1990s-produced an income stream that amounted to some U$2.2 billion per year, double the US$1 billion level during the decade's first half. Worker remittances advanced by a strong 6.5 Table 6.2 Turkey: Current Account Balance (US%billion,1990-2000) 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Trade Balance -9.6 -7.3 -8.2 -14.2 -4.2 -13.2 -10.6 -15.4 -14.2 -10.4 -22.4 Merchandise exports, FOB 13.0 13.7 14.9 15.6 18.4 22.0 32.4 32.6 31.2 29.3 31.7 Merchandise imports, FOB -22.6 -21.0 -23.1 -29.8 -22.6 -35.2 -43.0 -48.0 -45.4 -39.8 -54.0 Services Balance 2.4 2.5 3.2 4.0 3.8 6.4 3.7 7.9 10.5 3.9 7.3 Services, Credit 8.9 9.3 10.4 11.8 11.7 16.1 14.6 21.3 25.8 18.7 22.3 o/w Tourism 3.2 2.7 3.6 4.0 4.3 5.0 5.7 7.0 7.2 5.2 7.6 Services, Debit -6.5 -6.8 -7.3 -7.8 -7.9 -9.7 -10.9 -13.4 -15.3 -14.8 -15.0 o/w Interest -3.3 -3.4 -3.4 -3.6 -3.9 -4.3 -4.2 -4.6 -4.8 -5.5 -6.3 Transfer Balance 4.5 5.1 4.1 3.8 3.1 4.5 4.4 4.9 5.7 5.2 5.2 o/w Workers' remittances 3.2 2.8 3.0 2.9 2.6 3.3 3.5 4.2 5.4 4.5 4.6 Current Account Balance -2.6 0.3 -1.0 -6.4 2.6 -2.3 -2.4 -2.6 2.0 -1.4 -9.8 Current Account Balance (YOof GNP) -1.1 0.2 -0.6 -3.6 2.0 -1.4 -1.3 -1.4 1.0 -0.7 -4.5 Source: Central Bank. I financial markets. The average flow o f remittances increased by US$1.5 billion betweenthe first and second halves of the 1990s-from US$3 billion to US$4.5 billion. In contrast, flows o f official current transfers dropped sharply over the course o f the decade (Table 6.2) 211. Overall, Turkey's current account balance for the 1990s contrasts favorably with comparators. The average annual deficit o f 1.1% o f GDP) matches the outturn for middle- income countries. It represents a much less severe deficit than for Latin comparators, and i s only moderately greater than that for the East Asian countries. Prior to the 2001 crisis, Turkey's external debt profile was fairly well positioned among comparators, although debt/GDP and debt servicing ratios well exceeded the average for the middle income group. 212. Trade policy has played an important role in boosting Turkey's trade integration and Figure 6.4: Growth of Turkey's exports exceed the umwth nf the crnnrt market export performance. Important policies have 25 included removal o f quantitative restrictions on 20 imports adopted during the 1980s; the 1996 I 5 customs union with the EU, and more recently 10 FTAs accorded with Central European 5 countries and a process o f greater opening to 0 countries o f the Mediterranean region (EU- -5 MED Initiative). Diversification of the export -10 -15 mix has also served Turkish integration well. 1990 91 92 93 94 95 96 97 98 99 2000 1001 Rapid market penetration in key products has Source. WorldBonkdolo ond cdculonons. brought about a dramatic shift toward manufactures exports, from 63 percent o f total shipments during the 1980s to 84 percent by 106 2000, helping to boost overall export growth. Turkey's exceptional performance in trade integration has played a part in the longer term development o f the country's key sources o f foreign exchange revenue. Despite difficulties, the evolution has been a positive one, characterized by shifts in products and markets in goods trade; by diversification into services exports, and supported by continuing flows o f private transfers from abroad. Turkey's revenues on current account now stem from a broad and fairly balanced set o f sources Exports have become a key driver o f economic growth as Turkey has opened its economy and increased market penetration (Figure 6.4). Crisis and Recovery Since 2001 213, During2001,the large devaluationassisted inpropelling dollar-based goods exports to an advance o f 11.3 percent (15.2 percent in volume) from growth o f 8 percent in 2000. This performance occurred against a background o f decline in global economic activity, a dramatic contraction o f world trade growth and a 7 percentage point drop inthe growth o f Turkey's export (in volume terms). In effect, Turkey's export market penetration increased by some 11 percentage points, progress which will assist in underpinning export growth further as market activity recovers. Overall, Turkey recorded a current account surplus o f 2.4% percent o f GDP in 2001, compared to the deficit o f 4.9% percent o f GDP in 2000 which played a key role in undermining confidence in the exchange rate based disinflation program. For investors, the rapid correction in the current surplus has been an encouraging factor, although for the Turkish authorities and populace the associated recession and decline in import purchasing power was extremely painful. Strong export performance, buoyant tourism and renewed capital inflows- together with large IMF disbursementshave easedthe pressure on the BOP even as imports have expanded rapidly with the economic recovery. The modest current account o f some one percent of GNP in2002 was easily financed. Export growth has continued into the first half o f 2003, its growth rate was 30 percent indollar terms inthe January-May 2003 period despite the impact o f the Iraq war. However, the real exchange rate has rebounded strongly from its post-crisis lows, appreciating by some 25.5 percent from mid-2001 through June 2003. The strengthening Lira has not yet hurt export performance as it has been counterbalanced by a very sharp drop in real wages. Nevertheless, the widening current account deficit (now projected to reach 3.2 percent o f GNP in 2003) is an emerging cause for concern, particularly as it is being accompanied by a returno fshort-term capital inflows. 3 GLOBAL OUTLOOK 214. Turkey's nascent recovery from the 2001 crisis takes place against a background of an uncertain firming o f global growth. The year 2001 was characterized by sharp deceleration in economic activity and trade, affecting industrial and developing countries alike. World GDP growth fell from 3.9 percent in 2000 to 1.2 percent during 2001, while world trade collapsed from double-digit gains to a mere 0.4 percent growth. The recovery among the industrial countries, which commenced in late-2001 in the United States, followed by the Euro Area in early-2002, faltered inmid-2002, and quarterly real GDP growth inthe major economies slowed from 2 percent inthe first half o f 2002 to 1percent inthe fourth quarter. The recovery in global 107 growth was tempered, among other factors, by corporate-sector adjustment, volatility in equity prices and bothbusiness andconsumer uncertainty tied to concerns about the coming war inIraq. Early indicators suggest continued sluggishness continued into the first quarter o f 2003, due to the impact o f the war in Iraq and the SARS virus outbreak. As a consequence, the anticipated recovery inworld trade volumes i s expected to be delayed untilthe summer. The outlook for the year 2003 as a whole i s for growth to be spurred by firming real fixed investment spending and improving confidence, as the more immediate uncertainties, such as Iraq and SARS are resolved. A high-tech led turnaround in manufacturing sectors in East Asia and the United States is anticipated to spread to Europe through trade links. L o w interest rates should spur a revival o f consumer durable demand and investment over the coming months and quarters leading to a stronger profile for global growth in 2003-05. World GDP i s anticipated to expand at 2.3 percent during2003 over 2002, coming to stabilize near longer-term rates above 3 percent during 2004-05 (Table 6.3). Table 6.3: Global ConditionsAffecting Growth in DevelopingCountries, 2001-2005 (percentage change, multi-year periods compound i erage, except LIBOR) Current Estimates Current Forecasts 2001 2002 2003 2004 2005 I Forld trade (volume) I 0.4 3.0 6.2 8.1 8.1 [nflation(consumer prices) G-7 OECD countries" b' 1.5 1.o 1.4 1.3 1.3 UnitedStates 2.8 1.6 2.5 2.3 2.1 World GDP (growth) 1.2 1.7 2.3 3.2 3.1 Industrial countries 0.8 1.4 1.9 2.9 2.6 UnitedStates 0.3 2.4 2.5 3.5 3.O Japan 0.3 -0.3 0.6 1.6 1.4 Euro Area 1.5 0.8 1.4 2.6 2.6 Developingcountries 2.8 3.1 4.0 4.7 4.8 rurkey's export market growth** 1I 2.4 2.0 4.7 6.7 7.1 EU-15 (52% of market) 1.1 -0.6 2.9 6.3 7.3 CEECs (6%) 1.6 5.4 8.3 8.0 7.7 CIS (9%) 7.5 7.6 8.0 8.3 7.4 MENA (14%) 9.6 4.8 4.9 4.8 5.0 United States (9%) -3.6 4.0 6.9 7.8 7.0 Source: World Bank, EconomicPolicy andProspectsGroup, baseline, `lobalDevelopment Finance-2003 with uwdates. Notes: a? Canada, France, Germany, Italy, Japan, the United Kingdom, and the United States. b/ Inlocal currency, aggregatedusing 1995 GDP weights. c/ Unit value indexof manufacturesexports from G-5 to developingcountries, expressed in U.S.dollars. ** totalexportmarket growth is weightedsum o freal import demand inTurkey's export markets. 215. Growth in the Euro Area, which represents the largest share o f Turkey's export markets, i s expectedto recover slowly. Early indicators for the Euro Area during the first quarter o f 2003 show industrial activity losing momentum. Economic confidence indicators (industrial, retail and consumer confidence) reached a six year low inMarch 2003. Euro Area growth is projected to be 1.4 percent gain in 2003, and i s forecast to stabilize at 2.6 percent during 2004-05. After spiking to just over US$33 per barrel inearly March, 2003, crude oil prices fell to below US$25 per barrel in April 2003, after it became clear that there would be limited disruption to oil suppliesoutside of Iraq as a result ofthe war. Oilprices are forecast to rise slightly in2003 to an 108 average o f US$26 per barrel for the year from just under US$25 per barrel in 2002, and then to decline sharply in 2004 to US$21 per barrel and further to US$20 per barrel in 2005 (World Bank average crude oil price innominal terms). Declining oil prices will reduce the import bill for Turkey and help reduce inflationary pressures from pass-through effects. Capital flows to emerging markets declined in early 2003, reflecting the concern o f foreign investors due to uncertainty inthe lead-up to and during the Iraq war. Sentiment should improve now that major hostilities Iraqi have wound down and assuming related uncertainties are resolved. Given low industrial country yields, underpinned by increasing liquidity and lower risk aversion, the environment for capital flows to emerging markets appears conducive. Total world equity flows to emergingmarkets are forecast to firm moderately from an estimated US$152 billion in 2002 to US$l58 billion in2003.4 216. The crisis inneighboring Iraq and associated uncertainties created additional hurdles for Turkey's recovery. The buildup to war depressed domestic demand and tourism, in addition to dampening investor confidence worldwide. Although Turkey did not come to terms with the United States on a multiyear support package prior to the outbreak of hostilities, the resulting downward pressure on the Lira was limited and subsided fairly quickly, in part due to the fact that the Lira i s considered to be close to purchasing power parity. Subsequently, the apparent resolution o f the war, and the stability in northern Iraq, has mitigated some o f the concerns o f foreign investors. There i s also the likelihood that Turkish exporters will benefit from the reconstruction effort in Iraq. Further improving investor confidence, the IMF's fourth review was completed in April, 2003. Nevertheless, foreign investors will continue to watch for deviations from the IMF supported policies inthe coming reviews. The US Congress approved a US$75 billion war financing package, which includes a US$1 billion grant to Turkey (exchangeable for up to US$8.5 billion in loan guarantees). If disbursed, the grant will provide an additional boost to investor confidence. 217. The global recovery is coinciding with improved domestic conditions in Turkey in the wake o f implementation o f the crisis recovery program. Strengthening global production should support continuation o f Turkey's recovery through its effects on export growth and the follow- through to favorable developments inindustrial production, employment gains, and eventually to a firming o f domestic and foreign investment. Challenges remain as the volatility and uncertainty that have characterized Turkey's economy may work against a full realization o f opportunities in export markets. International financial markets continue to be highly risk averse, particularly inthe wake o f major sovereign default and corporate financial difficulties in the industrial countries, and, albeit apparently easing, political uncertainty in the neighboring Middle East. Overall balance o fpaymentsdevelopments will require careful monitoring. 4 PROSPECTSFOR TURKEY'S EXTERNAL REVENUES, 2001-2005 218. Prospects for Turkey's foreign exchange revenues, encompassing goods, services, transfers and income are encouraging for the period through 2005. Turkey's potential i s much Global Development Finance 2003: Strivingfor Stability in Development Finance, The World Bank, 2003. h~p:llwww.worldbank.org/prospects/gdf2003l 109 greater than has been realized to date, given its positive endowments o f labor, resources and physical location. Turkey has an opportunity to capitalize upon external strengths potentially leading to a strong rebound and sustainable economic expansion thereafter in which exports can play an important role. An indicative review and forecasting exercise for each o f these revenue sources reveals both driving and restraining forces to growth. Turkey's traditional and emerging strengths intrade, combined with emergence o f a more supportive external environment, should yield strong advances inexport earnings over the forecast period (Table 6.4) under a scenario o f continued reform5. 219. Based on improvements in Turkey's export market and continued competitiveness o f its exports, the growth o f goods exports could reach 9 percent per annum over the 2001-05 period. Penetration o f western European markets has been noteworthy, but demand for Turkish export products i s growing more rapidly among the CEEEIS group o f countries (by four-fold over 1996-2000). Efforts to target that market effectively would help sustain rapid Figure 6.5: Turkey's revenues: expansion strongest in growth o f exports. As for many tourism and autos countries, Turkey's exports can be (Millions US$) ................. ITransfenandincom segmented into emerging and IAutosandElecticalem1 Textiles and apparel diminishing sectors6. Among principal 0Tourism Othersetvices export products, the outlook for automotive and parts i s most encouraging (Figure 6.5). The sector i s anticipated to pace export growth over the medium term, driven by the current overcapacity and reorientation o f output from existing plants to exports, together with expectations of aggressive capital 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Source' IMF,Rolmce of Paymen19 SlouiNc~,EPPG dora expansion fueled by increased FDI and export plans o f major multinational firms. Electrical machinery and related goods, largely household (TVs, white goods) and industrial equipment (distribution and circuits) are projected to continue their dynamic export potential, having advanced by 18 percent per year over the 1990s and continuing double-digit gains during 2000-01. Reorientation o f agricultural, especially fmits (citrus) and fruit preparations shipmentstoward markets with increasing demand (CIS and MENA regions) could help to underpingrowth inthese products. This indicative projection exerciseis designedto give an ideaofTurkey's medium-termforeign exchange earnings potential; it is not a forecast. The figures presentedin Table 6.4 are broadly in line with, but not necessarilyequal to, the underlyingprojectionsunder sustainedreformscenariopresentedelsewhere inthe CEM. Among emergingsectors can be included:(i) automobiles and parts, domestic appliancesandtelevisions; (ii)niche marketsintextiles and apparel;(iii)potentialgrowth in non-ferrousmetalminingand export activity; (iv) niche marketsin fruit and preserves; (v) constructionandtransportreceiptstied to the Baku-Ceyhanpipeline; (vi) tourism-including alternativesto "sun and sand" packagetours; and (vii) passengertransport servicesand professionalandtechnical services. Among diminishingsectors can be included: (i)domestic productiono f cotton textile fabrics; (ii)simple apparelexports; and (iii)ferrous metalores and fabricated products; selectedironand steel products. 110 Table 6.4: Turkey: Indicative Projections of Foreign Exchange Revenue 2001-2005 Measured in T i. dollars, shares d growl rates in lercent Willions U.S. dollars 2000 2001E 2002F 2003 2005 Growth Growth Period averages 2000-01 2001-05 Merchandise Exports 31,665 35,255 37,930 40,970 46,690 11.3 7.3 Foodand beverages 3,400 3,970 4,265 4,710 5,570 16.8 8.8 AgriculturalProducts 1,805 2,130 2,315 2,580 3,110 18.0 10.0 Manufactures 23,025 25,950 28,060 30,375 34,630 12.7 7.5 Textiles, apparel& clothing 10,205 10,615 11,068 11,600 12,515 4.0 4.2 Metals, minerals, mtl mfgrs 3,400 3,910 4,145 4,475 5,120 15.0 7.0 Automotiveproducts 1,505 2,185 2,830 3,345 4,170 45.0 17.5 Electricalmachineryhelated 3,495 3,990 4,425 5,000 6,075 14.2 11.0 Other manufactures 4,420 5,250 5,590 5,950 6,750 18.7 6.5 Other goods 5,245 5,340 5,600 5,885 6,490 1.8 5.0 Services Exports 19,485 17,210 19,200 21,785 26,920 -11.7 11.8 Tourism [Travel] 7,636 8,090 9,060 10,350 12,700 6.0 12.0 Tourism arrivals `000 /b 9,586 10,775 11,745 12,870 14,680 12.4 8.0 Transport services 2,955 2,505 2,710 2,985 3,530 -15.2 9.0 Other services 8,895 6,610 7,430 8.450 10,680 -25.6 12.7 Constructionservices 1,032 1,150 1,195 1,260 1,455 11.6 6.0 OtherBusiness 4,580 2,090 2,300 2,730 3,560 -54.3 14.2 Personal.Cultural 2,590 2,420 2,920 3,340 4,375 -6.6 16.0 Income and Transfers 8,200 6,820 6,215 7,230 8,425 -16.8 8.3 Factor incomereceipts 2,880 2,770 1,810 2,385 2,860 -3.8 0.8 Worker remittances 4,560 2,785 2,965 3,320 3,925 -39.0 9.0 Officialtransfers 305 275 420 440 450 -10.8 13.5 Current Acct Revenues 59,345 59,280 63,350 69,985 82,035 -0.1 8.5 Share of GDP 29.8 38.4 ... Memo items: Merchandiseexport volume ($bn) 35.5 40.6 47.0 51.2 57.2 14.2 9.0 Exports o fGNFS (1995$ bn) 54.9 57.7 67.6 75.0 86.2 5.2 10.6 Realrevenues(adj XGS price) 63.7 65.2 75.0 83.6 96.1 2.5 10.2 $bn - ImportPP o f revenues/c 57.5 57.6 67.3 72.3 80.3 0.2 8.6 Population(mn) 65.3 66.2 67.1 68.0 69.7 1.4 1.3 Per-capitarevenues ($) 910 895 945 ,030 1,180 -1.5 7.0 per-capitaimportPP 880 870 1,000 ,065 1,150 -1.2 7.3 Importdeflator (GS,95=100) 103.2 102.9 94.2 96.8 102.2 -0.3 -0.2 Export deflator (GS,95=100) 93.2 90.9 84.5 83.7 85.3 -2.5 -1.6 Terms oftrade (1995) 90.3 88.3 Source: IMF BOPstatistics, UNCOM ADE database, Ban -`IS,Worlc 89.7 86.4 83.5 -2.2 -1.4 3ank, Wc 1Tourism Organization and EPPGestimates. Notes: /a current US.dollars. comDoundaverage growth rates: , . /b estimate oftotal tourism arrivals from World Tourism Organization; " 1 /c total current account revenues deflatedby the GNFS import deflator-a measure of capacity to import. World BankProiections. 220. Tourism carries strong multiplier links across broad sectors o f the domestic economy, and serves as a driving force for transport services and other streams o f revenues, trade-related and administrative. Increased marketing efforts are required in tourism-both to sustain "sun and sand" tourism against competitive pressures, and to diversify attractions by capitalizing further on Turkey's substantial historic, cultural and environmental endowments. This could help sustain a "year-long'' tourism industry vis-a-vis the current five month period o f activity concentrated in coastal provinces. Prospects for professional and technical services exports are 111 bright as these sectors increase in sophistication and demand resumes inthe industrial countries for such services as communications, informatics and related services. Finally, factor income receipts will be restrained by lower returns on holdings of overseas assets, while remittances should re-establish stronger rates of advance over the mediumterm. 221. Turkey's level of productivity in textile apparel and clothing (TAC)-a Figure 6.6: Turkish advantage in productivitylevels in key export sector-stands well above TAC is fading comparators, but that level is eroding. value-added per employee (%'OOO) Growth in output per employee (in USD terms) has faltered since the mid-1990s' I20 while it has risen rapidly in such countries as Mexico and Malaysia (Figure 6.6). Productivity in the TAC "'">---~ 15 I &/'--' sector should be increased through Malaysia. . 10 a d ' Q \/ restructuring and modernization in order 5 - h . . . & ' * - d ....-,.........-...-...-...-...Poland- -.... . . 5 to meet stiff competitive pressures. TAC * . exports to the EUmarket will continue to n face an environment of intense and 1990 91 92 93 94 95 96 97 98 1999 Sourer. L'nirrd Nortons: UNICOlndusry dolabosr. growing competition, increasingly from the Central and Eastern European countries, as well as the countries of the CIS and Asia, especially China. Central Europe has emerged as an alternative for EU production and, since July 2001, Mexico enjoys the same customs-free, quota free access to the EUmarket as Turkey. China, the world's largest TAC exporter (US$50 billion in 2000) has joined the World Trade Organization and will be seeking concessions inthis area. Under the Uruguay Round and multi- fibre Agreement (MFA), all textile quotas are to be liftedby 2005, benefiting to the lar est extent the low-cost producersof Asia (China, India,Pakistan and severalASEANproducers) . -B 222. Insummary, the "base case" outlook for Turkishexports over the 2003-05 period is quite favorable. This assessment is grounded in assumptions regarding the global outlook and continuation of Turkey's economic reform program. Risks in export marketsremain elevated in the near-term (especially the chance that recovery in the EU is somewhat delayed and lackluster); but a return to more evenly balanced global growth appears to be the most likely scenario. Following the extraordinary 12.8 growth of USD exports in 2001-on the strength of sharp currency devaluation-export growth accelerated further in 2002 to 14.1 percent. Projected export revenue growth over the 2001-05 period has the potential to parallel the performance of the secondhalf of the 1990s(7.3 percent annual growth). 'Taymaz (2002). 112 5 EXTERNAL VULNERABILITY AND RISKS External Vulnerability Among these figure the Iraqi invasion of Figure 6.7: A sequence of crises during the 1990s Kuwait in 1990-91 and the East Asian and GDP percent change particularly Russian crises o f the late 1990s. Source Turkey CIS ............... Even the 1994 crisis had a fundamental 10 balance o f payments dimension. Among 5 factors leading to increased external vulnerability have been: (i)a concentration 0 o f merchandise exports in the cotton/textiles/yam sectors, (ii) import high ............... -5 demand including reliance on imported intermediate and capital goods to support ............... -10 export growth across many sectors, (iii) 1990 91 91 93 94 95 96 91 98 99 1000 1001 CurWar Currency crisis EmfAsra/Russia Currency crrsis dependence on imported energy, and (iv) reliance on short-term capital inflows giving Source: World Bank rise to increased susceptibility to changes in-international financial conditions. Turkey also faces important geopolitical risks as highlightedby the external shock in the aftermath o f the terrorist attacks inthe United States on September 11,2001 and the Iraqwar. underlinedthat. Inaddition to Iraq, other geopolitical concerns very relevant to Turkey are developments in Cyprus and the process o f accession to the European Union. The geopolitical factors, while important to note, are not discussed further inthis section. 224. Turkey's highdegree of external vulnerability has exacerbated domestic imbalances and contributed to elevate the volatility o f Turkish output growth to levels twice as high as that for middle-income countries as a group. Inturn, the instability of growth andof the exchange Figure 6.8: Turkey: an "outlier" among rate, inflation and interest rates over the comparators in volatility of REER and GDP decade has made the macro climate for growth [1990-20001 ............................................................ investment difficult, hampering domestic I*+.;;.. T---- + capital spending and constraining foreign E l2................................................................. rn 1" Turkey investment flows to meager levels. Over the 8 ..................................... .......... 1990-2000 period, real exchange rate volatility' for Turkey (cv 10.8) stood some 4-6 points above that o f the East Asians Thailand Chile+ (Indonesia an exception); and volatility in GDP growth well exceeded that for Korea 0.5 0.15 1 1.15 1.s and Malaysia. Figure 6.8 highlights CV: GDP (Ch%) Soumi WoridBonk. JPUogan: valoali~measured by cmflmnr olvvnorion applmd Io gmwrh races o/rhe vanobiis Turkey's "outlying" position when Measured as the coefficient o f variation (sigma/mean) o f the growth rate o f the REER across countries. 113 comparing the volatility o f output growth with that o f the real effective exchange rate over the 1990s'. Other countries having experienced crises, rapid in-and-outflows o f capital and volatile exchange rate paths, including Mexico, Korea and Malaysia, have managed though policy and other methods to restrain volatility of both GDP growth and the real exchange rate to more moderate levels. It may be argued that in these cases, fiscal and monetary policies were more effective instabilizing the economy. For Turkey, it i s likely that sequential crises and recoveries ("boom and bust" phenomenon) tied to pro-cyclical fiscal policy and destabilizing short-term financial flows left the economy without a substantial "grace" period under which stabilizing macro conditions could take root and from which more sustainable growth could be generated. 225. An important factor in Turkey's external vulnerability i s structurally high Figure 6.9: Slackening pace of growth in per- 1 /"I.; import demand. The volume o f imports o f caDita imDort purchasing Dower levels in 1995 LIS dollars, 3-yr moving average[left scale] goods and services covered by Turkey's 1000 ..................................................................... growth in percent, 3-yr moving average[right scale] , foreign exchange earnings has declined by I 20 1.8 percent per year over the period. 800 .....___~___.__.......... ,, ....... ...-........... Figure 6.9 shows that a secular up-trendin growth *..e . ...b...._ ..L.. -5.. ........ .......___ .li. __.__.__. .....___._.- per-capita import purchasing power has 600 . . 10 been punctuated by cyclical and other short-medium term disturbances over the period since 1980. However, import purchasing power appears to have reached I 0 4 I I I I I I I I 1 I I I I I I I J. ,,'I I `\ I I -5 a plateau in 1998 and decelerated since 1980 82 84 86 88 1990 92 94 96 98 2000 underthe crisis. Source WorldBankdata and calculations I 226. One way o f quantifying the vulnerability o f the Turkish economy to more conventional external economic shocks, and how that vulnerability has changed over the past decade, i s presented in Figure 6.10. Inthis exercise, the economy is subjected to five "shocks": Figure 6.10: Turkey's exposure to international risks: (i)a 5 percentage point drop in export well above comparators markets; (ii) a 20 percent fall in the selling 24 ("scenario" effects: measuredas share of baseyear exports, %) Exchange rate price o f the key commodity export (cotton 0Interest rate for Turkey); (iii) a 20 percent rise inthe oil price; (iv) a 20 percent drop in the dollar (which raises the burden o f non-dollar debt); and (v) a 200 bp rise in LIBOR. The impact o f these shocks is quantified for both 1990 and 2000 and expressed in terms o f 1990 exports (left bar) and 2000 1990 2000 I990 ZOOO 1990 ZOOO 1990 2000 exports (right bar). A similar exercise was TURKEY MEXICO THAILAND KOREA also undertaken for three o f Turkey's peers Source World Bank, Economtc Policy andProspects Group - Mexico, Thailand and Korea. It shouldbe pointedout that for efficiencyof presentation,the figure excludesseveralcomparatorcountrieswhere volatility in exchangeratesand/or output growth has beenhigherthan inTurkey; specifically Indonesiaand Russia. 114 227. Four points are worthy o f note from this exercise. First, Turkey's overall external vulnerability has declined since 1990. Should all five shocks occur at once, then the hit to the economy would be equivalent to about 18 percent o f exports, down from 22 percent in 1990. Second, Turkey's external vulnerability remains relatively high by international standards. Mexico, for example, has been able to engineer a far greater improvement. Third, Turkey remains quite sensitive to commodity price shocks. The country's exposure to oil shocks has diminished somewhat over the decade given increased efficiency in energy use and the decoupling o f trade and transport arrangements with Iraq (crude and refined product imports have declined from 20 percent to 13 percent o f total imports over the decade), but it remains relatively high. Moreover, the rising prominence o f textiles inthe export product mix means that the impact of a decline inthe key export price has actually risen over the past 10 years (unlike the other three countries). Finally, the country's financial risks remainrelatively high. Notably, exposure to interest rate changes has risen, with interest-sensitive debt increasingas a proportion o f total international debt from 47 percent in 1990 to 58 percent in 2000. O f course, this i s not all bad in a falling global interest rate environment. However, relatively high exposure to interest-sensitive debt underlines Turkey's exposure to volatile international private debt flows which became particularly evident inthe early months o f 2001. 228. The vulnerability effects outlined above reflect only direct impacts which would be partially offset (or accentuated) by other economic factors, and by policy decisions geared toward addressing adverse consequences. For example, the effects o f rising interest rates would be mitigated somewhat by increased returns to holdings o f foreign reserves. The rise of the euro (or yen) would provide stimulus for home country exports to Europe (Japan) and to third country markets. InTurkey's case, an increase inoil price would be partially offset by improving market conditions among the oil exporting countries (including Russia), increased exports to those markets and potentially higher remittance inflows from workers finding employment inthe Gulf states and other hydrocarbon exporters. As trade and financial integration accelerate over the coming decade, decision-makers will need to become increasingly aware o f the accompanying risks and alternate approaches for managing these risks inorder that they may be better reflected inthe designandimplementationofpolicy. ExternalFinancingRisks 229. Aside from pure interest rates and exchange rate risks, the main external financial risks facing Turkey are the volatility o f capital flows evident across all emerging markets in recent years and the availability o f adequate private finance to cover medium-term external financing needs. Capital flows to emerging markets declined again in 2002 from already very subdued levels in 2001. Flows to Turkey edged up slightly in 2002, to US$7.3 billion, having collapsed from US$22.1 billion in 2000 to US$6.9 billion in 2001, Globally, bank lending to emerging economies has dropped precipitously, reflecting increased risk aversion in the aftermath o f the Argentine default as well retrenchment after losses stemming from Enron, Worldcom and other corporations in financial distress. Interest rate spreads rose through much o f 2002, peaking at very highlevels in September. They have since declined sharply inmost countries althoughthey have remained high and volatile for Turkey, partly reflecting uncertainties surrounding the November 2002 election, then the build up to the Iraq war and, most recently, the status o f the IMF program. Inthe aftermath o f Turkey's financial crisis there are a number o f uncertainties 115 and vulnerabilities regarding private capital market flows that require analysis and pro-active monitoring and evaluation on the part o f the monetary authorities. 230. The Turkish external capital flow cycle of recent years is similar to many others. Developments in Turkey echo developments in Mexico through the mid-1990s and the experience o f Mexico (and other countries that suffered severe boom-bust cycles in the 1990s) helps serve as a base for comparison. As countries struggle to adjust to the post-crisis world, there are three stylized facts that are relevant. First, countries have had to adjust to a lower volume of external financing and with corresponding adjustments to the current account. In macro terms, this requires that import demand growth be restrained and that exchange rate competitiveness (which generally improves sharply through exchange rate adjustments during the crisis) be retained. The appropriate blendo f policies that helps achieve the requiredshift of resources into the tradeable sector i s a tight overall monetary and fiscal stance, but with more fiscal than monetary tightness over time (which helps keep real interest rates lower and therefore the real exchange rate more competitive). Second, countries have had to adjust to a different blend o f financing. The collapse in portfolio and bank flows tends to persist beyond the crisis period in part because the country i s viewed as a higher risk by this category o f investor. This also reflects a conscious decision by domestic borrowers (the government included) to shy away from forms o f borrowing that have proved to be unstable. The key financing adjustment made by countries that have completed a successful turnaround from crisis is a shift towards FDI. Third, exceptional official funding typically disbursed during a crisis must be repaid fairly quickly. Brazil, Mexico and Korea all repaid emergency IMF financing within 2-4 years. They could do this as other sources o f finance were sufficiently strong. Table 6.5 Turkey: External Financing Requirements and Sources, 2000-2006 (in billions ofUS dollars) Actual Estimate Program Projection 2000 2001 2002 2003 2004 2005 2006 Gross financing requirements 25.9 10.5 24.5 29.6 25.5 18.3 18.2 Current account deficit (excl. official transfers) 10.0 -3.2 2.3 7.4 6.2 6.6 5.9 Amortization on debt securities, olw: 1.7 2.1 2.7 4.4 3.1 3.4 3.4 Medium and long-term debt amortization, olw: 13.8 14.3 13.4 16.4 14.1 14.5 14.7 Accumulation of gross reserves 0.4 -2.7 6.2 1.4 2.2 -6.2 -5.9 Available financing 25.9 10.5 24.5 29.6 25.5 18.3 18.2 Foreign direct investment(net) 0.1 2.8 0.9 1.3 0.9 1.1 1.4 Portfolio flows 3.4 -1.7 4.2 5.5 6.6 7.1 7.2 Mediumand long-term debt financing, olw: 18.1 13.2 15.6 16.3 18.6 19.1 20.3 Public sector 1/ 3.4 3.2 2.9 2.6 3.6 3.0 3.1 Private sector 12.8 9.2 11.4 13.1 13.4 14.3 15.2 Deposit money banks 1.9 0.9 1.3 0.6 1.6 1.8 2.c Short-term debt financing (net) 3.6 -12.1 -2.5 4.6 -1.2 -1.1 -1.4 Official transfers 0.2 0.2 0.5 0.5 0.5 0.5 0.t Other 21 -2.8 -2.1 -0.6 -0.1 0.0 0.0 0.c IMF (net) 3.4 10.2 6.4 1.5 0.1 -8.4 -9.t Purchases 3.4 11.3 12.5 2.1 1.9 0.0 0.c Repurchases -0.1 -1.1 -6.1 -0.6 -1.8 -8.4 -9.81 Source: CentralBank, IMF and WB estimates. liGeneralgovemmentandCentralBankofTurkey. 21Errors and omissions. 31 IMF 4th review figures adjustedby the World Bank. 116 231. Turkey faces a serious external financing challenge over the medium term. The challenge i s how to generate enough capital inflows to: (a) underpin the floating exchange rate regime, (b) finance a moderate current account deficit linked to sustained recovery, (c) service sizeable repayments o f IMF loans, and (d) avoid a renewed over-dependence on volatile short- term debt inflows. Increased FDI i s key to successfully meeting this challenge in order to provide adequate capital for private investment and export growth without adding to the already considerable external debt burden. Table 6.5 summarizes the projected medium-term external financing plan under the program which underscores several key points. Regarding the current account balance, the projections assume that the returnto deficits seen in2002-03 will continue, but that the deficits will be modest in line with financing constraints. The financing plan highlights the importance o f FDI as a medium-termsource of funding. An even larger surge in FDI would be desirable to head off the renewed buildup of short-term debt foreseen under the program. "collapse'' (1994-95) and "recovery" (l996-97) phases. ($ billion) 1994 1995 1996 1997 The collapse phase opened wlth .Current account balance -29.7 -1.6 -2.3 -7.4 -7.0 -0.6 -0.7 -1.9 the boom conditions %GDP O f 1994 Equity investments 15.1 10.0 12.0 16.0 which saw the current account Private debt, net 9.1 -14.8 1.6 2.8 deficit reach 7 percent o f GDP, Short-term, net 2.3 -16.2 -10.5 1.o massive inflows of equity Official debt, net 0.8 24.6 -10.6 -6.8 investments and increased short Other items, net -14.2 -7.7 1.9 4.7 term debt. At the same time, the Change inreserves 18.8 -10.6 -2.6 -9.4 again accumulating. 117 6 THEROLEOF FOREIGNDIRECTINVESTMENT 233. Foreign direct investment can be a vital resource to help a country recover from a financial crisis. A review o f experience in countries that went through crisis found that with the Malaysia, net FDI flows remained Figure 6.11: Ratio of FDI Inflows: 1and 2 years after positive through crisis and subseauent crisis relative to 1year prior to crisis ratios of FDI shares to GDP, ratio over 1 indicatespost-cn'sis years, unlike short-term capital flows, 1 FDI inflowswere above pre-crisis inflows such as portfolio investment. In 6 addition, net FDI flows in the year ............................ Crisis + 1 year / pre-crisis year subsequent to the onset o f crisis were .Crisis years /pre-crisis year ............................. - + higher relative to GDP than in the 3 ......................................................... year immediately preceding it. 2....................................... Countries that introduced economic reforms in general, and in particular toward FDI inflows-such as Argentina, Brazil, Mexico, Korea and Thailand-witnessed sustained increases in FDI inflows in the post- crisis period (Figure 6.11). ource: World Bank 234. Recent research suggests that FDI can increase productivity inhost countries and act as a catalyst for domestic investment and technological progress (de Mello, 1999). These effects can help countries sustain higher levels of growth. Borensztein et a1 (1995) show-in a cross- country regression framework with 69 developing countries-that FDI Figure 6.12: Total World FDI flows to Developing contributes more to domestic growth Percentage shares to Turkey and to CA than domestic investment. FDI can 18 200 spur productivity increases in -- domestic firms by stimulatingthem to 150 invest to meet the increased --100 competition associated with the new foreign entry. Borensztein et a1 ........................................ (1995) also find that FDI i s more productive than domestic investment 0 whenthe host country has a minimum threshold stock of human capital. +FDI flows to Turkey, percentageshare of total (left axis) This second qualifying point -FDI flows to ECA, exclTurkey,percentageshare of total (left axis) indicates, that sufficient absorptive &Total world FDIto developing countries,$ bn (rightaxis) capacity (e.g., educational attainment, R&D, infrastructure) is required for Source: Treasurv and World Bank databases FDIto lead to spillovers, suchas technology transfers. On average, FDIinflows represent about 2 percent of GDP in developing countries as a group. And as a share o f fixed investment, FDI represents 2.5 percent duringthe 1990s, up from 1.5 percent duringthe 1980s (Figure 6.12). 235. Survey data from the EBRD across the transition countries indicates that the depth o f reform appears to be strongly correlated with a country's ability to attract FDI. FDIi s especially 118 sensitive to uncertainty as investors have the option to place their resources in other countries. The greater the degree o f uncertainty, the lower the risk i s for an investor to delay gaining market share in the country. As uncertainty i s often driven by government policies, if reforms are pursuedsystematically and commitment is strong, this has a positive influence on the investment environment. Countries with high institutional performance ratings captured the highest cumulative net FDIinflows (dollars per capita). For example the top four rated countries (Czech Republic, Estonia, Hungary, and Poland) received an average o f close to US$1,800 in FDI inflows per capita during 1990-2000 compared to the US$l50 per capita average o f countries Figure 6.13: FDI inflows ranked by EBRD Institutional Performance Indicator from lowst to highest rating, rating w'th country name Turkey n a - Turhnistan 1 0 - Uzbekistan 1.8 - Armnia 1.8 - Russian Federation 1.9 - Azerbaijan 1.9 - Kyrgyz Republic 2.1 - Kazakhstan 2.2 - Romania 2.2 - Macedonia, FYR 2.3 - Lithuania 2.8 - Slovenia 2.8 - Slovak Republic 2.8 - $1,716 Czech Republic 3 2 - Poland 3 3 - Hungary 3 5 - $0 $500 $1,000 $1,500 $2,000 $2,500 Source: World Bank databases and EBRD Transition Report, 2000 which ranked as having initiated the least reforms over the decade (Figure 6.13). 119 for accession, FDI inflows increased to Figure 6.14: Annual FDIInflows to Turkey over US$1.7 billion. In 2001, FDI 3,000 , inflows peaked at US$3 billion, I boosted by substantially lower asset 2,500 prices and a mobile telecommunication 2,000 investment transaction (Telecom 1,500 Italia), before falling sharply again in 1,000 2002 (Figure 6.14). These figures are quite low especially compared with 500 FDI inflows in the realm o f 4 percent 0 o f GDP for advanced transition 1997 1998 1999 2000 2001 2002 237. On paper, Turkey has an attractive foreign investment policy. It has one o f the most liberal investmentregimes inthe OECD as almost all areas open to the Turkish private sector are also open to foreign participation. Turkey also has a very generous incentives regime and free zones." Additionally, Turkey has very positive locational factors given its proximity to the EU, Central and Eastern Europe, the former Soviet Union, the Middle East and North Africa and given its sizeable domestic market of over 67 million people. Turkey i s also an attractive host country for FDIbecause o f its skilled workforce and relatively modest labor costs. The Customs Union with the EU and the prospect o f eventual membership are important factors as well. Overall, the trade regime i s quite open and the infrastructure system i s relatively well developed. Due to expectations of high growth, and given a large population base, the U.S.Department of Commerce has designated Turkey as one o f the ten 'big emerging markets' (BEM) along with China, India, and Brazil. This group o f BEM countries i s expected to offer the greatest commercial opportunities due to high economic growth, the size o f the economy and rapidly growing population.l1 loThe incentives regime includesgrants ofup to 70 percentoftotal fixed investment, exemptions for customs duties,VAT refunds and subsidizedcredits.(www.treasury.gov.tr) 'IThe 10 bigemergingmarket countries are Argentina, Brazil,the Chinese EconomicArea, India, Indonesia, Mexico, Poland, SouthAfrica, South Korea, and Turkey. Together these countriesare expectedto account for over 40 percent oftotal global importsover the next 20 years accordingto the U S InternationalTrade Agency o fthe Departmentof Commerce.(Source: www.ita.doc.gov.) 120 238. Notwithstanding many positive features, a number o f factors have discouraged foreign investors from moving into Turkey. These factors include macroeconomic instability, political uncertainties, excessive bureaucracy, inadequate regulatory framework, the lack of a level playing field exacerbated by uneven enforcement o f tax collection and other rules and regulations, and an opaque and unsettled legal environment. The enabling environment for privatization, an important potential source for FDI, has been bogged down by delays, legal wrangling and pressure from vested interests. A signature example has been the disappointing results o f long-standing efforts to attract large-scale foreign investment in Turkey's energy sector. 239. Turkey has begunto assess the problem and develop a more comprehensive FDIstrategy. Assessments o f the FDIenvironment have beencarried out incollaboration with FIAS including: Diagnostic Review o f the FDI Environment (2000), Administrative Barriers to Investment (2001), Legal Framework for FDI (2002), and Institutional Strategy for FDI Promotion (2002). These assessments were supported by the private sector-most importantly the chamber o f commerce (TOBB), the foreign investors' association (YASED), the Turkish business assocation (TUSIAD) and the Istanbul Stock Exchange (ISE). A structured agenda to improve the investment environment was adopted by the Government in December 2001. This program centers around a Coordination Committee (YOIKK) and nine technical committees to reviewand advise the Government on such areas as FDI regulation, company establishment, sectoral licenses, investmentpromotion and taxes and subsidies. Chaired by a cabinet minister, YOIKK i s responsible to implement and supervise the reform initiatives developed by the specialized technical committees consisting o f representatives from relevant government agencies and the private sector. A new FDI law was enacted in June 2003 which incorporates input from FIAS and YOIKK. The law importantly removes all prior approvals and minimum capital requirements for foreign investments and guarantees free capital flows. A number o f other legislation were amended to improve FDI regime and investment climate. Enacted laws as a direct result o f the YOIKK process until June 2003 include: (i) law on employment o f foreign personnel; (ii) law on the investment incentives which enables a shift to an automatic state aid system intine with the EUrequirements; and (iii) on the company registrationthat simplifies law and streamlines the company registration process. Legislation to establish an independent investmentpromotion agency (IPA) has also beenprepared. 240. However, further action i s needed to improve the environment for FDI in Turkey. Key priorities include progress on macroeconomic stabilization, privatization and regulatory reform. These broad initiatives should be complemented by progress on the specific FDI agenda including: 0 Moving ahead with establishment o f an IPA on the basis o f the following guidelines: P mandate: solely investmentpromotion (no regulatory function). P financing: public sector as the main financing source with strong supplementary financing from business associations and private firms. P governance: Boardo f directors with majority private sector representation. P authority: The Board o f directors to govern the IPA with its president answering to the Board only. The IPA to provide regular financial reports to the Government andbe subject to public audit o fits accounts. 121 0 Activating the Investor Council consisting o f high-level representatives o f the international business community. 0 Initiating a full-scale review o f commercial law in Turkey to identify and address important obstacles to FDIinthe legal framework. 0 Continuing implementation of: (i) the medium-termtax strategy including improved fairness and transparency intax collection and (ii) national anti-corruption strategy. the 0 Developing a strategy to promote "green-field" foreign investmentinthe SME sector. 122 REFERENCES Alesina, Albert0 and Roberto Perotti (1995) "Fiscal Expansions andAdjustments inOECD Countries," Economic Policy ,21 (October) pp 205-48. Arslan, Ismailand Sweder van Wijnbergen (1995) "The Investment and Export Incentive System inTurkey, WorldBank, December. Basle Committee on Banking Supervision (2001) "Consultative Document, Overview ofthe NewBasle Capital Accord," Mimeo, January. Basle Committee on Banking Supervision (1998) "International Convergence of Capital Measurement and Capital Standards," Mimeo, July. Beng, GanWee and Soon Lee Ying (2001) "Credit Crunch during a Currency Crisis: The Malaysian Experience," ASEANEconomic Bulletin, 18(2): pp. 176-92, August. Betcherman, Gordon, Amit Dar, Amy Luinstra, andMakoto Ogawa (2001)"Active Labor Market Programs: Policy Lessons for East Asia," World Bank (available online at www.worldbank.org) Betcherman, Gordon, Amy Luinstra, andMakoto Ogawa (2001) "Labor Market Regulation: International Experience inPromoting Employment and Social Protection", World Bank (available online at www.worldbank.org) Burnside, Craig; Eichenbaum, Martin and Rebelo, Sergio (2001) "Government Guaranteesand Self-Fulfilling Speculative Attacks," manuscript, Northwestern University, September, Forthcoming Journal of Economic Theory. Burnside, Craig; Eichenbaum, Martin andRebelo, Sergio, (2001) "Prospective Deficits andthe Asian Currency Crisis," Journal of Political Economy, 109(6), December, pp. 1155-97. Burnside, Craig; Eichenbaum, MartinandRebelo, Sergio. (2003). "On the Fiscal Implications of Twin Crisis," inMichael P. Dooley andJeffrey A. Frankel, eds. "Managing Currency Crisis inEmerging Markets. Chicago: University of Chicago Press. Brahmbhatt, Milan (2001), "Lessons for Turkey from InternationalExperience with Financial Crisis andDisinflation," manuscript, World Bank. Chand, Sheetal K.,David Robinson and Jose da Silva-Lopes (1990) "Turkey: Reform of Financial Sector Taxation", International Monetary Fund, March . Claessens, Stijn, Simeon Djankov, and Daniela Klingebiel(l999) "Financial Restructuring in East Asia: Halfway There?", World Bank, Financial Sector DiscussionPaperNo. 3, September. Dado, MarinelaE. and Daniela Klingebiel (2002) "Decentralized Creditor LedRestructuring: Cross Country Experience," World Bank, mimeo, April . 123 Dar, Amit and Zafiris Tzannatos (1999) "Active Labor Market Programs: A Review of the Lessons from Evaluations", Social Protection DiscussionPaper Series 9901,World Bank, Washington, DC Demirguq-Kunt, Asli andEnrica Detragiache (2002) "Does Deposit Insurance IncreaseBanking forthcoming. System Stability? An EmpiricalInvestigation" Journal ofMonetary Economics, Demirguq-Kunt, Aslt andHarry Huizinga(2001) "Market Discipline and Deposit Insurance" World Bank mimeo. Demirguq-Kunt, Asli, and Edward Kane (2002) "Deposit InsuranceAround the Globe: Where Does it Work?" Journal of Economic Perspectives, Vol. 16 (2), Summer. Demirguq-Kunt, Aslt, Enrica Detragiache, andPoonam Gupta (2000) "Inside the Crisis: An EmpiricalAssessment of Banking SysteminDistress," International Monetary Fund Working Paper WP/OO/156. Ding, Wei, Ilker Domac, and Giovanni Ferri (1998) "Is There a Credit Crunch inEast Asia?" Asia Pacific Journal of Economics and Business, 2(2): pp.4-32, December. Dollar, David andMary Hallward-Driemeier (2000) "Crisis, Adjustment, andReform in Thailand's Industrial Firms," World Bank ResearchObserver, 15(1): pp.1-22, February. Domac, Ilkerand Giovanni Ferri (1999) "Did the East Asian Crisis Disproportionately Hit Small Businesses inKorea?" Economic Notes, 28(3): pp.403-29, November . Domelund, Dorte and Indermit Gill (2002) "Labor ReforminLatin America During the 1990s" inIndermitGill, Claudio Montenegro andDorte Domeland ( editors) Crafting Labor Policy: Techniques and Lessons from LatinAmerica World Bank, Washington DC: Easterly, William R. (1999), "When is Fiscal Adjustment an Illusion",Economic Policy, 28 (April), pp 57-86 Ebrill,Liam, Victoria Summers et a1(2000) "Turkey: The Agenda for Tax Reform-Next Steps, Fiscal Affairs Department, IMF, September. Ercan, Hakan (1998) "Non-wage labor costs inTurkishManufacturing Industry: An International Comparison", inT. Bulutay (ed) ILO-SIS series on Turkish Labor Markets (No 7),Ankara. Eroglu, Zeynep Ada (2001) "Resource Cost inthe TurkishBanking System," inTurkish, Working Paper, BRSA, Ankara. Erzan, Refik andAlpay Filiztekin (2002) Does Size Matter inGrowth, Productivity, " Employment and Vulnerability inTurkishManufacturing", CEM (2003) Background Paper, World Bank, June. Flanagan, Robert (1999) "Macroeconomic Performance and Collective Bargaining: An International Perspective", Journal of Economic Literature, Vol XXXVII (September) pp. 1150-1175. 124 Foreign Investment Advisory Service (2001), "Turkey: A Diagnostic Study ofthe Foreign Direct Investment Environment", International Finance Corporation and World Bank, February. Foreign Investment Advisory Service (2001), "Turkey: Administrative Barriers to Investment", International Finance Corporation and World Bank, June. Ghosh Swati R. andAtish R. Ghosh (1999), "East Asia inthe Aftermath: Was There a Crunch?", International Monetary Fund Working Paper 99/38. Gunes-Ayata, Ayse (200l), "Political Economy of Economic Reforms inTurkey,", Middle East Technical University, Ankara. International Monetary Fund(2002) Turkey: 2002 Article 1V Consultation Staff Report, Washington, DC (available at imf.org) Kawai, Masahiro, Ira Liberman, and William P. Mako (2000), "Financial Stabilization and Initial Restructuring of East Asian Corporations: Approaches, Results, and Lessons," in Charles Adams, Robert E. Litan, and MichaelPomerleano, eds. "Managing Financial and Corporate Distress: Lessonsfrom Asia," Brookings Institutions Press. Klingebiel, Daniela (2001) "The Use of Asset Management Companies inthe Resolution of Banking Crises: Cross-Country Experiences," World Bank, working paper 2284, February. Krueger, Anne and Aaron Tornell (1999) "The Role of Bank Restructuring inRecovering from Crisis: Mexico 1995-98,'' National Bureau of Economic ResearchWorking Paper 7042. LaPorta, Rafael, Florencio Lopez-de-Silanes,andAndrei Shleifer (2000) "Government Ownership of Banks," Harvad Univ. mimeo, August. Accessible at: http://www.economics harvard.edu/faculty/laporta/laporta. html. . Liviatan, Nissan (2002 Lessons for Turkey from International Experience with Financial Crisis " and Disiflation," CEM (2003) Background Paper, October McDermott, C. John andRobert F. Wescott (1996) "An EmpricalAnalsis of Fiscal Adjustments ",IMFWorkingPaper,WP/96/59.. Nurhan Sural (2001) "Labor Law", inTugrul Ansay andEric Schneider (eds) Introduction to TurkishBusiness Law, (September)pp 120-141, Kluwer Law International, Netherlands. OECD (1999) Employment Outlook, Paris. OECD (1999) Historical Statistics, 1970-1999, Paris. OECD (2000) Labor Market Policies and the Public Employment Service, Paris. OECD (2002) Employment Outlook. Paris. Park, Yung Chul and Jong-Wha Lee (2003), Recovery and Sustainability inEast Asia" in " MichaelP. Dooley and Jeffrey A. Frankel, eds. Managing Currency Crisis inEmerging Markets, Chicago: University of Chicago Press. 125 Pazarbasioglu, Ceyla (1997) "A Credit Crunch? Finlandinthe Aftermath of the BankingCrisis," International Monetary Fund StaffPapers, 44(3): pp. 315-327, September. Polackova Brixi, Hana and Allen Schick. (2002) Government at Risk: Contingent Liabilities " andFiscal Risk," World Bank. Ramey, Garey andValerie Ramey (1995) Cross-Country Evidenceonthe Link Between " Volatility and Growth,", AmericanEconomic Review, 85, 1138-51. Riboud, Michelle, Carolina Sanchez-Paramoand Carlos Silva-Jauregui (2002) "Does Eurosclerosis matter? Institutional Reform andLabor Market PerformanceinCentral and EasternEuropeanCountries" inLabor, Employment and Social Policies inthe EU Enlargement Process, BernardFunck and Lodovico Pizzat (editors), Washington, DC: World Bank. Riordan, Mick, Annette De Kleine, Philip Suttle, Hans Timmer (2002) "Turkey inGlobal Economy 2001-2005", CEM (2003) Background Paper, World Bank, July. Stiglitz, Joseph andAndrew Weiss (1981) "Credit Rationing inMarkets with Imperfect Information," American Economic Review, 71(3): pp. 393-410, June. Taskin, Fatma (2002) "The Automotive IndustryinTurkey : Overview of the Sector and Opportunities for Exports", CEM (2003) Background Paper, World Bank, October. Taylor, John (1998) "Staggered Price and Wage Setting inMacroeconomics", NBER Working Paper 6754, Cambridge, MA (available at nber.org) Taymaz, Erol(2002) "Competitiveness of the Turkish Textile and Clothing Insustries", CEM (2003) Background Paper, World Bank, September. Uygur, Ercan (2002) "The Two-Tier Crisis and Investment inTurkey", CEM (2003) Background Paper, World Bank, October . Woo-Cumings, Meredith(2001) "Miracle as Prologue: The State andthe Reform of the Corporate Sector inKorea," " inJosephE. Stiglitz andShadidYusuf ,eds." Rethinking the EastAsian Miracle, OxfordUniversityPress. World Bank (1994), "Averting the Old Age Crisis: Policies to Protect Old andPromote Growth", World Bank Policy ResearchReport, OxfordUniversityPress. World Bank (1998), "Argentina Financial Sector Review, ReportNo. 17864-AR ", Accessible at: http://www- ds.worldbank.org/servlet/WDSContentServer/WDSP/IB/l999/O6/O3/OOO178830-98 1117 03554833/Rendered/PDF/multiqage.pdf World Bank (2000) Turkey: Economic Reforms, Living Standards and Social Welfare Study, January. World Bank (2000) Turkey: Country Economic Memorandum, Structural Reforms for Sustainable Growth, September,2000. 126 World Bank (2001) Turkey: Public Expenditure and Institutional Review, August, 2001 (available online at www.worldbank.org) World Bank (2001), "Finance for Growth: Policy Choices inaVolatile World ",World Bank Policy ResearchReport. World Bank (2003), "Turkey: Corporate Sector Impact Assessment Report ", March. World Bank (2003), Non-Bank Financial Institutions and Capital Markets inTurkey ",April " World Bank (2003), " Turkey: Poverty and Coping After the Crises " 127 TECHNICALANNEX BANKINGCHAPTER - 3 Credit Crunchin Turkey? 1. Banking and currency crises often coincide with steep reductions in credit to the private sector, and Turkey's recent experience i s no exception (see Figure 1). It i s often unclear whether declines incredit are due to a reduction in demand, brought about by the sharp decline in economic activity and higher lending rates that typically accompany crisis, or to supply side factors. Inthis paper, we employ switching regression to explain the level o freal credit to the private sector inTurkey from 1991to 2001, The underlying intuition i s that in any given period the interest rate might not clear the market, and thus the observed credit level i s determinedby the minimum of the quantity supplied or the quantity demanded. By enabling the researcher to estimate separate equations for demand and for supply, switching regression helps to identify which o f the two regimes i s responsible for generating individual observations. When estimated supply and observed credit are similar, but both lie below estimated demand, we interpret this as evidence of credit crunch. Switching regression enables us not only to identify episodes o f credit crunch, but also to assess which o f the determinants o f the supply equation are the cause. Similarly, when depresseddemand is responsible for low observed credit, the determinants of the demand equation can point to the cause o f the problem. 1. INTRODUCTION 2. A credit crunch can be defined as a reduction in lending by financial intermediaries to an amount that i s less than the quantity that i s desired by borrowers, even if they are willing to pay for the cost o f credit and are able to satisfy the non-price terms o f credit requiredby the lenders. Definedin this way, a credit crunch represents a shift inthe supply o f bank credit. This reduction in supply o f bank loans may be caused by the erosion inthe intermediaries' capital position or deposit base, which we will refer to as a reduction inlendingcapacity. For example, tight monetary policy, by draining the banking system o f its reserves, reduces the amount o f funds available in the interbank market. Unless this reduction i s offset by an increase indeposits or capital, it will lead to a reduction inloans. 3. Banks might also cut back on credit due to greater regulatory scrutiny o f their portfolios and the imposition o f tougher regulatory standards. For example, credit may appear to be reducedwhen authorities remove the NPLs from banks' balance sheets as a result of restructuring operations. Bank lending could also be severely reduced when losses beginto erode the banks' capital to the point where they are unable to comply with regulatory capital adequacy ratios. In order to meet the regulatory requirements banks are forced to reduce their loan assets -which require 100 percent risk weighing. It i s this form o f reductionthat i s generally characterized as a capital crunch. 4. Informational asymmetries that accompany crisis can also induce credit rationing (Stiglitz and Weiss, 1981). Supply o f credit may shift (or bend back) because as borrower's net worth deteriorates following a crisis and collateral values decline, lenders are more inclined to charge a higher premiumon loans to compensate for higher risk and 128 higher cost o f monitoring (financial accelerator viewkredit rationing). Since ease o f monitoring and quality of the borrower i s an important factor here, small, unlisted and more specialized firms are more likely to bear the brunt o f the crunch. 5. International evidence on credit crunch i s mixed. Although there i s some evidence that the tighter rules for financial institutions affected the supply o f loanable funds in Indonesia, Korea, Malaysia, the Philippines, and Thailand (Ding, Domac and Ferri, 1998),there is also evidence that a large part o f the reduction incredit was due to demand factors. For example, Ghosh and Ghosh (1999) find that inKorea and Thailand it was the rising real interest rates and weakening economic activity that lowered credit demand, although in Indonesiathe supply o f credit was the bindingfactor briefly in 1997. Bengand Ying (2001) also show that inMalaysia, it was the decline indemand for credit that led to the decline in loans after the crisis. Survey results confirm the importance o f access to credit although again it i s not clear to what extent the lack o f access to credit reflected the poor prospects o f firms and the state of the financial sector (Dollar and Hallward-Driemeier, 2000). 6. Although all corporations are generally affected by a crisis, those with riskier financing patterns are more likely to be denied credit. Dollar and Hallward-Driemeier provide some support for this argument, showing that firms that did not rely on foreign exchange financing as much faced a less severe credit crunch. The degree o f informational asymmetries - as reflected in the type o f financing - i s likely to be important. For example Domac and Ferri (1999) also show that in Korea, small businesses were particularly vulnerable to shocks that affected smaller regional banks, which normally provided them with most o f their funding. Finally, investigating the aftermath o f banking crises in 36 countries over the 1980-95 period, Demirguc-Kunt, Detragiache and Gupta (2000) show that output recovery generally begins in the second year after the crisis and i s not led by a resumption in bank credit growth. Banks, including the stronger, well-capitalized ones, reallocate their portfolio away from loans following a crisis. Thus, the evidence casts doubt on the hypothesis that resumptiono f lendingis a necessary condition for output recovery. What generally happensis that once the macroeconomic outlook improves, firms are able to substitute other sources o f financing such as suppliers' credit, internal financing, foreign credit lines, equity or bonds. This behavior was observed in Mexico following the 1995 crisis (Krueger and Tornell, 1999). To sum up, while periods o f credit crunch can occur following crises, particularly for smaller enterprises, this episode i s likely to be short. Once the macroeconomic activity resumes, bank lendingalso increases. A DisequilibriumModelof Credit Crunch 7. Absent a credit crunch, or similar credit market imperfections, the interest rate not adjust sufficiently, the demand for credit @ will not equal the supply e`,and the adjusts so that the supply o f credit equals its demand. Ifinterest rates do not adjust, or do actual level o f credit will be: 129 8. Financial crises tend to coincide with steep reductions inthe levels and/or growth rates o f real credit. These reductions could be caused by either demand side factors or supply side factors, or both, which creates an identification problem for the researcher. We follow Pazarbasioglu (1997) and Ghosh and Ghosh (1999) ("GG') who use a switching regression framework which imposes a priori exclusion restrictions to identify separately a supply and a demand equation. For example, GG (1999) calculate a proxy for banks' lending capacity, and use it to identify the supply equation under the assumption that capacity affects supply o f credit but not demand. We employ a similar technique below. 9. Inthe supply equation, we follow GG and use separate proxies to control for the lending capacity o f the banking system versus its willingness to lend (at a given capacity). We assume that the real supply o f credit to the private sector depends, therefore, on the spread between lending and deposit interest rates, on sector lending capacity (L), and on an index of current industrial production (ip): 10. The index o f current industrial production i s a proxy for borrowers' ability to repay their loans which, inturn, serves as our proxy for bankers' willingness to lend. In some specifications, we replace the index o f industrial productivity with an index o f stock market prices (S). We use GG's definition o f lending capacity: total banking sector liabilities plus net worth minus cash in vault minus capital minus required reserves.' More detailed descriptions o f the construction and data sources for each o f the variables are inTable TA 5. 11. Demand for credit should decline when the real interest rate (r) and inflation (inf) rise. We use inflation as a proxy for general macroeconomic instability. Following GG, the demand equation also incorporates a number of proxies for firms' needs to finance current and future output. The index o f industrial production enters the demand equation to control for working capital requirements and as an indicator o f future output. The index o f stock market prices enters as another proxy for future output. Finally, the output gap, measured as the deviation o f current industrial production from its (Hodrick-Prescott filtered) long-run trend, controls for the short-term credit needs o f firms suffering an adverse shock that need help while waiting for economic conditions to improve. Our basic credit demand equation is: 12. We estimate the probability that a given observation comes from the supply (or demand) regime and the associated coefficients for each equation via maximum likelihood. The likelihood function i s derived inthe appendix. When it falls below our definition, GG use the "maximal banking sector assets implied by the prevailing capital adequacy requirement" as their definition of lending capacity. 130 Results 13. Maddala (1989) points out that the likelihood function can become unbounded which makes convergence to a maximum impossible. In our case, collinearity between regressors affects convergence in some specifications. There are strong associations between our proxies for borrowers' current credit needs and their current and future ability to repay their loans. For example, there i s a significant negative correlation (-.41) between the output gap (our proxy for current credit needs) and the index o f industrial production. In addition, there is a strong positive association between industrial production and the stock market index (,71), our proxies for borrowers' ability to repay current and future loans. 14. Our model follows closely those in GG (1999), except that to enable it to converge, we had to replace the index o f industrial production with the stock market index in the supply equation. In model 1 (Table TA 1) the overall fits o f the demand equation (q=. 115, P7.26) and the supply equation (os=.126, t=12.27) are good. The coefficients on the interest rate variables are o f the hypothesized signs, and the negative coefficient for the real lending rate i s significant in the demand equation. As hypothesized, inflation i s negative and significant in the demand equation, while the output gap and the index o f industrial production are both positive and significant. Inthe supply equation, both lending capacity and the interest rate spread are positive, as hypothesized. Only lending capacity i s significant. The stock market index i s not significant ineither equation.2 15. It is somewhat troubling that the spread variable, which theory suggests should be a key determinant o f credit supply, performs poorly in model 1. This appears to be attributable to the deposit interest rate. Although the interest rate spread used in Table TA 1is basedon deposit rates from IFS,we have also tried deposit rates from the Turkish Bankers' Association that yield similar results. In many cases, the adjustments to the deposit and lending interest rate series are asynchronous, resulting in multiple periods with negative spreads that are quite large inmagnitude. 16. We replace the spread variable inthe supply equation with the real lending rate in model 2.3 Once the noise associated with the deposit rates produces i s purged, coefficients for all variables except the stock market index are significant and o f the hypothesized sign. In particular, the interest rate variable i s negative in the demand equation and positive in the supply equation, which should inspire greater confidence in the model. In addition, the fit o f both the supply and the demand equations i s good - Q=.106, (F8.40) ando,=.126, (t=9.77). When the index o f industrialproductionreplacesthe stock market index inthe supply equation(as in GG 1999), we obtain similar qualitative results exceptthat the demand curve has poor fit and the coefficienton the spreadvariable is negative. Moreover, to enablethose modelsto converge, we must replacethe real interestrate variable with the nominal interestrate inthe demand equation. This substitution rendersthe inflationvariable insignificant. GG (1999) also appearsto use lendingratesratherthan spreads intheir supply equations. 131 17. These results are illustrated in Figure TA 2 The overall fit o f the switching regression may be judged by the extent to which the actual real credit level matches the minimum o f estimated demand and estimated supply in each period. For model 2, there are many periods for which actual and predicted credit levels are very similar. There i s evidence o f credit crunch (Le., estimated supply and actual credit lie near one another and below estimated demand) after the 1994 financial crisis, from late 1994 to mid-1998. However, from early 1999 onwards, slack demand appears to be constraining credit. That is, demand better explains observed credit levels, and both lie below estimated supply. 18. Across all models, the most important variable in the supply equation i s lending capacity, over which the government does have some control. Estimated credit supply dips in late 2001, which could be attributable to more stringent capital requirements. Figure TA 3 indicates that the ratios of both capital and required reserves to banking liabilities increased at this time.4 For capital, the jump was especially pronounced. However, the models here suggest that any bank supervisory-induced reductions in lending capacity were not the cause of credit decline in 2001. In that sense, efforts to increase lending capacity by easing capital requirements appear to be unnecessary and could be counter-productive. The current focus should be on improving the investment climate and thus restoring credit demand. Macroeconomic stability and the lower interest rates that accompany it should help increase demand for credit, which currently appears to be the more serious constraint. Creditto Households and Government 19. Although the foregoing models indicate no current supply-side constraints on aggregate credit from deposit money banks to the private sector, it i s highly probable that some parts of the Turkish economy are more affected by slow credit growth than others. For example, credit rationing a la Stiglitz and Weiss i s most likely for prospective borrowers with little net worth such as small businesses and individuals. To assess whether this concern has validity for Turkey, we re-run the switching regression with credit to households (including both consumer credit and credit cards) as our dependent variable. Although this measure does not include direct lending to small businesses, it is likely that many small business owners finance a high share o f their operations through personal loans and credit cards. Inthat sense, household credit should reflect some o f the difficulties that small businesses have had inweathering recent financial crises. 20. The householdcredit model appears inTable TA 4 and the estimated demand and supply are in Figure TA 4. The demand equation is very similar to that for aggregate private credit: inflation and the real lendingrate enter negatively and significantly while the coefficients for the output gap, industrialproduction, and the stock market index are all positive and significant. Inthe supply equation lending capacity i s the only significant variable, which i s also true for some o f the aggregate credit supply equations. Giventhe similarities between the household and aggregate credit models, it i s not surprising that the two types o f models offer similar qualitative results regarding credit crunch. In 4Not only the ratios but also real capital and real required reserves increased. 132 particular, there i s evidence in support o f a household credit crunch from 1994 to 1998. In the last six months of 1999, there is also some evidence of a credit crunch, as the estimated demand lies above the estimated supply equation, However, at the very end o f the period (late 2001), observed credit levels are similar to estimated demand and lie well below estimated supply. Supply does not appear to be constraining current credit to the household sector. 21. It is interesting to note that in percentage terms the mid-90s household credit crunch i s much larger than that for aggregate credit. For example, in models 1 and 2, observed aggregate credit i s 65-95% o f the estimated credit demand for almost all periods from 1994 to 1998. Over the same period, observed household credit i s only 45-70% o f estimated demand, which provides one indication that households and small businesses are more adversely affected by credit crunches than other private sector borrowers. This might provide a rationale for offering incentives to banks to lend to small businesses during credit crunch. 22. As a final empirical exercise, we run a switching regression with real credit to the government as the dependent variable (Table TA 4) to summarize state intervention in credit markets. As in the private credit regressions, inflation and high real interest rates are negatively associated with government demand for credit. However, as might be expected, the proxies for private sector credit needs (output gap, industrial production, and the stock market index) are not positively correlated with government credit demand. Coefficients for all o f those variables are negative, and the one for the stock market index i s significant. Perhaps the most striking feature o f the government credit model i s the poor fit of the supply equation (Figure TA 5). While real credit to the government remained relatively constant duringthis period and thus does not appear to have crowded out private borrowing, it i s interesting that government borrowing has no relation to the banking sector's lending capacity. Conclusions 23. The analysis presented here indicates that, although supply side constraints appear to have been responsible for a mid-1990s credit crunch, the recent credit decline in Turkey is attributable to demand side factors. This leads to (at least) two conclusions. First, since ingeneral we do not see a credit crunch andifthere i s one it is more likely to be there for the smallest firms, there is reason to do less rather than more, and simple transitory rules working on the firm side may be the most helpful approach. These can take different forms: a tax break for corporations; a scheme which gives a firm some government backed financing this year equal to a fraction o f taxes paid last year or two years ago (which i s a crude test o f whether the firms were making profits, and may thus be reasonable firms this year ifit were not for the crisis, and were obeying the tax code); some accounting relief, e.g., on foreign exchange losses; and for the SMEs, an automatic roll-over for very small loans, lower capital adequacy weights on SME loans, or some trade finance scheme may be helpful. Indeed, some o f these measures, particularly roll- overs and credit guarantees for SMEs were successfully applied inKorea. 133 24. Legal changes could also improve the financial contracting environment for firms and thus facilitate better access to credit. Most SMEs get the bulk of their external finance not from the banks, but from trade credit. This i s also true inTurkey. However, receivables financing i s not legally allowed in Turkey. Although it i s possible to get credit against post dated checks, it i s not possible to do this using invoices or accounts receivable. Such a change in the law to allow re-assignment o f intangibles would go a long way inrelaxing short term financing constraints o f SMEs. 25. Second, since demand side factors appear to be primarily responsible for the recent low level o f private credit, taking measures on the banking side may be counter- productive. In particular, current proposals to effectively relax capital adequacy requirements by permitting substantial upwards revaluation o f bank's real estate and other equity holdings i s unwarranted. Although the models presentedhere indicate that stricter capital requirements did coincide with reductions inthe estimated supply of credit in 2001, supply was not the binding constraint on observed credit levels. Therefore, relaxing capital requirementsto improve credit supply i s bothunnecessary and unwise, as it will lead to a more fragile banking sector over time. Moreover, resumption of bank credit is generally not a lead indicator; to the contrary, it follows output recovery. Requiringthe banks to lendtoo muchtoo soon may backfire and lead to non-performing loans. Table TA 1: Switching RegressionResults Model 1 Model 2 Coeff. t-statistic Coeff. t-statistic Credit Demand Constant 15.091 8.40 19.447 22.08 Real Lending Rate -.480 3.35 -.340 2.77 Inflation -5.939 3.03 -2.840 1.70 Output Gap .126 2.96 ,067 2.05 IndustrialProduction 3.488 6.37 2.222 9.23 Stock Market Index -.114 1.17 .045 0.91 OD .I15 7.26 ,106 8.40 ~ CreditSupply Constant 6.122 2.59 -3.842 0.80 Lending Capacity ,754 7.84 1.109 6.27 Stock Market Index .095 1.24 .002 0.02 Spread .065 0.52 Real LendingRate .I54 2.13 0 s .126 12.27 .126 9.77 Log Likelihood 88.57 91.94 Observations 114 114 134 Table TA 2: Summary Statistics Variable Mean Median High Low RealCredit (Trillions of 1995 Lira) 13.1 13.2 20.8 6.9 RealLendingRate 3.8% 6.1% 53.6% -200.2% NominalLendingRate 59.1% 61.5% 76.1% 37.2% Interest Rate Spread -14.6% -14.5% 23.6% -80.3% DepositRate 73.7% 76.6% 131.8% 33.8% Inflation 4.6% 4.4% 22.1% 0.5% (monthly) Output Gap -0.03 -0.3 1 19.78 -20.63 Index of IndustrialProduction 106.2 105.9 130.6 82.3 Stock MarketIndex 19686 15129 60207 4250 (Deflated by WPI) LendingCapacity 18,600 4,870 101,000 187 (trillions o f Lira) BankingSector Assets 23,600 6,320 118,000 248 (trillions o f Lira) BankingSector Liabilities 19,300 4,870 106,000 226 (trillions o f Lira) BankingSector Capital 3,550 928 26,200 36 (trillions ofLira) BankingSector Required 1,280 452 6,420 20 Reserves (trillions of Lira) BankingSector Cash inVault 180 68 758 4 (trillions of Lira) 135 Table TA 3: Correlations real output real nominal spread inflation industr stock lending credit gap lending lending produc- market capacity rate rate tion index real credit 1.oo output gap -.17* 1.oo real ,18* .ll 1.oo lending rate iominal -.os .25*** ,17* 1.oo lending *ate spread .08 -.lo .45*** .25*** 1.oo inflation -.19** -.02 -.94*** 18. -.36*** 1.oo industrial .84*** -.41*** .10 .09 18* -.07 1.oo production stock .79*** -.11 .17* -.04 .24**' -.18' .71*** 1.oo market index lending .89*** -.11 *11 -.17. -02 -.17* .68*** .83*** 1.oo capacity 136 Table TA 4: Credit to Householdsand Government Households Government Coeff. t-statistic Coeff. t-statistic Credit Demand Constant -7.723 3.99 13.055 5.93 Real Lending Rate -1.861 9.19 -.018 3.80 Inflation -19.811 6.89 -23.173 3.98 Output Gap ,077 1.90 -.054 0.66 Industrial Production 2.838 6.02 -.728 1.27 Stock Market Index ,292 2.48 -.220 1.99 .115 4.14 .330 8.49 Credit Supply Constant -47.335 7.69 -6.956 6.10 Lending Capacity 1.818 7.42 .786 13.51 Stock Market Index -.146 0.93 -1.017 16.59 Real Lending Rate .013 0.09 -.022 11.98 .239 8.50 ,016 0.08 Log Likelihood 26.4 -9.5 Observations 114 114 137 a Sources Symbol DeJinition Source Qt RealCreditto the PrivateSector.Nominal International Financial Credit deflatedby WPI; logarithm Statistics, line22 rt Reallendingrate; nominalmedium lending TurkishBankers' Association rate, adjustedfor annual inflation It NominalLendingRate; mediumterm Turkish Bankers' Association lendingrate inf, Consumer price inflation; Aln(CP1) International Financial Statistics, line 64 gapt Output gap; deviationof current index of Index of industrialproduction, industrial productionfrom its Hodrick- International Financial Prescottfilteredlong-runtrend; logarithm Statistics, line 66 ipt Indexof industrialproduction; logarithmof International Financial three-monthbackwardmovingaverage Statistics, line 66 Stockmarketprice index; deflatedby WPI; IsmailArslan, World Bank logarithm Office, Ankara Lt Reallendingcapacity;total commercialbank Liabilities:ZFS lines24-26, 45-47 capital- centralbankreserves;deflatedby liabilities+ net worth - cashinvault - WPI; logarithm Assets :IFS lines20-22,40-42 Cash inVault and Reserves: CentralBank o f Turkey Capital: IMF Country Desk for Turkey spreadt Nominalmediumterm lendingrate minus Lendingrate:Turkish nominaldeposit interestrate Bankers' Association; DepositRate:IFS In60L Dataare monthlyfrom December, 1991to December,2001. 138 Figure TA 1: RealCreditto the Private Sector 25 Figure TA 2: Credit Demand and Supply (Model 2) 45 40111-est. demand I 1 35 +est. supply n 30 25 20 15 10 5 139 Figure TA 3: Capital and Reserves at Central Bank 30% 20% 15% 10% 5% Figure TA 4: Creditto Households and SMEs '1 140 FigureTA 5: Creditto Government -A- real credit 141 Appendix. LikelihoodFunction The probabilitythat the observedquantity of creditat time t is determinedby the supply equation is: osandodare the estimatedstandarderrors ofthe credit supply and credit demandfimctions and a(*)is the cumulativenormaldistribution. Ifobservationt comesfromthe demandfunction, then Q, = QP < Q; andthe density of Q,,h(Q,),is givenby: where 4(.) i s thejoint density functionofQP and Q," Similarly, ifthe observedcredit levelis . determinedby the supply equation, then Q, =Q;