Romania - Fiscal policy in the transition (Vol. 1 of 2)

Report No. 11878-RO Romania Fiscal Policy in the Transition November 8, 1993 Country Operations Division Country Department I Europe and Central Asia Region FOR OFFICIAL USE ONLY Document of the World Bank This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  CURRENCY EOUIVALENTS Currency unit = Leu (plural Lei) 1986 1987 1988 1989 1990 1991 1992 Lei per US$ = Commercial 16.15 14.56 14.28 14.92 22.43 76.39 307.95 Non-commercial 11.34 9.76 8.75 8.99 - - - Romanian Fiscal Year January 1 through December 31 ABBREVIATIONS CEC - State Savings Bank CPI - Consumer Price Index CPS - Consolidated Public Sector EBF - Extrabudgetary Fund FEX - Foreign Exchange GC - Global Compensation GDP - Gross Domestic Product GIEAs - Gross Interenterprise Arrears GRF - Gold Revaluation Fund IEAs - Interenterprise Arrears MOF - Ministry of Finance NBR - National Bank of Romania NlEAs - Net Interenterprise Arrears NPL - Nonperforming Loans PPI - Producer Price Index RBFT - Romanian Bank for Foreign Trade RCB - Romanian Commercial Bank RF - Restructuring Fund RWA - Risk-Weighted Assets SOEs - State-owned Enterprises SRF - Stock Revaluation Fund VAT - Value Added Tax  FOR OFFICIAL USE ONLY TABLE OF CONTENTS Page No. EXECUTrIVESUMMARY . .......................................... i-x A. What is the Current Fiscal Policy Stance in Romania? .................. i B. Requirements for a Sustainable Fiscal Policy ....................... iv 1. Restraining Consumer Subsidies ............................ iv 2. Stemming Losses in the Enterprise and Banking Sectors .... ......... v 3. The Need for Tax Reform .............................. viii C. Conclusions ....................... x CHAPTER 1: DITRODUCTION ........... ............................ 1 CHAPTER2: TOWARDS A SUSTAINABLE FSCAL POLICY ................ . . 3 I. Introduction ..3 H. Revisiting the General Government Deficit ............... . 4 m. The Real Deficit including the Central Bank's Quasi-Fiscal Activities . ..................................... 6 A. The Nominal CPS Deficit .................. 7 B. The Real CPS Deficit . ................................... 10 IV. Fiscal Policy and Macroeconomic Consistency: The Financeable Deficit Over the Medium Term .................... 13 V. Implications for a Sustainable Fiscd Policy ......... .. .............. 15 CHAPTER 3: THE ENTERPRSE DIMENSION ........................... 17 I. Overview .............................................. 17 H. Enterprise Performance ..................................... 18 m. Fiscal and Quasi-Fiscal Implications of Interenterprise Arrears . . . 20 A. Interenterprise Arrears: Some Basic Facts ..21 B. Causes of Arrears ..24 1. Inefficiency of the Payment System .24 2. Lack of Adjustment by Firms .24 3. Game-Playing and Moral Hazard .24 4. Liquidity Problems and Inconsistencies in Macroeconomic Policies .24 5. Indications from an Enterprise Survey .26 C. Interenterprise Arrears as a Government Contingent Liability . . . .28 D. Interenterprise Arrears and Tax Revenues ..29 IV. Firms in Financial Difficulties ....... ................ 31 V. Recommendations ....................... 33 This report is based on the work of a mission that visited Bucharest, Romania between September 1-18, 1992 and a survey of enterprises carried out in November 1992. The following contributed to the report: Ritu Anand (Task Manager), Lawrence Bouton, Fabrizio Coricelli, David Sewell, Cornel Tarhoaca, Alfredo Thome and Sweder van Wijnbergen (Consultant). Faith Smith and Grace Sorensen are responsible for the production of the report. Peer reviewers were Wafik Grais and Homi Kharas. The report was discussed with the Romanian authorities in September 1993. The analysis is based on data for 1992. Major changes that have occurred since then have been reflected in the report. This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  CHAPTER 4: FISCAL IMPLICATIONS OF BANK RESTRUCTUR1NG . ........... 35 I. Introduction . . . 35 II. The Fiscal Cost of Bank Restructuring ....... ..................... 36 A. Identified Nonperforming Loans ............................. 36 B. The Government Schemes for Recapitalizing Banks and the Global Compensation Scheme .... ............ 38 C. Fiscal Cost of Alternative Bank Recapitalization Schemes ..... ........ 39 Im. Costs of Postponing Bank Restructuring ....... .................... 41 A. Increase in Nonperforming Loans ............. ............... 42 1. Evidence from Time Series ....... ...................... 42 2. Evidence from Individual Bank Audits ...................... 43 B. Interest Rate Spreads and Bank Real Net Worth . .46 IV. Pre-conditions for a Successful Bank Recapitalization . . .47 CHAPTER 5: THE TAX SYSTEM IN THE TRANSITION ..................... 49 I. Introduction ............................................ 49 H. Taxation of Personal Income ........ ........................... 51 A. Description .......................................... 51 B. Taxation of Persons: Comments and Recommendations ..... ......... 52 m. Corporate Taxes .......................................... 54 A. Description . ..................................... 54 1. Taxation of Domestic Firms ............................ 54 2. Treatment of Foreign Investment ......................... 57 B. Corporate Taxation: An Evaluation .......... .. ............... 57 C. Corporate Taxation: Recommendations ........ .. .............. 62 IV. Indirect Taxes ........................................... 66 A. Description .......................................... 66 B. Recommendations .................. .................... 67 V. Conclusions . . .68  List of Text Tables and Figures Table/Figure No. Page No. Chapter 2: TOWARDS A SUSTAINABLE FISCAL POLICY ...... ................. 3 Table 2.1 Consolidated General Govermment Balance ...... ...................... 5 2.2 Nominal Public Sector Deficit and Its Financing, 1991-1992 ..... ............. 7 2.3 Real Public Sector Deficit and its Financing, 1992 ...... ................. 10 Box 2.1 The Operational Deficit: Conceptual and Measurement Issues. 8 2.2 Corrections to the Reported Government Deficit in 1992 On account of Off-Budget Expenditures and Inflation Correction .12 Fig. 2.1 Real Interest Rate on Refinancing Credit .11 Chapter 3: THE ENTERPRISE DIMENSION ...................... 17 Table 3.1 Sources of Repayment of GC Credit by the 100 Firms in Arrears to Banks at the Beginning of 1992 .20 3.2 Real Wages .25 3.3 Arrears and the Real Burden of Taxation .30 Box 3.1 The Global Compensation Scheme .20 3.2 Interenterprise Arrears and Output .21 3.3 Interenterprise Arrears and Inflation .23 3.4 Net Arrears and the Chain of Arrears .29 Fig. 3.1 Real Bank Credit and Interfirm Arrears .22 3.2 Real Deposits of Enterprises .25 Chapter 4: FISCAL IMPLICATIONS OF BANK RESTRUCTURING .......... ...... 35 Table 4.1 Estimated Losses of Banks . ..................................... 36 4.2 Cash-Flow Effects of Alternative Bank Recapitalization Schemes .... .......... 41 4.3 Comparison of NBR's and Auditors' Estimates of Nonperforming Loans .... ..... 42 4.4 Breakdown of Total Loans and Deposits by Main Category ................. 44 4.5 Access of Bad Debtors to Global Compensation Financing .................. 45 4.6 Average Nominal and Real Lending Rates ........................... 46 4.7 Breakdown of Bank Spread by Sources of Costs ......................... 47 Fig. 4.1 Total Bank Overdue and Compensation Loans. .43 4.2 Enterprises' Bank Deposits and Loans .45 4.3 Breakdown of Bank Nominal Spread .47 Chapter5: THETAXSYSTEMINTHIETRANSITION ......... ................ 49 Table 5.1 Consolidated General Government Revenues, 1985-1992 ...... ............. 50 5.2 Revenues from Enterprise Profits .................................. 55 5.3 Marginal Effective Tax Rates .................................... 57 5.4 Financing of the Corporate Tax System ........... ................... 59 5.5 Revenue Gain after Indexing the Corporate Tax ....... .................. 63 Box 5.1 Marginal Effective Tax Rates .58   ROMANIA FISCAL POIUCY IN THE TRANSMON EXECUTIVE SUIMMARY i. The macroeconomic picture in Romania is puzzling at first sight. As in other countries in Eastern Europe, output has fallen (by over a third since 1989) and inflation has risen to about 200 percent per year for the last two years. However, contrary to the other high-inflation countries which run large fiscal deficits, Romania reports almost balanced government accounts. Clearly, the macroeconomic priorities for Romania are to reduce inflation and restore sustainable economic growth. But with such apparently small fiscal deficits, what can the government do? This question is the starting point of the report. It analyzes first whether Romania's fiscal policy stance has been as restrained as the reported figure seems to suggest. The report shows that a more careful analysis yields dramatically different results. It then estimates what a sustainable deficit is for Romania. Finally, the report indicates that reform in three key policy areas is crucial to fiscal balances: consumer subsidies," enterprise and financial sector reform and the tax system. Without major reform in these three areas macroeconomic stability will remain out of reach. A. What is the Current Fiscal Policy Stance in Romania? ii. The official budget position in 1992, with an overall deficit at about one percent of GDP,' looked remarkably good, but this provides a misleading picture of the fiscal stance. First, certain public expenditure items were not included in the budget, although fiscal resources will need to be found for them. Second, inflation not only distorts enterprise accounts but also government statistics. In Romania, where the public sector broadly defined is in the unusual position of having higher domestic currency denominated claims than debts, the net impact of correcting for the effect of inflation is to reveal a substantially larger deficit. Third, important revenue items that are now performing well, are likely to yield less in the near future. In particular, the enterprise sector is likely to yield less tax revenues and demand more transfers as time goes by. We discuss each issue in turn. Off-budget Items iii. A relatively large amount of expenditures on investments, subsidies and transfers is extrabudgetary, and some of these are not financed by current revenues. Prominent among these are the following three: (i) exchange rate subsidies to enterprises which were being financed by monetizing unrealized capital gains on gold; (ii) infrastructure investments and other expenditures if After this report was drafted, most consumer subsidies were abolished, although some were replaced by price controls. This, in effect, maintains the subsidies but shifts the costs to producers. If producers, in turn, get their losses covered by the government or the NBR, the government still pays the subsidies but in a less transparent way. 2/ This deficit is a consolidated position of the central administration budget, local governments, and the extrabudgetary funds presented to Parliament either as annexes to the government budget or, separately, for approval of expenditures.  - ii - fnanced from foreign loans. Foreign loans will eventually need to be repaid and thus should be considered as a means of financing a deficit rather than a source of income reducing the deficit. But the most important item outside the budget concerns (iii) the quasi-fiscal activities of the National Bank. In Romania, as in many other countries, it is common for the central bank to provide low- interest or subsidized loans to the enterprise sector through the banking system. This subsidy, as any other subsidy, should be registered as a fiscal expense. iv. Raising interest rates on such credit would stem the erosion of the NBR's real net worth,3' but this step would not solve the underlying problems that require subsidization; it would actually aggravate the financial situation of enterprises unless they adjust by cutting wages and other expenditures or increasing efficiency. Indeed, many banks hold a large proportion of non-performing loans and preferential credit, the costs of which are shifted to good borrowers via high intermediation charges. In such a context, raising the refinance rate could lead to unduly high lending rates or a repression in deposit rates. The latter is what happened in Romania when the NBR raised the refinance rate in mid-1992 (see Section B.2). This suggests that in the current circumstances merely increasing the refinance rate would not be sufficient; fundamental reform in the enterprise sector is calledfor iffiscal balance is to be restored. Correcting the Government Accounts for Inflation-Induced Distortions v. The nominal deficit, during periods of inflation, does not capture the public sector's claim on real resources. This is because, during inflation, a portion of the nominal interest payment is a compensation for the erosion in the principal that inflation brings about and therefore represents amortization, not a current expense. If the government is a net creditor to the rest of the economy, as is the case in Romania, the government will see the real value of its claims eroded unless it raises nominal interest rates with inflation to offset the erosion in the principal. Revenues are Likely to Decline in the Future vi. The government's budget performance depends heavily on state enterprise profitability which, surprisingly, in a period of declining output and labor productivity and increasing inventories, has been maintained. However, to a large extent these are 'paper' profits: enterprise profitability is artificially boosted by inadequate inflation accounting (historical cost accounting, inadequate depreciation allowances) and the large proportion of unserviced debts.' This is clearly not sustainable; the experience of other East European countries indicates that unreformed state enterprises within a year or two become a cash drain rather than a source of public revenue. vii. This process may already have started: government liabilities on account of unserviced debts of enterprises to banks (non-performing loans) are large and growing. A preliminary estimate 3/ But note that it would also reduce corporate tax revenues; interest payments are deductible under corporate tax so the net revenue effect is about half of the gross (the corporate tax rate is 45 %). v Inflation also reduces the tax burden through nominal interest deductibility; however, this is only effective to the extent interest is actually paid.  - iii - of the bad debts of the four major state-owned banks amounts to 7.5 percent of GDP.5' Moreover, the problem could snowball out of control if the government merely recapitalizes the banks, as has been the case in the past, without acting to fundamentally change the incentives of banks and enterprises so as to prevent further losses. viii. Besides their arrears to banks, enterprises have substantial arrears in payments to other supplier enterprises. While interenterprise arrears in principle should not have any quasi-fiscal implications, being claims of one firm against another, they do become a govermnent liability if the net debts of individual enterprises are implicitly or explicitly guaranteed by the government. This can be seen as a manifestation of continuing soft budget constraints of enterprises and implies that the government is the ultimate holder of the net debt toward the creditor enterprises. When it steps in to clear the net debts, as it did for instance in the global compensation scheme in December 1991, these debts become a quasi-fiscal expenditure. ix. The above factors taken together suggest that the general government accounts are fragile and that fiscal problems have been shifted from the budget to enterprises and banks. The first correction, that is including off-budget expenditures financed from foreign loans (2.6% of GDP) and monetizing unrealized capital gains on gold (4.2% of GDP), results in a nominal deficit of almost 8 percent of GDP. Considering the other two important corrections (incorporating the quasi-fiscal activities of the NBR and eliminating inflation distortions from the public sector accounts) leads to a significantly higher comprehensive (real or 'operational') deficit of more than 21 percent of GDP. The largest factor explaining this difference is related to real interest subsidies to the non-government sector financed by the NBR. In effect, the NBR has been transferring resources to enterprises through the provision of refinancing credit to banks at negative real rates of interest. This situation is equivalent to one in which interest subsidies (or any other subsidies or transfers) are included in the government budget but general government revenues are not sufficient to finance them and the resulting deficit is financed by the NBR. In both cases, the subsidies, whether they are budgetary or extrabudgetary, are financed by monetary expansion. x. Moreover, as economic restructuring advances, the budget will be under further pressure. Revenues are likely to shrink with the contraction of the state enterprise sector (on which the government budget so heavily depends), while the growth of the private sector cannot initially replace the state enterprises in the tax base, as most of the new enterprises are either tax-exempt or in the informal sector and difficult to capture in the tax nets. Moreover, the transition to a new tax system and the needed adaptation and strengthening of tax administration implies that lower-than- potential revenues can be expected for the next few years. xi. The weakening of the government's revenue base comes at a time of rising expenditure demands to meet the potentially heavy costs of enterprise and bank restructuring, as well as the requirements for social protection and social assistance in the wake of increasing unemployment. The government will have to consider important trade-offs between the rising .' While this estimate includes loans granted before 1991 which the banks have assessed as bad (Law 7), as well as non-performing loans classified by external auditors at end-1991, and overdue global compensation loans in 1992, it does not include potential bad debts as a result of fresh bank credit extended to enterprises with non- performing loans.  - iv - demands for a social safety net, a greater role for investment in infrastructure and human capital, and the need to restore balance between revenues and expenditures. B. Requirements for a Sustainable Flscal Policy xii. To assess the extent of adjustment required, the report derives a "financeable" deficit as a target for macroeconomic policy. The financeable deficit is defined as that level of deficit that does not require more financing than is compatible with sustainable external and domestic borrowing and the government's targets for inflation and output growth. xiii. Assuming an average inflation target of 35 percent and real GDP growth of 2 percent p.a. over the next five years, a real deficit of 4-5% of GDP over the medium term would be financeable (which translates for the more conventional nominal deficit in a target of 2-3%"). A tighter inflation target or lower growth would call for a lower deficit and therefore more restrictive policies. This suggests that, if the real deficit in 1992 is taken as a benchmark of the actual fiscal stance, a substantial reduction in the fiscal deficit is imperative. This deficit level depends at the same time on the availability of foreign financing. Should the latter not be forthcoming, the deficit would have to be reduced correspondingly. xiv. The gap between rising expenditure commitments and lagging revenues is the main threat to stability. Meeting this challenge requires a transition strategy involving expenditure cuts, revenue measures and the cautious use of external finance, while reducing the reliance on monetization as a source of public revenue. With the move to a market economy and private sector ownership the government will inevitably face a substantial decline in revenue. International comparisons for countries at a similar stage of development (outside eastern europe) indicate that government revenues are typically much lower as a proportion of GDP than is the case in Romania. Given the limits on sustainable deficits, this also means that the size of government will have to be cut back substantially, also in line with the move to a market economy. This report, rather than analyzing all government budgetary expenditures, indicates three areas where reforms are most urgently needed: (i) restraining consumer subsidies; (ii) stemming losses in the banking and enterprise sector; and (iii) reforming the tax system. B.1 Restraining Consumer Subsidies xv. At a staggering 9% of GDP in 1992, consumer subsidies are an important area of potential cost savings.' While by September 1992, subsidies on some food products were completely eliminated, there was an increase in subsidies for a number of items (notably energy, fuels, and public transportation which accounted for more than half of total budgetary subsidies) for which 6/ This conversion assumes that the target real deficit is maintained throughout the period and that the inflation and output targets are achieved. "/ This refers to the budgetary costs and does not include the extrabudgetary exchange rate subsidies to enterprises compensating them for the controlled prices on their output, many of which are subsidized products to households.  intermediate price increases were not passed on to the consumers. And food subsidies still accounted for no less than 3 percent of GDP in 1992. There is no doubt that a part of these subsidies reaches those intended. But a large share will unavoidably go to those not in direct need of such assistance. Such mis-targeting ultimately threatens the assistance to those most in need as skyrocketing costs make wholesale cutbacks necessary. On May 1, 1993, budgetary consumer subsidies were eliminated. However, they were replaced in some cases by price controls. This, in effect, will pass the cost of subsidies on to producers and more likely entail a reduction in the supply of those items and/or require off-budget transfers from the government or the NBR to the producers to cover their losses. In effect, then, the government would still be paying the subsidies, but in a less transparent way. A better way to protect the poor and contain budgetary costs is to target subsidies more sharply. An example would be to replace blanket subsidies reaching poor and rich alike by ration shops transferring infra-marginal quantities to low income families only. B.2 Stemming Losses in the Enterprise and Banking Sectors xvi. The analysis in this report indicates that the underlying fiscal problems are much more serious than what is signalled by the general government budget deficit as measured by the Romanian authorities. This underlying fiscal pressure is ultimately connected with the difficulties and problems affecting the enterprise and banking sector. We consider the enterprise and bank problems in turn, although ultimately of course problems in banks stem from those in their borrowers, the enterprises. Enterprse Arrears: A Source of Concern? xvii. The financing of enterprise losses could come to represent a large source of quasi- fiscal expenditures. However, in the present Romanian economy, losses are hard to measure. Accounting losses, in an inflationary environment, may significantly underestimate the economic losses. For instance, valuing inventories at replacement cost rather than historical cost would reduce accounting profits of industrial firms by about 30 percent (see Chapter 3). Inadequate depreciation allowances also artificially boost enterprise profitability. Finally, in a situation of undefined property and control structure of enterprises, firms do not have an incentive to maximize profits. As a result, profit/losses are not a good indicator of enterprise performance. xviii. Nevertheless, the performance and behavior of enterprises have important quasi-fiscal implications through their effects on government and central bank liabilities. This is apparent in the accumulation by enterprises of bad debts, both with banks and with other enterprises. State enterprise losses can be financed by banks and other enterprises, losses that may be monetized in the future. Interenterprise arrears become a fiscal problem when creditors maintain their claims while debtors are partially or completely bailed out by the government. This may have happened recently in Romania under the "global compensation scheme." Under this program, bank credit was extended to cover the net arrears of enterprises to other enterprises. This provided an opportunity to identify the unrecoverable amount of net arrears and the debtor enterprises that would have been unable to repay the global compensation credit had they not had access to fresh financing from banks or other enterprises. However, since these enterprises were able to borrow again, their bad loans would be buried in the banks' or other enterprises' balance sheets again. xix. Almost all Romanian state enterprises have arrears. This results in a loss of firm-  - vi - level information as the balance sheets of firms become interlinked. It also reduces firm responsiveness to prices and forms a barrier to adjustment. The problem of finns bas also been exacerbated by credit policy, excessive increase of real producer wages and a combination of exchange rate and price policy which creates large subsidies for some firms and large implicit taxes for other firms. In their present form, arrears are a very inefficient formn of financing of output. However, enterprises in net arrears are not necessarily doing worse than the net creditor enterprises. xx. The report shows that the problem of firms in distress is highly concentrated both at the sectoral and enterprise level. In fact, more than interenterprise arrears, arrears with banks seems to be a better indicator of loss-making firms, with a large proportion of unsold production. In particular, we identified 100 firms which account for most of the arrears with banks and also absorb a significant part of net interenterprise arrears. Indeed, these firms, by absorbing almost fifty percent of total bank credit crowd-out funds for other firms, thus contributing to the accumulation of interenterprise arrears. xxi. The size of the arrears and their dynamics may certainly be affected by incentives for collusive behavior among enterprises. Arrears are probably at least partially related to fraudulent behavior by enterprises. The corporate tax system, through its asymmetric treatment of costs and revenues (the former can be counted as soon as the liability is incurred, while the latter only need to be reported when actually received) opens the possibility of tax avoidance through mutually agreed upon interenterprise arrears. At least as important, are the incentives linked to expectations of a bail- out by the government. Since the government has already stepped in-before and after reforms-to clean up interenterprise arrears, the expectation of a bail-out is widespread in the economy, as was confirmed by our survey of enterprises. Moreover, the same firms which obtained net global compensation credit account for a large proportion of the new net interenterprise arrears. Therefore, arrears may become a self-sustaining process as the growth of arrears signals to enterprises that the government is likely to step in, and thus arrears wilt grow even more. xxii. However, simply eliminating or outlawing arrears will not help much. The solution of the problem of arrears rests with the introduction of a set of consistent and credible macroeconomic policies, an improved allocation of bank credit and measures to foster a genuine market for trade credit. A clear legal system defining the procedures against delinquent firms and measures such as bankruptcy to enforce contracts are also required. In addition, the determination of taxable profits on an accrual basis, which will become effective with the implementation of the new accounting system, will eliminate an incentive to accumulate arrears. Various other measures, such as interest rate penalties and the modification of Law 76, have also been recently taken by the authorities. Modifications of wage policy as a deterrent - a punitive tax on wage increases in firms with arrears past one month, for instance, - could be considered. However, measures to impose penalties will only be effective if it is believed that they will in fact be enforced and the underlying incentives are changed. Most important, therefore, is the introduction of owners with their own money at risk; this is after all the most powerful incentive for avoiding losses. This suggests that, ultimately, problems in the enterprise sector need to be addressed by a substantial acceleration of the privatization effort. Prerequisites for Preventing Future Bank Losses xxiii. The banks' stock of bad debts indicates losses that have already been incurred.  - vii - Recognizing these losses, therefore, does not imply an additional cost; these bad debts are being implicitly serviced, by bank customers through higher interest rate spreads or, eventually, financed by the inflation tax. Government recognition of the bad debts would make the costs transparent and require budgetary provision from general government revenues for servicing the debts. xxiv. The more important problem, therefore, is the recurring flows of bank losses and the continuing direction of resources to non-productive uses that is implied. If no action is taken to stop these, there is no point in an infusion of government funds into the banks for their recapitalization. For, although it would make costs transparent and, possibly, shift the burden to less distortionary taxes, it would signal that the government would continue to pick up future bad debts. And, as long as enterprise losses continue to be financed by banks, either because of irresponsible credit practices or because of direct state interference in bank management, there would be a continued requirement for funds. xxv. In the course of the last two years, the Government has taken three measures which have had important effects on the banks' asset portfolio. The schemes are: first, substituting government bonds for pre-1991 non-performing bank loans to state enterprises (Law 7); second, providing government deposits to banks to be on-lent on preferential terms to enterprises before being converted into capital of the banks; and third, requiring banks to provide credit to enterprises so as to clear their debts to other enterprises (the global compensation scheme). xxvi. These schemes had very serious deficiencies. First, none of these schemes included enterprise restructuring as a condition. So, in effect, the government has assumed responsibility for past losses, without changing anything to prevent future losses. In fact, it has provided perverse incentives by signalling that it will step in to bail-out loss-making activities repeatedly. Second, government resources for recapitalizing banks have not been transparently budgeted. After the surpluses of extra-budgetary funds were exhausted (for Law 7 bad debts), the means of recapitalizing the banks has been increasingly to resort to money creation. xxvii. Third, the global compensation scheme has worsened the position of banks, since a large share of the global compensation credit went to former bad debtors of banks. At the same time, since there was a tightening of monetary policy, banks financed the GC credit by reducing credit to viable enterprises. The repayment of GC credit was then financed by the provision of fresh bank credit to the same enterprises. This, together with a large increase in preferential credit to enterprises in some sectors, squeezed normal performing credit even further. Banks increased their spreads as a means of passing on the costs of the large and growing share of non-performing loans and preferential credit. In order to limit the increase in lending rates, deposit rates were suppressed and deposits fell sharply as depositors shifted to holding foreign currency. This, in turn, led to a decline in real credit. As a result, credit to viable enterprises has been squeezed, the rate of subsidy on credit has increased, and the deposit base has fallen. These developments would inevitably have an adverse effect on output and inflation, as credit to productive sectors has declined and so has the demand for base money. xxviii. It is, therefore, crucial that the behavior of banks is altered so that future lending is sound. For this reason, partial and frequent 'recapitalization' by the government should be avoided. Recapitalization should be once and for all and must be accompanied by a change in incentives. Such a change can be brought about through privatization of banks once there are effective and strong  -viii - prudential regulations in force or, until such time, through a change in corporate governance of banks whereby managers have the incentives to act responsibly and there is no government intervention (the banks are run on an 'arms length' basis). At the same time, however, since the root of the problem lies in the state enterprises, enterprise restructuring, leading to privatization, has to be addressed together with banking reform. Unless enterprises face a "hard budget constraint", banks face no risk. B.3 The Need for Tax Reform xxix. In order to meet its rising expenditure commitments, the government needs to secure non-inflationary sources of finance. Fiscal revenues have already fallen by a dramatic 10 percentage points of GDP since the beginning of tax reforms and it is likely that they will fall further, in line with the reduced role of the government in the economy. Romania, like most countries in Eastern Europe, still has a higher revenue ratio (about 40 percent of GDP) than countries with similar income and economic structures. To some extent, tax revenues have been propped up by the effect of inflation, which has artificially boosted enterprise profitability. This has been reflected in an increase in profit tax revenues in real terms as well as in the taxation of inflationary gains from inventory revaluation of state enterprises. However, the inventory effect is likely to fade away as firms adjust to the damaging interaction between inventory holdings, inflation and the tax burden. Therefore, the net effect of inflation is likely to go the other way in the future. Moreover, through its impact on the tax burden, inflation is an arbitrary and unpredictable source of revenue, while masking the underlying fiscal problem. It is likely that profit tax revenues will fall; as economic restructuring advances, the contraction of the state enterprise sector will further undermine the government's traditional tax base. xxx. Thus, Romania faces the major task of constructing a modern tax system at the very time it also needs to increase revenues. The authorities have already moved far in this direction with a profit tax, an individual income tax and substantial modifications of the turnover tax to facilitate the introduction of a broad-based value added tax (VAT)." It is uncertain, however, if Romania has the requisite administrative machinery in place to operate such a complex tax effectively. The choice of a single VAT rate (18%), and a sufficiently high threshold level to reduce the number of small taxpayers, will reduce the administrative burden, although the negative distributional impact of a uniform VAT rate may lead to serious resistance. There are, however, a large number of proposed exemptions; exemptions should, at the least, be replaced by inclusion at a zero rate since exempting prevents enterprises from reclaiming VAT paid over inputs. It would not materially add to complexity since there already is a zero rate category (exports). Also, exemptions for small entrepreneurs could, as experience in other countries suggests, be replaced by a form of presumptive taxation, for example, based on sales and an assumed profit-to-sales ratio. Those falling under the threshold for the regular VAT could be offered the choice to opt for the presumptive tax or for full VAT treatment with its attendant reporting requirements.' Although the proposed rate of the VAT has apparently been set to provide a cushion above revenues from the turnover tax that it will replace and the excise taxes will become supplementary, recent experience in administration of the new / MThe VAT was introduced in July 1993. 9/ There is no need, for instance, to provide a blanket exemption to the self-employed; they may automatically opt out of the VAT system if they fall below the threshold, opting for the presumptive tax instead.  - ix - customs duties and excises strongly suggest that the projected revenues may not be collected. Widespread evasion of these taxes has been noted by the authorities themselves. In this context, it is important to put more emphasis on simplification and enforcement of the existing taxes."