Report No. 27599-TUN Republic of Tunisia Strategy for Public Debt Management January 2004 Social and Economic Development Group Middle East and North Africa Region The World Bank Document of the World Bank FISCALYEAR 2004 July 1-June 30 Vice President: ChristiaanPoortman ~ CountryDirector: TheodoreAhlers Sector Director MustaphaNabli Sector Manager Pedro Alba Task Team Leader AristomeneVaroudakis This report was prepared by a team comprising: Sergei Shatalov, Aristomene Varoudakis, and Jean-Luc Steylaers (consultant). Lars Sondergaardand Masakazu Someyaprovided valuableassistance. Peer reviewers were Phillip R D Anderson and Thordur Jonasson. Guidancewas provided by Mustapha Nabli (Sector Directorand ChiefEconomist), and PedroAlba (Sector Manager). Valuable supportfrom the counterpartteam in the Ministryof Finance in Tunisia is gratefullyacknowledged.The team also benefited from discussions with the Central Bank of Tunisia, the Direction de la Prevision of the Ministry of Development and International Cooperation, the Securities Commission, the Direction of the Stock exchange, and a number of banks and "Specialistesen Valeursdu Tresor". The team benefited from commentsby: DomenicoFanizza (IMF) ,AbdourahmaneSam (IMF), and Dimitri Vittas. PREFACE Over the past two decades, the Government o f Tunisia successfully controlled its debt burden, gradually decreasing it, despite the country's exposure to unfavorable external shocks. Tunisia successfully developed its access to international capital market, and is among the few emerging market borrowers that have investment grade rating. At the same time, a domestic market for government fixed-income instruments was launched, and the Government is gradually implementing a shift from exclusively external to domestic sources of fimding. To maintain these positive trends and consolidate Tunisia's position as one o f the best- quality emerging market borrowers, the authorities plan to strengthen the public debt management strategy, with the aim o f better managing risks while reducing the medium- term cost o f the government debt portfolio. Introducing an action plan for active public debt management i s part of the Government's strategy to maintain a stable macroeconomic framework, and is being supported by the Third Economic CompetitivenessAdjustment Loan (ECAL 111). The need for an effective public debt management strategy has increased with Tunisia's stronger presence in the international financial markets and the larger exposure to changing borrowing conditions and exchange rate fluctuations. At the same time, there are better conditions today for public debt management, with the deepening o f the secondary market for Tunisian debt instruments on the international bond market, while the public debt management strategy would be greatly strengthened by the steps taken to develop the domestic government securities market. This study discusses options to the reform of the government public debt management practices, with the aim o f increasing their efficiency, consolidating further the country's market access and containing the costs and risks o f borrowing in both external and domestic markets.The study is intended to facilitate the introduction of an action plan for the implementation o fthe public debt management strategy, as part o fthe set o fmeasures aiming at strengtheningthe macroeconomic framework inECAL 111. Chapter One presents debt sustainability scenarios and discusses the underlying vulnerability factors. Chapter Two examines key principles o f a strategy for public debt management, and presents a discussion o f how active risk management can be progressively introduced in Tunisia. The analysis benchmarks Tunisia's debt situation against other emerging economies with comparable characteristics. Chapter Three addresses reform options to step up the development o f domestic government securities market-a key component of the strategy. Finally, Chapter Four considers options in institutional reforms that would facilitate the implementation o f the desired innovations indebt management strategy andoperations. i REPUBLIC OF TUNISIA STRATEGYFORPUBLICDEBTMANAGEMENT Table of Contents PREFACE ........................................................................................................................... i EXECUTIVE SUMMARY ............................................................................................ IV I COPINGWITHRISKSTOMEDIUM-TERMDEBTSUSTAINABILITY . .....1 1.1 CREATING AN ENABLINGENVIRONMENT FORACTIVE DEBTMANAGEMENT: THE ROLE OF A SOUND MACROECONOMIC POLICY FRAMEWORK INTHE FACE OF RISKS ...........1 1.2 THEMEDIUM-TERM BASELINE SCENARIO: DEBT SUSTAINABILITY SECURED BY FAST GROWTH .................................................................................................................. 3 1.3 A STRESS TEST: PERSISTENTEXTERNALSHOCKSWOULD PUTMEDIUM-TERM DEBT SUSTAINABILITY AT RISK ................................................................................................. 5 1.4 CREATING MORE FISCAL SPACEWOULD HEDGE PUBLIC DEBTFROMDOWNSIDE MACROECONOMIC RISKS AND THE CONTINGENT LIABILITIES OF THE PUBLIC SECTOR .......9 I1 . DEBTAND RISKMANAGEMENTSTRATEGY ......................................... 14 11.1 OBJECTIVESOF GOVERNMENT DEBTMANAGEMENT ......................................... 14 11.2 SCOPEOF A PUBLIC DEBT MANAGEMENT STRATEGY ........................................ 17 11.3 STRATEGIC BENCHMARKS FORPUBLICDEBT MANAGEMENT ............................ 19 11.4 R I S K S INTUNISIA'S DEBT PORTFOLIO ............................................................... 21 11.5 DECISIONSUPPORTMECHANISMS: THEROLEOFINFORMATIONANDANALYSIS .31 I11 . PROMOTINGTHE DEVELOPMENTOFDOMESTICGOVERNMENT SECURITIESMARKETS ............................................................................................. 33 111.1 STATUSOFTHEDOMESTICGOVERNMENT SECURITIESMARKETS........................ 33 111.2 AN ACTIVE MONEY MARKET IS AN IMPORTANTUNDERPWNMGOFDOMESTIC SECURITIESMARKETS ..................................................................................................... 35 111.3 THEPRIMARY MARKET FORGOVERNMENT SECURITIES ...................................... 38 111.4 THESECONDARYMARKET FORGOVERNMENT SECURITIES................................. 41 111.5 REINFORCING THE INVESTOR BASE .................................................................... 44 111.6 OPTIONSFOR REFORM ........................................................................................ 46 III.6.I Structural initiatives.................................................................................. 47 11I.6.2 Technical measures................................................................................... 49 11I.6.3 Phasing in the reforms .............................................................................. IV. INSTITUTIONSAND INFORMATIONFOREFFECTIVEDEBT MANAGEMENT ............................................................................................................ 55 ANNEX 1 MEDIUM-TERMDEBTSUSTAINABILITY SCENARIOS . .................64 BASE CASESCENARIO ................................................................................................... 64 LowCASESCENARIO .................................................................................................... 65 FISCALADJUSTMENT SCENARIO .................................................................................... 66 .. 11 ANNEX 2 FISCAL RISKBEYONDTHE BUDGET . ................................................ 67 DeJning Quasi-Fiscal Activities and Hidden Deficits.............................................. 67 Approaches to the management of quasi-fiscal risks................................................ 69 ANNEX 3 FUNCTIONAL COMPONENTS OF THE GOVERNMENT DEBT ........................................................................................................................... . OFFICE 76 ANNEX 4 BENCHMARKING TUNISIA AGAINST OTHER EMERGING . ECONOMIES .................................................................................................................. 79 REFERENCES ................................................................................................................ 80 Listof Figures Figure1.1 .A counter cyclical fiscal stance, despite apersistent structuralbudget deficit.2 Figure1.2 .The baseline scenario: Drivers o f growth and macroeconomic imbalances.....4 Figure I-4.The low-case scenario: A downturn drivenbypersistent extemal shocks ........5 FigureI-3.Debt dynamics inthe baseline scenario............................................................. 6 Figure I-5.Debt dynamics inthe low-case scenario............................................................ 8 Figure 1.6 . Inthe low-case scenario external vulnerability would increase ....................... 8 Figure1.7 .A fiscal consolidation scenario to maintain debt sustainability inthe face o f external shocks.......................................................................................................... 10 Box 11.1 .Benchmark Tunisia: Progress Towards Sound Debt Management Practice.....16 Box 11.2 . Cost-at-Risk and Budget-at-Risk ...................................................................... 21 Figure11.1 .Short-term debt as percent o f (a) total external debt and (b) foreign reserves ................................................................................................................................... 23 Figure11-2.Average Maturity o fDebtfrom Private Sources............................................ 24 Figure 11.3 . Structure o f Tunisia's sovereigndebt by creditor, percent o f GDP ..............25 Table 11.1 .Structure o f external debt outstandingby creditor.......................................... 24 Figure 11.4 . Currency composition of Tunisia's long-term extemal debt and trade flows26 Table 11.2 .Currency composition o f long-term extemal debt, end o f 2000..................... 27 Figure 11.5 . Share o f variable rate debt inexternal long-term debt .................................. 29 Figure 111-1.Maturity profile o f domestic debt and o f its issuance in2002...................... 34 Figure 111-2.Moneymarket interest rates and interbank loans.......................................... Box 111.1 . The role o fbank incentives inthe development o f an active bondmarket .....36 37 Box 111-2.Primarydealers ingovernment securities markets ........................................... 39 Box 111.3 .The costs of a reform strategy o f domestic debt markets ................................ 40 Box 111.4 .Benchmark Tunisia: development o f an efficient government debt market....46 Box IV.l .Evolution o fDebt Management Institutions inthe OECD Countries ............. 59 Table 11.1 .Structure o f extemal debt outstanding by creditor.......................................... List of Tables 24 Table 11.2 .Currency composition o f long-term external debt, endo f 2000..................... 27 ... 111 Executive Summary Over the past two decades, the Government of Tunisia successfully controlled the public debt burden, gradually decreasing it, while pursuing a very prudent debt management policy. During this period, the share o f private market borrowing in the government debt portfolio has grown substantially, while aid flows have stagnated and capital grants declined, reflecting the country's progress indevelopment andhigher levels o f income per capita. Tunisia successfully developed its access to intemational capital market, and managed to avoid serious contagion from the emerging market crises. Tunisia is among the few emerging market borrowers that have investment grade rating. In the same period, Tunisia had launched a domestic market for government fixed- income instruments, and is gradually implementing a shift from exclusively extemal to domestic sources o f funding. Tunisia's financing program is anchored in the Five-Year Development Plan and is further detailed in annexes to its annual budgets. To secure fiscal sustainability, Tunisia follows a simple but efficient principle, keeping the total amount o f sovereign debt under 60-65 percent o f its GDP. Its liquidityposition has been comfortable during most o f the past decade, with debt service to exports ratio staying below 20 percent, and in recent years hovering around 15 percent. More recently, it has become evident that to maintain these positive trends and to consolidate its position as one of the best-quality emerging market borrowers, the Government of Tunisia will need to restructure is debt management policies and institutions. The share of extemal debt remains large, and, together with the structural deficit on the current account, makes the country vulnerable to the volatility of intemational financial flows. At the same time, structural characteristics o f Tunisia's financial sector and the still narrow domestic investor base make the deepening o f domestic capital markets a challenging task. The current approach managed to stabilize Tunisia's overall risk profile, but the efficiency of public debt management could be enhanced. Outstanding issues remain with regardto the strategic vision andthe scope o f Tunisia's public debt management; the measurement and disclosure o f risk; and the reinforcement o f debt management institutions and domestic market infrastructure. Complementary actions should be envisaged on a number of fronts: (i) Anchoring the strategy in a medium-term, rolling macroeconomic and fiscal framework; (ii) Formulating an integratedrisk management strategy, based on appropriate risk benchmarks, taking into account all types o f risks the government portfolio o f debts and related assets may contain; iv (iii) Taking steps to strengthen the domestic government securities market- especially its medium- and long-term ends-as this is the only viable means to manage the exchange androllover risks while reducing borrowing costs. (iv) Working towards the reorganization o f the debt management institutions, to improve focus and flexibility, so as to enable the implementation o f an effective public debt management strategy. Because Tunisia is vulnerable to shocks, debt management should be cast in a medium-term framework that helps reckon the underlying risksfor debt sustainability. Risks to medium-term debt sustainability may arise from persistent external shocks, and may be exacerbated by existing rigidities inthe budget. Promoting a robust medium-term fiscal policy framework will bolster the confidence o f domestic and foreign investors on Tunisia's ability to cope with the surrounding risks, thus facilitating the move indue time towards active public debt management. International experience suggests that maintaining a sound macroeconomic policy framework is a prerequisite to reducing government financing costs over the medium to long term, as it enables the risk premia on government securities to decline. Such a framework also facilitates the development o f govemment securities markets-especially extending the yield curve on domestic government securities into the medium and the long term-by strengthening investor confidence and improving the incentives to invest in government securities at reasonable market conditions. Tunisia has a good track record of prudent macroeconomic management, but the transition to greater integration into global markets will raise challenges for debt sustainability due to competitive pressures and increased exposure to volatility. A prudentmacroeconomic policy framework is a valuable asset inpublic debt management and the development of the domestic government securities market. A prudent monetary policy has held inflation subdued, while helping keep external imbalances to sustainable levels. Fiscal consolidation has progressed, and despite a persistent primary structural budget deficit, the fiscal stance has been generally counter cyclical. Tunisia's choice to promote greater integration into global markets will lay the ground for faster growth, furtherbolstering investor confidence, and supporting debt sustainability. However, trade liberalization will make it more difficult to reduce both current account and budget deficits, while growth and the external balance may suffer setbacks from the increased volatility o f tourism andtravel. I n a baseline scenario that reflects Tunisia's good medium-term growth prospects, despite some tension on the external and internal balances, public debt would be sustainable. Total public debt would decline to about 54 percent o f GDP by 2012, from 61 percent in 2002. Domestic debt would increase to about 30 percent o f GDP, partly substituting for foreign public debt. Total foreign debt would decline to about 44 percent o f GDP-after accounting for foreign non guaranteed private debt rising to 20 percent o f GDP, owing to the persistent current account deficit and reflecting the improved access o f the private sector to international capital market financing. The favorable total public and foreign debt dynamics in the baseline scenario depend critically on the projected strong growth o f the Tunisian economy. V The downside risks to GDP growth, stemming from Tunisia's transition to a more competitive international environment, represent the main threat to debt sustainability in the medium term. A "stress test" for debt sustainability highlights the downside risks associated with a worst-case combination o f unfavorable external shocks and a slow export recovery. Insuch a low-case scenario, total public debt could increase by as much as 20 percentage points o f GDP compared to the baseline. Most o f the increase inpublic debt would come inthe form o f domestic debt (13 percentage points of GDP), due to the increase in the primary budget deficit. Owing to a significant widening o f the primary current account deficit, foreign debt would become unsustainable, increasing by about 20 percentage points o f GDPby 2012 (including private non guaranteed debt). Creating more fiscal space would help accelerate the reduction of public debt while hedging public debt against the surrounding downside risks. Tunisia's foreign debt ratio i s higher than in the average o f other countries with a similar sovereign rating, suggesting that a faster reduction o f public debt could yield substantial benefits in terms o f access to international capital markets. Moreover, adjustment to preserve debt sustainability in a low case scenario would call for an important fiscal consolidation effort-estimated to 3.6 percentage points by 2012, assuming away the use o f other adjustment options. The fiscal adjustment that would be required to stabilize debt levels underscores the needed additional room inthe budget to prevent an increase indomestic andforeign debt ifthe highlightedshocks were to occw inthe medium term. Main steps in this direction would call for improving the flexibility of the budget and further fostering public expenditure efficiency-as this will free up budgetary resources and help create the fiscal space needed to hedge the downside risks to public debt. Taking early steps in fiscal consolidation would also hedge public debt from future calls on the budget that may arise from the contingent and implicit liabilities of the public sector. Contingent liabilities may arise from foreign exchange guaranties issuedby the government-especially if policy adjustment to persistent external shocks were to partlyrely on exchange rate adjustment. Fundedguarantee schemes-such as the "Fonds depkrkquation des changes ",whose liabilities are increasing rapidly, now amounting to 5 percent o f GDP-may also generate losses, as suggested by the significant decrease in the Fund's reserves. In the absence o f reform, the pension system may soon become a source o f implicit liabilities for the government due to its generosity (especially the CNRPS). Contractual guarantees and other explicit contingent liabilities must be viewed aspart of the government liability porlfolio, and risks comingfrom them should be factored into the overall public debt strategy.. The management of contingent liabilities requires the adoption o f more sophisticated tools not only by debt managers, but by the Budget Office, and other government agencies. These tools allow the Ministry o f Finance to make an informed choice across the whole spectrum o f financial instruments, and, when deemed necessary, to set up appropriate provisioning levels for explicit contingent liabilities o f the state. Public debt management, therefore, could consist o f three "concentri~'~activities: management o f the government contractual obligations proper; and government authorization and monitoring regime for other types o f public sector vi liabilities not covered by the sovereign guarantee; and primarily analytical follow-up o f the implicit liabilities o fthe government. For a small sovereign issuer such as Tunisia, the challenge is to define a set of strategic, market-neutral benchmarks, which will allow to reduce the overall exposure to portfolio risk. This will also be an important first step towards active public debt management. But active debt management should not be confused with high volume o f transactions which constantly "chase" the benchmark. Active debt management calls primarily for a comprehensive, strategic vision o f risk faced by a sovereign borrower. Periodic rebalancing of the portfolio i s advisable only to the extent it serves this strategic vision. The fimctions o f a debt benchmark include: incorporation o f strategic objectives, limitation o f risk, and measurement o f performance. Recent intemational practice has defined five mainprinciples for buildingportfolio benchmarks: 0 Benchmarks should be tested and be efficient through a large number o f market scenarios, with as little as possible reliance on assumptions about the future economic and financial environment; 0 defined "for the longrun", possibly over the lifetime o f government debt; 0 provide the lowest cost for the chosen level o f risk; 0 simplicity and transparency-to be understood and followed by the common sense; 0 reflect the current structural and institutional constraints o f a sovereign issuer Tunisia isperceived by investors as one of the safer emerging market borrowers, but its rollover (refinancing) risk is still tangible. For extemal debt, the share o f short term debt in the total debt portfolio is now well below 15 percent. However, due to its current account vulnerability, Tunisia has little choice but to be prudent, as its ratio o f short term debt to intemational reserves i s on the high end o f other countries in MENA. Domestic rollover risk is higher, despite the fact that demand from Tunisian investors, primarily commercial banks is stable. Measures aimed at increasing the average maturity o f the public debt portfolio and avoiding bunching o f maturities (using buy-backs and exchanges) decrease the rollover risk. However, in examining such measures, tradeoffs betweendifferent risks mustbe carefully considered. Tunisia also faces a considerable exchange rate risk, despite the stability of the real effective exchange rate of the Dinar. The country has to cope with both types o f exchange rate risk: (a) possible change in the value of the foreign currency debt due to the fluctuations o f the exchange rate o f the domestic currency; and (b) cross currency risk reflecting the risk exposure due to the currency composition o f the foreign debt portfolio. These risks require that two separate benchmarks are defined. The currency benchmark should be formulated, with the aim o f defining a desired proportion between foreign and domestic currency debt. This should reflect the Treasury's capacity to carry exchange rate risk, after taking into consideration the attractive characteristics o f the foreign currency portfolio (often longer maturity, lower: interest level, and much broader investor base). The second benchmark should address the cross currency risk in the foreign debt portfolio. Since the JPY risk is already swapped away, the main value o f the cross currency risk benchmark would be in gauging the risk coming from other major vii currencies (USD and Euro) as well as from the smaller currency components. In taking decisions on swapping away exchange rate risk o f specific liability positions, it has to be ensured that counterpart ceilings are set at prudentlevels. An interest rate benchmark would help determine the acceptable level of interest rate risk moreprecisely. On the extemal debt side, the share o f floating-rate debt has grown steadily since mid-l970s, fi-om about 15 percent o f the total extemal long-term debt to almost 35 percent in recent years. Since the overall level o f debt service is significant, and since the budget is cash-based, fixed rate debt should be generally favored, even at the expense o f increased cost o f the borrowing program, because it makes interest expenditure more predictable over the medium run. However, Tunisia's exposure to interest rate risk is generally in line with MENA, but i s much lower than in other comparable market borrowers. Hence, there may exist some potential to increase this exposure somewhat hrther, after accounting prudently for expected interest rate trends. A fix-floating mix benchmark should be formulated to measure the risk and the performance of the overall portfolio. The benchmark should be formulated with full awareness o f Tunisia's likely limited and costly access to the interest swap market and should also take into account acceptable limits on counterpart risk. I n March 2003, Tunisia became thefirst member county of the World Bank to sign a Master Derivatives Agreement with it. This Agreement will assist the Government o f Tunisia in reducing its currency and interest rate risks via a range o f hedging products offered by the World Bank, including currency swaps, interest rate swaps, caps and collars and, on a case by case basis, commodity swaps. In providing these financial products, the World Bank stands between market institutions and its borrowers, entering into separate financial contracts with each o f them. Tunisia therefore would benefit from financial terms that reflect the Bank's AAA credit rating. The Tunisian Authorities have taken important steps to develop the market of the domesticpublic debt, since this is the only viable way to cope with exchange risk and towards lower rollover risk. Should the access o f Tunisia to the international capital markets be temporarily restricted, the existence o f a domestic liquid market would be essential to limit financing stress. Moreover, the development o f a robust and liquid domestic market would give rise to a domestic yield curve, which could be used as reference for domestic issuers, the diversification o f risks and the reduction o f the long- term cost o f the debt. However, despite adequate technical infrastructure and regulatoy framework, issuance in the primay market remains limited while the secondary market remains fairly inactive. The average maturity o f domestic debt remains short-about 41 percent o f total outstanding domestic debt has remaining maturity o f less than one year and new issuance is skewed towards shorter maturities even more. Moreover, bonds outstanding are concentrated inthe hands o f a limitednumber o f investors, the majority o f which are public entities-such as, for example, the CNSS. The concentrationo f issues on a limited number o f public sector investors, suggests that the primary market is still somewhat artificial.. The number o f transactions in the secondary market is still very limited, and ... Vlll most of these transactions take place within investment groups or in an over-the-counter market, where transparency is lacking. The absence o f secondary market prevents the emergence o f a yield curve. The reasons for the inadequate functioning of the Tunisian public debt market are multiple and mutually reinforcing. Mainbottlenecks can be traced back to five factors: 0 the limited activity inthe money market-an important underpinning o f the bond market-partly due to the monetary policy operating procedures and partly to the incentives structure inthe banking system; 0 the narrow investor base; 0 the issuing strategy o f the Treasury on the primary market; 0 institutional deficiencies that hamperthe secondary market; 0 psycho sociological factors The stability and the predictability of the money market rates have hampered the development of an active money market. To achieve intermediate targets for credit growth, the Central bank o f Tunisia (CBT) relies on the control o f the volume o f bank refinancing. The instruments used create dependence o f banks to CBT refinancing facilities, and convey to the CBT a dominant position in the money market. Inaddition, the absence o f a yield curve prevents adequate assessment o f credit risks, leading to generalized credit indexing on the money market rate (TMM). Since the interest rates o f many deposits and loans agreements are indexed on the money market rate (TMM), changes inmonetary policy are likely to be quickly reflected in the financial system and the economy as a whole. This has created rigidity ininterest rates, promoting dependable and easy access o f banks to liquidity at a predictable rate. The easy and foreseeable access o f Banks to CBT refinancing has discouraged the development o f an active interbank market. The delay in the introduction o f repurchase agreements has also inhibited money market development. interest rate rigidity also discourages the development o f secondary bond markets, wherein transactions rest primarily on differences inanticipatedinterest rate changes. The Treasury, despiteits stated willingness to adhere to the rules of the market in order to meet the government's financing needs, finds it hard to build its credibility as an issuer. Treasury issuance i s perceived as lacking transparency, because the issuance calendar undergoes many changes, such as cancelled auctions, sometimes for a long period. One reason i s that Treasury does not participate inthe interbank market. The lack of transparency and predictability makes market participants uncertain about the issuing policy and reduces the perceived liquidity o f the market. The Treasury's credibility is also hampered by a reluctance to issue at yields deemed higher than what it considers as the appropriate market levels. As a result, bond auctions meet little success. The Treasury encounters particular difficulty to issue in the long term end o f the yield curve. Along with adhering to the rules o f the market, it is important to ensure that sufficient competition from market participants prevails in the Treasury's offerings. Regulatory safeguards that prevent collusive behavior inthe market should, thus, be enhanced, while collaboration between the Central Bank and the "Comit6 duMarch6 Financier" should be strengthened. 1x Developing an efficient domestic public debt market is a long-term endeavor and entails a cost at the beginning. This is suggestedby the experience o f countries which acquired the status o f a high grade sovereign issuer, along with the associated advantages interms o f issuing conditions, reduction of public debt cost ando f the financing cost in the national economy as a whole. The benefits come only later on, when the credibility o f the Treasury as an issuer has been well established. A rule to enhance credibility would call for the Treasury to issue not when it needs but when it can. Moreover, the limited amount o f securities outstanding do not support the liquidity and the development o f the secondary market. The BTA buy back policy o f the Treasury does not contribute to increase neither the outstanding stocks nor the liquidityo f the market. Although the secondary market remains inactive, there exists a "parallel" secondary market for intra-group transactions. The secondary market transactions have to take place in the framework o f the Stock exchange (Bourse de Tunis) but these transactions are non-existent. Nevertheless, over-the-counter, intra-group transactions exist. These transactions lack transparency, as they are not reported to the market authority (the Stock exchange), while their financial conditions may also be questionable, as they may be underpinned by accountancy and balance sheet reasons. Improving the transparency o f these transactions, mainly carried out by the banks, would help the development of the secondary market, but would call for close coordination between the supervisory authorities o f the banking system and the capital market. Moreover, the obligation of quotation i s not constrainingenough to support the development o fthe secondary market. The development of domestic markets government securities has been held back by the absence of domestic or international investors with a strong appetite for long maturities. The pension funds are not largely developed in Tunisia, while he global financial health o f the insurance sector is precarious and life insurance is still in its infancy. By contrast, the Collective Investment Institutions are reasonably well developed, with total assets o f SICAVs surpassing 5 percent o f GDP and growing fast. The SICAV could contribute to the development o f the secondary market and the establishment of a yield curve if the "intra-group" transactions were properly regulated. Further strengthening o f supervision would also be needed, especially to ensure proper valuation of assets and compliance with custodial and auditing rules. With an appropriate framework inplace, the development o f a capitalization pillar for the pension system would contribute to increased institutional savings and reinforce the development o f a domestic bondmarket. . Partially opening of domestic debt issuance toforeign investors could be seen as a step to strengthening the demandfor government securities. However, this opening implies a meticulous preparation to guarantee its success, since a false step might prove expensive interms of investor confidence (investors do not make the difference between loans in currency and domestic loans) thereby influencing Tunisia's access to international capital markets. An option would be to open a new line of three-month BTCT ,which would be entirely open not only to residents but to foreign investors as well. That would help familiarize foreign investors with the local market given that liquidity is less o f concern for short-term paper than long-term bonds, while Treasury X would retain control o f stocks outstanding. Dependingon experience, this access could be gradually broadened. I n Tunisia, debt management practices are strong, but will benefit from greater centralization and functional focus. Government debt management is scattered across several agencies. The current setup provides good control over the individual debt instruments, but does not support the integrated view o f public debt portfolio, and could obstruct active risk management. The debt management function must be centralized; and debt management units must be given more operational flexibility. A new government function for risk management must be created, and a sustainable medium-term debt strategy must be formulated, relying upon clearly defined quantitative benchmarks. These needs can be met most efficiently via the establishment of a specialized Debt Agency, to manage the whole liability portfolio of the Tunisian government, and a High Debt Committee to steer the Agency and to set the medium-term debt strategy. The Committee should be chaired by the Minister of Finance, and have senior-level representation from all relevant government institutions as well as the BCT. Internally, it is advisable to organize the Debt Agency by function, as follows: the Back Office (debt recording and accounting), the Middle Office (risk management and analysis) and the Front Office (borrowing and hedging operations). Within the new structure o f the Debt Agency, the Middle Office should be responsible for drafting a comprehensive medium- term liability management strategy. Once the strategy is approved by the High Debt Committee, the Middle Office should supervise its implementation. In doing this, the Middle Office shouldnot engage indebt transactions, but serve as a risk "watchdog". Strong capacityfor active debt management requires an adequate system of incentives. Tunisian government debt managers should have attractive remuneration and promotion structures, comparable with the private sector investment institutions, which need similar skillprofiles. The Debt Agency should have greater autonomy inmanaging its pay scales and hiringprocedures. Institutional reform requires strong political leadership and a new legal framework. The reform would entail deep shifts in the mandates o f various government agencies. Strategic vision and leadership is needed to escape bureaucratic impasse, and to realize long-term gains from the new institutional structures. A coherent legal framework, giving the newly created Debt Agency a strong risk control mandate, must shore up the envisaged reforms. It i s advisable to pass a new Law on Public Liability Management, endorsing the contemplated reform o f debt management institutions andprocesses. x i 9 e, b cu .3 x e E id k 0 g .e Y % 3 W e e, ec e, 3 Ac0 * R I. Copingwithriskstomedium-termdebtsustainability I.1 Creating an enabling environmentfor active debt management: the role of a sound macroeconomicpolicyframework in theface of risks 1. Public debt management and the development of a government securitiesmarket is more successful when a consistent medium-term macroeconomicpolicy framework is in place. Domestic and foreign investors will be reluctant to hold government securities, especially medium- and long-term maturities, when there are expectations o f macroeconomic imbalances, leading to high inflation, overvalued exchange rates, and raising even remote risks o f default. Public debt management is, therefore, supported by a sound macroeconomic policy framework with a credible commitment to prudent and sustainable fiscal policies, stable monetary conditions, and a credible exchange rate regime (World Bank/IMF, 2002). Maintaining a sound macroeconomic policy framework is a prerequisiteto reducing government financing costs over the medium to long term, as it enables the risk premia on government securities to decline. Moreover, such a framework facilitates the development o f government securities markets because it bolsters investor confidence and improves the incentives to invest in government securities at reasonable market conditions. Stable expected monetary conditions- including a credible exchange rate-is a precondition for extending the yield curve on domestic government securities into the medium andthe longterm. 2. Promoting a robust medium-term fiscal policy framework is key to strengthening policy credibility and bolstering the confidence of domestic and foreign investors. Creating an enabling environment for active debt management calls for consistency among fiscal, monetary, exchange rate, and capital account policies, and also requires a financially healthy banking system. However, maintaining a robust medium-term fiscal framework i s the key underpinning o f debt sustainability and also determines the ability of the government to cope with macroeconomic risks and unexpected calls on the budget. Such a fiscal framework should be able to efficiently collect direct and indirect tax revenues; provide an effective budgeting and expenditure control system that ensures flexibility without compromising expenditure efficiency; and be able to account for contingent liabilities o f the public sector-including public pensions and diverse government guarantees. 3. Tunisia has a good track record of prudent macroeconomic management, which is a valuable asset in the move towards active public debt management and the development of the domesticgovernment securities market. A prudent monetary policy has held inflation subdued, while helpingkeep external imbalances to sustainable levels. Fiscal consolidation has progressed, and despite a persistent primary structural budget deficit, the fiscal stance has been generally counter cyclical-as suggested by the positive correlation between the estimated primary structural budget deficit and the output gap (Figure AI).' When Tunisia faced severe external shocks, as in 2002, swift fiscal 'The structural budget deficit was estimated after removing the cyclical component from the actual deficit, due to the operation of fiscal automatic stabilizers. Automatic fiscal stabilizers inTunisia are related to the 1 adjustment has been forthcoming to help absorb part o f the external financing gap. For example, in 2002, the withdrawal of fiscal stimulus amounted to 0.6 percent o f GDP, while the 2003 budget maintains a similar pace o f fiscal consolidation. FigureI-1.A counter cyclicaljkal stance, despite apersistent structural budget deJicit . . ... ~Inpercentof NonAgnculuialpotentialGDP 4 .. . . . . ... ..-.. . .-.. .--.-. . .. -. R 2 0 2 4 -6 8 / Pnmaryjstructmlbldgetdeficit -10 . . - 8m Sm m m m m m m m a m m m o m m m m m m o o o o ~ 8 ~ 9 8 6 5 ~ 8 S ~ 8 Z 8 8 6 8 8 ~ ~ ~ ~ i r r i r r r r r r r r r r r r r - r i N N " Source: World Bank staff 4. Over the past two decades, Tunisia successfully controlled the public debt burden, but faster reduction of foreign debt could yield substantial benefits. Prudent macroeconomic policies have helped Tunisia improve its access to international financial markets, allowing borrowing at long maturities and under relatively favorable conditions (see Chapter 2). However, Tunisia's foreign debt remains high in international comparison, especially in view o f the country's exposure to extemal shocks. Because o f the structural current account deficit and the persistent primarybudget deficit, public debt hovered at around 60 percent during the 1990s (of which, 39 percent foreign and 22 percent domestic debt in2002). Tunisia's foreign debt ratio i s about 12 percentage points of GDP higher than in the average o f 10 other countries with a similar sovereign rating (BBB-, BBB, and BBB+), suggesting that a faster reduction o f public debt could yield substantial benefits interms o f access to international capital markets.2 5 . Tunisia has made thefirm choice of integrating more closely the world economy- a strategic move that will lay the ground for faster growth and support debt sustainability. Ongoing trade liberalization, spurred by the Association Agreement with the EU, will facilitate integration o f Tunisian firms into EU cross-border production networks, thus further improving Tunisia's attractiveness as a hub for export-oriented foreign direct investment. Increased domestic and foreign investment and development o f new competitive niches will create an enabling environment for faster growth, helping Tunisia to meet the growth challenges o f the lothDevelopmentPlan andbeyond. Fast and procyclical changes in direct and indirect taxes. However, because the tax base o f direct taxes-which are far more sensitive to GDP changes-is narrow, the cyclical component o f the budget i s relatively small. The cyclical component hovers around an estimated -10.5 percent o f GDP, so that the estimated structural budget deficit comes very close to the actual deficit. * The comparators are Latvia, Poland, China, Lithuania, Slovak Republic, Trinidad and Tobago, Croatia, Mexico, South Africa. and Thailand. 2 sustained growth will further bolster investor confidence as it will be critical in supporting Tunisia's debt sustainability inthe mediumand long term. 6. But the transition to greater integration into global markets will also raise challenges, due to competitivepressures and exposure to volatility. The tariff reductions scheduled in the Association Agreement with the EU, combined with the need for large imports o f equipment and intermediate goods to support fast growth, will make it more difficult to reduce current account deficits. The loss of tax revenues on extemal trade will enlarge the financing needs inthe budget. Also, with the gradual phasing-out o f the MFA quota system by 2005, Tunisia's textile exporters will face stiffer competition from low- cost producers in traditional European markets (see World Bank, 2001). Deeper integration into EUmarkets will heighten Tunisia's dependency on economic swings in the EU, where growth has proven less resilient to global downturns, as seen inthe 2001- 2002 slowdown. Anemic growth in the EU, which absorbs more than 75 percent o f Tunisia's exports and i s home to more than 70 percent o f the country's foreign visitors, is swiftly felt in Tunisia, as in the sharp slowdown seen in 2002. In the emerging intemational environment with heightened security concems, growth and the extemal balance may also suffer setbacks from the increased volatility o ftourism receipts. 7. Because of Tunisia's vulnerability to shocks, debt management should be cast in a medium-term framework that helps reckon the underlying risksfor debt sustainability. Risks to medium-term debt sustainability may arise from persistent extemal shocks, and may be exacerbated by existing rigidities in the budget. The strategy o f public debt management will thus need to go in tandem with initiatives that further strengthen the fiscal framework, with the aim of increasing the government's ability to cope with the surrounding risks. To highlight how such risks may affect Tunisia's current account and fiscal balances, and the dynamics o f foreign and domestic debt, two scenarios are reviewed below. Both cover a ten-year period, up to 2012, to allow for long-term adjustments o f debt inthe face o f shocks. A baseline scenario, reflecting Tunisia's good medium-term growth prospects, underpinnedby deeper integration into global markets, sound fundamentals, and a good track record o fprudent macroeconomic management inthe face o f shocks. A low-case scenario, reflecting the downside risks and Tunisia's vulnerabilities in the transition towards deeper intemational integration. This scenario should be considered as a benchmark, reflecting a combination o f risks that would give rise to a "worst-case" outcome. A number o f intermediate scenarios would become likely inthe presence o f milder shocks. 1.2 Themedium-term baseline scenario: debt sustainability secured byfast growth 8. The baseline scenario assumes steady GDP growth, but accounts for current accountpressures and loss offiscal revenues in the run up to thefree-trade zone with the EU. Real GDP grows by an annual rate o f 5.6 percent on average, leading to a nearly doubling o f Tunisia's per capita GDP by 2012 (see Annex 1 for a detailed description). Growth is underpinned by healthy investment and productivity growth, while export growth, spurred by better competitiveness and export-oriented FDI, provides added 3 impetus to GDP growth (Figure I-2a). However, reflecting steady growth o f equipment imports and the dismantlement o f tariff protection, the primary current account turns from a small surplus into a deficit o f nearly 1 percent o f GDP (Figure I-2b). Because imports from the EU account for more than 75 percent o f Tunisian imports, the completion of the Association Agreement with the EUwill entail a sizeable loss o f fiscal revenue from external trade taxes. Despite improved tax collection, the primary budget deficit worsens by 0.6 percentage points o f GDP compared to 2002 (Figure I-2b). FigureI-2. The baseline scenario: Drivers of growth and macroeconomic imbalances Growth i s underpinned by healthy investment and ...but tariff dismantlement exerts pressure on increased integrationwith global trade.. . external and internal balances InpercentofGDP hpercentofGDP I 30 2 1 6 0 T 55 -- 50 -- 45 -- ,--.-+. ........ Currentacmuntdeficit /' *.__ .._ 40 -. =... '%,, ..... ----./' ,. , 1 Budgetdeficit ( 1 35 20 -5 ................... . . (4 Source: World Bank staff 9. Despite some tension on the external and internal balances, debt would be sustainable. Total public debt declines to about 54 percent o f GDP by 2012, from 61 percent in2002-although at a much slower pace after 2006 due to the wideningprimary current account and fiscal deficits. At the same time, the persistent fiscal deficit entails an increase indomestic public debt to about 30 percent o f GDP. Domestic debt i s thus partly substituted for foreign public debt (Figure I-3.a). By contrast, foreign private debt increases to 20 percent o f GDP, from 12 percent in 2002, owing to the persistent current account deficit and reflecting the improved access o f the private sector to international capital market financing. However, total foreign debt declines to about 44 percent o f GDP, from 51 percent in 2002 (Figure I-3.b). Owing to the favorable foreign debt dynamics, external vulnerability somewhat recedes, with the debt service ratio hovering around 11percent o ftotal exports, down from about 14percent in2002. 10. Debt sustainability is secured by high growth rates. The favorable total public and foreign debt dynamics in the baseline scenario depend critically on the projected strong growth o f the Tunisian economy. With growth projected at 5.6 percent per year, significantly above the real interest rates on foreign and domestic debt (see Annex l), the likelihood for debt sustainability improves dramatically, provided the primary current 4 account and fiscal deficits remain below reasonable margins.3 The downside risks to GDP growth, stemming from Tunisia's transition to a more competitive international environment, represent the mainthreat to debt sustainability inthe medium term. Figure I-3.Debt dynamics in the baseline scenario Total public debt is reduced while foreign private debt ...butindicators o fexternalvulnerability hpercentofGDP Inpercent 80 16 70 14 60 12 10 50 8 40 6 30 4 20 2 10 0 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 /I iGDP(leftam)+Debt Totalforeigndebt servlceltotal expoits(nghtans) Source: World Bank staff 1.3 A stress test:Persistent external shocks wouldput medium-term debt sustainability at risk 11. The low-case scenario highlights the downside risks of combined unfavorable external shocks and a slow export recovery. Inthis scenario, Tunisia suffers a decline in export demand for goods and services in 2005, due to a combination o f adverse developments: A possible slowdown ingrowth o fits maintrade partners; w A loss ofmarket shares intextile andclothing exports due to the removal o fMFA quotas; A drop in tourism receipts, similar to that seen in 2002, possibly due to slow economic activity in the EU and/or to renewed security concerns in international travel. The drop in export revenues lasts for two years, while sluggish export growth lowers exports by about 6 percentage points of GDP by 2012 (Figure I-4.a). The lowering o f tariffs and the completion o f the Association Agreement with the EUwould be likely to boost import demand despite the slowdown o f activity. The primary current account deficit widens by about 2 percentage points o f GDP compared to the baseline scenario, to reach 2.5 percent o f GDP by 2012 (Figure I-4.b). As established in previous studies (World Bank, 2001), with a growth rate o f 5.5 percent and a real implicit interest rate o n foreign debt o f 3.3 percent, Tunisia could afford a primary current account deficit o f up to 1.1percent of GDP without compromising foreign debt sustainability. 5 12. Persistent external shocks would spread over to domestic demand, thus exacerbating the downturn. Domestic investment drops, thus, by about 2.5 percentage points o f GDP compared to the baseline scenario (Figure I-4a). Foregone growth is about 5 percent per year on average during the first three years after the shock, with growth only gradually resuming thereafter, to reachthe baseline by 2012. FigureI-4.The low-case scenario: A downturn driven bypersistent external shocks Reflecting trade liberalization, shrinking exports are ...leading to anunsustainable current account only partly offset by a decline inimports.. . deficit, while the primary budget deficit temporarily increases Differencestobaseline scenario,Inpercent Differencesbbaselinescemno,inpercentofGDP____________ -y--T--iri " j y m z ) ( p i n d i t u r e s ..................... ...................... ExportsiGDP Pnmarycumnt acmunt deficit -8 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: World Bank staff 13. Sluggish growth would broaden the budget deficit, thus further boosting public debt, Slower economic growth would exert pressure on the primary budget deficit, on boththe expenditure andthe revenue sides. . On the expenditure side, the main vulnerability stems from the sizeable government wage bill-accounting for about 56 percent o f primary current public expenditures (about 12.3 percent o f GDP). Because wage agreements span over three-year periods, while employment in the public sector still constitutes a significant source o f job creation (World Bank, 2003), the wage bill remains relatively inflexible and it would likely display inertia in case o f a downtum. The low-case scenario assumes that, eventually, the government wage bill keeps pace with slower average GDP growth, so that by 2012 it remains constant as a proportion to GDP. However, because o f lower wage and employment flexibility inthe civil service, the wage bill increases as a share of GDP during the initial stages o f the downtum, by as much as 1 percentage point compared to the . baseline (FigureI-4b). On the revenue side, tax revenues are generally procyclical as they are proportional to incomes or expenditures. But the progressivity o f some direct taxes, especially the tax on personal income, makes such tax revenues elastic with respect to GDP growth. The share of direct income tax revenues in GDP varies 6 thus positively with the rate o f income gr~wth.~ the low-case scenario the In slowdown spurs a transitory decline in income taxes by about 0.7 percentage points o f GDP, with the tax ratio gradually returning to its initial level thereafter (Figure I-4b). Reflecting the limited flexibility o f current public expenditures and the foregone direct income tax revenues, the primary budget deficit increases by as much as 1.3 percentage points of GDP, before returning to the baseline level by 2012. 14.I n the low-case scenario, total public debt could increase by as much as 20 percentage points of GDP compared to the baseline. Most o f the increase inpublic debt would come in the form o f domestic debt (13 percentage points o f GDP), due to the increase in the primary budget deficit (Figures I-5a and I-5b). Due to the increase in public debt, the burdenof interest payments rises to 4 percent o f GDP by 2012, compared to 3 percent in the baseline. Interest payments would thus end up absorbing nearly 16 percent o f tax revenues, up from 11.5 percent in the baseline. Once again, it should be stressed that the low-case scenario is meant to be a "stress test", associated with a worst- case combination o f unfavorable shocks. A number of intermediate scenarios would be likely under milder downside circumstances. 15. Owing to a significant widening of theprimary current account deficit,foreign debt would become unsustainable. Foreigndebt would increase by about 20 percentage points of GDP above the baseline scenario by 2012 (Figure I-5b). Most o f the increase reflects private foreign debt, due to assumed greater private sector access to international capital markets (Figure I-5a). Larger foreign public debt accounts for about 6 percentage points of the increase in total foreign debt. The unfavorable debt dynamics would exacerbate external vulnerability, with the debt service to exports ratio rising by 6 percentage points compared to the baseline (Figure 1-6). The spiraling foreign debt reflects the unsustainable primary current account deficit, which-contrary to the primary budget deficit-would fail to close down under the low-case scenario assumptions. For this level of the primary current account deficit to be sustainable, growth would have to resume at a steady rate much higher than in the baseline, o f about 8 percent per year (see World Bank, 2001). Reflecting the progressivity of the tax rates, the elasticity of direct taxes to GDP was estimated at 2.3, while the elasticity o f indirect taxes was estimated at 0.9. Corporate tax revenues are also elastic, despite a flat statutory corporate tax rate, owing to the numerous exemptions that reduce taxable corporate income, thus makingthe effective corporate income tax progressive. 7 FigureI-5.Debt dynamics in the low-case scenario Total public debt and foreign private debt spiral up... ...by about 20percentagepoints ofGDP h percentof GDP In percent of GDP 120 1 Lowcase 100 - Total publicdebt 80 - ~ ForeianPnmtedebt 60 30 I 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 1 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 201; (4 (b) Source: World Bank staff FigureI-6.I n the low-case scenario external vulnerability would increase Debt sewce inpenent of total exports ~ Low case ,:. .,.,,'.:,, 16 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: World Bank staff 16. Short of an exceptionally fast resumption of growth, restoring external sustainability would callfor a significant improvement in competitivenessto revitalize exports, along with policy adjustment to absorb part of the financing gap. Policy adjustment would be needed, because Tunisia's external vulnerabilities highlighted inthe low-case scenario may give rise to quasi-permanent shocks, in the sense that the effects o f such shocks could persist for a significant period o f time. Complementary policy options would have to be envisaged on different fronts: Tightening fiscal policy to reduce the primary fiscal deficit, which would further dampen domestic demand and the demand for imports, thus preventing the deterioration inthe current account balance; Tighteningmonetary conditions to further rein inthe growth o f domestic demand and holdback the growth o f imports; 8 Adjusting the exchange rate to offset the loss of competitiveness, due to the structural shifts in the demand for Tunisian exports and the increased penetration of imports inthe context o ftrade liberalization. 17. Policy adjustment would reduce fiscal and current account imbalances, but the effect of different adjustment options on debt dynamics would be uneven. Inparticular, monetary restraint and exchange rate adjustment would also have an effect on public debt. The increase in domestic interest rates would increase interest payments on domestic debt, enlarging the fiscal deficit, and making it harder to reduce the debt ratio. Exchange rate depreciation would raise the burden o f foreign debt and debt service. On impact, a 10 percent exchange rate depreciation would increase the public debt by an estimated 3.8 percentage points in proportion to GDP, while the debt service would increase by 0.7 percentage points o f GDP. Moreover, fresh debt may need to be issued to pay for contingent liabilities related to the foreign exchange guarantees granted by the government to foreign borrowing bypublic enterprises andbanks (see below). 18. The requiredfiscal adjustment toprevent an increase in domestic andforeign debt in a low-case scenario would depend on the parallel use of other policy options. As a benchmark, the fiscal adjustment that would be required to keep the public and total foreign debt constant in percent o f GDP was estimated assuming away other complementary options o f adjustment. Placing the burden o f adjustment on fiscal policy would increase the size o f needed action, but would be free o f drawbacks in terms o f larger interest payments on public debt and higher foreign debt burden associated with other policy options. Intermediate cases o f milder fiscal consolidation could be considered, depending on the parallel use o f other adjustment options. 1.4 Creating morefiscal space would hedgepublic debtfrom downside macroeconomic risks and the contingent liabilities of thepublic sector 19. I n a low-case scenario, optionsfor fiscal consolidation would have to rely primarily on expenditure reduction. The required fiscal adjustment was estimated assuming the concurrent use of three fiscal options: . Holding primary current expenditures constant in real terms throughout the projection, from 2005 to 2012. Primary current expenditures would be reducedby about 2.5 percentage points o f GDP by 2012, with the onus o f the adjustment (2 percentage points) falling on the government wage bill. Reducing capital expenditures by one percentage point inproportion to GDP.' Taking measures aimed at broadening the tax base, to offset the temporary drop in the share o ftax revenues inGDP that would occur ina low-case scenario.6 Reducing public capital expenditures would be facilitated by greater private sector participation in infrastructure, supported by an accelerated opening up of network industries and infrastructure services to competition. This would also enhance non debt creating sources o f external financing-thus relieving pressure from the capital account and allowing a more rapid reduction o f external debt-as foreign direct investment inflows have been greatly boosted by services liberalizationacross developing countries. 9 20.Adjustment to preserve debt sustainability would call for an important fiscal consolidation effort. The required adjustment in the primary fiscal deficit would gradually rise to 3.6 percentage points by 2012 (Figure I-7a). With private sector investment and savings assumed to remain constant in proportion to GDP, the fiscal consolidation would be matched by a decrease in the current account deficit that would prevent the unsustainable rise inthe foreign (private and public) debt associated with the low-case scenario. Reflecting fiscal consolidation, both debt ratios would converge to their baseline levels by2012 (Figure I-7b). Figure I-7.AJiscal consolidationscenario to maintain debt sustainability in theface of external shocks Adjustment would call for a significant fiscal effort.. . ...that would prevent an increase intotal public and foreign debt Differencesto thelowcase scenario; in percentof GDP Inpercentof GDP 4 ____ ~ 80 -1 3 1 Pnmatybudget deficit , 70 - Total publicdebt Lowcase - _J Capita expenditures ---A 1 I Pnmatycuirent expenditures 40 I I -3 - 30 `$ I1 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 (a) Source: World Bank staff 21.Accelerating privatization, and using the receipts to reduce public debt, may provide an additional cushion to hedge downside risks, thus reducing the size of needed fiscal consolidation. However, privatization receipts may be volatile, because they depend on market conditions, and do not provide a sustained source of fiscal revenue. Though this option may provide some relief, it should not be viewed as a viable alternative to fiscal consolidation. 22. Creating more fiscal space would insulate the public debt from the surrounding downside risks. The fiscal adjustment required to stabilize debt levels in a low case scenario underscores the additional flexibility in the budget that would be needed to hedge public debt management from persistent extemal shocks during Tunisia's transition to a more open trade and financial regime. Taking early steps in fiscal consolidation would create enough room inthe budget to prevent an increase indomestic and foreign debt ifthe shocks highlighted inthe low-case scenario were to occur in the mediumterm. Creating enough fiscal space would be appropriate not only to secure debt sustainability in the face o f risks, but also to possibly allow some counter cyclical fiscal Such measures could, for example, come in the form o f streamlining the generous system o f tax incentives to investment, which in 2001 entailed an estimated fiscal cost o f 1.8 percentage points o f GDP (DT 528 million, on account of foregone fiscal revenues for VAT, tariff duties, and corporate income tax rebates-World Bank, 2002). 10 stance to offset part o f the downside impact o f external shocks. Main steps in this direction would call for further strengthening public expenditure management with the aim o f fostering the efficiency o f public expenditures (achieving better outcomes per unit of money spent) and improving the flexibility o f the budget (Box 1-1). Box I-1. Enhancing Expenditure Effectiveness and Budget Flexibility: Mutually Reinforcing Steps to SecureDebt Sustainability in the Face of Downside Risks Promoting greater efficiency o f public expenditures will free up budgetary resources and help create the fiscal space needed to hedge the downside risks to public debt. Improving the efficiency o f public expenditures generally calls for arrangements that facilitate expenditure reallocation towards the most valuable uses, by ensuring that there is no under-provision o f key public services, or over-provision, inthe form o f programs that are unnecessary or fail to adapt to changing circumstances. Greater flexibility in expenditure allocation will make overall public expenditure easier to control, thus increasing the capacity of fiscal policy to meet unforeseen developments and external shocks that may put into risk an otherwise sound public debt management policy. Greater public expenditure efficiency and enhanced budget flexibility are, therefore, mutually reinforcing. Policy options may be considered along a number o f different (but not exclusive) directions: Performance-based budgeting. Efficiency of public expenditures calls for spending public money where it has the greatest impact, within the limits o f the global resource envelop available. This is facilitated by performance-based budgeting, which requires linking reallocation with information concerning the effectiveness o f expenditures, based o n multi-year impacts and program outcomes. Such a framework would need to be underpinned by institutional arrangements between central budget authorities and line departments that facilitate expenditure reallocation towards the most valuable projects. T o this end, appropriate performance indicators should be developed, within the context o f a Medium-term Expenditure Framework. Owing to the complexity o f the issues involved, a shift towards performance-based budgeting would have to progressive. It thus becomes important that steps in this direction be taken from an early stage. A Medium-Term Expenditure Framework. A MTEF would help anchor annual expenditure appropriations in medium-term projections. This would involve establishing expenditure targets within a baseline projection, which would help measure the fiscal impact o f policy changes and reckon the implications of downside risks. The baseline possibly covers three to five years, and is rolled over with each annual budget. Experience suggests that a MTEF helps recognize the implications o f current budgetary decisions for government finances inthe future, thus limiting inefficiencies arising from annual appropriations for multi-year capital projects. The establishment o f multi-annual budgets within a comprehensive macroeconomic framework, based o n realistic projections, would help prioritize projects that shouldbe carried out immediately andprogram those that might be spread over several years. Promoting efficiency in the operation of the civil service. Experience suggests that flexible incentives in the civil service help strengthen management o f the wage bill and improve public service provision. Options may involve more market-oriented and flexible approaches to public sector pay determination; conditions o f employment; redeployment; and staffing levels; with the aim o f creating scope for efficiency gains through reduced operating expenditures and improved public services. Options may also involve contracting out, while the introduction of modern accounting and reporting systems greatly facilitates performance assessment. Source: World Bank staff. 23. Taking early steps in fiscal consolidation would also hedgepublic debtfrom future calls on the budget that may arise from the contingent and implicit liabilities of the public sector. Contingent and implicit liabilities pose fiscal risks, as they may burden the 11 budget directly (when cash payments are required), or indirectly-through higher interest payments, when they leadto an increase inpublic debt. Contingent and implicit liabilities arise from a variety o f sources (see also table in Annex 2): (i) non-provisioned non- The performing loans accumulated by public banks. Non performing loans would be likely to spike in a low-case scenario, owing to the growing exposure o f banks to the tourism industry and to exporting sectors. (ii)The unfunded liabilities o f the pay-as-you-go pension system; (iii)Guarantees for credits extended by financial institutions to borrowers (such as those provided by the "Fonds National de Garantie" or the FOPRODI); (iv) guarantees on the foreign-currency debt issued by public enterprises; (v) foreign exchange guarantees issued by the "Fonds de pCrCquation des changes"; (vi) Performance guarantees to private providers o f infrastructure services, through BOT or BLT agreements, may also give rise to contingent liabilities inthe future. 24. I n a low-case scenario, contingent liabilities may be boosted by the foreign exchange guaranties issued by the government. Foreign exchange guarantees may generate fresh debt if the policy adjustment to persistent external shocks were to partly rely on exchange rate adjustment. This would be the case with the non-provisioned foreign exchange guarantees providedby the government to external borrowing bypublic non financial enterprises. But funded guarantee schemes-such as the "Fonds de pCrCquation des changes", which issues foreign exchange guaranties for loans by local or foreign banks operating in Tunisia-may also generate 10sses.~The Fund's liabilities increased by about 40 percent in 2002, reaching TD 1,4 billion (5 percent o f GDP). Moreover, there is a significant currency mismatch in the Fund's portfolio o f guaranties: about one-third o f its liabilities is inYen, 16 percent in dollar and the rest inEuro, while Tunisia's export earnings are mainly in Euro. Reflecting the Dinar's depreciation, outflows have outpaced the Fund's revenues every year since 2000, and the Fund's reserves have been almost halved since 1999, amounting to only 5 percent o f outstanding liabilities in October 2002. Using market-based instruments to hedge foreign exchange risk would be a superior option, despite the possibly higher cost for the borrowers. This would also promote the development o f better foreign exchange risk management skills and credit culture ina context ofprogressive opening up o fthe capital account. 25. I n the absence of reform, thepension system may soon become a source of implicit liabilities for the government. Because o f Tunisia's relatively young and fast growing population, the fundamentals o f the pay-as-you-go pension system will not become critically unfavorable for some time to go. The dependency ratio for those aged more than 60 is projected to remain constant to around 17 percent o f the working-age population until2010, and then start risingto reach some 27 percent by 2030-still below the strongly unfavorable trends projected for the rapidly ageing societies in the EU. However, owing to the generosity o f the current pension system, financial liabilities are projected to accumulate very soon (Ben Braham, 2002). Contribution periods remain relatively short and replacement ratios are high, close to 80 percent o f the last salary for 'The "Fonds de pe`riquation des changes" is h d e d by commissions on the guaranties issued to the beneficiary financial institutions; foreign exchange gains upon loan reimbursement; a 0.5 percent commission on bank loans included in lending interest rates; a participation in the net revenue of the Central Bank of Tunisia. 12 both the public pensions fund (CNRPS) and the civilian pensions fund (CNSS). A typical employee may expect to recover duringretirement an amount o fpensions about 4.5 times larger than the amount of his contributions inthe case o f the CNRPS, and 5.5 times larger in the case of the CNSS. The urgency o f reform is, however, greater for the CNRPS, because this fund (covering about 30 percent o f social security contributors) i s projected to be in deficit o f more than DT 800 million by 2012-with accumulated liabilities reaching about DT 3.5 billion by then. This financing gap could be covered only partly by the (declining) surplus o f the CNSS. On current policies, the liabilities o f the overall pension system by 2012 are estimated at about DT 1.5 billion (5 percent o f GDP). 13 11. Debt and risk managementstrategy 26. The first Chapter provided a broad sustainability framework for Tunisia's public debt management. This Chapter focuses on Tunisia's debt strategy, evaluating it from two perspectives - macro/fiscal policy perspective and portfolio management perspective. The macro/fiscal policy perspective looks at what the government debt management should do, defining its scope and sustainability goals; the portfolio management perspective looks at how to achieve these goals. Both perspectives are vital, and both should be pursued concurrently. The chapter argues that on both counts Tunisia's debt strategy can be enhanced. To achieve this, senior-level decisions are needed to reshape the formulation andthe implementationmechanisms ofthe sovereign debt strategy. 11.1 Objectivesof Government Debt Management 27. The central objective of public debt management is to finance the budget at the lowest possible cost with the acceptable level of risk over long-term. Both goals (cost and risk) are equally important, and the attainment o f one at the expense o f the other i s a weakness o f a debt strategy. The right cost/risk balance i s achievable only over the medium-to-long term horizon. Otherwise, the choice o f borrowing instruments may be skewed towards shorter maturities, which do have lower costs, but carry higher refinancing risk for the borrower. In the past decade, many governments adopted debt management practices from the corporate world, and sought increasingly standardized and quantitative solutions to the cost/risk dilemma.. O f course, this standardizationo f risk control tools does not mean the standardization o f debt strategies themselves, which remain very country-specific. 28. No single strategy can deliver the correct cosurisk balancefor every country under any circumstances. Substantial differences remain even within one sovereign creditworthiness rank. These can be traced to the constraints imposed by a country's macroeconomic, fiscal and monetary policies, maturity o f its domestic financial markets, its access to international markets and participation in the global division o f labor. Nevertheless, all good debt management practices abide by the following key principles: H The public debt manager's risk tolerance should be low, mirroring risk tolerance o f the taxpayers. The debt manager handles public funds, and may not gamble with them; Consequently, the debt manager would be illadvised to take positions against the market, or to favor seemingly cost-saving short-term debt instruments, as these may turn out to be excessively risky over medium-term; H Equally, government debt manager should not attempt to eliminate risk from the debt portfolio completely, as this would inevitably be very costly, if, indeed, possible at all. 14 29. Over the past decade, Tunisia kad earned its reputation of a cautious and responsible sovereign borrower, which alreadyfollows many of the good international practices in sovereign debt management. This has paid off: the country's sovereign rating has been repeatedly upgraded, and now stands at BBB - one o f the not so numerous investment grade ratings inMENA.Tunisia's financing program i s anchored in the Five-Year Development Plan and i s further detailed in annexes to its annual budgets. From the macro/fiscal sustainability perspective, Tunisia follows a simple but efficient principle, keeping the total amount o f sovereign debt under 60-65% o f its GDP. Its liquidityposition has been comfortableduringmost o f the past decade, with debt service to exports ratio staying below 20 percent, and inrecent years hovering around 15 percent. This is effected with the macro/fiscal policy tools, rather than with the debt portfolio management tools, the latter beingusedonly on selective risks and infrequently, which i s also a good practice. 30. However, the strengthening of sovereign debt management practices isfar from complete. Inorder to preserve and further strengthen Tunisia's sovereign rating reforms inthis fieldofpublic policymustcontinue, embracing allthree keydimensions -policy, institutions andmarket infrastructure. Outstanding issues remainwith regards to the strategic vision and the scope of Tunisia's public debt management; the measurement and disclosure o frisk should be upgraded; and debt management institutions and domestic market infrastructure should be strengthened further. Box 11-1takes stock o f Tunisia's current position, benchmarking its debt management practices against the Guidelines for Public Debt Management developed by the World Bank and IMF, andhighlightingthe major tasks that lie ahead. 31. TheefJiciency of Tunisia'spublicdebt management can be enhanced on both counts -Jiscalfram ework andportfolio management: Although the medium-term fiscal framework exists, it i s too rigid to support active debt management. The central goal o f public debt management currently is to control the overall debt burden; in this context, important medium term considerations may get neglected. Efficient public debt management should rely upon more flexible updates o f fiscal scenarios (at least on an annual, best if on quarterly basis), to reflect the changing composition o f debt portfolio risks. Furthermore, the anchoring o f debt management inthe fiscal policy i s incomplete, as the fiscal framework covers the budgetary sphere stricto sensu, while the risks coming from the broader public sector are not factored into the borrowing decisions. Contractual guarantees and other explicit contingent liabilities must be viewed as part o f the government liability portfolio, and risks coming from them should be factored into the overall public debt strategy; On the portfolio management count, while the Government is aware o f specific risks present inits debt portfolio, it has neither a full picture o fportfolio risks, not a comprehensive approach to control them. Tunisia is, and will remain the price- taker in the international markets, due to the small size o f its borrowing program, and the structural nature of the current account deficit, which makes the country vulnerable to external shocks. Consequently, debt portfolio management must quickly assess and flexibly respond to the changing risk composition. 15 Furthermore, an integrated vision o f public liabilities as one portfolio i s absent, which may distort decisions with regardto handling risks o f individual classes o f liabilities. For example, Fonds de pe're'quation des changes continues to extend coverage for Yen-denominated contracts, despite the fact that on the side o f direct debt Yen risks arebeingswapped away. Box 11-I.Benchmark Tunisia: Progress Towards SoundDebt Management Practice I COMPONENT OF SOUND PRACTICE STATUS QUO AND OUTLOOK FORTUNISIA Ensure that the government's financing and payment Tangible progress, further obligations are met at the lowest possible cost, taking into improvements are possible account the cost-risk tradeoff. Clearly define and disclose the objectives o f the debt strategy T o be done and measures o f cost andrisk. Unambiguous delineation o f responsibilities and close Clearer delineation o f mandates and coordination among the agencies involved (e.g., Central Bank stronger interagency cooperation are and Ministryo f Finance). needed Regular publication o f stock and profile o f debt and financial Impressive record to date, but firther assets including their currency, maturity and interest structure. progress is needed, with greater attention to portfolio structure Develop an accurate, comprehensive and timely management Management information should be information system. improved with risk analysis and scenario simulations Establish code o f conduct and conflict o f interest guidelines T o be done as part o f the broader for staff inthe management o f their financial affairs. institutional reform agenda Take into account the risks associated with foreign-currency Successfulhandling o f some risks; and short-term or floating rate debt. more systematic approach needed Conduct regular stress tests o f the debt portfolio o n the basis Some analysis under medium-term o f the economic and financial shocks to which the macroeconomic scenarios; portfolio government and the country are susceptible. scenarios to be formulated Borrowing decisions should take into account the impact o f Limited analysis; consistent policy contingent liabilities on the government's financial position and control framework absent and its overall liquidityposition. Operations should be consistent with the development o f an Limitedprogress to date, efficient government securities market. substantially more can be done Achieve a broader investor base with due regardto cost and Greater progress o n external than o n risk andtreat investors equitably. domestic markets; further efforts needed Use market-based mechanisms, including competitive Commendable initial steps, more to auctions and ensure transparency and predictability inthe be done primarymarket. I Promote the development o f secondary markets. Major near-term agenda Ensure adequate controls regarding stock o f debt and Strong controls and documentation strengthen documentation. overall, hrther progress needed 16 32.Most importantly, specific risks are not quantified and not managed as one portfolio. The lack o f an articulated risk management approach is only partly due to Tunisia's structural constraints (insufficient flexibility in its access to external markets and underdeveloped domestic markets). This deficiency has more to do with the fragmentation o f the public debt management function (see Section Five). Only some elements o f a guarantee management strategy exist, and these are not integratedwith the management o f direct government obligations; the guarantee risk i s perceived as essentially project risk. Similarly, there exist no specific procedures to control the risk o f the onlending operations. The quality of information available to debt managers and decision-makers can also be improved. 33. Recommendation: In order to increase the efficiency of public debt management, formulate an integrated risk management strategy, based on the rolling medium-term fiscal framework, and considering all risks that affect the government's liabilities and related assets (e.g. onlending) as one portfolio. 11.2 Scope of a PublicDebt Management Strategy 34. The scope of public debt management varies from country to country, but the overall trend of the past decades has been to broaden it. In traditional practice, governments only cared about their direct debt obligations. However, the history o f financial crises demonstrates that a govenunent's debt position i s vulnerable to financial situation inthe broader public sector. Indeed, the financial performance o f different parts o f the public sector is intricately connected - both via budgetary and off-budget mechanisms. A recent study by Kharas and Mishra (2001) suggests that actuarial fiscal deficit' i s a much better explanatory variable for the accumulation o f public debt and the incidence of currency crises, than the conventional budget deficit. While formulating its debt management strategy, governments should not ignore financial operations o f other parts o f the public sector - local authorities, social security system, SOEs, and state- owned financial institutions, even if these are not covered by state guarantees. Active debt management should take into account intricate connections between different parts o fthe public sector. An increasingnumber o f governments view their debt strategy inthis broader perspective - as a fiscal risk mitigation tool, redefining the task o f government debt manager as public liabilitymanagement. 35. Poor management of contingent liabilities has led to significant losses for governments, and many now seek to manage them in a moreprudent and systematic fashion. Some governments have given their debt management units an important role in managing contingent liabilityrisks, often inclose coordination with the Budget Office. In some other cases, Ministries o f Finance have enhanced their monitoring and analysis o f potential fiscal risks arising from financial operations o f local governments, SOEs, other parts o f the public sector and the broader economy. The management o f contingent liabilities requires the adoption o f more sophisticated tools not only by debt managers, Actuarial deficit i s defined as the sumof conventional budgetary deficit and hidden deficits o fthe broader public sector. For further detail see Annex 2. 17 but by the Budget Office, and other government agencies. For example, proper management of the costs o f a guarantee i s impossible under cash-based budgeting, while accrual budgeting and accounting allow to compare a contingent obligation with a direct one. These tools allow the Ministry o f Finance to make an informed choice across the whole spectrum o f financial instruments, and to set up appropriate provisioning levels for both direct and contingent obligation o f the state. 36. There exist, however, certain dangers in extending the scope of public debt management. Both public sector borrowers and their creditors may interpret this behavior as an implicit government guarantee, and may be tempted to practice adverse selection, favoring more risk-laden operations and weaker prudential practices. It i s never possible to eliminate moral hazard altogether, but good practices contain it with clear separation (via legal, operational and informational means) o f the government's own fiduciary responsibilities and those o f the broader public sector. For example, there may be a provision in the law that the central government is not liable for the default o f any other public sector entity with a separate balance sheet, unless the debt in question is covered by a sovereign guarantee. Inreality such "Chinese wall provision" is never fully enforceable, due to the ownership and control relationships within the public sector. The second best solution i s for the government to have a comprehensive ex ante monitoring system, comprising all forms o fpublic sector debt. 37. Public debt management could consist of three "concentric" activities: active management o f the government contractual obligations proper(comprising direct debt and guarantees); government authorization and monitoring regime for other types o f public sector liabilities not covered by the sovereign guarantee; and primarily analytical follow- up o fthe implicit liabilities o fthe government. This concentric strategy canbe illustrated bythe fiscal riskmatrix presented inAnnex 2, with items to beunder direct management regime highlighted with blue, items for which an authorization regime i s more suitable highlighted with yellow, while other items mostly require good monitoring and analysis. Comprehensive control o f fiscal risks present inall three concentric areas can be done by the government as a whole, with the debt management office taking the lead on the inner circle o f direct government debt and guarantees, and coordinating closely with other parts o f the government responsible for fiscal and macroeconomic policy, and with the Central Bank, in order to better delineate the goals o f government debt management and monetarypolicy. 38. For Tunisia, like for other countries which consider gradually liberalizing the capital account, another relevant question is the impact of private sector borrowing on the government's creditworthiness. Once the cross-border movement o f capital is liberalized, external debt o f Tunisian private sector may rise significantly. While it is not advisable for the Government to manage liabilities beyond its own balance sheet, close monitoring and analysis o f non-guaranteed external debt would be needed, both on the macro- andmicro level. Onthe macro level, since private borrowing operations may have a lasting impact on the balance o f payments, debt sustainability scenarios conducted by the Treasury should include private non-guaranteed debt. This means that good detailed information on this class o f external debt should be available to the CBT and the Government o f Tunisia. On the micro level, any large borrowing operation of a private sector entity i s likely to affect the government's own access to external markets, since 18 any unusual bunching o f issues from one country will inevitably push up the cost o f borrowed funds for everyone, including the sovereign. Even top-rated sovereign borrowers such as Denmark reserve the right to advise its private sector borrowers on the timingo ftheir debt placements. Finally, since investors tendto perceive any government monitoring and analysis o f private debt as a sign o f implicit guarantee, the Government and the CBT must practice a firm policy o f no bailouts to troubled private borrowers. Moral hazard and fiscal costs are minimized only if the boundaries o f government responsibility are clearly marked and do not extend beyondits direct debt and guarantees. 39. Tunisia's public debt management strategy could benefitfrom broadening its scope and from making itsformulation and implementation more cohesive. So far, the debt management's scope is largely confined to direct debt, whose monitoring i s comprehensive, but i s done only on nominal basis, with no market valuation attempted. Precise valuation o f direct debt portfolio i s impeded by the delays with which loan disbursements are reflected in the central debt database (SIADE). Likewise, the probabilistic analysis o f guarantee portfolio i s absent, which prevents extending the scope of public debt management to this class o f public liabilities. No attempts are made to value and manage risks o f other off-budget liabilities, such as those generated by social protection and the pension system and the broader public sector. 40. Recommendation: Focus the government debt strategy on the contractual obligations o f the government, but anchor it in a comprehensive view o f fiscal risks arising from the public broad sector and the economy as a whole. 11.3 Strategic Benchmarksfor Public Debt Management 41. Balancing cost containment with portfolio risk mitigation points to the need for strategic benchmarks for the public debt portfolio. In practical terms, benchmarks can be defined as market-neutral compositions o f debt portfolio, which allow the debt manager to avoid excessive risk. Market neutrality is the key feature o f a government debt benchmark. It is unrealistic to assume that government debt managers possess superior information or judgment compared to that o f other market participants and/or able to transact more efficiently than the latter, which i s required for being able to lower borrowing costs without incurring more risk. 42, If the government is viewed as manipulating the market, trust will evaporate, and investors will not be willing to reward the issuer by accepting lower yields. Even on the domestic market, where a sovereign is often the largest issuer, and the regulator o f the financial system, an opportunistic stance may undermine the strategic priorities of developing the market and the sovereign issuer's investor base. The markets' perception would become even more risky if interest rate or currency positions taken by a government were to signal its view on the future direction o f interest rates or the exchange rate. An opportunistic public debt management may thus severely undermine the implementation o f monetary policy. In foreign markets, government debt managers have no "edge" on other market participants. Inrecent years, several governments and/or central banks (e.g. Thailand) have lost billions o f dollars in the derivatives market. Among the OECD countries, experiences are more diverse, with some actively managing 19 their foreign currency debt inorder to generate savings, or to allow their debt managers to accumulate greater market knowledge; other governments refrain wholly from tactical trading. Taking a position on the future market movement contradicts the fundamental principle that the government debt manager's market stance should be risk-averse - aligned with the expectations o f the taxpayers.' Among the minority of OECD sovereigns undertaking active debt management, the positions taken are controlled by strict limits and represent a small proportion of the total debtportfolio. 43. An active debt management stance should not be confused with high volume of transactions which constantly "chase" the benchmark. Active debt management calls primarily for a comprehensive, strategic vision o f risk faced by a sovereign borrower. Periodic rebalancing o f the portfolio i s advisable only to the extent it serves this strategic vision. The more aggressive the portfolio "churning" aimed at risk-mitigation is, the higher i s its cost. Finally, data required to measure the effectiveness o f a proactive operational stance may be fragmented, andor unavailable, making it difficult to assess the true cost o f frequent portfolio rebalancing. Recently, some o f the most proactive debt managers inthe OECD, such as the Swedish SNDO and IrishNTMA heralded a "return to the basics" - to the strategic dimension o f debt management. Indeed, a government may gain more (interms o f reducingthe long-term risks o f its debt portfolio) from further strengthening the link o f debt management with its fiscal policy, thus ensuring debt sustainability. For a small sovereign issuer such as Tunisia, the challenge is not to choose between merely taking the market's terms and aggressive trading, but to define a set of strategic market-neutral benchmarks which will help reducing its overall risk exposure. 44. Thefunctions of a debt benchmark include: incorporation of strategic objectives, limitation of risk, and measurement of performance. International best practice stresses five mainprinciples for buildingportfolio benchmarks: Robustness: Benchmarks should be tested and be efficient through a large number o f market scenarios, with clear and narrow limits for opportunistic market operations. Benchmarks should have as little as possible reliance on assumptions about the future economic and financial environment. It i s wrong when a certain currency mix o f the foreign debt portfolio is based on certain assumptions about future changes in the value o f different currencies or the proportion o f floating rate debt is changed based on assumptions about future yield movements; Long-term horizon: Benchmarks should be defined for the long run, preferably .over the lifetime o f government debt. Decisions targeting short-term portfolio gains may increase the risks and costs inthe longrun. Efjciency: provide the lowest cost for the chosen level o f riskeffectively guide decision-making in respect o f tradeoffs between expected cost and risks, taking into account the government's risk preferences and important macroeconomic policy objectives; The World BankJIMF Guidelines stress that the "debt managers who seek to manage actively the debt portfolio to profit from expectations o f movements ininterest rates and exchange rate.. should be aware of the risks involved and accountable for their actions. These risks include possible financial losses, as well as conflicts o f interest, and adverse signaling with respect to monetary and fiscal policies." 20 9 Transparency: - understood and followed by the common sense. and clearly specify the risk parameters within which the portfolio should lie. 9 FeasibiZity/reaZism: the benchmark should reflect the structural and institutional constraints o f a sovereign issuer - the maturity o f domestic debt markets, capacity o f the Debt Office, structure o f the country's financial flows, and relative role o f and compatibility betweenthe domestic and external capital markets. These factors define what i s feasible when the benchmark i s set up; it should not be sought for the sake o f doing it. 45. I n recent years, several simple methodologies were developed that simplified the construction of market-neutral performance benchmarks .Two examples are Cost-at- Risk, formulated by the Danish National Bank, described in Box 11-2, and Budget-at- Risk, usedby the ItalianMinistryo f Finance. Box 11-2. Cost-ut-Riskand Budget-ut-Risk In Denmark, Government Debt Management applies a Cost-at-Risk (CaR) model, which is a stochastic simulation model for the analysis o f trade-off between expected costs and risk in government debt portfolio. Inmethodological terms, CaR i s related to VaR, a risk-management instrument which is widely used by corporate portfolio managers, and which expresses the maximum loss in a portfolio's market value with a given probability over a given horizon. CaR expresses the maximum increase inthe annual interest payments o n the debt with a given probability inthe medium and long term. CaR entails quantification o f the risk, depending on the probability distribution o f the future market development. The purpose o f the CaR analysis i s to assess the differences between various strategies in the longer t e q i.e. the strategies' average characteristics. The risk related to short-term fluctuations in market interest rates is countered by e.g. spreading borrowing across the year. This reduces the government's exposure to short-term volatility. In CaR model, actual quantification o f the risk requires assignment o f different probabilities to the cost scenarios, allowing to quantify the trade-offbetween costs and risk. The basis i s information o n the existing debt portfolio (outstanding amount o f all government securities and swaps) and the government's expected future budget surplus. Another input into the model concems the strategic focus o f the government debt policy, including the distribution o f future borrowing on various maturity segments, the frequency at which new securities series are opened, and the volume o f future swaps and buy-backs o f government securities. This information gives the government's current borrowing requirement, the distribution o f new borrowing o n various securities and the government's interest-rate exposure. New redemptions and interest payments are subsequently included in the future borrowing requirement. The interest costs are calculated on the basis o f simulated zero- coupon-yield curves. The yield curves are applied in the model to determine coupon rates for new loans, the swap interest rates, and prices for bonds subject to buy-back. The calculations are made for 2,500 scenarios. Each describes a particular development in interest rates on a quarterly basis over 10 years, and thereby a specific scenario for the development in the interest costs o n the debt. On the basis o f the 2,500 scenarios, a probability distribution for the annual interest costs during the simulated period can be determined. The expected future annual costs o f a given strategy are calculated as the mean value of the calculated costs, The risk i s summarized in two measures: absolute CaR and relative CaR. Absolute CaR for a given year states the maximum costs with a probability o f 95 per cent. Relative CaR i s the difference between absolute CaR and the mean value. Relative CaR is thus a measure o f the maximum increase incosts fiom the mean value for a given year, with a probability o f 95 per cent. Source: DanishNational Bank (2003) 11.4 Risks in Tunisia's Debt Portfolio 46. Rollover (refinancing) risk remains the key concern in emerging markets. While all types o frisks are present o f all debt management environments, a fundamental difference between emerging and mature markets is that in emerging markets the rollover risk 21 dwarfs all other types o f risk. Rollover risk defines an emerging market, which is classified as such because it does not possess sufficient depth and flexibility, and thus does not offer the borrower the ease o f refinancing. Volatility is typically higher in an emerging market, which further constrains a govemment in its refinancing options. Rollover risk may cause a dramatic increase o f interest cost to the budget, and in the worst-case scenario it can lead to sovereign default. Other types o f risk, particularly currency risk, often act in concert with rollover risk to precipitate or exacerbate a crisis. Measures aimed at increasing the average maturity o f the public debt portfolio and avoiding bunching o f maturities (using buy-backs and exchanges) decrease the rollover risk; however, in recommending such measures, tradeoffs between different risks much be carefully considered. For example, reducing rollover risk by using indexed instruments can significantly increase market risk, and worsen, not buttress the debt sustainability. 47. Tunisia is perceived by investors as one of the safer emerging market borrowers, but her rollover risk is still tangible. For external debt, the simplest measure o f rollover risk - the share of short term debt in the total debt portfolio - confirms an earlier observation that government debt strategy has been very prudent: this indicator peaked in 1996-1999, and i s now well below 15 percent". Except for short periods the share o f short term debt in total debt was substantially lower than MENA average (Figure I-1.a). However, Figure I-1.b makes also clear that Tunisia has little choice but to be prudent: due to its current account vulnerability, its ratio of short term debt to international reserves is, on average, substantially higher than average for other MENA (30-year averages are: 52 percent for Tunisia but 21percent for other MENA)-although this ratio has declined steadily during the 1990s.' Tunisia i s less well-equipped to cope with short- * term external shocks and Figure I-1.b also reveals a cyclical nature o f short term debt indicator if measured as a percentage of foreign exchange reserves. This suggests that rollover risk i s still very much present, but that it is primarily related to weak current account position andnot to financial difficulties as such. loThis analysis focuses only on government and government-guaranteed debt. For short term debt, a DRS definition is used - debt with an original maturity o f less than one year. This definition may understate the rollover risk, since it does not include longer-term debt which fully matures in the coming year. World Bank DRS data. Other Bank and Fund sources report 1996-2001 levels that are 2-3 percentage points higher than reported by DRS, [possibly since the latter numbersinclude maturinglonger-term debt. Morocco, Oman, Syria, and Yemen, and Non-MENA Countries (NMC) - Chile, Dominican Republic, The two comparator groups are: Other MENA - Algeria, Djibouti, Egypt, Iran, Jordan, Lebanon, Hungary, Malaysia, Pakistan, Philippines and Romania. The composition o f groups i s discussed in Annex 4. 22 Figure &I. Short-term debt aspercent of (a) total external debt and (b)foreign reserves I130 / I 150% ~ 25 20 100% 15 10 50% 5 0% 1870 1975 1980 1985 1990 1995 2000 1 (a> Notes: indicators for both groups are simple averages. Source:The Treasury of Tunisia, World Bank Debtor Reporting System(DRS). 48. Average maturity of Tunisian external debt has remained broadly stable over the past 20 years. It increased by less than one year in the 1990s compared to the 1980s (from 16 years to 16.8 years), and in both decades remained close to MENA average. This indicates strong control over the external rollover risk. The success o f Tunisia's prudent approach to debt management is most vividly demonstrated by the extension o f average maturities it commanded on the private markets: while for other MENA countries these maturities have declined (from 9.6 years in 1980ies to just under 8 years inthe 1990ies' strengthening creditworthiness helped Tunisia to extend the maturities of its long-term debt from 9.3 to 12.1 years over the same period (see trendlines inFigure 11- 2). 49. Another measure of rollover risk is the diversification of sources of finance. All else equal, the more these sources are diverse, the less vulnerable a country would be. Table 11-I suggests that overall, Tunisia was less successful in diversifying its rollover risk than some other MENA countries. The share o f private creditors in Tunisia's debt has fallen deeper in 1990s than in the rest o f the region; over time, Tunisia has become more, not less dependent on a small group of official creditors. Source diversification declined innon MENA group as well, but from much more comfortable levels. The only type o f instruments, for which Tunisia's efforts to promote its sovereign risk have paid off were intemational bonds. On the contrary, debt to commercial banks ended the past decade sharply lower, which is generally believed to be a sensible rollover strategy. Tunisia i s one o f the few MENA countries, which had successfully tapped the intemational bond markets over the past decade. Still, the share o f bonds in Tunisia's total external debt i s well below the level achievedby nonMENA countries. 23 FigureII-2.AverageMaturity of Debtfrom Private Sources omer MENA 1970 1975 1980 1985 1990 1995 2000 Note: for definitions o fgroups see Annex 4. Source: The Treasury o fTunisia, World Bank Debtor Reporting System (DRS). TableII-I.Structure of external debt outstanding by creditor Tunisia Other MENA N M C average average Average average average average 1981-1990 1991-2000 1991-2000 1981-1990 1981-1990 1991-2000 Official creditors 69% 78% 62% 73% 43% 59% Private creditors 31% 22% 38% 27% 57% 41% Of which: Bonds 1% 7% 1% 2% 8% 24% Banks 10% 5yo 13% 10% 41% 13% Other private 20% 10% 22% 14% 8% 4% Source: The Treasury o f Tunisia, World Bank Debtor Reporting System (DRS). 50. Domestic rollover risk is higher, despite the fact that demand from Tunisian remains short - about 41 percent o f total outstanding has remaining maturity o f less than investors,primarily commercial banks is stable. The average maturity o f domestic debt one year; new issuance i s skewed towards shorter maturities even more (Figure 111-1). This suggests a considerable exposure to rollover risk. The only way to lower it lies in lengthening maturities o f the traditional fixed rate nominal bonds in the domestic market12. On the contrary, mitigating domestic rollover risk by switching to external funding sources (which arguably offer much longer maturities) would be merely a shift from rollover risk to currency risk, running counter the strategic vision o f Tunisian Government. 51. Tunisia still faces considerable exchange rate risk, despite the relative stability of Dinar exchange ratel3. Characteristically, even much more sophisticated OECD economies always viewed currency risk exposure as very undesirable, and worked to l2This can only be a long-term goal. Over shorter time horizon, incountries where investors have not been willing to invest in longer-dated instruments, the authorities have used indexed instruments. Clearly this exchanges one type o f risk (rollover) for another (market), but in some circumstances may be a valid tradeoff. Chapter 3 discussesissues indeepening the domestic market ingreater detail. l3 Reflecting the policy o f constant real effective exchange rate followed by the Central Bank, average fluctuations o f real effective exchange rate in 1996-2001 averaged only 1 percent over each 12 month period. Among the key factors that threaten the Dinar's exchange rate, Tunisia's persistent current account deficit remains the most important. On the other hand, stable debt levels and low inflation help stabilize the currency. 24 substitute external debt with domestic fixed income instruments. Tunisia, whose economy and financial system are more vulnerable to external shocks, i s also taking this route, but until the substitution i s complete, Tunisia will continue to face exchange rate risk in two ways: (a) possible change inthe value of the foreign currency debt due to the fluctuations o f the exchange rate o f the domestic currency; and (b) cross currency risk reflecting the risk exposure due to the currency composition o f the foreign debt portfolio. 52. Thegovernment of Tunisia isfully aware of significant exchange rate risk present in itsportfolio and takes mitigation measures. The main option to reduce this risk i s by deepening the domestic debt market. Inthis policy, Tunisia's conduct is similar to other developed and middle-income countries, which are phasing out their external debt portfolios; there i s a widening realization that long-term funding costs are always lower in the domestic market. Recent experience of less-developed EU member countries (Ireland, Portugal, and even Spain), which have launched specific market-development programs in the run-up to the EU accession, could be quite useful for Tunisia. More mature OECD economies, which faced balance o f payments problems in the 1990s (e.g. UnitedKingdom and Sweden), coped with them partlyby enhancing their domestic debt issuance programs. Figure11-3. Structure of Tunisia's sovereign debt by creditor,percent of GDP :I 80% 60% 40% 120% I I 1999 2000 2001 10External 0 Domestic ~ ~ 2o02 Note: consolidated central government debt only. Source: The Treasury o f Tunisia, World Bank staff estimates. 53. Due to the limited investor demand in the domestic capital market, financing of Tunisian public debt is still dependent on international borrowing, which remains at about 2/3 of the totalpublic debtportfolio. While the cost o fthese instrumentsmay seem lower than the cost the Treasury has to pay for domestic currency borrowings, external market access for Tunisia remains less than certain. The risk profile o f foreign borrowing instruments is difficult to determine properly, their costs become evident only in the mediumterm, and the Treasury's freedom o fmaneuver incovering the exchange rate risk i s limited. One way to assess the risk o f foreign borrowing would be to undertake devaluation stress tests on the debt sustainability scenarios prepared by the Government, be it for the next borrowing operation or for the five-year plan. Since the domestic swap market i s non-existent the only way to manage this exposure is by changing the issuance structure (i.e. issue more domestic currency denominated debt). 25 54. Cross-Currency Risk appears to be significant. The natural currency target for Tunisia is Euro, which predominates in the structure o f the country's net exports. This Cross-currency risk i s well understood by the Government and the CBT: the Yen component i s routinely swapped away by the CBT. Thus, the main cross-currency risk is coming from the relative shares o f two other large components - USD and Euro. For Tunisia, which receives only small US$-denominated current inflows, the present share o f US$-denominated debt appears excessive (Figure 11-4). Apart from these three major currencies, Tunisia has loans outstanding in other currencies and currency baskets (e.g. SDR). l4 represents approximately 29 percent ofthe total externalportfolio, and This adversely affect the size o f the public debt. O f course, "other" currencies. can also be swapped, but the size o f such operations would not necessarily deliver economies o f scale. The residual cross-currency risk remains substantial. Such cost considerations should be more explicitly factored into Tunisia's debt strategy. Currently, the cost o f swapping Yen-denominated debt i s simply absorbed as inevitable, and is not assessedas part o f the total cost o fthe portfolio. Figure11-4. Currency composition of Tunisia's long-term external debt and tradejlows Currency structureof long-term externaldebt Currency structureof tradeflows (XGS-MGS) hlti CHR 4 % 7 1% USD USD) Arab - 6% 5% Note: currency structure o f debt shown before the Yen swaps. Source: The Treasury o f Tunisia, World Bank DRS. 55. Currency risk can be mitigated both by natural hedges and by specific portfolio rebalancing action^.'^ Since the natural hedges do not entail any costs, this is a highly preferable way for the government to cover a given risk. For example, in other MENA countries, large US$-denominated debt (Table 11-2) better matches their US$- denominated foreign trade because contracts in oil and oil products are denominated in dollars. In Tunisia, however, net exports as a natural hedge may be only o f limiteduse, l4Mostly debt to the World Bank and the IMF. A natural hedge is a situation when a composition o f a country's foreign exchange inflows or stocks or reserves are close to the composition o f currency structure o f its debt service payments and other outflows, or ,and no special actions such as currency swaps are requiredto limit the cross-currency risk. For oil- exporting MENA countries, US$-denominated oil exports provide a natural hedge against risks from US$ component o f their external debt . 26 due to the persistent current account deficit: full advantage o f natural hedge can be realized only inthe situation o f a current account surplus or equilibrium.16 Table11-2. Currency composition of long-term external debt, end of 2000 Tunisia Other MENA Non-MENA U.S. Dollar 30.4 53.2 53.8 Other currencies 24.6 19.3 6.6 Yen 21.6 7.8 17.9 EURO 16.3 10.8 8.1 Multiple currencies 5.8 5.5 13.6 Swiss Franc 0.6 1.1 0.3 Pound Sterling 0.5 1.2 0.3 SDR 0.2 1.1 0.6 Notes: Only public andpublicly guaranteed debt. For definitions o f groups see Appendix 4. Currency structure shown before swap operations. Figures for comparator groups are non-weighted averages. Figures may not necessarily add to 100due to rounding. Source: World Bank DRS. 56. I n March 2003, Tunisiabroke new ground in external debt management, becoming thefirst member country to sign a Master Derivatives Agreement with the World Bank. The Agreement will allow the Government to use a range o f hedgingproducts linked to existing World Bank loans and will assist Tunisia in reducing its currency and interest rate risks via a range o f hedgingproducts offered by the World Bank, including currency swaps, interest rate swaps, caps and collars and, on a case by case basis, commodity swaps. The hedging products offered by the World Bank allow borrowers to use standard market techniques to transform the risk characteristics o f their outstanding World Bank loans. In providing these financial products, the World Bank stands between market institutions and its borrowers, entering into separate financial contracts with each o f them. Tunisia therefore would benefit from financial terms that reflect the Bank's AAA credit rating. 57. To help dejhe an optimal proportion betweenforeign and domestic currency debt, Tunisian Treasury should formulate a currency benchmark. This would reflect the Treasury's capacity to carry exchange rate risk, but also taking into consideration the attractive characteristics o f the foreign currency portfolio (often longer maturity, lower interest level, and much broader investor base). Tunisia's investment rating makes the issuance o f large, medium- and long-term bonds inthe international market an attractive alternative to the domestic issuance even if the issuer has to pay a significant premium over the reference government bonds and, o f course, takes over the foreign exchange risk from the investor. By definition, all foreign currency related issues should be taken into account in this benchmark. Including SDR denominated loans there was significant net l6 The actual potential for natural hedge may be somewhat greater than suggested by Figure 11-4, which demonstrates a sharp imbalance between currency structure of debt and trade flows, since the full current account deficit is lower, as a proportiono f GDP that the trade balance, due to large net inflows on tourism and workers remittances, which are denominated almost exclusively in Euro. The analysis of another natural hedge - Tunisia's foreign exchange reserves was impeded by the lack o f data on their currency composition 27 extemal financing in both 2001 and 2002 leading to an increase o f share o f the foreign currency debt (to about 60% o fthe total). 58. The second benchmark concerning foreign currencies should address the cross currency risk in the foreign portfolio. Tunisia's foreign currency debt i s primarily serviced by the foreign currency cash flows generated by exports, remittances, tourism income and the financial income eamed by foreign currency reserves denominated mainly inU S D and Euro. Since the Yen risk i s already swapped away, the main value of the benchmark would be in gauging the risk coming from other major currencies - USD and Euro, and from the smaller currency components. It is equally important to examine the exchange risk inherent inSDR-denominated and other currency basket loans. 59. Recommendations: . . Review the mix o f extemal vs. domestic finance, takinginto account full extent o f foreign exchange risk; Formulate a currency benchmark to optimize the mix o f extemal/domestic financing; and II Formulate a cross-currency benchmark to manage risk in three key currencies (USD/EUR/JPY); while defining this benchmark, consider the net currency . exposure o fthe government inUSD and EUR; Quantify the cross-currency risk coming from "smaller" and basket currencies, . and ifthe analysis suggests so, phase them out from the foreign portfolio; In taking decisions on swapping away exchange rate risk of individual liability positions, ensure that counterpart ceilings are set at prudent levels. 60.Interest Rate Risk. The Treasury does not have a procedure to define acceptable levels o f risk from floating interest rates. On the extemal debt side, the share o f floating- rate debt has grown steadily since mid-l970ies, from about 15 percent o f the total extemal long-term debt to almost 35 percent inrecent years (Figure 11-5). The choice of fixed vs. floating instruments should be determined by how much volatility in cash outflows the budget can tolerate. Since the overall level of debt service i s not insignificant, and since the budget i s cash-based, the rule o f thumb would be to favor fixed rate debt, which makes interest expenditure more predictable over the medium run, even if it increased the cost o f borrowing program. This precisely has been the approach o f the Government: the share o f variable rate debt was growing but this trend was gradually reversed, and i s significantly lower than in other countries with similar sovereign rating (BBB). 61. Tunisia's exposure to interest rate risk is generally in line with other MENA countries, but is much lower than in other comparable market borrowers. The current market conditions favor locking in historically low fixed interest rates; however, this decisions should be considered carefully, because runs counter the need to maintain market neutrality (discussed above). From this perspective, Tunisia's recent issuance 28 strategy could be revisited, as it continues to increase the share o f floating rate debt, contrary to the trends inother MENA countries. Figure11-5. Share of variable rate debt in external long-term debt 30% 20% 10% Notes: Definition o f comparators are given inAnnex 4; numbers for them are non-weighted averages. Series begin in 1977, when Tunisia issued external debt o n variable rates for the first time. Source: The Treasury of Tunisia, World Bank DRS. 62. Much like in the case of rollover risk, the interest rate risk on the domestic market is significant. Over half o f new domestic debt issuance is under one year and carries floating rates. Indeed, short-term Treasury Bills should be counted as floating instruments, since their yield may change several times ina year. Currently this exposure is dealt with by strict liquiditymanagement policies o f the Central Bank. However, as the domestic debt markets develop, and interest rates become more volatile, large portfolio o f floating debt may hurt the Government's balance sheet. Over the mediumterm, it should aim at minimizing risks for the budget by shifting the portfolio structure towards longer- term fixed rate instruments inthe domestic market. 63. Recommendations: Undertake analysis and formulate interest-rate risk benchmarks for the domestic and foreign-currency portfolios. The benchmarks should be formulated with full awareness o f Tunisia's likely limited and costly access to the interest swap market (as compared to much easier access to currency swaps); take also into account acceptable limits on counterpart risk (see next paragraph). Select new borrowing instruments with explicit reference to the interest rate benchmark; recourse to interest rate swaps only in a limited fashion, if at all, to achieve the desirable proportions between floating rate and fixed rate debt. 64. Counterparty and Credit Risk. Credit risk can be defined as the inability o f the counterparty to perform its contractual duties. In this narrow sense the Treasury faces limited credit risk as its main counterparties are the central bank and the primary dealers. Indealing with the latter the credit risk is limited due to the settlement system based on delivery against payment (DVP). In a broader sense the counterparties o f the Treasury may be different entities in the public sector that are a recipient o f onlent loans or guarantees. Though in their respect the Treasuries powers are limited, the use o f credit enhancement tools (collateral, reserve fund) may limit the exposure to credit risk. On the external debt side, counterparty risk should be assessed for the CBT currency swap 29 transactions. It is understood that the CBT as the fiscal agent o f the Government exercises prudence while choosing its counterparties, but there should exist comprehensive guidelines for this, in order for the Treasury to have a more complete picture o frisk. 65. Another form of credit risk is present in onlending operations, and may also be signipcant. Often, onlending appears more attractive to a government than direct subsidies, since onlent hnds are expected to be repaid. However, experience shows that the recovery of budgetary loans i s typically poor, and has high administration costs. Where these funds are lent to inefficient public enterprises, they skew the playing field against more profitable new businesses. MOFs rarely have the institutional capacity to monitor and manage project risks inherent in the budgetary loans, which are often irrecoverable and thus turn into subsidies. Subsidies and budget lending serve similar economic goals, but the former are preferable on transparency grounds. While on- lending, the government incurs risks typical for a banking institution, which it is ill equipped to deal with. On-lending requires that the government actively manages the credit risk o f debtor enterprises and has a strong recovery process. Efficient solutions are rare; for example, contracting this service out to an agent (such as a development bank) i s also unlikely to reduce the risk. In Tunisia's current situation, where macroeconomic performance i s threatened by a variety o f adverse factors, the government should be wary o f building too large an on-lending portfolio, and in any event should take steps to strengthen its risk management capacity inthis area. 66. Risks of contingent liabilities and other off-budget operations should be better assessed and controlled. Off-budget risks are typically less well monitored and managed by governments, both developed and developing. These risks are not transparent and/or may reveal themselves only over the medium term; often they strike unexpectedly, being triggered by external developments or natural disasters. Such risks force the government to spend scarce budgetary resources to cover the losses o f state owned financial institutions and enterprises, state social support institutions and other public entities, and sometimes even private sector entities, where the failure of such entity may endanger fragile macroeconomic stability. It i s worth noting that fiscal risks may affect both expenditures and revenues, both assets and liabilities o f the government. As a matter o f priority, the Tunisian government should consider strengtheningits budget procedures to cover unexpected losses from quasi-fiscal activities; expand its monitoring system to cover broader universe of fiscal risks; and extend the horizon for fiscal management, which helps to reveal longer-term costs of quasi-fiscal activities. Annex 2 provides further detail onthe policies andinstitutionsthat help to control better off-budget risks on the liability side o f the government's balance sheet. 67.Recommendations: The Treasury should establish and monitor counterparty limits based on the integrated vision o f risk. While the relevant departments responsible for different types of issuance ("Front Offices") may make suggestions as to possible counterparties, the overall system o f exposure limits should be established by the unit in change of debt strategy formulation, i.e. by the "Middle Office". The approval o f the limits should be done by the Minister o f Finance. 30 68. Operational risk should be carefully reviewed. Based on the findings o f the World Bank missions, operating procedures are sufficiently strong and security o f sensitive information appear to be adequate. However, despite the strong culture o f debt monitoring, there remain precedents of different government agencies coming out publicly with diverging debt information, which may negatively affect investor relations strategy and unduly increase the cost of borrowing. Another aspect o f operational risk is related to the effectiveness of interagency communication, planning and coordination; this needs to be significantly strengthened. Chapter 4 o f this report discusses the institutional aspects ingreater detail. 69. To efficiently manage the diverse range of riskspresent in Tunisia's public liability portfolio, the most important measures are: Compile a full view o f different risks and introduce methods, such as accrual budgeting and quantitative valuation o f indirect liabilities, to rationalize procedures for and rebalancing the structure o f risks inorder to prevent excessive exposures andto maintainoverall sustainability o f government finances; . Formulate procedures for proper budgetingto cover fiscal risks; Examine the suitability o f different techniques diversify the fiscal risk o f onlending operations and contingent liabilities and to pass part o f it back to the market. 11.5 Decision support mechanisms: the role of information and analysis 70. A government risk management strategy is only as good as the information used to formulate it. Tunisia has one o f the best information systems among the countries at similar level o f economic development. Databases on individual classes o f government debt are modern, and support well transaction control and standardized reporting. However, there exists no unified debt database; consolidated debt reports are difficult to compile and even more difficult to customize, which makes it impossible to proactively respondto the changingmarket conditions. 71. On the external debt side, the Central Bank of Tunisia and the Ministry of Finance have sponsored the development of a centralized database (SIADE), which can justly be characterized as one of the best practices among the emerging market economies. The system has a modern open architecture design and is supported by a strong dedicated team o f system specialists. SIADE covers all categories o f public debts, including those not guaranteed by the state, as well as onlending operations. It includes a strong forecasting module, which generates loan-by-loan cash flow profiles. The database also supports the preparationo f various aggregate reports. 72. While SIADE comfortably meets most of the transaction management needs, government experts still express concerns about the quality and timeliness of information, particularly with regards to the disbursements under project loans. Disbursement information is supposed to be provided by the sector ministries and entered in the database by its administrators - Centre Informatique du Ministkre des Finances (CIMPF). Sector ministries often delay such updates, which results in SIADE producing 31 less than reliable cash flow forecasts. These concerns, however, can be addressed within the existinginformation management framework. 73. Analytical systems supporting the risk assessment and strategyformulation, which are the Middle Office functionalities, are less developed. The system tools are insufficient to support a more disaggregated view o f portfolio risks. Only nominal valuation o f loan instruments i s supported, and no mark to market valuation is available. Likewise, the guarantee portfolio i s recoded in nominal figures, and no procedures exist for the valuation o f guarantee risk. The report generation facilities cannot easily create customized reports required for more active debt management. Onthe domestic debt side, both information requirements and systems implemented are more basic. Data are kept and core parameters o f domestic debt are calculated insimple spreadsheets. 74. The analytical systems are more basic and do not support the calculation of specific portfolio risk exposures. SIADE does not have such analytical capacities beyond cash flow forecasting for individual loan instruments. The Ministry o f Finance does not regularly prepare debt sustainability scenarios, relying inthis task on other agencies such as the CBT, Ministry o f International Cooperation, and Ministry o f Economic Development. As mentioned above, the existing debt system do not allow to quantify guarantee risk; similarly, no quantitative procedures exist to evaluate the risks o f onlending operations and the risks o f non-guaranteed debts o f the broader public sector. Private external debt, which will have a growing impact in the future is tracked by the CBT, but this information i s not regularlyincluded into the debt sustainability scenarios. Finally, an important deficiency is the lack o f consolidation for the total government portfolio (external plus domestic) except on highest level o f aggregation, and only in nominal figures. Mark-to-market valuation o f the whole government portfolio i s still impo~sible'~. 75. I n order to support the public debt management strategy, the following upgrades are advisablefor government information systems: Ensure regular in-year (at least quarterly) consolidation o f all relevant debt data, including on contingent liabilities o f the government, public sector debts on onlending operations, to allow regular updating o f debt sustainability scenarios andportfolio benchmarks; Implement analytical systems supporting the measurement and forecasting o f different types o f portfolio risks, including those stemming from contingent liabilities o f the government and non-guaranteed debts, to support an integrated asset/liability view o f the government financial position. l7This observation should not be interpreted as a recommendation o f unifying all debt information (both external and domestic) in one database. International experience has shown that such integration is very costly and not necessarily efficient. One off-the-shelf solution advertised recently is a new verision o f CS- DRMS, shipped by the Commonwealth Secretariat, which can register both external and domestic debt. Implementation record o f this version o f CS-DRMS is still too short for conclusions about its efficiency in comprehensive debt portfolio recording. An alternative solution would be to equip the analytical modules with specially designed "data extraction channels" that would query the central databases and generate information inthe format meaningful for portfolio risk analysis. 32 111. Promotingthe developmentof domestic governmentsecurities markets 76. The Tunisian Authorities are aware of the need to develop the market of the domesticpublic debt and have taken important steps in this direction. Should the access of Tunisia to the international capital markets be temporarily restricted, the existence o f a domestic liquid market would be essential to cope with financing stress. Moreover, elements such as the development o f a domestic yield curve, which could be used as reference for domestic issuers, the diversification o f risks and the reduction o f the long- term cost o f the debt also plead in favor o f the development o f a robust and liquid domestic market. This chapter examines more in depth options for the development o f domestic debt markets - an essential precondition for a successful low-risk debt strategy. IILl Status of the domesticgovernment securitiesmarkets 77. The distribution of borrowing between domestic and foreign debt does not only depend on macro-economic constraints (such as the balance of payments deficit), but also on the material possibilities of issuance. As regards the breakdown between the issues o f long-term loans in TD and in currencies, it is conditional upon the refinancing o f the maturing loans, the deficit being in theory financed half on the domestic market and half on the international market. Thus, 922 MDT were issued in 2000 in the form o f BTA ("Bons du Trdsor assimilables" - loans in TD with final maturities over one year) and that 617.4 M D T come from the international financial market, that i s roughly speaking a breakdown o f 60 % in TD - domestic- and 40 % in currencies- international. But the financing needs o f the State on the domestic market are note entirely met, because the Treasury does not succeed in raising on the primary market at market conditions the amounts that it would like to issue or that it should issue. Inparticular, it finds it difficult to issue longmaturities (Figures 111-1a and111-1b). 78. Treasury bond issuance is concentrated in the hands of a limited number of investors, the majority of which arepublic entities. For example, at end October 2002, 80 % o f the securities issued inthe BTA 12-year bond were sold to the CNSS ("Caisse Nationale de Securitk Sociale").'* For the BTCT (short-term Treasury bill- maximum 52 weeks maturity), the situation i s less clear, but the CNSS usually buys a share o f roughly 40 % o f the auctions. This concentration of issues on a very limited number of public sector investors, suggests that the primary market i s somewhat artificial, rather consisting inshiftingpublic funds from one sector ofthe State to another. The bulk ofthe financing needs o f the State is thus covered by a transfer o f capital within the public sector, either The issues inthe 12-year bond(BTA 8,25 % July 2014) represented end October 2002 54 % o f the total issuance o f a maturity over 1year. 33 via auctions or other mechanisms o f provision o f funds, without any real call to the financial markets." FigureIII-1.MaturityproJile of domestic debt and of its issuance in 2002. I Over IOycarS over I O years 7% 17% 7 1.2yearr 22% Note: Data are for October 2002. Source: The Treasury o f Tunisia, World Bank staff estimates. 79. The secondary market is still fairly inactive. In November 2002, the secondary market o f the BTA entered 106 transactions, including 105 in "intra-groups", that is to say simple movements o f accounts within the same financial group. The situation i s even worse on the secondary market o f the BTC: 555 intra-groups transactions, for a single transaction between different counterparts. The secondary market is thus more akin to intra-group transactions o f limited transparency. 80. Theabsence of a secondary marketprevents the emergenceof a yield curve. A yield curve for government securities is key to enabling the "pricing" o f other issuers' loans. It also makes it possible to evaluate the portfolios at market value and assess market risks realistically. The absence of a yield curve hampers more generally the perception by the public o f the value o f a fixed income asset. 81. An active money market is a perquisite for developing a fmed-income securities market. An efficient money market supports the valorization o f liquidity, providing necessary benchmarks for the valorization o f fixed income assets that differ in terms o f maturity, liquidity, and other risk characteristics. That provides the short-run anchoring necessary to the yield curve (Schinasi and Smith, 1998). At the same time, an active money market ensures the refinancing positions and the effective cash management for a number o f participants, such as banks, stock exchange intermediaries, and non-financial enterprises. The money market supports the bond market by increasingbond liquidityand providing bond traders with a means o f fimding their temporary inventories. A liquid money market helps the financial institutions meet their needs for short-term liquidities and reduces the risks and cost o f holding bonds in order to meet an investor's demands l9 Another indicator o f this situation i s the importance o f deposits with the general Treasury. The deposits o f the C.N.S.S. represented 900 MTD at the end o f 2000; the direct deposits with the general Treasury were o f 1,195.4 MTD at the end o f October 2002, that is to say more than double the total o f BTA issues at the same date (580.6 MDT for the issues over 52 weeks). 34 and finance their trading portfolios. By ensuring liquidity, the money market becomes thus a catalyst o f bond market development. Despite some progress, the Tunisian money market is still inhibited by the reliance o f banks on the CBT's refinancing facilities at stable andpredictable interest rates (see below). 82. The reasonsfor the inadequatefunctioning of the Tunisian public debt market are multiple, but can be traced back tofive mainfactors: ....the Central Bank's monetary policy operating procedures and the incentives structure inthe banking system; the narrow investor base; the issuing strategy o fthe Treasury on the primary market; the absence o f secondary market; psycho sociological factors, bynature difficult to apprehend. 83. The various causes of the dysfunction interact and are mutually reinforced. It i s thus difficult to isolate one single element, the reform o f which would lead to a drastic improvement o f the market. Improving the functioning o f the domestic public debt market would call for a coordinated strategy across these five dimensions. Elements o f such a strategy are outlined inthe last section, after reviewing more indepth the nature o f the bottlenecks highlightedabove. 111.2 An active money market is an important underpinning of domestic securities markets 84. The monetary policy framework plays an important part in the emergence of an active money market. A rigid monetary framework, which does not support a sufficient flexibility o f interest rates, hinders money market development. The framework o f monetary policy in Tunisia i s anchored on the defense o f the external and internal value o f the currency. The Central Bank o f Tunisia (CBT) i s responsible for controlling economic activity in order to preserve external balance (thus the currency's external value), while keeping inflation under control (in order to preserve the currency's internal value). With this intention, CBT's policy i s centered on maintaining a constant effective real exchange rate. At the same time, credit growth, which is narrowly correlated with inflation and domestic demand growth, i s used as an intermediate policy target. The existence o f capital controls confers certain autonomy on monetary policy and makes it simultaneously possible to target the real effective exchange rate and domestic credit. 85. To reach intermediate credit growth targets, the CBT relies on the control of the banking system's refinancing volume by using diverse instruments. Weekly liquidity auctions (at the official intervention rate) constitute the main instrument that CBT uses to manage the banking system liquidity. A refinancing facility for a period o f seven days, and up to 100 points above the official intervention rate, permits banks to satisfy some supplementary liquidity needs, when weekly liquidity auctions fall short o f required reserves. The end-of-day operations finally permit a fine tuning of liquidity at a narrow rate, at &1/32 the official intervention rate (World Bank and IMF, 2001). The dependence o f Banks on easy refinancing by CBT has conveyed to the CBT a leading role in the 35 money market. This dependence, coupled with the narrow gap between the rate o f deposits and the rate o f refinancing by the CBT, discouraged the development o f an active interbank market. 86. The absence of a yield curveprevents adequate evaluation of credit risks and leads to the indexing of the cost of credit to the TMM. Owingto the indexingo f interest rates of many deposits and loans agreements on the TMM ("taux du marche` mone`taire"), the changes in the monetary policy are quickly reflected in the financial system, then in the economy as a whole. Ifthis flexibility has the advantage to allow a fast response o f the system to a change in interest rates, it also makes the CBT and the political authorities reluctant to changes in monetary policy, because such changes would quickly be passed on to the economy with the risk o f increasing its funding cost ingeneral and inparticular that o f the "strategic sectors". This has created rigidity in interest rates (Figure 111.2a), along with dependable and easy access to liquidity by the banks at a predictable rate. Therefore, although monetary policy seeks, in principle, to control growth in domestic credit as its main intermediate goal, in practice, monetary policy ends up focusing on interest rates as targets rather than as instruments to achieve the credit growth intermediate goals. FigureIII-2.Money market interest rates and interbank loans hpercentperyear 7 1 - . . .. Interbankloans inpercentof totalbank credit to the economy- 5.6 -Tarade rappeldoffres TMM +Pnse en pereionA 7 p u n 5.4 Dec. Dee. Dec March Apri May June July Auals! 1999 2000 2001 2002 2002 2002 2002 2002 2002 (a), (a) Source: The Treasury of Tunisia, World Bank staff. 87. Although the control of interior credit growth is the intermediate target of monetay policy, in practice, monetay policy implementation results in a limited flexibility of interest rates. Implementation o f monetary policy effectively emphasizes interest rates as targets and not instruments used to reach intermediate objectives o f credit growth. Besides, operational procedures don't rely enough on market instruments like open market operations, which could support the short-term market o ftreasury securities. 88. Thestability andpredictability of the money market rates inhibit the development of an active money market The money market remains narrow, because banks can easily refinance themselves at stable and predictable conditions from the CBT, rather than through the interbankmarket. Since 1999, interbank loans have shown a declining trend as a percentage o f total bank credit to the economy, despite some recovery (Figure 11-2b). 36 The rigidity o f rates also discourages the development of secondary bond markets, where transactions rest extensively on different expectations of interest rate variations*'. 89. I t is important that the structure of the banking system provides adequate incentivesfor banks to activelyparticipate in the bond market. Banks which have short- term liabilities should play an active part in the short-end, rather than in the long-end of the bond market. But as the experience o f developing countries, including Tunisia, has shown, if the banking system is dominated by state-owned financial institutions, incentives are not conducive to the development of active money markets. Thus, technical measures to promote the development of active money and bond markets may not pay off, unless there is parallel progress inprivatization andbank restructuring (Box 111-1).InTunisia, even thoughthere has been some progress inbank participation inthe money market, it has been long in coming, and it may b e related to the successful privatization o f some of the smaller banks. Box III-1. Therole of bank incentives in the development of an active bond market Experience suggests that, although many developing countries have succeeded in bolstering the domestic bond markets, others have been less successful as, in a sense, they tended to put the "cart before the horse" intryingto promote the market. Reform efforts often focused on technical issues (such as creating primary dealers and signing detailed duty agreements) or on establishing an institutional structure (creating a debt office within or without the central bank), but failed to address the more fundamental issues o f possible conflicts between monetary policy and debt strategy and the incentives facing potential market participants. Of course, the technical and institutional aspects are very important. It would be wrong to ignore them. But the fundamental issues must also be addressed. The incentives of potential market participants are linked to the dominant presence o f state- owned institutions in banking, insurance, and social security. Maintaining some moderate state presence in banking and insurance may have some merit in terms o f promoting certain financial policy objectives or delivering broad-based financial services. However, experience suggests that the managers o f state-owned institutions do not have strong incentives to develop an active presence inthe money and bond markets. They do not favor innovation and competition. Usually their performance i s not evaluated on the basis o f the profitability and solvency o f their institutions. Of course, one or other state-owned bank could replicate the organizational structure o f a private institution and encourage its managers to support an active presence in money and bondmarkets. But it i s difficult to see all state-owned institutions doing so. This raises serious doubts that efficient money and bond markets can be developed in a financial system that i s dominated by state-owned institutions. To `some extent, this is due to the way monetary policy i s conducted. Indeed, the willingness o f the central bank and the authorities to provide accommodation to banks on demand i s not independent o f the fact that it i s the very same public banks that are inneed o f such accommodation. *'Reflecting the stable and predictable conditions of refinancing at the central bank, banks become somewhat less concerned about the quality o f debtors, especially since granted loans could be used under certain conditions as collateral for refinancing from the CBT. That can have consequences for the stability o f the whole banking system. 37 90. A number of other institutional rigidities discourage the emergence of an active money market. Main factors include: m The absence o f legislation on the repurchase agreements ("repos"). A draft o f regulation on the matter was recently approved (11/12/2002) by the Council o f Ministers, but it has yet to be implemented; The prohibition for the banks and other issuers to issue short term commercial paper was recently abolished (authorization was granted to issue certificates o f deposit at 10days), but this possibility should be extended; The CBT accepts other collateral than Government paper for its refinancing operations. Recently, the loans on certain high-quality debtors, belonging to the so called "strategic sectors" were accepted as collateral. Even though emphasis has recently shifted, from sectoral considerations to the creditworthiness o f the debtors, this practice may not favor the emergence o f an active government securities market since private credits are regarded as having a quality equivalent to sovereignpaper. It i s necessary to be sure that admissible collateral is not only o f good quality, but that comes from debtors who meet strict transparency and auditing criteria. There i s also the question o f the limit imposed on the central bank for the holding o f government securities (10 percent o f Government tax revenues o f the preceding year), with the aim o f limiting the risk o f debt monetization. This ceiling may inhibit the implementation o f monetary policy based on open market operations. The rigidity andpredictability o f the money market interest rates has prevented so far the banks from developing their expertise as regardsliquidity management. 111.3 Theprimary marketfor government securities 91. The technical organization of the primary market in Tunisian public debtfollows international bestpractice. The Tunisian Treasury has indeed very largely inspireditself from the French experience-which, in tum, drew largely on the US model, whose success i s well known. However, the primary market suffered from an initial weakness because the systemo f primary dealers in government securities ("Spe`cialistes en VaZeurs du Tre`sor"; SVT), to whom was granted, at the beginning, the monopoly o f participation to the auctions of the Treasury, was set up with the brokers only. Indeed, the banks showed little interest for a reform o f the public debt market since the Treasury's former method o f issuance ("Bons du Tre`sor cessibles ") suited them well and they were no willing to change. 92. Implementation of the reform led to stiff competition between the SVT and the banking sector. SVT were distributing the public debt to investors, while banks were proposing other more remunerative products. The banks, as it should be, won the competition, taking advantage o f their stronger capital basis and a better knowledge o f the final investors needs. The SVT thus withdrew, one after the other (only one or two 38 still take part in the auctions), since they could not turn profitable the heavy investments they had to make to acquire the status o f SVT. Thus, the Treasury had to call on to other market players, but the banks turned out to be less enthusiastic, since the reform o f the domestic public debt market was initiated without their collaboration. Box III-2.Primary dealers in government securities markets The experience of the "developed" markets suggests indeed that final investors do not participate directly in the auctions The reasons for this are basically two : (I) secondary market usually drops at the the moment o f the auctions, so that it i s less expensive for final investors to buy on the secondary market than at the auctions. The banks have the same concern, butthey are more o f less forced to buy : Primary Dealers are "remunerated" (in several different ways) in association with their presence inthe markets, so they have an incentive to buy.It i s said that some bigbanks have even a budget to "buy market shares"; (II) the timing and the auctioned bonds o f the Treasury's auctions do not match per se the investor's needs and strategies. The typical scheme i s as follows: the "primary dealers'' participate to the auctions and then sell to final investors o n the secondary market the securities they acquired o n the primary market. However, only the financial intermediaries that have a sufficient financial standing, the banks, can afford to take o n their books the securities they acquired at the auctions. The brokers, by definition, do not have this capital. Primary dealers (S.V.T.) in Western Europe are quasi exclusively banks. This is true for Belgium, Netherlands, Portugal, France, etc. Italy accepts as primary dealers some brokers, with a special legal status. These brokers, however, are only a few, and are disappearing, due to the consolidation o f the Italian financial sector. 93. The Treasury, despite its stated willingness to conform to the rules of the market in order to meet the government'sfinancing needs,finds it hard to build its credibility as an issuer. A first reason is the perceived lack o f transparency o f Treasury issuance. The issuance calendar i s published, but it undergoes many changes, such as cancelled auctions, sometimes for a long period. There are objective reasons for these changes, such as insufficient offers o f the bidders at the auctions, or the timing o f the international issues. However, lack o f transparency and predictability makes the participants uncertain about the issuing policy. Establishing transparency and credibility calls for making intentions known to the market and for conforming as much as possible to the announced plans. The Internet could be an appropriate means for the Treasury to publish its auction calendar and, before each auction, the range o f amounts inwhich it intends to auction as well as other information relevant for the market .To be sure, the Internet i s only a means o f communication. There are others. What matters i s the result: communication 21. As a rule o f thumb, to enhance its credibility, the Treasury would have to issue debt on a regular basis, not when it needs, but when it can-respecting its commitments in the auctions' calendar. 94. Owing to the absence of a yield curve and thus of reliable benchmarks, theprocess of "price discovery" is random. This undermines the credibility o fthe Treasury-who i s also reluctant to issue at yield levels that it feels exceed anticipated market levels, thus preventing bidders from making offers in line with market yelds.. A systematic 21 Itis highly desirable that the Treasury have its own Internet site, whether integrated into the Ministry o f Finance's site or not; in any case, the Treasury must independently control its own content and its regular updates. 39 difference o f appreciation with the market leads the Treasury to allocate amounts much lower than those preannounced, thus damaging its transparency and its credibility. The Treasury's stated main concern is to limit its funding costs and not to drive the other issuers out o f the capital market. However, reluctance to pay the price asked by the markets casts doubt about the declared will to play the game o f the financial markets. Of course, before adhering completely to market rates, the Treasury should make sure that sufficient competition from market participants prevails in its offerings. It is, therefore, important to enhance regulatory safeguards-including penalties in the SVT duties agreements-that prevent collusive behavior in the market. A formal protocol o f collaboration between the Central Bank and the "Comite' du Marche' Financier" (see below) should consider monitoring the participation o f the banks inthe auctions to avoid possible collusion. 95. As the Treasuy is reluctant to issue at the levels above those deemed appropriate, bond auctions meet little success. The Treasury encounters particular difficulty to issue inthe long-term end ofthe yieldcurve. For example, up untilOctober 2002, issuance for the year inmaturities longer than 12 months amounted to TD 580.6 million, against TD 1,267.3 million inmaturities equal or lower than 12 months (Figure 111-lb).As discussed earlier, long-term issuance relies heavily on a few public sector entities, such as the CNSS. As a consequence, the primary market remains still artificial, since on the long- term segment the final investors are mainly State institutions, while on the short-term segment the State i s still very present-with the balance o f the issues taken by the banking system, which needs government paper to participate inthe liquidity auctions o f the centralbank. 96. Developing an efficient domestic public debt market is a long-term endeavor and entails a start up cost. This is suggested by the experience o f countries which acquired the status o f a highgrade sovereign issuer, along with the associated advantages interms o f issuing conditions, reduction o f public debt cost and o f the financing cost in the national economy as a whole. The benefits come only later on, when the credibility o f the Treasury as an issuer has been well established. Political will is needed to accept the short-term cost o f a strategy with a medium-term pay off. Inthis respect, Tunisia's good economic fundamentals and medium-term growth prospects, allow sufficient flexibility to set up significant reforms under favorable conditions without having to fight against time nor being confronted with insurmountable budgetary needs. This element is nonetheless precious andit i s important not to waste it. Box 111-3. Thecosts of a reform strategy of domestic debt markets An essential element o f a reform strategy is the cost o f the reforms. The experience of the countries o f Westem Europe suggests that the implementationof a performing public debt market has a cost at the start. T o take the example o f Belgium, at the time of the reform o f the domestic public debt market end 1980, begin o f 1990 (introduction o f the OLOs - linear bonds, Belgian equivalent o f the French OAT and the treasury certificates, equivalents o f the French BTF), the Belgian Treasury financed itself during long months at Bibor +, that i s to say at conditions less favorable than those obtained by the domestic banks between themselves. There are reasons for such a situation: the banks consideredthe Treasury as a "customer" like any other one and were not ready to grant it preferential conditions. And it is only with the gradual standardization of its issues, the concretization in the facts o f its commitment to play the game o f the financial markets, its transparency, the importance attached to the liquidity and well targeted "marketing" steps that the Belgian 40 Treasury built its credibility as a sovereign issuer (with as a consequence that the last 10 year benchmark bond has been issued in January 2003 (5 billion euros) at Bund + 16 basis points- Belgium comes from levels of roughly Bund+ 300 bp at the beginning of the 1990s). Source: World Bank staff. 97. The primary market is also inhibited by some institutional weaknesses. Most importantly, the participants to the auctions (SVT, banks) are not remunerated for their efforts. In accordance with the French model, the principal advantage o f the SVT is the authorization to introduce (in certain limits and according to certain conditions) non- competitive bids, after the auction, at the weighted average price o f the auction. However, taking into account the current situation (absence o f a yield curve, lack o f a secondary market, restrictive issuing policy), this advantage i s rather formal and has little interest for them, so that inpractice the Treasury has no incentive to support the success o f its auctions. An additional factor o f weakness o f the primary market i s the importance of deposits within the "trksorerie gknkrale "-TD 1,195 million at end October 2002, that i s approximately half o f the total BTA outstanding. While this issue was not investigated thoroughly, part o f these hnds could be injected on the primary market and contribute to improve the issuance . 98. Moreover, the rapidity of the systems of bids to the auctions and of communication of the results could be improved. The current procedures are still somewhat slow, while practices such as bids to the auctions by mail could increase trading risks. Due to the lack o f a performing secondary market, the situation is not too alarming at present but one should be aware that an upgrade o f procedures and techniques must be considered quickly so that these deficiencies do not weigh on the development o fthe market. 111.4 The secondary marketfor government securities 99. The secondary market of the Tunisian public debt is inactive, with absence of a yield curve and thus no reliable benchmarks for "rice discovery". This handicaps all the issuers, including the Treasury. Due to the absence o f a yield curve, portfolios, investment funds, etc. are not valued at market value ("marked to market"), disregarding the legal provisions on the matter, but at face value. Moreover, the "yield" i s calculated on the basis o f the rate of the coupons, or with other ad hoc methods, rather than on the basis o f the actualizationo f cash flows. 100. It exists neverthelessa `@parallel"(over-the-counter, OTC)secondary market, in the sense that there are transactions between different entities within firzancial groups. This parallel secondary market lacks transparency, since the transactions in question are carried out inside the same financial group in an over-the-counter market (OTC), non supervised by any particular authority. It is thus hard to establish the transparency o f this OTC market. Transactions between banks or SVT are obviously rare, owing to the opacity o f this parallel secondary market, which prevents the emergence o f a true yield curve. The financial conditions o f these intragroup transactions might also be questionable, because of the legal relationship between various group entities, and their underlying reasons, often more for accountancy and balance sheet matters than for purely financial ones. Nevertheless, it will be noted that the Treasury has started to receive a 41 detailed list o f these intra-groups transactions, which could behelpful as a basis o f a yield curve. 101. The transparency of this parallel secondary market should be improved by appropriate initiatives. The institutional situation poses some problems. Banks are subject to prudential control by BCT, something lacking in the case o f transactions carried out in the OTC market. As for CMF, the Article 16 o f the Convention between banks and the Treasury could provide a basis for the control o f these transactions. However, this opportunity remains unexploited for lack o f an agreement with the CBT, that would allow to implement this control. So there is a gap in the law supporting intra- group transactions. 0 as a first step, the Treasurer should publishdaily inthe media the yield curve that can be established from these intra-groups transactions. 0 Then, it would be necessary to establish a general obligation of transactions reporting (all transactions, intra-group and OTC market transactions) to a "reporting" authority in charge o f establishing market transparency, since "reporting" to the Treasury i s only an imperfect option. e As the prerogatives o f the reporting authority only involve receiving transaction reports and establishing market transparency, with virtually no control over participants, it is probably unproductive to establish a new body-especially given that the Tunis Stock Exchange has both the material infrastructure and the requisite knowledge to perform these functions. 102. Once the general obligation to report transactions has been introduced, supervisors will have to ensure that it is enforced by all players in the OTC market. In this regard, there must be an agreement between CMF and CBT, in view o f the institutional gaps mentioned above. This agreement should include, at least (i) clear a distribution o f competencies between CBT and CMF with regard to verifying the integrity and content of reporting and (ii) procedure o f communication between a authorities concerned about any "irregular" transactions, and the attempt to elude transactions reporting. The creation of a "Market Authority" for these OTC transactions could also be considered, but an effective agreement between BCT/CMF and the reporting authority will make the deal easier if the creation o f yet another structure i s to be avoided. Furthermore, there is the issue o f identifying the market participants subject to the obligation o f transaction reporting. The question merits carefully study, but, at the very least, transaction reporting would be essential for the financial intermediaries who act as counterparts, brokers, agents or commission agents. 103. The obligation of quotation is not constraining enough. It is limited to transactions above TD 20,000 and even at this level, it i s not always respected. One should be aware that the development o f the secondary market depends on the capacity o f the financial intermediaries to bring liquidity to the market and thus to quote prices. If they do not quote, there cannot be a market. Implementation o f an effective quotation requirement must go hand-in-hand with the establishment o f OTC market transparency, since secondary market liquidity will be built on the capacity o f players to propose competitive market prices, thereby ending the phenomenon o f captive intra-group transactions. 42 104. The limited amount of securities outstanding do not support the development of the secondary market. International experience demonstrates that the degree o f bond liquidity is largely related to the amount o f stocks outstanding. It is recommended to target a critical mass in bond issuance in order to create a liquid Treasury bonds secondary market. But the BTA issue with the most significant stock outstanding as o f October 2002 was the BTA 3/2009, with an outstanding amount o f TD 509.5 million. Follows the 10/2004 issue (TD 483.8 million) and the 7/2014 issue (311.1 MDT). These limited amounts outstanding do not support the liquidity o f the market. By comparison, Portugal sets an objective o f 5 billion euros for its benchmarks (thus approximately 10 times more) and in Belgium, the objective is a minimum of 10 billion euros. The BTA buy back policy o f the Treasury does not contribute to increase neither the outstanding stocks nor the liquidityo f the market. 105. Restrictions on the ability of banks to make direct transactions with jinal investors do not support the secondary market. These restrictions are reflected in Article 7 of the convention between Treasury (with the exception o f notification o f adjudication results and client payment). It is questionable whether such a restriction i s wise, insofar as it sets up an obstacle to player intervention inthe secondary market. One o f the key bottlenecks in the secondary market i s precisely its lack o f competition. The problems caused by the intra-group transactions should be regulated by general measures regarding transparency and market integrity rather than by barriers on secondary market competition. Similarly, the right to make secondary market transactions should be granted to all financial intermediaries without necessarily having to sign an agreement with the Ministryo fFinance. 106. Other more technical factors also hamper the development of the secondary market. Prominent among them are the following: The absence o f regulations on repurchase agreements transactions. The draft law authorizing and regulating the repurchase agreements i s currently under discussion at the Assembly. A rapid implementation is necessary. The procedures as regards payment-delivery system could still be improved. Advanced projects exist in this field, inparticular the multi-cycle and the shift to settlement inD+3. The system o f "liquidity contracts" concluded between the banks and private investors are not likely to favor the emergence o f an active secondary market for the domestic debt because : (I) are not transparent (dependingon the captive they relations between a bank and its customers); and (11) they do not increase the liquidityo fthe market as a whole. 107. Some progress has been accomplished recently in tackling the obstacles to the development of an active government securities market. In particular, since the assessment o f these issues in the context o f the FSAP for Tunisia (November 2001- World Bank / IMF, 2001b), progress has been achieved inthe following areas: The legal framework o f the repurchase agreements i s under discussion at the Assembly; 43 Implementation o f the repurchase agreements o f the 3 months Treasury bills by way o f auctions o f the T.C.B. from 6/11/2001; Modification o f the term o f the end-of-day rep0 (from 1to 7 days) by the T.C.B. from 2/1/2002; The interdiction for banks and other issuers to issue short term commercial paper was abolished; Reporting o f intra-group and OTC market transactions to the Treasury; A decisiono fprinciple to openthe local debt market to foreign investors. New banking law establishing the generic concept o f credit establishment, which regroup at the same time the banks and financial establishments; Various other technical measures relating to the working o f the banking activity (reinforcement o f the regulation concerning the granting o f credits and risk measurements, creation o f a liquidity ratio, publicity of the conditions o f the banks). 111.5 Reinforcing the investor base 108. Tunisia lacks domestic or international investors with a strong appetitefor long maturities. There exist three categories o f domestic long-term investors: pension finds, insurance companies and long-term savings. As discussed earlier, the pension funds are not largely developed in Tunisia, with the exception o f some rare public pension finds. The financial health o f the insurance sector is precarious. An in-depth rehabilitation is imperative for this sector to become an active player on the public debt market. Life insurance is not very well developed, for the same reasons as the pension schemes. 109. The Collective Investment Institutions are, on the other hand, reasonably well developed, and investment companies (SICAv) invest in Government paper. The sector is dominated by finds managed by banks, representing 25 out o f the 28 existing funds in 2000. Total assets o f the SICAVs grew fast since their introduction in 1992, amounting to 5.1 o f GDP at the end o f 2000.22The large majority o f SICAV (90 percent o f total assets) are bond hnds, investing about evenly on government securities (43 percent o f their portfolio) and corporate bonds-including commercial paper. However, these purchases do not contribute, however, to the activity o f the secondary market because they remain mainly "intra-group" transactions: the bank who buy securities at the auction simply transfer them to the in-house SICAV, The new accounting standards on mutual funds, introduced in 1999, set out modem valuation rules, including "mark-to-market" pricing o f assets. However, in practice, most o f debt instruments are still priced at their book value. The supervision o f collective investment institutions by the C M F has been 22 The Tunisian law provides for the operation o f both open-ended investment companies ("SociLtLs d'investissement a capital variable"-SICAV) and unit trusts ("Fonds communs de placement"--FCP), butonly the former have beenestablishedso far. 44 strengthened, but further steps are needed, as there are concems that the valuation o f assets may be deficient, while compliance with custodial and auditing rules may be inadequate (World Bank / IMF, 2001b). 110. Foreign investors cannot hold public debt denominated in TD, thus cutting the Treasury off apotentially important market. Tunisia has taken some steps inliberalizing the capital account and the government hopes to consolidate this progress and achieve full convertibility o f the Dinar in the future. But a number o f challenges should be addressed irrespective o f the sequencing o f capital account liberalization, among which the development o f a robust government securities market holds a prominent place (IMF, 2002). A partial opening o f the domestic market o f govemment securities is being examined by the Treasury, (see below). Broader opening up o f domestic debt issuance to foreign investors could be associated with further steps to open up the capital account in the future, but would call for ambitious steps to reinforce the bankingsystem. 111. However, extendingpart of domestic debt issuance toforeign investors callsfor careful planning because setbacks can be costly. In this field Tunisia cannot afford mistakes because a disappointed investor would be a lost investor, while setbacks inthe domestic debt market will be also sanctioned in the intemational financial markets, raising the cost o f Tunisia's foreign borrowing. Should non-resident investors be granted access to domestic debt, these investors must be able to exit the market as easily as they enter. Limited market liquidity on the domestic market would inevitably lead the foreign investors "to remain stuck" with their securities and the Tunisian debt would quickly acquire a bad reputation on the intemational markets. This would affect rapidly Tunisia's foreign currency borrowing because it i s very unlikely that foreign investors distinguish between loans in local currency and loans in currencies when liquidity i s at stake.23 The "illiquidity premium" on sovereign borrowing would thus increase the cost o ffinancing. 112. There are advanced projects aimed at opening the domestic government securities market to foreign investors (up to 5% of domestic debt outstanding) but it must bepointed out that this domesticmarket opening scheme is not very suitable in its current form. On the one hand, the system proposed is complicated and difficult to manage. Intemational experience shows that rationing measures are not applicable inthe secondary market, while foreign investors find it unattractive spending time trying to figure out complicated structural mechanisms. In addition, quota techniques are likely to disappoint foreign investors eager to buy or sell debt in local currency, even risking to drive them out o f the domestic market and, consequently, out o f the foreign currency bond market. For this reason, even though the gradual opening o f domestic debt to foreign investors goes inthe right direction, the currently envisaged application methods would need to be revisited. 23 The internationalization o f the holding o f the debt in domestic currency i s the key factor o f liquidity. Non-resident investors have other views than the domestic players and it is precisely this diversity o f interests that increases liquidity o f the domestic market. The example o f Belgium is very relevant: before the Euro, the OLOs (= Belgian OAT) were held 90 % inside and 10 % outside Belgium. The market was not very liquid, the Belgian players going generally at the same time in the same direction to buy or sell. Since the Euro, and due to constant efforts o f marketing abroad, more than 50 % o f the OLO are now held by foreign investors, this proportion reaching even 80 % for the OLO issued by way o f syndication. The liquidityofthe Belgian market has increasedand is now similar to that o f the Frenchmarket. 45 113. Psycho-sociologicalfactors would also have to be addressed, because the are signijkant in shaping domestic investor appetitefor public debt. A frequently cited cause for the failure of the public debt market is the lack of financial culture o f the market players, in particular with regard to the evaluation o f the fixed income assets. Private investors are used to think in nominal value and do not actually realize that the market value o f a bond varies according to market conditions They do not seem to be prepared yet to accept that the valuation of their bond assets i s lower than their nominal value. The fact that the investment vehicles are not evaluated at market value thus does not help the public to become aware of the financial risks o f its investment. III.6 Optionsfor reform 114. Although countries have adopted different approaches in the timing and sequencing of measures to develop a government securities market, experience suggests that some key steps are important prerequisitesfor success. The main elements o f programs that have proven successful are summarized in the World Bank and IMF Guidelines for Public Debt Management. Based on the diagnostics in this chapter, to assess in a snapshot Tunisia's progress inthis process, a 4-level rating is attempted with regard to each o f these criteria: (i) fulfilled; (ii) entirely fulfilled; (iii) progress; not some (iv) pending (Box III-4).24 Box 111-4.Benchmark Tunisia: development of an eficient government debt market Establishing a legal framework for securities issuance Fulfilled Developing a regulatory environment to foster market development and Fulfilled, but some further enable sound supervisory practices to be enforced improvements desirable, such as reporting o f intra-group transactions Introducing appropriate accounting, auditing, and disclosure practices for Not entirely fulfilled due to the financial sector reporting absence o f mark-to-market evaluation o f fixed income assets Introducing trading arrangements suitable for the size o f the market, which Fulfilled include efficient and safe custody, clearing, and settlement procedures Encouraging the development o f a system o f market-makers to enable buyers Some progress and sellers to transact efficientlv at orices reflecting fair value removing any tax or other regulatory impediments, which may hamper Fulfilled trading ingovernment securities fostering, at a later stage, the scope for other money market and risk Some progress (in view o f the management instruments, such as repos and interest rate futures and swaps imminentintroduction o frepos) Central bank operations to manage market liquidity Some progress Removing regulatory and fiscal distortions, which inhibit the development o f Some progress-reflecting institutional investors (e.g., pension reform) reasonable SICAV development, but absence o f pension reform Eliminatingbelow-market-rate funding through captive investor sources Some progress (role o f the general 24 The benchmarking in Box 111-4 should b e interpreted carefilly because in a number of cases, though the requiredregulatory steps have been taken and best-practice infrastructure has beenput inplace, the practice ofmarket participants does not fully comply with the existing regulatory framework. 46 treasury) Implementing appropriate rules and regulatory regime affecting participation Pending(some progress inview) by foreign investors inthe domestic market I n developing the suppry of government securities the key elements for establishing an efficient primary market include: 115. Moving forward, strengthening the market of the Tunisian public debt would call for measures in complementary directions. Three areas o f reform seem more prominent: organizational; structural, and technical. Organizational measures are discussed in the last section on Institutions for active debt management, so that the discussion below focuses on the two other areas o f measures and the phasing in o f reforms. 111.6.1 Structuralinitiatives 116. A properly functioning government securities market has to be underpinned by a healthy bank sector. The reform o f the public debt market is, therefore, linked to the reform o f the banking sector as a whole, and inparticular to the solution o f the problem o f the "bad loans". With the assistance o f the World Bank and other International Financial Institutions, the authorities have taken measures since the mid-1990s for the rehabilitation o f the banking system, and have significantly tightened prudential supervision. However, the reforms in this field should be continued and intensified, to reduce the dependency o f banks from Central Bank refinancing and allow more flexibility inmoneymarket interest rates. 117. Creating an enabling environment for the emergence of an active money market would call for a new framework of monetary policy. Monetary policy should move away from controlling interest rates, towards controlling base money. The current system o f credit targeting through the control o f the amount o f bank refinancing appears sub-optimal, due to the unstable relationship between bank refinancing and credit growth (IMF 2003). By contrast, base money seems more closely related to credit. Adopting a monetary framework built around the control o f base money would allow interest rates to fluctuate more freely and enabling the emergence o f an active interbank market and a short-term yield curve. This should go in tandem with a move toward operating procedures for monetary policy based on indirect instruments, such as open market operations. 118. More specific measures to reform theframework of the monetary policy could facilitate the emergence of a government securities market. First, and until the 47 government securities market has been established, it will be important to insure that only private paper o f good quality, matching strict criteria o f transparency and audit o f debtors accounts is eligible as collateral for the CBT refinancing operations. Second, the authorities could consider easing the rule o f 10 percent o f the annual revenues from taxes o f the preceding year as limit to the holding o f Government paper by the CBT. Evenifit is advisable that the implementation o f these two reforms takes place in a progressive way, their decision would give a clear signal to the market. 119. Promoting a capitalization pillar for retirement benefts would put thepension system on a more soundfooting and would strengthen the basisfor the development of domestic securities markets. International experience suggests that the development o f funded private pension schemes that supplement the traditional pay-as-you-go pillar o f the pension systems may potentially have large benefits for the development o f domestic financial markets. Under some conditions, funded pension schemes may increase the rate o f savings in a country, thus directly stimulating the development o f domestic financial markets. But even without a direct impact on the rate o f savings, funded pension schemes contribute to a shift in the composition o f financial assets, by increasing the supply o f long-term contractual savings. Better mobilization o f long-term savings can support the development o f the long-term end o f domestic government securities markets, because private pension funds primarily invest in government securities (36 percent o f pension fund assets inChile, and as much as 60 percent inPoland and 70 percent inBolivia). But the benefits from the emergence of pension funds depend on a number o f preconditions. Attainment o f critical mass; an enabling regulatory environment that frees up pension funds from overly constraining investment regulations; a supervisory framework that preserves integrity; prevalence o f competitive market structures; are important preconditions for securing a positive impact on the financial system and increased efficiency (Vittas, 2000). 120. However, creating even a partially funded private pension system entails initial fiscal costs. Moving to a partially funded pension system involves foregone revenues for the pay-as-you-go pillar. Public finances will, thus, bear temporary fiscal costs, to pay for the retirement benefits o f those who will not be covered by the capitalization pillar during the transition to the new system (existing retirees and old-age contributors). To keep such costs at a minimumit is important that the transition be implemented from an early stage, well before the projected deterioration o f the demographic fmdamentals and the associated worsening o f the financial situation o f the pay-as-you-go system. The Tunisian authorities are aware o f the benefits o f developing a properly regulated capitalization pensions pillar and have established various working groups for this purpose, although no initiative has beentaken to date. 121. Opening up progressively domestic debt issuance to foreign investors would help strengthen the investor base for government securities. However, to the prerequisite for such a move would be a robust domestic government debt market, and more specifically an improved liquidityof the secondary market. Failure to do so would risk to damage investor confidence, not only on the domestic bond market but, most importantly, on the market for foreign currency Tunisian debt. 122. The establishment of a new Code of duties for the SVT could renew the momentum of the domestic debt market. The drawing-up o f a good Code o f duties is a 48 delicate undertaking, which is more related to negotiation than to a rational construction. The Treasury must be prepared to make concessions and, thus, to pay the short-term cost for the development o fthe market. The initiative for this maneuver lies with the Treasury, but the recommendation is to carry out the broadest possible consultations with all the players concerned and not only with the (future) SVT. 111.6.2 Technical measures 123. TheSVT Code of dutiesshould be reviewed. 0 As seen previously, the current system o f non-competitive bids i s o f limited value in the absence o f a secondary market. Something else should thus be found to remunerate the S.V.T. in exchange o f their obligations. One option could be the access o f the S.V.T. to a refinancing preferential facility with the C.B.T., possibly modulated according to their own merits (activity on the primary and secondary markets, respect o f the obligation o f quotation, etc). The granting o f a preferential refinancing facility to the S.V.T. could facilitate the adjustment of the monetary policy while offering them a remuneration, which would have some value in this case. 0 Moreover, the SVT must post bid and ask prices and must quote, o f course within reasonable limits, when they are asked for a price. This is a key precondition, as it is indeed impossible for the secondary market to emerge if the market players do not quote prices to the investors. 0 Another important point o f the Code o f duties must be the requirement o f a coherent internal structuring o f SVT for the development o f capital markets. The SVT must indeed have a trading room, centralized management o f their own portfolios, their activities for a third party account, their activities in the international bond markets, their possible private bonds and options, etc., in order to support a permanent arbitrage between various financial instruments. In this regard, existing gaps in the Code o f duties must be filled. The emphasis should be laid on ensuring an SVT structure that enables a more dynamic presence on the capital market. 0 It is finally necessary to remove from the Code o f duties all measures that might restrict competition on the secondary market. All SVT (and more generally all financial intermediaries, including the banks) must have the freedom to buy or sell to anyone inthe secondary market. if124. All intra-group transactionsshould be reported to the reporting authority, even they are not carried out on the stock exchange. The declaration should at least mention the following elements: identity o f the counterparty, nature o f the transaction (purchase, sale), underlyingsecurity, exchanged volume, price (or rate), and value date o f the transaction. The reporting authority would thus have the necessary elements to establish the transparency of the market and to release a yield curve, even if these transactions do not take place on the stock exchange. 125. Measures should also be considered concerning the intra-group transactions. One option is to put inplace "best execution rules", intandem with the reporting o f the 49 transactions to a reporting authority (which could be the stock exchange), even if the transactions are not executed in the stock exchange but OTC. O f course, control o f the rules of "best execution" should be enforced and enforcing such control would call for close coordination between the CBT and the CMF. Another option i s the interdiction o f the intra-group transactions and the centralization o f these transactions on the stock exchange. This could be a way to bring these transactions to light, and consequently to create a secondary market, while limiting the ambiguous relationship that may exist between various entities o f the same group. The market would be much more transparent ifaSICAVwas obligedtobuyfrom anotherbankthanthebankofitsgroup. 126. Suppressing the indexation of the cost of credit to the economy on the TMM would addflexibility to the money market. This indexation can be justified within the framework o f a stable monetarypolicy inwhich the rates remain constant but it will raise problems when the secondary market develops. The cost o f credit to the economy must be determined: (i) comparison with the yield curve o f the Government securities- in rather than on basis o f the money market rates, even if a variable margin is applied to it according to the term o f the credit; and, (ii)according to the quality and the creditworthiness o fthe debtor, as opposed to other economic policy priorities. 127. To build its credibility, the Treasury must respect its calendar and no longer cancel auctions, whatever the circumstances or the timing of the external issues. The Treasury must distinguish between its cash position management and its long-term funding policy. These are two different fields which must be managed separately and which should not interact. In other words, the Treasury must absolutely avoid regularly modifying its long-term fundingpolicy infunction o f cash imperatives. 128. I n this approach which separates Treasuryfinancing and daily management of cashflows, the Treasury must study the possibili@ of becoming an interbank market player. The Treasury will thus have the ability to actively manage its daily cash needs. Indeed, its cash surpluses should not be "frozen" at the central bank but could become actively managed, thus reducing the cost o f State financing. This reform would not affect BCT monetary policy, because Treasury will be a market player just like the BCT and others, and therefore monetary policy should be adapted according to the currency held bythe Treasury, likeany other marketplayer. 129. The Treasury must respect the range of the needs it publishes and allocate at the price asked by the market. The Treasury should announce a minimum and a maximum amount for each maturity offered. Of course, initially, this policy will cost, but itwill greatlyhelpbuildcredibility, thus favoring the emergence o f amarket o fthe public debt. Ifnecessary, the Treasury must also re-examine the amount o f the needs it wants to fund by way o f auctions, in order to adjust its volume to the expected demand o f the market. Itwill become therefore predictable inthe amounts it allocates. 130. Limiting the number of issuance lines of BTA and increasing the outstanding amounts will improve market liquidity. Issues should be concentrated on a limited number o f BTA, the benchmarks, to be chosen according to market demand. The optimal issuance size o f a benchmark should be examined jointly with the market participants. Roughly, the Treasury should aim at an average outstanding o f roughly one billion TD per benchmark line. Similarly, the buy back policy can be justified only insofar as the 50 Treasury buys back the old illiquid and cheap debt so as to issue more liquid and more expensive new debt. A dialogue with the market i s necessary before deciding o f a repurchase policy. 131. Upgrading the public debt market would call for a "mise d niveau " of it's the main players. This would require a general effort for a better technical training: participation in organized seminars either locally, or in partnership (co-operation o f managers o f public debt and bankers o f other countries for "training courses in total immersion", participation inthe seminars and meetings organized by certain international banks, training course, etc.). From this point o fview, the opening o f the banking sector to foreign banks would be particularly useful, because it would enable the exchange o f experience and o f ideas and the broadening o f the financial domestic culture from which Tunisia would have muchto gain. 132. I t is necessary to implement the mark-to-market valuation of fixed income assets, not only for government securities, but also for the SICAV, investmentfunds, etc. The obligation exists legally but it not applied in practice. It will be necessary nevertheless to act cautiously given that a brutal passage to valuation in market value is likely to reveal losses and depreciations inthe portfolios, with the consequences that can be imagined, not only legal, but mainly psychological since the population is not accustomed to this method o fvaluation. 111.6.3 Phasinginthe reforms 133. The phasing of the reforms would depend on the political agenda of the Government and on the market situation. However, as a general guideline, the sequencing o f reforms could proceed as follows: The Treasury should make the distinction between (i)its mission o f public authority, where its credibility has been well established, and (ii) its State financing mission as a market player, like the others. Inthis second mission, the touchstone to its strategy must be the long-term partnership with the financial sector. The task at hand i s to work out beneficial opportunities for all market participants, as it i s very unlikely that a market could be established by decree. The Treasury must build its credibility as "sovereign issuer," particularly in the following areas: (1) issuance predictability (timetable and amounts to be raised), (2) payment o fmarket prices at auctions, and (3) limitation o fits issuance to some lines selected in dialogue with the market, so as to achieve a critical amount o f stocks outstanding. The drafting-up o f a new Code of duties and the consultations this work requires are on the way. Particular attention should be paid to : 0 The SVT remuneration within the framework o f a long-term partnership with the Treasury. 0 The SVT's quotation obligation (with limits in centimes rather than in basic points for the BTA) and in reasonable margins to be determined in dialogue with the market and according to development o f the secondary market (the current margins o f 50 and 20 basis points appear too broad). 51 0 The SVT internal structuring according to capital market needs - installation o f a marketroom. 0 The opening o f the secondary market to the competition o f all financial intermediaries, which means that banks (and more generally all financial intermediaries) must be able to be active on the secondary market without preliminary authorization. 0 The term o f S.VT. mandate should be shortened. Three years seems too long, compared to the practice in "developed" markets o f one year or sometimes two. 0 The SVT selection should be reviewed. Not all banks or brokers have the capabilities to become SVT. It is necessary to aim for quality rather than quantity. The Treasury could - and must - be demanding in its choices, assuming that the SVT's activity is lucrative in view o f the way they are remunerated. Other technical measures (legal framework for repurchase agreements, improvement o f compensation system and payment, revision o f auction procedures, faster introduction o f bids and faster communication o f results), etc.) must also be implemented quickly. Initiate a consultation on the structural reforms (opening o f the banking sector, reforms o f the monetary policy, promotion o f longterm savings etc). The reforms in these fields are delicate and will not bring their effects before long; it is therefore better to begin or to continue the at an early stage, even if the concretization o fthe reforms takes time. Make the secondary market transparent: 0 In the immediate future, the Treasury must make public, every day before the market opening, the yield curve which it can establish from transactions declared bybanks and S.V.T. 0 Then, it is necessary to quickly set up a general obligation o f OTC transactions reporting. Effectiveness will hinge on good sharing o f competencies between B.C.T. and C.M.F. The Tunis Stock Exchange could be incharge o fthis mission o freporting authority. 0 In a short period, reporting of all transactions, not only those related to government bonds, should be generalized. 0 If the compulsory transactions reporting is not enough to increase the secondary market transparency, it would be necessary to adopt more radical measure, such as the banning o f intra-group transactions or their centralization on the stock exchange. 0 Reinforce the obligation o f quotation in order to increase competition between market players and to enable investors obtain better prices on the market than from their "in-house bank". This reinforcement o f obligation must go inparallel with the establishment o f secondary market transparency. Re-examine methods o f domestic debt opening to foreign investors. An option would be to open a new line o f 3-month BTCT to both foreign and resident .nvestors: (i)the difference innotation are less important on the short term than on ;he long term, (ii)it is important to familiarize foreign investors with the local 52 market. In this respect, foreign investors prefer initially buying short-term paper. (iii) bondliquidityislessproblematicfora3-monthBTCTthanfora12- Limited year BTA, and (iv) The Treasury can maintain control o f outstanding stocks. A target o f 300 to 400 million DT can be aimed at in a first step. Such an opening to foreign investors could also have a favorable influence on short-term market rates. The Treasury must consider the possibility o f becoming a player in the interbank market, inorder to manage its daily cashtransactions. When the secondary market starts developing and there is a yield curve, consider the valuation at market value o f the fixed income assets for the SICAV, investment funds, etc. Evenifthis measure should not be immediate, market participants could start its study without delay, in particular with regard to phasing in and the transitional measures. 10)Initiate a program o f "mise ri niveau of the financial culture inthe near future, and " makeplans with regardto financing (bilateral grants or borrowing from multilateral institutions) and practical feasibility (technical assistance, training courses organizedby international banks, choice of seminars, etc). 134. Immediate timetable. An actionplan that can be carried out immediately can be summarized asfollows: The Treasury must set up its Internet site, which will be the privilegedtool o f its relation with the market. To create and update an Internet site i s not expensive and does not require a disproportionateworkload. The Treasury must calculate a yield curve based on data that it receives. must then make the yield curve public everyday before the markets open (that i s important, as no market i s interested in a three-day old curve). Examples include a publication inthe daily newspaper, on the Tunis Stock Exchange Internet site, the bulletin o f quotations o f this stock exchange, etc. It is necessary to ensure the coordination between B.C.T. and C.M.F. in order (i) regulate the transactions that are carried out by banks on the to OTC and intra-group markets, and (ii) to establish a general mode o f transactions "reporting". As for the authority o f "reporting", the Tunis Stock Exchange appears to be a good choice, thanks to its good infrastructure and know-how. It would be recommended that the new SVT Code o f duties goes in the direction outlined above. This does not require large investments but a political will to tie a new partnership with the market. Regulations on repurchase agreements transactions. The bill is under discussion in the Parliament, but the Treasury can already anticipate by preparing texts o f application and by organizing the "mise ri niveau" of the players. An action plan for the "mise ri niveau of financial market players (banks " and institutional investors) must be elaborated. Several options are possible: Technical assistance bythe Treasury, international organizations; courses and seminars organized by international banks; training sessions 53 organized by schools and universities; trainings given by private sector specialists. The opening o f the local market (in DT) to foreign investors should start with the opening o f issuance lines o f 3-month BTCT without limitation for foreign investorsneither on the primary nor on the secondary market. Then, according to the acquired experience, the access could be extended - BTCT to 6 months, etc. By all means, the operationalization o f this market opening to foreign investors must be accompanied by an offensive "marketing" o f appropriate scale. 54 IV. Institutions and Informationfor Effective DebtManagement 135. The two most visible trends in reforming public debt management institutions have beenfor the governments to: (a) centralize all aspects of public debt management function in one unit (the "Debt Agency'?, allowing comprehensivemanagement of all risks of government liabilities; and (b) organize the debt work byfunction. Merely two decades ago, even in the most advanced economies, the function o f debt management was scattered among government agencies, with domestic and foreign currency debt managed separately. Since different government units focused narrowly on their particular responsibilities, it was impossible to develop a portfolio view o f government financial liabilities and the overall strategic risk orientation. The portfolio perspective was further clouded by bureaucratic rivalries. Meanwhile, the rapid pace o f financial globalization had exposed the OECD governments to greater range o f risks and higher market volatility. Loose fiscal policies led to over-borrowing, while poor debt management further increased the risks o f rapidly growing government debt portfolios. By mid-1980s many OECD governments became concerned about excessive levels o f debt, and initiated reforms to centralize government debt management operations; and obtain the right mix o f market-focused skills. Inthe past decade, most OECD states have completed such institutional restructuring; more recently, many emerging economies have followed suit. 136. Functional organization of a government Debt Agency, similar to the organization of an investment institution, has largely replaced traditional organization by product25.Under it, three operationalblocks are distinguished: . The Front Office (FO) implements the debt management strategy day-to-day, including the execution o f all transactions. It participates in the design o f the funding strategy, and takes the decisions concerning the amount and terms o f loans to be obtained (choice o f borrowing instruments), as well as new product development. It contacts with the FO o f the counter parties (lenders), and concludes the transactions. The FO i s the main vehicle for investor relations, marketing to investors both domestic and international, and communication with other Treasuries, the public, the press, etc. If a Government has a substantial on- lending operation, the Front Office would execute it. The FO is also incharge o f portfolio management - it implements decisions to rebalance the portfolio bringing it closer to strategic benchmark. The same body may also manage the Treasury's liquid assets indomestic and foreign currency, and do risk hedging. In some countries the FO manages the daily cash position o f the Govemment in cooperation with the Central Bank. The Middle Office (MO) takes the lead in developing the debt management strategy. It analyzes public liabilities in terms o f risk structure, develops 25Intraditional organizational structure, operationally separate departments would exist for external and domestic debt; fragmentation would often go deeper, with separate departments for investment project loans andborrowing from the financial markets. 55 borrowing scenarios, and proposes medium-term and in-year risk mitigation measures. If the Debt Agency undertakes the management o f financial assets o f the government, the Middle Office would formulate assetAiability management methodologies and develop guidelines for onlending. The Back Office (BO) settles the transactions undertaken by the Front Office. It checks them with counterparties, issues payments instructions, and makes sure that sufficient funds are available in the Treasury account. It maintains debt databases, making sure that information on public debt is correct and up-to-date. It uses its databases to validate data from the FO, send instructions to the netting, clearing and settlement systems, and to report transactions to the Authorities and to the audit bodies. The Back Office would normally takes care o f debt accounting and translate debt transactions into budget execution entries. The boundaries o f the Front OfJice and the Back Office are quite standard in different national environments. On the contrary, the boundaries o f the Middle OfJice in a sovereign debt management structure vary from country to country, reflecting different institutional setup for the broader public finance function. 137. Thefocal point of the Government's Debt Management operation is the Middle Office, whose core competenciesare drafting and supervision of the medium-term debt strategy. The strategy i s adopted by the Government's decision endorsed by the Parliament). The MO translates general debt management objectives into manageable and quantifiable guidelines. The debt strategy i s comprised o f the primary market programs (what instrument to be issued andor bought-back, at what maturity, inwhat currency and inwhat proportion, when, etc.); communicationandmarketingstrategy; andnewproduct and infrastructure development program. Portfolio risk management objectives (criteria to decide on the choice between domestic and foreign borrowing, on average maturity, duration, maturity profile, etc.) should also be part of the debt strategy. First, simple benchmarks should be developed. More advanced measures like "BudgetKost-at-Risk", which quantify the government's risk, should be introduced when the environment becomes more sophisticated. The strategy is then translated into general directives to the debt management unit, to be signed by the Minister o f Finance. These directives constitute the framework for the debt managers' actions. The MO, in close coordination with the analytical units in the MOF, Ministry o f the Economy, and the Central Bank, analyzes the government borrowing policy to assure the consistency with the medium- term macroeconomic, fiscal, and monetary policies. The MO also monitors off-budget risks, particularly related to other public debt issuers such as the municipalities andpublic enterprises, government guarantee portfolio, and the like. With regards to the contingent liabilities o f the State, MO competencies may range from minimal (collecting information) to coercion ifnecessary, depending on the political set up, the seriousness o f risks, and the effectiveness o f control instruments. In some countries, the MO monitors and offers advice even to large private sector borrowers. 138. The technical competence of the MO that underpins all its other functions lies in risk analysis and malzagement. This competence is the key to strategy formulation, especially if the MO uses mathematical models (e.g. benchmark portfolios). In such a model, risks, which cannot be managed, are treated as "constraints". These constraints 56 must be accurately defined and politically accepted, as these set the boundaries o f the feasible debt strategy. Risk analysis must be undertaken in a multiyear environment, and be comprehensive, Le. cover all types o f government liabilities, and risks on the asset side as well. 139. Related to risk analysis is the risk control function of the Middle Office. I t should exercise thisfunction both within the broader government and, in particular - vis-d-vis the main borrowing desk of the Government, which is the Front Office. The latter, as all market operators, tends to have a very detailed perspective on each borrowing operation. It also has to make expedient transaction decisions. As a consequence, the Front Office often has neither the capacity nor the time to analyze the entire portfolio o f risks, and may exceed prudent risk limits. The Middle Office's role i s defining these limits, embed them in the debt strategy document, and to enforce them once this document i s adopted. The risk control function should prevent the FO from taking positions deviating from the risk profile approved by the Government. At the same time, the MiddleOffice assists the FO indeveloping new hnding instruments. 140. The Middle Office alsoperforms reporting on the government debtportfolio. It prepares and publishes the Annual Debt Report (annexed to the budget execution report to the Parliament), quarterly updates, and operational reports to the Management of the Treasury and other decision makers (the Cabinet, the MOF, etc.). A specific form o f reporting i s the MO's inputs into budgeting: it prepares inputs into the state Budget and updates the MOF regularly on the implementation o f its borrowing program. A strong link with the Budget Formulation Department of the MOF is required, to monitor the evolution o f the State's funding needs. These projections should include estimates of budget costs o f contingent liabilities. Finally, the MO conducts research on public liabilitymanagement and relatedissues. 141. The location of other functions may vary. The Back Office usually includes a dedicated information technology group, but it also can be a separate unit. A dedicated legal group is needed, to maintain documentation for foreign debt, syndications, and adaptations to the current legislation with regard to the development o f the domestic market, collateralization, and ISDA contracts. These activities may be part o f Front Office, Back Office, or it may be located elsewhere and service more than just the Debt Agency. However, an essential requirement for effective formulation and execution o f a public debt management strategy is that all three units described above function as one organism. 142. Another important challenge is the degree of autonomy of the government Debt Agency. In the OECD countries, the recent trend has been to increase the operating autonomy o f the Debt Agency, inorder to protect it from political pressures that may lead to over-borrowing. This shift towards greater operational autonomy was pioneered by Sweden, Denmark and Ireland, and took different forms in different countries. Some developed and emerging market economies have opted for a completely separate debt agency (Hungary, Ireland, New Zealand). This happened simultaneously with the increased independence of central banks, required for clearer delineation between monetary policy and government debt management. Increased autonomy brought about the need to stronger coordination o f debt management with other areas of government policy. The government debt portfolio is usually the largest liability portfolio in a 57 country, and can have major systemic repercussions on the macro/fiscal outcomes, the monetary policies o f the central bank, and on the financial markets as a whole. Greater coordination helps prevent negative policy outcomes. 143. For a given national institutional setup, the degree of autonomy of the government Debt Agency and effectiveness of its institutional model depend on the features of policy and market environment. Greater autonomy can be justified only where the Debt Agency's actions fully influences the debt policy outcomes, and can be held accountable for its actions. Maximum operational autonomy makes sense only at certain level o f maturity o f the economy as a whole, and the financial markets in particular. Furthermore, the degree o f autonomy o f the Debt Agency refers to its ability to implement an agreed strategy - to transact without higher approval while remaining the risk boundaries set by a debt management strategy. Higher-level decisions about these risk boundaries, as well as other key fiscal parameters are made by Ministers o f Finance, or policymakers o f even higher level. 144. I n Tunisia, debt management practices are strong, but will benefit from greater centralization andfunctional focus. Institutional structure existing in Tunisia i s akin to what existed inmost OECD countries 10-15 years ago, with debt management scattered across several agencies and fragmented along borrowing product lines. The Treasury i s in charge o f domestic debt issuance; the Ministry for Intemational Cooperation deals with loans from multilateral financial institutions; and CBT handles borrowing on foreign markets. The Treasury exerts overall control over the debt servicing payments for direct government obligations, but does not play an active role in formulating the debt strategy. The Central Bank o f Tunisia is the banker o fthe State, and is responsible for debt service payments in foreign currency. While the CBT's professional capacity i s impressive, coordination between the MOF and the CBT indebt management could be strengthened. Tunisia's institutional setup provides adequate control over the individual debt instruments,but does not support the integrated view o fpublic debt portfolio, and could obstruct more active management o fits risks. 145. The need for greater cohesion of different components of Tunisian debt management is evident on thepolicy, operational, and information levels: . Policy: The macro-framework for Tunisian debt policy has been strengthened over the past decade, allowing the government to contain its debt burden within sustainable boundaries. However, this macro-framework i s formulated for the Five Year Plan, and is too generic and inflexible to supplythe debt manager with meaningful operational benchmarks. Efficient debt management requires more flexible policy setting, synchronizedwith the budget process. This problem canbe resolved by greater coordination between the Budget department o f the M O F and the Treasury in formulating the borrowing plans as part of the rolling medium- term fiscal framework; Operations: Debt management is organized by product, allowing robust control over the risks o f individual financial products; what i s lacking is the adequate control on the overall level o f risk in the government debt portfolio. Fragmented and ad hoc perspectives on risk can lead to erroneous choice o f borrowing instruments, resulting in the unjustified increase in the cost o f borrowing and 58 portfolio management decisions. This calls for a decisive reformulation o f approaches to debt operations: each should be viewed as part o f the portfolio o f government liabilities; Information and analysis: While the quality o f information on sovereign debt is good overall, it i s also fragmented by type o f debt and i s thus neither sufficient nor up-to-date to support active debt management. Box IV-I. Evolution of Debt ManagementInstitutions in the OECD Countries Inthe 198Os,public debt levels and financial risks inthe debt portfolio o f various OECD countries rose considerably. This generated a strong impetus toward modernizing and reforming government debt management. DMO functions were consolidated and gradually public debt management was separated from the implementation o f monetary policy. Debt management was increasingly seen as an instance o f portfolio management having distinct objectives in terms o f cost minimization within risk limits. In an attempt to increase the efficiency o f debt management, a number o f governments delegated the operational dimensions o f debt management to separate debt management offices. Other OECD countries did not see the need for a separation between DMO and Ministry o f Finance and questioned whether it was advisable to operationally isolate public debt management from public policy. These countries favor a balance between public policy and financial portfolio management in their approach to debt management. Though debt management functions remained in the Ministry o f Finance, existing departments were consolidated and modernized and clear objectives, guidelines and accountability mechanisms were adopted. Regardless of the location o f the DMO, four issues were identified as crucial for the success o f active debt management: m The formulation o f a risk-averse, market-neutral debt strategy in the context o f sustainable fiscal policies, m Creation of mechanisms for successful delegation o f debt management function to and accountability to the Ministry o fFinance and Parliament, m Strengthening institutions and cadre o f debt management for both fmancial portfolio management and sustainable debt policy; Modernizing debt information systems. In Hungary, two important institutions were created in 1996, to improve budget execution and debt operations: the Treasury and the Debt Management Agency (Allamadhug Kezelo Kozpont, AKK). The k K prepares the financing strategy o f the Treasury, which is approved by the Ministryof Finance, and carries out the borrowing decisions. It has two other important functions: organizing the domestic market and providing information for market participants. The management o f the foreign portion o f the public debt was transferred from the National Bank o f Hungary (NBH) to the AKK in 1997, after one year of discussions and preparations. The Hungarian State Treasury is an independent organization operating under the supervision of the Minister o f Finance. In practice, the Treasury and its branch network was built around budget implementation functions that were carried out by the State Development Institute and the NBH. The ledger system o f NBH, containing the accounts o f government agencies, was transferred to the Treasury. To facilitate the recruitment and retention o f qualified staff, the Treasury obtained a special salary scale for its public employees and absorbed experienced professionals from the State Development Institute, the NBHandits 19 county directorates. InMarch2001, ina move to modernize debt management, k K was establishedas ajoint stock company, organizationally independent but under the supervision of the Ministry o f Finance. The tasks o f the new agency include, among others, the fine-tuning o f instruments for the issuing o f public debt, and the systematic use o f benchmarking in order to minimize risk and costs associated with securities, both denominated in HUF and in foreign exchange. The k K monitors debt risks in order to ensure the long- term sustainability o f the Hungarian debt. Source: Source: Currie et al. (2003), Buzas, L.(2002). 59 146. Most important, the risk management function is insufficiently developed in Tunisia. While each o f the government agencies responsible for a certain type o f debt conducts assessment o f risks present in it, these assessments are aggregated only in very general exposure indicators prepared by the MOF for the budget execution reports and annually for the Budget Economique. Some risks (for example, those o f contingent liabilities) are not quantified at all. 147. A government risk management strategy is only as good as the information used toformulate it. Tunisia's government finance information system (CFIS) is one o f the best among countries at similar level o f economic development. Databases on individual classes o f government debt are modem, and support well transaction control and standardized reporting. On the extemal debt side, the Central Bank o f Tunisia and the Ministry o f Finance have sponsored the development of a centralizeddatabase (SIADE), which i s among the best practices in the emerging market economies. The system has a modern open architecture design and is supported by a dedicated team o f system specialists. SIADE covers all categories o f public debts, including those not guaranteed by the state, as well as onlending operations. It includes a strong forecasting module, which generates loan-by-loan cash flow profiles. The system allows generating a variety o f aggregate reports. However, there exists no unified debt database; consolidated debt reports are difficult to compile and even more difficult to customize, which makes it impossible to proactively respond to the changing market conditions. While SIADE meets most o f the transaction management needs, government experts express concerns about the quality o f information, particularly with regards to the disbursements under project loans. This information i s supposed to be provided by the sector ministries and entered in the database by its administrators - Centre Informatique du MinistCre des Finances (CIMF). Sector ministries often delay such updates, which affect the reliability o f SIADE's forecasts. 148. Analytical systems supporting debt strategy formulation are insufficient for active management of Tunisia's risk exposures. SIADE does not have analytical capacities beyond cash flow forecasting for individual loan instruments. Only nominal valuation o f loan instruments i s supported, and no mark to market valuation is available. The report generation facility cannot easily create customized reports required for active debt management. Onthe domestic debt side, both information requirements and systems implemented are quite basic; data are kept and core parameters are calculated in simple spreadsheets. Ministry o fFinance does not regularly prepare debt sustainability scenarios, relyinginthis on other agencies such as the CBT, Ministryo f International Cooperation, and Ministry o f Economic Development. The existing debt information system does not allow quantifying guarantee risk; no quantitative procedures exist to evaluate the risks o f onlending operations and the risks o f non-guaranteed debts o f the broader public sector. Private extemal debt, which will have a growing impact o f the balance o f payments trends, is tracked by the CBT, but this information i s not included into the debt sustainability scenarios. Another important deficiency i s the lack o f consolidation for the total government portfolio (external plus domestic) except on the highest level o f 60 aggregation, and only in nominal figures. Mark-to-market valuation o f the whole portfolio is impossible26. A BLUEPRINT FOR INSTITUTIONAL REFORM 149. The Government of Tunisia realizes that active debt management strategy begins with institutional change. One o f the key reasons for preparingthis Report was the need to determine viable institutional reforms, requiredto bring debt management up to the new level o f efficiency. The mandates o f key debt institutions must be reinforced; and they should be given more flexibility in conducting their operations. A new government function for risk management must be created, and a sustainable medium- term debt strategy must be formulated, relying upon clearly defined quantitative benchmarks. The performance o f government debt managers must be assessed against these benchmarks. These needs can be met most efficiently via the establishment o f a specialized Debt Agency, which will manage all contractual liabilities o f the Tunisian government, and will monitor fiscal risk from the general public sector and beyond. 150. For the Debt Agency to be effective, it shouldput its risk mitigation expertise at the service of senior government decision-makers. In the international practice, a combination of interagency body setting the strategy (High Debt Committee) and a technical body supplying decision support (Debt Agency) has delivered the best risk mitigation results. It i s thus advisable to establish a senior-level High Debt Committee, and have the Debt Agency submit its proposals and assessments to this Committee. The HighDebt Committee should have representation ofall relevant government institutions, including the Ministry for Development and International Cooperation, as well as the BCT. The Minister o f Finance as the top official should chair it personifying the sovereign borrowing authority. CBT's active participation in the High Debt Committee would guarantee good coordination between monetary policy and government debt management. 151. I n Tunisia, the proposed Debt Agency can only be semi-autonomous (most likely attached to the MOF or the Treasuy), since government debt management is closely integrated with broader public policy, and thus the Agency cannot be heldfully accountablefor its actions. Tunisian economy is less diversified and more vulnerable to external shocks than a typical OECD economy. The fiscal sustainability stance remains vulnerable, particularly on the external side. In the event o f serious external shock, Tunisia's debt service i s likely to jump substantially. Another constraint i s limited recourse to domestic borrowing. On both counts, Tunisian debt managers face real limitations on what active market operations they can undertake to change the composition and the risk profile o f their debt portfolio. The arrangement under which a 26This observation should not be interpreted as a recommendation o f unifying all debt information (both external and domestic) in one database. International experience has shown that such integration is very costly and not necessarily efficient. One off-the-shelf solution advertised recently is a new version o f CS- DRMS, shpped by the Commonwealth Secretariat, which can register both external and domestic debt. Implementation record of this version o f CS-DRMS is still too short for conclusions about its efficiency in comprehensive debt portfolio recording. An alternative solution would be to equip the analytical modules with specially designed "data extraction channels" that would query the central databases and generate information inthe format meaningful for portfolio risk analysis. 61 semi-autonomous Debt Agency reports to the High Debt Committee chaired by the Minister o f Finance would offer an optimal combination o f policy centralization and operational flexibility. 152. The preferable path of reform would be to establish the Debt Agency in one move. The Tunisian Government already has formidable Back and Front office units, which should be reassigned to the new Debt Agency. However, the nucleus o f the Debt Agency would be the newly established dedicated risk management unit - the Middle Office. Building up this risk control unit is the key priority. Since the quality o f risk control depends on the quality o f debt information, it i s also important to reassign to the Debt Agency the debt-recording units (Back Office), which are now part o f the Treasury. The Front Office function now exists as several desks, which work with different classes o f lenders and have to follow distinct procedures for each o f them. For a while, they may be allowed to operate inthe current environment - provided their activities are subject to risk control exercised by the Debt Agency. Eventually, the Front Office desks will need to be reassigned to the Debt Agency, and the sooner this happens, the better. 153. Within the new structure o f the Debt Agency, the Middle Office will be responsiblefor drafting a comprehensive medium-term liability management strategy, and submitting it to the High Debt Committee for review and approval. Once the Government endorses the medium-term strategy, the Middle Office should supervise its implementation. In performing this function, it should have sufficient authority to look into borrowing and debt management practices o f all government agencies, and to evaluate risks to the centralbudget stemming from these operations. As priority tasks, the MO should: (a) prepare weekly reports to the Management o f the Treasury MOF, CBT, and Ministry o f Economic Development and cooperation, covering the whole portfolio o f public liabilities, and presenting a comprehensive and detailed risk analysis o f this portfolio; and (b) publish quarterly debt reports, as the main reporting vehicle to the broad investorhreditor community, the legislature, and the broad public. To make these reports authoritative, the Middle Office should undertake regular (at least quarterly) consolidation o f all relevant debt data, including on contingent liabilities o f the govemment, public sector debts on onlending operations. This will facilitate regular updating o f debt sustainability scenarios and portfolio benchmarks. The Middle Office should also implement quantitative analytical systems supporting the measurement and forecasting o f various portfolio risks. I n sum, the Middle Office should not engage into individual debt transactions, but serve as a risk control "watchdog".27 I54. Interagency coordination is crucial, particularly between the Debt Agency?the Treasury and the MOF. This i s required for smooth coordination between treasury cash management and liability management: inefficient cash management practices ofien distort debt issuance and increase the cost o f borrowing to the Government. Such coordination allows improving other portfolio management functions, in particular - when the debt management unit also manages the Government's financial assets. The Budget Department o f the Ministry o f Finance and the Ministry o f Economic 27 It i sadvisable that the Middle Office's risk mitigation mandate extends to commercial public enterprises, since inthe event of their excessively riskybehavior the Government balance sheet may be exposed even in the absence of an explicit guarantee. InDenmark, for example, the Crown's Debt Manager had the right to advise even private sector borrowers o n the timing and terms of their external debt issuance. 62 Development are other important partners o f the Debt Agency: government debt strategy must be set and implemented in the context o f medium-term fiscal framework. Of even greater importance i s the Debt Agency's coordination with the CBT. As the Front Office functions are eventually reassigned to the Debt Agency, the Central Bank will limit its role in state debt management to providing banking services. The need for the harmonization o f monetary policy and government debt management will remain as strong as ever, requiring regular - at least monthly - consultations between the Debt Agency and the CBT. The development o f the domestic government debt market would be impossible without the Central Bank taking the lead in strengthening the market infrastructure and encouraging the development o fmoney andrep0 markets. 155. Strong capacity for active debt management requires an adequate system of incentives. Tunisian government debt managers should be offered attractive remuneration structures and other incentives, such as training programs and improving promotion prospects. The skills mix required by the Debt Agency is very'close to the skills mix needed inprivate investment banks, investment companies etc., andthe Debt Agency will be competing for talent with the private sector. The Agency's salary levels should reflect the importance and complexity o f the Debt Agency jobs inmanaging the largest liability portfolio in the country. It should, therefore, have greater autonomy in managing its pay scales and hiringprocedures. The hirindpay scale arrangements for the newly established Debt Agency should be no less, and possibly even more flexible as those existing for the CBT, or the Cour des Comptes. 156. Institutional reform requires strong political leadership and modern legal framework. The reform would entail deep shifts in the mandates o f various government agencies. Some may raise objections to these shifts, arguing that the long-established government practices "work well enough". Strategic vision and leadership i s needed to escape bureaucratic impasse, and to realize long-term gains from the new institutional structures. A coherent legal framework, giving the newly created Debt Agency adequately strong risk control mandate, must shore up the envisaged reforms. It is advisable to pass a new Law on Public Liability Management, endorsing changes in institutions and processes, defining the roles o f different government agencies, and establishing broad efficiency criteria against which the new debt management practices will be evaluated. 63 ANNEX 1.MEDIUM-TERMDEBT SUSTAINABILITY SCENARIOS BASECASESCENARIO TABLE 1: MACROECONOMICFUNDAMENTALES 2002 2003 2004 20% 2006 2007 2nO8 2309 2C1C 2CII 2012 I" % Growth Rateof GDP 1 9 5 4 6 0 5 7 5 6 5 6 5 6 5 6 5 6 5 6 5 1 ConsumptionGrowth 4.8 5 5 5.4 5 6 5 8 5 8 5 9 5 7 5 36 55.16 4 87 5 Public Consumption 5 2 4.9 6 0 5 7 5 6 5.6 5 6 5 6 PrivateConsumption 4 7 5 7 5.3 5 6 5 9 5 9 6 0 5 7 5.3 5.0 4.7 lnveshnentGrowth (GDI) 4 4 7 6 7 6 6 9 6 5 6 5 6 5 6 5 6 5 6 5 6.5 Impor?(GNFS) growth -1.5 8 8 7.8 7 5 7 3 7 5 7 7 7 8 7.8 7.5 7 5 Export (GhTS) growth -2.8 7 7 8 1 7 1 6 6 6 8 6 9 7 4 8 0 8 0 8.5 Inflation (pa.) 2 7 2.6 2.5 2 4 2 4 2.4 2 4 2 4 2 4 2 4 2.4 in % ofGDP Real InterestRateForeign 2 1 2 5 1 9 2 1 2 3 2 3 2 4 2 5 2 7 3 0 3 3 RealInterestRateDomestic 3 1 3 5 3 7 3 7 3 7 3 7 3.7 3 7 3 7 3 7 3 7 Investment 26 2 26 7 27 1 27 4 27.6 27.9 28 1 28 4 286 288 291 Exports 447 456 46 4 46 9 47 2 47.1 48.2 49 0 50 1 51 2 525 Imports 488 497 50 3 50 9 51 7 52.6 53 6 54 7 558 568 577 CAB (excl all officialtransfers) -3 5 -3 5 -2 9 -2 8 -3 1 -3 3 -3.7 -40 -41 4 0 -37 Non interestcurrentaccount 0 9 0 6 1 0 0 8 0.4 0 0 -0 5 -0 8 -09 -0 7 -03 TABLE 2 : GOVERYMENT FINANCES 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 I" x Dmct Taxes(0ther than SSC) 6 7 6 9 6 9 7 0 7 0 7 0 7 1 7 1 7 1 7 1 7 1 socialsecuntyconhbutions 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 Taxes-IntemationalTrade 2 u 1 9 1 7 1 4 1 1 n y n 8 8 8 0 8 a x 0 8 Others Indirect Taxes 1 2 7 126 1 2 6 126 1 2 6 126 126 1 2 6 126 116 126 Non Tax Revenue 3 1 2 7 2 7 2 7 2 7 2 7 2 7 2 7 2 7 2 7 2 7 pnvatmatlon receipts I 1 0 2 n u o n n o on o n n o n o n n o n Total Revenues (excl grants) 308 294 291 289 286 284 283 283 283 283 283 Consumption 139 137 137 137 137 137 137 137 137 137 137 InterestPayments 3 1 3 1 2 8 2 7 2 7 2 7 2 7 2 8 2 8 2 9 2 9 Subsidies 2 5 2 4 2 4 2 4 2 4 2 4 2 4 2 4 2 4 2 4 2 4 Currenttransfers 5 4 5 5 5 4 5 4 5 4 5 4 5 4 5 4 5 4 5 4 5 4 CurrentExpenditures 249 247 243 242 242 242 242 242 243 244 244 Capital Expenditures 8 4 8 0 8 0 8 0 8 0 8 0 8 0 8 0 8 0 8 0 8 0 Total Expenditures 332 328 323 322 322 322 322 322 323 324 324 Deficit excludinggrants andpnvatization -36 -35 -32 -33 -36 -38 -39 -39 -40 -40 -41 PRIMARY DEFICIT (excl grantsandpnvatizat -0 5 -04 -0 4 -0 6 -0 8 -1 0 -I1 -11 -1 1 -1 1 -I1 TABLE 3 : DEBT in X memoTotal debt 61 3 596 577 56 5 55.8 553 549 54.6 544 540 53 8 memo Domestic Public debt 225 218 21.5 21 0 21 1 228 240 257 27.1 28.6 2 9 9 memo ForeignPublic debt 389 37.8 362 355 34.7 325 309 290 273 25.4 23.9 PrivateDebt 125 11.8 10.2 9 1 8 8 103 117 140 161 185 200 Total Foreignpublic andprivate debt 51 3 496 464 446 43 5 428 42.1 43 0 43 5 43.9 439 Debt Service/ Total Exports 147 132 141 131 125 11.7 9 8 11.3 125 113 1 1 6 Debt Service 7 4 6.8 7 3 6 8 6 6 6 2 5.2 6.1 7 0 6.4 6 7 64 LowCASESCENARIO TABLE 1: MACROECONOMICFUNDAMENTALES 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 in % Growth Rate of GDP 1 9 5 4 6 0 0 9 0 2 I O 2 1 2 9 3 8 4 8 5 7 ConsumptionGrovnh 4 8 5 5 5 3 3 4 1 4 2 5 3 3 3 9 4 1 4 2 5 2 PublicConsumption 5 2 4 7 6 0 2 7 2 6 2 6 2 6 2 7 2 6 2 7 2 6 PrivateConsumption 4 7 5 7 5 1 3 6 1 2 2 5 3 4 4 1 4 4 4 5 5 7 InvestmentGrowth(GDI) -44 7 6 8 4 -36 -06 1 8 2 6 3 5 4 2 5 0 6 1 Import(GNFS) growth -15 8 8 S O -20 -10 3 0 4 5 5 6 7 1 6 7 7 6 Export(GNFS) growth -28 7 7 8 1 -40 -30 0 0 2 5 4 0 7 0 8 1 8 6 Inflation(p a) 1 6 1 8 2 2 2 2 2 4 2 4 2 5 2 5 2 4 2 4 2 4 in % ofCDP RealInterestRateForeign 2 1 2 5 1 9 2 1 2 3 2 3 2 4 2 5 2 7 3 0 3.3 Real InterestRateDomestic 3 1 3 5 3.7 3 7 3 7 3 7 3.7 3 7 3 7 3 7 3 7 InveSment 26.2 267 27.3 26.1 25.9 26 I 262 264 265 265 266 ExpoRs 44.7 45 6 464 440 426 42 1 422 426 43 9 453 466 Imports 48.8 49.7 504 486 479 489 50 1 51 4 53 1 54 1 552 CAB (exclall official transfers) -3 5 -3.5 -3 0 -3 3 -3.9 -5.0 -6 0 -6 9 -74 -7 3 -7 3 No" interestcwent account 0 9 0 6 0 9 0 5 0 0 -12 -20 -2.8 -3.1 -27 -24 TABLE2 : GOVERNMENTFINANCES 2002 2003 2004 2005 2006 200- 3008 2009 2010 7011 2012 in % Direct Taxes(Gther than SSC) 6 7 6 9 6.9 6 5 6 3 6 5 6 7 6 8 6.9 7.0 7.0 socialsecurity contributions 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5 1 5.1 5 1 Taxes-IntemationalTrade 2 0 1.9 1 7 13 1 1 0 9 0.8 0.8 0 8 0 9 0.9 Othen Indirect Taxes 12.7 12 6 12 6 126 126 126 126 126 126 12.6 12 6 NonTax Revenue 3 1 2 7 2.7 2 7 2 7 2 7 2 7 2 7 2 7 2 7 2.1 privatizationreceipts 1 1 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Total Revenues (excl grants) 30 8 29 4 29 1 283 278 279 280 281 282 284 28 4 Consumption 13 9 13 7 13 7 13.9 142 145 145 145 14.4 14 I 13 7 InterestPayments 3 1 3 1 2.8 2 9 3 0 3 2 3 4 3 6 3 8 4 0 4 0 Svbsidirr 2 5 2 4 2.4 2 4 2 4 2 4 2 4 2 4 2.4 2 4 2 4 Currenttransfen 5.4 5 5 5 4 5.4 5 4 5 4 5 4 5 4 5.4 5 4 5 4 CurrentExpenditures 24 9 24 7 24.2 24.5 25 I 25 5 25 8 25.9 260 25 8 25 5 CapitalExpenditures 8 4 8 0 8.0 8 0 8.0 S O 8 0 8 0 8 0 8 0 8.0 Total Expenditures 33 2 32 7 32 2 325 33 1 33.5 33 8 339 340 33 8 33 5 Deficit excludinggrants andprivatization -3 6 -3.5 -3 2 -42 -5 2 -5 6 -5.8 -5.8 -5.8 -5 4 -5.1 PRZMARYDEFICIT (excl grants and privatization) -0 5 -0 4 -0 4 -1 4 -2 2 -2 4 -2.3 -2.2 -1.9 -1 5 -1.1 TABLE3 : DEBT 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 in % Total Public debt 61.3 59.6 57.7 59.9 63.6 67.1 69.9 72.2 73.6 73.7 73.2 DomesticPublic debt 225 21 8 21 5 227 25.3 296 33 0 367 396 41 9 43.1 ForeignPublic debt 389 37.8 362 372 383 375 369 355 34.1 31 9 30 I PrivateDebt 125 118 10.3 9 1 8.9 11.8 151 198 248 297 336 Total Foreign public and private debt 51.3 49.6 46.5 46.3 47.3 49.3 52.0 55.3 58.8 61.5 63.1 DebtService/ Total Exports 14.7 13.2 14.1 144 149 148 127 151 175 165 175 DebtService 7 4 6 8 7.3 7.2 7 2 7 2 6 2 7 5 8.9 8.6 9 3 65 FISCALADJUSTMENTSCENARIO TABLE 1 : GOVERNMENTFINANCES 2007 2008 2009 2010 2011 in % Direct Taxes(0ther than SSC) 6.7 6.9 6.9 7.0 7.0 7.0 7.1 7.1 7.1 7.0 7.0 social security connibutions 5.1 5.1 5.1 5.1 5.1 5.1 5.1 5.1 5.1 5.1 5.1 Taxes-Intemational Trade 2.0 1.9 1.7 1.3 1.1 0.9 0.8 0.8 0.8 0.9 0.9 Others IndirectTaxes 12.7 12.6 12.6 12.6 12.6 12.6 12.6 12.6 12.6 12.6 12.6 NonTax Revenue 3.1 2.7 2.7 2.7 2.7 2.7 2.7 2.7 2.7 2.7 2.7 privatization receipts 1.1 0.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Total Revenues(excl. grants) 30.8 29.4 29.1 28.8 28.6 28.4 28.4 28.4 28.4 28.4 28.4 Consumption 13.9 13.7 13.7 13.5 13.5 13.4 13.1 12.7 12.3 11.7 11.1 Interest Payments 3.1 3.1 2.8 2.9 3.0 3.1 3.2 3.2 3.2 3.2 3.1 Subsidies 2.5 2.4 2.4 2.4 2.4 2.4 2.4 2.4 2.4 2.4 2.4 Current "fen 5.4 5.5 5.4 5.4 5.4 5.4 5.4 5.4 5.4 5.4 5.4 Current Expenditures 24.9 24.7 24.2 24.2 24.3 24.3 24.1 23.8 23.3 22.8 22.0 Capital Expenditures 8.4 8.0 8.0 7.9 7.7 7.6 7.5 7.4 7.3 7.1 7.0 Total Expenditures 33.2 32.7 32.2 32.1 32.0 31.9 31.5 31.1 30.6 29.9 29.0 Deficit excluding grants andprivatization -2.8 -3.6 -3.4 -3.5 -3.7 -3.7 -3.5 -3.1 -2.5 -1.7 -0.9 PRIMARY DEFICIT (excl. grants and privatii -0.5 -0.4 -0.4 -0.4 -0.5 -0.3 0.0 0.4 1.o 1.8 2.5 TABLE2 : DEBT 2007 2009 2010 2011 12 b x Total Public debt 61.3 59.6 57.7 58.9 60.9 62.3 62.8 62.3 60.8 57.8 54.0 Domestic Public debt 22.5 21.8 21.5 21.7 22.6 24.8 25.9 26.8 26.8 25.9 23.9 Foreign Public debt 38.9 37.8 36.2 37.2 38.3 37.5 36.9 35.5 34.1 31.9 30.1 Private Debt 12.5 11.8 10.3 8.1 6.3 7.1 8.3 10.7 13.1 15.4 16.8 Total Foreign public and private debt 51.3 49.6 46.5 45.3 44.6 44.7 45.2 46.2 47.2 47.3 46.8 Debt Service /Total Exports 14.7 13.2 14.1 14.1 14.2 13.8 11.1 12.6 13.9 12.0 11.9 Debt Service 7.4 6.8 7.3 7.2 7.2 7.0 5.6 6.5 7.4 6.5 6.6 66 ANNEX2. FISCAL RISKBEYONDTHE BUDGET DEFINING QUASI-FISCAL ACTIVITIES AND HIDDEN DEFICITS Quasi-fiscal activities (QFAs) can be defined as activities not explicitly executed via budgetary mechanisms, but that have, or may have in the future, measurable fiscal implications. Inanalyzing QFAs, we apply a comprehensive methodology proposed by Brixi and Schick (2002), classifying risks from government liabilities into four categories. Government direct explicit liabilities are obligations that fall due with certainty and are defined by law or contract. Government direct implicit liabilities represent a political, rather than legal commitment o f the government that will occur with certainty. They arise as a presumed consequence o f public expenditure policies. Explicit contingent liabilities represent the government's legal obligations to make a payment if a particular event occurs, such as state guarantees for non-sovereign borrowing. Finally, implicit contingent liabilities are those that are not officially recognized until a failure occurs. The trigger, the value at risk, and the size o f the government outlay are uncertain for this latter category. The following table offers a systematic view o f different types o f government liabilities: Liabilities Direct (Obligation in any Contingent (obligation ifa particular event occurs) event) Explicit 9Stateguaranteesfor nonsovereignborrowing and obligations Govemment issuedto subnational govemmentsandpublic and privatesector liability is entities (development banks) recognized 1 Umbrellastate guaranteesfor varioustypes ofloans (e.g. bylaw or mortgages, studentsstudyingagriculture, and smallbusinesses) contract . State guarantees (for trade and the exchangerate, borrowing by a Budget expenditureslegally foreign sovereign state, privateinvestments) bindinginthe longte State insuranceschemes (for deposits, minimum rcturns from (civil service salaries .. privatepension funds, crops, floods, war risk) servicepensions) . Futurerecurrent costs of Default of a subnational govemment andpublic or private entity public investmentprojects on non-guaranteeddebt and other liabilities Futurepublic pensions(as Cleanup of the liabilities of privatizedentities opposed to civil service Bank failure (beyondstate insurance) pensions)ifnotrequiredby Investment failure of a non-guaranteedpensionfund, law ... Implicit . A "moral" obligationof the govemment employment fund, or social security fund (social protectionof Social security schemes if small investors) not requiredby law Default of the centralbank on its obligations (foreign exchange Futurehealthcare financing contracts, currencydefense, balance o fpaymentsstability) ifnot specifiedbylaw .. that reflects public expectations and/or pressures by Bailouts following a reversal in privatecapital flows interest Residual environmentaldamage, disaster relief, military groups financine, and the like Quasi-fiscal activities give rise to hidden (off-budget) deficits, that cannot be captured by conventional indicators of fiscal deficit, which represent aflow concept. budgeting and reportingpractices. To redress this deficiency, a recent study by Kharas and Mishra 67 (2001) introduce a notion o f actuarial deficit, defined as the change in the stock o f government liabilities, i.e. the entire stock o fpublic debt and the money base: Dta = (Bt -Bt-I) + (Mt-Mt-1), where Dta i s the actuarial budget deficit inperiod t; Bt i s the stock o fthe public debt; and Mtis the base money. The difference between the actuarial andthe conventional deficits i s the hidden (off-budget) deficit. On the sample o f 32 countries, Kharas and Mishra demonstrated that there exists a much closer link between the level o f actuarial deficits and macroeconomic outcomes (approximated by the incidence of currency crises), than between the conventional deficit and these macro/fiscal outcomes. The following figures illustratethis: No. of Currency Crises versus Avg. No. of Currency Crises versus ConventionaI Budget Deficit Average Actuarial Budget Deficit 0 -5 0 5 10 15 -5 0 5 10 15 Average Conventional Budget Deficit Average Actuarial Budget Deficit Source: Kharas H.,Mishra, D. (2001) QFAs are a major problem in both developing and developed countries, since they utilize public resources in an inefficient and non-transparent manner, undermining accountability and economic growth. Often, as the budget deficit i s beingbrought under control, on-budget subsidies are replaced with quasi-fiscal ones, reflecting the unwillingnessandlor inability o f governments to deny support to structurally or socially important enterprises, regardless o f their performance. Quasi-fiscal activities generate unpredictable ex post outflows that crowd out vital public spending, particularly in the social sectors. Typically, connected and privileged groups use their high leverage to distort resource flows to their advantage via quasi-fiscal instruments. Off-budget operations can cause a major blow to the transparency o f public spending. The civil society (as well as the decision-makers) may be unaware that contingent obligations are driving the fiscal sphere beyond sustainable boundaries. The poor and disadvantaged 68 social groups in whose name some o f the quasi-fiscal operations are being conducted (e.g. investment projects inpoor regions that benefit from government guarantees) have no means to cope with the fallout o fthese operations once the fiscal crisis erupts. Quasi-fiscal operations are particularly dangerous for emerging economies. Governments inthese countries are naturally tempted to issue contingent liabilities, given that they are constrained in issuing additional direct obligations by their weak creditworthiness. The pressure to engage in quasi-fiscal activities comes from the vast needto generate growth andprovide social protection with scarce available resources. Traditional cash-based budgeting and accounting practices encourage the governments to use quasi-fiscal tools, since their cost appears to be zero under these practices, making quasi fiscal instruments politically more attractive than cash expenditure items. A particular weakness o f conventional budgets is that off-budget activities often create long-term obligations, while traditional budgets reflect only obligations incurred during only one year, and only from the narrow cash flow perspective. Cash flow budgeting typically enables fiscal opportunism: contingent liabilities are not recognized in the budget until the trigger (risk) event occurs. Conventional budgeting creates incentives to shift the costs to future generations, and/or from one social group to another, thus encouraginguncontrollable spending. APPROACHES THE MANAGEMENTOFQUASI-FISCAL RISKS TO The first condition is to identifi, classifi, and understand the fiscal risks facing the government. This Comprehension o f the fiscal risks and their consequences will encourage the government to avoid risks that are bound to surface in a politically meaningful time horizon. For risks that extend beyond that timeframe, achievement o f fiscally sound behavior may depend on coercion. In particular, policymakers are more likely to gravitate to fiscally sound decisions if the media, the public, investors, credit- rating agencies, and multilateral institutions understand the government's fiscal performance in its entirety and if there are sanctions when the government exposes the state to excessive risks and conceals those risks. To achieve sound fiscal performance, the government needs to include in its fiscal analysis and decision-making the fiscal risks relating to future possible obligations o f the state and to consider those fiscal risks inthe context o f its risk preference, risk financing, and risk management capacities. The following steps are critical. Considerfiscal performance infull, that is, beyond the budget and debt. Fiscal analysis i s complete only if it factors in the cost o f the implicit subsidies in the government's contingent support programs. In particular, the government cannot separate the analysis o f its fiscal position from the obligations it has undertaken outside the budget. Moreover, the government may have mismanaged some institutions to finance and implement its policies outside the budget system. A string o f years with a balanced budget and low public debt suggests neither that the government has been fiscally prudent nor that there will be fiscal stability. Indeciding between alternative forms o f support, the government needs to consider the medium-term fiscal impact and allocative and operational efficiency of programs pursuedoutside the budget to the same extent as with the budget. 69 Medium-term fiscal forecasts, the budget itself, and government financing and borrowing plans are truly viable only ifthey provide for contingent and other fiscal risks. Identij), classij), and analyze all fiscal risks in a single portfolio. To understand and prepare for the entire range o f potential fiscal pressures, policymakers will have to take stock o f all programs and promises and identify and classify the main sources o f fiscal risks, as shown in the fiscal risk matrix. For each item o f the fiscal risk matrix and in order o f significance, the government needs to analyze the risk factors and ways to control and reduce its exposure to the risks. Qualitative analysis o f risks would help the government formulate and design sound new programs andpromises. The government should consolidate the stock of contingent liabilities into a single portfolio, along with state debt and other public liabilities, so that it can evaluate correlations, sensitivity to macroeconomic and policy scenarios, and overall risk exposure. A single portfolio allows the government to relate its contingent liabilities to its comprehensive risk strategy and guidelines regarding risk exposure, asset and liability management, hedging, and benchmarking. As an input into the analysis o f risk exposure, the government would also analyze information about budget arrears, state guarantees, state insurance programs, subnational government borrowing, obligations o f state-owned and state-guaranteed institutions, effects o f private capital flows, and similar factors. In contrast to the deficit and debt constraints, indicators reflecting a comprehensive analysis o f the government's exposure to fiscal risks would have greater predictive value for fiscal stability. Determine the government's optimal risk exposure and reservepolicy based on its risk preference and risk management capacity. The government needs to base its risk and reserve strategy on its overall risk exposure, risk preference, and ability to manage risk and absorb contingent losses. Ideally, the risk strategy would be tied to the risk preference o f the median voter. Similarly, the government would assess new programs based on their marginal impact on overall risk exposure and fiscal outlook. It would agree to further contingent and implicit forms o f financial support only to the extent it i s able to evaluate, regulate, control, and prevent the risks. Ifthe government has a low capacity to evaluate and manage risks, the best approach is to favor direct subsidies and provision o f services rather than guarantees. That is, assuming that the government's intervention in a particular area i s justifiable, the government would opt for budgetary financing o f its intervention rather than ensuring that particular outcomes will be delivered by the private sector. To this end the government has to enact guidelines for prudent and sound fiscal management. An institutionalframework for publicfinance will encourage thegovernment topursue soundfiscal performance only if it encompassesboth direct and contingentfiscal risks. A framework for public finance management that ignores the hture fiscal implications o f contingent liabilities and other off-budget commitments will only make such forms of government support look inexpensive and politically attractive. An adequate institutional framework requires that the government treat any non-cash program involving a contingent fiscal risk as it does other budgetary or debt items from the viewpoint o f aggregate fiscal stability and allocative and technical efficiency, control, public 70 disclosure, and accountability. Rules for issuing, monitoring, and handling state guarantees and insurance programs and for monitoring and financial management o f public, state-guaranteed, and subnational government institutions are also needed. As the role o f the state transforms from direct provision o f services to guarantees against residual risks, governments need to follow the example o f the private sector indeepening their capacities for fiscal analysis andmanagement beyond the state budget and debt. Internalize and disclose the full fiscal picture. The rules and practices applied in the budget process, financial management, and public accountability framework determine how much flexibility the government has to assume immediate and future direct and contingent unbudgetedobligations. Optimally, government choices will reflect qualitative and, where possible, quantitative evaluation o f the future outlays and risks associated with alternative foms o f government support, including programs outside the budget such as guarantees and activities of state-guaranteed agencies. To address theproblem of government accountability andfiscal discipline outside the budget, public disclosure is more important than full-fledged accrual-based accounting, budgeting, and risk measurement systems. Particularly for governments with lower institutional capacities, the system should require them to assess risk factors, make rough provision for contingent risks in the budget, and publish a statement o f contingent liabilities and overall risk exposure. Such a system is more sensible than the optimal institutional framework, which involves accrual-based accounting and budgeting standards and sophisticated risk measurement methodologies. Public disclosure of fiscal information extending beyond the budget and direct debt enables the public and markets to monitor the government's full fiscal performance, including the fiscal risks accumulated outside the budget. Market agents such as investors and credit rating agencies are then able to take both direct and contingent fiscal risks into account intheir analysis and investment decisions. Their ability to do so intum indirectly encourages budgetary and overall fiscal discipline. In addition, greater fiscal transparency facilitates parliamentary scrutiny and monitoring by the market, particularly byinvestors and sovereign credit ratingagencies, andbyinternationalinstitutions. Monitor, regulate, and disclose fiscal risks to the public and private sectors. Governments reduce the fiscal risks when they strive to prevent market failures and minimize the moral hazard associated with their programs, commitments, and residual responsibility for market failures. To reduce moral hazard and failures inthe markets, the government maintains regulatory and law enforcement systems, monitors the systemic risks in both the private and public sectors, and enforces transparency about the risk exposure o f both financial and non-financial institutions in the markets. Well-developed regulatory and public disclosure systems are particularly important when government embarks on privatization while assuming an explicit or implicit obligation to cover residual liabilities and ensure that private agents achieve particular outcomes. Prevention offiscal risks depends on a combination of analytical tools, incentives, and the capacities of parliamentarians, civil servants, regulators, supervisors, international institutions, and market agents. Potentially the best place to develop risk monitoring 71 capacities is inthe central bank, given its role incollecting balance-of-payments data and, inmany instances, supervising banks. Specific regulatory and supervisory agencies such as the securities and exchange commissions may best handle the monitoring o f specific risks. Ultimately, the ministry o f finance and the office responsible for public liability management would handle the monitoring and prevention o f the government's overall risk exposure. Undertake measures to reduce thefiscal risk of individual government programs and promises. Whether the government's programs, promises, and exposure to fiscal risks are appropriate depends on their consistency with government policies and actions. The following aspects o f consistency in particular influence a government's fiscal performance: 9 Consistency o f government programs and promises with the stated role and strategic priorities o fthe state 9 Consistency in the eligibility and management standards applied across government programs over time 9 Consistency o f the risks assumed and reserves provisioned under a program with the risk management capacities o f the government Consistency between the authority o f policymakers to assume contingent fiscal risks and their accountability. Steps to control the risk of individual government programs could be asfollows. I.Before GovernmentAdmitsan Obligation Assess how the obligation fits thepronounced role and strategicpriorities of the state. What types o f support the government decides to offer both outside or through the budget define the actual role o f the state. Programs outside as well as inside the budget should, in principle, be subject to the same type o f policy analysis and consideration. Inthe case o f contingent support programs such as guarantees for state institutions and funds, the government must consider whether their objectives fit within its announced role and priorities and whether theyjustify the potential long-term fiscal costs. Consider the choices of policies and forms of support relative to the associated financial risks and government's risk preference and risk management capacity. The quality o f fiscal performance benefits when the government acknowledges the cost o f uncertainty about future public financing requirements in considering alternative programs and forms o f support for particular policy objectives. As with corporations, an unexpected requirement for financing disrupts financial planning and increases the cost of borrowing or, in a worse case, runs the risk that no credit financing i s available. Governments need to evaluate alternative ways to implement their policies not only on the basis o f their potential cost andbenefits but also on the extent o f the uncertainty they involve for future public financing. The government would judge contingent forms o f support interms of the extent o f the asymmetric information and transaction costs. These 72 considerations would be made in the context o f the government's own risk preferences andriskmanagement capacities andthe reliability o fits access to ad hoc borrowing. Define and communicate the standards for and limits of government involvement to minimize the moral hazard. It i s not so much the budgeted expenditures but the contingent liabilities, particularly the implicit ones, as understood by the public and markets, that define the outer limits o f state responsibilities and affect the behavior o fthe public and market agents. The more formally and precisely the government defines and signals its responsibilities (its area o f commitment), the more distinct are the explicit liabilities and the smaller are the implicit liabilities. The more credibly the government defines its responsibilities and the pain market agents will bear in cases o f their failure and reliance on government rescue, the less i s the problem o f moral hazard. Take the example o f a society where the government has a strong tradition o f extensive public services. In such a case the central government may be expected to take over any obligations o f subnational governments introubles. Such an expectations raises a scope for moral hazard on the side o f subnational governments. The central government can reduce the moral hazard by signaling that it will only ensure the delivery o f core services to citizens o f insolvent subnational governments. At the same time, it can state that it will not bail municipalities out from their debts and non-core expenditure obligations. Evaluate the risks of programs individual& and in a singleportfolio that also contains existing risks, estimate the potential fiscal cost of each obligation, and set additional reserve requirements. Qualitative analysis o f the risk factors in alternative government programs and estimates o f their potential long-term fiscal costs and hidden government subsidies prior to any commitment helps optimize the choice and design o f government programs. Rough quantification o f the risk and potential fiscal cost o f government contingent liabilities and commitments requires good qualitative analysis o f the underlying risks. Specialized methodologies such as option pricing and value-at-risk are o f great value in deriving a more precise estimate o f the potential costs o f a particular program. According to government reserve policy, the risk exposure o f a proposed program added to overall government exposure determines the amount of additional resources that should go into the governmentreserve fund. Design the program well to protect the government against risks. Based on the qualitative risk analysis, the government needs to identify those risks it can control reasonably well, decide which risks to cover under its proposed program o f contingent support, and develop effective risk-sharing, regulatory, and control mechanisms to monitor the performance o f the parties under the program. Apart from exogenous risks such as drought, the government faces endogenous risks that are mainly a function o f program design. A poor design can create varying levels o f market distortion and moral hazard, whereas a good design can reduce the potential fiscal cost o f the program. An example is a partial guarantee that covers only non-commercial risks, less than 100 percent o f the value of the potential loss, and the last rather than the first portion o f the loss. Programs that involve implementation by an intermediary agency that itselfmust be established, such as a guarantee fund o f any sort, are more difficult to design, particularly interms ofmanagement incentives andperformancemonitoringbythe government. 73 2. When the GovernmentAccepts and Holds an Obligation Stick to thepre-set limits of government responsibilities.After the government approves a program or commitment, the main challenge i s to ensure that the markets and public do not expect any state support beyond the announced limits over the life o f the obligation. Any indication that the government might provide financial support beyond the announced limits will raise the moral hazard for and distort the behaviors o f the parties potentially benefiting from the program. Budget, account, and disclose the obligation. On the institutional side, the government faces the challenge of budgeting, accounting and provisioning for, and disclosing the obligation adequately. How does it ensure that no unknown contingent liability appears only after it i s triggered? For instance, the public finance law can state that an obligation i s valid only if it was assessed, budgeted, accounted, and, above all, disclosed at the time o f its adoption by government. Monitor the program risk factors and reserve-fund adequacy. Over the life o f an obligation, the government needs actively to monitor the program's risk factors, the performance o f the agents under the program and, inthis context, also the adequacy o f its reserve funds. Monitoringo f intermediary agencies, such as banks and various credit and guarantee funds that the state uses to implement its policy objectives and guarantees, is particularly important. If the government lacks a good monitoring capacity, it can contract this task out for a performance-based fee. The cost o f monitoring and administering programs o f contingent support may be relatively high and should be reflected inthe ex-ante calculations o fthe potential fiscal cost o f a program. 3. After a Liability Falls Due Execute the obligation within its pre-set limits and identifv lessons for future policy choices. It is critical that the government meet an obligation when it falls due within the stated limits, particularly in terms o f the credibility o f future announcements and the scope for future moral hazard in the markets. For instance, paying depositors more than the specified deposit insurance levels tells the markets that the government will submit easily to political pressure, tells depositors that banks offering higher yields are "safe," andtells the bankingsector that excessive risks are worth taking. By applying the lessons from its involvement with direct and contingent liabilities, both explicit and implicit, the government is able to adjust its role incrementally, rather than abruptly, in a crisis. A timely and credible explanation o f any adjustment in the state's role that will affect future policy choices will prompt the public and market agents to adjust their expectations and behavior. For example, by explaining that the public pension scheme i s not fiscally sustainable and that future governments will have to reduce the pensionbenefit significantly, the government influences the saving behavior o fpeople. Ifanobligation isimplicit, assesswhetheritcoincideswiththestate'sannouncedrole and promotes desired market behaviors. When public interest groups or market agents suddenly call on the government to extend more support than was originally specified, 74 policymakers need to ask whether extending that support coincides with its announced role and how it affects future behavior in the markets. The long-term damage to the government o f acting upon an ad hoc request may sharply exceed the potential short-term benefit. Acting upon ad hoc requests may, however, be politically attractive, and the government is often able to find ways to improperly use financial institutions and funds outside the public sector to implement and finance its actions. Thus the public, investors, and international authorities needto monitor the government's responses to ad hoc claims o f an implicit government liability and apply sanctions for fiscally irresponsible choices. Compare and report the estimated and actual cost of government support, evaluate performance, and apply sanctions for failures. The requirement that the government report and compare the ex-ante risk evaluation and actual layouts for a program i s critical to government accountability. Performance evaluation applies to government departments and officials as well as to the parties under a program. Sanctions may involve government officials (the case where particular interests distorted the ex-ante risk analysis), the managers o f state-guaranteed and intermediary agencies implementing the government's programs (such as for exposing the government to unnecessary and excessive risks), and the parties under the program (where they breached an agreement). 75 ANNEX3. FUNCTIONAL COMPONENTSOF THE GOVERNMENT DEBTOFFICE A prerequisite for efficient management o f sovereign liabilities is a well specified organizational stnicture. Debt management i s most effectively organizedby function, not by product namely into Front, Middle, and Back Offices. This organizational structure must be supported by clear objectives, delegation o f authority and reporting, controls, performance monitoring, strong code o f ethics and soundjudgrnent.The boundaries o f the Middle Office (MO) in a sovereign debt management structure vary from country to country, and are typically defined "by default": the competency o f the MO covers all issues which are inthe competence of neither the Front Office (FO), nor the Back Office (BO). On the contrary, the boundaries o f these two departments are quite standard in differentnational environments andare typically as follows: Front Office ......... . Fundingmanagement -implementationo fdebt management strategy Participates inthe design of the fundingstrategy Concludes the transactions: follow-up o f the financial markets, preparation, conclusion Contacts with the FO o f the counter parties (lenders) Investor relations andmarketing(to investors bothdomestic and international) Communication (with other Treasuries, the public, the press, etc.) Newproduct development (inclose coordination with the MiddleOffice) Portfolio management - implementation o f decisions to rebalance the portfolio bringingit closer to strategic benchmark, and o fthe derivatives contracts, In some countries the FO manages the daily cash position o f the Government (contacts with the CentralBank as the fiscal agent o f the Government) Competence o f the FO stops when the ticket is written and transmitted to the Back Office Insomecountries, adedicated legalunitoperates fromthe FO. Back Office .. ... Executes the transaction Makes sure that sufficient money i s available inthe Treasury account Validates data from the FO Sends instructions to the netting, clearing and settlement systems Reports transactions to the legal reporting Authority and to internal authorities (audit, Board) 76 fl Takes care o f accounting and translates the transactions into budget execution entries fl . Keeps the securities accounts Insome countries, adedicated ITunitisplacedwithinthe BO. The main reason for separating the Front and Back Office functions is to ensure that staff executing transactions are different from staff entering them into the accounting system, eliminating incentives to buildan excessively risky exposure. Middle Offlce The core competence o f the MO is the formulation of the State's medium-term debt management strategy, to be adopted by the Authorities (e.g. by a Government's decision endorsed by the Parliament). The MO translates general objectives for debt management into manageable and quantifiable targets and guidelines. The debt strategy is comprised o f the issuance (and eventually buy-back) programs (what instrument to be issued andor bought-back, at what maturity, in what currency and in what proportion, when, etc.) set inline with the overall borrowing objectives o f the State; communication and marketing strategy; and new product and infrastructure development program. Portfolio risk management objectives (criteria to decide on the choice between domestic and foreign borrowing, on average maturity, duration, maturity profile, etc.) should also be part o f the debt strategy. First, simple benchmarks should be developed. More advanced measures like "Budget/Cost-at-Risk", which quantify the government's risk, should be introduced, when the environment becomes more sophisticated. The strategy i s then translated into General Directives to the debt management unit, to be signed by the competent Authorities. These Directives constitute the framework for the debt managers' actions. Other MO competences could include: . Risk analysis and management. This competence is closely linked to the strategy formulation, and they should be viewed together, especially if the MO makes use o f a mathematical model (benchmark debt portfolio, standard benchmark, etc.). In such a model, risks which cannot be managed are treated as "constraints". It is . strongly recommended that these constraints be accurately defined and politically accepted. Risk analysis must be undertaken ina multiyear environment; Risk control: Monitoring the activities o f the FO on the basis o f pre-determined risk exposure ceilings, to prevent the FO from positions that are out o f line with the riskprofile approved by the Government; Sustainability and consistency analysis o f the government borrowing strategy (in close coordination with the analytical units inthe M O F and the Central Bank), to assure the consistency o f the government borrowing policy with the priorities o f medium-term macroeconomic, fiscal, and monetary policy; 77 m Legal: maintaining documentation for foreign debt, syndications, adaptations to the current legislation with regard to the development o f the domestic market, collateralization, ISDA contracts, etc. New product development (In close partnership with the FO: proposals for improvements to the primary dealers system, secondary market regulations, development of retail debt products, etc.) Managing off-budget risks (monitoring other public debt issuers such as municipalities and public enterprises; managing the government guarantee portfolio, etc.). M O competencies can range from minimal (collecting information) to coercion if necessary, depending on the political set up, the seriousness o f risks, and the feasibility o f control instruments. In some countries, the MO collects information and offers advice on the timing o f large private sector issuers; 4 Inputs into Budgeting: Preparing the inputs into the state budget and updating the M O F regularly about the implementation o f its borrowing program. A strong link with the Budget Formulation Department of the MOF is required, to monitor the evolution o f the financing needs o f the State. These projections should include estimates o fbudget costs on contingent liabilities; Reporting: preparing the Annual Debt Report (as part of the budget execution report to the Parliament), quarterly updates, and operational reports to the Management o f the Treasury and other decision makers (the Cabinet, the MOF, etc.); Research on public liability management andrelated issues. 78 ANNEX4. BENCHMARKING TUNISIAAGAINST OTHEREMERGINGECONOMIES The two comparator groups used in this study are: Other MENA (Algeria, Djibouti, Egypt, Iran, Jordan, Lebanon, Morocco, Oman, Syria, and Yemen), and Non-MENA debtor countries (Chile, Dominican Republic, Hungary, Malaysia, Pakistan, Philippines andRomania). The reasons for selecting these groups are as follows. Other MENA group. Region-wide data aggregation proved to be impossible due to irregular data quality, and to major structural differences in financial position o f oil- exporting countries. Instead, a representative sample o f MENA countries was constructed, in which all three analytical categories o f debtor countries are represented: severely indebted countries (Jordan and Syrian Arab Republic); moderately indebted (Algeria, Lebanon, Yemen, and Tunisia itself) and less indebted (Djibouti, Egypt, Iran, Morocco, and Oman). Tunisia i s classified bythe World Bank as moderately indebted middle income country. All countries are classifiedbythe World Bankbythe levelo ftheir external debt into: Severely indebted means either o f two key debt sustainability ratios is above critical levels: present value o f debt service to GNI (80 percent) andpresent value o f debt service to exports (220 percent). m Moderately indebted are countries where either o f the two severely indebted key ratios exceeds 60 percent of, but does not reach, the critical levels; and Less indebted are countries in which either o f the two key ratios i s below 60 percent. Middleincome economies are those inwhich 2000 GNIper capitawas between $756 and $9,265. Low income economies are those inwhich 2000 GNIper capita was $755 or less. Indicators for the two comparator groups are simple (non-weighted) averages. 79 REFERENCES Bento, Vitor (1999). Setting-up a Debt Agency, Mimeo. BenBraham, Mehdi(2002). Quellertiformepour lesystBmede retraite en Tunisie?Lecons des expkriences ktrang&es, mimeo, Caisse des DCpBts et Consignations, Paris. B6z6s, L.(2002) Hungarian Public Debt Management, World Bank http://www.worldbank.ore;/pdmlpdf/buzasdebt.pdf Currie, E., Dethier, J.-J., Togo, E. (2003). Institutional Arrangements for Public Debt Management. WorldBank DanishNational Bank (2003). Danish Government Borrowing and Debt. IMF.(2002). "Liberalization ofthe Capital Account inTunisia-Progress Achieved and Prospects for FullConvertibility", Selected Issues, Article IV Consultation, May, Washington DC. IMF (2003). "Tunisia: Strengtheningthe Monetary Framework and Instruments", SelectedIssues, Article IV Consultation, July, Washington DC. Kharas, Homi and DeepakMishra. 2001. Fiscal Policy, Hidden DeJicit, and Currency Crises. World Bank Economists' Forum: 31-48. Cassard, M.,Folkerts-Landau, D.(1998) Risk Management of Sovereign Assets and Liabilities in Global Financial Markets . IMF. Polackova Brixi, H.,and Schick, A. (2002). Government ut Risk. OxfordUniv.Press. Schinasi G.J., and SmithR.T.(1998). "Fixed-Income Markets inthe UnitedStates, Europe, and Japan: Some Lessons for Emerging markets", IMF, Working Paper 981173, dkcembre, Washington D.C. Vittas, Dimitri(2000). Pension Reform and Capital Market Development: "Feasibility and " "Impact Preconditions, The World Bank, Washington D.C. " Wilton, David (1999). Developingthe Money Market, mimeo, The World Bank, Washington D.C. World Bank (2001). Republic of Tunisia: Review of External Sustainability in a Changing Environment, MENA Region, WashingtonD.C., January. World Bank I IMF (2001a). Guidelinesfor Public Debt Management World Bank I IMF (2001b). Tunisia-Financial Sector Assessment Program, Washington D.C., November. World Bank (2002). Tunisie: Promotion de l'investissementprivk et le r d e de lapolitique d 'incitations: Analyse et recommandations, FIAS,Washington D.C., September. World Bank I IMF(2002). Developing Government Bond Markets: A Handbook, Washington D.C. World Bank (2003). Republic of Tunisia: Employment Strategy, Preliminary draft, Washington D.C., February. 80