risk management for sovereign debt

Sovereign Debt Risk Management Part 1 – Risk Management

Chapter 1 – Risk Management for Sovereign Debt 1-1

Chapter 1

Risk Management for Sovereign Debt


1-2 Chapter 1 – Risk Management for Sovereign Debt



Chapter 1


1.1 Introduction ......................................................................................................1-5

1.2 Setting the Goal and Establishing Constraints ...................................................1-5

1.3 Classes of Risk ..................................................................................................1-7

1.4 Risk As a Macroeconomic Problem .....................................................................1-9

1.5 Risk Management of Contingent Obligations ....................................................1-10

1.6 Risk Tolerance and Performance Measurement ................................................1-15

1.7 Procedures for Managing Market Risks and Returns ........................................1-19

1.8 A Structured Approach to Organization and Information Architecture ..............1-20

1.9 Organization Structure ................................................................................1-22

1.10 Information Architecture Using Separate but Integrated Databases ................ 1-28



Chapter 1


1.1 Introduction

This document provides a framework to help sovereign debt managers meet their goal – sustained ability to fund fiscal needs – while remaining within funding policy objectives and risk targets. This is the context for assessing whether the compensation for risk exposure is adequate. This is an optimizing framework.

This framework helps promote efficient management of the sovereign’s debt. Given the constraints and risk limits, are borrowing costs as low as feasible and is risk exposure rewarded appropriately in light of market conditions? But the relevant concept for the sovereign is not simply being compensated for risk exposure. It is far beyond that because sovereign debt has important macro-economic impact. At one extreme, an inappropriate exposure structure in debt and sovereign guarantees has caused some countries’ economic downturns to become crises, whereas a well-implemented optimized debt and guarantees structure can actually help promote cyclical stability. Moreover, a well designed debt structure can help the sovereign promote its own local currency capital market, with associated benefits for home ownership and business financing, and it can help provide instruments for effective monetary policy transmission.

This chapter describes the thinking process that underpins an optimizing approach to sovereign debt management. It describes the important first step, setting the goal and establishing the limits. It then describes the types of risks. Effective risk management is one of the important tests of a borrower’s credit rating, which ultimately influences the borrower’s availability and cost of funds relative to other borrowers.

Two of the principal risks facing all debt portfolio managers are market risk and operational risk. Market risk can be measured, which requires a coherent data structure and tools for measuring and managing risk. Operational risk arises largely from a weak organizational structure. Operational risk can be managed through an organizational structure that provides for appropriate separation of responsibilities and uses well-defined procedures for accurate and timely accounting.

This chapter discusses the various types of risk and how they can be managed. It then introduces an approach to assessing and reviewing the structure of the organization and its information requirements. It proposes an information management system that suits each of the two main functions for managing debt — accounting/budgeting and policy/risk management—in a separate but integrated way. This chapter also introduces stress testing for risk measurement and management. Both of these are discussed in more detail in technical chapters that follow this one.

1.2 Setting the Goal and Establishing Constraints

Sovereign debt is an explicit contractual obligation of the sovereign to repay, usually with interest and at a specified time, funds that have been provided to it. The obligation ultimately takes the form of a contract to borrow. In one case, the obligation might be ‘direct’, meaning that there are only two parties involved: the sovereign (as borrower) and the provider of funds (the lender), and the full faith and credit of the sovereign supports the agreed repayment terms. In another case, the



obligation might be ‘indirect’, meaning that a third party, such as an agency of the sovereign, might be the nominal borrower and there might not be a contractual underpinning of the full faith and credit of the sovereign. A third is that the obligation might be ‘contingent’, meaning that a third party is the borrower with the direct obligation to repay, and the sovereign’s obligation to repay contractually arises only upon some event such as the failure of the direct borrower to repay. A guarantee is a contingent obligation.

In any case, the obligation to repay might be in the sovereign’s domestic currency (“domestic currency debt”) or it might be in one or more external currencies (“foreign currency debt”). Foreign currency debt may have all the risks and rewards of domestic currency debt. Moreover, it has an additional risk/reward component: there is the potential of either an adverse or a beneficial change in the foreign exchange rate relative to what was accepted when the exposure was agreed, resulting in debt servicing being either more onerous or less costly than anticipated.

We can think of sovereign debt management as having two different dimensions that must be considered. One dimension has all the aspects of any portfolio management problem. Where those pertain, the best approach for asset-liability management as practiced in any financial institution can be used. There is one small technical modification, however, in that the appropriate measure for a sovereign is the cash flows — the sources of funds to service the debt and the amount of funds needed to service the debt. The second dimension is the macro-economic implications of sovereign debt management. Both are discussed below.

1.2.1 Risk as an Asset-Liability Management Problem: From a Cash Flow Viewpoint

Sovereign debt management shares some features common to all asset-liability and portfolio management for financial institutions. However, one significant difference between sovereign and corporate debt is that sovereign debt management is viewed in terms of cash flows. But this does not detract from the

usefulness of the asset-liability management approach. The features of that approach applicable to sovereign debt management include the need for a clear strategic objective as well as establishment of constraints and risk tolerances. They also include a need for data and procedures to measure debt portfolio market risk and returns; to manage debt portfolio market risk and performance; to manage the legal and other risks of issuing debt; and to limit debt portfolio operational risk. Viewed this way, there is a vast literature and array of ‘best practices’ and regulatory standards available to serve as guidelines.

1.2.1.1 The Strategic Objective and the Setting of Constraints and Risk Tolerances

Any asset-liability management structure should be framed in terms of the organization’s objective and mission. This frame ultimately establishes the goal for portfolio management, and in the process helps define the strengths and weaknesses of the entity. The constraints similarly provide the boundaries for the portfolio managers, as do the risk tolerances. Portfolio management ultimately is an optimization problem, even though that explicit term is rarely used. The approach to getting clarity on the objective, constraints, and risk tolerances involves achieving consensus, or at least working agreement, among the stakeholders. This consensus among stakeholders is particularly important in managing public debt. Achieving that clarity of objective and consensus also requires expertise and technical sophistication to be able to evaluate and measure potential risks and returns, as well as the likely consequences of risk exposures entailing possibly large losses. Skills also are needed to communicate these results to non-technical stakeholders in order to establish meaningful risk tolerances. A framework of the strategic objective, constraints and risk tolerances allows the portfolio managers to have a real basis for evaluating choices among instruments or markets, and it provides their superiors and the public the basis for meaningful evaluation of performance.



1.3 Classes of Risk

The classic definition of ‘risk’ refers to exposure to an event that has some chance of occurring, causing a gain or a loss. The chance of its occurrence can be measured reasonably accurately.1 ‘Uncertainty,’ in contrast, is the term used for an occurrence that can result in unusual gain or loss, but whose probability of occurrence cannot be measured reliably.

Measurement of ‘risk’ is a key feature of risk management as applied to financial portfolios and financial intermediaries. A reliable range of the probability of gain or loss is an essential ingredient of successful risk taking and risk management. That information also permits a reliable estimate of risk bearing capacity. Such estimates serve as useful decision tools for evaluating pricing and risk-adjusted returns on capital.

Many kinds of financial risk lend themselves to measurement and therefore to this quantitative approach to risk management. Market risk is a prime example. It includes a range of interest rate risks — yield curve risk, yield levels risk, basis risk. It also includes fluctuations in the value of currencies against each other, liquidity risk and market price risk. Funding risk is another measurable financial risk. Some types of credit risk also can be managed using quantitative techniques if the number of observations and data records allow. For example, in the U.S., many types of consumer loans –for home mortgages, cars and credit cards, for example— lend themselves to quantitative approaches.

Since market risk can be measured and managed, it is suited to intelligent, focused, potentially profitable risk taking. This undertaking is supported by estimates of both an adequate capital base and an adequate risk-adjusted return on capital. This is the foundation for successful financial intermediation and portfolio management.

Other kinds of risk can involve losses severe enough even to cause failure but they do not lend themselves to a quantitative approach

to measurement or management. They are classes of risks that are best managed through control procedures, and are referred to as ‘operational risk’. Controls and reputation risks are prime examples. They might arise, for example, through conflict of interest or from ‘rogue behavior’ not properly managed or controlled. They might reflect moral hazard, where something in the reward or performance structure encourages risk taking without commensurate risk consequences. Examples often emerge in the transactions areas, where they might be as simple as mistakes that go undetected until too late or other, self-serving, mistakes that allow an improperly supervised employee to reap undeserved gains. Operational risks almost never result in gains to the organization. They are usually associated with losses, losses that range from the trivial to the fatal. As such, they do not lend themselves to the calculated risk-taking associated with most market risks, where rewards can be correlated with risk and where good portfolio management through techniques such as diversification and notional risk-capital allocation can mitigate exposure. Most institutions view operational risk as a risk to be avoided or at least minimized. They tend to do this through organizational structures designed to make the control mechanisms as effective as possible.

Finally, there is a class of risks termed “country risk”—risk that is systemic and beyond the immediate purview of the manager of the portfolio or financial intermediary. These represent changes in the external environment that can affect the success of the portfolio. Unexpectedly high or volatile inflation is one example. Another example might be a deterioration in a country’s own sovereign debt rating, since that serves to underpin the debt rating of any entity from that country that also is issuing debt in external markets. Regulatory risk can be systemic, since decisions taken by regulators can affect the viability of whole classes of portfolios or intermediaries. The widespread failures of savings and loan associations in the United States in the 1980s were a very expensive example of regulatory risk.

1 Frank Knight, Risk, Uncertainty and Profit, 1921 republished in Midway Reprint, Chicago: The University of Chicago Press, 1985.



In summary, market risk is best managed quantitatively through reliance on data and on the tools available to analyze data. Operational risks are best managed by avoidance where possible and otherwise by limiting techniques. Appropriate organization structure tends to be one of the most effective tools for avoiding or limiting operational risks. Even though country risks such as systemic and regulatory risk might be seen as outside the control of individual portfolio managers, individual managers can nonetheless make a contribution to the management of these ‘macro’ risks. Their contribution would be made by evaluating the consequences of potential decisions through the same good ‘what if’ simulations as are used to manage market risks and then communicating their findings to policymakers.

It is helpful to disaggregate the fundamental risks into specific risks that permit measurement and effective mitigation actions. For sovereign debt management, the risks discussed above take the following forms:

• Funding risk - the risk that the ability to fund new cash needs or to rollover existing debt will be sharply curtailed as to quantity, or quantity at terms that can be accepted.

• Market risk - risk that arises from normal operations in the interest rate and foreign exchange markets.

• Credit risk - risk that partners in business transactions will not pay as promptly or as fully as anticipated when the transaction was agreed.

• Operational or controls risk - risk that arises from administrative processes in debt issuance, trading, accounting and redemption operations.

• Country risk - risk arising from the overall financial economic and political condition of the country or from a major failure of controls that damages the country’s reputation.

These broad categories of risks are elaborated below.

Funding risk

Volume. The risk that the Ministry of Finance will be unable to provide sufficient funds to meet the government’s expenditure and debt reduction targets, on time.

Refunding. The risk that the Ministry of Finance will be unable to rollover maturing obligations.

Market risk

Currency. The risk that an adverse change in the exchange rate would increase the current debt servicing requirements of a loan denominated in foreign currency relative to the funds available to service that debt.

Liquidity. The risk that securities cannot be readily converted to cash for their full face value, despite having short maturities and a domestic currency denomination.

Marketability. The risk that securities cannot be readily sold at a reasonable price, even if not their full face value, either because of discontinuities in the marketplace or because the market for those particular instruments is too thin.

Interest Rate. The risk that, relative to the funds available to service debt, debt service costs will increase directly or the opportunity to reduce debt service costs will be foregone as a result of adverse price movements in financial markets.

Instrument. The risk that an option embedded in a complex security will produce costs in excess of those anticipated at the time the option was created.

Credit risk

Counter-party. The risk of nonperformance by the obligor to a forward contract or derivative.

Other Partner. The risk that a partner in a contract other than that described above,



such as a borrower, will not pay the full amount agreed in the contract in a timely way.

Contingent. The risk that a contingent obligation would be called, resulting in the sovereign taking over the debt servicing obligations of a guaranteed party. The common cause for a contingent obligation to be called is the guaranteed entity becoming unable to service the debt.

Operational (controls) risk

Technical. The risk of inadequate internal control in debt management operations, which would allow embezzlement, fraudulent transfer, or other loss to occur undetected or with ease.

Settlement. The failure of a business partner to settle on agreed terms, for reasons other than the partner’s credit deterioration.

Country risk

Reputational. The risk that the credit standing of the sovereign will deteriorate, resulting in an increase in the interest cost it would have to pay relative to that paid by other borrowers. This event could be associated with a lowered credit rating. At the extreme, this refers to the risk that in the wake of a default, an issuer will find market access severely curtailed or quite expensive, or both.

Systemic. The risk arising from a widespread failure, such as failure of the banking system that leads to widespread collateral failures. It might be caused by a failure of the regulatory system or inadequate stop-loss safety mechanisms, for example.

1.4 Risk As a Macroeconomic Problem

There are three principal additional considerations for sovereign debt management over and above standard asset-liability

management as practiced in a financial institution. The first the nature of the sources of funds available to service sovereign debt. The second is the role of sovereign debt in country crisis management. And an important third is the role of sovereign debt issued in domestic markets as a tool for open market monetary policy operations and as a basis for developing capital markets.

Sources of Funds Available to Service Sovereign Debt

Debt risk management and optimization are viewed in the context of the flow of funds available to service the debt. Preserving or improving the sovereign’s highest feasible credit standing— important to achieving the main goal for sovereign debt management— can be considered in cash flow terms. Credit standing is strongest when there is a record of making debt service payments in full and on time and giving market participants confidence in the country’s continued capacity to do so. In that regard, the currency composition of funds flows available to service debt, and the potential volatility associated with domestic and external market conditions, are important pieces of information, as are the size and growth trend of these funds. Measuring sources of funds available to service debt is typically a coordinated effort. The sovereign’s budget and economics statistical functions would be expected to make forecasts of funding sources and provide them to the sovereign debt management team. The central bank is often the repository of information on quantities of foreign currency available to make payments on external debt, and would coordinate this information with debt management. Consequently, while important, the sources of funds available to service debt are not discussed further in this document as the focus is on managing the requirements for servicing debt.



The Role of Sovereign Debt in Country Crisis Management

Sovereign debt, including guarantees issued by the sovereign, can play many roles in country crisis management. Inappropriate debt structures, such as those not well matched to the cyclical patterns of funds available to service debt, can in fact cause or exacerbate crises. Exposure to contingent obligations such as guarantees can create added stresses on the sovereign’s cash needs during a crisis unless these potential calls have been properly anticipated. Tests of the adequacy of the margin by which funds sources of the guaranteed entity exceed debt servicing requirements provide an indication of the probable calls on the sovereign. The capacity to issue new debt as a source of liquidity to weather a crisis is an important defense: one goal of appropriate risk measurement and management is to preserve that capacity. These considerations need to be incorporated into the overall goal and constraint setting of sovereign debt management.

Sovereign Debt as a Policy Instrument

The sovereign issues debt primarily for the purpose of financing a current or past deficit. The goal of debt management is to fund the fiscal requirement. But debt management also operates under a number of constraints that arise from many aspects of sovereign debt that influence an economy. Agreement on the nature and scope of these constraints is a significant part of the debt planning process. Since government securities are also an important vehicle for open market operations conducted by central banks to implement monetary policy, the central bank would contribute to this process. Other interests and agencies of government also are stakeholders in the consensus building. For example, a desire to develop a long-term home mortgage market or longer-term corporate finance markets also might be represented in establishing the constraints for sovereign debt management. The taxing and legislative authorities could have important contributions to make in constraint

setting and performance monitoring. The stage of development of the capital market may be a constraint. In the process of agreeing on the goal and setting constraints, countries should also bear in mind that sovereign debt issuance and management can play a role in fostering development of viable, transparent domestic money and capital markets. Sovereign debt provides a benchmark yield curve and instruments for trading. It should also set a standard for market operations.

During the early stages of a sovereign debt market, market development is a principal focus of the debt manager. A key first step is to reduce impediments to market development, such as exposure to crises, liquidity strains, and excessive volatility. One simple example is managing the refunding portfolio to avoid exceptionally large amounts of principal falling due on a single day. Market development consists of broadening the universe of investors,, lengthening the yield curve, removing barriers to development of the secondary market, and deepening the primary market so as to permit full financing of the deficit in the domestic market in the year in which the deficit occurs. Once these conditions are met, the primary focus of the debt manager can shift from market development to risk management in the debt portfolio. The goal at this point is further reduction of risks to the state budget, beyond those described as impediments to market development.

1.5 Risk Management of Contingent Obligations

Contingent obligations are a good example of the intersection of financial instruments, financial risk management, and the special macroeconomic considerations unique to the sovereign. In evaluating the benefits and costs of contingent obligations, the sovereign weighs them not just in a static analysis but also in terms of their dynamic or cyclical properties. The sovereign assesses its contingent obligations in the context of its responsibility to manage the economy efficiently, to avoid severe cycles and crises, and to have in place effective tools to manage such events if they do occur.



1.5.1 Sovereign Guarantees

A guarantee is a contract involving three parties. The lender is one party to the guarantee. The direct borrower, the guaranteed party, is the second. And the guarantor is the third. In this discussion, the guarantor is the sovereign. The direct borrower is obliged to repay the lender, as with any loan. However, the guarantee provides that if certain specified conditions arise, the guarantor is obliged to repay the lender in place of the direct borrower. The specified condition represents the contingency of the obligation from the guarantor’s viewpoint. Once the contingency has occurred, the guarantor’s obligation to repay is conceptually like any other form of debt issued directly by the guarantor.

A common type of contingent obligation is a credit substitution guarantee under which a sovereign agrees to pay one or more of the debt service payments owed by a third party to a lender if that third party does not make the payment on time and in full. This is a simple substitution of the sovereign’s credit standing for that of the direct borrower, the guaranteed party. The lender finds this guarantee valuable because it is less likely that the sovereign will miss a debt payment than that the third party would. The usual reason for the missed payment is a shortage of funds. An example is a sovereign with AAA credit rating guaranteeing a third party with a CCC rating. The contingency to which the sovereign is exposed in this guarantee is that the direct borrower does not pay, on time and in full, the debt-servicing obligation covered by the guarantee. The size of the payment covered by the guarantee at the time the contingency materializes is the amount to which the sovereign is exposed. That amount is specified in the guarantee contract, and it might be limited to one debt servicing payment or it might extend to as much as the full outstanding principal balance.

Another class of contingent obligations covers specified conditions, such as certain acts beyond any of the three parties’ control but which in the end would likely adversely affect the ability of the guaranteed entities to cover their debt servicing obligations. Examples are

shortages of commodities, whether manifested as literal shortages or extreme price increases, foreign exchange rate instabilities, and civil strife. The distinction between these two types of guarantees is that in the former, credit quality deterioration manifested by missed debt service payments would lead to a call on the guarantee without regard to cause.

In the domestic market, a credit substitution guarantee might occur when small businesses, agricultural enterprises or homebuyers borrow from private financial institutions. These are examples of common three-party guarantees, in which the borrower is the guaranteed party, the sovereign is the guarantor, and a private financial institution is the lender. A private institution is specified because it faces the consequence of bankruptcy if it makes too many loans to borrowers who do not repay principal and interest. In a particular country, weaknesses in sectors such as farming and small business might imply frequent interruptions in timely and full debt servicing from such borrowers, which would prompt private lenders to restrict amounts loaned, limit repayment periods and charge high interest rates reflecting the high credit risk premium. Those restrictions might be eased if the sovereign were to backstop weaker credit sectors with loan guarantees on the assumption that the guarantees would generate net gains for the public at large. This would require that the gains to society resulting from the guarantees more than cover the sovereign’s costs from the guarantees.

Finally, a fourth entity is sometimes involved in sovereign guarantees when the sovereign’s own credit standing requires bolstering. This is often the case with cross-border transactions. An example might be a AAA international corporation seeking financing for a project to be located in a country in which the sovereign’s own credit rating is significantly below AAA. An investor’s or lender’s project financing terms typically reflect the probability of an interruption in debt servicing associated with the weakest credit, which in this example is the sovereign. Under such conditions, the sovereign’s own guarantee would be backstopped by a guarantee from an international financial institution.



The sovereign might well conclude that the public benefit from the additional guarantee would exceed that from any alternative use of financing from the international financial institution.

This illustrates another very important aspect of the risk in contingent obligations — the currency in which the called guarantee must be paid. In three-party sovereign guarantees linked to the debt of domestic farmers or small businessmen and owed to domestic private financial institutions, the currency of the contingent obligation typically is the domestic currency. The sovereign has control over the supply of that currency, and is therefore able to pay those claims whether by increasing the money supply, displacing other government expenditures, borrowing more or taxing more. The macroeconomic effects of those choices are not the same, of course, and the preferred choice may not be the simplest politically. But nonetheless, in three-party sovereign guarantees, the sovereign has access to the domestic currency to pay domestic contingent obligations.

1.5.2 The Sovereign’s Contingent Obligations Costs and Risk Exposure

Event risk is the term used to refer to the risk that the contingency will materialize. Contingent obligations are best managed when the probability associated with event risk can be forecast with reasonable assurance. Even so, there inevitably is some margin of error in calculating the probability of occurrence of the kinds of events likely to be covered by sovereign guarantees. Moreover, events triggering calls on contingent obligations are linked to certain macroeconomic conditions and neither they, nor their linkages with calls on guarantees, can be forecast with complete accuracy.

Some of the benefits of a guarantee arise precisely because of its contingent nature. Because the guarantor has made a contractual promise to pay, the guarantor rightly measures his exposure to the risk of the guarantee being called as equal to the amount promised. However, for something less than a one-to-

one chance of actually having a cash funding obligation— that is, for only the probable amount called— the sovereign can provide its support to worthy projects or entities thereby improving their access to funds or to favorable funding terms. These benefits from the guarantee might translate to improved success rates for the projects guaranteed, or to lowered borrowing costs.

Probable amount called refers to the expected cash obligation of the sovereign if event risk materialized — that is, if the guarantee were called. The amount that would be owed typically is specified contractually under the guarantee, and is date-specific. The term ‘probable’ in this case therefore reflects the product of the probability of the event’s occurrence (event risk) and the specified contractual cash obligation owed at that time by the sovereign should the event occur (exposure).

Probable loss refers to the probable amount called on the guarantee, net of the probable amount of funds, if any, subsequently recovered by the sovereign from the guaranteed party. The loss calculation might also include expenses and some cost for the time value of money depending upon the nature of the recovery.

Exposure refers to the total amount the sovereign has promised to guarantee, and therefore conceivably could be obligated to fund. It is the maximum possible (not probable) amount that can be called at any point currently or in the future. Exposure is an important risk measurement concept in the management of contingent obligations. It is the measure by which the guarantor tracks its aggregate outstanding promises. As the maximum guaranteed amount, it represents the outer extreme of the total amount that can possibly be called, and possibly lost. Its probability is not zero. If there were absolutely no chance a guarantee ever would be called, the parties in the transaction would find the guarantee fee to be not worth paying, and therefore the transaction would not have been executed.

Guarantee fee is the compensation paid to the guarantor for accepting the risks and costs of a contingent obligation. Sometimes there are enough independent guaranteed parties



and enough historical data to compute a fee using an actuarial quantitative approach. Such an approach would use a large number of observations to measure probable calls and losses, their time horizon and potential impact on funding requirements. It would also help determine not only the adequacy of the guarantee fee structure but also the associated reserves. A long-standing program of sovereign guarantees on home mortgages would lend itself to that actuarial type of analysis, for example. The sovereign also might add a supplement to the guarantee fee in order to limit any distortion of competition in a market such as might arise from extending sovereign guarantees to some but not all participants.

1.5.3 Sovereign Contingent Obligations: The Dynamics of Risk Exposure

The sovereign would want to evaluate the benefits and costs of contingent obligations in terms of their dynamic or cyclical properties, not just in a static analysis. The sovereign assesses its contingent obligations in the context of its responsibility to have in place effective tools to manage such events if they did occur. Moreover, probable calls on contingent obligations should be included in projecting a sovereign’s borrowing program. Each macroeconomic scenario used to project future borrowing needs would likely entail a different volume of calls on contingent obligations.

Any contingent obligation is expected to convey a benefit to the guaranteed party in terms of a reduced cost of financing, better financing terms, and/or improved access to funds. That by itself is only a small part of the benefit test for the sovereign when evaluating these instruments, although it is a prerequisite. Benefits from some contingent obligations are difficult to quantify. The partial guarantee, for example, leaves some risk exposure for private lenders. This improves the credit terms that they can provide while leaving them an incentive to make sound credit quality lending allocation choices. Another such benefit is the civil stability that can accrue from extensive home ownership as a result of sovereign home mortgage guarantees. Other benefits more

readily lend themselves to measurement, however. For example, extending the sovereign guarantee to the debt of small businesses might improve their viability by lowering their debt servicing obligations. This might generate enough associated increases in employment and income tax revenue to more than compensate for the sovereign’s costs and risk exposure from extending the guarantees.

The more important benefit test for the sovereign is the likely behavior of contingent obligations in the context of the unique role the sovereign’s policymakers play on behalf of the public in managing the macroeconomic conditions in a country. Under adverse cyclical conditions, for example, would calls on contingent obligations exacerbate the cycle or exert a counter-cyclical influence? In a crisis, would calls on contingent obligations confound crisis management or would they act as deterrents to further downward spirals? Under such extreme circumstances, which sectors in the economy would be protected by the presence of sovereign contingent obligations and which sectors or activities might be damaged? Finally, what is the counter case: using the same analysis of the extreme cyclical circumstances or crisis conditions, what would the results be for the payments system and overall macroeconomic management if there were no contingent obligations?

1.5.4 Managing Sovereign Contingent Obligations

Questions about the costs and benefits of exposure to contingent obligations can be answered through the insight gained from a ‘stress testing’ analysis of a portfolio of contingent obligations. This approach helps in evaluating the quantifiable benefits, risks, cost and ultimate value to an economy of a sovereign guarantee program. Stress testing is most fully accomplished through a simulation model, but if that is not feasible then even working through the logic of a stress test is helpful. A stress test analysis allows examination of two very important attributes of potential calls from contingent obligations: their covariance with other cyclical patterns in the economy, and the



degree of concentration of these potential claims that is optimal for macroeconomic management. It illustrates these attributes because the stress test takes an extreme case, even if it seems unlikely, and works through its dynamics and feedback effects.

When the sovereign contingent obligation requires a payment in a foreign currency, the risk analysis of the extreme or crisis condition is a very crucial part of the sovereign’s evaluation because the foreign currency to honor the called guarantee must be acquired at a market price. In a crisis, that price can be very high in terms of the domestic currency, and this has been shown even in cases where the domestic currency had been pegged under pre-crisis conditions to a strong foreign currency. Moreover, even when a contingent obligation has the backstop of a fourth party, such as an international financial institution that would lend the foreign currency to the sovereign in order to pay the claim, there is a need for very careful extreme-case risk analysis. The analysis would consider the conditions that would contractually permit exercise of the contingent obligation alongside the plausible extreme — not just the likely — macroeconomic setting under which those conditions might co-exist.

Evaluating and managing sovereign contingent obligations requires careful dynamic risk analysis. Not every contingent obligation structure necessarily produces benefits that outweigh risks, although some clearly do. The successful programs, even when costly at times, have featured good macroeconomic risk containment.

Successful management of sovereign contingent obligations rests on a good quantitative base built upon a solid information structure that includes both called and contingent obligations. That solid information structure underpins risk analysis, reserves adequacy analysis and setting the level of the guarantee fee. It helps measure the linkage between the business cycle and the cycle of calls.

The Analytical Database recommended for risk management and described later contains modules for contingent obligations that are designed to facilitate this analysis and record keeping. The database would contain information on each uncalled contingent obligation as follows: the amounts associated with the underlying debt guaranteed, the guarantee obligation itself, the associated dates and identities in the contract, and references to the events covered by the contingent contract. A time series section of the database would store all this information for potential analytical and ad hoc uses. The database also would contain information on each called (exercised) contingent obligation as follows. On each measurement date, the amount owed directly by the sovereign through calls would be detailed in the appropriate section of the database depending upon how the sovereign honored the call. For example, if the sovereign borrowed directly from a multilateral institution to honor the foreign currency obligation of a four-party guarantee, the relevant details would be stored as a borrowing from a multilateral. The cash flows associated with honoring each called contingent obligation, as well as the relevant details of the guaranteed entity and all dates and amounts due, would be stored in the time series section of the database.

Chapter 2 describes a simulation model that further enhances risk analysis. These database and risk measurement structures would lend themselves to answering questions like:

Under alternative macroeconomic, fiscal and market conditions, how much additional funding would be required from probable calls on guarantees? What would be the impact on the funding plan?

How many calls on guarantees in a particular economic sector (e.g., small business) have occurred in the past? How does the cyclical pattern of those calls relate to the cyclical pattern of sovereign funding needs in general? How much has been recovered in the end from amounts disbursed on called guarantees?



1.6 Risk Tolerance and Performance Measurement

Sovereign debt management involves using the best generic portfolio and asset-liability management standards. But in addition, there are considerations of public trust and macroeconomic management as well as potentially complicated stakeholder involvement associated with managing on behalf of the sovereign. These all come together in the topics of risk tolerance and performance measurement.

For example, a private portfolio would be managed to maximize return for any given risk exposure, and will be structured to take on certain risk exposures in order to earn the associated returns. This is calculated risk taking. But involved with that is some probability of loss, which of course might materialize. Also involved with risk exposures are the ‘uncertain’ events that can trigger losses - these might include extreme political stresses, financial crises, international emergencies or strife.2 When a private portfolio incurs extreme loss, failure results and for the most part the consequences of the failure are contained to the stakeholders of that particular portfolio. There is no such containment, of course, when the portfolio is a public portfolio. Serious losses can have major repercussions throughout the economy. Moreover, households have no means to hedge risks taken by the sovereign, but it is they who ultimately bear the consequences of the sovereign’s risk exposure. With these considerations, the public debt managers would set risk tolerance goals that would incorporate stricter loss avoidance and risk aversion than a typical private portfolio.

The risk tolerances for public debt management are complex. They are likely to reflect a goal and a hierarchy of constraints developed through a consensus-building process among the major stakeholders in the economy who stand to gain from good public debt management. These include the fiscal authorities; elected

politicians; the central bank and the parties involved in the conduct of monetary policy; those elements of the economy who would be most jeopardized by a financial crisis; and the segments of the public who would gain most from stable, efficient domestic capital markets. These of course, and others, also stand to lose from public debt management that entails too much exposure to crises or to volatility in debt servicing requirements

For example, a set of risk tolerances for a sovereign might be structured as follows. First, minimize the risk of a funding shortfall: that is, borrow the required volume of funds. Second, minimize, to the extent feasible, the risk exposure to undue volatility in debt service requirements relative to funds available to service debt: the relevant exposures would include liquidity risk, market interest rate risk, currency risk, counterparty and other business partner risks. Third, undertake borrowing operations consistent with those risk tolerances that also limit rollover risk. Fourth, consistent with the above, borrow funds at the lowest feasible cost.

These, all reasonable statements and a reasonable set of conditions for a sovereign, would ultimately be focused on the cash flow requirements. The particular focus would be the near term, but their dynamic path over the medium term, and their elasticity relative to that of funds available to service debt, are very important points of focus. This is one reason why a ‘benchmark’ that is constructed to focus on an outstanding portfolio of debt can be a very difficult tool for managing toward the sovereign’s goal.

There are several other reasons why a ‘benchmark’ constructed in terms of an outstanding debt portfolio can cause difficulties for sovereign debt managers.3 The sovereign’s debt portfolio is unlikely to be actively managed, so ‘rebalancing’ an existing portfolio of funded debt to accord with a benchmark focused on portfolio characteristics is likely to be

2 The word “crisis” implies a special or infrequent condition. Financial crises associated with sovereign debt in the 1980s and 1990s were by no means infrequent, however.

3 Paul Sullivan when he was director of Strategic and Risk Management for the National Treasury Management Agency of the Republic of Ireland, said, “all benchmarks should carry a government health warning.”



costly at the least. The likelihood of a need to rebalance may be low if a sovereign has a well-established debt management operation and stable operating constraints. The chance of needing to rebalance may be quite high, on the other hand, for a sovereign whose starting debt position is quite different from the desired, and it is very likely to be high for a sovereign in a transition status wanting also to accomplish efficient market development. Moreover, rebalancing a domestic debt portfolio could interfere with market development. Even with established, marketable, traded portfolios, wrongly-constructed ‘benchmarks’ can result in perverse incentives and rewarding performance that does not necessarily help to achieve the goal; hence the title of a presentation at a conference on benchmarks, “Lies, Damn Lies and Benchmarks”.

Each country should establish its own risk tolerances, goal statement and hierarchy of constraints because the achievement of macroeconomic objectives produces long-range benefits to the economy not accounted for in the standard short-term cost/risk formulations that benchmark commercial portfolio performance. For example, domestic sovereign debt is a macroeconomic tool that can affect the functioning of domestic money and capital markets. Also, a sovereign’s foreign currency debt can have implications for the exposure of the economy to international market conditions. Macroeconomic objectives such as these become constraints in the framework for the sovereign debt manager.

Especially where there is a transition implied in market development, performance measurement is more likely to be consistent with the complex interaction of constraints if a single ‘benchmark portfolio’ concept is avoided. Each sovereign would have its unique hierarchy of constraints depending upon its own starting position and time path for market development. Well-constructed performance measures can provide positive incentives and a helpful framework for making choices that are most consistent, all things considered, with the sovereign’s objectives. Given this complexity, and the possibility that at times a

set of constraints might contain some internal conflicts, performance measures that are structured more like optimizing statements than typical commercial portfolio benchmarks are likely to be most useful. An example of the overall statement is provided in the top section of Table 1. This statement would likely be unchanged for many years and is amenable to being made public.

The sovereign’s funding operation risk tolerances and management guidelines are likely to be quite specific, on the other hand, and would evolve as needed or desired. The specific tolerance ranges need not be publicly stated, nor would they be expected to be as stable as the overall policy statement. Perhaps risk appetites or risk-bearing capacity might change, for example, and that change would be reflected in the specific tolerances. The specific macroeconomic considerations also would be detailed, focused on achieving a near term goal or satisfying near term constraints. Examples of specific risk tolerance and macroeconomic policy guideline topics are in Table 1. The statements would be specific enough to provide an operating framework for issuing debt and evaluating the cost/risk tradeoff of alternative funding choices. The statements likely would be formulated by the risk management and policy functions and endorsed by senior management. These statements are the basis to measure performance.

A tool for establishing these specific tolerances is the dynamic simulation model, as described in the following section and more fully in Chapter 2. Extreme situations can be tested and measured so that the potential risk exposure linked to a cost reduction can be evaluated. An example is the additional interest risk exposure and funding risk exposure associated with reducing cost, in a typical yield curve environment, by issuing all short-term debt.

The work underpinning the statement of specific tolerances also can be helped by a thought exercise. “What is the case with no exposure to funding and market risks?” Each sovereign will have its own answers to establish its own ‘no



risk’ hypothetical position. That starting point helps in evaluating how much true value comes from the acceptance of additional risk in return for reduced cost, viewed in light of the country’s own capacity to cope with risk exposure. The context is the ability to honor each debt service obligation, in full and on time.

Funds Available to Service Debt. How sensitive is the net fiscal deficit, excluding interest payments, to economic conditions? How tightly are interest payments on outstanding sovereign debt linked to current market conditions?

Funding Risk Exposure. What is the size of the current and projected fiscal deficit, and the need for net new funding? How evenly are principal payments due on outstanding debt spread throughout the next ten years and what is their size relative to the typical volume of debt issued in that country’s capital markets?

Market Risk Exposure. What is the currency composition of the country’s source of funds available to service debt? How much does that vary under different market conditions? What is the currency composition of the country’s debt servicing obligations? What is the net difference between the sources and uses of funds in each currency, and how does each net amount respond to changing economic and market conditions?4 Similar questions are asked about the net responsiveness of the funds available to service debt and debt servicing with respect to market interest rate changes

Credit Risk. Are there counterparties in swaps or related transactions? Are there contingent obligations outstanding but not yet exercised? Are there on-lending contracts?

4 This measure of risk, the currency-by-currency net difference between sources available to service debt and debt service payments owed, is illustrated by the definition of a sovereign debt crisis: when the sovereign has insufficient funds to supply or to purchase the currency needed for a debt service payment.



Table 1

Sovereign Goal and Policy Guidelines

Example of Public Statement of Goal for Sovereign Debt Management:

To provide the required funding volume in a fiscal year, in an orderly way that does not compromise the ability to do that again in the coming year, and at lowest cost consistent with guidelines and policies.

(1) Funding operations would be conducted within risk tolerances as defined by policy guidelines. (Guidelines would be specified.)

(2) Funding operations would be subject to macro-economic considerations established by policies with regard to cyclical stability and domestic market development. (Policies would be specified.)

Example of Internal Guidelines for Risk Tolerances:

Subject to remaining within [specific tolerance range] of policy/risk management limits for:

(a) the currency composition of future debt servicing requirements over the medium term horizon

(b) the interest rate resetting pattern over the coming [five] years

(c) the overall annual debt amortization profile over the coming [five] years; and

Other [specified in detail] considerations, as established by risk tolerances and other macroeconomic objectives, such as:

(a) using instrument designs and marketing procedures that enhance liquidity and marketability

(b) limiting credit risk exposure [with respect both to volume and minimum credit rating], especially with regard to external entities

(c) using settlement and internal operating procedures that limit exposure to operational risks.

Example of Policies for Macroeconomic Considerations:

Subject to [specified approaches] to funding so that it is consistent with offering debt regularly and predictably in order to:

(a) enhance primary and secondary markets

(b) maintain market liquidity

(c) maintain the risk free status of sovereign obligations



1.7 Procedures for Managing Market Risks and Returns

Using the tools and conceptual framework described above, asset-liability managers are able to establish processes and incentives to help achieve their portfolio goal within their constraints. They are able to measure the performance of portfolio managers in that framework, having been able to inform the managers ahead of time what that framework is — i.e., with clear definition of the goal and limits. They can measure performance returns adjusted for risk, and ultimately can judge how close their managers are to the ‘efficient frontier’ in their market risk class.

Thinking about sovereign debt management as having some features that are macroeconomic in nature but others that are shared by managers of any financial portfolio helps to identify sources of value added from providing technical assistance to sovereign debt managers. The aspect of sovereign debt management that carries all the same hallmarks as any generic portfolio management can benefit from ‘best practices’ and tools that have been most reliable during the recent general upgrading in quantitative financial risk management.

There is a strong case for relying on those existing systems to the extent feasible. They have a proven track record. Moreover, introducing sovereign debt managers to best practices, and helping them to use these tools in an appropriate way, is in the end a very efficient form of technical assistance. The tools and approaches may involve a considerable

initial investment, both in money and technical effort, however. Accordingly, some of the technical assistance to sovereign debt managers would likely take the form of helping them plan the steps in the transition to a fully quantitative risk management and measurement system.

The market risk measurement system has several components. Two components are the structure of the data architecture and the use of hardware and software appropriate for the risk management and accounting/transactions needs. The tools for making projections and risk evaluations are a third component, and the optimizing tool for measuring tradeoffs is another. These are discussed below.

Projections for planning purposes, for evaluating the risk/cost attributes of alternative strategies, and for stress testing all can be accomplished through a simulation model. This procedure is becoming standard for financial risk management.5 The simulation model is fed from the Analytical Database and uses exogenous variables to depict market conditions such as for interest and currency exchange rates, and to depict other exposures such as funding, liquidity and credit risks. The simulation model is the tool for measuring the value at risk, the extremes of risk exposures, and the dynamic testing of risk-bearing capacity. Chapter 2, Dynamic Simulation Model for Sovereign Debt Risk Measurement, describes this procedure. A simulation model can be designed as a generic tool for sovereign debt risk measurement. However, any such model and its use can be complicated, and the best technical support would come from helping

5 “Optimal Bank Supervision in a Changing World” by Alan Greenspan, “The Declining? Role of Banking” May 1994, Thirtieth Annual Conference on Bank Structure and Competition, Chicago Federal Reserve Bank.

“A Survey of Stress Tests and Current Practice at Major Financial Institutions: report by a Task Force established by the Committee on the Global Financial System of the Central Banks of the Group of Ten Countries,” April 2001. Bank for International Settlements. available at www.bis.org. Forum for Central Banks, Committee on the Global Financial System.)

“Stress Testing by Large Financial Institutions: Current Practice and Aggregation Issues”, April 2000, Committee on the Global Financial System, Bank for International Settlements, Basel, Switzerland (available on www.bis.org. Forum for Central Banks, Committee on the Global Financial System.)

“Standard of Practice on Dynamic Solvency Testing for Life Insurance Companies,” Canadian Institute of Actuaries, June 1991. available at www.actuaires.ca.

Bank for International Settlements release, May 21,1996, Basel Committee/IOSCO “Ensuring that Banks Have Adequate Capital.” www.bis.org.

“A Universal Approach to Credit Analysis.” Moody’s Investors Services at www.moodys.com.



the country team to ultimately ‘own’ its model. This would be accomplished by providing a generic set of equations, a fairly standard set of exogenous variables, and perhaps even a standard set of output tables and charts. But these are best provided in a way that helps assure that the country team does its own programming and tailoring of the generic to suit its own needs and learn from the ground up how the model works.

A fully quantified risk measurement system will have a structure for evaluating choices and for trading off among alternative cost/risk portfolio positions. This structure for evaluating choices could be achieved by an explicit model that measures these tradeoffs and that solves for an ‘optimal’ portfolio taking into account the goal, constraints and available choices. Such a model is complicated to build and maintain, however. It moreover requires a fairly high technical proficiency among decision-makers. And most importantly, an explicit optimizing model is of no practical use if the sovereign’s own debt data, and the sovereign’s own strategic objective, are not available to be incorporated. This point highlights the notion that there is a critical path in developing good measures for portfolio management.

Even if there is no explicit model to ‘solve’ for the optimal portfolio, it is possible to structure the approach to making choices within this risk management framework. Articulation of the goal and constraints, and the desired time path to achieve a certain goal, as described in Section 2 of this chapter, are the first step in any case. The sovereign’s own debt data must be available in a sufficiently detailed form for risks and opportunities to be measurable. The current choices available from the market —in other words, the interest costs for each maturity class, for each type of instrument, in each currency — also are needed in any case. Successive simulations, while not

directly solving for the ‘optimal’ portfolio, can show the dynamic time profiles of costs and risks under many alternative future market conditions. These results can then be used to make choices within the tolerance ranges of the existing framework; they also can be used to inform further development of policies and risk tolerances.

1.8 A Structured Approach to Organization and Information Architecture

There is a methodology with a well-proven track record used to analyze the data and information technology needs of an organization that also provides an analysis of the organization structure for conducting business. “Experience has shown that this business process can be applied to all institutions in the public sector and all industries in the private sector because the requirements for developing information systems are similar regardless of the business served or the products and services provided.”6 This information architecture methodology starts from the premise that information technology objectives should support an entity’s business objectives. For any given business objective, the business process and data requirements are relatively stable.

The structured approach discussed here identifies objectives, needs and priorities largely through interviews with managers at all levels. Results of these interviews are used to establish a framework of the processes and data requirements, and ultimately the information resources, needed to serve the business objective. Many organizations have found that this structured approach to information strategy planning also has yielded valuable insights into business organizational improvements. This process yields results needed for both quantitative market risk management and for operational risk management.

6 From “Information Systems Planning Guide,” which offers assistance to teams conducting a Business Systems Planning study, copyright International Business Machines Corporation (IBM), Third Edition, July 1981. This approach was used initially in the 1960s for internal corporate business planning at IBM. It has been refined and used successfully among a wide group of entities. By now, many professional practitioners in strategic information planning and information engineering use procedures emanating from this original approach and an extensive literature is available.



The interviews determine each manager’s opinions on the following:

1. The key responsibility of each manager; the most important objective in the key areas he/she manages relating to output of the organizational unit and management processes.

2. How personal success is measured and defined.

3. Critical success factors and the information needed to enable success.

4. Major obstacles or problems inhibiting achievement of the objective, and his or her suggested remedies.

5. Specific items of information needed to support each business process in which the manager is involved; where the information originates; adequacy of information; and an opinion of the benefits of gaining access to available information not already received.

6. Likely future changes in the objective, business process, organizational structure and outputs in the manager’s area of responsibility.

The compilation of the results of these open-ended interviews produces output including:

1. Mission statement.

2. Detailed list of functions performed and the associated processes. For example, “borrow money” is a function and the associated processes could be “develop borrowing needs”, “issue debt,” “manage outstanding debt portfolio,” “pay debt service,” etc.

3. List of business process definitions. For example, “report financial condition” could be defined as “prepare mandated and routine information and reports on financial condition and performance for external and internal uses such as audited financial statements and prospectuses.”

4. Data entity type definitions. For example, “borrowing instrument” could be “business agreement by which funds are raised and which results in a liability”.

5. Business area identification and definitions. For example, “risk management” is a business area that could be defined as “the processes and data used to manage financial risk exposures.”

6. Data collection definitions or sub-databases. For example, “common data” might be defined as “a data set commonly used by many functional areas, processes and application systems, such as exchange rates, information about counterparts, and macroeconomic data.”

7. Inventory of application systems. For example, those that compute daily market values for debt obligations, using daily market prices.

8. Matrix of business processes and the application system each uses.

9. Matrix showing each business process and the organizational unit involved and its degree of responsibility.

10. Matrix showing each business process, the data associated with the business process and whether the data involvement is to create the data or only to use it.

The matrix referred to in (10) above can be thought of as a ‘map’ showing where the data are generated and where the data are used. This map is essential in building an integrated data/information system that will support quantitative risk management as well as the requirements for the General Ledger and transactions processing.

These structured interviews bring to light areas of responsibility for a subject or process and the linkages of that organizational unit with other parts of the organization that have subsidiary responsibilities or dependencies. This output, the matrix in (9) above, also can be thought of



as a ‘map’. In this case, the map yields valuable insights into the nature of operational risks and any gaps or redundancies in the organizational structure that would need to be corrected to minimize operational risk. This lends itself to an action plan with appropriate assignments. In particular, the process of producing these ‘maps’ through interviews with all managers helps to generate their feeling of ownership of the action plan and their support for its successful conclusion.

This methodology can be especially valuable in a setting in which such a structured approach has not yet been applied. Many countries now working on improved methods for measuring and managing their sovereign debt are examples. Typically, they are starting from a position of fragmented institutional responsibility for issuing sovereign obligations; in many cases multilateral debt, guarantees, foreign currency private market debt and domestic currency obligations are all handled by different units of government. Those divergent responsibilities often are associated with incompatibilities in data structures and data handling standards across organizational units, and with difficulties in data sharing at many levels.

The potential to use this methodology for providing technical assistance to countries in their sovereign debt management can be summarized as follows. Each country will have its own unique starting position in terms of data quality and data processing tools, its own unique data map and its own unique organizational arrangements. The same tool— this structured, interview-based approach— would be applied to each country, but the specific information derived for each country will be a unique guide to the requirements of the debt management organization in that country. The data map will show where data items originate, the unit responsible for them, and other units that rely on that data for their own work. Additional information on this structured approach to information and organizational architecture is provided in Part 2.

1.9 Organization Structure

The debt management system of any country is complex. Many functions are interdependent and performed by a number of government agencies. Effective debt management requires a clear allocation of responsibility and tasks among agencies and smoothly functioning relationships between the units involved.

Fortunately, however, many of the considerations that are most significant in designing an appropriate organization structure are common to any financial management function. Accordingly there are well-defined ‘best practices’ and a broad base of regulatory, finance and academic literature on which to draw. For example, an entirely different approach to risk management comes into play in managing exposure to the legal, regulatory, underwriting and control risks of issuing debt. A robust organization structure helps limit exposure to some of these risks, as described below. Well-designed accounting, cash management and transaction databases and procedures also help. Experienced legal advice is necessary, as is advice on individual markets and managing ‘market presence’.



1.9.1 Organizational Structures to Limit Portfolio Operational Risk

Some types of organizational structures have proven to be effective at providing appropriate incentives and other approaches to limit operational risk. For example, the ‘risk management’ unit needs to report to the apex of the management team, independently of the managers directly involved with the portfolio decisions. The risk management unit needs some staff with strong quantitative analytical skills and some with experience in markets and transactions. The performance measurement function should reside in a separate management stream from the personnel responsible for transactions. Accounting and transactions processing should be segregated from the personnel responsible for generating transactions. The internal audit function should report directly to top management and should be appropriately staffed with experienced, qualified individuals. Policy and risk management should be independent from the transactions group, and authority should be commensurate with responsibility throughout the organization. A number of classic “red flags” that indicate a flawed organization structure are described in Table 2. Further examples relate to organizational checks and balances in data entry and validation. Yet others refer to the incentives and controls structure for the traders. Many of the largest external auditing firms as well as the commercial bank regulators provide information on ‘best practices’ in this broad area.7 Table 3 lists principles for sound operational risk management that are drawn from recommendations for banks 8

7 For example, see documents available from the U.S. Board of Governors of the Federal Reserve System at www.federalreserve.gov. Supervision Manuals (1) Bank Holding Company, (2) Commercial Bank Examination Manual, and (3) Trading and Capital-Market Activities. Another source is the Bank for International Settlements in Basel, whose web site www.bis.org lists many publications, especially those from the Committee on Bank Supervision and Regulation.

8 Basel Committee on Banking Supervision, “Sound Practices for the Management and Supervision of Operational Risk”, February, 2003. Bank for International Settlements, Basle, Switzerland.



Table 2Red Flags that Indicate Lack of

Organizational Controls

Stacked One-on-One Reporting Relationships. This concerns relationships among managers, supervisors and workers. Assume a Manager has an Assistant Manager as a sole report who in turn has an Assistant Deputy Manager as a sole report. The Assistant Deputy Manager supervises three first-line supervisors, each of whom has responsibility for several actual workers. In the example, the Manager must deal with three additional layers of supervision before reaching anyone who does the real work.

Information Cubby Holes. The problem here is that while information is created by or provided to the organization, it is not transmitted to the user. A bureaucrat who retains it in a power-aggrandizement scheme, or who simply fails to pass it along, may have captured the information.

Circuitous Information Pathways. This often results from repeated reorganizations. The information flows were not considered when people were moved, with the result that information flows are not direct. This often results in slow transmission of decision-critical data or lack of focus in the data. Late data have the same value as no data. Obscure data may escape the attention of the decision-maker.

Delivery of Time-Critical Data to the Wrong Address. Here, the information is clear and timely

but does not reach the user. The effect is the same as if the data were never created or observed.

Crisis Management as the Norm. When everyday operations require the intervention of senior officials, as in a true crisis, the organization has failed. Crisis management should be the rare exception and be invoked as the result of an external event.

Forced Coordination. Complex tasks require complex organizations. The various organizations must communicate, share data, and accept one another’s roles for both effectiveness’ and efficiency’s sake. Failures to communicate or support one another in routine operations that require the intervention of a senior manager indicate an organizational failure.

Internal Bartering. Complex tasks often require input from more than one office. Sometimes, offices are reluctant to carry their end of the load. Evidence of this phenomenon is a statement by a person in another office in response to a normal request to the effect that “I’ll do it, but it will cost you!” instead of “When can I have that done for you?”

Negotiated Report Results. Reports should reflect data gathered from the real world, analyzed by objective professionals. When reports are tailored to fit preconceived notions from within the organization, there is a problem. A grant of independence to the analysis unit is a cure.

Static Organization. No organization structure is perfect for all times and all circumstances. Things change. Markets evolve. Needs arise, then go away. New ways of doing things become possible through technology. And so on. The organization must change in response to these stimuli. When such stimuli exist, an organization that has not changed is the problem, not the proposal for change.

Responsibility Without Authority. This is the classic trap of the dysfunctional bureaucracy; a worker is held accountable for performance of a task without being given the authority to complete it. The usual cause is someone’s desire to avoid responsibility.

Excess Staff. This is not just an efficiency problem. It has effectiveness aspects. The old saying among managers “If you need something done, give it to a busy person” is applicable. Underutilized people are demoralized and dissatisfied. Most people would prefer to be challenged than to be told, in effect, “We don’t expect very much from you, so we are paying you eight hours to accomplish four hours of work.”

Lack of Career Progression. A healthy organization is self-regenerating. People have the opportunity to enter, gain skills and experience, and progress in responsibility and compensation. People should also have the opportunity to move on to other organizations. Dead-end jobs attract deadhead people. Debt management is too important to be entrusted to the intellectually dead.

Risk Management Unit With No Authority. The risk management unit should be considered an integral function, reporting directly to senior management and with authority commensurate with responsibility. Red flags include risk managers reporting through transactions personnel; risk managers with responsibility but only nominal authority; tension between risk managers and transactions personnel allowed to fester by senior management.

Independent Department Limited to One Type of Transaction. This is a classic reason a portfolio becomes too concentrated in one type of instrument or one foreign currency. A red flag is lack of coordination, at a relatively low organizational level, across the units charged with transacting the various instruments available to the portfolio so that the relative volume of each instrument meets an overall portfolio goal.



Table 3

Principles for Sound Practices for Managing and Supervising Operational Risk

1. The Senior Ministers should:

Approve and periodically review the operational risk management framework. It should have a wide definition of operational risk and lay down the principles of how operational risk is to be identified, assessed, monitored, and controlled/mitigated.

Ensure that the operational risk management framework is subject to effective and comprehensive internal audit by operationally independent, appropriately trained and competent staff. The internal audit function should not be directly responsible for operational risk management.

2. Senior management should:

Be responsible for implementing the operational risk management framework consistently throughout the Ministry.

Identify operational risk in all material activities, processes and systems.

Be responsible for developing policies, processes and procedures for managing operational risk in all material activities, processes and systems to control and/or mitigate operational risks..

Implement a process to regularly monitor operational risk profiles and material exposures to losses and report pertinent information to senior Ministry officials.

Establish adequate operational risk assessment procedures before new activities, processes and systems are introduced or undertaken.

3. Ministry management and staff should:

Periodically review their risk limitation and control strategies and adjust their operational risk profile accordingly using appropriate strategies, in light of their overall risk appetite and profile.

Have in place contingency and business continuity plans to ensure their ability to operate on an ongoing basis and limit losses in the event of severe business disruption.

Make sufficient public disclosure to allow market participants to assess their approach to operational risk management.



1.9.2 Structured Approach to Business Functions and Processes

The main participants and their roles, and the stakeholders and their goals and constraints, are either known or emerge through the structured approach to organization architecture described in this chapter. The functions and business processes associated with sovereign debt management are likely to include the items listed below, among others. Each country will organize its own linkages of functions and processes for debt management. “Best practices” can be implemented in an organizational structure that will accommodate the particular existing setting, skill base, cultural sensitivities, budget status, and other realities in a particular country. That is, one way or another all the functions and processes listed below will be performed but not necessarily with the same organizational structure in each country.

Functions

• Policymaking

• Planning and Program Management

• Transacting Borrowings

• Cash Management

• Risk Management

• Analysis

• Legal

• Operations and Controls

• Accounting

• Budgeting

• Auditing

Business Processes

• State management objectives and risk tolerances

• Develop policies

• Authorize financing instruments

• Approve funding program

• Plan currency mix

• Develop financing plan

• Establish frequency, volumes and instruments of issuance

• Implement portfolio limits

• Implement counterparty risk limits

• Advise on policy and risk management levels

• Design new instruments

• Establish and maintain relationships with investors and underwriters

• Borrow: bring transactions to market via auction, syndicate, or private placement

• Select and manage primary dealers

• Manage funded debt outstanding

• Manage day-to-day relationships with domestic financial institutions

• Manage day-to-day relationships with external financial institutions

• Report compliance with portfolio guidelines to policy/risk management

• Monitor performance

• Manage market risks

• Maintain official deal documentation

• Manage cash flows: Pay and collect

• Control currency flows

• Monitor cash flows

• Project short-term cash balances and requirements

• Provide timely information on amount and timing of funds availability

• Monitor government securities markets and advise on market preferences

• Follow macroeconomic forecasts and trends for their impact on debt management

• Maintain relationships with rating agencies

• Prepare prospectuses and other similar debt issuance materials

• Monitor and receive data on budget revenue and performance versus plan

• Analyze and report debt data for risk/performance management purposes

• Maintain Analytical Database

• Develop and maintain portfolio model

• Build assumptions



• Run portfolio model for medium term projections, stress testing and benchmarking

• Ensure compliance with domestic law

• Ensure compliance with laws in other countries where debt is issued

• Prepare contracts and agreements

• Approve disclosure and other similar financing documents

• Hire external counsel where needed

• Manage delivery of securities or maintenance of book-entry registry

• Monitor arrangements for DVP— delivery of securities against full payment of funds due

• Conduct debt administration

• Maintain relationships: fiscal agents, registrar, paying and transfer agents

• Maintain General Ledger and its databases

• Maintain official records

• Manage input to Analytical Database

• Manage information staff and systems

• Audit

• Train staff

• Manage for appropriate professional work environment

• Review organization structure periodically

In addition to the best practices, incentives, controls, and other organizational issues that face any debt management function, there are several other aspects unique to debt management undertaken on behalf of the sovereign. For one thing, the portfolio is managed on behalf of public trust, so that serious losses have significantly more far reaching consequences than those associated with a private portfolio. For another, sovereign debt is a very important macroeconomic tool, with consequences ranging from development and maintenance of efficient markets, to the conduct of monetary policy, to crisis avoidance and crisis management. These attributes of sovereign debt management add to the list of functions and business processes, with the following items likely to be among those included. The standards for best practices

may be more culturally sensitive, and therefore country-specific, than the standards for the preceding set of functions and business processes.

Sovereign-Related Functions

• Market Development

• Central Banking

• Legislation

• Managing Indirect and Contingent Obligations

• Market Regulation

Sovereign-Related Business Processes

• Advise policymakers on all aspects of debt management and their impact on the budget, macroeconomic stabilization, crisis exposure and readiness

• Build political consensus

• Monitor government securities markets and advise on market preferences

• Avoid crises to the extent possible

• Manage crises to minimize their damage

• Monitor auction results and arrangements for pricing efficiency

• Manage indirect issuers as to timing, volume and nature of issuance

• Manage exposure and public benefits from indirect and contingent obligations

• Provide material and timely information to the public, the legislature, the media, and other stakeholders in the public debt

• Assure transparency and fairness in the markets

• Develop markets for efficient capital allocation in the economy

• Develop legal and regulatory environment supporting market efficiency



1.10 Information Architecture Using Separate but Integrated Databases

It is not possible to manage a portfolio efficiently if risks and compensation for risk exposure are not properly measured. A good quantitative foundation is very helpful.

A valuable starting point is a review of the existing informational and organizational structure of the national debt management and supporting Information Technology functions, with a plan for restructuring if needed. The preceding section describes a comprehensive approach to this process in which the results for each country would be unique. This process can be time consuming at the front end, and requires skilled leadership, but it is an important component of producing high quality results.

‘Best practices’ for the organizational structure consistent with healthy incentives and effective institutional risk management/controls structure have been promulgated by, among others, commercial bank regulators in the U.S. and other countries, the Bank for International Settlements and by the major external auditing firms. These “best practices” provide a helpful comparison to the map of the existing organizational structure in each country that would be an outcome of the information and organizational mapping process described here. This comparison will help to guide any institutional changes that might be warranted for improving controls and incentives toward achieving the sovereign’s goal.

The other outcome of this process is the map of information flows. This would be the bedrock of the design of the data system to support risk measurement. To the extent feasible, it is also desirable to use ‘best practices’ to design the information strategy and to use tools with proven track records. The advantages are: (1) high quality of data as to its completeness and its accuracy; (2) improved chance for success in systems design and reliability; and (3) transfer of technical capacity at a very high level. The drawbacks are likely to be: (1) potential large initial investment for licensing; and (2) potentially intense technical training

in systems and quantitative support for the sovereign debt management team.

The structured interviews on data creation, use, sharing and disposal will show that data needs fall into two broad categories. One requires continually updated information for recording, accounting, cash management and active liability management. The other requires data snapshots taken at the close of the books for time series and other analysis. The descriptions of data uses also will show that these two categories are associated with different requirements for processing and data handling tools.

With these differences, the most efficient information structure to manage either a single financial portfolio, such as sovereign debt, or a set of portfolios such as in the asset-liability management of a financial institution, is to use two separate databases. One is the ledger record of all transactions, essential for paying bills and for conducting all accounting and control functions as well as active liability management. It is referred to here as the Ledger Database. The other, supporting quantitative risk management, is referred to here as the Analytical Database.

While the two would be distinct, they are integrated in that they draw from the same core data source — the data input to the ledger by the authority responsible for the “data create” function. This structure is one in which each data item, or set of data items, is associated with a business process that has the authority and responsibility for ‘creating’ it. For example, an accounting function supporting the funding transactions might be responsible for entering into the system all of the data attributes for each new debt issue. Other organizational units might ‘use’ that data item, even if they are not responsible for entering it (and hence not responsible for its validity). To continue the example, there may be a central accounting unit that is responsible for reporting the debt balances to the public, even though it is not responsible for having entered or validated the data in the first place. The unit responsible for entering the data item also must be viewed as the authority for the accuracy and completeness



of the data entered, and as such it takes responsibility for its achieving the highest feasible quality level.

This approach obviously enhances the efficiency of data management. For one thing, it assures that there is no duplication of effort: each data item is entered once and only once. For another, it helps to assure that the validation procedures for the Analytical Database can rely on tables produced from the transactions or accounting systems, which often are the source for published and audited information. Furthermore, this linkage of responsibility and authority can be important if, for example, the terms of a particular debt issue or swap transaction are unusual and subject to misinterpretation or perhaps miscalculation of market value or market yield.

This data structure and information flow usually transcends organizational units. A senior-level information management function is needed for effectiveness.

Described briefly here, with a more complete description in Part 2, is an information architecture that consists of a core data system that is modular, that supports a transactional/accounting database and a relational/analytical database. This approach has long been used in businesses, and accordingly there are existing generic database tools that operate in this way. Choices can be made among these database management systems, considering their robustness in meeting defined needs, efficiency of upgrades and ultimate cost-effectiveness as well as their initial cost. The standards by which to judge ledger systems, whether building one using a generic tool or purchasing one specifically designed for debt management, would depend upon clear technical specifications to meet users’ needs. Similarly, the design for the analytical/relational database likely can draw on existing databases that have good track records and that contain many of the same types of instruments, including contingent obligations and derivatives that would be found in a sovereign debt portfolio. In addition, these existing databases can provide useful guides to data table structures, data dictionary contents, validation procedures, and other data management issues. The table structures and

data dictionary for the Sovereign Debt Analytical Database described in Part 2 provide a sound basis for ensuring that all data elements are captured in the information architecture.

The data handling tools for assembling data on existing portfolio into tables, charts and other transactions/accounting and analytical formats are another key component of the quantitative portfolio management. Application of these tools is relatively straightforward if the databases have been structured properly. First, a simple query language for straightforward fact finding and a robust analytical language for more complicated analyses should be available. The Analytical Database is structured to be compatible with these query and analytical tools.

Many analytical and risk management uses are supported by the integrated data structure. One category of these uses is statistical analysis such as time series analysis or cross section analysis. For example, time series analysis might be used to answer the question of the payments performance record for groups of borrowers within the country (e.g., farmers) receiving loans on-lent from the proceeds of a sovereign borrowing. Or cross section analysis might be used to measure exposures and portfolio concentration, such as the relative volume across currencies of debt principal payments contractually scheduled in selected future time periods.

The Analytical Database also is the primary source of information on the existing debt portfolio used with a model that projects future cash flows and portfolios, such as for medium-term planning purposes and for risk measurement. A simulation model that allows projections of many different scenarios also expands use of the quantitative base. This allows ‘stress testing’ of the portfolio(s) and is described in Chapter 2.

risk management for sovereign debt

Documents

sovereign debt management

risk exposure

effective risk management

risk managementchapter

risk management1

risk limits

risk targets

risk tolerance