## _wp12241 ## Source details **Canonical URL:** [_wp12241](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12241.pdf) ## Other formats - [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12241.pdf.md) - [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12241.pdf.json) --- ### Introduction: rationale and scope - Financial crises of the last thirty years show unattended public and private sector risks can be at crises’ origins. - Sovereign exposures vary by development: - Emerging market sovereigns: increased market exposure of net foreign asset and debt portfolios as access to domestic and international capital markets expands. - Frontier market sovereigns: exposure to terms-of-trade shocks and changes in debt refinancing terms; greater vulnerability to exogenous shocks. - Advanced markets: market risk from reliance on capital market access and systemic public debt issuance; additional risks from aging populations, structural reforms, contingent liabilities from systemically important financial or corporate firms, and weak financial sectors and/or subnational entities. - Realization of these risks could cause a significant fiscal and financial drain and a consequent fall in domestic absorption and potential output. ### SALM purpose, objectives, and core activities - Purpose: - SALM (Sovereign Asset and Liability Management) is a balance-sheet approach used to identify and manage key financial exposures and inform macroeconomic and financial stability policy design. - Objectives: - Preserve a sound balance sheet to support a sustainable policy path and economic growth. - Preserve positive sovereign financial net worth. - Support sustainable macroeconomic policies. - Core activities: - Monitor and quantify impact of movements in exchange rates, interest rates, inflation, and commodity prices on sovereign assets and liabilities. - Identify debt-related vulnerabilities and contingent claims. - Inform long-term macroeconomic and developmental objectives (economic diversification, broadening export markets, reducing dependence on key imports). - Identify long-term fiscal challenges such as unfunded social security liabilities implying future claims on resources. ### Portfolio perspective and risk detection - SALM enables: - Analysis of financial characteristics of the sovereign balance sheet. - Identification of sources of costs and risks and quantification of correlations among sources. - Focus on unmatched portions (net financial positions) when assets and liabilities only partially match. - Short- to medium-term strategy: - Develop financial risk management to reduce exposures; for liability net positions, develop strategies to manage associated risks. - Tools and approaches mentioned: - Financial risk models. - Contingent claims approach (CCA). - Intuitive risk management approaches. - Figures and tables referenced in source supporting analysis: - Figure 1. Sovereign Asset and Liability Management Framework. - Figure 2. Overview of Contingent Claims Analysis for the Sovereign. ### Constraints on SALM implementation - Policy and institutional constraints: - Monetary policy objectives affect SALM strategies via market interest rate and exchange rate management and balance sheet size. - Fiscal policy objectives that aim to limit annual debt service costs may constrain debt duration and currency composition. - Market-structure constraints: - Illiquid and/or shallow domestic debt markets and lack of reliable local investor base can prevent domestic debt issuance. - Attempts to issue domestic-currency external debt may be poorly received due to vulnerability to shocks, restrictions on foreign investors, poor transparency, and lack of hedging instruments. - Potential institutional conflicts: - Combining sovereign asset and liability management under one framework may create conflicts between monetary policy and debt management (central bank roles in debt markets). - Examples include central bank reluctance to raise interest rates to contain inflation because of debt rollover cost implications, and intervention policies conflicting with managing foreign currency debt liquidity. - Sovereign balance sheet constraints (as enumerated): - Domestic Constraints: Initial sovereign debt level and composition; Initial sovereign asset level and composition; Policy interest rates and inflation; Contingent liabilities; Institutional organization; Market development. - External Constraints: Exchange rates; External interest rates; Commodity prices; Asset prices; Market risk appetite. ### Application notes and country relevance - Commodity-exporting countries: - High commodity prices increase current revenues and strengthen unadjusted fiscal positions. - Commodity depletion leads to lower future revenues; appropriate asset management can mitigate to ensure future fiscal expenditures are sustainable. - SALM is part of overall macroeconomic management; the paper will define SALM (Section II), survey country experiences (Section III), and conclude (Section IV). ### Comparative perspective: sovereigns vs private sector (Box 1) - Shared risks: - Macro-level: business cycle, demographic, natural disaster risk. - Micro-level balance sheet exposures: interest rates, exchange rates, credit risk, inflation rates, systemic risk factors, commodity prices. - Key differences: - Balance sheet reporting: private sector benefits from clear accounting; sovereigns lack a clear sovereign balance sheet definition and full public sector consolidation. - Market instruments: some sovereigns lack capacity or political latitude to use derivatives; purchase of credit derivatives can be politically controversial. - Reputational/political risk: sovereign derivative losses face public scrutiny (example cited in source: Ecuador, early 1993, Daniel (2001)). - Systemic importance: sovereign distress can trigger financial and economic crises; well-managed sovereigns can lower private sector risk premia. - International reserves and foreign-currency assets: - Role: meet balance of payments and intervention needs; assets must be sufficiently liquid to smooth intertemporal consumption. - Mundell liquidity index example: with income velocity of money equal to four, the rule implies an optimal level of reserves equal to three months of imports. - Alternative reserve adequacy measures: in emerging markets with open capital accounts, reserves may be measured relative to short-term public and private debt by residual maturity; one study shows adequate reserves should cover one year’s worth of debt obligations. - IMF (2011) proposes a framework taking into account external debt level, exports, and broad money. - Reserve composition guidance: match currency composition of reserves to future import needs or to currency composition of short-term debt. - Managing ample foreign-currency assets: use excess reserves to buy back external debt, reduce balance sheet size, lower currency and maturity mismatch and carry cost; excess beyond that could be allocated to SWF-type arrangements with clear rules. - SWF taxonomy and dynamics: - Five types of SWFs by policy objective: (i) stabilization funds; (ii) savings funds; (iii) development funds; (iv) pension reserve funds; (v) reserve investment corporations. - Countries can maintain multiple SWFs concurrently; objectives and asset tranching may evolve as assets grow. ### SALM objectives under specific cases (from Box 1) - Net debtor / tax-smoothing: - Objective: minimize need to raise taxes or curtail services after shocks. - Instruments recommended: avoid exchange rate exposure and refinancing risk — preference for domestic-currency denomination and long-term debt/balanced maturity structure. - Smoothing fiscal balances: - Optimal debt structure depends on correlations among output, inflation, and interest rates; instrument choice varies with policy regime (inflation-targeting vs output stabilization). - Practical objective: minimize cost of financing and debt service while incurring a prudent degree of risk. - Reducing balance sheet risks: - Foreign currency mismatch management: define optimum share of foreign-currency debt minimizing projected debt servicing costs subject to risk constraints. - Maturity mismatches: interest rate exposure depends on average maturity, coupon linkage, and asset structure; account for derivatives and off-balance-sheet exposures. - Other objectives: - Align SALM objectives with macroeconomic framework; additional goals can include attaining better credit rating, correcting externalities, providing public services, income redistribution, and poverty reduction. ### Defining relevant assets and liabilities - Scope distinctions: - Narrow SALM: international reserves and foreign currency debt. - Broader SALM: all sovereign financial assets and liabilities (excluding nonfinancial ones). - Accounting valuation issues: - Market (mark-to-market) vs historical valuation produce different asset/liability values. - Contingent liabilities valuation depends on discount rate, life expectancy, inflation assumptions. - Example: equity in loss-making state-owned companies may be much lower at market value than historical value. - GFSM (Government Finance Statistics Manual, 2001) definitions: - Assets: those over which general government has ownership rights and may derive economic benefits; classified as financial or nonfinancial. - Liabilities: obligations to other institutional units. - Net worth = assets − liabilities; net financial worth excludes nonfinancial assets. - GFSM excludes central bank balance sheet from general government but encourages compiling central bank statistics and permits integrated treatment for risk analysis when appropriate. - Merton (2007) economic-risk balance sheet contents (as listed): present value of incomes, nondiscretionary expenses, balances (cash, currency reserves, investments), government-owned enterprises and other assets, monetary base, government debt, pension liabilities, contingent claims, and net worth/net financial worth. - Contingent liabilities: - Explicit (credit guarantees, state insurance schemes) and implicit (potential financial system bail-outs, corporate bail-outs, natural disasters); difficult to quantify and more likely to be realized during recessions. - Contingent Claims Analysis (CCA): - Treats international reserves, external debt, net fiscal assets, equity, value of money issuance, domestic debt, and base money as balance sheet items; derives market-value and volatility of sovereign assets from option pricing using market prices/volatilities of liabilities. ### SALM methodologies (select) and uses - Intuitive frameworks: - Practical recommendations include keeping low levels of foreign currency and short-term debt and linking reserves to external debt service in developing countries. - Mean–variance approach: - Determine allocation maximizing portfolio return subject to prespecified risk level; allows constraints such as minimum share of risk-free asset, maximum duration, maximum exposure to each currency, upper limits on weighted downgrade and default probabilities, and upper limits on transaction costs. - Value-at-Risk (VaR): - Purpose: calculate a “worst case” loss; typical severe loss often termed as one that has a 1 percent probability of occurring over a certain horizon. - Common normality assumption: 1 percent chance that losses exceed 2.33 standard deviations. - Parametric 99 percent VaR formula: VaR = 2.33 × σ × √T (where σ is standard deviation of portfolio value; T is time period). - Alternative tail probability: 2.5 percent tail implies 1.96 standard deviations. - Three calculation methods: (i) parametric VaR; (ii) historical simulation; (iii) Monte Carlo simulation. - Cost-at-Risk (CaR): - Purpose: compute debt-servicing costs and associated risks using stochastic simulation of macro variables. - Outputs: distribution of debt level, expected costs, Absolute CaR often measured as the 99th percentile of debt-servicing costs, Relative CaR as difference between 99th percentile and expected costs. - Efficient frontier: plot cost-risk combinations and select optimal strategy at a prespecified risk level. - Multivariate Adaptive Regression Splines (MARS): - Purpose: estimate government objective function when high-dimensional and nonlinear. - Empirical findings: MARS can outperform OLS, NKR, and PPR in complex settings. - Illustrative results for financing options (3-month, 1-year, 5-year securities): - Minimizing cost → higher weight to shorter-term maturities (up to one year). - Risk constraints or risk-averse utility → increased weight on longer-term maturities. - Minimizing cost and probability of budget deficit → share of longer-term maturities increases. - Very high risk aversion → optimal portfolio can be polar (e.g., only three-month and five-year securities). - Limitations: treats macro variables as given; data requirements remain a challenge. - Dynamic stochastic optimization model: - Use Markov decision processes, sparse stochastic trees for exogenous variables, solve dynamic stochastic optimization, and estimate outcome densities for policymaker preferences. - Allows inclusion of rare large-loss events and flexible scenario/stress testing. - Contingent Claims Analysis (CCA): - Uses market information on sovereign liabilities to infer sovereign asset value/volatility via option pricing (Black–Scholes). - Outputs distance-to-distress, default probabilities, credit spreads. - Distress barrier (DB) per Gray and Malone (2008): DB = STD + α·LTD + IPTM (STD = short-term external debt ≤ 1 year; LTD = long-term external debt > 1 year; α between 0 and 1, usually around 0.5; IPTM = interest payments on external debt due in the first year). - Indicators estimable: 1-year and 5-years distance to distress (Φ_t^1), 1-year and 5-year risk neutral default probabilities RNDP_t = 1 − N(d_2), 1-year and 5-year credit default spreads. - Data requirements: for each debt security—coupon, amount outstanding, maturity, current market price, market (external or domestic), currency, coupon type (fixed/floating/indexed), coupon frequency; and time series on foreign/domestic interest rates and FX rates. ### Institutional, operational, and market prerequisites - Interagency coordination: - Effective SALM requires high institutional coordination among entities managing sovereign financial assets and liabilities; legislative backing is desirable. - If central bank conducts quasi-fiscal operations or bailouts, consolidated management including central bank and government balance sheets is prudent. - Debt capital market development: - SALM presumes minimum domestic capital market development (depth and liquidity) and financial intermediation (investor base). - Constraints in market infrastructure, yield curve development, benchmark issues, maturities, or limited investor base may render SALM-suggested allocations infeasible. ### Appendix II — institutional practices and risk management metrics - Sovereign balance sheet typical items: - Assets: "cash deposits, other special purpose funds—including pension reserve funds and sovereign wealth funds, foreign reserves, and receivables." - Liabilities: "the government debt stock, and in some cases, public guarantees and deposits by local authorities and commercial banks.." - Liability management objective: "ensure financing of the budget at the lowest possible cost subject to an acceptable level of risk." - Asset management objective: "ensure that cash balances meet commitments and maximize the purchasing power of the long-term capital given moderate level of risk." - Partially coordinated SALM practices (examples cited): integrated net position management (Finland, Greece, Turkey); asset allocation reflecting broader priorities (Australia); consolidated interest-rate risk management when funds hold government bonds (Denmark); consolidated balance sheet preparation when reserve increases require foreign issuance (Hungary); reserve duration/currency decisions based on government FX liabilities (Turkey). - Risk management metrics and techniques: - "Duration and modified duration are used commonly in both asset and liability management." - "VaR is used more commonly by asset managers, while CaR by liability (debt) managers." - "Asset managers also use scenario analysis and stress tests more often than liability managers; counterparty limits are common to prevent credit risk." - "Liability managers commonly use ratio analysis — in particular the refixing ratio — to monitor risks." - Forecasting: - Asset and liability managers rarely produce in-house forecasts for exchange rates, interest rates and inflation; when done, based on historical econometric models or forward curves; forecast data sourced from national statistical agencies or specialized units. - Institutional arrangements: - Management typically by a debt management office, the central bank, or statutory agencies. - Coordination mechanisms include delegation to central bank with ministry oversight (Canada), treasury/finance ministry mandates (Australia), or regular interagency meetings (Hungary, Canada). - SWFs sometimes on central government balance sheets and have issued debt at low levels; effects include functioning financial markets and improved liquidity management (Australia, Norway references in source). - Concluding practical points: - Transitioning from treasury management to SALM is lengthy. - Benefits: joint analysis of assets and liabilities, comprehensive sovereign portfolio stress testing, optimal debt and asset portfolio management to attain desired balance sheet risk exposure cost-effectively. - Requirements: consolidated view of sovereign balance sheet, coordinated portfolio strategy decisions, clarity about macroeconomic policy goals, central role for risk assessment, elaboration of alternative scenarios (short- and long-term costs) and stress tests, quantification of contingent liabilities (e.g., via CCA), and publication of risk information to enhance transparency and accountability. - Institutional trade-offs: combining asset and liability management under one framework may not always be desirable or feasible; clear roles, objectives, and resources across institutions are necessary to maintain independence and coordination. ### Appendix I & II — technical methodologies and modeling examples (key points) - Mean–variance: maximize expected next-period excess return subject to portfolio variance σ^2 ≤ W_t' V_t W_t; budget constraint sum ω_{i,t+1} = 1; nonnegativity ω_{i,t} ≥ 0; additional constraints (minimum risk-free share, maximum duration, maximum currency exposure, upper limits on weighted downgrade/default probabilities, upper limits on transaction costs). - VaR specifics: - Typical severe loss: 1 percent probability over horizon. - 1 percent → 2.33 standard deviations; parametric 99 percent VaR: VaR = 2.33 × σ × √T. - 2.5 percent tail → 1.96 standard deviations. - CaR specifics: - Absolute CaR often the 99th percentile of debt-servicing costs; relative CaR = 99th percentile − expected costs. - MARS results: - Optimal financing tenor shifts with objective: cost minimization → shorter maturities; risk aversion → longer maturities. - Dynamic stochastic optimization: - Generate sparse trees of stochastic variables (interest rates, exchange rates, GDP, deficits, commodity prices, rare events), formulate and solve model with decision variables (reserves, debt levels, maturities, currency shares), estimate outcome density functions for policy preferences. - CCA technical elements: - Distress barrier DB = STD + α·LTD + IPTM with α ≈ 0.5 typical. - RNDP_t = 1 − N(d_2) and distance-to-distress and credit spreads derivable from option-pricing relationships; data-intensive (coupon, amount outstanding, maturity, market price, market, currency, coupon type, coupon frequency; time series on interest rates and FX rates). - Interest rate and exchange rate modeling examples: - Term-structure models: CIR, CIR2, NS (Nelson–Siegel), HW (Hull–White), HJM, DK. - CIR and CIR2 dynamics (mean reversion, Brownian motions) and use to decouple short- and long-term rates. - NS: r_t(τ) expressed as β1,t + β2,t*(...) + β3,t*(...), with β factors interpreted as level, slope, curvature and decay parameter λ_t. - Exchange rate CKLS variant and real-to-nominal mapping: real exchange rate stochastic process and nominal exchange rate defined by nominal = (external price index / domestic price index) × real rate in source equations. - Estimation and data constraints: - Parameter estimation requires historical data series; emerging markets may lack long-time data series on interest rates and other variables, limiting model adoption. - Foreign interest rate parameter estimation easier due to deep and liquid markets. *Source: _wp12241 - References; Box 1; Appendix I; Appendix II.* ### References ............................................................................................................. ### _wp12241 - References ............................................................................................................. ### Introduction: rationale and scope - Financial crises of the last thirty years have shown that unattended public and private sector risks can be at crises’ origins. - Sovereigns are exposed to multiple risks depending on economic and financial development: - Emerging market sovereigns: increased market exposure of net foreign asset and debt portfolios as access to domestic and international capital markets expands. - Frontier market sovereigns: exposure to terms-of-trade shocks and changes in debt refinancing terms; greater vulnerability to exogenous shocks. - Advanced markets: market risk from reliance on capital market access and systemic public debt issuance; additional risks from aging populations, structural reforms (health and pension), contingent liabilities from systemically important financial or corporate firms, and weak financial sectors and/or subnational entities. - Realization of these risks could cause a significant fiscal and financial drain and a consequent fall in domestic absorption and potential output. ### SALM (Sovereign Asset and Liability Management) purpose and uses - SALM is a balance-sheet approach used to identify and manage key financial exposures and inform macroeconomic and financial stability policy design. - SALM objectives: - Preserve a sound balance sheet to support a sustainable policy path and economic growth. - Preserve positive sovereign financial net worth. - Support sustainable macroeconomic policies. - Core SALM activities: - Monitor and quantify impact of movements in exchange rates, interest rates, inflation, and commodity prices on sovereign assets and liabilities. - Identify debt-related vulnerabilities and contingent claims. - Inform long-term macroeconomic and developmental objectives (economic diversification, broadening export markets, reducing dependence on key imports). - Identify long-term fiscal challenges such as unfunded social security liabilities implying future claims on resources. ### Portfolio perspective and risk detection - SALM allows analysis of financial characteristics of the balance sheet, identification of sources of costs and risks, and quantification of correlations among sources. - If assets and liabilities only partially match, risk management should focus on unmatched portions (net financial positions). - Short- to medium-term strategy: develop financial risk management to reduce exposures; for liability net positions, develop strategies to manage associated risks. - SALM tools mentioned: - Financial risk models. - Contingent claims approach. - Intuitive risk management approaches. - Figures and tables referenced in source supporting SALM analysis: - Figure 1. Sovereign Asset and Liability Management Framework - Figure 2. Overview of Contingent Claims Analysis for the Sovereign - Tables and boxes enumerated (e.g., Government Balance Sheet definition, Stylized Economic-Risk Balance Sheet, comparison box). ### Constraints on SALM implementation - Policy and institutional constraints: - Monetary policy objectives affect SALM strategies via market interest rate and exchange rate management and balance sheet size. - Fiscal policy objectives that aim to limit annual debt service costs may constrain debt duration and currency composition (e.g., high shares of short-term debt perceived to lead to greater volatility of service costs). - Structure of international and domestic capital markets shapes SALM implementation: - Illiquid and/or shallow domestic debt markets and lack of reliable local investor base can prevent domestic debt issuance. - Attempts to issue domestic-currency external debt may be poorly received due to vulnerability to shocks, restrictions on foreign investors, poor transparency, and lack of hedging instruments. - Potential institutional conflicts: - Combining sovereign asset and liability management under one framework may create conflicts of interest between monetary policy and debt management (central bank operational roles in debt markets). - Examples: central bank reluctance to raise interest rates to contain inflation because of debt rollover cost implications; central bank intervention policies may conflict with managing foreign currency debt liquidity. - Sovereign balance sheet constraints enumerated (as listed in source): - Domestic Constraints: Initial sovereign debt level and composition; Initial sovereign asset level and composition; Policy interest rates and inflation; Contingent liabilities; Institutional organization; Market development. - External Constraints: Exchange rates; External interest rates; Commodity prices; Asset prices; Market risk appetite. ### Application notes and country relevance - SALM can highlight asset management challenges for commodity-exporting countries: - High commodity prices increase current revenues and strengthen unadjusted fiscal positions. - Commodity depletion leads to lower future revenues; appropriate asset management can mitigate to ensure future fiscal expenditures are sustainable. - SALM framework forms part of overall macroeconomic management strategy with references to country experiences and literature (authors cited in source text). - The paper reports it will: - Define the SALM approach and main framework elements (Section II). - Survey select country experiences in SALM application (Section III). - Provide concluding remarks (Section IV). ### Methodological and conceptual elements (preview) - Section II focuses on theoretical underpinnings: - Describes relevant assets and liabilities to include in analysis (Das, 2011). - Presents main methodologies used within an SALM framework. - Uses these conceptual and methodological issues to compare and assess country experiences with SALM. *Source: _wp12241 - References .............................................................................................................* ### Box 1. Comparison of Risk Management Approaches: Sovereigns and the Private Sector ### Box 1. Comparison of Risk Management Approaches: Sovereigns and the Private Sector ### Comparative overview of risks - Both sovereigns and the private sector face macro-level risks: business cycle risk, demographic risk, and natural disaster risk. - Both face micro-level balance sheet exposures to: interest rates, exchange rates, credit risk, inflation rates, systemic risk factors, and commodity prices. - Consequences of unmanaged risks: - Sovereigns: external vulnerability and/or debt stress. - Private sector: inadequate cash flow for operations or debt obligations, undercapitalization, or insolvency. ### Key differences in practices, tools, and constraints - Balance sheet reporting and consolidation: - Private sector: clear specification of assets, liabilities, and equity; international accounting standards facilitate comparisons. - Sovereigns: lack of a clear sovereign balance sheet definition; incomplete public sector balance-sheet reporting; difficulty consolidating all public sector balance sheets—assets and liabilities often hard to quantify—making risk management difficult. - Availability and use of market instruments: - Some sovereigns are not well-equipped to use financial derivatives. - Domestic interest rates, exchange rates, and inflation are linked to macroeconomic policy objectives; sovereigns may be unable to actively use derivatives because markets could misinterpret their actions. - Purchase of credit derivatives to cover sovereign credit exposures can be politically controversial. - Reputational and political risks: - Sovereigns face public scrutiny and reputational risk from large derivative losses (example cited: Ecuador, early 1993, Daniel (2001)). - Systemic importance: - A sovereign under debt distress can trigger a financial and economic crisis. - A well-managed sovereign with sound macroeconomic policies, investor-friendly environment, and high credit ratings can reduce private sector risk premia and mitigate private sector risk management burdens. - Some sovereigns may act as risk managers of last resort, providing implicit guarantees on private sector exposures (especially foreign exchange). ### Role of international reserves and foreign-currency assets (Case 1 and Case 2) - Primary role of foreign-currency sovereign financial assets: - Meet balance of payments and intervention needs to smooth intertemporal consumption; such assets must be sufficiently liquid to redistribute shocks through time. - Mundell liquidity index example: - For a country with income velocity of money equal to four, the rule implies an optimal level of reserves equal to three months of imports. - The three-month rule is cited as a benchmark for reserve adequacy, especially with relatively closed capital accounts and limited market exposure. - Alternative reserve adequacy measures: - In emerging markets with open capital accounts, reserves may be measured relative to short-term public and private debt by residual maturity; one study shows adequate reserves should cover one year’s worth of debt obligations. - IMF (2011) proposes a framework taking into account external debt level, exports, and broad money. - Reserve composition guidance: - Match currency composition of reserves to future import needs or to currency composition of short-term debt when reserves provide coverage for short-term debt. - Managing ample foreign-currency assets: - Excess reserves can be used to buy back external debt, reduce balance sheet size, and lower currency and maturity mismatch risk and carry cost. - If excess remains after redemptions/buy-backs, assets could be allocated to Sovereign Wealth Fund (SWF)-type arrangements with clear funding and withdrawal rules. - Cases where investment funds are sensible include: - (i) a local economy that cannot absorb additional spending without overheating; - (ii) windfall profits likely to disrupt the domestic economy (“Dutch disease” effects); - (iii) uncertain future in countries rich in non-renewable natural resources; - (iv) a country needing to diversify its asset base. - Five types of SWFs (based on stated policy objectives): - (i) stabilization funds; - (ii) savings funds; - (iii) development funds; - (iv) pension reserve funds; - (v) reserve investment corporations. - SWF institutional and objective dynamics: - Countries can establish multiple SWF types concurrently. - As SWF assets grow, objectives may change and assets may be split into tranches or separate funds (example pattern: initial stabilization fund later broadened—Norway referenced). ### SALM objectives when sovereign is net debtor (Case 3) - Tax-smoothing objective: - SALM literature objective: minimize need to raise taxes or curtail services when faced with unexpected shocks causing permanent fiscal revenue falls or rises in liabilities. - Optimal debt structure hedges non-diversifiable risks to reduce taxation costs; depending on tax cost function, strategies differ: - If rise in taxes yields proportional welfare losses: choose portfolio with highest return, even high leverage. - If taxation costs increase more than proportionally: minimize variance of taxation path by issuing debt negatively correlated with public spending and positively correlated with tax returns. - Recommended instruments to insure against government revenue, spending, and interest-rate shocks: debt instruments avoiding exchange rate exposure and refinancing risk; preference for domestic-currency denomination to avoid exchange rate exposure and long-term debt/balanced maturity structure to reduce refinancing risk. - Smoothing of fiscal balances objective: - Under a budget-deficit constraint, optimal debt structure depends on correlations among output, inflation, and interest rates. - If output and inflation are low, issuing short-term debt can achieve fiscal stability given a normal upward sloping yield curve. - If output and inflation are positively correlated, issue inflation-indexed or variable-rate bonds to offset cyclical deficits. - In an inflation-targeting environment: issue long-term and inflation-indexed debt. - If monetary authorities prioritize output stabilization: issue short-term bonds; inflation-indexed bonds may be inappropriate. - Practical debt management objective in many countries: minimize cost of government financing and debt service while incurring a prudent degree of risk; reducing variability of debt-servicing costs tends to smooth the budget deficit (IMF and the World Bank, 2001). ### Reducing balance sheet risks (Case 4) - Foreign currency mismatches: - Two dimensions of foreign currency exposure: (i) liquidity risk (ease of obtaining foreign exchange); (ii) balance sheet risk (sensitivity of net wealth/income to exchange rate changes). - Strategic benchmark should define optimum debt structure including optimum share of foreign-currency debt in total debt. - Optimum currency composition typically minimizes projected debt servicing costs subject to constraints on foreign exchange-, interest rate-, credit-, and liquidity-risk levels. - SALM approach often seeks to match the currency composition of debt (liabilities) with that of assets (e.g., international reserves and projected primary balances). - Maturity mismatches: - Interest rate exposure depends on average maturity, coupon linkage of long-dated debt to short-term rates, and the asset structure of the borrower’s balance sheet. - Importance of accounting for derivatives and off-balance sheet exposures when assessing interest rate and currency mismatches. ### Other SALM objectives (Case 5) - SALM objectives should align with the macroeconomic policy framework. - Additional objectives sometimes included: attaining a better credit rating, correcting externalities, providing public services, income redistribution, and poverty reduction. ### Defining relevant assets and liabilities (Section B) - Definitions and scope: - Narrow SALM: international reserves and foreign currency debt. - Broader SALM: all sovereign financial assets and liabilities (excluding nonfinancial ones). - Accounting measure challenge: - Market (mark-to-market) vs historical valuation leads to different asset/liability values; contingent liabilities valuation depends on discount rate, life expectancy, inflation assumptions. - Example: equity in loss-making state-owned companies may be much lower at market value than historical value. - GFSM (Government Finance Statistics Manual, 2001) framework: - Assets: those over which general government has ownership rights and may derive economic benefits; classified as financial or nonfinancial. - Liabilities: obligations to other institutional units. - Net worth = assets − liabilities; net financial worth excludes nonfinancial assets. - GFSM excludes the central bank balance sheet from the general government, but encourages compiling central bank statistics and allows integrated treatment of general government and central bank for risk analysis when appropriate. - Broader sovereign balance sheet definitions: - Merton (2007) economic-risk balance sheet includes present value of incomes (taxes, fees, seigniorage), present value of nondiscretionary expenses, balances (cash, currency reserves, investments), government-owned enterprises and other assets, monetary base, government debt (domestic and foreign), pension liabilities, contingent claims, and net worth/net financial worth. - Contingent liabilities: explicit (credit guarantees, state insurance schemes) and implicit (potential financial system bail-outs, corporate bail-outs, natural disasters); difficult to quantify and more likely realized during recessions. - Contingent Claims Analysis (CCA) approach: treats international reserves, external debt, net fiscal assets, equity, value of money issuance, domestic debt, and base money as balance sheet items; derives market-value and volatility of sovereign assets from option pricing using market prices/volatilities of liabilities. ### SALM methodologies (Section C) - Intuitive frameworks: - Many governments identify risks and vulnerabilities without formal economic balance sheets; common practical recommendations include keeping low levels of foreign currency and short-term debt and, in developing countries, linking reserves directly to external debt service (matching currency and maturity structures). - Select methodologies and their uses: - Mean-variance approach: find optimal portfolio allocation given risk tolerance; requires instrument return paths and covariance matrix; sovereign balance sheet data needed. - Value-at-Risk (VaR): determine maximum volatility of net asset returns consistent with tolerable loss; requires well-defined distribution of net returns and historical return time-series. - Cost-at-Risk (CaR): determine maximum volatility of debt portfolio costs consistent with tolerable debt cost; requires bond yield time-series. - Multivariate Adaptive Regression Splines (MARS): estimate government objective function form and determine optimal portfolio weights given objective parameters; requires time-series of bond yields and asset returns. - Dynamic stochastic optimization model (Claessens and Kreuser, 2004): determine stochastic paths of exogenous variables and optimal allocations; allows for low probability large-loss events and does not require a specific utility function. - Contingent Claims Analysis (CCA): compute risk-adjusted mark-to-market balance sheets, estimate distance to distress, default probability, and credit spreads; requires market information about value and volatility of sovereign liabilities and assumes normal distribution of financial market prices. - Note on mean-variance family: optimal solutions depend on exogenously determined (ad hoc) risk levels chosen by the government. ### Institutional and market prerequisites (Section D) - Interagency coordination: - Effective SALM requires high institutional coordination among entities managing sovereign financial assets and liabilities; legislative backing is desirable to reduce frictions. - In stable macro environments, a CaR approach using basic indicators (GDP, inflation, interest rates) may suffice; more complex integrated approaches yield low payoff relative to resource cost. - If central bank conducts quasi-fiscal operations or bails out institutions, consolidated management including central bank and government balance sheets is prudent; central bank liabilities issued for sterilization/monetary policy can have systemic implications and may require coordination. - Debt capital market development: - SALM effectiveness presumes minimum development of domestic capital markets (depth and liquidity) and financial intermediation (investor base). - Constraints in market infrastructure, yield curve development, benchmark issues, maturities, or limited investor base may render SALM-suggested allocations infeasible, especially for developing countries. *Source: Box 1, Comparison of Risk Management Approaches: Sovereigns and the Private Sector (excerpt).* ### Appendix II). ### _wp12241 - Appendix II) ### A. Sovereign Balance Sheet and SALM Objectives - Asset side typically contains: "cash deposits, other special purpose funds—including pension reserve funds and sovereign wealth funds, foreign reserves, and receivables." - Liability side typically contains: "the government debt stock, and in some cases, public guarantees and deposits by local authorities and commercial banks.." - Main objective of sovereign liability management: "ensure financing of the budget at the lowest possible cost subject to an acceptable level of risk." - Objectives of sovereign asset management: "ensure that cash balances meet commitments and maximize the purchasing power of the long-term capital given moderate level of risk." - Examples of partially coordinated SALM practices: - Integrated management of the net position on central government debt and cash reserves (Finland, Greece, and Turkey). - Allocation of assets between alternative portfolios and funds may take account of broader government priorities, but not specifically of balance sheet risks (Australia). - When government funds hold primarily government bonds, interest-rate risk is managed on a consolidated basis (Denmark). - Coordination when an increase in international reserves requires debt issuance in foreign currencies; central bank may prepare a consolidated balance sheet (Hungary). - Decision about duration and currency composition of international reserves may be based on maturity structure and currency composition of government foreign exchange liabilities, including off-balance sheet foreign exchange cash flows (Turkey). ### B. Risk Management Framework - SALM objectives are frequently specified based on a narrow set of balance sheet items—e.g., "net central government debt and cash reserves." - From an asset management perspective, risk management considerations naturally involve: - "minimizing liquidity and credit risk and, in some cases, interest rate risk." - From a liability management perspective, considerations prevail for: - "minimizing interest rate risk, refinancing risk, and currency risk" among debt managers. - Commonly used metrics and techniques: - "Duration and modified duration are used commonly in both asset and liability management." - "VaR is used more commonly by asset managers, while CaR by liability (debt) managers (see Section II. C and Appendix I)." - "Asset managers also use scenario analysis and stress tests more often than liability managers; and counterparty limits are common to prevent credit risk by asset managers." - "With respect to liability management, ratio analysis—and in particular the refixing ratio—is commonly used to monitor risks." - Forecasting practices: - "Asset and liability managers rarely produce in-house forecasts for the exchange rates, interest rates and inflation." - When forecasts are made in-house: "typically based on historical data using econometric models or are derived from forward curves." - Forecast data typically received from: "national statistical agencies or specialized units in the ministry of finance or the central bank." ### C. The Institutional Framework - Sovereign assets and liabilities are managed by: "a debt management office, the central bank, or statutory agencies charged with the management of specific funds." - Coordination observations: - "Given the prevailing fragmentation in the operational management of sovereign assets and liabilities, coordination between various institutions involved in SALM is not common." - Most integrated case: decision making authority for both assets and liabilities assigned to one agency or ministry—e.g., the ministry of finance—"who delegates responsibilities for day-to-day management—e.g., to the central bank—and coordinates the planning and execution of the borrowing and investing programs (Canada)." - Other mandates set following coordination or delegation by a responsible ministry—treasury or the ministry of finance (Australia). - A more common coordination mechanism: "instituting regular meetings between the agencies involved in SALM (Hungary, Canada)." - Sovereign wealth fund (SWF) practices and effects: - Some countries maintain SWF assets and liabilities on the central government’s balance sheets and have "also issued debt, albeit at low levels." - Effects: "helped maintain effectively functioning financial markets with important consequences for the level of interest rates (Australia) and the effectiveness of liquidity management (Norway)." ### IV. Concluding Remarks - Transitioning from treasury management to SALM is lengthy: "Admittedly, going from managing the treasury function of a government to SALM is a long way." - Benefits of a comprehensive SALM framework: - "joint analysis of the characteristics of financial assets and liabilities on the sovereign balance sheet allows decision makers to take into account more fully the interrelationships and correlations among sources of risks when formulating strategies and policies." - "The SALM framework also provides an appropriate setting for a comprehensive sovereign portfolio stress testing." - Should allow countries to "optimally manage their sovereign debt and asset portfolios and attain a desired level of balance sheet risk exposure in a prudent and cost effective manner." - Requirements and recommended practices for a wider SALM approach: - "A wider SALM approach in managing sovereign risks will require a consolidated view of the sovereign balance sheet and its risk exposures, as well as coordinated decisions on the appropriate portfolio strategies to be adopted to manage that risk effectively." - Design and operationalization challenges: "the design of an effective framework that can be operationalized is a crucial issue, while successful operationalization of such a framework may entail institutional challenges, though not insurmountable." - Precondition: "having clarity about macroeconomic policy goals." - Risk assessment role: "given the high level of uncertainty, risk assessment should play a central role in informing the SALM strategy." - Risk assessment methods: "elaboration of alternative scenarios that encompass short- and long-term costs and stress test for various macroeconomic shocks." - Contingent liabilities: "Potential contingent liabilities in a sovereign balance sheet should also be analyzed, using, for example, quantitative methods such as the CCA framework." - Transparency and accountability: "information about the risk associated with the management of the sovereign balance sheet need to be published; as transparency and accountability help public stakeholders better understand and maintain confidence in SALM." - Institutional trade-offs: "involvement of multiple institutions that comprise the sovereign balance sheet could blur accountability." - Institutional arrangements and flexibility: - "From an institutional point of view, combining the management of sovereign assets and liabilities under one framework may not always be desirable or feasible." - As long as "the roles, objectives and resources of institutions involved in SALM are clearly defined and consistent, an SALM approach could ensure appropriate coordination in the management of assets and liabilities across institutions while maintaining the independence of each institution." - Depending on structural and institutional conditions, an SALM approach could "help better analyze and manage a country’s sovereign balance sheet and comprehensively review possible policy interconnections taking into account identified risks,mitigate potential distortionary spillovers, and better prepare for systemic events." *Source: _wp12241 - Appendix II).* ### APPENDIX I: SELECT SALM METHODOLOGIES ### APPENDIX I: SELECT SALM METHODOLOGIES ### Mean–variance approach - Based on Markowitz (1952); used to determine strategic asset allocation of international reserves by maximizing portfolio return subject to a prespecified risk level. - Allows restriction of investments by currency, maturity and credit rating to preserve value and liquidity of sovereign assets. - Stylized dynamic maximization problem (as presented): - Objective (equation (1)): maximize expected next-period excess return using weights ω_{i,t} and expected returns E_{t}(r_{i,t+1}) relative to risk-free return r_f. - Constraint on portfolio variance (equation (2)): σ^2 ≤ W_t' V_t W_t. - Budget/weight constraint (equation (3)): for all t, sum_{i=1,...,I} ω_{i,t+1} = 1. - Nonnegativity constraint (equation (4)): ω_{i,t} ≥ 0 for all i,t. - Definitions preserved exactly: - t R is the gross return on the sovereign portfolio of reserve assets in period t; - r_f is the return on the risk-free (or liquid) asset; - E_{t+1}(r_{i,t+1}) is the expected in period t next-period return on asset i; - ω_{i,t} is the weight of asset i in the portfolio in period t; - W_t is a vector of portfolio weights; - V_t is a covariance matrix; - σ^2 is an upper limit on the portfolio variance. - Additional constraints that can be imposed: - (i) minimum share of the risk-free (liquid) asset; - (ii) maximum duration of the portfolio; - (iii) maximum exposure to each currency; - (iv) upper limits on the weighted downgrade and default probabilities of the portfolio; - (iv) upper limits on transaction costs. - Conceptual point: government assumes a certain (nonzero) risk level and enforces a maximum acceptable variance rather than minimizing variance to zero. ### Value-at-Risk (VaR) method - Purpose: calculate a “worst case” loss in sovereign asset and/or liability portfolio value during severe adverse market fluctuations. - Typical definition: a severe loss often termed as a loss that has a 1 percent probability of occurring over a certain horizon (for daily losses, implies loss worse than VaR two to three days per year). - VaR computation relies on current holdings and estimated probability distribution of price changes (historical returns or risk factors). - Common normality assumption: 1 percent chance that losses exceed 2.33 standard deviations. - Parametric 99 percent VaR formula shown (equation (5)): - VaR = 2.33 × σ × √T (presented as TT VaR σ × = 33.2 in source formatting) - where σ is the standard deviation of the portfolio’s value; T is the time period over which the standard deviation of returns is calculated. - VaR horizons: monthly or yearly horizons often employed when referring to potential losses. - Alternative tail probability: 2.5 percent tail implies 1.96 standard deviations. - Three common VaR calculation methods: - (i) parametric VaR; - (ii) historical simulation; - (iii) Monte Carlo simulation. - Multiday VaR approximation (footnote): TVaR_{T} = √T × VaR_{1}, assuming (i) changes in market factors are normally distributed; (ii) one-day VaR is constant; (iii) no serial correlation. ### Cost-at-Risk (CaR) method - Purpose: compute debt-servicing costs and associated risks for financing strategies and a given debt profile using stochastic simulation of macroeconomic variables (term structure, exchange rates). - Outputs: - Distribution of the debt level at the end of a time period. - Expected costs measured as the average of the debt-servicing cost distribution. - Absolute CaR often measured as the 99th percentile of debt-servicing costs. - Relative CaR measured as the difference between the 99th percentile and the expected costs. - Methodology: - Plot combinations of cost and risk on a cost-risk plane. - Define efficient frontier as strategies that minimize cost for a given risk level. - Optimal strategy selected at a prespecified risk level from the efficient frontier. ### Multivariate Adaptive Regression Splines (MARS) method - Purpose: estimate government objective function to achieve computational completeness in SALM problems; particularly useful when objective is high-dimensional and nonlinear with limited observations. - Empirical findings (from simulations cited): - MARS outperforms OLS, nonparametric kernel regression (NKR), and projection pursuit regression (PPR) in high-dimensional or complex-function settings. - Illustrative empirical results for optimal debt portfolio (given debt stock and financing options of 3-month, 1-year, and 5-year securities): - If government objective is to minimize cost of debt service: - Optimal strategy gives higher weight to shorter-term maturities—up to one year. - If government faces a risk constraint or maximizes a risk-averse utility: - Weight of longer-term maturities increases. - Method quantifies increase in debt service cost as risk constraint tightens and when other constraints (e.g., limits on instrument shares) are imposed. - If objective is to minimize cost and minimize probability of a budget deficit: - Share of longer-term maturities increases; greater weight on reducing deficit probability → higher optimal share of longer-term debt. - If government is highly risk averse: - Optimal portfolio contains a high share of short-term securities. - Share of long-term maturities increases with risk aversion parameter; for very high risk aversion the government may choose only three-month and five-year securities (a polar portfolio). - Limitations: - MARS treats macroeconomic variables as given; requires separate modeling of their paths. - Data requirements remain a challenge. ### Dynamic Stochastic Optimization Model - Covers stochastic processes where time is essential; emphasis on Markov decision processes and probabilistic dynamic programming. - Uses dynamic stochastic optimization to identify trade-offs with expected present values in time-varying setups. - Proposed ALM framework (Claessens and Kreuser, 2004, 2007) three main steps: - Generate sparse trees of stochastic variables: - Exogenous variables: interest rates, exchange rates, GDP, fiscal deficits, current account deficits, bond prices, commodity prices, rare events triggering contingent liabilities. - Endogenous/decision variables: foreign exchange reserves, debt levels, maturity structures, reserve currency shares, etc. - Scenario generation via multi-factor stochastic partial differential equations; representative form provided (equation (6)): - Σ (...) = μ(...), σ(...), b(...), and stochastic term ϖ(...), as in source presentation. - Parameters (σ, μ, b) estimated from historical data. - Formulate and solve a dynamic stochastic optimization model: - Separate model formulation from stochastic tree/processes to permit changes in constraints, objectives, tree size, or stress tests without mutual interference. - Define events per time period; define decision variables for each event (levels of assets, liabilities, currencies, derivatives). - Decision variables may allow short selling, transaction costs, spreads, and limits. - Assets usually defined in broad classes (short-, medium-, long-duration or currency clusters) rather than individual instruments. - Estimate density functions of outcomes from the model: - Solve model (e.g., maximize expected return at horizon) and estimate density functions using maximum likelihood methods. - Policymakers can express preferences over shapes of density functions (e.g., “a less fat left tail”) rather than specifying utility parameters directly. - Advantage: allows asset and liability allocations based on analyzed density functions resulting from combinations of techniques. ### Contingent Claims Analysis (CCA) approach - Uses observable market information about value and volatility of sovereign liabilities to infer non-observable sovereign asset value and volatility (uses Black and Scholes option formula). - Outputs credit risk indicators: distance-to-distress, default probabilities, credit spreads. - Seniority assumptions: external debt generally considered most senior; domestic debt and base money considered junior contingent claims. - Default definition: when implied asset value falls below a distress barrier. - Distress barrier (DB) definition following Gray and Malone (2008) (equation (7)): - DB = STD + α·LTD + IPTM - STD = short-term external debt (maturity ≤ 1 year); - LTD = long-term external debt (maturity > 1 year); - α is a parameter between 0 and 1 (usually around 0.5); - IPTM = interest payments on external debt due in the first year. - Sovereign asset and volatility implied equations (equations (8) and (9) as presented) to solve for $Sov V and $Sov σ using option-pricing relationships and observed $LCL (base money plus domestic debt) and its volatility $LCL σ. - Black–Scholes terms d_1 and d_2 defined by equations (10) and (11) (preserved in source form). - Volatility of $LCL formula shown (equation (12)) with component definitions: - DdLC σ = volatility of domestic debt in local currency; - FX σ = volatility of the 1-year forward exchange rate; - dF_DX ρ = correlation between domestic debt in local currency and 1-year forward exchange rate; - MLC σ = volatility of domestic debt in local currency (alternate notation in source); - LCF_MX ρ = correlation between domestic debt in local currency and 1-year forward exchange rate; - d_MD ρ = correlation between domestic debt in foreign currency and base money in foreign currency. - Indicators that can be estimated: 1. 1-year and 5-years distance to distress (Φ_t^1), giving number of standard deviations of sovereign asset value away from foreign debt distress barrier (equation (15) preserved in source form). 2. 1-year and 5-year risk neutral default probabilities (t RNDP) (equation (16)): RNDP_t = 1 − N(d_2). 3. 1-year and 5-year credit default spreads (t_spread) (equation (17)) preserved in source form. - Data requirements for application: - For each debt security: (i) coupon; (ii) amount outstanding; (iii) maturity; (iv) current market price; (v) market (external or domestic); (vi) currency; (vii) coupon type (fixed, floating, or indexed to inflation); (viii) coupon frequency. - Time series on foreign and domestic interest rates and foreign exchange rates. *Source: APPENDIX I: SELECT SALM METHODOLOGIES (IMF working paper content).* ### APPENDIX II. TECHNICAL ASPECTS OF SALM MODELING: SOME EXAMPLES ### APPENDIX II. TECHNICAL ASPECTS OF SALM MODELING: SOME EXAMPLES ### Overview - Purpose: Modeling interest rates, exchange rates, and other macroeconomic variables is important for estimating a sovereign’s risk exposure because the net worth of the country is driven by evolutions of these variables. - Risk indicators: Generate risk indicators from stochastic simulations driven by changes in interest rates, exchange rates, and other macroeconomic variables; indicators are determined by generated probability distributions and specific confidence levels to estimate gains and losses originating from market movements with a certain probability. ### A. Interest Rate Modeling - Commonly used term structure models: - One-factor CIR (Cox, Ingersoll, and Ross, 1985) model and extensions (CIR2). - NS (Nelson and Siegel, 1987) model. - HW (Hull and White, 1994a and 1994b) model. - HJM (Heath, Jarrow, and Morton, 1992) model. - DK (Duffie and Kan, 1996) model. - Market usage examples: CIR used by Denmark, Canada, Brazil; NS extension by Sweden; HJM by Hungary. - Cox, Ingersoll, and Ross (CIR) model: - Instantaneous interest rate follows (equation numbering retained): - (18) )()()]([)(tdztrdttrktdr - Definitions: - )(tr is interest rate at instant t - k is mean-reversion speed parameter -  is long-term interest rate average -  is interest rate volatility - )(tz is Brownian motion - Conditional mean and variance of r(t) given Fs (s