## _061906

## Source details

**Canonical URL:** [_061906](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2006/_061906.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2006/_061906.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2006/_061906.pdf.json)

---

### Executive summary — overview and policy instruments
- Member countries face shocks that raise volatility in aggregate output and can cause crises, creating demand for "country insurance."
- Protective measures:
  - Self-insurance: sound policies, robust financial structures, adequate reserve coverage.
  - Regional risk-pooling arrangements.
  - Multilateral IMF financing (collective arrangements to be addressed in subsequent staff papers).
- Focus of this paper: mechanisms of self-insurance, with emphasis on national balance sheets (external liability structures and reserves). Collective arrangements left to later work.

### Nature and costs of output events and shocks
- Event identification:
  - Events start at first year of a decline in GDP per capita and end when GDP per capita returns to pre-event level.
  - Event filters: duration at least two years; total output loss at least 5 percent of pre-event GDP per capita.
- Empirical highlights:
  - Output drops are more frequent, long-lasting, and costly for emerging market and developing countries than for advanced economies.
  - For emerging markets and developing countries, output events start about every 16 years on average.
  - Event duration:
    - Emerging markets: events last for six years.
    - Developing countries: events last twice as long as in emerging markets.
  - Median cumulative output loss for concluded events:
    - Emerging markets: 15 percentage points of GDP.
    - Developing countries: 38 percentage points of GDP.
  - Output losses are two to three times larger in developing countries than in emerging markets.
  - Consumption falls more in developing countries than in emerging markets for a given output decline.
- Taxonomy of shocks analyzed:
  - Financial and macroeconomic: currency crises, banking crises, debt crises, sudden stops (defined as worsening in the financial account balance by more than 5 percentage points of GDP compared with the previous year).
  - Country-specific external: terms of trade and disasters.
  - Sociopolitical: wars and political turbulence.
  - Global: large increases in international interest rates and oil prices.
  - Boom-bust cycles: end of lending booms and growth booms.
- Expected-cost computation (insurance perspective):
  - Expected cost = unconditional probability × conditional probability of output event given the shock × median cost of the event when it occurs.
  - Key expected-cost findings (based on concluded events):
    - Emerging markets: largest expected cost from financial and macroeconomic shocks—especially sudden stops (¾ percent of GDP per capita annually based on concluded events) and currency crises.
    - Developing countries: terms-of-trade shocks most costly (2½ percent of GDP per capita annually when concluded and ongoing events are considered), followed by debt crises and global interest rate hikes.
  - Note: expected costs for multiple shocks are not additive due to overlap.

### Liability-side (self-insurance through sound fundamentals and liability structures)
- First line of defense: sound macroeconomic policies and supporting institutions to reduce risk and severity of currency, debt, and banking crises.
- Desirable liability-structure features:
  - Long-term, domestic currency debt to lower rollover and foreign exchange risk.
  - Equity-like liabilities (FDI, portfolio equity) to strengthen links between external payments and countries’ ability to pay.
- Underutilized private instruments for country insurance:
  - Catastrophe bonds and insurance against natural disasters for smaller, disaster-prone countries.
  - Commodity price futures or other hedges for commodity-concentrated producers.
  - GDP growth-indexed bonds (growth-indexed instruments; e.g., GDP-warrants) for broader membership.
- Considerations for financial innovation:
  - Larger economies provide "critical mass" for deep secondary markets.
  - Transparent institutions and statistics needed for instruments like growth-indexed bonds.
- Empirical findings on sudden stops and flows:
  - FDI has played essentially no role in financial flow reversals; portfolio equity limited role; portfolio debt prominent but recovers relatively quickly; bank lending and official flows drop sharply and remain depressed for years.
  - Volatility ranking (accounting for size of flows):
    - Low volatility: FDI and portfolio equity.
    - Medium volatility: portfolio debt and official flows.
    - High volatility: bank flows.
  - Cross-country evidence: equity-like liabilities share positively associated with educational attainment and institutional quality.
- Public debt management and policy recommendations:
  - Target debt-service correlation with ability to repay for insurance purposes.
  - Develop domestic debt markets to extend maturities and allow domestic-currency issuance.
  - Denominate international debt in domestic currency or index to real variables (e.g., economic growth).
  - Preconditions: credible fiscal and monetary policies, institutional quality, pension reforms, use of inflation-indexed bonds during transitions.
  - Growth-indexed bonds could provide insurance benefits but present implementation challenges.

### Liability-side developments in emerging markets (highlights)
- Move toward greater reliance on long-term, domestic-currency debt; domestic-currency bonds more attractive to non-residents.
- Examples of sovereign issuance in domestic currency on international markets and private global bonds denominated in or indexed to local currencies.
- Growth-indexed instruments observed in Argentina’s 2005 exchange (market capitalization estimated at US$5.8 billion in mid-April 2006; Argentina’s GDP-linked securities detached market capitalization US$2.9 billion on November 30, 2005 and US$5.8 billion by mid-April 2006).
- Aggregate assessment: liability-structure improvements reduce but do not eliminate need for other self-insurance forms.

### Asset-side self-insurance — international reserves: rationale and empirical role
- Rationale for holding reserves (insurance perspective):
  - Speed and ease of use during balance of payments pressure.
  - Support credibility of monetary and exchange regimes; help prevent currency crises.
  - Provide liquidity to foreign exchange and domestic financial markets; buffer domestic economy against shocks.
- Empirical role in sudden stops:
  - Typical sudden stop pattern: average net inflows of 5 percent of GDP (year prior) turn to net outflows of 4 percent of GDP (year of stop).
  - Accompanied by a large drop in reserves (by 4 percentage points of GDP).
  - Observed fall in domestic absorption held to 2½ percentage points of GDP versus counterfactual 6½ percentage points of GDP if reserves had been constant.
- Reserves trends:
  - Total international reserves more than tripled since 1990; Asian emerging economies’ share of global reserves increased by more than 20 percentage points between 1990 and 2005.
  - Data for 2005 refer to stocks at the end of the second quarter of 2005.
- Reserves and crisis probability:
  - Some studies find higher reserves relative to short-term debt associated with lower probability of currency crisis or sudden stop, though robustness and reverse causality concerns exist.

### Costs of holding reserves — measurement and regional evidence
- Opportunity cost typically measured as difference between return on short-term foreign currency assets and returns on more profitable investments.
- Decomposition of emerging-market bond yield into four components:
  1. Short-term rate on U.S. Treasuries;
  2. Term premium (long vs short U.S. rates);
  3. Default premium;
  4. Risk premium (correlation with global cycle).
- Two opportunity-cost variants used:
  a) U.S. term premium (baseline).
  b) Full difference between emerging-market yields and U.S. short-term Treasuries (term premium plus sovereign spread).
- Regional comparison (2001–05) — total cost of holding reserves (in percent of GDP) using the term premium:
  - Latin America: 0.2, 0.3, 0.4, 0.4, 0.2 (2001–2005)
  - Asia: 0.3, 0.7, 0.8, 0.8, 0.3 (2001–2005)
  - Others: 0.3, 0.6, 0.6, 0.7, 0.3 (2001–2005)
- Regional comparison (2001–05) — total cost using term premium plus sovereign spread (EMBIG):
  - Latin America: 1.0, 1.6, 1.4, 1.4, 0.8 (2001–2005)
  - Asia: 0.9, 1.1, 1.2, 1.2, 0.7 (2001–2005)
  - Others: 1.1, 1.5, 1.2, 1.0, 0.5 (2001–2005)
- Reserves/GDP by region (2001–2005):
  - Latin America: 12.1, 11.6, 12.5, 13.5, 14.0
  - Asia: 21.2, 22.6, 25.3, 28.9, 29.1
  - Others: 17.5, 20.0, 21.3, 21.8, 21.3
- Interpretation:
  - By term premium, cost of holding reserves substantially lower in Latin America than in Asia in 2001–05.
  - By term premium plus sovereign spread, costs relatively similar across regions because Latin America faces higher sovereign spreads while Asia holds higher reserves/GDP.

### Policy options to reduce reserve costs and state-contingent assets
- Options:
  - Invest reserves in longer-term, higher-yielding foreign assets.
  - Use favorable inflow periods to progress toward financial account liberalization via securitization of inflows (closed-end mutual funds issuing domestic-currency shares, purchasing foreign exchange, investing abroad).
  - Hold state-contingent assets that pay off when needed (e.g., instruments inversely related to main export prices; instruments that provide liquidity when international conditions tighten).
- VIX-based contingent reserves (Box 3 summary):
  - Proposal: invest part of reserves in call options on the VIX to deliver returns during sudden stops.
  - Empirical basis: probability of observing a statistically significant “jump” in the VIX conditional on a sudden stop ≈ 70 percent.
  - Claimed portfolio effect: optimal hedging portfolio (risk-free assets + VIX calls) could increase expected reserves during sudden stops by up to 40 percent vs risk-free assets only.
  - Constraints: small current VIX market size (notional value of open futures ≈ US$1.5 billion), uncertainty about persistence of VIX–sudden stop correlation, counterparty risk.

### Analytical framework for optimal reserves (model and calibration)
- Framework objective: trade off consumption-smoothing benefits of reserves against financial costs; reserves optimal level depends on crisis likelihood and country fundamentals.
- Model calibration (Jeanne and Rancière, 2005) — panel of 33 middle-income countries, 1980–2003:
  - Sudden-stop observations: 77 → unconditional probability π = 0.10 per year.
  - Mean size of sudden stops λ = 0.11 (11 percentage points of GDP).
  - Output cost γ = 0.06 (6 percent of GDP).
- Baseline parameters (selected):
  - Potential Output Growth g = 0.033
  - Term Premium δ = 0.015
  - Risk Free Rate r* = 0.05
  - Risk Aversion σ = 2 (range [1,10])
  - Size of Sudden Stop λ = 0.11 ; range [0,0.3]
  - Probability of Sudden Stop π = 0.10 ; range [0,0.25]
  - Output Loss γ = 0.06 ; range [0,0.2]
- Behavioral results:
  - Optimal reserves rise near-linearly with size of sudden stop beyond 2.5 percent of GDP.
  - Optimal reserves rise non-linearly with sudden-stop probability; example: increase in annual probability from 5 to 10 percent raises optimal reserves from 4 to 10 percentage points of GDP.
  - An increase in opportunity cost by 1.5 percentage points above baseline reduces optimal reserve ratio by more than 6 percentage points of GDP.
  - Higher risk aversion matters mainly at low σ; less impact at higher σ.
- Benchmark/comparative numbers:
  - Benchmark economy initial optimal reserves = 8.2 percent of GDP (about 3 months of imports; 105 percent of short-term debt).

### Comparative-statics and quantitative impacts of fundamentals (Table 4 results)
- Baseline initial optimal level: 8.2 percent of GDP.
- Effects of specified changes (estimated change in sudden stop probability; change in optimal reserves as percentage points of GDP; change in months of imports; change in reserves/short-term debt in percent):
  - Exchange Rate Overvaluation 0.0 → 20 (percent)
    - Sudden stop probability: 4.3 percentage points
    - Change in optimal reserves: 2.7 percentage points of GDP
    - Change in months of imports: 1.3 months
    - Change in reserves/short-term debt: 34.6 (percent)
  - Public Debt/GDP 40.0 → 60 (percent)
    - Sudden stop probability: 2.2 percentage points
    - Change in optimal reserves: 1.7 percentage points of GDP
    - Change in months of imports: 0.8 months
    - Change in reserves/short-term debt: 21.8 (percent)
  - Foreign Liabilities/Money 45 (Thailand, 1990) → 262 (Thailand, 1997)
    - Sudden stop probability: 7.0 percentage points
    - Change in optimal reserves: 3.7 percentage points of GDP
    - Change in months of imports: 1.8 months
    - Change in reserves/short-term debt: 47.4 (percent)
  - Financial Openness (|gross inflows|)/GDP 5.5 → 10.4 (percent)
    - Sudden stop probability: 5.4 percentage points
    - Change in optimal reserves: 3.2 percentage points of GDP
    - Change in months of imports: 1.5 months
    - Change in reserves/short-term debt: 41.0 (percent)
  - Exchange Rate Regime Floating → Fixed
    - Sudden stop probability: 5.3 percentage points
    - Change in optimal reserves: 3.4 percentage points of GDP
    - Change in months of imports: 1.6 months
    - Change in reserves/short-term debt: 43.5 (percent)
- Interpretation:
  - Plausible changes in fundamentals can increase optimal reserves substantially, in some cases by one third to one half of the initial optimal level.

### Trends in optimal reserves for country groups (1980–2003)
- Computation details:
  - Optimal reserves computed for each country-year for 33 middle-income countries (1980–2003); regional averages for Latin American and Asian emerging markets.
  - Risk aversion set to σ = 6 to match Asia 1991 mean reserves/GDP in baseline exercise.
- Asian emerging markets:
  - Model implies reserves should have declined between early and late 1980s (slowdown in financial flows).
  - Beginning early 1990s, model envisages rapid increase in optimal reserves due to rising international financial integration.
  - Post-1997 observed reserve buildup may have been excessive: to rationalize actual reserves increase 1997→2003, expected size of sudden stop or output cost would have had to more than double relative to 1990s averages.
  - Increasing risk aversion from 6 → 10 raises warranted reserves by less than 3 percentage points of GDP—insufficient to explain post-crisis buildup.
- Latin American emerging markets:
  - Warranted reserves well in excess of actual reserves in the 1980s.
  - Predicted reserves lower in early 1990s (improved fundamentals).
  - Rise over past decade aligns with heightened financial integration and increasing de facto dollarization.
  - Close match between model and data 1991–2003 notable given no individual-year Latin America reserve levels used in calibration.
  - Caveat: model does not account for potentially larger output losses for dollarized economies.

### Model of sudden-stop probability — determinants and probit estimation
- Determinants increasing sudden-stop probability:
  - Currency real appreciation (Real Effective Exchange Rate deviation from HP-trend: coefficient ≈ 1.52–1.85).
  - Public Debt/GDP (coefficients ≈ 0.72–0.81).
  - Financial openness (|Gross Inflows|/GDP: coefficients ≈ 9.11–11.15).
  - Ratio of foreign liabilities of financial sector to money in banking sector (coefficients ≈ 0.20–0.23).
- Determinants decreasing sudden-stop probability:
  - GDP growth (coefficients ≈ -1.25 to -1.87; significance varies).
- Fixed exchange rate regime effect not robust once fundamentals included.
- Probit estimation statistics (representative):
  - Observations across specifications: 537–707.
  - Pseudo R2 values: 0.06, 0.14, 0.17.
  - Significance notation used: * significant at 10%; ** significant at 5%; *** significant at 1%.

### Policy conclusions and prioritization
- Key measures and relevance (summary):
  - Sound macroeconomic and financial policies: All countries; short-term impact.
  - Improving institutional quality: All countries; medium term; expected to increase equity-like liabilities (FDI, portfolio equity).
  - Appropriate debt management: Emerging market countries; short-term impact; facilitates longer-term domestic-currency debt issuance.
  - Innovative financial instruments (growth-indexed bonds, commodity hedges, disaster insurance): Emerging markets and developing countries; immediate in principle but practical obstacles; may require international coordination.
  - Appropriate reserve holdings: Emerging markets and developing countries; especially important for dollarized and financially integrated economies; short-term impact.
  - Collective arrangements (high-access contingent financing instruments; IMF support for regional pooling): Emerging market countries; to be considered in subsequent papers.
- Overall finding: For emerging markets, financial shocks (sudden stops) are the most significant shock type; reserves as self-insurance and other measures (debt management, institutional reform, innovative instruments, regional/collective arrangements) play complementary roles.

*Source: Executive Summary and selected sections, _061906*

### Executive Summary ......................................................................................................

### Executive Summary

### Overview
- Member countries face a range of shocks that raise volatility in aggregate output and, in extreme cases, cause economic crises; this creates demand for "country insurance."
- Protective measures: self-insurance (sound policies, robust financial structures, adequate reserve coverage), regional risk-pooling arrangements, and multilateral IMF financing.
- This paper analyzes mechanisms of self-insurance, focusing on national balance sheets (external liability structures and reserves). Collective arrangements are left to subsequent papers.

### Nature and Costs of Shocks
- Output drops are used as the proxy for economic costs; events start at the first year of a decline in GDP per capita and end when GDP per capita returns to pre-event level.
- Event filters: duration must be at least two years; total output loss must be at least 5 percent of pre-event GDP per capita.
- Types of shocks and findings:
  - Output drops in developing countries are associated primarily with real shocks, notably terms of trade declines.
  - Financial shocks (such as "sudden stops") play a lead role in emerging market countries.
  - Wars and episodes of political turmoil, though relatively infrequent over the sample, are extremely costly when they occur—particularly for developing countries.
- Empirical features (Table 1 summary as presented in source):
  - Advanced economies / Emerging markets / Developing countries
  - Sample frequency: All1.96.56.7
  - Concluded1.53.02.3
  - Median duration: All4612
  - Concluded555
  - Median cumulative loss: All154189
  - Concluded131538
  - (in years) / (in percent of pre-event GDP per capita) / (in percent of country-years)
  - Sources: Maddison (2003) and staff calculations.
  - Notes: “All” events include concluded, on-going and sub-events. Concluded events are fully observed within the sample period whereas ongoing events had not ended by 2001 and the duration and loss for these events are calculated assuming that the events ended on 2002. Sub-events are new events that start before the previous event has ended.

### Sound Fundamentals and Liability Structures (Self-Insurance: Liabilities)
- First line of defense: sound macroeconomic policies and supporting institutions to reduce the risk and severity of currency, debt, and banking crises.
- Sound policies can:
  - Facilitate issuance of long-term, domestic currency debt, lowering rollover and foreign exchange risk.
  - Foster increases in the share of equity-like liabilities (such as FDI) in external liability structures, strengthening links between external payments and countries’ ability to pay.
- Underutilized private instruments that can provide country insurance:
  - Catastrophe bonds and insurance against natural disasters for smaller, disaster-prone countries.
  - Commodity price futures or other instruments to hedge commodity price fluctuations for countries with heavily concentrated production.
  - GDP growth-indexed bonds for a broader segment of the IMF membership.
- Considerations for financial innovation:
  - Larger economies more likely to provide the necessary "critical mass" for deep and liquid secondary markets.
  - Transparent institutions and statistics are especially relevant to overcome measurement challenges posed by growth-indexed bonds.

### Self-Insurance through International Reserves (Assets)
- Reserves are a flexible and reliable form of insurance against a wide variety of shocks; particularly desirable for emerging markets exposed to sudden stops in financial flows or large unhedgeable shocks.
- Optimal-reserve framework:
  - Trades off consumption-smoothing benefits of reserves against financial costs of holding them.
  - Takes into account costs of reserve holdings, consumption-smoothing benefits, and role of country fundamentals in crisis likelihood.
- Key findings from framework and regional assessment:
  - Reserve accumulation in Asian emerging markets since the early 1990s was initially commensurate with increasing insurance needs, but more recently appears to have been in excess of what could be justified on the basis of plausible changes in fundamentals.
  - Latin American emerging markets seemed underinsured in the early 1990s; their reserves are now assessed as providing a broadly appropriate degree of self-insurance given regional fundamentals.
- Framework components and empirical supplements (as organized in source):
  - Section IV.A: Trends in Reserve Accumulation, and Benefits and Costs of Reserve Holdings.
  - Section IV.B: A Framework for Assessing the Optimal Level of Reserves.
  - Section IV.C: Trends in Optimal Reserves for Country Groups.
  - Boxes referenced in source: The Asian Bond Funds; Argentina’s GDP-Linked Securities; State-Contingent Reserves Based on the VIX.

### Organization and Further Work
- Paper structure:
  - Section II: Nature and economic costs of shocks across membership segments.
  - Section III: Roles of sound fundamentals and liability structures (external capital structure, public debt management).
  - Section IV: Asset-side analysis and analytical framework for desirable level of self-insurance through official reserves.
  - Section V: Summary and conclusions.
- Subsequent staff papers will address collective insurance arrangements—regional reserve pooling arrangements and a possible new IMF instrument for high-access contingent financing for countries with strong policies and transparent reporting.

*Source: Executive Summary, _061906 - Executive Summary*

### 12.      Output events are more frequent, long-lasting, and costly for emerging market and

### _061906 - 12.      Output events are more frequent, long-lasting, and costly for emerging market and

### Key empirical findings on output events
- Output events are more frequent, long-lasting, and costly for emerging market and developing countries than for advanced economies (Table 1).
- On average, both emerging market and developing countries have output events starting about every 16 years (or approximately twice during the three decades considered in the sample period).
- Event duration:
  - Emerging markets: events last for six years.
  - Developing countries: events last twice as long as in emerging markets.
- Median cumulative output loss over the event (for concluded events):
  - Emerging markets: 15 percentage points of GDP.
  - Developing countries: 38 percentage points of GDP.
- For both emerging markets and developing countries, the frequency, duration, and especially median loss of concluded output events is significantly lower than for all events.
- Output losses are two to three times larger (for any type of event) in developing countries than in emerging markets.
- Relative impact on consumption: for a given output decline, consumption falls more in developing countries than in emerging markets, possibly reflecting liquidity constraints or more pronounced impacts on permanent income.

### Taxonomy of shocks associated with output drops
- Shocks analyzed include:
  - financial and macroeconomic—currency crises, banking crises, debt crises, and sudden stops in financial flows;
  - country-specific external—terms of trade and disasters;
  - sociopolitical—wars and political turbulence;
  - global—large increases in international interest rates and oil prices;
  - boom-bust cycles—the end of lending booms and growth booms.
- Two simplifying assumptions in the analysis:
  - The analysis does not address causality between shocks and output declines (it does not ask whether shocks cause declines in output rather than the other way around).
  - The analysis does not separate effects of individual shocks when events are associated with more than one shock (one third of output events is associated with more than one shock).
- Sudden stop definition used: a worsening in the financial account balance by more than 5 percentage points of GDP compared with the previous year.

### Expected cost of shocks (insurance perspective)
- Expected cost is computed from three inputs estimated on observed frequencies in 1970–2001:
  1. the probability of the shock (unconditional frequency);
  2. the conditional probability that the shock will lead to a loss in output;
  3. the output cost associated with the event (median cumulative output loss).
- Expected cost equals unconditional probability × conditional probability of output event given the shock × median cost of the event when it occurs.
- Key expected-cost findings (based on concluded events):
  - Emerging markets: largest expected cost is for financial and macroeconomic shocks—especially sudden stops (¾ percent of GDP per capita annually based on concluded events) and currency crises.
  - Developing countries: terms-of-trade shocks are the most costly (amounting to 2½ percent of GDP per capita annually when concluded and ongoing events are considered), followed by debt crises and global interest rate hikes.
- Note on aggregation: expected cost refers to the impact of one type of shock regardless of co-occurrence; expected costs of multiple shocks are not additive because of overlap when shocks occur simultaneously.
- Figure and table based results:
  - Table 2 reports unconditional frequency of shocks (in percent of country-years), frequency of output event conditional on shocks (in percent), and cumulative output loss conditional on shocks (in percent of pre-event GDP per capita). Missing values indicate fewer than five occurrences or fewer than five concluded events.
  - Figure 2 illustrates expected cost in percent of pre-event GDP per capita for various shocks, showing large differences between concluded-only and all events (including ongoing), especially for developing countries due to costly ongoing and very long-lasting events.
- Omitted segment: advanced countries are omitted from Figure 2 because the value of insurance for this segment appears to be very low.

### External capital structure and crisis dynamics
- Two arguments on why composition of external liabilities matters:
  - Equity-like finance (portfolio equity, FDI) has desirable cyclical properties: payments are lower when economic performance is worse, enabling risk sharing and stabilizing domestic consumption.
  - Some forms of flows behave more desirably during crises: sudden stops have often been driven by reversals in debt flows rather than equity-like flows; FDI is traditionally viewed as more stable than portfolio financial flows.
- Empirical results on sudden stops (33 episodes, annual data since 1980):
  - FDI has played essentially no role in financial flow reversals.
  - Portfolio equity plays a limited role in sudden stops.
  - Portfolio debt plays a more prominent role in sudden stops but recovers relatively quickly.
  - Bank lending flows and official flows experience severe drops and remain depressed for several years after sudden stops.
- Broader volatility and comovement evidence:
  - For emerging and developing countries, volatility ranking (accounting for size of flows) suggests:
    - Low volatility: FDI and portfolio equity.
    - Medium volatility: portfolio debt and official flows.
    - High volatility: bank flows.
  - Comovement across countries is not markedly higher for portfolio or bank flows than for FDI; differences in persistence, procyclicality, or responsiveness to G7 growth or US interest rates are not particularly striking.
- Determinants of liability structure:
  - Cross-country evidence suggests equity-like liabilities (FDI, especially portfolio equity) as share of total external liabilities (or GDP) are positively and significantly associated with indicators of educational attainment and institutional quality.
  - Institutional quality improvements can, over time, shift external liability structures toward more equity-like financing.

### Public debt management and debt-structure policy recommendations
- Insurance perspective for debt management: for a given cost of borrowing, debt managers should seek to induce high correlation between debt service costs and the borrower’s ability to repay.
- Policy measures to improve debt resilience:
  - Develop domestic debt markets to extend maturities and allow issuance of debt denominated in domestic currencies.
  - Denominate international debt in domestic currency or index it to real variables (e.g., economic growth).
- Preconditions and supporting reforms:
  - Credibility of fiscal and monetary policies is key for investor willingness to hold long-term local-currency bonds; credibility depends on institutional quality and reputation for sound policymaking.
  - Empirical observations: share of foreign currency denominated or indexed debt is far higher in Latin America than in Asia, reflecting different inflation histories.
  - Policy steps that have helped improve debt structures include:
    - Macroeconomic stabilization and institutional reforms (e.g., central bank independence, inflation targeting).
    - Pension reforms that widen the domestic investor base.
    - Use of inflation-indexed bonds during transition to lengthened maturity structures.
- Debt instruments with equity-like features:
  - Instruments that reduce payments in adverse shocks (e.g., growth-indexed bonds) could improve debt sustainability and international risk-sharing.
  - Growth-indexed bonds would likely provide substantial insurance benefits to a broad range of countries, though they present implementation challenges.

*Source: IMF staff analysis in the provided chapter text.*

### 25.      These considerations may have underpinned a number of developments in emerging

### _061906 - 25.      These considerations may have underpinned a number of developments in emerging

### Developments in emerging-market debt structures
- Greater reliance on long-term, domestic currency debt by a number of emerging market issuers.
- Domestic currency bonds have become more attractive for non-residents, owing in part to the development of local emerging market indices (IMF, 2005d, chapter 2).
- A number of emerging market sovereigns have issued domestic currency debt on international markets (examples cited: Uruguay in 2003 and 2004, Colombia in 2004, and Brazil in 2005). Private entities have issued global bonds denominated in or indexed to local currencies in Mexico and Brazil.
- The share of long-term, domestic currency debt issued on domestic markets has recently increased somewhat, with considerable variation across regions (Figure 5; source: Guscina and Jeanne (2006)).
- Increasing use of innovative financing instruments:
  - Growth-indexed instruments, notably GDP-warrants used in Argentina’s debt exchange of early 2005. The market capitalization of these growth-indexed instruments was estimated at US$5.8 billion in mid-April 2006 (Box 2).

### Aggregate assessment of liability-side evolution
- These developments represent a gradual move toward greater reliance on long-term domestic currency debt and the use of innovative forms of financing, such as growth-indexed instruments.
- Despite progress, countries’ liability structures are unlikely to have evolved sufficiently rapidly to obviate the need for other forms of self-insurance against sudden stops and other shocks.

### Asian Bond Market Initiative (Box 1)
- Objectives of the Asian Bond Funds (ABFs):
  - Invest Asian reserves in Asian bonds rather than European or U.S. bonds.
  - Provide a new regional channel of financial intermediation with reduced costs compared with intermediation through non-regional banks.
  - Act as a catalyst for private investors—especially regional institutional investors—to consider investment in Asian securities.
  - Promote development of regional bond markets and facilitate issuance of domestic currency bonds by participating countries.
  - Identify and remove market impediments such as capital controls, withholding taxes, and deficient clearing and settlement infrastructures.
- Institutional context: Initiated by the EMEAP group of 11 central banks (Australia, China, Hong Kong SAR, Indonesia, Japan, the Republic of Korea, Malaysia, New Zealand, the Philippines, Singapore, and Thailand). Complementary initiatives by ASEAN and APEC aim at regulatory and infrastructure improvements and increasing supply of local instruments.
- ABF implementation to date:
  - ABF1 (2003): $1bn supplied by EMEAP central banks to be invested in U.S. dollar denominated bonds; managed by the BIS.
  - ABF2 (2005): nine separate funds (one pan-Asian fund, PAIF, and eight local market funds) for a total of US$2bn to be invested in local currency denominated debt. These are open-ended exchange traded funds managed by private sector institutions and will gradually be opened to institutional and retail investors.

### Argentina’s GDP-linked securities (Box 2) — features and market outcomes
- Context: Included in Argentina’s global debt restructuring settled on June 10, 2005. Eleven new bonds issued, each with a detachable GDP-linked security. For each US$1 of defaulted debt tendered and accepted, creditors received new restructured bonds and one unit of GDP-linked security.
- Market capitalization and pricing developments:
  - After detachment and independent trading beginning November 30, 2005, total market capitalization amounted to US$2.9 billion.
  - Subsequent upward revisions to growth expectations led prices to approximately double—to 9.3 cents per dollar of notional value in mid-April 2006—with total market capitalization reaching US$5.8 billion.
- Payment and contract design:
  - Payments on the GDP-linked securities are contingent on Argentina’s economic performance: securities pay only if both the level and growth of GDP exceed a specified threshold in the relevant reference year, beginning in 2005 and ending in 2034.
  - The sum of payments over the life of the security cannot be higher than 48 percent of its notional value.
  - Payments are made on December 15 of each year following the relevant reference year.
  - The first payment of about 0.2 percent of GDP is expected on December 15, 2006.
- Valuation and market perception:
  - At the time of the exchange, the overall restructuring offer was priced by market participants in the range of 34-35 cents per dollar of principal claims.
  - Market analysts estimated the “theoretical value” of the GDP-linked securities to be about 4 points (cents per dollar of notional value) but recommended a conservative valuation at half that theoretical value due to novelty and other obstacles.

### Self-insurance through international reserves — trends
- Total international reserves have more than tripled since 1990, partly reflecting rapid deepening in trade and financial integration.
- The increase in reserves has been particularly pronounced in Asian emerging economies, whose share of global reserves increased by more than 20 percentage points between 1990 and 2005.
- Data references: International Financial Statistics, IMF. Data for 2005 refer to stocks at the end of the second quarter of 2005.

### Rationale for holding reserves — benefits (insurance perspective)
- Reserves provide:
  - Speed and ease of use during balance of payments pressure.
  - Support for credibility of monetary and exchange regimes, helping to ward off external pressures and currency crises.
  - Liquidity to the foreign exchange market, aiding the efficiency of floating regimes.
  - A buffer for the domestic economy against balance of payments shocks.
  - Liquidity to domestic financial markets and the banking sector, especially where significant dollarization exists.
- Empirical role in sudden stops:
  - Typical sudden stop: average net inflows of 5 percent of GDP (year prior) turn to net outflows of 4 percent of GDP (year of the stop).
  - Such a stop is accompanied by a large drop in reserves (by 4 percentage points of GDP).
  - This drop helps to hold the observed fall in domestic absorption to 2½ percentage points of GDP, much smaller than the counterfactual fall in domestic absorption (6½ percentage points of GDP) that would have been observed if reserves had been held constant during each episode.
- Reserves and crisis probability:
  - An increase in the ratio of reserves to short-term debt has been found in some studies to be associated with a lower probability of a currency crisis or sudden stop (Bussière and Mulder, 1999; Garcia and Soto, 2004), though robustness has been questioned (Berg and others, 2004) and concerns about reverse causality exist (Detragiache and Spilimbergo, 2001).

### Analytical framework and modeling
- A calibrated model for emerging market countries (Jeanne and Rancière, 2005) is used to gauge benefits and costs of international reserves.
- Model focus: an open economy hit by financial flow reversals that lead to falls in output. The optimal level of reserves equates the opportunity cost of holding foreign reserves with the marginal benefit from smoothing domestic consumption in the event of a sudden stop.

### Costs of holding reserves
- The cost is typically measured as the difference between the return on short-term foreign currency assets and returns on more profitable alternative investment opportunities.
- Debate on the appropriate definition of alternative investment opportunities (repayment of foreign debt vs. higher-yielding domestic investments or public infrastructure).
- Decomposition of the full yield on emerging-market bonds into four components:
  1. The short term rate on U.S. Treasuries;
  2. The term premium—the difference between long-term and short-term U.S. interest rates;
  3. The default premium;
  4. The risk premium due to possible correlation between the tendency for defaults and the global cycle or global asset prices.
- Considerations in including/excluding components:
  - Evidence suggests the risk premium may be small, but segmentation of investor classes could make it significant.
  - The treatment of the default premium matters: it could be excluded as a fair reflection of the probability of non-repayment, or included because authorities that do not intend to default will include it when computing reserve-holding costs.
- The opportunity cost of holding reserves is estimated using two variants:
  a) The U.S. term premium (difference between long and short rates excluding default and risk premia)—the baseline in the paper.
  b) The full difference between emerging market yields and U.S. short-term Treasuries.

*Source: Excerpts from IMF chapter/section _061906 (pages 17–24). *

### 36.      On average, the total cost of holding reserves was substantially lower in Latin

### _061906 - 36.      On average, the total cost of holding reserves was substantially lower in Latin

### Cost of Holding Reserves: Regional Comparison (2001–05)
- Using the term premium, the total cost of holding reserves (in percent of GDP) averaged:
  - Latin America: 0.2, 0.3, 0.4, 0.4, 0.2 (for 2001, 2002, 2003, 2004, 2005 respectively)
  - Asia: 0.3, 0.7, 0.8, 0.8, 0.3 (for 2001–2005)
  - Others: 0.3, 0.6, 0.6, 0.7, 0.3 (for 2001–2005)
- Using the term premium plus the sovereign spread (EMBIG spread), the total cost of holding reserves (in percent of GDP) averaged:
  - Latin America: 1.0, 1.6, 1.4, 1.4, 0.8 (for 2001–2005)
  - Asia: 0.9, 1.1, 1.2, 1.2, 0.7 (for 2001–2005)
  - Others: 1.1, 1.5, 1.2, 1.0, 0.5 (for 2001–2005)
- Reserves/GDP (in percent) by region (2001–2005):
  - Latin America: 12.1, 11.6, 12.5, 13.5, 14.0
  - Asia: 21.2, 22.6, 25.3, 28.9, 29.1
  - Others: 17.5, 20.0, 21.3, 21.8, 21.3
- Interpretation:
  - On average, total cost of holding reserves was substantially lower in Latin America than in Asia in 2001–05 when measured by the term premium (Latin America: between 0.2–0.4 percentage points of GDP; Asia: between 0.3–0.8 percentage points of GDP).
  - When using term premium plus sovereign spread, costs were relatively similar across regions (Latin America: between 1.0–1.6 percentage points of GDP; Asia: between 0.7–1.2 percentage points of GDP).
  - Explanation: Asian countries have reserves/GDP ratios about twice those of Latin American countries, but sovereign spreads are substantially higher in Latin America.
  - Trend note: The cost of holding reserves increased moderately for emerging market countries over the past few years, partially offset by a reduction in sovereign spreads.

### Policy Options to Reduce Costs of Reserves
- Invest reserves in longer-term, higher-yielding foreign assets to reduce the cost of holding reserves.
- Use favorable external circumstances (large financial inflows and reserve accumulation) to progress toward financial account liberalization by:
  - Securitizing a portion of financial inflows through closed-end mutual funds that issue shares in domestic currency, use proceeds to purchase foreign exchange from the central bank, and invest abroad.
  - Expected benefits listed:
    - Eliminate fiscal costs of sterilizing inflows.
    - Give domestic investors opportunities for international diversification.
    - Stimulate development of domestic financial markets.
    - Allow central banks to control timing and quantity of outflows.
- Hold state-contingent assets that provide liquidity when needed, e.g.:
  - Instruments with returns inversely related to the price of a country's main exports (for commodity exporters).
  - Instruments that provide liquidity when international liquidity conditions tighten.
- Box 3—VIX-based contingent reserves (summary):
  - Proposal: Central bank invests part of reserves in contingent securities (call options on the VIX) that deliver returns when sudden stops occur.
  - Empirical basis: Probability of observing a statistically significant “jump” in the VIX conditional on a sudden stop is about 70 percent.
  - Claimed portfolio effect: The optimal hedging portfolio (risk-free assets + VIX call options) increases expected reserves during sudden stops by as much as 40 percent compared with a portfolio of risk-free assets only.
  - Advantages: VIX is correlated with sudden stops and is independent of country policies (limits moral hazard).
  - Constraints and concerns:
    - Current size of the VIX market is relatively small (notional value of open futures contracts about US$1.5 billion; options over the counter).
    - Uncertainty whether VIX–sudden stop correlation will persist.
    - Counterparty risk if VIX option sellers are exposed to emerging market losses and cannot honor obligations in a sudden stop.

### Analytical Framework for Assessing Optimal Reserves
- Purpose: Develop an analytical framework where reserves function as self-insurance against vulnerabilities from changes in the financial account balance (sudden stops).
- Model features and calibration:
  - Main benefit: Reserves smooth domestic consumption during decreases in financial inflows and output.
  - Warranted level of reserves is an explicit function of:
    - Probability of a sudden stop.
    - Size of a sudden stop.
    - Output cost of a crisis.
    - Opportunity cost of reserves.
    - Authorities’ risk aversion vis-à-vis crisis episodes.
  - Probability of a sudden stop estimated via panel probit regressions relating sudden stops to country fundamentals: exchange rate regime, financial openness, level public debt, de facto dollarization, output growth, exchange rate overvaluation.
  - Model assumptions (as specified):
    - Size of sudden stop and output cost assumed constant and identical across countries (estimated as observed averages for the 33 countries in sample over 1980–2003).
    - Risk aversion assumed constant across countries and over time (set equal to six in baseline calibration to match mean level of reserves to GDP in Asia in 1991).
    - Opportunity cost of holding reserves is the term premium (varies over time but not across countries).
  - Comparative-statics exercise based on a benchmark economy (initial optimal level of reserves equivalent to 8.2 percent of GDP; about 3 months of imports; 105 percent of short-term debt).

### Quantitative Results: Impact of Changes in Fundamentals on Optimal Reserves (Table 4)
- Baseline note: The initial optimal level of reserves is equivalent to 8.2 percent of GDP (about 3 months of imports, or 105 percent of short-term debt).
- Effects of specified changes in fundamentals (reported as: estimated change in sudden stop probability; change in optimal ratio of reserves to GDP; change in optimal level of reserves in months of imports; change in optimal ratio of reserves to short-term debt):
  - Exchange Rate Overvaluation: 0.0 → 20 (percent)
    - Estimated change in sudden stop probability: 4.3 percentage points
    - Change in optimal ratio of reserves to GDP: 2.7 percentage points
    - Change in optimal level of reserves (in months of imports): 1.3 months
    - Change in optimal ratio of reserves to short-term debt: 34.6 (percent)
  - Public Debt / GDP: 40.0 → 60 (percent)
    - Estimated change in sudden stop probability: 2.2 percentage points
    - Change in optimal ratio of reserves to GDP: 1.7 percentage points
    - Change in optimal level of reserves (in months of imports): 0.8 months
    - Change in optimal ratio of reserves to short-term debt: 21.8 (percent)
  - Foreign Liabilities / Money: 45 (Thailand, 1990) → 262 (Thailand, 1997)
    - Estimated change in sudden stop probability: 7.0 percentage points
    - Change in optimal ratio of reserves to GDP: 3.7 percentage points
    - Change in optimal level of reserves (in months of imports): 1.8 months
    - Change in optimal ratio of reserves to short-term debt: 47.4 (percent)
  - Financial Openness defined as (|gross inflows|) / GDP: 5.5 → 10.4 (percent)
    - Estimated change in sudden stop probability: 5.4 percentage points
    - Change in optimal ratio of reserves to GDP: 3.2 percentage points
    - Change in optimal level of reserves (in months of imports): 1.5 months
    - Change in optimal ratio of reserves to short-term debt: 41.0 (percent)
  - Exchange Rate Regime: Floating → Fixed
    - Estimated change in sudden stop probability: 5.3 percentage points
    - Change in optimal ratio of reserves to GDP: 3.4 percentage points
    - Change in optimal level of reserves (in months of imports): 1.6 months
    - Change in optimal ratio of reserves to short-term debt: 43.5 (percent)
- Interpretation:
  - Plausible observed changes in fundamentals can increase optimal reserves substantially, in some cases by one third to one half of the initial optimal level of reserves.
  - Example magnitudes:
    - 20 percent exchange rate overvaluation → +2.7 percentage points of GDP in optimal reserves.
    - Public debt rising from 40 percent to 60 percent of GDP → +1.7 percentage points of GDP.
    - Large foreign liabilities buildup (Thailand 1990→1997) → +3.7 percentage points of GDP.
    - One standard deviation increase in financial openness → +3.2 percentage points of GDP.
    - Moving from floating to fixed exchange rate → +3.4 percentage points of GDP.

### Trends in Optimal Reserves for Country Groups (1980–2003)
- Computation details:
  - Optimal level computed for each country-year for 33 middle-income countries over 1980–2003.
  - Regional averages obtained for Latin American and Asian emerging markets.
  - Probability of sudden stop computed for each country-year using probit estimates.
  - Size of sudden stop set to realized mean in each region and decade.
  - Risk aversion set equal to six for matching Asia 1991 mean reserves/GDP.
- Asian emerging markets:
  - Model suggests reserves should have declined somewhat between early and late 1980s due to slowdown in financial flows after the early 1980s debt crisis.
  - Beginning early 1990s, model envisages rapid increase in optimal reserves owing to rising international financial integration.
  - Slight decline in optimal reserves following the Asian crisis attributed primarily to reduction in public debt and financial flows.
  - For the post-1997 buildup in reserves, the model suggests the buildup may have been excessive: to rationalize the increase in actual reserves between 1997 and 2003, the expected size of either the sudden stop or the output cost would have had to more than double relative to its average level observed in the 1990s.
  - Increasing risk aversion parameter from 6 to 10 would raise warranted reserves by less than 3 percentage points of GDP—insufficient to explain the post-crisis buildup in emerging Asia’s reserves.
- Latin American emerging markets:
  - Warranted reserves appear to have been well in excess of actual reserves in the 1980s.
  - Predicted reserves are lower in the early 1990s (improved fundamentals: public debt and economic growth).
  - Rise over the past decade aligns with heightened international financial integration and increasing de facto dollarization.
  - Close match between model and data in 1991–2003 is notable given that no individual-year level of reserves for Latin America was used in calibration.
  - Caveat: Model includes de facto dollarization among determinants of sudden-stop likelihood but does not account for potentially larger output losses for dollarized economies in a crisis.

### Conclusions and Policy Implications
- Countries face a wide variety of shocks with major, enduring output consequences; mechanisms to prevent excessive disruption can provide sizable benefits.
- Key policy measures and their relevance (summary from Table 5):
  - Sound macroeconomic and financial policies: All countries; impact in the short term.
  - Improving institutional quality: All countries; medium term; expected to increase share of equity-like liabilities (e.g., FDI, portfolio equity).
  - Appropriate debt management: Emerging market countries; short-term impact; sound policies may facilitate longer-term domestic-currency debt.
  - Innovative financial instruments (growth-indexed bonds, commodity hedges, disaster insurance): Emerging markets and developing countries; immediate in principle but practical obstacles; may require international coordination.
  - Appropriate reserve holdings: Emerging markets and developing countries; especially important for dollarized and financially integrated economies; impact in the short term.
  - Collective arrangements (high-access contingent financing instruments; IMF support for regional pooling): Emerging market countries; to be considered in subsequent papers.
- Overall finding: For emerging markets, financial shocks (sudden stops) are the most significant shock type; reserves as self-insurance and other measures (debt management, institutional reform, innovative instruments, regional/collective arrangements) play complementary roles.

*Source: _061906 - 36.      On average, the total cost of holding reserves was substantially lower in Latin*

### 50.      For all countries, the first line of defense against adverse shocks is the pursuit of

### _061906 - 50.      For all countries, the first line of defense against adverse shocks is the pursuit of

### Sound policies and liability-structure reforms
- First line of defense against adverse shocks: pursuit of sound policies.
- Fiscal policy recommendation: seek to improve sustainability, taking into account that sustainable debt levels seem to be lower in emerging and developing countries than in advanced countries.
- Debt management recommendation: increase the share of long-term, domestic currency debt in total debt.
- Liability-structure recommendation: increase the share of equity-like liabilities, such as FDI.
- Evidence: macroeconomic stability and supporting institutional reforms may facilitate the move toward safer liability structures.

### Financial instruments and country insurance
- Financial instruments that may provide country insurance:
  - Catastrophe bonds and insurance against natural disasters for smaller, disaster-prone countries.
  - Commodity price futures or other instruments aimed at hedging against commodity price fluctuations for countries with heavily concentrated production structures.
  - GDP growth-indexed bonds for a broader segment of the IMF’s membership.
- Market-structure consideration: new instruments have a greater chance to succeed with large issuers, which can provide the necessary “critical mass” for a deep and liquid secondary market.

### Official reserves and model-based assessment of reserve adequacy
- Maintaining appropriate cushions of official reserves is an essential part of countries’ self-insurance toolkit.
- Analytical framework: useful for clarifying trade-off between the costs of holding reserves and the benefits in smoothing external shocks.
- Empirical application: applied to data for the past two and a half decades for a sample of more than thirty middle income countries.
- Key findings for Asian emerging markets:
  - Rapid reserve accumulation since the early 1990s may now have resulted in a degree of self-insurance that is above what can be justified on the basis of plausible changes in fundamentals.
  - Increased international financial integration and the related potential for financial flow reversals are consistent with the rapid increase in reserves observed in earlier years.
  - For the period since the late 1990s, reserve accumulation seems to be excessive compared with what the model would estimate as appropriate self-insurance.
- Key findings for Latin America:
  - The model suggests a substantial degree of under-insurance at least until the early 1990s.
  - On average, reserves may have reached broadly adequate levels in recent years.
  - Open issue: the model may not fully account for larger output losses in the presence of de facto dollarization.

### Collective insurance arrangements
- Role: collective insurance arrangements are likely to continue playing a key role in cushioning countries from the impact of shocks.
- Forms:
  - Regional or other pooling arrangements such as the Chiang Mai initiative.
  - Global insurance through international financial institutions.
- Rationale: pooling risks makes collective arrangements potentially more cost-effective than self-insurance.
- Note: These issues are to be taken up in separate papers over the coming months.

### Annex — Data sources and definitions (selected)
- Sample period and coverage:
  - Per capita GDP data are purchasing power parity adjusted (1990 international Geary-Khamis dollars), drawn from Maddison (2003).
  - Sample period is limited to 1970–2001, yielding 4,882 country-year observations.
  - At the end of the sample period, the data cover 167 countries.
- Financial and macroeconomic shocks:
  - Currency crises identified using three conditions (as in Frankel and Rose, 1996):
    - (i) devaluation/depreciation of at least        25 percent cumulative over a 12-month period;
    - (ii) devaluation/depreciation rate by at least 10 percentage points greater than in the preceding 12 months;
    - (iii) a minimum of 3 years since last crisis.
  - Banking crisis dummy: value 1 if at least one of Kaminsky and Reinhart (1999), Vila (2000), Bell and Pain (2000), Caprio and Klingebiel (2003), and Demirguc-Kunt and Detragiache (2005) identifies the country-year as an outbreak of a banking crisis.
  - Debt crisis dummy: records a 1 if at least one of Detragiache and Spilimbergo (2001), Manasse and Roubini (2005) and Reinhart, Rogoff, and Savastano (2003) identifies the country-year as the beginning of a debt crisis.
  - Sudden stops definition: worsening in the financial account balance by more than 5 percentage points of GDP.
- Country-specific external shocks:
  - Terms of trade shocks: defined as a 10 percent worsening in the terms of trade for goods, based on WEO data.
  - Disaster dummy: takes the value 1 if the number of injured times 0.3 plus the number of killed is greater than 0.01 percent of the country’s total population; data drawn from EMDAT, published by CRED (www.em-dat.net).
- Socio-political shocks:
  - War dummy: data from the Correlates of War project; records a 1 in the first year of a war.
  - Political-system shocks: deterioration by 3 points or more in the Polity index published by the Polity IV project.
- Global shocks:
  - Global interest rate shock: takes the value 1 when the U.S. federal funds rate increases by more than 150 basis points in one year.
  - Oil price shocks: refer to the first year of these episodes (i.e., 1973, 1978, 1989, 1999), drawn from World Economic Outlook, Appendix 1.1, September 2004, International Monetary Fund.
- Boom-bust cycles:
  - Lending boom dates drawn from Gourinchas, Valdés and Landerretche (2001, Table A1); dummy takes value 1 in the year after a lending boom ends.
  - Growth boom defined as a three-year period with average growth exceeding by two standard deviations the country’s average growth rate estimated over the entire sample; dummy takes value 1 in the first year after such an episode.

*Source: _061906 - 50.      For all countries, the first line of defense against adverse shocks is the pursuit of*

### ANNEX TO SECTION III:  BEHAVIOR OF DIFFERENT TYPES OF FINANCIAL FLOW

### ANNEX TO SECTION III:  BEHAVIOR OF DIFFERENT TYPES OF FINANCIAL FLOW

### Data description
- Sample: 1970–2003, full sample includes 153 countries.
- Flows analyzed (net, in current U.S. dollars, normalized by GDP):
  - Financial account
  - Foreign direct investment (FDI)
  - Portfolio debt investment (PDI)
  - Portfolio equity investment (PEI)
  - Official flows (government and monetary authorities)
  - Bank flows
  - Other investment
- GDP sources: World Bank’s World Development Indicators database, supplemented with IMF’s World Economic Outlook database.
- Data cleaning: quality checks, dropping outliers and unusable observations.
- Country grouping:
  - Advanced, Emerging, Developing (see Table 6 for lists and classification rules).
  - Note: Korea classified as emerging for this empirical analysis.

### Summary statistics — average net flows (1970–2003)
- Cross-country median financial account balances by group (averages of capital flows, normalized by GDP):
  - Advanced: 1.34
  - Emerging: 2.12
  - Developing: 3.15
- FDI average net inflows:
  - Advanced: 0.07
  - Emerging: 1.30
  - Developing: 2.24
- Other component averages (all normalized by GDP):
  - PDI: Advanced 0.25; Emerging 0.21; Developing -0.07
  - PEI: Advanced -0.20; Emerging 0.15; Developing 0.00
  - Official flows: Advanced 0.01; Emerging 0.10; Developing 0.86
  - Bank flows: Advanced 0.41; Emerging 0.01; Developing -0.14
  - Other investment: Advanced 0.10; Emerging 0.04; Developing 0.57
- Robustness: results similar when considering 1990–2003 only.

### Volatility
- Standard deviation of financial account (cross-country median, percent of GDP, 1970–2003):
  - Advanced: 2.88
  - Emerging: 4.38
  - Developing: 4.55
- Standard deviation by component (medians):
  - FDI: Advanced 1.35; Emerging 1.50; Developing 2.11
  - PDI: Advanced 2.09; Emerging 1.20; Developing 0.92
  - PEI: Advanced 1.56; Emerging 0.63; Developing 0.33
  - Official flows: Advanced 0.80; Emerging 1.89; Developing 2.22
  - Bank flows: Advanced 2.31; Emerging 1.32; Developing 1.14
  - Other investment: Advanced 1.86; Emerging 1.95; Developing 1.90
- Coefficient of variation (median across countries):
  - Financial account: Advanced 1.34; Emerging 1.83; Developing 1.32
  - FDI: Advanced 2.22; Emerging 1.00; Developing 0.92
  - PDI: Advanced 2.28; Emerging 3.18; Developing 3.07
  - PEI: Advanced 3.25; Emerging 1.96; Developing 2.14
  - Official flows: Advanced 2.66; Emerging 3.16; Developing 2.39
  - Bank flows: Advanced 5.02; Emerging 7.22; Developing 3.62
  - Other investment: Advanced 5.55; Emerging 3.61; Developing 2.15
- Key volatility findings:
  - Financial flows are substantially more volatile in emerging and developing countries.
  - FDI is the least volatile flow (especially for advanced countries, except official flows) and has a coefficient of variation near 1 in emerging and developing countries.
  - PDI and PEI are 2 and 5 times more volatile in advanced countries than in developing countries, respectively.
  - For developing countries, official flows and FDI are among the most volatile types.

### Persistence
- AR(1) pooled estimates (persistence; medians reported):
  - Financial account AR(1):
    - Advanced: 0.68
    - Emerging: 0.52
    - Developing: 0.51
  - By flow, advanced countries: AR(1) coefficients range between 0.30–0.40 across flows.
  - Emerging markets:
    - FDI AR(1): 0.51 (most persistent)
    - PDI AR(1): 0.02 (least persistent)
  - Developing countries:
    - FDI AR(1): 0.35
    - PDI/PEI/Other: coefficients between 0.2 and 0.5
- Methodology notes:
  - AR(1) estimated using pooled fixed effects with first lag.
  - Alternative country-by-country AR(1) regressions yield similar results.

### Correlations with macro variables
- Correlations reported (medians across countries within each group):
  - Correlation with domestic growth (financial-account and components):
    - Financial account: Advanced 0.08; Emerging 0.23; Developing 0.15
    - FDI: Advanced -0.02; Emerging 0.10; Developing 0.17
    - PDI: Advanced -0.10; Emerging -0.03; Developing -0.08
    - PEI: Advanced 0.03; Emerging 0.09; Developing 0.03
    - Official flows: Advanced -0.02; Emerging -0.03; Developing -0.02
    - Bank flows: Advanced 0.17; Emerging 0.17; Developing 0.04
    - Other investment: Advanced -0.02; Emerging 0.20; Developing 0.11
  - Correlation with G7 growth:
    - Financial account: Advanced 0.03; Emerging -0.09; Developing 0.02
    - PEI shows notable positive correlations: Emerging 0.10; Developing 0.20
  - Correlation with US 1 year T-bill:
    - Financial account: Advanced 0.24; Emerging 0.08; Developing 0.07
    - FDI correlations with US T-bill:
      - Emerging: -0.29
      - Developing: -0.16
- Key correlation findings:
  - Financial flows are mildly procyclical in emerging market and developing countries.
  - In developing countries, FDI is most correlated with domestic growth (0.2).
  - U.S. interest rate correlates with inflows into advanced countries (0.2) but is virtually uncorrelated with emerging and developing countries’ financial accounts.
  - Different types of flows are generally uncorrelated or weakly negatively correlated with each other.

### Principal components analysis
- First principal component (share of total variance explained; medians):
  - Financial account:
    - Advanced: 0.30
    - Emerging: 0.24
    - Developing: 0.28
  - By flow, first principal component shares vary:
    - Advanced: FDI and PEI display largest common component.
    - Emerging: FDI has largest common component.
    - Developing: FDI, PDI, and PEI are roughly similar.
- Key finding: First principal component accounts for about 25–30 percent of variation for total financial flows across groups; no pronounced differences across flow types in relative importance of the common component.

---

### III. ANNEX TO SECTION IV: A MODEL OF OPTIMAL RESERVES

### Model overview (Jeanne-Rancière, 2005)
- Purpose: analyze optimal reserve accumulation in a growing open economy to reduce frequency/impact of sudden stops and smooth domestic consumption.
- Mechanism:
  - At sudden stop, authorities lend reserves to domestic agents cut off from private lenders, mitigating consumption drop.
  - Cost of holding reserves is the difference between return on reserves and the long-term interest rate, reducing non-crisis consumption.
  - Optimal reserves equate marginal costs and benefits of holding reserves.
- Calibration sample: panel of 33 middle-income countries, 1980–2003.
- Sudden stop definition: worsening in the financial account balance by more than 5 percent of GDP.
  - Observations: 77 sudden stops in sample → unconditional probability of sudden stop = 10 percent per year.
  - Mean size of sudden stops: 11 percentage points of GDP.
  - Output cost of a sudden stop: γ = 0.06 (6 percent of GDP), defined as difference in average real GDP growth between sudden-stop year and following year versus remaining years.

### Calibration parameters (Table 8)
- Baseline values and ranges:
  - Size of Sudden Stop λ = 0.11 ; range [0,0.3]
  - Probability of a Sudden Stop π = 0.10 ; range [0,0.25]
  - Output Loss γ = 0.06 ; range [0,0.2]
  - Potential Output Growth g = 0.033
  - Term Premium δ = 0.015 ; range [0.0025, 0.05]
  - Risk Free Rate r* = 0.05
  - Risk Aversion σ = 2 ; range [1,10]
- Cost of holding reserves: opportunity cost = term premium (benchmark δ = 1.5 percent), variants include sovereign spread.
- Assumption: benevolent authorities maximize representative domestic consumer welfare.
- Potential output growth set at 3.3 percent (mean real GDP growth in middle-income countries, 1980-2003, excluding sudden stop years).

### Simulation results and comparative rule
- Policy comparison: Jeanne-Rancière optimal reserves vs Greenspan-Guidotti rule (plots in Figure 9).
- Behavioral results:
  - When size or probability of sudden stop is sufficiently low, optimal reserves = 0 (costs exceed benefits).
  - Optimal reserves rise near-linearly with size of sudden stop beyond 2.5 percent of GDP.
  - Optimal reserves rise non-linearly with probability of a sudden stop; e.g., increasing annual probability from 5 to 10 percent raises optimal reserves from 4 to 10 percentage points of GDP.
  - Changes in opportunity cost of holding reserves have substantial impact:
    - An increase in opportunity cost by 1.5 percentage points above baseline reduces optimal reserve ratio by more than 6 percentage points of GDP.
  - Impact of risk aversion:
    - Major when starting from low σ values; limited when starting from higher σ values.

### Determinants of vulnerability to sudden stops
- Method: probit estimation of sudden-stop probability using panel of 33 middle-income countries, 1980–2003.
- Explanatory variables: averages of first and second lags (predetermined); selection via general-to-specific approach; results robust to time effects and fixed effects.
- Variables increasing probability of sudden stop:
  - Currency real appreciation (Real Effective Exchange Rate deviation from HP-trend coefficient in preferred specifications ≈ 1.52 to 1.85; significant)
  - Public Debt/GDP (coefficients around 0.72–0.81; significant)
  - Financial openness (| Gross Inflows | /GDP) (coefficients ~9.11–11.15; significant)
  - Ratio of foreign liabilities of the financial sector to money in the banking sector (coefficients ~0.20–0.23; significant)
- Variables decreasing probability:
  - GDP growth (negative coefficients ~-1.25 to -1.87; significance varies at 10%)
- Fixed exchange rate regime:
  - Associated with higher probability of sudden stop only if financial openness and exchange rate overvaluation are omitted; otherwise effect not robust.
- Interpretation:
  - Vulnerability rises with degree of international financial integration (financial openness and foreign liabilities relative to money).
  - Trade openness not significant once financial openness included.

### Probit estimation — key coefficients and statistics (Table 9)
- Representative coefficient estimates (columns vary by specification; reported values shown):
  - Real Effective Exchange Rate Deviation from HP-trend: 1.52; (z-statistics in parentheses, significance levels reported)
  - GDP Growth: -1.25 to -1.87
  - Public Debt/GDP: 0.72 to 0.81
  - Ratio of Foreign Liabilities to Money in the Banking Sector: 0.20 to 0.23
  - Financial Openness (| Gross Inflows | )/GDP: 9.11 to 11.15
  - Dummy for Fixed Exchange Rate Regime: 0.29 (in one specification)
- Observations across specifications: 537–707
- Pseudo R2 reported in specifications: 0.06, 0.14, 0.17 (varies by column)
- Significance notation:
  - * significant at 10%; ** significant at 5%; *** significant at 1%.

*Source: IMF staff calculations and analysis as presented in the Annex to Section III and Annex to Section IV of the source PDF.*

### REFERENCES

### _061906 - REFERENCES

### International reserves, reserve management, and precautionary vs. mercantilist views
- Aizenman, Joshua, Yeonho Lee, and Yeongseop Rhee, 2004, “International Reserves Management and Capital Mobility in a Volatile World: Policy Considerations and a Case Study of Korea,” NBER Working Paper No. 10534.
- Aizenman, Joshua, and Jaewoo Lee, 2005, “International Reserves:  Precautionary Vs. Mercantilist Views, Theory and Evidence,” IMF Working Paper 05/198.
- Flood, Robert, and Nancy Marion, 2002, “Holding International Reserves In an Era of High Capital Mobility,” Brookings Trade Forum 2001, No. 1, pp. 1–68.
- Garcia, Pablo S., and Claudio Soto, 2004, “Large Hoarding of International Reserves: Are They Worth It?” Working Papers Central Bank of Chile 299.
- Hauner, David, 2005, “A Fiscal Price Tag for International Reserves,” IMF Working Paper 05/81.
- Lee, Jaewoo, 2004, “Insurance Value of International Reserves: An Option Pricing Approach,” IMF Working Paper 04/175.
- Hviding, Ketil, Michael Nowak, and Luca Antonio Ricci, 2004, “Can Higher Reserves Help Reduce Exchange Rate Volatility?” IMF Working Paper 04/189.
- Dooley, Michael P., David Folkerts-Landau, and Peter Garber, 2004, “The Revived Bretton Woods System: The Effects of Periphery Intervention and Reserve Management on Interest Rates and Exchange Rates in Center Countries,” NBER Working Paper No. 10332.
- Caballero, Ricardo J., and Stavros Panageas, 2005, “A Quantitative Model of Sudden Stops and External Liquidity Management,” NBER Working Paper No. 11293.
- Caballero, Ricardo J., and Stavros Panageas, 2004, “Contingent Reserves Management: An Applied Framework,” Working Paper No. 04-32.

### Financial crises, banking crises, early warning, and systemic distress
- Bell, James, and Darren Pain, 2000, “Leading Indicator Models of Banking Crises—A Critical Review,” Financial Stability Review (Bank of England) Issue 9 (December) pp. 113–29.
- Berg, Andy, Eduardo Borenzstein, and Catherine Pattillo, 2004, “Assessing Early Warning Systems: How Have They Worked in Practice,“ IMF Working Paper 04/152.
- Demirguc-Kunt, Asli, and Enrica Detragiache, 2005, “Cross-Country Empirical Studies of Systemic Bank Distress: A Survey,” National Institute Economic Review, No. 192 (April) pp. 68–83.
- Detragiache, Enrica, and Antonio Spilimbergo, 2001, “Crises and Liquidity: Evidence and Interpretation,” IMF Working Paper 01/2.
- Caprio, Gerard, and Daniela Klingebiel, 2003, “Episodes of Systemic and Borderline Financial Crises,” World Bank Research Dataset.
- Kaminsky, Graciela, and Carmen Reinhart, 1999, “The Twin Crises: The Causes of Banking and Balance-of-Payments Problems,” American Economic Review, Vol. 89, No. 3, pp. 473–500.
- Vila, Anne, 2000, “Asset Price Crises and Banking Crises: Some Empirical Evidence,” Bank for International Settlements, BIS Conference Papers, No. 8 (March 2000).

### Capital flows, sudden stops, volatility, and financial globalization
- Broner, Fernando, and Roberto Rigobon, 2004, “Why are Capital Flows so much more Volatile in Emerging than in Developed Countries?” Economics Working Papers 862.
- Claessens, Stijn, Michael P. Dooley, and Andrew Warner, 1995, “Portfolio Capital Flows: Hot or Cold?” The World Bank Economic Review, Vol. 9. No. 1, pp. 153–74.
- Fernandez-Arias, Eduardo, 1996, “The New Wave of Private Capital Flows: Push or Pull?” Journal of Development Economics, Vol. 48, pp. 389–418.
- Prasad, Eswar S., Kenneth Rogoff, Shang-Jin Wei, and Ayhan Kose, 2003, Effects on Financial Globalization on Developing Countries: Some Empirical Evidence, IMF Occasional Paper No. 220.
- Albuquerque, Rui, Norman Loayza, and Luis Serven, 2005, “World Market Integration through the Lens of Foreign Direct Investors,” Journal of International Economics, Vol. 66, pp. 267–295.
- Gourinchas, Pierre-Olivier, Rodrigo Valdes and Oscar Landerretche, 2001, “Lending Booms: Latin America and the World,” Economia, Vol. 1, No. 2, pp. 47–99.
- Prasad, Eswar S., and Raghuram G. Rajan, 2005, “Controlled Capital Account Liberalization: A Proposal,” IMF Policy Discussion Paper No. 05/07.
- Broner/others and related work on volatility and sudden stops: Caballero and Panageas (2005, 2004).

### Sovereign debt structure, GDP-indexed bonds, and debt crises
- Borensztein, Eduardo, Marcos Chamon, Olivier Jeanne, Paolo Mauro, and Jeromin Zettelmeyer, 2004, Sovereign Debt Structure for Crisis Prevention, IMF Occasional Paper No. 237.
- Borensztein, Eduardo, and Paolo Mauro, 2004, “The Case for GDP-Indexed Bonds,” Economic Policy, Vol. 38, pp. 165–216.
- Manasse, Paolo, and Nouriel Roubini, 2005, “‘Rules of Thumb’ for Sovereign Debt Crises,” IMF Working Paper 05/42.
- Reinhart, Carmen, Kenneth Rogoff, and Miguel Savastano, 2003, “Debt Intolerance,” NBER Working Paper No. 9908.

### Exchange rate regimes, currency crashes, and lender-of-last-resort issues
- Frankel, Jeffrey A., and Andrew Rose, 1996, “Currency Crashes in Emerging Markets: An Empirical Treatment,” Journal of International Economics, Vol. 41, pp. 351–66.
- Mussa, Michael, Paul Masson, Alexander Swoboda, Esteban Jadresic, Paolo Mauro, and Andrew Berg, 2000, Exchange Rate Regimes in an Increasingly Integrated World Economy, IMF Occasional Paper No. 193.
- Rogoff, Kenneth, Aasim M. Husain, Ashoka Mody, Robin J. Brooks, and Nienke Oomes, 2004, Evolution and Performance of Exchange Rate Regimes, IMF Occasional Paper No. 229.
- Reinhart, Carmen, and Kenneth Rogoff, 2004, “The Modern History of Exchange Rate Arrangements: A Reinterpretation,” Quarterly Journal of Economics, Vol. 119, No. 1, pp. 1–48.
- Jeanne, Olivier, and Charles Wyplosz, 2003, “The International Lender of Last Resort: How Large Is Large Enough?” in Michael Dooley and Jeffrey Frankel, eds., Managing Currency Crises in Emerging Markets, University of Chicago Press, pp. 89–118.
- Goldstein, Morris, 2004, “Adjusting China's Exchange Rate Policies,” Working Paper No. 04-1.

### Institutions, governance, political economy, and development policy
- Kaufmann, Daniel, Aart Kraay, and Massimo Mastruzzi, 2003, ”Governance Matters III: Governance Indicators for 1996–2002,” World Bank Policy Research Working Paper No. 3106.
- Johnson, Simon, Jonathan D. Ostry, and Arvind Subramanian, 2006, “Levers for Growth: Policy Lessons from Earlier Bouts of Growth in Developing Countries,” Finance and Development, Vol. 43, (March), pp. 28–31.
- Rogoff, Kenneth, 1999, “International Institutions for Reducing Global Financial Instability,” Journal of Economic Perspectives, Vol. 13, pp. 21–42.
- Krueger, Anne, 2006, “A Response to Allan Meltzer,” Review of International Organizations, Vol. 1, No. 1, pp. 61–64.
- Meltzer, Allan H., 2005, “New Mandates for the IMF and the World Bank,” Cato Journal, Vol. 25, No.1, pp. 13–16.
- Rajan, Raghuram, 2006, “The Ebbing Spirit of Internationalism and the International Monetary Fund,” Krasnoff Lecture at the Stern School, New York University, March 8.

### Methodology, datasets, and measurement
- Guscina, Anastasia, and Olivier Jeanne, 2006, “The Structure of Government Debt in Emerging Market Countries: A New Data Set,” IMF Working Paper 06/98.
- Maddison, Angus, 2003, The World Economy—Historical Statistics, OECD (Paris). Data available at http://dx.doi.org/10.1787/456125276116.
- Marshall, Monty G., and Keith Jaggers, 2002, “Polity IV Project, Political Regime Characteristics and Transitions, 1800-2002—Dataset User’s Manual.”
- Ogaki, Masao, Jonathan D. Ostry, and Carmen Reinhart, 1996, “Saving Behavior in Low- and Middle-Income Developing Countries: A Comparison,” Staff Papers, International Monetary Fund, Vol. 43, No. 1, pp. 38–71.
- Bulíř, Aleš, and A. Javier Hamann, 2003, “Aid Volatility: An Empirical Assessment,” Staff Papers, International Monetary Fund, Vol. 50, No. 1, pp. 64–89.

### Regional cooperation, bond markets, and crisis prevention instruments
- Kuroda, Haruhiko, and Masahiro Kawai, 2003, “Strengthening Regional Financial Cooperation in East Asia,” PRI Discussion Paper No. 03A-10.
- Ma, Guonan, and Eli M. Remolona, 2005, “Opening Markets through a Regional Bond Fund: Lessons from ABF2,” BIS Quarterly Review, June, pp. 81–92.
- Cordella, Tito and Eduardo Levy-Yeyati, 2005, “A (New) Country Insurance Facility,” IMF Working Paper 05/23.
- Borensztein et al., 2004, Sovereign Debt Structure for Crisis Prevention, IMF Occasional Paper No. 237.

### IMF reports and working papers cited
- International Monetary Fund, 2005a, “The Managing Director’s Report on the Fund’s Medium-Term Strategy.”
- International Monetary Fund, 2005b, World Economic Outlook, Chapter 3, September.
- International Monetary Fund, 2005c, Global Financial Stability Report, September.
- International Monetary Fund, 2004, “Crisis Prevention and Precautionary Arrangements—Status Report.”
- International Monetary Fund, 2003, World Economic Outlook, September.
- Ostry, Jonathan D., and Jeromin Zettelmeyer, 2005, “Strengthening IMF Crisis Prevention,” IMF Working Paper 05/206.
- Chamon, Marcos, and Paolo Mauro, 2005, “Pricing Growth-Indexed Bonds,” IMF Working Paper 05/216.

*Source: _061906 - REFERENCES*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2006/_061906.pdf_