## Box 1: Why Scale Reserves by Gross External Liabilities?

## Source details

**Canonical URL:** [Box 1: Why Scale Reserves by Gross External Liabilities?](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08192.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08192.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08192.pdf.json)

---

### I. Introduction and motivation
- Examines whether foreign exchange reserves in emerging Asia have grown beyond what is needed to support financial stability.
- Reserves provide liquidity for temporary shortfalls in exports or capital inflows and help avoid disruptive changes in the exchange rate, investment, and consumption.
- Reserves can protect the domestic banking system and domestic credit markets from outflows of domestic or external resources (Obstfeld, Shambaugh, Taylor (2007)).
- Rapid reserve accumulation in Asia parallels the sharp expansion of trade and capital flows and increased volatility of gross capital flows, raising the disruptive potential of sudden stops.
- Build-up of reserves also reflects heightened risk aversion following the 1997–99 crises and the use of large current account surpluses and capital inflows to build buffers.

### II. Why scale reserves by gross external liabilities (conceptual rationale)
- Traditional indicators (reserves to imports, reserves to short-term debt) have limitations given the strong expansion of non-debt cross-border flows.
- The ratio of reserves to gross external liabilities is preferred for assessing vulnerability because of the growth in gross capital flows—including portfolio and direct investment—that increase gross external liabilities and the potential for disruptive outflows.
- Financial deepening and development of local bond and equity markets increase the domestic financial system stakes in case of outflows, reinforcing the argument for scaling by broader external liabilities.
- It is not suggested that reserves should fully cover total gross liabilities (the Greenspan-Guidotti 100% rule). Appropriate reserve coverage should vary by liability type and by volatility of the flow/liability.

### III. Developments in reserve adequacy indicators (key findings and statistics)
- Foreign exchange reserves of Asian economies have quadrupled since the end of the 1997–99 financial crisis.
- Excluding China, reserves more than doubled in nominal terms during 2000–07.
- Relative to GDP, reserves increased by about 10 percentage points over 2000–07, to 36 percent of GDP (excluding China).
- Emerging Asia’s reserves reached the equivalent of about 5 percent of global GDP (in nominal terms) at end 2007.
- The ratio of reserves to imports (excluding China) has declined modestly since 2003 but remains higher than the traditional benchmark of three months of imports.
- The ratio of reserves to external debt coming due within the next year exceeds the recommended 100 percent under the Greenspan-Guidotti rule by a wide margin in all economies, though these very high ratios may overstate insurance against sudden stops given increases in portfolio and direct investment flows.
- Reserves currently cover less than one-third of external liabilities in emerging Asia (excluding China).
- On average, reserves cover about one-third of broad money in emerging Asia; this ratio is above the Wijnholds and Kapteyn (2001) recommended holdings of 5–20 percent.
- In 2005, reserves covered about 15 percent of financial sector deposits, bonds and equities (World Bank database series). For most economies the ratio was around 10 percent; Singapore and Taiwan Province of China were in the 30–40 percent range.
- Financial deepening has outpaced reserve growth in emerging Asia (excluding China) since 2002; the ratio of reserves to financial system deposits, equities, and bonds declined for all economies in 2002–05.
- Volatility measures: gross capital flow volatilities increased in the 2001–2006 period compared with 1987–1996 in many Asia-Pacific subgroups; net inflow volatility decreased, indicating better synchronization of gross inflows and outflows, but the persistence of that trend is uncertain.

### IV. Insurance model of optimal reserves (framework and drivers)
- Reserves provide liquidity to cushion sudden stops in capital flows, at an opportunity cost equal to the return difference between capital and reserves.
- Optimal reserves derive from minimizing opportunity cost of reserves plus expected welfare cost of a crisis; formal loss function stated as: Loss(R) = δR + π f(.)
  - δ is the opportunity cost of reserves.
  - R is reserve holdings.
  - π is the probability of a crisis or sudden stop.
  - f(.) is the welfare cost of a crisis, increasing in the size of the sudden stop (L) and the output loss (ΔY).
- Under constant risk aversion (σ) and exogenous crisis probability, the optimal reserves expression given in the source (formula notation preserved) implies optimal reserves are larger when:
  - Size and output cost of a crisis are greater.
  - Probability of a sudden stop is higher.
  - Opportunity cost of holding reserves is lower.
  - Degree of risk aversion is higher.

### V. Key parameter estimates and calibration choices
- Output loss benchmark:
  - The Asian crisis (1997–99) used to estimate output loss.
  - Cumulative output loss for the six most affected Asian economies was 19 percent of GDP on average.
  - Indonesia and Thailand cumulative cost around 30 percent of GDP.
  - Example estimates: contemporaneous output loss in Indonesia 22 percent of GDP; total loss including beyond crisis 42 percent of GDP.
  - Paper assumes a potential output loss of 19 percent of GDP in the benchmark calibration (higher than Jeanne (2007) benchmark of 10 percent of GDP).
- Probability and size of a sudden stop:
  - Average probability of crisis set to 10 percent (unconditional frequency of sudden stops in sample of emerging economies, 1975–2003).
  - Average historical size of capital outflows estimated around 10 percent of GDP; used in calibrations except where specific country ratios applied.
  - Note: in some economies gross external liabilities exceed 700 percent (Hong Kong SAR) and 400 percent (Singapore) of GDP.
- Opportunity cost of holding reserves — three measures discussed:
  1. Sovereign risk premium — baseline uses interest spread on foreign debt. Table 3: Interest Rate Spreads (basis points), average for 2007:
     - China 71
     - Hong Kong SAR -44
     - India 332
     - Indonesia 197
     - Korea 73
     - Malaysia 87
     - Philippines 178
     - Singapore -175
     - Taiwan Province of China -230
     - Thailand -4
  2. Term premium (U.S. term premium in 2007 was less than 1/4 percent).
  3. Fiscal cost of sterilizing reserves (domestic financing rate paid by central bank minus yield on foreign reserves). Table 4 country estimates (yield on foreign reserve holdings assumed equal to one-year U.S. Treasury note = 3.34 percent at end-2007):
     - China: Domestic Financing Costs 3.24 — Net Carry 0.10 — Carry Income on Total Reserves (% GDP) 0.05
     - India: Domestic Financing Costs 6.00 — Net Carry -2.66 — Carry Income on Total Reserves (% GDP) -0.7
     - Indonesia: Domestic Financing Costs 8.00 — Net Carry -4.66 — Carry Income on Total Reserves (% GDP) -0.6
     - Korea: Domestic Financing Costs 5.00 — Net Carry -1.66 — Carry Income on Total Reserves (% GDP) -0.5
     - Malaysia: Domestic Financing Costs 3.50 — Net Carry -0.16 — Carry Income on Total Reserves (% GDP) -0.1
     - Philippines: Domestic Financing Costs 5.25 — Net Carry -1.91 — Carry Income on Total Reserves (% GDP) -0.4
     - Singapore: Domestic Financing Costs 0.98 — Net Carry 2.36 — Carry Income on Total Reserves (% GDP) 2.5
     - Taiwan Province of China: Domestic Financing Costs 3.38 — Net Carry -0.04 — Carry Income on Total Reserves (% GDP) 0.0
     - Hong Kong SAR: Domestic Financing Costs 5.75 — Net Carry -2.41 — Carry Income on Total Reserves (% GDP) -1.8
     - Thailand: Domestic Financing Costs 3.25 — Net Carry 0.09 — Carry Income on Total Reserves (% GDP) 0.0

### VI. Calibration results and main findings
- The insurance model explains a large share of observed reserves in many emerging Asian economies:
  - Model explains about 80 percent of current reserves in Thailand.
  - Model explains about 70 percent of total reserves in India.
  - Model explains about 60 percent of reserves in Taiwan Province of China.
  - Current reserve levels in Hong Kong SAR, Korea, the Philippines, and Singapore appear at or very close to optimal.
- Notable exceptions:
  - Model explains about half of reserve accumulation in China.
  - Reserves in Malaysia at end-2007 exceeded predicted optimal levels by more than twice.
  - Reserves in Indonesia appear lower than model-predicted optimal levels, suggesting scope for further accumulation on insurance grounds.
- Timing of overshooting (selected entries from reported table "Foreign Reserves Levels: Optimal Vs. Actual (in percent of GDP)"):
  - China: 26 47 40 2004
  - Hong Kong SAR: 66 75 70 2001
  - India: 16 22 20 2003
  - Indonesia: 17 14 11 ...
  - Korea: 26 28 27 2004
  - Malaysia: 25 57 53 1998
  - Philippines: 21 22 17 ...
  - Singapore: 10 29 81 03 ...
  - Taiwan POC: 43 71 73 2002
  - Thailand: 29 37 32 2005
- Overall interpretation:
  - A simple insurance model, with higher assumed output loss (19 percent of GDP) and modest opportunity costs of reserves, performs relatively well in explaining reserve holdings in emerging Asia except for certain outliers (China, Malaysia).
  - Results differ from Jeanne and Rancière (2006) and Jeanne (2007) largely due to different parameter choices (notably larger output loss and lower opportunity cost).
  - Most economies in Asia were closer to optimal levels in 2003 than in 2007; breaches of optimal levels mainly occurred between 2002 and 2005 in many cases.
  - Evidence of some reversal in reserves where levels had overshot (e.g., Taiwan Province of China, Korea) and the emergence of sovereign wealth funds (SWFs) are consistent with slowing accumulation recommendations.

### VII. Spreads–reserves threshold model (methodology and empirical results)
- Methodology:
  - Sample: panel of 34 emerging economies for 1997–2006.
  - Approach: threshold estimation as in Hansen (1996, 2000) to detect non-linear relation between J.P. Morgan’s EMBI spreads and reserve ratio indicators (SR).
  - Controls: 10-year U.S Treasury rate; CSFB High Yield spread; country GDP growth rate; ratio of debt to GDP. Variables in logs and lagged one period; country-specific fixed effects included.
  - Sequential threshold regressions used to detect one or two thresholds γ.
- General empirical conclusions:
  - Holding reserves has a significant impact in reducing spreads and lowering interest rate bills.
  - This effect continues to be important even at relatively high levels of reserves.
  - Estimated thresholds beyond which there are no gains in reduced borrowing costs are significantly above standard rules of thumb and closer to the optimal reserve levels found earlier.
  - The insurance motive plus the spreads-reduction benefit helps explain observed reserve levels in emerging Asia and other emerging economies.

### VIII. Threshold estimates and spreads–reserves elasticities (selected numeric estimates preserved exactly)
- Reserves to GDP:
  - Threshold = 49 percent of GDP.
  - Elasticity below threshold = 0.43 (reported as 43 percent).
- Reserves to months of imports:
  - Threshold = 6.3 months of imports.
  - Elasticity below threshold = 0.33 (reported as 33 percent).
- Reserves to broad money:
  - Threshold = 28 percent.
  - Elasticity below threshold = 0.46 (reported as 46 percent).
- Reserves to short-term debt:
  - Thresholds = 125 and 534 percent of short-term debt.
  - Elasticity in intermediate regime = 0.31 (reported as 31 percent).
- Reserves to total foreign liabilities:
  - Thresholds = 12 and 52 percent.
  - Elasticity in intermediate regime = 0.42 (reported as 42 percent).
- Reserves to financial system deposits and equities:
  - Thresholds = 13 and 30 percent.
  - Elasticity in intermediate regime = 0.61 (reported as 61 percent).

### IX. Robustness, caveats, and precision
- Confidence intervals for some threshold point estimates are large, implying uncertainty about true values.
- New reserve adequacy indicators (e.g., reserves to gross foreign liabilities) yield estimates with higher precision than traditional indicators.
- The spreads–reserves exercise analyzes impact on sovereign debt service costs and thus provides a lower bound on benefits (it does not incorporate private sector borrowing-cost benefits).
- Results are sensitive to choice of model parameters; in some cases regional averages were used for consistency or due to data limitations.

### X. Policy implications and conclusions
- The precautionary motive largely explains Asia’s large reserve accumulation since the 1997–98 crisis.
- Current reserve holdings in most of Asia (excluding China) are not excessive when compared with model-predicted optimal levels.
- Reserves provide insurance benefits that can more than compensate for opportunity costs of holding liquid assets.
- As foreign liabilities and capital flows grow, reserves may continue to rise in nominal terms, but optimal-response predictions suggest stabilization or moderate decline in reserve ratios going forward.
- The paper casts doubt on the continued use of traditional rules of thumb (3 months of imports, 100 percent of short-term external debt, 20 percent of broad money) for assessing reserve adequacy:
  - Country-specific optimal ratios are frequently well above these benchmarks.
  - Traditional indicators fail to capture vulnerabilities associated with increased financial globalization and large capital flows.
  - Measuring reserves against total foreign liabilities better captures the precautionary aspect.
  - Country-tailored reserve adequacy indicators are preferable given heterogeneity in exposure and opportunity costs across economies.
- The analysis cannot fully explain China’s large stock of reserves; motivations beyond precautionary and spread-reduction benefits remain to be explored in future research.
- Reserve accumulation has positive region-wide spillovers by reducing external vulnerabilities and potentially lowering regional contagion risk.
- Implication for sovereign wealth funds (SWFs): largest funds in Asia are in China, Singapore, and Malaysia—consistent with results that these economies have reached or exceeded recommended reserve levels. Rapid accumulation could lead to large SWF growth if excess reserves are directed to SWFs; impacts on global economy and financial stability warrant future research.

*Source: _wp08192 - Box 1: Why Scale Reserves by Gross External Liabilities?*

### Box 1: Why Scale Reserves by Gross External Liabilities?......................................................7

### Box 1: Why Scale Reserves by Gross External Liabilities?

### I. Introduction and motivation
- Paper examines whether foreign exchange reserves in emerging Asia have grown beyond what is needed to support financial stability.
- Reserves are held to provide liquidity in case of temporary shortfalls in exports or capital inflows and to avoid disruptive changes in the exchange rate, investment, and consumption.
- Reserves can protect the domestic banking system and domestic credit markets from outflows of domestic or external resources (Obstfeld, Shambaugh, Taylor (2007)).
- Rapid reserve accumulation in Asia parallels the sharp expansion of trade and capital flows and increased volatility of gross capital flows, raising the disruptive potential of sudden stops.
- The build-up of reserves in Asia also reflects heightened risk aversion following the 1997–99 crises and the use of large current account surpluses and capital inflows to build buffers.

### II. Why scale reserves by gross external liabilities (conceptual rationale)
- Traditional indicators (reserves to imports, reserves to short-term debt) have limitations given the strong expansion of non-debt cross-border flows.
- The ratio of reserves to gross external liabilities is preferred for assessing vulnerability because of the growth in gross capital flows—including portfolio and direct investment—that increase gross external liabilities and the potential for disruptive outflows.
- Financial deepening and development of local bond and equity markets increase the domestic financial system stakes in case of outflows, reinforcing the argument for scaling by broader external liabilities.

### III. Developments in reserve adequacy indicators (key findings and statistics)
- Foreign exchange reserves of Asian economies have quadrupled since the end of the 1997–99 financial crisis.
- Excluding China, reserves more than doubled in nominal terms during 2000–07.
- Relative to GDP, reserves increased by about 10 percentage points over 2000–07, to 36 percent of GDP (excluding China).
- Emerging Asia’s reserves reached the equivalent of about 5 percent of global GDP (in nominal terms) at end 2007.
- The ratio of reserves to imports (excluding China) has declined modestly since 2003 but remains higher than the traditional benchmark of three months of imports.
- The ratio of reserves to external debt coming due within the next year exceeds the recommended 100 percent under the Greenspan-Guidotti rule by a wide margin in all economies (Figure 4), though these very high ratios may overstate insurance against sudden stops given increases in portfolio and direct investment flows.
- Reserves currently cover less than one-third of external liabilities in emerging Asia (excluding China) (Figure 5).
- On average, reserves cover about one-third of broad money in emerging Asia; this ratio is above the Wijnholds and Kapteyn (2001) recommended holdings of 5–20 percent.
- In 2005, reserves covered about 15 percent of financial sector deposits, bonds and equities (World Bank database series). For most economies the ratio was around 10 percent; Singapore and Taiwan Province of China were in the 30–40 percent range.
- Financial deepening has outpaced reserve growth in emerging Asia (excluding China) since 2002; the ratio of reserves to financial system deposits, equities, and bonds declined for all economies in 2002–05.
- Volatility measures: gross capital flow volatilities increased in the 2001–2006 period compared with 1987–1996 in many Asia-Pacific subgroups (Table 1); net inflow volatility decreased, indicating better synchronization of gross inflows and outflows, but the persistence of that trend is uncertain.

### IV. Empirical assessment and modeling
- Much, though not all, of the reserve increase in Asia can be explained by an optimal insurance model under which reserves provide liquidity to cushion the impact of a sudden stop in capital inflows on output and consumption (Jeanne, 2007, model adapted for Asian conditions).
- The insurance model accounts for the opportunity cost of holding reserves but does not incorporate the impact of reserves in lowering spreads on an economy’s privately held external debt.
- Using threshold estimation techniques, the analysis shows that most of Asia can still benefit from higher reserves in terms of reducing borrowing costs (reducing spreads).

### V. Policy implications and interpretation
- Despite moves toward inflation targeting and floating exchange rates, Asian economies may still need substantial reserves because sudden stops in a financially integrated world can cause greater volatility, especially in relatively thin and institutionally less developed financial markets.
- Reserves in most Asian economies do not appear excessive against historical levels given the size of gross liabilities and the increased volatility of gross flows, although China may be a special case.
- Some convergence is evident: economies with very high reserve adequacy indicators have begun to see declines in those indicators as they move toward exchange rate flexibility and accumulate reserves more slowly; economies with relatively lower reserves continue to increase holdings.
- Traditional operational indicators (reserves to imports, reserves to short-term debt) remain in use in IMF staff reports but may understate vulnerability given increases in non-debt gross liabilities; scaling by gross external liabilities provides a broader measure of potential external drains.

*Source: _wp08192 - Box 1: Why Scale Reserves by Gross External Liabilities?*

### Box 1: Why Scale Reserves by Gross External Liabilities?

### Box 1: Why Scale Reserves by Gross External Liabilities?

### Rationale for scaling reserves by gross external liabilities
- The ratio of reserves to imports was developed to measure resilience to trade shocks that predominated before liberalization of financial systems and capital accounts.
- With increased cross-border capital flows and the rising possibility of sudden stops and capital outflows, the ratio of reserves to external debt maturing within a year became a key indicator of reserve adequacy.
- The ratio of reserves to short-term debt was highly suitable for assessing vulnerability to currency and maturity mismatches and was a good predictor of crisis in the 1990s.
- Given the evolution of capital flows and instruments, total gross external liabilities can better capture vulnerability to reversals than short-term debt alone, because:
  - The share of debt, including short-term debt, has decreased.
  - Portfolio flows have proved to be the most volatile form of capital flow; volatility of gross inflows and outflows has risen sharply since 2000.
  - Long-term liability holders can rapidly build hedges when balance of payments problems arise; total liabilities capture the potential for such behavior.
- It is not suggested that reserves should fully cover total gross liabilities (the Greenspan-Guidotti 100% rule). Appropriate reserve coverage should vary by liability type (e.g., FDI, portfolio equity, short-term debt) and by volatility of the flow/liability.

### Insurance model of optimal reserves (framework and drivers)
- Reserves provide liquidity to cushion sudden stops in capital flows, at an opportunity cost equal to the return difference between capital and reserves.
- Optimal reserves derive from a cost–benefit minimization: opportunity cost of reserves plus expected welfare cost of a crisis.
- Formal loss function stated as: Loss(R) = δR + π f(.)
  - δ is the opportunity cost of reserves.
  - R is reserve holdings.
  - π is the probability of a crisis or sudden stop.
  - f(.) is the welfare cost of a crisis, increasing in the size of the sudden stop (L) and the output loss (ΔY).
- Under constant risk aversion (σ) and exogenous crisis probability, the optimal reserves expression given (as in source):
  - ( ) 1/ 11 R L Y σ δ π − = + Δ − − +
- Optimal reserves are larger when:
  - Size and output cost of a crisis are greater.
  - Probability of a sudden stop is higher.
  - Opportunity cost of holding reserves is lower.
  - Degree of risk aversion is higher.

### Key parameter estimates and calibration choices
- Output loss benchmark:
  - The Asian crisis (1997–99) used to estimate output loss.
  - Cumulative output loss for the six most affected Asian economies was 19 percent of GDP on average.
  - Indonesia and Thailand cumulative cost around 30 percent of GDP.
  - Example estimates: contemporaneous output loss in Indonesia 22 percent of GDP; total loss including beyond crisis 42 percent of GDP (from Cerra and others (2005) cited in text).
  - Given these magnitudes, the paper assumes a potential output loss of 19 percent of GDP in the benchmark calibration (higher than Jeanne (2007) benchmark of 10 percent of GDP).
- Probability and size of a sudden stop:
  - Average probability of crisis set to 10 percent (unconditional frequency of sudden stops in sample of emerging economies, 1975–2003).
  - Average historical size of capital outflows estimated around 10 percent of GDP; used in calibrations except for:
    - Hong Kong SAR and Singapore (short-term liabilities significantly exceed 10 percent of GDP): actual ratio of short-term external debt to GDP used.
    - Indonesia (short-term debt estimated around 6.5 percent of GDP): actual ratio used.
  - Note: in some economies gross external liabilities exceed 700 percent (Hong Kong SAR) and 400 percent (Singapore) of GDP.
- Opportunity cost of holding reserves — three measures discussed:
  1. Sovereign risk premium (spread between private foreign borrowing costs and yields on reserve assets) — baseline scenario uses interest spread on foreign debt.
     - Table 3: Interest Rate Spreads (basis points), average for 2007:
       - China 71
       - Hong Kong SAR -44
       - India 332
       - Indonesia 197
       - Korea 73
       - Malaysia 87
       - Philippines 178
       - Singapore -175
       - Taiwan Province of China -230
       - Thailand -4
     - Note: EMBI spreads for Indonesia; spreads for others are ten-year government bond spreads for Malaysia, Philippines.
  2. Term premium (difference between long-term foreign rate and short-term foreign rate); U.S. term premium in 2007 was less than ¼ percent.
  3. Fiscal cost of sterilizing reserves (domestic financing rate paid by central bank minus yield on foreign reserves). Table 4 provides country estimates (as of December 2007), where the yield on foreign reserve holdings is assumed to be the one-year U.S. Treasury note (3.34 percent at the end of 2007).
     - Table 4 entries (Domestic Financing Costs (Sterilization Rate) — Net Carry (Interest on Foreign Reserves Minus Sterelization Rate) — Carry Income on Total Reserves (% GDP)):
       - China 3.24 0.10 0.05
       - India 6.00 -2.66 -0.7
       - Indonesia 8.00 -4.66 -0.6
       - Korea 5.00 -1.66 -0.5
       - Malaysia 3.50 -0.16 -0.1
       - Philippines 5.25 -1.91 -0.4
       - Singapore 0.98 2.36 2.5
       - Taiwan Province of China 3.38 -0.04 0.0
       - Hong Kong SAR 5.75 -2.41 -1.8
       - Thailand 3.25 0.09 0.0

### Calibration results and main findings
- The insurance model explains a large share of observed reserves in many emerging Asian economies:
  - Model explains about 80 percent of current reserves in Thailand.
  - Model explains about 70 percent of total reserves in India.
  - Model explains about 60 percent of reserves in Taiwan Province of China.
  - Current reserve levels in Hong Kong SAR, Korea, the Philippines, and Singapore appear at or very close to optimal.
- Notable exceptions:
  - Model explains about half of reserve accumulation in China.
  - Reserves in Malaysia at end-2007 exceeded predicted optimal levels by more than twice.
  - Reserves in Indonesia appear lower than model-predicted optimal levels, suggesting scope for further accumulation on insurance grounds.
- Timing of overshooting:
  - Table 5 reports year when actual reserves breached the optimal threshold (values in percent of GDP; year when optimal since):
    - China 26 47 40 2004
    - Hong Kong SAR 66 75 70 2001
    - India 16 22 20 2003
    - Indonesia 17 14 11 ...
    - Korea 26 28 27 2004
    - Malaysia 25 57 53 1998
    - Philippines 21 22 17 ...
    - Singapore 10 29 81 03 ...
    - Taiwan POC 43 71 73 2002
    - Thailand 29 37 32 2005
  - (Table presented in source as "Foreign Reserves Levels: Optimal Vs. Actual (in percent of GDP)")
- Overall interpretation:
  - A simple insurance model, with higher assumed output loss (19 percent of GDP) and modest opportunity costs of reserves, performs relatively well in explaining reserve holdings in emerging Asia except for certain outliers (China, Malaysia).
  - Results differ from Jeanne and Rancière (2006) and Jeanne (2007) largely due to different parameter choices (notably larger output loss and lower opportunity cost).
  - Most economies in Asia were closer to optimal levels in 2003 than in 2007; breaches of optimal levels mainly occurred between 2002 and 2005 in many cases.
  - Evidence of some reversal in reserves where levels had overshot (e.g., Taiwan Province of China, Korea) and the emergence of sovereign wealth funds (SWFs) are consistent with slowing accumulation recommendations.

### Concluding assessment
- The ratio of reserves to gross external liabilities appears to best capture vulnerability to sudden stops and capital account reversals, especially given growing complexity of capital market instruments and the potential for both short-term and long-term liability holders to contribute to rapid capital outflows.
- Different reserve coverage intensities should apply to different liability types, reflecting their liquidity, volatility, and rollover risk.

*Source: Box 1, "Why Scale Reserves by Gross External Liabilities?", extracted from the provided IMF content unit.*

### conclusions emerge from Figure 11. First, the optimal levels predicted by the model are

### _wp08192 - conclusions emerge from Figure 11. First, the optimal levels predicted by the model are

### Key findings on optimal reserves and rules of thumb
- The optimal levels predicted by the model are above the standard rules of thumb in most economies.
- The average optimal level of reserves for Asia is estimated at around 6 months of imports, twice as large as the traditional benchmark.
- Estimated optimal ratios for Indonesia, India, China, and Taiwan Province of China are above 6 months.
- With regards short-term external debt, all economies’ optimal reserve levels are above the 100 percent Greenspan-Guidotti rule. The average optimal level in Asia is more than three times this level.
- The average optimal level of reserves to broad money is around 30 percent, above the 5–20 percent range usually proposed in the literature.
- Implication: standardized rules of thumb may no longer be relevant; economy-specific indicators that account for economy-specific vulnerabilities and opportunity costs may be preferable.

### Observed reserve levels versus adequacy
- Actual reserve ratios are not substantially above adequate levels, except in economies discussed explicitly (e.g., China, Malaysia, Taiwan Province of China).
- Differences between actual and adequate levels are smaller when expressed in:
  - months of imports,
  - broad money,
  - short-term debt, or
  - gross external liabilities,
  than when presented in nominal terms or in percent of GDP.
- Much of the recent increase in reserves is explained by the precautionary motive and paralleled the sharp expansion of trade and capital flows and the increase in the volatility of gross capital flows.

### Threshold model of spreads–reserves elasticity (methodology)
- Sample: panel of 34 emerging economies for the period 1997–2006.
- Approach: threshold estimation as in Hansen (1996, 2000) to detect non-linear relation between J.P. Morgan’s EMBI spreads and reserve ratio indicators (SR).
- Model specification allows regression parameters to differ depending on reserve ratio R and endogenously determines threshold level(s) γ.
- Control variables: (i) two exogenous global factors: the 10-year U.S Treasury rate and CSFB High Yield spread; (ii) country GDP growth rate and ratio of debt to GDP. All variables estimated in logs and lagged one period; country-specific fixed effects included.
- Procedure: sequential threshold regressions—estimate first threshold and, if warranted and observations allow, test for a second threshold by splitting sample above first threshold.

### Empirical results — general conclusions
- Holding reserves has a significant impact in reducing spreads and lowering interest rate bills.
- This effect continues to be important even at relatively high levels of reserves.
- Estimated thresholds beyond which there are no gains in reduced borrowing costs are significantly above standard rules of thumb and closer to the optimal reserve levels found earlier.
- The insurance motive plus the spreads-reduction benefit helps explain observed reserve levels in emerging Asia and other emerging economies.

### Threshold estimates and spreads–reserves elasticities (by reserve indicator)
- Reserves to GDP:
  - Relevant threshold at 49 percent of GDP.
  - Below this level, elasticity of spreads w.r.t. reserves is 43 percent (a 1 percent increase in reserves/GDP → 0.43 percent decline in spreads).
  - Reserves in excess of 49 percent of GDP do not longer have an impact on spreads.
  - Most observations fall below this threshold (except Hong Kong SAR, Malaysia, Lebanon, and Algeria for some recent years).
- Reserves to months of imports:
  - Threshold above which reserves no longer reduce spreads is 6.3 months of imports (twice the traditional rule of thumb).
  - Spreads–reserves elasticity below 6 months of imports is 33 percent.
  - In 2006, reserves of Indonesia, Thailand, the Philippines, and Hong Kong SAR fell short of the estimated threshold.
- Reserves to broad money:
  - For reserves/broad money below 28 percent, elasticity is 46 percent.
  - Beyond 28 percent, no evidence that further accumulation reduces spreads.
  - Estimated threshold close to average optimal level for Asia (32 percent of broad money) and above current reserve levels of China, Indonesia, India, Hong Kong SAR, and Vietnam.
- Reserves to short-term debt:
  - Two significant thresholds estimated at 125 and 534 percent of short-term debt, both higher than the Guidotti rule.
  - Effect is negative for the intermediate regime (elasticity of 31 percent) and insignificant for the first and third regimes.
  - In Asia, only China’s reserves are above the second threshold.
- Reserves to total foreign liabilities:
  - First threshold at 12 percent; second threshold at 52 percent.
  - Intermediate range shows very negative and significant impact on spreads (elasticity of 42 percent).
  - Thresholds estimated with high precision (narrow 95 percent confidence intervals).
  - Under this indicator, only current levels in China and Taiwan Province of China cannot be explained by spread-reduction benefits.
- Reserves to financial system deposits and stock market capitalization:
  - Two thresholds at 13 percent and 30 percent.
  - Elasticity in the intermediate regime is 61 percent; not statistically significant in the other two regimes.
  - Thresholds estimated with great precision.

### Selected numeric threshold and elasticity estimates (preserved exactly as reported)
- Reserves to GDP: threshold = 49 percent of GDP; elasticity below threshold = 0.43 (reported as 43 percent).
- Reserves to months of imports: threshold = 6.3 months of imports; elasticity below threshold = 0.33 (reported as 33 percent).
- Reserves to broad money: threshold = 28 percent; elasticity below threshold = 0.46 (reported as 46 percent).
- Reserves to short-term debt: thresholds = 125 and 534 percent of short-term debt; elasticity in intermediate regime = 0.31 (reported as 31 percent).
- Reserves to total foreign liabilities: thresholds = 12 and 52 percent; elasticity in intermediate regime = 0.42 (reported as 42 percent).
- Reserves to financial system deposits and equities: thresholds = 13 and 30 percent; elasticity in intermediate regime = 0.61 (reported as 61 percent).

### Robustness, precision, and caveats
- Confidence intervals for some threshold point estimates are large, implying uncertainty about true values.
- New reserve adequacy indicators (e.g., reserves to gross foreign liabilities) yield estimates with higher precision than traditional indicators.
- The spreads–reserves exercise analyzes impact on sovereign debt service costs and thus provides a lower bound on benefits (it does not incorporate private sector borrowing-cost benefits).
- Results are sensitive to choice of model parameters; in some cases regional averages were used for consistency or due to data limitations.

### Policy implications and conclusions
- The precautionary motive largely explains Asia’s large reserve accumulation since the 1997–98 crisis.
- Current reserve holdings in most of Asia (excluding China) are not excessive when compared with model-predicted optimal levels.
- Reserves provide insurance benefits that can more than compensate for opportunity costs of holding liquid assets.
- As foreign liabilities and capital flows grow, reserves may continue to rise in nominal terms, but optimal-response predictions suggest stabilization or moderate decline in reserve ratios going forward.
- The paper casts doubt on the continued use of traditional rules of thumb (3 months of imports, 100 percent of short-term external debt, 20 percent of broad money) for assessing reserve adequacy:
  - Country-specific optimal ratios are frequently well above these benchmarks.
  - Traditional indicators fail to capture vulnerabilities associated with increased financial globalization and large capital flows.
  - Measuring reserves against total foreign liabilities better captures the precautionary aspect.
  - Country-tailored reserve adequacy indicators are preferable given heterogeneity in exposure and opportunity costs across economies.
- The analysis cannot fully explain China’s large stock of reserves; motivations beyond precautionary and spread-reduction benefits remain to be explored in future research.
- Reserve accumulation has positive region-wide spillovers by reducing external vulnerabilities and potentially lowering regional contagion risk.
- Implication for sovereign wealth funds (SWFs): largest funds in Asia are in China, Singapore, and Malaysia—consistent with results that these economies have reached or exceeded recommended reserve levels. Rapid accumulation could lead to large SWF growth if excess reserves are directed to SWFs; impacts on global economy and financial stability warrant future research.

*Source: IMF Working Paper content in the supplied PDF excerpt.*

### References

### References

### Reserve accumulation and optimal reserves
- Aizenman, J., and J. Lee, 2006, “Financial Versus Monetary Mercantilism: Long-Run View of the Large International Reserves Hoarding,” IMF Working Paper No. 06/280 (Washington: Internationanl Monetary Fund).
- Durdu, C.B., E. Mendoza, and M. Terrones, 2007, “Precautionary Demand for Foreign Assets in Sudden Stop Economies: An Assessment of the New Mercantilism,” IMF Working Paper No. 07/146 (Washington: International Monetary Fund).
- Jeanne, O., 2007, “International Reserves in Emerging Market Countries: Too Much of a Good Thing?” Brooking Papers on Economic Activity, 1:2007.
- Jeanne, O., and R. Rancière, 2006, “The Optimal Level of International Reserves for Emerging Market Countries: Formulas and Applications,” IMF Working Paper No. 06/229 (Washington: International Monetary Fund).
- Gonçalves, F., 2007, “The Optimal Level of Foreign Reserves in Financially Dollarized Economies: The Case of Uruguay,” IMF Working Paper No. 07/265 (Washington: International Monetary Fund).
- Wijnholds, J.O., A. Kapteyn, 2001, “Reserve Adequacy in Emerging Market Economies,” IMF Working Paper No. 01/143 (Washington: International Monetary Fund).
- Wyplosz, C., 2007, “The Foreign Exchange Reserve Buildup: Business as Usual?,” Paper presented at the Workshop on Debt, Finance and Emerging Issues in Financial Integration, at the Commonwealth Secretariat in London, March 2007.

### Costs, taxation, and fiscal aspects of reserves
- Hauner, D., 2005, “A Fiscal Tag for International Reserves,” IMF Working Paper No. 05/81 (Washington: International Monetary Fund).
- Levy Yeyati, E., 2006, “The Cost of Reserves,” Universidad Torcuato Di Tella, Business School Working Paper No. 2006–10.
- Rodrik, D., 2006, “The Social Cost of Foreign Exchange Reserves,” NBER Working Paper No. 11952.
- Summers, L.H., 2006, “Reflections on Global Account Imbalances and Emerging Markets Reserve Accumulation,” L. K. Jha Memorial Lecture. Reserve Bank of India, Mumbai, March 24.

### Vulnerability, liquidity, and crisis dynamics
- Bussière, M., and C. Mulder, 1999, “External Vulnerability in Emerging Market Economies: How High Liquidity Can Offset Weak Fundamentals and the Effects of Contagion,” IMF Working Paper No. 99/88 (Washington: International Monetary Fund).
- Caprio, G., and D. Klingebiel, 2003, “Episodes of Systemic and Borderline Financial Crises,” mimeo, World Bank.
- Cerra, V., and S. Saxena, 2005, “Did Output Recover from the Asian Crisis?, IMF Staff Papers, Vol. 52. No. 1 (Washington: International Monetary Fund).
- McCulley, P., and R. Toloui, 2007, “Perils of Plenty: Can Foreign Reserves Growth Forever?” Global Central Bank Focus, PIMCO.
- Obstfeld, M., J.C. Shambaugh, and A.M. Taylor, 2008, “Financial Stability, the Trilemma, and International Reserves,” CEPR Discussion Paper Series No. 6693.

### Analytical and data resources on external positions and reserves
- Edison, H., 2003, “Are Foreign Reserves Too High?” Staff Studies for the World Economic Outlook, September 2003 (Washington: International Monetary Fund).
- Green, R., and T. Torgerson, 2007, “Are High Foreign Exchange Reserves in Emerging Markets a Blessing or a Burden?” Department of the Treasury, Office of International Affairs, Occasional Paper No. 6.
- Lane, P. and G.M. Milesi-Ferretti, 2007, “The External Wealth of Nations Mark II: Revised and Extended Estimates of Foreign Assets and Liabilities,” Journal of International Economics, 73, 223–250.
- International Monetary Fund, 2007, Asia and Pacific Regional Economic Outlook, September 2007 (Washington).

*Source: _wp08192 - References*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08192.pdf_