## _060316 - EXECUTIVE SUMMARY

## Source details

**Canonical URL:** [_060316 - EXECUTIVE SUMMARY](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2016/_060316.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2016/_060316.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2016/_060316.pdf.json)

---

### Purpose and scope
- Provides operational guidance to staff on reserve adequacy discussions in the IMF’s bilateral and multilateral surveillance.
- Based on the policy paper Assessing Reserve Adequacy—Specific Proposals and related Board discussion; draws on ARA 2011, ARA 2013, and ARA 2015.
- Aim: deeper and more consistent discussion of reserve issues in Article IV consultations and the External Sector Report; also relevant where reserve targets anchor Fund programs.
- Date: June 2, 2016. Approved by Siddharth Tiwari. Prepared by the Strategy, Policy and Review Department.

### Core questions addressed
- Expected coverage of reserve issues at Policy Note, mission, and Staff Report stages [Section II, Box 3].
- Which reserve adequacy tools best fit economies by financial maturity, economic flexibility, and market access [¶s 8, 9, Annex II].
- Reserve needs for mature markets and assessment approaches [Section III A.].
- Tailoring reserve adequacy discussions for emerging/deepening financial markets, including:
  - commodity-intensive economies [¶s 23–25],
  - countries with capital flow management measures (CFMs) [¶s 21, 22],
  - partially and fully dollarized economies [¶s 26–30] [Section III B.].
- Metrics for countries with limited access to capital markets [Section III C., Annex IV and VI].
- How potential drains on reserves should be covered [Box 1, ¶7].
- Measures of the cost of reserves for countries with and without market access [Section IV].

### Operational emphasis
- Provide regular, forward-looking discussion on reserve adequacy to identify emerging external vulnerabilities.
- Encourage authorities to compile and disseminate Reserves Data Template (RDT) data where not currently provided.
- Box 1 clarifies: Gross Reserves, Net Reserves, Metrics, Country classification and reserve adequacy tools.
  - Gross Reserves: follows IMF (2009, p. 111) definition; includes market value of financial derivatives including swaps and forward positions.
  - Net reserves: subtract predetermined short-term drains (on- and off-balance sheet) from official reserve position.
  - Metrics: measure potential FX liquidity needs in adverse circumstances; short-hand scenario analysis.
  - Country classification: tools selected by market access, market liquidity resilience, and economic flexibility.

---

### Guidance on Article IV and multilateral surveillance coverage
- Article IV consultation reports should:
  - Discuss authorities’ stated objectives for holding reserves (precautionary and non-precautionary).
  - Where warranted, discuss domestic and external cost of reserves.
  - Tailor depth, emphasis, and choice of methodologies to country circumstances.
  - Integrate reserve issues into general policy discussion.
- Multilateral surveillance reports (e.g., External Sector Report) present multilaterally consistent assessment of largest economies’ international reserve positions and policies.

---

### Principles for reserve adequacy discussions
- Precautionary
  - Assessment should reflect external liquidity vulnerabilities and tools appropriate to country characteristics (external liabilities, capital controls, economic flexibility, depth and resilience of financial markets), and other buffers.
  - Article IV staff reports should explain why a particular set of tools was chosen.
  - If Policy Note suggests substantial deviation from appropriate reserves (e.g., persistent large-scale intervention), conduct thorough assessment and discuss with authorities during mission.
  - Account for alternative external liquidity sources (e.g., swap lines) and be forward-looking (Box 4 exemplifies dynamic assessment).
- Non-precautionary
  - When reserves are well above precautionary levels, discuss non-precautionary motives (exchange rate regime features, actions to achieve inflation target, intergenerational savings in resource-rich economies, by-product of export-led strategies).
- Cost
  - Domestic financial and opportunity costs of reserves should be discussed for some countries in Article IV consultations.
  - For countries whose reserve policies may significantly influence the operation of the IMS (outward spillovers), Article IV and possibly multilateral surveillance should discuss external costs in line with the Integrated Surveillance Decision (ISD).
- FX Intervention
  - For countries with reserves below adequate levels, advise rebuilding reserves; borrowed foreign currency to hold as reserves is an option if borrowings have relatively long maturity.

---

### Country classification and tool selection

### Economy types and tool guidance
- Choice depends on extent of market access, resilience of market liquidity, and economic flexibility; correlated but not identical to advanced, emerging, low-income classifications.
- Three economy types organized for guidance:
  - Mature Markets (MMA) — Section III A.
  - Deepening Financial Markets (DFM) — Section III B.
  - Credit-Constrained Economies — Section III C.
- Staff should justify tool choice for each country.

### Currency unions
- Reserve needs can be assessed at consolidated union level depending on ability of common central bank to allocate liquidity.
- For unions able to issue a reserve currency (e.g., Euro area), consolidated considerations align with reserve currency issuers.
- For unions of EMs and LICs, describe reserve adequacy in line with those economy types.
- Financial architecture (absence of banking union, inefficient liquidity allocation) and synchronization of shocks can limit pooling and increase pooled reserve needs.

---

### Mature Market Economies (MMA): key considerations and scenario analysis
- MMAs without a reserve currency or standing swap line may need precautionary reserve buffers.
- Reserve needs relate to acute financial market stress and dysfunction; buffer should reflect:
  - potential FX funding needs of the financial system,
  - extent and availability of financial system buffers,
  - potential duration of market dysfunction.
- Past experiences suggest dysfunctional funding and FX markets typically last between two and three weeks, but sometimes extend longer (ARA 2013).
- Standing liquidity swap providers listed: U.S. Federal Reserve, Bank of Canada, Bank of England, Bank of Japan, ECB, Swiss National Bank.
- Providing reserves against financial sector risks can create moral hazard; supervisory and prudential policies are principal defenses; a fee could be charged by central bank.
- Due to modelling challenges and data gaps, no consensus on standardized reserve levels for MMAs; use scenario analysis in Article IV consultations:
  - Discuss authorities’ scenarios and staff complementary scenarios centered on market dysfunction and FX funding shortfalls, calibrated with historical market liquidity and turnover.

---

### Role of reserve adequacy metrics and starting points

### General guidance
- Metrics are a useful starting point but require judgment; do not apply mechanically.
- Traditional metrics: import coverage, reserves to broad money, reserves to short-term debt.
- Optimality models (e.g., Jeanne and Rancière (2006)) equate marginal benefits and costs of reserves.
- Discuss additional risks/drains (swap/forward transactions, encumbered reserves), availability of external buffers, and dynamic use of metrics.

### ARA EM metric
- Motivation: capture multiple channels of market pressure across export income loss, resident outflows, rollover risks, nonresident outflows.
- Coverage: potential loss of export income, risk of resident outflows (broad money), rollover risks (short-term debt), risk of nonresident equity and MLT debt outflows (other liabilities).
- Reporting: report alongside other relevant analysis in Article IV staff reports for applicable economies.
- Weights for the metric (Revised Weights and regime-specific weights reported exactly):
  - Revised Weights (in percent):
    - Short-term Debt: 30
    - Other Liabilities: 20
    - Broad money: 10
    - Exports: 10
  - Fixed regime:
    - Short-term Debt: 30
    - Other Liabilities: 15
    - Broad money: 5
    - Exports: 5
  - Floating regime:
    - Short-term Debt: 30
    - Other Liabilities: 15
    - Broad money: 5
    - Exports: 5
- Reserves in the range of 100-150 percent of the composite metric are considered broadly adequate for precautionary purposes.
- Weights on “other liabilities” raised by 5 percentage points in ARA (2015).

### Dynamic Reserve Assessment (Box 4)
- Forward-looking approach using projections of metric components (reserves, short-term debt, broad money, exports) from desks’ projections (e.g., WEO).
- Project portfolio equity and MLT debt from IIP data and WEO projections; construct projected ARA metric.
- Applications: examples for India and Thailand show relative positions typically stable in projection period; useful in Article IV surveillance.

---

### Country-specific considerations: CFMs, commodity intensity, dollarization

### Countries with Capital Flow Management Measures (CFMs)
- Empirical findings: CFMs tend to reduce resident outflows and probability of exchange market pressure events.
- Implication: adjust weight on broad money in ARA EM metric where CFMs affect resident outflows; present both standard and adjusted metrics.
- Empirical rule for identifying important CFMs:
  - Countries with at least two of the three standard capital control indicators (Chinn-Ito, Quinn and IMF share) with a value less than or equal to 0.25 (indices normalized between 0 and 1).
- Numerical guidance on broad money weights (in percent):
  - Broad money weights without and with CFMs:
    - Without CFMs: Fixed 10, Floating 5
    - With CFMs: Fixed 5, Floating 2.5
- Note: where countries maintain clear controls on exit of non-resident assets, weight on “other liabilities” could be halved as discussed in ARA 2013 and ARA 2015.

### Commodity-Intensive Economies
- Maintain an additional buffer against terms-of-trade shocks rather than mechanically adjust reserve adequacy metric.
- Buffer calculation: Buffer = (price gap) × (trade value); price gap = percent difference between current prices and specified point of forecast price distribution.
- Typical confidence interval: 68 percent (one standard deviation if Gaussian).
- Practical application: compute one-year ahead forecast at 68 percent confidence interval; percent gap versus current price; multiply percent gap by current-year dollar value of commodity exports or imports to obtain buffer.
- Present reserves relative to metric augmented by buffer and unaugmented metric; report nominal buffer size.
- Illustrative example: example price gaps and numbers cited (examples use WTI prices and dates in example methodology).

### Dollarized Economies
- Fully dollarized economies:
  - Need liquidity buffers in adopted foreign currency to support domestic financial institutions and as fiscal buffer.
  - Wiegand (2013) proposes one month of government spending as a fiscal reserve yardstick.
  - Use ARA EM components as a conservative starting point; assess availability of public and private buffers.
- Partially dollarized economies:
  - No strong empirical basis for systematically larger reserves; ARA EM is a useful starting point.
  - Article IV should explain whether privately held buffers or dollarization imply higher or lower reserve needs and which metric is most appropriate.
  - Consider risks from non-resident foreign exchange holdings and typical larger FX reserve requirements.

---

### Credit-Constrained Economies and the ARA-CC approach

### Context and motivation
- Credit-constrained economies (mostly LICs) vulnerable to exogenous current account shocks; limited access to capital markets makes reserves important.
- Traditional approaches useful (three-month import rule, 20 percent broad money, 100 percent short-term debt) but can miss country specificities.

### ARA-CC methodology and logic
- Adequate reserves derived by equating marginal benefit and marginal cost:
  - Marginal benefit: absorption-smoothing benefits estimated by two empirical regressions capturing (i) role of reserves in reducing crisis likelihood and (ii) impact of reserves in reducing crisis severity.
  - Marginal cost: proxied by external funding costs, sterilization cost, or marginal product of capital (all net of returns from holding reserves).
- Adequate level defined where additional marginal benefit equals marginal cost.
- Complementary metrics: ARA EM, short-term debt rule, broad money rule; small islands metric for natural disaster vulnerability.
- Dollarized LICs may need more reserves depending on degree and nature of dollarization.

---

### Measuring the cost of reserves (Section IV)

### Overview
- Reserves are costly self-insurance; opportunity cost is an important consideration when reserves are in or above adequate ranges.
- Guidance differs by country type.

### A. Countries with Market Access (MMA and DFM)
- Sterilization and yield-based opportunity costs are useful measures.
- Two main cost components:
  - Foregone return on alternative use/asset of local authorities.
  - Cost of issuing paper for sterilization less return on reserves.
- When reserves are inadequate or barely adequate:
  - Marginal cost of financing accumulation is most relevant (cost of borrowing reserves or sterilizing interventions).
  - Quasi-fiscal cost of sterilizing accumulation can be high, including anticipated exchange rate valuation losses.
- When reserves are ample:
  - Use “net financing” or opportunity cost: difference between local yield and return on reserve assets, adjusting yields for effect of higher reserves lowering yields.
- For market access economies with high externally issued FX debt:
  - Use external FX denominated yield as opportunity cost proxy (proxied by the EMBI spread less endogenous impact of higher marginal reserves lowering spreads).
- Yield adjustments should count anticipated FX depreciation and align with expected exchange rate path in Article IV consultations.

### B. Credit-Constrained Countries
- Opportunity cost approximated by market yield on sovereign bond or estimated marginal product of capital (MPK).
- Guidance:
  - If a sovereign bond issued in last five years, use the yield proxy.
  - If no recent data, use average market yield for peer LICs or cross-country currency interest swap.
  - If securities markets not heavily distorted, use return on government securities adjusted for exchange risk premium (compute premium using average historical REER depreciation where no forward markets).
  - Where unavailable, use MPK computed from output-to-capital evolution, with attention to measurement and data issues; regional or LIC averages can be used.

---

### Tools, templates, and resources (Section V and Annexes)
- Tools available at: http://www.imf.org/external/np/spr/ara/index.htm.
- Tools include:
  - Cross-country reserve data and underlying data for metrics.
  - EM ARA tools tailored for commodity-intensive economies and adjusted metrics for CFMs.
  - Main template for EMs: plots reserves against adequacy metrics, tracks monthly reserves relative to ARA metric, provides one-year ahead forecast of metric components.
  - Stand-alone tool for commodity buffer (Section III.B and Annex V).
- ARA-CC tool workflow:
  - (i) determine whether country is “credit constrained”;
  - (ii) identify economic classification and exchange rate regime;
  - (iii) compile data for empirical regressions;
  - (iv) choose proxy for cost;
  - (v) quantify net cost of holding reserves;
  - (vi) calibrate reserve adequacy.
- Examples of surveillance applications: India, Colombia, Russia, Croatia, South Africa, Sweden, Mozambique.

---

### Annex III — Commodity Buffer methodology (highlights)
- Commodity intensity thresholds referenced: exporters above 50 percent of goods exports; importers 20 percent (per ARA 2015).
- Commodity buffer formula: Buffer = (price gap) × (trade value), with price gap percent difference between current price and one-year ahead forecast at 68 percent confidence interval.
- Practical application instructions and a Commodity buffer template in Annex V; template requires current-year dollar value of commodity imports/exports and latest commodity price projections from IMF Research Department.

---

### Annex VI — Assessing Reserve Adequacy Tool for Credit-Constrained Economies (operational summary)

### Template features and workflow
- Automatically obtains data from desk macro-framework and WEO; includes all developing countries; user selects country.
- Designed sheets: “Home Sheet”, “Approach”, “Computation”, “Sensitivity Analysis”, background/input sheets.
- Key operational steps:
  - Step 1: Choose the country.
  - Step 2: Choose economic classification (resource rich / non-resource rich defaults from IMF 2012 paper), exchange rate regime (AREAER de facto), and country paths; staff may override.
  - Steps 3–6: Select data to estimate marginal benefits using template regression coefficients and country data; staff can override data codes and coefficients.
  - Step 7: Estimate cost of holding reserves (template provides ARA (2013) and ARA (2015) estimates; staff may pick proxy or enter custom costs).
- Results:
  - Adequate reserves calibrated in “Computation”; presented in “Home” sheet alongside traditional three-month rule.
  - Sensitivity Analysis sheet allows variable adjustments and scenario comparisons.

### Estimated coefficients for marginal benefits computations (Table AVI-1)
- Likelihood of a crisis — coefficients by classification (values reported exactly):
  - Reserve, months of imports:
    - Resource Rich: -0.093
    - Non Resource Rich: -0.090
    - Updated Resource Rich: -0.093
    - Small State: -0.093
    - Fragile: -0.090
    - Frontier: -0.090
  - Government balance, % of GDP:
    - Resource Rich: -0.031
    - Non Resource Rich: -0.032
    - Updated Resource Rich: -0.031
    - Small State: -0.035
    - Fragile: -0.032
    - Frontier: -0.032
  - CPIA:
    - Resource Rich: -0.314
    - Non Resource Rich: -0.309
    - Updated Resource Rich: -0.314
    - Small State: -0.246
    - Fragile: -0.309
    - Frontier: -0.309
  - Flexible exchange rate regime, 1 if flexible:
    - Resource Rich: -0.381
    - Non Resource Rich: -0.380
    - Updated Resource Rich: -0.381
    - Small State: -0.440
    - Fragile: -0.380
    - Frontier: -0.380
  - IMF program:
    - Resource Rich: -0.291
    - Non Resource Rich: -0.302
    - Updated Resource Rich: -0.291
    - Small State: -0.394
    - Fragile: -0.302
    - Frontier: -0.302
  - Constant:
    - Resource Rich: 0.922
    - Non Resource Rich: 0.865
    - Updated Resource Rich: 0.922
    - Small State: 0.782
    - Fragile: 0.865
    - Frontier: 0.865

- Magnitude of absorption drop — coefficients by classification (values reported exactly):
  - Reserves, months of imports, in logs:
    - Resource Rich: -2.257
    - Non Resource Rich: -2.240
    - Updated Resource Rich: -2.239
    - Small State: -1.945
    - Fragile: -1.945
    - Frontier: -1.945
  - Flexible exchange rate regime, 1 if flexible:
    - Resource Rich: -8.623
    - Non Resource Rich: -8.698
    - Updated Resource Rich: -8.611
    - Small State: -9.418
    - Fragile: -9.418
    - Frontier: -9.418
  - External demand growth, percent:
    - Resource Rich: -1.002
    - Non Resource Rich: -0.932
    - Updated Resource Rich: -0.833
    - Small State: -0.901
    - Fragile: -0.901
    - Frontier: -0.901
  - Terms of trade growth, percent:
    - Resource Rich: -0.086
    - Non Resource Rich: -0.084
    - Updated Resource Rich: 0.000
    - Small State: 0.000
    - Fragile: 0.000
    - Frontier: 0.000
  - Change in FDI to GDP, percentage point of GDP:
    - Resource Rich: -0.023
    - Non Resource Rich: -0.016
    - Updated Resource Rich: 0.000
    - Small State: 0.000
    - Fragile: 0.000
    - Frontier: 0.000
  - Change in aid to GDP, percentage point of GDP:
    - Resource Rich: 0.000
    - Non Resource Rich: 0.053
    - Updated Resource Rich: 0.000
    - Small State: 0.000
    - Fragile: 0.000
    - Frontier: 0.000
  - Fixed effects:
    - Resource Rich: 7.569
    - Non Resource Rich: 3.784
  - Terms of trade growth, percent*resource rich:
    - Resource Rich: 0.000
    - Non Resource Rich: 0.000
    - Updated Resource Rich: -0.181
    - Small State: 0.000
    - Fragile: 0.000
    - Frontier: 0.000
  - Resource Rich (interaction / indicator):
    - Resource Rich: 0.000
    - Non Resource Rich: 0.000
    - Updated Resource Rich: 5.297
    - Small State: 0.000
    - Fragile: 0.000
    - Frontier: 0.000

### Template customization and staff judgment
- Staff may override coefficients and data codes, choose cost proxies from ARA (2013) and ARA (2015), enter custom costs, and adjust classification and regime settings.
- Judgment required in interpreting calibrated adequate reserves; sensitivity and scenario analysis encouraged.

---

### Key operational recommendations (concise)
- Use a tailored mix of metrics and scenario analysis; do not apply any single metric mechanically.
- Present ARA EM metric alongside traditional metrics and any country-specific adjustments (CFMs, commodity buffers, dollarization).
- For MMAs, emphasize scenario analysis and financial-sector buffers; for DFMs use ARA EM and complementary metrics; for credit-constrained LICs apply ARA-CC with sensitivity analysis.
- Discuss both precautionary adequacy and non-precautionary motives; where reserves exceed precautionary needs, assess opportunity and fiscal costs.
- Report and discuss potential drains (swap/forward positions, encumbered reserves) and Net International Reserves (NIR) where relevant.
- Use available templates and tools, and encourage dissemination of RDT data.

*GUIDANCE NOTE ON RESERVE ADEQUACY, International Monetary Fund, Executive Summary (June 2, 2016).*

### EXECUTIVE SUMMARY

### _060316 - EXECUTIVE SUMMARY

### Purpose and scope of the guidance note
- Provides operational guidance to staff on reserve adequacy discussions in the IMF’s bilateral and multilateral surveillance.
- Based on the policy paper Assessing Reserve Adequacy—Specific Proposals and the related Board discussion, and draws on ARA 2011, ARA 2013, and ARA 2015.
- A key aim is to provide a deeper and more consistent discussion of reserve issues in IMF bilateral and multilateral surveillance, particularly in Article IV consultations and the External Sector Report.
- Guidance also relevant where reserves targets form an anchor in Fund programs.
- June 2, 2016.
- Approved by Siddharth Tiwari. Prepared by the Strategy, Policy and Review Department in consultation with other departments.

### Core questions the note addresses
- What is the expected coverage of reserve issues at different stages of the bilateral surveillance process (Policy Note, mission, and Staff Report) [Section II, Box 3]?
- Which reserve adequacy tools best fit different economies based on their financial maturity, economic flexibility, and market access [¶s 8, 9, Annex II]?
- What do possible reserve needs in mature markets relate to, and how can their adequacy be assessed [Section III A.]?
- How can reserve adequacy discussions for emerging and deepening financial markets [Section III. B.] be tailored and applied to better evaluate reserve levels in: (i) commodity-intensive economies [¶s 23–25]; (ii) countries with capital flow management measures (CFMs) [¶s 21, 22]; and (iii) partially and fully dollarized economies [¶s 26–30]?
- What reserve adequacy considerations hold for countries with limited access to capital markets? How can metrics for these economies be tailored to evaluate their reserve needs [Section III C., Annex IV and VI]?
- How should potential drains on reserves be covered [Box 1, ¶7]?
- What are the various measures of the cost of reserves for countries with and without market access [Section IV]?

### Guidance on Article IV and multilateral surveillance coverage
- Individual Article IV consultation reports should:
  - Discuss the authorities’ stated objectives for holding reserves, including for precautionary and non-precautionary purposes.
  - Where warranted, discuss the domestic and external cost of reserves.
  - Tailor depth and emphasis, and choice of methodologies, to country circumstances and aspects relevant for the country’s external stability and global stability.
  - Integrate reserve issues into the general policy discussion.
- Multilateral surveillance reports (e.g., the External Sector Report) present a multilaterally consistent assessment of the largest economies’ international reserve positions and policies.

### Principles for reserve adequacy discussions (overarching considerations)
- Precautionary:
  - Assessment should reflect external liquidity vulnerabilities and tools appropriate to the country’s characteristics (external liabilities, capital controls, economic flexibility, depth and resilience of financial markets), and other buffers.
  - Article IV staff reports should explain why a particular set of tools was chosen.
  - If a preliminary Policy Note suggests a substantial deviation from appropriate reserves—e.g., persistent large-scale intervention in one direction in the exchange market—there should be a thorough assessment and discussion with authorities during the mission.
  - Take into account alternative complementary sources of external liquidity (e.g., swap lines) and be forward-looking (Box 4 exemplifies a dynamic assessment).
- Non-precautionary:
  - When reserves are well above precautionary levels, staff should discuss non-precautionary motives (e.g., exchange rate regime features, actions to achieve an inflation target where other transmission channels are ineffective, intergenerational savings in resource-rich economies, by-product of export-led strategies).
  - Discussion helps understand authorities’ reserve holdings and implications for policy.
- Cost:
  - For some countries, domestic financial and opportunity costs of reserves should be discussed in Article IV consultations.
  - For countries whose reserve policies may significantly influence the operation of the IMS (outward spillovers), Article IV and possibly multilateral surveillance should discuss external costs in line with the Integrated Surveillance Decision (ISD).
- FX Intervention:
  - For countries with reserves below adequate levels, policy advice should encourage authorities to build reserves.
  - Rebuilding need not be through intervention; borrowed foreign currency to hold as reserves is an option, provided foreign currency borrowings have relatively long maturity.

### Country classification and tool selection
- Reserve assessment tools and considerations depend on:
  - Extent of market access, resilience of market liquidity, and economic flexibility.
  - These characteristics are broadly correlated with advanced, emerging, and low-income classifications but not identical; staff should justify tool choice in each case.
- The note organizes adequacy considerations around three economy types:
  - Mature Markets (MMA) — Section III A.
  - Deepening Financial Markets (DFM) — Section III B.
  - Credit-Constrained Economies — Section III C.
- Annexes provide metrics, country classification, commodity buffer calculation, ARA-CC application, and tailored tools:
  - Annex I: The IMF Composite EM Reserve Adequacy Metric and Other Metrics Relevant for Deepening Financial Markets.
  - Annex II: Country Classification.
  - Annex III: Commodity Intensive Economies and Calculation of the Commodity Buffer.
  - Annex IV: Application of the ARA-CC Approach to Credit-Constrained Economies.
  - Annex V: Assessing Reserve Adequacy Tool for EMs.
  - Annex VI: Assessing Reserve Adequacy Tool for Credit-Constrained Economies and LICs.

### Operational recommendations and resources
- Staff should provide a regular, forward-looking discussion on reserve adequacy to identify emerging external vulnerabilities; country authorities commonly assess reserve needs frequently as a useful exercise to test assumptions.
- Box 1 clarifies important concepts: Gross Reserves, Net Reserves, Metrics, Country classification and reserve adequacy tools.
  - Gross Reserves: follows IMF (2009, p. 111) definition; includes market value of financial derivatives including swaps and forward positions.
  - Net reserves: subtract predetermined short-term drains (on- and off-balance sheet) from official reserve position.
  - Metrics: measure potential FX liquidity needs in adverse circumstances; short-hand scenario analysis.
  - Country classification: tools selected by market access, market liquidity resilience, and economic flexibility.
- The note points to available reserve adequacy data and tools, including a template for countries with limited access to capital markets, and provides examples of application in Fund surveillance (Section V).
- Staff should encourage authorities to compile and disseminate Reserves Data Template (RDT) data where not currently provided.

*GUIDANCE NOTE ON RESERVE ADEQUACY, International Monetary Fund, Executive Summary (June 2, 2016).*

### 7. Significant short-term FX liabilities or other potential short-term drains on a central

### _060316 - 7. Significant short-term FX liabilities or other potential short-term drains on a central

### Short-term FX liabilities and other potential drains
- Significant short-term FX liabilities or other potential short-term drains on a central bank’s reserves warrant discussion when the impact could be material.
- Such drains limit the usability and availability of reserves buffers and need to be discussed with the authorities.
- Examples:
  - Provisions allowing commercial banks to meet reserve requirements in foreign exchange boost gross international reserves, but those reserves “may not be available for other balance-of-payments purposes, as they would fall if deposits were to decline.”
  - Central bank drains can reflect swap/forward positions and short-term FX liabilities to residents; these positions should be discussed relative to the level of gross reserves.
- Where relevant, Net International Reserves (NIR) (Box 1) could be compared with relevant reserve adequacy metrics.

### Composition and liquidity considerations
- The composition of international reserves can substantially affect a central bank’s reserves during stress periods, mainly through:
  - liquidity of reserve assets, and
  - valuation effects from fluctuations in exchange rates.
- Sovereign assets not directly required for liquidity purposes are generally more appropriately managed through longer-term Sovereign Wealth Funds (SWFs).
  - To be counted as official reserve assets, SWF assets would need to be both readily available to, and controlled by, the monetary authorities to meet balance of payments needs (IMF 2009).

### Guidance on coverage of reserve adequacy issues in surveillance (Box 3)
- Depth, nature and emphasis of reserve adequacy issues should depend on country circumstances.
- Bilateral surveillance (Article IV consultations) — staff should highlight before the mission (Policy Note (PN)):
  - A preliminary discussion of precautionary reserve adequacy using tools appropriate to the country; choice of tools depends on external vulnerabilities, depth of financial markets, reliance on capital controls, commodity intensity, and other structural characteristics.
  - For emerging and deepening financial market economies, include the Fund’s ARA EM metric among other appropriate tools.
  - Where reserve policies may significantly influence the effective operation of the IMS, the PN should discuss these policies.
  - Where a country faces potentially sizable drains in reserves, staff should discuss implications for the NIR position.
  - If authorities have important non-precautionary motives for holding reserves, the PN should discuss reserve holdings associated with meeting these needs.
  - When reserves are more than adequate for precautionary purposes, based on relevant standard metrics and scenario analysis, the PN should provide a detailed assessment of the opportunity costs of reserve holdings.
  - When using the tool designed for low-income credit constrained economies, the PN should discuss the cost of reserves.
  - Where a recent Article IV consultation (e.g., in the last two years) has detailed these issues and considerations are broadly unchanged, a brief update in the PN is sufficient.
- During the mission staff should discuss with authorities:
  - Staff’s assessment of reserve needs for precautionary purposes, including relevant metrics and adjustments for country-specific factors like commodity intensity and CFMs.
  - Authorities’ reserve policies, views on desirable reserve holdings, nature of vulnerabilities and risks, desirable external liquidity buffers and how they are met between reserves and alternative external buffers or policies, and non-precautionary motives to hold reserves.
  - Opportunity cost of reserve holdings when reserves are assessed to be at a comfortable level, and potential outward spillovers if reserve policy may significantly influence the effective operation of the IMS.
- The Staff Report:
  - Article IV staff reports should normally include a discussion and bottom line on staff’s assessment of reserve adequacy for precautionary purposes and a presentation of the authorities’ views where differences exist.
  - Discussion should, where appropriate, include non-precautionary needs and the cost of reserves, with emphasis reflecting country circumstances.
  - Assessment should include externalities associated with reserve policies if these policies have global systemic implications.
  - Where relevant, include discussion of potential drains, NIR, non-reserve buffers and appropriateness of FX intervention.
- Multilateral surveillance (e.g., ESR):
  - Adequacy of international reserves is a major pillar in staff assessment of the overall external position in the ESR.
  - Methodology and results in the ESR need to be consistent with the Article IV consultation staff report for each country.
  - ESR country assessments should reflect only staff views (teams should discuss the assessment with the authorities during relevant policy discussions).

### Role of reserve adequacy metrics and starting points for assessment
- Assessing reserve adequacy for precautionary purposes provides a useful starting point to ground the discussion on reserve issues.
- No universally accepted framework exists; several metrics are widely used as simple guides on reserve strength relative to specific risk factors.
- Choice of metrics depends on country circumstances; metrics provide a practical starting point beyond which country-specific risk factors complement the discussion.
- Staff should apply judgment in interpreting assessments derived from relevant metrics.

### Framework for selecting analytical tools by economy type
- Tools useful across economies are outlined; Annex II discusses considerations of market maturity and economic flexibility relevant for tool choice.
- Staff may base tool choice on traditional classification of countries: advanced, emerging, and low-income economies, noting some economies straddle categories (e.g., frontier markets).
- Section organization:
  - Part A: mature—or advanced—market economies (MMA) — principal tool is scenario analysis.
  - Part B: deepening financial—or emerging—market economies (DFM) — considerations and metrics outlined.
  - Part C: tool developed by the IMF for credit-constrained, mostly low-income economies vulnerable to current account shocks.

### Currency unions
- Currency unions span economies of all maturity levels; individual members can face balance of payments shocks.
- If a common central bank holding adequate reserves can allocate liquidity within the union, individual members may have less need for their own reserves.
- Reserve adequacy considerations at the consolidated union level depend on the nature of the currency union:
  - For unions able to issue a reserve currency (like the Euro area), consolidated considerations align with reserve currency issuers.
  - For unions comprising emerging market and low-income countries, reserve adequacy at the union level should be described in line with those economy types.
- Financial architecture and synchronization of shocks may limit reserve pooling:
  - Financial architecture: absence of a banking union or inefficient allocation of liquidity within the union limits pooling.
  - Synchronization: lack of economic diversification and correlated shocks (surges in food and fuel prices, plunges in FDI and terms of trade, drops in external demand of common trading partners) may require pooled reserves to be higher and potentially reach the aggregate reserve needs of each member.

### Mature Market Economies (MMA) — key considerations and scenario analysis
- MMAs without a reserve currency or automatic access to reserve currencies through standing swap lines may need reserve buffers for precautionary motives.
- Reserve needs often relate to acute financial market stress and dysfunction; required buffer should reflect:
  - potential FX funding needs of the financial system,
  - extent and availability of buffers in the financial system, and
  - potential duration of market dysfunction.
- The need to assist banking systems depends on size of FX asset-liability mismatches and maturities.
- Past experiences suggest dysfunctional funding and FX markets typically last between two and three weeks, but sometimes extend longer (ARA 2013).
- Standing liquidity swap arrangements are available from the U.S. Federal Reserve, the Bank of Canada, Bank of England, Bank of Japan, the ECB, and the Swiss National Bank.
- Providing reserves against financial sector risks can create moral hazard; supervisory and prudential policies are principal defenses. A fee could be charged by the central bank based on each institution’s contribution to risks covered by reserves to limit moral hazard.
- Due to modelling challenges and data gaps, standardized approaches or metrics for MMAs are difficult to develop; no consensus exists on appropriate reserve levels for precautionary purposes in MMAs.
- Where deeper assessment is warranted, Article IV consultations should use scenario analysis to assess potential FX funding needs relative to existing bank and nonbank buffers and test sensitivity:
  - Authorities’ scenario analysis: staff should discuss authorities’ reserve adequacy framework, dominant risks, possible reserve needs, and costs.
  - Staff scenario analysis: staff can complement with their own scenarios centered on market dysfunction risks and FX funding shortfalls, calibrated using historical information on trading liquidity, turnover, market participant behavior, and short-term funding needs.
  - Financial market stress: Article IV staff reports should discuss prudential and regulatory frameworks and buffers held in financial institutions as relevant.

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2016/_060316.pdf*

### 18. Staff should use all relevant reserve metrics in considering reserve needs for countries

### 18. Staff should use all relevant reserve metrics in considering reserve needs for countries

### Overview and guidance
- Appropriate metrics (Annex I provides a list of metrics) should be used with judgment in gauging reserve adequacy; there should not be any mechanical application of any metric.
- Traditional metrics include import coverage, and reserves to broad money or short-term debt.
- Optimality models, such as Jeanne and Rancière (2006), explicitly equate the marginal benefits and costs of reserves.
- Metrics provide a starting point for discussion of reserve adequacy for precautionary purposes because they summarize vulnerabilities to particular types of liquidity shocks and act as a simple—standardized—type of scenario analysis.
- Discussions should consider additional risks or drains on reserves (e.g., swap or forward transactions, or “encumbered” reserves); availability of additional external buffers (e.g., contingent financing lines and borrowing arrangements); and use of metrics in a dynamic setting to highlight expected evolution of reserves and vulnerabilities.

### ARA EM metric
- The ARA EM metric was proposed by the Fund as an additional metric against the backdrop of traditional metrics (Annex I). In the Board discussion, most Directors agreed with the methodology to calculate the ARA EM metric.
- Motivation: past balance of payments crises were characterized by multiple channels of market pressure, suggesting a need for a metric encompassing a broad set of risks. Empirical work and surveys (ARA 2011, ARA 2013) suggest reserves are held against multiple vulnerabilities.
- Coverage: the ARA EM metric covers the potential loss of export income, the risk of resident outflows (broad money), rollover risks (short-term debt), and the risk of nonresident equity and MLT debt outflows (other liabilities).
- Reporting: this metric should be reported along with other relevant analysis in Article IV consultation staff reports for these economies.
- Weights for the metric (as reported in the text table):
  - Revised Weights (in percent):
    - Short-term Debt: 30
    - Other Liabilities: 20
    - Broad money: 10
    - Exports: 10
  - Fixed regime:
    - Short-term Debt: 30
    - Other Liabilities: 15
    - Broad money: 5
    - Exports: 5
  - Floating regime:
    - Short-term Debt: 30
    - Other Liabilities: 15
    - Broad money: 5
    - Exports: 5

### Dynamic Reserve Assessment (Box 4)
- Proposal: a forward-looking approach that evaluates relevant metrics in light of the expected path of reserves and key risks, illustrated using the ARA EM metric and short-term debt.
- Methodology:
  - Use desks’ projections of metric components: changes in foreign exchange reserves, short-term debt, broad money, and exports (e.g., from the latest WEO).
  - Project portfolio equity and medium- and long-term debt based on latest IIP data and WEO projections for associated balance-of-payment flows (assuming no valuation effects); incorporate valuation effects where important.
  - Construct a projected ARA metric from these estimates.
- Applications: examples shown for India and Thailand — for most emerging markets, relative position vis-à-vis the ARA EM metric does not change much in the projection period, though changes can occur due to projected changes in reserves or country-specific components.
- Use: forward-looking diagnostics can serve as an important surveillance tool in policy discussions during Article IV consultations.

### Country-specific considerations — summary
- Staff should address a number of country-specific considerations when using the ARA EM and other metrics, including:
  - countries with long-standing capital flow management measures (CFMs);
  - commodity-intensive economies;
  - dollarized economies.
- Staff should outline how other relevant country-specific factors are built into the assessment.

### Countries with Capital Flow Management Measures (CFMs)
- Empirical findings:
  - CFMs tend to reduce resident outflows; Binici, Hutchinson, and Schindler (2010) find capital controls can substantially reduce equity outflows.
  - Saborowski and others (2014) report that effectively tightening controls reduced capital outflows in EMs when tightening is supported by strong fundamentals or good institutions, or existing restrictions are comprehensive.
  - Staff studies found CFMs can reduce the probability of exchange market pressure (EMP) events and reduce the ability of residents to transfer assets abroad (proxied by broad money) and the amount of resident outflows during EMP events.
- Implication for ARA EM metric:
  - Where countries have CFMs on residents, the weight on broad money (a proxy for risk from resident outflows) should be adjusted.
  - Presentation: both the standard (unadjusted) ARA EM metric and the metric with an adjusted weight on broad money—together with other relevant metrics—should be presented in Article IV staff reports.
  - Judgment: staff should use this information to form a judgment on overall adequacy, considering the nature of controls, their effectiveness, and progress towards liberalization.
  - If liberalization is expected, discuss speed of liberalization and need for supportive macroeconomic policies during transition; consider using the dynamic reserve assessment methodology.
- Empirical rule for identifying important CFMs:
  - The empirical work in ARA 2015 considered countries with at least two of the three standard capital control indicators (Chinn-Ito, Quinn and IMF share) with a value less than or equal to 0.25 (indices normalized between 0 and 1) as having important CFMs.
  - Teams should consider these indicators and may bring in additional indicators (including deviations from covered interest parity) in forming a view.
- Numerical guidance on adjusting weights (as presented):
  - Broad money weights without and with CFMs (in percent):
    - Without CFMs: Fixed 10, Floating 5
    - With CFMs: Fixed 5, Floating 2.5
  - Note: where countries maintain clear controls on exit of non-resident assets, the weight on “other liabilities” could be halved as discussed in ARA 2013 and ARA 2015.

### Commodity-Intensive Economies
- Risks and implications:
  - Commodity-intensive economies face more volatile terms of trade and greater difficulty adjusting since commodity imports or exports are relatively price inelastic.
  - They may have higher precautionary liquidity needs to smooth adjustment against commodity price changes.
- Treatment:
  - Maintain an additional buffer against terms-of-trade shocks rather than adjust the reserve adequacy metric per se.
  - The additional liquidity needs can be met through hedging, longer-term contracts, or savings under a sovereign wealth fund not included in reserves.
  - Staff paper should report the nominal size of the buffer and can supplement with ratios of reserves to the metric(s) augmented by the buffer when authorities intend to meet the buffer through reserves.
- Buffer calibration:
  - The buffer should be calculated on the basis of forward-looking adverse price movements at a given confidence level.
  - In general, a one standard deviation price “shock” (e.g., 68 percent confidence interval) should be sufficient to cover commodity price risks for smoothing temporary shocks or slowing adjustment to a permanent terms-of-trade shock.
  - If different assumptions are used, staff should clearly elaborate and justify reasoning in the staff report.
  - When reserves are used to meet the buffer, staff should report reserves relative to the metric augmented with the buffer, reserves relative to the metric without the buffer, and the nominal size of the buffer. Annex III provides methodology and an illustrative example.

### Dollarized Economies
- General:
  - Reserve adequacy should account for fully and partially dollarized economies’ specific circumstances.
  - Partially dollarized economies with their own currency and a central bank that can meet local currency needs: reserve adequacy considerations do not differ conceptually from non-dollarized ones.
  - Fully dollarized economies do not face exchange rate fluctuation risk or currency mismatches, so their FX liquidity buffer role differs.
- Fully dollarized economies:
  - May need liquidity buffers in the adopted foreign currency to support domestic financial institutions and as a fiscal buffer.
  - Liquidity pressures in banks can arise from export declines, sudden stops in external financing, non-resident flight, or resident runs.
  - Several fully dollarized economies complement high bank liquidity ratios with centralized reserve buffers or LOLR-type facilities.
  - Staff reports should assess whether available public and private buffers are sufficient to meet liquidity needs; components of the ARA EM metric capture most identified risks and may provide a conservative starting point.
  - For fiscal reserve buffer, Wiegand (2013) proposes one month of government spending as a standard yardstick.
  - Staff could discuss needs for additional tools (e.g., fiscal rules) to secure international reserve levels.
- Partially dollarized economies:
  - No strong empirical basis that partially dollarized economies need larger reserve buffers; the ARA EM metric is a useful starting point.
  - Literature is mixed: no clear positive association between degree of dollarization and crisis likelihood or cost; partially dollarized EMs have not seen larger outflows during market pressure events than others, though non-resident deposits may be more likely to leave in highly dollarized economies.
  - Article IV discussions should explain whether privately held buffers or dollarization imply higher or lower reserve needs compared with standard metrics, and which metric estimate is most appropriate.
  - Consider non-resident foreign exchange holdings risks, and larger FX reserve requirements often present in these economies.

### Credit-Constrained Economies (LICs)
- Context:
  - LICs are particularly vulnerable to exogenous shocks and have limited access to capital markets; reserves provide an important buffer against external stability risks.
- Approaches:
  - Traditional approaches remain useful: 3-month import rule, 20 percent broad money coverage, and 100 percent short-term debt coverage.
  - These simple methods may not reflect country specificities or capture multiple motives for holding reserves.
- ARA-CC approach:
  - The new Fund approach for credit-constrained economies (ARA-CC) seeks to balance marginal benefits and costs and address gaps of traditional methods.
  - This section, together with Annexes IV and VII, addresses practical issues in implementing reserve adequacy assessments for LICs in Article IV consultations using the ARA-CC approach.

*Source: _060316 - 18. Staff should use all relevant reserve metrics in considering reserve needs for countries_*

### 33. The ARA-CC approach is a useful complement to existing approaches as it takes into

### _060316 - 33. The ARA-CC approach is a useful complement to existing approaches as it takes into

### ARA-CC approach: scope and purpose
- Designed to assess adequacy of reserves for countries with limited or no access to international capital markets.
- Focuses primarily on shocks emanating from the current account side (e.g., adverse terms of trade, remittance and aid shocks).
- Balances the absorption-smoothing benefits of reserves against the opportunity cost of holding reserves, taking into account country specificities.

### Methodology for deriving adequate reserves (ARA-CC)
- Adequate level derived through an algorithm that equates marginal benefit and marginal cost:
  - Marginal benefit: absorption smoothing benefits estimated by two empirical regressions capturing:
    - (i) the role that reserves play in reducing the likelihood of a crisis, and
    - (ii) the impact of reserves in reducing the severity of a crisis.
  - Marginal cost: proxied by one of the external funding costs, sterilization cost, or the marginal product of capital (all net of returns from holding reserves).
- Adequate level of reserves = level where additional (marginal) benefit from one more unit of reserves = additional (marginal) cost to obtain and maintain that level of reserves.

### Country specificities and complementary metrics
- Attention to country-specific factors recommended (e.g., size and composition of debt, nature of shocks).
- For countries facing rollover risks or capital flight risks, complementary use of:
  - ARA EM metric,
  - short-term debt rule,
  - broad money rule.
- Small islands metric (drawing on ARA-EM) useful for economies vulnerable to natural disasters.
- Dollarized LICs may need to hold more reserves; required level depends on nature and degree of dollarization (see Box 6 in ARA 2013).

### IV. Measuring the cost of reserves — overview
- Reserves are a costly form of self-insurance; opportunity cost of reserve accumulation is an important input in bilateral policy discussions.
- Cost discussions relevant when opportunity cost has been rising and reserves are in or above adequate ranges.
- Guidance on cost measures varies across country types.

### A. Countries with Market Access (MMA and DFM)
- Sterilization and yield-based opportunity costs provide useful measures of marginal cost of reserves.
- Two main cost components:
  - Foregone return on an alternative use/asset of local authorities.
  - Cost of issuing paper for sterilization less the return on reserves.
- Indicator choices should reflect country circumstances, adequacy of reserves, alternative uses, and exchange rate level relative to fundamentals.

Key points:
- When reserves are inadequate or barely adequate:
  - Marginal cost of financing accumulation most relevant (cost of borrowing reserves or cost of sterilizing interventions, based on maturity of sterilization paper).
  - Quasi-fiscal cost of sterilizing accumulation can be high; includes anticipated exchange rate valuation losses.
- When reserves are ample:
  - Use “net financing” or opportunity cost: difference between local yield and return on reserve assets, adjusting yields for the effect of higher reserves in lowering yields.
- For market access economies with high externally issued foreign currency debt:
  - Use external FX denominated yield as opportunity cost proxy (proxied by the EMBI spread less the endogenous impact of higher marginal reserves in lowering spreads).
  - ARA (2011, 2015) found that for the median EM the rise in reserve holdings had essentially eliminated the impact of reserves on lowering marginal cost; convergence of net financial cost measure with EMBI spread noted.
  - Yield adjustments should count any anticipated FX depreciation and align with expected exchange rate path in Article IV consultations.
- For market access economies with adequate reserves and local currency debt that could be retired:
  - Local currency bond yield commonly used as proxy (opportunity cost of retiring local currency debt or using savings for projects).

### B. Countries that are Credit Constrained
- Opportunity cost can be approximated by market yield on a sovereign bond, or estimated marginal product of capital (MPK).

Guidance:
- If a credit-constrained country/LIC has issued a sovereign bond in the last five years, use the yield as a proxy.
- If no recent data or market biases exist, use average market yield for a subset of LICs or the cross-country currency interest swap (cross currency swap).
- If government securities markets are not heavily distorted, use return on government securities adjusted for exchange risk premium.
  - For LICs without forward exchange markets, compute premium using an average of historical real effective exchange rate depreciation.
  - Use return on longer-term securities (less influenced by monetary policy) while ensuring they are not illiquid.
- For other credit-constrained countries/LICs, use evolution of output-to-capital ratio as proxy for opportunity cost of foregone fixed investment (MPK).
  - MPK depends on investment as percent of GDP, output, depreciation rate, and share of capital stock (see Box 5 in IMF 2013 for details).
  - For resource-rich economies with significant FX resources at the central bank, MPK of public capital can be useful.
  - Due to measurement and data issues, MPK estimates could use averages (e.g., regional or LIC average).

### V. Tools and resources
- A number of the tools outlined are available at: http://www.imf.org/external/np/spr/ara/index.htm.
- Tools include:
  - Detailed cross-country data on reserve holdings and underlying data to calculate metrics.
  - EM ARA tools tailored for different types of EMs (buffers for commodity-intensive economies and adjusted metrics for countries with CFMs).
  - Main template applicable to all EMs; features include:
    - plotting reserves against several adequacy metrics,
    - tracking recent monthly reserves relative to ARA metric,
    - providing one-year ahead forecast of metric components.
  - Stand-alone tool for commodity-intensive economies to calculate additional buffer (see Section III. B and Annex V).
- ARA-CC tool workflow:
  - (i) determine whether country is “credit constrained”;
  - (ii) identify economic classification and exchange rate regime;
  - (iii) compile data for empirical regressions;
  - (iv) choose most appropriate proxy for cost;
  - (v) quantify net cost of holding reserves;
  - (vi) calibrate reserve adequacy.
  - Annex IV and Annex VI provide detailed information on using the ARA-CC tool.
- Examples of good practice in surveillance:
  - EM examples: India (commodity importer), Colombia and Russia (commodity exporters) — highlighted need to hold reserves against volatile terms of trade, risk of capital flight, and intergenerational equity considerations.
  - Croatia and South Africa — complementary role of regulatory FX liquidity buffers in banks, and fiscal cost of reserve holdings, respectively.
  - Sweden — FX liquidity shortages in the banking system after Lehman and corresponding reserve needs (MMA example).
  - Credit-constrained example: Mozambique — reserve needs weighed against high cost.
  - Full list and summaries available at http://www.imf.org/external/np/spr/ara/index.htm.

### Annex I — Traditional metrics and ARA EM metric highlights
- Traditional metrics:
  - Import coverage (months of prospective imports; three months’ coverage typically used).
  - Reserves to short-term debt (Greenspan-Guidotti rule: 100 percent cover of ST debt).
  - Reserves to broad money (M2) — upper end of prudent range typically 20 percent, threshold around 5 percent more typical.
  - Combination metrics: expanded Greenspan-Guidotti (ST debt plus current account deficit if in deficit); Wijnholds and Kapteyn (2001) combining ST debt and M2.
- Optimal reserve models:
  - Jeanne and Rancière (2006): optimal reserves balance potential loss in output/consumption given sudden stop probability with opportunity cost.
  - Jeanne and Rancière (2006) suggests many EMs optimally hold reserves around 80–100 percent of ST debt plus current account deficit and between 75 to 150 percent of the ARA EM metric, though results sensitive to assumptions.
- ARA EM metric:
  - Comprises four components reflecting potential drains on balance of payments:
    - (i) export income,
    - (ii) broad money,
    - (iii) short-term debt,
    - (iv) other liabilities.
  - Relative risk weights based on 10th percentile of observed outflows from EMs during exchange market pressure episodes.
  - Does not include current account deficit as a separate component; financial financing of deficit should be captured in ST debt or other liabilities.
  - Reserves in the range of 100-150 percent of the composite metric are considered broadly adequate for precautionary purposes.
  - Weights on “other liabilities” raised by 5 percentage points in ARA (2015).

### Annex II — Country classification and application
- Country tolerance to external risks influenced by flexibility of economy, maturity of markets, and extent of market access.
- For countries without well-established market access, reserve needs based around current account risks are most appropriate.
- Indicators for market access durability and depth include magnitude and frequency of public issuance over past three-to-five years, sovereign debt rating, borrowing in international markets, and use of government or external guarantees.
- Provided maturity-based country classification (data as of January 2016) and cluster-based classifications for policy differentiation.

*GUIDANCE NOTE ON RESERVE ADEQUACY  
INTERNATIONAL MONETARY FUND*

### Annex III. Commodity Intensive Economies and

### Annex III. Commodity Intensive Economies and Calculation of the Commodity Buffer

### Commodity intensity of economies
- Staff may use judgment on whether an additional commodity buffer is relevant for a particular country given the commodity intensity of their imports and exports.
- ARA (2015) reported commodity intensity using COMTRADE data for 2007–13; exporters with commodity intensity above 50 percent of total goods exports broadly match the WEO classification of commodity exporters (with Colombia added as a fuel exporter and Argentina, Brazil and Peru as non-fuel exporters).
- For importers, 20 percent was proposed in ARA (2015).
- Source of commodity shares and calculations: WITS (UN Comtrade) and IMF staff calculations.
- Commodity categories defined in this annex:
  - Fuels: coals, petroleum, natural gas, and electric current.
  - Primary commodities (non-fuel): food and live animals, beverages and tobacco, crude materials (except fuels), animal and vegetable oils, and non-ferrous metals.
- Note: Advanced economies and PRGT-eligible countries are excluded from the commodity-intensity charts.

### Methodology to calculate the commodity buffer
- The additional commodity buffer is calculated as:
  - Buffer = (price gap) × (trade value)
  - Price gap = percent difference between current prices and a specified point of the forecast price distribution (percent change from baseline price projections).
- Distribution and confidence interval:
  - ARA (2013) used a future price at the 68 percent confidence interval (equivalent to one standard deviation if the future distribution is Gaussian).
  - The distribution of future prices can be based on a model (e.g., VAR) or on option/futures prices.
  - Prior to the global financial crisis, futures prices—particularly for energy—provided an unbiased projection of commodity prices; since then model-based forecasts have outperformed futures for recent periods.
- Practical application:
  - For a given country, compute the one-year ahead forecast at the 68 percent confidence interval; compute percent gap versus the current price; multiply that percent gap by the current-year dollar value of commodity exports or imports to obtain the buffer.
  - If authorities intend to meet the need through reserves rather than hedging or other methods, the buffer can be added to the chosen reserve metric; the unaugmented metric should also be shown.
- Operational note:
  - Countries with institutions that produce long-term commodity price distributions (e.g., Chile) may use those distributions if assumptions are discussed with authorities.

### Illustrative example (Box AIII-1): Buffers for the ARA EM metric in commodity intensive economies
- Example countries considered: Colombia, Kazakhstan, Russia (fuel exporters) and India (net fuel importer).
- General result: Reserves remain well above 100 percent of the metric, and above the metric augmented by the commodity buffer in the illustrative examples.
- Fuel exporters (example methodology and numbers):
  - The additional buffer captures the risk of lower future oil prices.
  - Example price gap: 19 percent = difference between the one year ahead price forecast at the 68 percent confidence interval (US$35 per barrel as of Nov 2016) and the current WTI price (US$44 per barrel as of Nov 2015).
  - The price gap is multiplied by the value of oil exports in 2015 to calculate the additional buffer.
- Fuel importers (example methodology and numbers):
  - The additional buffer captures the risk of higher future oil prices.
  - Example gap: about 49 percent = difference between the one year ahead price forecast at the 68 percent confidence interval (US$65 per barrel as of November 2016) and the current WTI price (US$44 per barrel as of November 2015).
  - The price gap is multiplied by the value of oil imports in 2015 to calculate the additional buffer.
- Presentation guidance:
  - If augmenting an ARA metric with the calculated commodity buffer, present both the augmented and unaugmented ratios of reserves to the metric.

### Implementation resources and tooling
- Annex V provides a commodity buffer template and tool (Commodity buffer template) that:
  - Includes illustrative examples for Colombia (fuel exporter) and India (fuel importer).
  - Automatically updates gross reserves and the ARA EM metric.
  - Requires user inputs: current year dollar value of commodity imports/exports and the latest commodity price projections from the IMF Research Department commodities team.
  - Computes the price gap as the percent difference between the current commodity price and the one-year ahead forecast at the 68 percent confidence interval, and then computes the buffer by multiplying the price gap by the current value of commodity exports/imports.
- Practical note on obtaining projections used in the template: the template example is based on the November 2015 commodity price outlook and risks; the latest commodity price projections are available from the IMF Research Department commodities team.

*Source: Annex III, "Commodity Intensive Economies and Calculation of the Commodity Buffer," Guidance Note on Reserve Adequacy (IMF).*

### Annex VI. Assessing Reserve Adequacy Tool

### Annex VI. Assessing Reserve Adequacy Tool
### Overview and purpose
- Tool: ARA tool for Credit-constrained Economies (LICs) to calculate reserve adequacy metrics for credit-constrained economies.
- Template features:
  - Automatically obtains data from various sources including a country desk macro-framework and the World Economic Outlook (WEO) databases.
  - Includes all developing countries and allows selection of countries perceived to be credit constrained.
  - Mirrors procedures in the template; flow summarized in Figure AVI-1 and template structure in Figure AVI-2.
  - Designed sheets referenced: “Home Sheet”, “Approach”, “Computation”, “Sensitivity Analysis”, and background/input sheets.

### Operational steps (template workflow)
- Step 1: Choose the country
  - Template includes all developing countries; user selects the country to analyze.

- Step 2: Choose economic classification, exchange rate regime and country paths
  - Economic classification automatically populated into either “resource rich” or “non-resource rich” economy using defaults drawn from the 2012 IMF board paper “Macroeconomic policy frameworks for resource-rich developing countries”.
  - Default exchange rate regime drawn from the AREAER database (de facto regime).
  - Staff options:
    - Manually adjust economic classification and exchange rate regime (to capture frontier, fragile, or small state sub-samples or reflect regime changes).
    - Use “reset” button to return to default settings.
    - Adjust database path and/or variable codes used in the analysis.
  - Adjustments affect coefficients in the marginal benefit regressions.

- Steps 3–6: Choose data to estimate marginal benefits of holding reserves
  - Template includes regression coefficients (described in IMF, 2011 and 2013) and relevant country data linked to comprehensive datasets and databases (including country desks).
  - Country teams can override data codes and coefficients as needed to best reflect country situation.
  - Data inputs for benefits update when file is refreshed; economic classification and cost data are based on IMF (2013, 2015) Board paper classifications and updated annually.

- Step 7: Estimate the cost of holding reserves (net of real returns)
  - Template provides estimates drawn from ARA (2013) and ARA (2015); automatically populated.
  - Staff options:
    - Pick which proxy to use and the sample (e.g., the average cost for that approach from ARA (2015), the average for countries in similar region, economic size, or degree of fragility).
    - Manually insert custom cost estimates.

### Results presentation and interpretation
- Calibration and outputs:
  - Adequate level of reserves calibrated in sheet “Computation”.
  - Results presented in the “home” sheet.
  - Template provides a comparison with the traditional three month rule (visuals: graphs, a table, and text).
- Judgment:
  - Judgment is needed in interpreting results (see previous section on judgment).
- Scenario and sensitivity analysis:
  - “Sensitivity Analysis” sheet allows adjustment of variables to assess impacts on adequate reserves.
  - Examples: adjust government balance to assess adverse fiscal outturn effects; alternative macroeconomic frameworks (adverse fiscal balance, less benign external environment).
  - A separate sheet mirrors the “home” sheet to view scenario results side-by-side.
  - Teams may alter one variable and examine sensitivity (e.g., one standard deviation in fiscal balance).

### Estimated coefficients for marginal benefits computations (Table AVI-1)
- Likelihood of a crisis — coefficients by classification
  - Reserve, months of imports:
    - Resource Rich: -0.093
    - Non Resource Rich: -0.090
    - Updated Resource Rich: -0.093
    - Small State: -0.093
    - Fragile: -0.090
    - Frontier: -0.090
  - Government balance, % of GDP:
    - Resource Rich: -0.031
    - Non Resource Rich: -0.032
    - Updated Resource Rich: -0.031
    - Small State: -0.035
    - Fragile: -0.032
    - Frontier: -0.032
  - CPIA:
    - Resource Rich: -0.314
    - Non Resource Rich: -0.309
    - Updated Resource Rich: -0.314
    - Small State: -0.246
    - Fragile: -0.309
    - Frontier: -0.309
  - Flexible exchange rate regime, 1 if flexible:
    - Resource Rich: -0.381
    - Non Resource Rich: -0.380
    - Updated Resource Rich: -0.381
    - Small State: -0.440
    - Fragile: -0.380
    - Frontier: -0.380
  - IMF program:
    - Resource Rich: -0.291
    - Non Resource Rich: -0.302
    - Updated Resource Rich: -0.291
    - Small State: -0.394
    - Fragile: -0.302
    - Frontier: -0.302
  - Constant:
    - Resource Rich: 0.922
    - Non Resource Rich: 0.865
    - Updated Resource Rich: 0.922
    - Small State: 0.782
    - Fragile: 0.865
    - Frontier: 0.865

- Magnitude of absorption drop — coefficients by classification
  - Reserves, months of imports, in logs:
    - Resource Rich: -2.257
    - Non Resource Rich: -2.240
    - Updated Resource Rich: -2.239
    - Small State: -1.945
    - Fragile: -1.945
    - Frontier: -1.945
  - Flexible exchange rate regime, 1 if flexible:
    - Resource Rich: -8.623
    - Non Resource Rich: -8.698
    - Updated Resource Rich: -8.611
    - Small State: -9.418
    - Fragile: -9.418
    - Frontier: -9.418
  - External demand growth, percent:
    - Resource Rich: -1.002
    - Non Resource Rich: -0.932
    - Updated Resource Rich: -0.833
    - Small State: -0.901
    - Fragile: -0.901
    - Frontier: -0.901
  - Terms of trade growth, percent:
    - Resource Rich: -0.086
    - Non Resource Rich: -0.084
    - Updated Resource Rich: 0.000
    - Small State: 0.000
    - Fragile: 0.000
    - Frontier: 0.000
  - Change in FDI to GDP, percentage point of GDP:
    - Resource Rich: -0.023
    - Non Resource Rich: -0.016
    - Updated Resource Rich: 0.000
    - Small State: 0.000
    - Fragile: 0.000
    - Frontier: 0.000
  - Change in aid to GDP, percentage point of GDP:
    - Resource Rich: 0.000
    - Non Resource Rich: 0.053
    - Updated Resource Rich: 0.000
    - Small State: 0.000
    - Fragile: 0.000
    - Frontier: 0.000
  - Fixed effects:
    - Resource Rich: 7.569
    - Non Resource Rich: 3.784
  - Terms of trade growth, percent*resource rich:
    - Resource Rich: 0.000
    - Non Resource Rich: 0.000
    - Updated Resource Rich: -0.181
    - Small State: 0.000
    - Fragile: 0.000
    - Frontier: 0.000
  - Resource Rich (interaction / indicator):
    - Resource Rich: 0.000
    - Non Resource Rich: 0.000
    - Updated Resource Rich: 5.297
    - Small State: 0.000
    - Fragile: 0.000
    - Frontier: 0.000

### Template customization and staff judgment
- Staff may:
  - Override coefficients and data codes to reflect particular country variables or alternative variables.
  - Choose cost proxies from ARA (2013) and ARA (2015) or enter custom costs.
  - Adjust country classification, exchange rate regime, data sources, and variable codes.
- Template behavior:
  - Data inputs for benefits update automatically when refreshed from linked databases.
  - Economic classification and cost data follow IMF (2013, 2015) Board paper classifications and are updated annually.

*Source: Annex VI. Assessing Reserve Adequacy Tool*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2016/_060316.pdf_