## _dp1401

## Source details

**Canonical URL:** [_dp1401](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/dp/2014/_dp1401.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/dp/2014/_dp1401.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/dp/2014/_dp1401.pdf.json)

---

### Overview of EBA/CGER Coverage and Methodology
- The pilot External Balances Assessment (EBA) methodology provides estimates for current account and exchange gaps and was used as inputs into external assessments for 29 large economies in the 2013 Pilot External Sector Reports (ESR).
- EBA applies to 49 economies; the IMF membership consists of 188 countries.
- EBA improvements over Consultative Group on Exchange Rates (CGER) methods:
  - Strips out cyclical factors and includes a greater range of explanatory variables.
  - Estimates the impact of actual policies (fiscal policy, monetary policy, capital controls, reserve accumulation, proxies for social protection and financial policies) on CA and real exchange rate, and identifies policy gaps.
- Coverage limitations:
  - EBA covers approximately 50 advanced and emerging market economies and does not generally include many “special case” countries (e.g., countries with concentrated external income sources, small/low-income economies, many energy-specialized economies).

### Practices and Survey Findings on External Assessments
- Standard CGER-type approaches used where EBA is not applied:
  - Macrobalance (MB) approach: deviations relative to a CA norm.
  - Equilibrium real exchange rate (ERER) approach: deviations from an exchange rate norm.
  - External sustainability (ES) approach: deviations from a sustainable CA balance.
- Survey of 88 staff reports:
  - About 50 major advanced and emerging market economies are included in CGER and EBA exercises, representing more than 90 percent of global output.
  - Nearly three-quarters of reports used adjusted CGER methods; three-quarters used all three CGER-type approaches; 20 percent used two; all but five used at least one.
  - Only five reports did not use CGER or adjusted CGER methods, due to data limitations.
- Table 1 — bottom line exchange rate assessment in Article IV Reports (Percent of the reports surveyed):
  - CGER methods (83 reports): Overvalued 31.8, Equilibrium 56.8, Undervalued 6.8, Undefined 0.0, All 95.5
  - Adjusted CGER methods (65 reports): Overvalued 23.9, Equilibrium 43.2, Undervalued 5.7, Undefined 0.0, All 72.7
  - Other methods (5 reports): Overvalued 2.3, Equilibrium 0.0, Undervalued 0.0, Undefined 2.3, All 4.5
  - Total: Overvalued 34.1, Equilibrium 56.8, Undervalued 6.8, Undefined 2.3, All 100.0

### External Assessments in Special Cases and Typical Adjustments
- Most common “special” characteristic: concentration of external income in one or more sectors (nonrenewable commodity exports, tourism, financial services, foreign aid, remittances).
- Challenges:
  - EBA methods take special account of oil and natural gas balances but do not generally account for other commodity exports or services; many countries with concentrated external income are outside EBA coverage.
  - Country-specific adjustments can introduce concerns about overfitting, unequal treatment, and multilateral inconsistency.
- Typical staff adjustments:
  - Adding variables relevant in special cases (e.g., aid flows).
  - Estimating regressions on subsets of countries (e.g., tourism exporters).
  - Changing regression specifications or sample periods (e.g., modifying underlying exchange rate model).
- Tradeoffs:
  - Balancing errors from omitting relevant variables against errors from using selective subsamples or alternative specifications.

### Toolkit, Templates, and Frameworks for Non-EBA Countries
- Exchange rate assessment toolkit:
  - Extends CGER-type methods to a wider panel of countries, provides up-to-date coefficient estimates, allows for special country circumstances while aiming to preserve multilateral consistency.
  - Implemented variants of PPP, MB, ERER, and ES approaches on an annual panel dataset spanning 184 economies from 1973 to present.
  - Reports exchange rate misalignment estimates together with intermediate results (panel regression coefficients, multilateral consistency adjustments).
  - Includes measures of aid and remittance inflows for emerging and developing economies.
  - Note: Toolkit generally does not take account of policy distortions emphasized by EBA.
- External indicators template:
  - Structured template for comparing external indicators (PPP, unit labor costs, reserves, competitiveness measures, economy-specific indicators such as tourist arrival market share).
  - Encourages competitiveness scorecards and panels for external stability when quantitative estimation is constrained.
- Framework for capital-intensive, foreign-owned sector:
  - Distinguish transactions with foreign-owned sector; separate commodity exports, related capital imports, and income payments to foreign investors from CA to assess domestic-sector competitiveness.
  - Data limitations often complicate separation; illustrative separation approach discussed.

### Characteristics and Key Statistics for Country Groups with Concentrated External Income (Table 2 summary)
- Countries with concentrated external income account for "just over 10 percent" of world output; about half are small states; few are included in CGER.
- Key statistics (Nominal GDP (billions of USD); Average Per Capita Income (USD); Population (millions); CAB (in percent of GDP); Fiscal Balance (in percent of GDP); NFA):
  - Exporters of non-renewable comm.: 117.7; 10,003; 22.3; 7.1; 4.5; −15.2
  - Exporters of financial services: 256.4; 29,027; 7.0; −0.5; −2.6; 33.6
  - Exporters of tourism services: 4.8; 8,404; 40.6; −13.3; −4.6; −92.0
  - Recipients of remittances: 7.7; 2,540; 3.6; −6.7; −2.8; −65.0
  - Recipients of aid: 3.1; 1,090; 6.2; −7.4; −1.0; −53.4
  - All countries: 340.1; 10,719; 37.5; 0.1; −1.0; −30.5

### Risks from Adjustments and Empirical Evidence on Subsample vs Full-Sample Estimates
- Common adjustments: modifying panel sample, explanatory variables, and econometric methodology.
- Risk: Estimating on subsamples of similar economies can reduce estimated CA gaps or exchange rate misalignments, potentially reflecting sample selection bias rather than better fit.
- Empirical illustration (Table 3: Average Directional Difference in Exchange Rate Misalignment Estimates; Full Sample - Subsample; in percentage points):
  - MB Approach / ERER Approach
    - Exporters of nonrenewable commodities: 11.3 / −1.9
    - Exporters of financial services: 4.2 / 2.1
    - Exporters of tourism services: 14.3 / 9.7
    - Recipients of aid: 6.2 / 7.6
    - Recipients of remittances: 6.5 / 2.0
    - Average: 8.6 / 4.5
  - Example interpretation: For tourism exporters, using the full sample (184 countries) without customization yields an exchange rate gap estimate 14.3 percentage points further from zero than an estimate based on a subsample of 19 tourism exporters (MB approach).

### Allowing for Country-Group Heterogeneity and Quantified Impacts
- Allowing slope heterogeneity across country groups materially affects exchange rate assessments; not allowing heterogeneity may overstate exchange rate gaps.
- Quantified impacts (Average Directional Difference of Estimated Exchange Rate Gaps; Full sample, with no customization minus Full sample, with customization, in percentage points):
  - MB (REER) / ERER (REER)
  - Exporters of nonrenewable commodities: 5.8 / 5.6
  - Exporters of financial services: 1.7 / 2.6
  - Exporters of tourism services: 1.1 / 2.5
  - Recipients of aid: 2.5 / 1.9
  - Recipients of remittances: 0.5 / 0.7
  - Average (Customized groups): 0.9 / 2.7
  - Average (All other countries): 2.3 / 0.1
- Example interpretation: for nonrenewable commodity exporters, average estimated exchange rate gaps for both MB and ERER approaches are about 6 percentage points further from zero in the absence of slope heterogeneity.

### Data Issues, Measurement Distortions, and Separation of Foreign-Owned Resource Sector
- Data and measurement issues:
  - Large foreign ownership can generate flows that dwarf domestic-sector flows, distorting exchange rate elasticities and competitiveness assessments.
  - Standard CA statistics can distort measured external wealth accumulation when income is generated in one location but ultimately owned elsewhere (example: Swiss National Bank analysis suggests measurement issues could add roughly 3½ percent of GDP per year to Switzerland’s measured CA surplus).
  - Incomplete reporting of investment income breakdowns hinders accounting for reinvested earnings.
- Separation approach (where foreign-owned resource sector dominates):
  - Define domestic component excluding resource value added; exclude resource revenues and royalties from fiscal balance and treat them as remittances to domestic non-resource sectors.
  - Separate commodity exports, related capital imports, and income payments to foreign investors from the CA.
  - Assumptions often needed: example assumptions used in Appendix 6 illustrative case—income payments account for 40 percent of resource export revenues; resource-related imports amount to 40 percent of FDI inflows declining over time.
- Illustrative comparative results (Table A6.1 — CA Norm, Underlying CA, CA Gap (percent of GDP)):
  - 2011: Overall Economy CA Norm 4.97; Underlying CA 7.62; CA Gap 2.7. Adjusted Economy CA Norm −4.7; Underlying CA −10.6; CA Gap −5.9.
  - 2012: Overall: CA Norm 5.16; Underlying CA 6.6; CA Gap 1.5. Adjusted: CA Norm −8.8; Underlying CA −12.3; CA Gap −3.5.
  - 2013: Overall: CA Norm 4.85; Underlying CA 5.6; CA Gap 0.7. Adjusted: CA Norm −9.8; Underlying CA −10.4; CA Gap −0.6.
  - 2014: Overall: CA Norm 3.7; Underlying CA 4.0; CA Gap 0.3. Adjusted: CA Norm −10.4; Underlying CA −7.8; CA Gap 2.5.
  - 2015: Overall: CA Norm 3.13; Underlying CA 3.3; CA Gap 0.2. Adjusted: CA Norm −9.4; Underlying CA −5.8; CA Gap 3.6.
  - 2016: Overall: CA Norm 2.42; Underlying CA 2.8; CA Gap 0.4. Adjusted: CA Norm −7.2; Underlying CA −4.2; CA Gap 3.0.
  - 2017: Overall: CA Norm 2.0; Underlying CA 2.9; CA Gap 0.9. Adjusted: CA Norm −7.0; Underlying CA −2.8; CA Gap 4.2.
- Implication: Adjusted domestic-economy assessments can diverge sharply from aggregate assessments, with important consequences for policy advice and growth projections.

### Empirical Estimation Results and Robustness (Selected)
- MB approach — selected final model coefficients (standard errors in parentheses):
  - C: −0.006 (0.009)
  - C for exporters of tourism services: −0.083*** (0.018)
  - Relative income for exporters of financial services: 0.027** (0.012)
  - Output growth: −0.944*** (0.292)
  - Oil trade balance: 0.152*** (0.057)
  - Oil trade balance for recipients of aid: 0.589** (0.302)
  - Fiscal balance for exporters of nonrenewable commodities: 0.541** (0.231)
  - NFA: 0.030*** (0.009)
  - Grant/aid: −0.301* (0.180)
  - Grant/aid for recipients of aid: −0.386*** (0.160)
- MB adjusted R-squared values by sample:
  - Benchmark: 0.393
  - Aid Recipients: 0.382
  - Exporters of Nonrenewable Commodities: 0.444
  - Financial Centers: 0.397
  - Tourism-dependent Countries: 0.422
  - Final Model: 0.436
- ERER approach — selected final model coefficients (standard errors in parentheses):
  - C: 0.115*** (0.040)
  - Terms of trade for exporters of nonrenewable commodities: 0.396** (0.192)
  - Relative productivity for recipients of remittances: 0.393* (0.209)
  - Relative government consumption: 1.407*** (0.417)
  - Relative government consumption exporters of nonrenewable commodities: 10.575*** (2.543)
  - Aid flows: 0.816 (0.695)
  - Remittance flows: 0.202 (0.280)
- ERER Adjusted R-squared and observations:
  - Adjusted R-squared: 0.421 (Benchmark/Final Model), 0.307 (Exporters of Nonrenewable Commodities), 0.421 (Remittance Recipients)
  - Observations: 1951 across reported models
- Robustness:
  - Parameter sensitivity: perturbing classification thresholds by plus or minus 20 percent does not materially affect coefficient estimates and exchange rate gaps.
  - Estimates robust to leaving-one-out samples; 95 percent confidence intervals shown for coefficients across 184 countries.

### Financial Centers: Features, Adjustments, and Guidance
- Defining features:
  - Large financial services exports, FDI and investment flows, and/or net foreign assets (NFA); sizable positive NFA positions and relatively high investment income.
- Empirical and policy considerations:
  - Swiss CA may be overstated by measurement of retained earnings; SNB estimates effect could add roughly 3.5 percent of GDP per year to measured CA surplus.
  - Gross external assets and liabilities can be far larger than CA and may better reflect exchange rate determination and external stability (example: United Kingdom gross external liabilities > 550 percent of GDP in 2012 while CA was −3½ percent of GDP).
- Exchange rate gap examples (real effective exchange rate over/undervaluation ranges in percent from staff reports):
  - Switzerland: Undervalued [(–15.2)–6.1]
  - Singapore: [(–8)–(–1)]
  - Belgium: Equilibrium [(–1)–6]
  - Ireland: [(–3.9)–6]
  - Luxembourg: [2.05–8.7]
  - Panama: [1–3]
  - United Kingdom: [(–4)–0]
  - Hong Kong SAR: [(–10)–6]
  - Barbados: Overvalue [5–10]
  - Cyprus: [11.86–21.23]
- Guidance:
  - Emphasize analysis of gross foreign assets and liabilities and valuation effects to supplement CA-based assessments.
  - Quantitative measures tend to exhibit wider ranges for financial centers; regressions restricted to subsamples may introduce biases.

### Best Practices and Policy Recommendations for Staff Practice
- Transparency:
  - Disclose adjustments, economic rationale, and implications when changing CGER methodologies or correcting for one-off factors and measurement biases.
  - If customized methodologies are used, provide clear discussion on how assessments differ from standard methods.
- Coverage and indicators:
  - Move beyond quantitative point estimates; include broader indicators: capital flows, reserve adequacy, foreign exchange intervention, capital controls, external assets and liabilities.
  - Include structural indicators of competitiveness, especially where data are limited.
- Methodological recommendations:
  - Prefer a wide country sample and either measure/model omitted structural characteristics directly or employ dummy/interaction terms for country groups.
  - Document steps to derive REER misalignment estimates, variable contributions to CA norms or equilibrium REER, and time horizons used.
  - Avoid non-transparent derivations, reliance on one empirical method, and basing ER assessments on very limited samples.
- Data improvements suggested:
  - Correct for distortions related to investment income.
  - Focus on domestic competitiveness by netting out flows to large foreign-owned sectors.
  - Enhance data provision to enable separation of foreign-owned sector flows.

*Source: IMF Departmental Paper content unit _dp1401 (Appendices and staff calculations as provided).*

### 1.  Foreign  exchange  rates.  I.  Ter-Martirosyan,  Anna.  II.  International

### 1.  Foreign exchange rates. I. Ter-Martirosyan, Anna. II. International

### Overview
- The pilot External Balances Assessment (EBA) methodology provides estimates for current account and exchange gaps for a group of advanced and emerging market economies and was used as inputs into external assessments for 29 large economies in the 2013 Pilot External Sector Reports (ESR).
- EBA applies to 49 economies; the IMF membership consists of 188 countries.
- The EBA builds on and improves Consultative Group on Exchange Rates (CGER) methods by:
  - Stripping out cyclical factors and including a greater range of explanatory variables.
  - Estimating the impact of a country’s actual policies (fiscal policy, monetary policy, capital controls, reserve accumulation, proxies for social protection and financial policies) on its CA and real exchange rate, and identifying policy gaps.
- Coverage limitations: EBA covers approximately 50 advanced and emerging market economies and does not generally include many “special case” countries (e.g., countries with concentrated external income sources, small/low-income economies, many energy-specialized economies).

### Practices and Issues in External Assessments
- Country teams continue to rely on CGER-type approaches where EBA is not applied; the three standard CGER approaches are:
  - Macrobalance (MB) approach: estimates deviations relative to a current account (CA) norm.
  - Equilibrium real exchange rate (ERER) approach: estimates deviations from an exchange rate norm.
  - External sustainability (ES) approach: estimates deviations from a sustainable CA balance.
- Survey and usage findings:
  - A survey of 88 staff reports examined methods used in recent Article IV Reports and compared them to CGER-type methods.
  - About 50 major advanced and emerging market economies are included in CGER and EBA exercises, representing more than 90 percent of global output.
  - CGER methods (including adjusted versions) are widely used in bilateral surveillance for countries outside EBA.
  - Only five reports in the survey did not use CGER or adjusted CGER methods, due to data limitations.
  - Nearly three-quarters of the reports used adjusted CGER methods (i.e., customized CGER methods to account for special characteristics).
  - Three-quarters of the reports used all three CGER-type approaches; 20 percent used two; all but five used at least one.
- Table 1 — Comparison of bottom line exchange rate assessment in Article IV Reports (Percent of the reports surveyed):
  - CGER methods (83 reports): Overvalued 31.8, Equilibrium 56.8, Undervalued 6.8, Undefined 0.0, All 95.5
  - Of which: Adjusted CGER methods (65 reports): Overvalued 23.9, Equilibrium 43.2, Undervalued 5.7, Undefined 0.0, All 72.7
  - Other methods (5 reports): Overvalued 2.3, Equilibrium 0.0, Undervalued 0.0, Undefined 2.3, All 4.5
  - Total: Overvalued 34.1, Equilibrium 56.8, Undervalued 6.8, Undefined 2.3, All 100.0

### External Assessments in Special Cases
- The most common “special” characteristic is concentration of external income in one or more sectors (nonrenewable commodity exports, tourism, financial services, foreign aid, remittances).
- Challenges:
  - EBA methods take special account of oil and natural gas balances but do not generally account for other commodity exports or services; many countries with concentrated external income are outside EBA coverage.
  - Country-specific characteristics can limit applicability of standard methods and complicate multilateral consistency.
  - Teams’ country-specific adjustments can introduce concerns about potential overfitting, unequal treatment, and multilateral inconsistency.
- Typical staff adjustments to standard methodologies:
  - Adding other variables relevant in special cases (e.g., aid flows).
  - Estimating regression models on subsets of countries (e.g., tourism exporters).
  - Changing regression specifications or sample periods (e.g., modifying underlying exchange rate model).
- Tradeoffs highlighted:
  - Balancing errors from omitting relevant variables against errors from using selective subsamples or alternative specifications.

### Tools and Approaches Presented
- Exchange rate assessment toolkit:
  - Extends CGER-type methods to a wider panel of countries, pending further development and sample extension of EBA.
  - Provides more up-to-date coefficient estimates, allows for special country circumstances while aiming to preserve multilateral consistency.
  - Is adjustable to accommodate country-specific characteristics.
- External indicators template:
  - A structured template for comparing various external indicators across countries to supplement model-based assessments.
  - Encourages use of alternative indicators (purchasing power parity, unit labor costs, reserves, competitiveness measures, economy-specific indicators such as tourist arrival market share).
- Framework for analysis of a capital-intensive, foreign-owned sector:
  - Where a resource sector is largely foreign-owned and dominates external flows, distinguishing transactions with the foreign-owned sector can affect assessments and policy advice.
  - Data limitations often complicate separation of foreign-owned sector activities; an illustrative approach to separation is discussed in the paper.

### Survey Findings on Method Use and Customization
- Use of CGER methodologies across country groups shows customization is common, especially for:
  - Non-renewable commodity exporters, renewable commodity exporters, aid/remittances recipients, tourism-dependent economies, and financial centers.
- Complementary indicators commonly used include purchasing power parity, unit labor costs, reserves, and competitiveness measures; some indicators are specific to the economy (e.g., market share of tourist arrivals).
- IMF staff are assessing how to extend EBA innovations to the rest of the membership either by expanding the EBA sample or applying EBA aspects to other analyses; data availability and structural differences are constraints.

### Adjustments in Special Cases — Conceptual and Empirical Considerations
- Adjustments can improve model fit and alignment with a country’s economic structure.
- However, customized adjustments raise questions regarding:
  - Accuracy of estimates.
  - Multilateral consistency across IMF assessments.
  - Evenhandedness in treatment across countries.
- The paper reviews various country experiences and emphasizes that no one-size-fits-all approach exists; judgment remains essential when choosing methods and assessing external positions.

*Source: IMF Departmental Paper prepared by staff from the IMF’s Strategy, Policy, and Review Department, led by Anna Ter-Martirosyan (contents and excerpts from the provided PDF content unit).*

### Appendix 2, more than 80 economies

### Appendix 2, more than 80 economies

### Overview
- Countries with concentrated sources of external income account for "just over 10 percent" of world output.
- About half of these countries are small states; only a few—mainly financial centers—are included in CGER.
- Country groups display distinct external outcomes, suggesting distinct structural characteristics relevant to external assessments.

### Characteristics by country group
- Exporters of nonrenewable commodities (including oil)
  - Tend to have relatively high positive current account (CA) and fiscal surpluses.
  - Intergenerational equity considerations encourage saving resource earnings, increasing CA balances.
  - Alternative views: with constrained external borrowing, optimal use of resource income to fund real investment could yield smaller CA surpluses.
  - Some IMF country teams use a model-based framework accounting for optimal consumption and investment behavior and structural characteristics of resource-rich developing countries to produce CA norms.
- Exporters of financial services
  - Tend to have high CA surpluses, sizable positive net foreign assets (NFA) positions, and high gross financial assets and liabilities.
  - High CA balances may partly reflect accounting issues in measuring investment income and/or accumulation of precautionary balances against volatility in large gross asset and liability positions.
- Exporters of tourism services
  - Often have chronically high CA deficits and high debt levels.
  - Financing of tourism-related infrastructure can attract sizeable capital flows, notably FDI, which lowers the CA balance due to large import content of FDI.
- Recipients of aid and remittances
  - Tend to have higher-than-average CA deficits.
  - Aid inflows may induce capital-intensive investment.
  - Remittances can weaken the CA if they boost domestic demand enough to raise domestic costs and reduce competitiveness; however, empirical evidence for persistent RER appreciation from aid/remittances is weak.

### Key statistics (Table 2 summary)
- Nominal GDP (billions of USD), Average Per Capita Income (USD), Population (millions), CAB (in percent of GDP), Fiscal Balance (in percent of GDP), NFA
  - Exporters of non-renewable comm.: 117.7; 10,003; 22.3; 7.1; 4.5; −15.2
  - Exporters of financial services: 256.4; 29,027; 7.0; −0.5; −2.6; 33.6
  - Exporters of tourism services: 4.8; 8,404; 40.6; −13.3; −4.6; −92.0
  - Recipients of remittances: 7.7; 2,540; 3.6; −6.7; −2.8; −65.0
  - Recipients of aid: 3.1; 1,090; 6.2; −7.4; −1.0; −53.4
  - All countries: 340.1; 10,719; 37.5; 0.1; −1.0; −30.5

### Adjustments to standard methodologies and associated risks
- Common adjustments: modifying the panel country sample, explanatory variables, and/or econometric methodology.
- Examples of methodological/customization choices by staff:
  - Modifying sample or explanatory variables for MB or ERER approaches.
  - Modifying ES methodology to ensure intergenerational equity.
  - Using different cointegration techniques for ERER estimation.
- Risk: Adjustments may underestimate external gaps. Specifically:
  - Estimating regressions on subsamples of apparently similar economies can reduce estimated CA gaps or exchange rate misalignments—potentially reflecting bias from sample selection rather than better fit.
  - Conducting assessments on subsamples implicitly assumes the group is on average at an appropriate, non-distorted position; this may fail to account for a group-level misalignment relative to the rest of the world.
- Empirical illustration (Table 3: Average Directional Difference in Exchange Rate Misalignment Estimates; Full Sample - Subsample; in percentage points)
  - MB Approach / ERER Approach
    - Exporters of nonrenewable commodities: 11.3 / −1.9
    - Exporters of financial services: 4.2 / 2.1
    - Exporters of tourism services: 14.3 / 9.7
    - Recipients of aid: 6.2 / 7.6
    - Recipients of remittances: 6.5 / 2.0
    - Average: 8.6 / 4.5
  - Example interpretation: For exporters of tourism services, using the full sample (184 countries) without customization yields an exchange rate gap estimate 14.3 percentage points further from zero than an estimate based on a subsample of 19 tourism exporters (MB approach).

### Good practice for external assessments
- Transparency
  - Disclose adjustments, economic rationale, and implications when changing CGER methodologies or correcting for one-off factors, measurement biases, or other relevant factors.
  - If customized methodologies are used, provide clear discussion on how assessments differ from standard methods.
- Coverage
  - Move beyond quantitative point estimates of external gaps; include broader indicators: capital flows, reserve adequacy, foreign exchange intervention, capital controls, external assets and liabilities.
  - Include relevant structural indicators of competitiveness, especially for countries with severe data limitations.
  - Acknowledge uncertainties stemming from econometric methodologies and the valuation of some fundamentals (e.g., prospective reserves for oil exporters).

### Toolkit for Non-EBA Countries and alternative indicators
- An internal toolkit has been developed to apply CGER and EBA analytical approaches (MB, ERER, ES) to a larger set of economies.
  - The toolkit:
    - Implements variants of PPP, MB, ERER, and ES approaches.
    - Uses an annual panel dataset and econometric programs spanning the 184 economies covered by the World Economic Outlook database from 1973 to present.
    - Reports exchange rate misalignment estimates together with intermediate results (panel regression coefficients, multilateral consistency adjustments).
    - Includes measures of aid and remittance inflows for emerging and developing economies.
  - Note: The CGER methodology and toolkit generally do not take account of policy distortions emphasized by EBA.
- When quantitative estimation is constrained, use qualitative and group-specific alternative indicators:
  - A competitiveness scorecard and panel charts for external stability can supplement exchange rate assessments.
  - Competitiveness scorecard components: relative prices (internal terms of trade, CPI-REER over last five years), export indicators (export volumes and global market shares), production costs (internet users per 100 people, gasoline and diesel prices), institutional quality (Doing Business ranking, Corruption Perception Index).
  - Panels for external stability examine CA balance, NFA position, international reserves, tourist arrivals (for tourism-dependent economies), CPI-REER, FDI inflows, and other group-relevant time series.
  - Templates can be customized for structural characteristics and data availability; missing indicators can be proxied (e.g., external debt for NFA).

### Evidence on structural heterogeneity and model adjustments
- Staff developed an extension to the toolkit to account for differences across country groups with concentrated external income sources.
- Empirical exercise beginning with CGER regression specifications examined differences in slope and intercept values and selected modified regression specifications; results indicate value in modeling structural differences.
- Findings on slope coefficient differences (no significant intercept differences):
  - Nonrenewable commodity exporters
    - Higher slope coefficients on the fiscal balance (MB regression) and on the terms of trade and government consumption (ERER regression).
    - Rationale: fiscal balance dominated by oil revenue swings correlates strongly with CA; exports are less diversified, exposing countries to large terms of trade fluctuations that feed into domestic prices and RER; fiscal transfers from oil revenues tend to raise domestic prices and appreciate the RER.
    - Consistent with prior findings for oil exporters.
  - Aid-dependent economies
    - Higher slope coefficients on the oil trade balance and aid inflows in the MB regression.
    - Rationale: high vulnerability to terms of trade fluctuations and high sensitivity to aid inflows given limited intertemporal consumption smoothing due to low savings and credit constraints.

*Italic: Source — Appendix 2, more than 80 economies (content unit: _dp1401 - Appendix 2, more than 80 economies).*

### Appendix 4). The sign of  the coeffi cient on

### _dp1401 - Appendix 4). The sign of  the coeffi cient on

### Allowing for country-group heterogeneity in exchange rate regressions
- The sign and magnitude of coefficients on aid inflows depend on whether inflows are official grants included in the CA or loans that are excluded.
- Financial centers: In the MB regression, the slope coefficient on relative income is higher for financial centers, possibly reflecting high earnings.
- Remittance-dependent economies: The slope coefficient on relative productivity in the ERER regression is higher for remittance-dependent economies; remittances may support a higher relative price of nontraded goods than relative productivity alone would imply, raising domestic prices and appreciating the RER.
- Allowing slope heterogeneity across country groups can materially affect exchange rate assessments; when slope coefficients are not allowed to differ across country groups, regression residuals may be larger and MB- and ERER-based estimates may overstate the exchange rate gap.

### Quantified impacts from Table 4 (Average Directional Difference of Estimated Exchange Rate Gaps; Full sample, with no customization minus Full sample, with customization, in percentage points)
- MB (REER) / ERER (REER)
- Exporters of nonrenewable commodities: 5.8 / 5.6
- Exporters of financial services: 1.7 / 2.6
- Exporters of tourism services: 1.1 / 2.5
- Recipients of aid: 2.5 / 1.9
- Recipients of remittances: 0.5 / 0.7
- Average (Customized groups): 0.9 / 2.7
- Average (All other countries): 2.3 / 0.1

- Example interpretation: for nonrenewable commodity exporters, average estimated exchange rate gaps for both MB and ERER approaches are about 6 percentage points further from zero in the absence of slope heterogeneity.

### Trade-offs: sample selection versus coefficient heterogeneity
- Restricting the sample to structurally similar countries may understate the magnitude of misalignment.
- Not allowing coefficients to vary across structurally different economies may increase residuals and overstate misalignment.
- The meaningfulness of either issue depends on whether the common effect in the subsample reflects an undesirable distortion or an omitted structural factor that should be controlled for.
- Recommended practice: prefer a wide country sample and either measure/model omitted structural characteristics directly or employ a dummy variable for the country group; adjustments must be accompanied by plausible hypotheses and grounded in economic theory.
- Continuous interaction terms (e.g., interacting terms of trade with size of net oil exports) are an alternative to binary dummies (noted as used in the EBA methodology).

### Data issues and supplemental current account measures
- Large foreign ownership in natural resource or financial sectors can generate external flows that dwarf domestic-sector flows, complicating estimation of exchange rate elasticities and assessments of competitiveness.
- Standard CA statistics can distort measured external wealth accumulation when income is generated in one location but ultimately owned elsewhere (example: FDI profits retained in host country push measured CA above actual net external wealth change).
- Incomplete data reporting (aggregated total investment income without breakdowns by functional category or instrument) hinders correct accounting of reinvested earnings in CA estimates and complicates cross-country comparisons.
- Example: Swiss National Bank analysis suggests measurement issues could add roughly 3½ percent of GDP per year to Switzerland’s measured CA surplus.

### Separating the foreign-owned resource sector (case study approach)
- Rationale: where foreign-owned resource sector is not well integrated, resource-related flows can dominate external flows and mask the non–resource sector’s competitiveness and fundamentals.
- Recommended separation (subject to data availability):
  - Economic activity: define domestic component to include all non–resource-related activities, public revenues, private wages, and profits accrued by domestic agents in the resource sector.
  - External balance: separate commodity exports and related capital imports, and income payments to foreign investors in the resource sector, from the CA balance; treat resource-related fiscal revenue as remittances/grants accrued by the domestic non-resource sectors.
- Illustrative result: in Appendix 6 example, overall-economy fundamentals point to exchange rate undervaluation, while the adjusted (domestic) economy suggests overvaluation—highlighting potential divergence in policy advice and inconsistent projections for non-resource sector output growth.
- Limitations: such analysis requires significant data and strong assumptions (particularly for resource-related imports, FDI, and income flows); enhanced data provision by members would be needed.

### Conclusions and implications for staff practice
- IMF country teams often adjust methodologies for economies not included in EBA and CGER samples by: (1) adding relevant omitted variables, (2) using coefficient estimates based on a subset of similar countries, or (3) changing regression specifications.
- Trade-offs: excessive customization risks multilateral inconsistency and statistical overfitting, potentially understating CA or exchange rate gaps; subsample estimation implicitly assumes the group is not misaligned with the rest of the world.
- Practical recommendation: when using customized methodologies, provide a clear discussion of how assessments may differ from standard methods.
- An internally developed exchange rate assessment toolkit can:
  - Facilitate customized external assessments in a multilaterally consistent way.
  - Allow accounting for heterogeneity across groups of structurally similar countries within a global panel regression framework (including different slope coefficients via interaction terms).
  - Provide a template to analyze information relevant for countries with concentrated external income sources.

### Additional data improvements suggested
- Correct for distortions related to investment income.
- Focus more specifically on domestic competitiveness by netting out flows to large foreign-owned sectors.
- Staff intends to continue refining tools in the context of the EBA approach to external assessment and to adapt methods for countries outside the EBA sample; many issues discussed remain relevant.

*Source: IMF staff calculations as presented in the provided content unit.*

### Appendix 3 .  Survey of Exchange Rate Assessments

### Appendix 3 .  Survey of Exchange Rate Assessments

### Overview
- Nearly 95 percent of the 88 reviewed reports for countries with special characteristics use CGER or adjusted methods for exchange rate assessments.
- Appendix provides details on typical customizations across special groups, how customization has been done, likely effects on assessments, and lessons for future practice.

### Customization Across Special Country Groups: How Has It Been Done?

- General note: All three CGER-based approaches (MB, ERER, ES) have been customized for various special country groups.

- Exporters of Non-renewable Commodities
  - MB approach
    - Customization occurs at two stages: the CA regressions and the elasticity of the CA to the REER.
    - CA regressions
      - Standard CGER regressions may fail to capture oil-exporting specificities: effects of oil wealth, degree of maturity in oil production, and intergenerational equity concerns.
      - Expanded set of fundamental determinants includes oil wealth and degree of maturity of oil production.
      - Non-oil fiscal balance is considered the most relevant fiscal variable to separate effects of oil revenues and fiscal policy on the CA; focus on sustainability of the non-oil CA.
    - Derivation of CA elasticity to the REER
      - For oil exporters, GDP shares of oil and non-oil exports and elasticities of non-oil exports and imports to the REER are explicitly accounted for.
    - Examples in Article IVs
      - MB approach uses CA regressions from Bems and de Carvalho Filho (2009) based on a sample of 24 oil-/gas-exporting countries for Middle Eastern and Central Asia Department reports (e.g., Algeria, Bahrain, Kuwait, Oman, Saudi Arabia).
      - For Mongolia, nonfuel mining products are treated as part of the oil balance in the MB approach.
  - ERER approach
    - Customized regressions use explanatory variables not necessarily matching CGER’s.
    - Two frameworks commonly used:
      - Multivariate relationship: REER regressed on relative government consumption, relative productivity, and terms of trade; additional regressors include openness and investment income.
      - Bivariate relationship: REER and real oil prices using high-frequency data (used for Bahrain, Kuwait, Oman, Saudi Arabia).
  - ES approach
    - Methodology change: instead of finding a CA that stabilizes NFAs at a benchmark, derive a path for future NFA satisfying an intertemporal income-allocation rule for non-renewable resources.
    - Three allocation rules used: constant real annuity, constant real per capita annuity, and constant annuity as a share of GDP.
    - Modified ES used for Middle Eastern and Central Asia Department countries, Norway, Gabon, and the Republic of Congo.
  - Frameworks and model-based approaches
    - IMF staff frameworks apply model-based approaches with optimal private and public investment decisions and frictions: absorptive capacity constraints, investment inefficiencies, borrowing constraints relaxed by resource windfalls.
    - Alternative framework can yield short- and medium-run CA benchmarks lower (higher deficits) than MB and ES norms when calibrated (example: Ghana).
    - Model-based approach calibrated to Ghana yielded short- and medium-run CA benchmarks that were lower (higher deficits) than MB and ES norms (see Table A3.1).

- Tourism-Dependent Economies
  - MB approach
    - Customized regressions include FDI inflows as a determinant of equilibrium CA balance.
    - Adjustments in some cases:
      - Underlying CA adjusted by excluding one-off imports and/or other exceptional temporary factors (Mauritius, Seychelles).
      - Extensions include aid and remittance inflows as explanatory variables and indicators of armed conflicts and extraordinary oil production where relevant (used for Cape Verde and other sub-Saharan Africa countries).
  - ES approach
    - Customized to account for capital account transfers including grants (assumed zero in standard CGER), assumptions on foreign assets and liabilities to derive a range of NFA benchmarks (Seychelles).
    - For some countries (St. Lucia, St. Kitts and Nevis), assessment based on IMF’s external debt sustainability assessment when net international investment position data are absent.

- Aid/Remittance-dependent Countries
  - Exchange rate assessments primarily use regressions augmenting standard CGER variables with aid and/or remittances (Vitek (2009) widely used).
  - Some reports use cointegration-based analyses (Chudik and Mongardini (2007); Mongardini and Rayner (2009)); Malawi reports reference Mongardini and Rayner (2009).

- Financial Centers
  - Most external assessments for financial centers do not make adjustments.
  - Exceptions:
    - Hong Kong SAR (2011 and 2012): estimate CA and exchange rate gaps using various subsamples that eliminate gaps found in standard CGER exercise.
    - Switzerland: reduce estimated CA surplus to account for investment income attributed to Swiss residents that actually accrues to nonresident owners of subsidiaries of Swiss multinationals; this reduces but does not eliminate the estimated positive CA gap and undervaluation.

### Empirical Example: Ghana (Table A3.1)
- MB | ES | Model
  - Short-run (2012)
    - CA projection–12.2%–12.2%–12.2%
    - CA benchmark–7.7%–2.8%–9.8%
    - CA gap–4.4%–9.4%–2.4%
  - Medium-run (2017)
    - CA projection–8.6%–8.6%–8.6%
    - CA benchmark–3.5%–2.8%–4.0%
    - CA gap–5.1%–5.8%–4.6%
- Sources: Staff projections and calibration based on Araujo and others (2012).
- Note: CA = current account; ES = external sustainability; MB = macrobalance.

### How Is Customization Likely to Change Results vs. Standard CGER and Why?
- Customization can explain developments that might otherwise raise concerns; direction of change versus standard CGER is specific to country groups and country developments.
- Typical effects by group:
  - Oil exporters / non-renewable commodity exporters
    - Customization can address sustainability given resource depletion and capacity development needs.
    - Can justify:
      - Large CA surpluses for extended periods to accumulate NFA for smoothing consumption after resource depletion.
      - REER strengthening when real world price of the export commodity increases (bivariate REER–commodity price relationship); raises question whether such appreciation is desirable.
      - Weaker CA balances in resource-rich LICs that invest windfalls in domestic productive capacity, reducing external savings temporarily.
  - Aid-dependent and tourism-dependent economies
    - Customization may justify large absorption and resulting CA weakening because aid and inward FDI finance public investment and private tourism projects that raise absorption.
    - Calibrated model simulations show full spending and absorption of aid can cause temporary RER appreciations and medium-term positive output effects via higher public capital.
  - Remittance-dependent economies
    - Viewing remittances as exhaustible over the long term may highlight need to save some remittances, justifying a relatively stronger CA.
    - Use of PIH-like rules for intertemporal allocation applied (example: Philippines Article IV 2011).

### Lessons and Best Practices
- Practices that improve transparency and quality of external rate assessments:
  - Provide information on the customization.
  - Document steps performed to arrive at REER misalignment estimates, including whether multiyear averages or annual values of explanatory variables are used and the time horizon for computing the CA norm or equilibrium REER.
  - Document the contribution of each variable to the CA norm or “equilibrium” REER to inform policy discussions.
  - Exclude large one-off factors from the CA level whose appropriateness is being assessed.
  - Present other indicators of competitiveness (structural indicators like cost of doing business; economic indicators like market shares or growth of nontraditional exports).
- Practices that may undermine reliability and should be avoided if possible:
  - Apply coefficients from older regressions, especially if regressions did not include the country being assessed.
  - Rely on one empirical methodology and/or overlook structural competitiveness indicators that may matter more for CA dynamics than the ER, particularly for LICs.
  - Use non-transparent derivations of CGER or CGER-like estimates.
  - Base ER assessments on a very limited sample of countries.

*Source: Appendix 3 .  Survey of Exchange Rate Assessments (from the supplied content unit).*

### Appendix 2 for benchmark country groups). In particular, we consider

### _dp1401 - Appendix 2 for benchmark country groups). In particular, we consider

### Parameter sensitivity and robustness of coefficient estimates
- Perturbations in classification thresholds required to change the number of countries in each group by plus or minus 20 percent do not materially affect coefficient estimates and exchange rate gaps.
- The parameter estimates under the MB and ERER approaches are robust to outliers: estimated coefficients remain stable when dropping individual countries from the sample of 184 countries one at a time.
- Figures A4.2 and A4.3 show 95 percent confidence intervals around estimates and demonstrate robustness to leaving-one-out samples (horizontal axis lists 184 countries; vertical axis shows estimated coefficients).

### Macroeconomic Balance (MB) approach — panel regression estimation results (selected)
- Final model coefficients and significance (standard errors in parentheses). Symbols ***, **, and * denote significance at 1%, 5%, and 10%, respectively.
  - C: −0.006 (0.009)
  - C for exporters of tourism services: −0.083*** (0.018)
  - Old age dependency: −0.010 (0.089)
  - Population growth: −0.326 (0.392)
  - Relative income: 0.000 (0.006)
  - Relative income for exporters of financial services: 0.027** (0.012)
  - Output growth: −0.944*** (0.292)
  - Oil trade balance: 0.152*** (0.057)
  - Oil trade balance for recipients of aid: 0.589** (0.302)
  - Fiscal balance: 0.137 (0.134)
  - Fiscal balance for exporters of nonrenewable commodities: 0.541** (0.231)
  - NFA: 0.030*** (0.009)
  - Grant/aid: −0.301* (0.180)
  - Grant/aid for recipients of aid: −0.386*** (0.160)
- Adjusted R-squared values by sample:
  - Benchmark: 0.393
  - Aid Recipients: 0.382
  - Exporters of Nonrenewable Commodities: 0.444
  - Financial Centers: 0.397
  - Tourism-dependent Countries: 0.422
  - Final Model: 0.436
- Number of instruments reported across sample columns: 9, 11, 10, 10, 10, 13.

### Equilibrium Real Exchange Rate (ERER) approach — panel regression estimation results (selected)
- Final model coefficients and significance (standard errors in parentheses). Symbols ***, **, and * denote significance at 1%, 5%, and 10%, respectively.
  - C: 0.115*** (0.040) [Final Model]
  - Terms of trade: 0.121 (0.080) [Final Model]
  - Terms of trade for exporters of nonrenewable commodities: 0.396** (0.192)
  - Relative productivity: 0.126 (0.135) [Final Model]
  - Relative productivity for recipients of remittances: 0.393* (0.209)
  - Relative government consumption: 1.407*** (0.417) [Final Model]
  - Relative government consumption exporters of nonrenewable commodities: 10.575*** (2.543)
  - Aid flows: 0.816 (0.695) [Final Model]
  - Remittance flows: 0.202 (0.280) [Final Model]
- Goodness-of-fit and observations:
  - Adjusted R-squared: 0.421 (Benchmark/Final Model), 0.307 (Exporters of Nonrenewable Commodities), 0.421 (Remittance Recipients), 0.306 (Final Model other)
  - Observations: 1951 across reported models.

### Comparisons of MB and ERER misalignment estimates
- Figure A4.1 regression summaries (selected lines and R²):
  - Customized vs. Benchmark (MB/ERER comparisons) exhibit high correlation in several groupings:
    - Example: Exporters of Nonrenewable Commodities (MB Approach): y = 0.793 x –2.8175, R² = 0.8422
    - Tourism-Dependent Countries (MB Approach): y = 0.933x + 1.4934, R² = 0.9739
    - Aid-Dependent Countries (MB Approach): y = 0.7171x + 9.5397, R² = 0.2366
    - Financial Centers (MB Approach): y = 0.9534 x –2.3172, R² = 0.9608
    - Exporters of Nonrenewable Commodities (ERER Approach): y = 0.8098 x –12.918, R² = 0.6669
    - Remittance-Dependent Countries (ERER Approach): y = 0.6753 x –8.1221, R² = 0.5863

### Financial centers — defining features and implications for external assessment
- Financial centers typically exhibit:
  - Large financial services exports, FDI and investment flows, and/or net foreign assets (NFA).
  - Sizable positive NFA positions and relatively high investment income leading to large current account (CA) surpluses (average, 1980–2011 shown in Figure A5.1).
- Table A5.1 selection criteria and list (data as of 2010):
  - External assets > 350% of GDP
  - Financial Service Exports: Exports > 1% and Financial Service Exports: Output > 1%
  - Countries listed include Bahrain, Barbados, Belgium, Cyprus, Hong Kong SAR, Ireland, Lebanon, Luxembourg, Malta, Panama, Netherlands, Singapore, Sweden, Switzerland, United Kingdom, Vanuatu (columns correspond to different criteria).
- Empirical observations and adjustments:
  - Country teams adjusted external assessments for financial centers in a few cases; of 13 countries listed under Column 2 in Table A5.1, 2 had adjusted external assessments: Hong Kong SAR and Switzerland.
  - Swiss CA may be overstated by accounting treatment of retained earnings; SNB estimates this effect increases Switzerland’s CA by roughly 3.5 percent of GDP per year.
  - Foreign assets and liabilities (gross positions) can be much larger than CA balances and may better reflect exchange rate determination and external stability (example: United Kingdom’s gross external liabilities > 550 percent of GDP in 2012 while CA was −3½ percent of GDP).
- Exchange rate gap estimates for financial centers (Table A5.2), real effective exchange rate over/undervaluation ranges (in percent) derived from staff reports:
  - Switzerland: Undervalued [(–15.2)–6.1]
  - Singapore: [(–8)–(–1)]
  - Belgium: Equilibrium [(–1)–6]
  - Ireland: [(–3.9)–6]
  - Luxembourg: [2.05–8.7]
  - Panama: [1–3]
  - United Kingdom: [(–4)–0]
  - Hong Kong SAR: [(–10)–6]
  - Barbados: Overvalue [5–10]
  - Cyprus: [11.86–21.23]

### Considerations for use of quantitative tools with financial centers
- Quantitative measures tend to exhibit wider ranges for financial centers; regressions restricted to subsamples may introduce biases.
- Additional emphasis should be placed on analysis of gross foreign assets and liabilities and valuation effects to supplement CA-based assessments.
- Net external liabilities (primarily net debt) have a strong effect on probability of external crisis per Catão and Milesi-Ferretti (forthcoming).

### Illustrative example: separation of foreign-owned resource sector (Appendix 6)
- Purpose: Separate domestic economy from foreign-owned resource sector to estimate CA norm for adjusted domestic economy using CGER MB approach; highlights potential Dutch disease effects.
- Method:
  - Output: adjusted economy excludes value added of the resource sector.
  - Fiscal balance: excludes resource revenues and royalties; these treated as remittances in adjusted external account.
  - CA balance: receipts from commodity exports, corresponding capital imports, and income payments to external investors treated as “foreign resource–related balance.”
  - Assumptions where detailed data unavailable: income payments account for 40 percent of resource export revenues; resource-related imports amount to 40 percent of FDI inflows, declining over time as sector matures.
- Key comparative results (Table A6.1) — CA Norm, Underlying CA, CA Gap (percent of GDP) for Overall Economy vs. Adjusted Economy:
  - 2011: Overall Economy CA Norm 4.97; Underlying CA 7.62; CA Gap 2.7. Adjusted Economy CA Norm −4.7; Underlying CA −10.6; CA Gap −5.9.
  - 2012: Overall: CA Norm 5.16; Underlying CA 6.6; CA Gap 1.5. Adjusted: CA Norm −8.8; Underlying CA −12.3; CA Gap −3.5.
  - 2013: Overall: CA Norm 4.85; Underlying CA 5.6; CA Gap 0.7. Adjusted: CA Norm −9.8; Underlying CA −10.4; CA Gap −0.6.
  - 2014: Overall: CA Norm 3.7; Underlying CA 4.0; CA Gap 0.3. Adjusted: CA Norm −10.4; Underlying CA −7.8; CA Gap 2.5.
  - 2015: Overall: CA Norm 3.13; Underlying CA 3.3; CA Gap 0.2. Adjusted: CA Norm −9.4; Underlying CA −5.8; CA Gap 3.6.
  - 2016: Overall: CA Norm 2.42; Underlying CA 2.8; CA Gap 0.4. Adjusted: CA Norm −7.2; Underlying CA −4.2; CA Gap 3.0.
  - 2017: Overall: CA Norm 2.0; Underlying CA 2.9; CA Gap 0.9. Adjusted: CA Norm −7.0; Underlying CA −2.8; CA Gap 4.2.
- Findings and implications:
  - The adjusted domestic economy exhibits significant divergence from aggregate assessments, often showing negative and significant CA gaps for 2011–2013 indicating overvaluation and potential Dutch disease.
  - Projections that assume buoyant medium-term output growth for the adjusted domestic economy may be implausible given low competitiveness; CA gap expected to change sign in 2013 or 2014 and widen toward undervaluation over the medium term.
  - Separating the foreign-owned resource sector helps identify domestic policy gaps and serves as a consistency check for projections.

*Source: IMF staff calculations and content from the supplied Appendix excerpts.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/dp/2014/_dp1401.pdf_