## _wp1043 - Appendix I. Sample Characteristics.

## Source details

**Canonical URL:** [_wp1043 - Appendix I. Sample Characteristics.](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1043.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1043.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1043.pdf.json)

---

### I. INTRODUCTION
- The real exchange rate is a key relative price signaling production choices and structural adjustment in the medium term.
- Methodologies to estimate equilibrium real exchange rates are broadly classified as:
  - bivariate (e.g., absolute and relative purchasing power parity),
  - multivariate (e.g., macro-balance and external sustainability approaches).
- The Consultative Group on Exchange Rate Issues (CGER) uses an equilibrium notion based on internal and external balance:
  - Internal balance: zero output gap.
  - External balance: an “equilibrium” current account balance estimated as a function of fundamentals and a sustainable net foreign asset or liability position.
- Multivariate approaches include:
  - (i) quantity-based (macro-balance): medium-term current account balances as a function of medium-term fundamentals;
  - (ii) price-based (fundamental approach): reduced-form equilibrium real exchange rate as function of fundamentals like sectoral productivity differentials, government size, terms of trade shocks;
  - (iii) balance sheet-based (external sustainability): equilibrium current account consistent with a benchmark net foreign asset position.
- This paper applies the macro-balance approach exclusively to emerging market economies with two contributions:
  - By using only emerging markets, it abstracts from developed-economy influences on current account norms, providing a robustness check versus CGER estimates.
  - Provides estimates for eleven smaller emerging market economies not included in the CGER country sample.
- Paper organization: Section II — variable choice, sample description, main determinants; Section III — regression estimates and comparison with CGER; Section IV — derive current account norms and compare with underlying projections; Section V — conclusion.

### II. SAMPLE AND MAIN REGRESSORS
- Sample definition and coverage:
  - Emerging economies that have placed sovereign bonds in international financial markets at least once since 1985.
  - Sovereign risk ratings range from single A to B3 as of June 2009.
  - Resulting group: 33 countries.
  - GDP per capita in 2008 U.S. Dollars ranges from US$1,000 (India and Pakistan) to more than US$28,000 (Israel).
  - About two thirds of the sample are also part of the CGER sample.
- Model specification:
  - Linear reduced form model: (1) CA = Xβ + ξ
    - CA: yearly current account balances in percent of GDP.
    - X: fundamentals including fiscal balance, young and old dependency ratios, population growth, lagged NFA, oil balance, GDP per capita growth, per capita income relative to the United States, and FDI.
    - ξ: standard error terms, assumed independent and normally distributed.
- Expected signs on regressors:
  - Fiscal balance: negative effect on current account balance (larger fiscal deficit → lower CA).
  - Young dependency ratio: expected negative impact on CA.
  - Old dependency ratio: expected negative impact on CA.
  - Population growth: expected negative relationship with CA.
  - Net foreign assets (NFA, one-period lag): ambiguous a priori.
  - Oil balance: increases CA for oil-exporters; decreases CA for oil-importers.
  - GDP per capita growth: expected negative effect on CA.
  - Per capita income relative to the United States: expected negative relationship with CA.
  - Ratio of FDI to output: expected negative coefficient (FDI-financed deficits more sustainable).
  - Asian crisis dummy (Indonesia, Republic of Korea, Malaysia, Philippines, Thailand): equals one after 1997 to proxy structural increase in savings after the 1997 crisis.
- Sample period and data:
  - Annual data, 1970 to 2008.
  - Uses October 2009 World Economic Outlook data, including oil price projections.

### III. ESTIMATION APPROACH
- Stationarity and method:
  - Standard pooled OLS used, conditional on stationarity of dependent and explanatory variables.
  - Unit root tests indicate most variables are stationary, with a couple deviations; demographic variables may trend but are bounded.
  - Regressions include time dummies and exclude country fixed effects:
    - Time dummies isolate year-specific effects and net out cyclical factors.
    - Country fixed effects omitted because chosen determinants are intended to distinguish current account fluctuations across countries.
- Main regression findings (preferred specification = Column (3), baseline):
  - Oil balance:
    - Coefficient statistically significant in all cases.
    - Reported coefficients: 0.223*** (All), 0.358*** (All minus oil exporters), 0.359*** (Baseline), 0.23*** (All), 0.19*** (CGER).
    - Standard errors for first three: (0.035), (0.075), (0.071).
  - Fiscal balance:
    - Coefficients around 0.482***, 0.526***, 0.492***, 0.492***; CGER: 0.19***.
    - Standard errors: (0.039), (0.044), (0.040), (0.040).
    - Interpretation: stronger fiscal balance effect in this emerging market sample than CGER.
  - Demography:
    - Young-age dependency ratio consistently significant.
    - Young-age dependency coefficients: -0.0556**, -0.0395, -0.0580***, -0.0520** with standard errors: (0.022), (0.024), (0.022), (0.022).
    - Old-age dependency not consistently significant.
  - Net foreign assets (NFA):
    - NFA coefficients: 0.0631***, 0.0595***, 0.0658***, 0.0630***, CGER: 0.02***.
    - Standard errors: (0.007), (0.008), (0.007), (0.007).
    - Column (3) implication: every 10 percentage point of GDP increase in NFA → current account balance improves by 0.578 percent of GDP (reported text).
  - GDP per capita (output) growth:
    - Output growth coefficients: -0.184***, -0.194***, -0.184***, -0.187***; CGER: -0.21**.
    - Standard errors: (0.039), (0.042), (0.039), (0.032).
    - Column (3) estimate around -0.2 (text truncated on exact interpretation).
  - FDI:
    - FDI coefficients: -0.159, -0.147, -0.199*, -0.148 with standard errors: (0.101), (0.107), (0.102), (0.101).
  - Population growth:
    - Coefficients: -0.0695, 0.235, -0.0741, -0.0286, CGER: -1.22**; standard errors: (0.241), (0.279), (0.244), (0.242).
  - Relative income:
    - Coefficients: -0.00554, -0.00502, -0.00567, -0.00391, CGER: 0.02*; standard errors: (0.003), (0.003), (0.003), (0.003).
  - Asian crisis dummy:
    - Coefficients: 0.0468***, 0.0457***, 0.0447***, 0.0468***, CGER: 0.06***; standard errors: (0.010), (0.012), (0.011), (0.011).
  - Differential oil balance (D_Oilbalance_OilX):
    - Coefficients: 0.199***, -0.196** with standard errors: (0.043), (0.089).
- Sample and fit:
  - Observations: 1044 (All), 921 (All minus oil exporters), 1044 (Baseline), 1044 (All), 1044 (CGER column).
  - Number of countries: 33, 29, 33, 33, 54.
  - R-squared: 0.379, 0.352, 0.366, 0.382, 0.52.
  - Significance codes: *** p<0.01, ** p<0.05, * p<0.1.
- Robustness:
  - Regression estimations robust to different specifications; alternative specifications available on request.

### IV. CURRENT ACCOUNT NORMS AND PROJECTIONS
- Methodology:
  - Equilibrium assumed at 5 years into the future (i.e. 2014), consistent with IMF World Economic Outlook forecast.
  - Current account norms computed using projected values for relevant regressors in 2014.
  - Demographic projections use U.S. Census Bureau International Database growth rates.
- Aggregate results using baseline model:
  - About two thirds of equilibrium current account balances are in deficit.
  - Deficit norms range from minus 0.1 percent of GDP (Uruguay) to over minus 9 percent of GDP (Lebanon).
  - Surplus norms range from 0.2 percent of GDP (Ecuador) to over 17 percent of GDP (Algeria).
- Comparison with CGER:
  - Asian crisis countries on average have current account norms more than 4 percentage points of GDP higher than other countries in the sample (column (3) of Table 1).
  - CGER yields a positive effect of about 6 percent of GDP for Asian crisis countries but assumes the effect dies out over time.
  - For non-Asian crisis countries with CGER estimates available, results are similar.

### V. SELECTED COUNTRY-LEVEL BASELINE NORMS (baseline vs CGER where available)
- Selected entries (Norms in percent of GDP):
  - Costa Rica: Baseline -2.2, CGER n.a.
  - Dominican Republic: -4.0, n.a.
  - Ecuador: 0.2, n.a.
  - El Salvador: -2.7, n.a.
  - Guatemala: -3.4, n.a.
  - Jordan: -7.9, n.a.
  - Lebanon: -9.2, n.a.
  - Panama: -5.8, n.a.
  - Paraguay: 0.8, n.a.
  - Sri Lanka: -1.4, n.a.
  - Uruguay: -0.1, n.a.
  - Algeria: 17.2, 14.2
  - Argentina: 1.9, -0.6
  - Brazil: -0.5, -0.4
  - Chile: 1.3, -2.8
  - China: 5.6, 2.9
  - India: -2.6, -3.3
  - Indonesia: 5.4, 0.0
  - Korea: 8.3, 0.4
  - Malaysia: 8.0, 0.7
  - Mexico: -2.8, -0.8
  - Philippines: 5.2, -0.1
  - Thailand: 6.5, -2.9
  - Venezuela: 5.6, 6.6
- Note: 1/ Estimated by using methodology and coefficients from Lee et al. (2008).

### VI. DECOMPOSITION OF DRIVERS FOR SELECTED COUNTRIES (percent of GDP)
- Algeria (Norm 17.2)
  - Fiscal balance: 0.8
  - Age dependency (young): -2.0
  - NFA: 7.5
  - Oil trade balance * Oil producers: 6.3
- Malaysia (Norm 8.0)
  - Fiscal balance: -3.2
  - Age dependency (young): -2.5
  - NFA: 4.4
  - Asian crisis: 4.5
- Guatemala (Norm -3.4)
  - Fiscal balance: -0.9
  - Age dependency (young): -3.8
  - NFA: -3.2
- Jordan (Norm -7.9)
  - Fiscal balance: -2.7
  - Age dependency (young): -2.6
  - NFA: -6.0
- Panama (Norm -5.8)
  - Fiscal balance: 0.1
  - Age dependency (young): -2.4
  - NFA: -5.8

### VII. CONFIDENCE INTERVALS, DEVIATIONS, AND OUT-OF-EQUILIBRIUM COUNTRIES
- Confidence bands and rule for disequilibrium:
  - Confidence bands constructed from standard errors of fitted values.
  - A four standard deviation band contains over 95 percent of possible observed sample estimates under normality and unbiasedness assumptions.
  - Rule for potential disequilibrium: underlying current account more than two standard deviations away from estimated norm may warrant a real exchange rate adjustment.
- Countries with underlying balances more than two standard deviations away from the norm (using 2014 projections): Algeria, Chile, Ecuador, Korea, Indonesia, Philippines, South Africa, Thailand.
- Cautions:
  - For natural-resource-rich countries, the macro-balance approach may be inappropriate because it may not capture stock of wealth from natural resources; results for Algeria, Ecuador, and Chile should be interpreted with caution.
  - If the Asian crisis dummy effect fades in the medium term, underlying balances for Korea, Indonesia, Philippines, and Thailand would not be statistically different from norms; Malaysia would fall outside the 95 percent confidence interval.
- Selected Table 4 baseline entries (Norm, SD, 2014 Projection; percent of GDP):
  - Colombia: Norm -0.4, SD 1.3, 2014 Proj. -2.1 (Included in 95 percent confidence interval)
  - El Salvador*: -2.7, 1.3, -2.7
  - Brazil: -0.5, 1.3, -0.8
  - India: -2.6, 1.4, -1.8
  - Turkey: -4.9, 1.4, -3.1
  - Chile: 1.3, 1.4, -2.2
  - Ecuador*: 0.2, 1.4, -3.2
  - South Africa: -3.4, 1.4, -7.5
  - Peru: -0.9, 1.5, -1.6
  - China: 5.6, 1.5, 8.4
  - Malaysia: 8.0, 1.5, 10.8
  - Indonesia: 5.4, 1.5, -1.0
  - Korea: 8.3, 1.5, 2.1
  - Argentina: 1.9, 1.5, 2.0
  - Thailand: 6.5, 1.6, 1.6
  - Philippines: 5.2, 1.6, 0.9
  - Panama*: -5.8, 1.6, -6.0
  - Venezuela: 5.6, 1.6, 5.7
  - Jordan*: -7.9, 1.6, -7.0
  - Algeria: 17.2, 1.8, 28.2
  - Lebanon*: -9.2, 2.0, -9.3
  - Uruguay*: -0.1, 2.8, -0.1
  - Legend: * Not included in CGER sample. 1/ Included in the 95 percent confidence interval. 2/ Included in the 99 percent confidence interval. 3/ Statistically different from norm at 99 confidence level.

### VIII. CONCLUSIONS AND POLICY IMPLICATIONS
- Key conclusions:
  - Fundamental determinants of equilibrium current account balances for 33 emerging market economies are similar to those identified by CGER, with notable differences:
    - Fiscal balance has a considerably stronger impact on current account norms in this emerging market sample.
    - Evidence supports a permanent shift toward higher savings in Asian crisis countries (Indonesia, Korea, Malaysia, Philippines, Thailand).
  - For eleven smaller emerging market economies not included in the CGER sample, almost all have underlying current account balances not statistically different from their norms, suggesting policies broadly consistent with external and domestic balance in the medium term.
- Caveats and recommended next steps:
  - Estimates subject to significant uncertainty; goodness of fit is less than 40 percent, reflecting large variation in current account balances across countries and time.
  - A common specification may not capture heterogeneity across diverse countries.
  - Results for natural-resource-rich countries should be interpreted with caution; alternative methodologies addressing resource wealth stocks are warranted.
  - The macro-balance approach should be complemented with other methods for a more rigorous assessment of policy appropriateness.

*Source: IMF staff estimations and analysis in the provided content unit.*

### Appendix I. Sample Characteristics. ....................................................................................

### _wp1043 - Appendix I. Sample Characteristics.

### I. INTRODUCTION
- The real exchange rate is a key relative price in the economy, signaling production choices and structural adjustment in the medium term.
- Methodologies to estimate equilibrium real exchange rates are broadly classified as:
  - bivariate (e.g., absolute and relative purchasing power parity),
  - multivariate (e.g., macro-balance and external sustainability approaches).
- The Consultative Group on Exchange Rate Issues (CGER) has conducted biannual exchange rate assessments since 1997, using a notion of equilibrium based on internal and external balance:
  - Internal balance: zero output gap.
  - External balance: an “equilibrium” current account balance estimated as a function of economic fundamentals and a sustainable net foreign asset or liability position.
- Multivariate approaches classified as:
  - (i) quantity-based (macro-balance): estimate medium-term current account balances as a function of medium-term fundamentals;
  - (ii) price-based (fundamental approach): reduced-form equilibrium real exchange rate as function of fundamentals like sectoral productivity differentials, government size, terms of trade shocks;
  - (iii) balance sheet-based (external sustainability): determine equilibrium current account consistent with a benchmark net foreign asset position.
- The CGER has focused mainly on the macro-balance and external sustainability approaches, and to a lesser extent the fundamental approach, using a sample of 55 developed and emerging economies and providing formal estimates for 27.
- This paper applies the macro-balance approach exclusively to emerging market economies with two contributions:
  - By using only emerging markets, it abstracts from developed-economy influences on current account norms, providing a robustness check versus CGER estimates.
  - It provides, for the first time, estimates for eleven smaller emerging market economies not included in the CGER country sample.
- Paper organization: Section II — variable choice, sample description, main determinants; Section III — regression estimates and comparison with CGER; Section IV — derive current account norms and compare with underlying projections; Section V — conclusion.

### II. SAMPLE AND MAIN REGRESSORS
- Sample definition:
  - Emerging economies that have placed sovereign bonds in international financial markets at least once since 1985.
  - Sovereign risk ratings range from single A to B3 as of June 2009.
  - Resulting group: 33 countries (see Appendix 1).
  - GDP per capita in 2008 U.S. Dollars ranges from US$1,000 (India and Pakistan) to more than US$28,000 (Israel).
  - About two thirds of the sample are also part of the CGER sample.
- Model specification:
  - Linear reduced form model: (1) CA = Xβ + ξ
    - CA: vector of yearly current account balances in percent of GDP for every country.
    - X: set of fundamental variables including fiscal balance, two dependency ratios (young and old), population growth, initial stock of foreign assets (NFA), oil balance, GDP per capita growth, per capita income relative to the United States, and FDI.
    - β: coefficient vector.
    - ξ: standard error terms, assumed independent and normally distributed.
- Fundamental variables and expected signs:
  - Fiscal balance: larger fiscal deficit reduces national savings and lowers current account balance (negative relationship).
  - Dependency ratios:
    - Young ratio (under 16) and old ratio (over 65), both over economically active population.
    - Higher share of elderly tends to reduce national savings and current account balance.
    - Larger share of young population should enhance future productivity growth and facilitate future repayment of current account deficits; both young and old dependency ratios expected to have negative impact on current account balance.
  - Population growth: higher population growth allows larger future workforce and higher current account deficits (implying a negative relationship between population growth and current account balance).
  - Net foreign assets (NFA): net holdings by domestic residents of foreign assets minus foreign liabilities; enters with a one-period lag.
    - Two opposing effects: higher NFA implies wealth enabling weaker current account balance; but also yields income returns implying more positive current account balance. Sign is an empirical question.
  - Oil balance: higher oil prices increase current account balance for oil-exporters and decrease it for oil-importers.
  - GDP per capita growth: fast-growing economies have higher income potential, allowing lower savings today; GDP per capita growth expected to have a negative effect on current account balance.
  - Per capita income relative to the United States: lower ratio indicates earlier development stage and lower savings–investment balance; expected negative relationship with current account.
  - Ratio of FDI to output: current account deficits financed by FDI are more sustainable, so higher FDI should be associated with weaker current account balances (negative coefficient).
  - Asian crisis dummy: for Indonesia, Republic of Korea, Malaysia, Philippines, and Thailand, a dummy equals one after 1997 to proxy a structural increase in savings after the 1997 crisis.
- Sample period and data:
  - Sample period: 1970 to 2008 (annual data).
  - Sample includes data from the October 2009 World Economic Outlook, including oil price projections at that moment.

### III. ESTIMATION
- Stationarity and econometric approach:
  - Standard analysis assumes stationarity of dependent and explanatory variables, allowing pooled ordinary least squares (OLS).
  - Unit root tests indicate most variables are stationary, with a couple deviations (see Appendix III); tests have low power and some variables (e.g., demographic variables) may be trending but are bounded.
  - Regressions include time dummies to account for the business cycle and exclude country-level fixed effects (see Appendix IV).
    - Time dummies isolate year-specific effects without computing five-year averages and net out cyclical factors.
    - Chosen determinants are intended to distinguish current account fluctuations across countries; thus country fixed effects are not included.
- Main regression results (summary of findings; detailed Table 1 referenced):
  - Oil balance:
    - Coefficient is statistically significant in all cases and ranges between 0.2 for oil exporters and almost 0.4 for oil importers.
    - Separate oil effects for importers and exporters yield an effect for exporters similar to CGER, but a higher effect for importers; preferred specification is Column (3) in Table 1.
  - Fiscal balance:
    - Coefficient highly significant and fairly stable around 0.5 across specifications.
    - This is more than twice the coefficient estimated by the CGER.
    - Stronger fiscal balance effect may reflect less developed financial markets in emerging economies, limiting efficient channeling of government savings to the private sector.
  - Demography:
    - Demography matters; of three demographic variables used, only the youth dependency ratio is consistently significant across all specifications.
    - Estimated coefficient for youth dependency ratio is around minus 0.06 in Column (3), the preferred specification.
    - This contrasts with CGER findings, which find population growth and old-age dependency ratio significant.
    - The effect of population growth may be captured indirectly by the young dependency ratio.
  - Net foreign assets (NFA):
    - NFA has a positive effect on current account norms.
    - Coefficient in Column (3) implies that for every 10 percentage point of GDP increase in NFA, the current account balance improves by 0.578 percent of GDP.
    - This effect is stronger than CGER’s estimate of a 0.2 percent of GDP improvement.
  - GDP per capita growth:
    - Per capita output growth impacts negatively the current account.
    - Estimated coefficient in Column (3) is around minus 0.2, meaning higher per-capita output growth of 1 percent would lower the current account balance by (text truncated in source).
- Additional notes:
  - Regression estimations are robust to different specifications; alternative specifications available on request.
  - Figure 1 (Key Variables) shows time series averages for regions (Latin America, Asia, Middle East) for variables including Current Account (% of GDP), Young Dependency Ratios, Old Dependency Ratios, Oil Trade Balance (% of GDP), Government Balance (% of GDP), and Net Foreign Assets (% of GDP).

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

### 0.2 percentage points of GDP. Both the sign and direction of the estimated coefficients are

### _wp1043 - 0.2 percentage points of GDP. Both the sign and direction of the estimated coefficients are

### Key findings on determinants and specification
- Estimated coefficients are consistent with results obtained by the CGER.
- Evidence supports the hypothesis that Asian crisis countries experienced a structural change in savings behavior after 1997.
  - The Asian crisis countries group has, on average, a current account norm that is more than 4 percentage points of GDP higher than other countries in the sample (see column (3) of Table 1).
  - This result is consistent with the estimations performed by the CGER, which yield a positive effect of about 6 percent of GDP, though CGER assumes that the effect dies out over time.
- The baseline regression (column (3) of Table 1) is the preferred specification and treats oil exporters differently via a dummy variable.

### Regression results (summary from Table 1, baseline and alternatives)
- Oil balance coefficients: 0.223*** (All), 0.358*** (All minus oil exporters), 0.359*** (Baseline), 0.23*** (All), 0.19*** (CGER)
  - Standard errors (in parentheses) for the first three listed: (0.035), (0.075), (0.071)
- FDI coefficients: -0.159, -0.147, -0.199*, -0.148 (standard errors: (0.101), (0.107), (0.102), (0.101))
- Fiscal balance coefficients: 0.482***, 0.526***, 0.492***, 0.492***, 0.19*** (standard errors: (0.039), (0.044), (0.040), (0.040))
- Population growth coefficients: -0.0695, 0.235, -0.0741, -0.0286, -1.22** (standard errors: (0.241), (0.279), (0.244), (0.242))
- Young-age dependency coefficients: -0.0556**, -0.0395, -0.0580***, -0.0520** (standard errors: (0.022), (0.024), (0.022), (0.022))
- Old-age dependency coefficients: -0.0404, 0.00880, -0.0526, -0.0646, -0.14** (standard errors: (0.107), (0.117), (0.108), (0.107))
- NFA coefficients: 0.0631***, 0.0595***, 0.0658***, 0.0630***, 0.02*** (standard errors: (0.007), (0.008), (0.007), (0.007))
- Relative income coefficients: -0.00554, -0.00502, -0.00567, -0.00391, 0.02* (standard errors: (0.003), (0.003), (0.003), (0.003))
- Asian crisis coefficients: 0.0468***, 0.0457***, 0.0447***, 0.0468***, 0.06*** (standard errors: (0.010), (0.012), (0.011), (0.011))
- Output growth coefficients: -0.184***, -0.194***, -0.184***, -0.187***, -0.21** (standard errors: (0.039), (0.042), (0.039), (0.032))
- Differential oil balance (D_Oilbalance_OilX): 0.199***, -0.196** (standard errors: (0.043), (0.089))
- Constant terms: 0.0200, -0.00737, 0.0231, 0.0184 (standard errors: (0.030), (0.033), (0.030), (0.030))
- Sample and fit:
  - Observations: 1044 (All), 921 (All minus oil exporters), 1044 (Baseline), 1044 (All), 1044 (CGER column)
  - Number of countries: 33, 29, 33, 33, 54
  - R-squared: 0.379, 0.352, 0.366, 0.382, 0.52
- Standard errors in parentheses. Significance codes: *** p<0.01, ** p<0.05, * p<0.1.

### Current account norms and projections (methodology and results)
- Equilibrium assumed at 5 years into the future (i.e. 2014), consistent with the IMF‘s World Economic Outlook forecast.
- Current account norms computed using projected values for relevant regressors in 2014; demographic projections use U.S. Census Bureau International Database growth rates.
- Using the baseline model:
  - About two thirds of equilibrium current account balances are in deficit.
  - Deficit norms range from minus 0.1 percent of GDP for Uruguay to over minus 9 percent of GDP for Lebanon.
  - Surplus norms range from 0.2 percent of GDP for Ecuador to over 17 percent of GDP for Algeria.
- Compared to CGER norms:
  - Asian crisis countries on average have current account norms over 4 percent of GDP higher than non-Asian crisis peers in the sample.
  - For non-Asian crisis countries with CGER estimates available, results are similar.

### Country-level baseline norms (selected from Table 2, baseline vs CGER)
- Costa Rica: -2.2 (Baseline), n.a. (CGER)
- Dominican Republic: -4.0, n.a.
- Ecuador: 0.2, n.a.
- El Salvador: -2.7, n.a.
- Guatemala: -3.4, n.a.
- Jordan: -7.9, n.a.
- Lebanon: -9.2, n.a.
- Panama: -5.8, n.a.
- Paraguay: 0.8, n.a.
- Sri Lanka: -1.4, n.a.
- Uruguay: -0.1, n.a.
- Algeria: 17.2, 14.2
- Argentina: 1.9, -0.6
- Brazil: -0.5, -0.4
- Chile: 1.3, -2.8
- China: 5.6, 2.9
- India: -2.6, -3.3
- Indonesia: 5.4, 0.0
- Korea: 8.3, 0.4
- Malaysia: 8.0, 0.7
- Mexico: -2.8, -0.8
- Philippines: 5.2, -0.1
- Thailand: 6.5, -2.9
- Venezuela: 5.6, 6.6
- Note: 1/ Estimated by using methology and coefficients from Lee et al. (2008).

### Decomposition of drivers for selected countries (Table 3; percent of GDP)
- Algeria (Norm 17.2)
  - Fiscal balance: 0.8
  - Age dependency (young): -2.0
  - NFA: 7.5
  - Oil trade balance * Oil producers: 6.3
- Malaysia (Norm 8.0)
  - Fiscal balance: -3.2
  - Age dependency (young): -2.5
  - NFA: 4.4
  - Asian crisis: 4.5
- Guatemala (Norm -3.4)
  - Fiscal balance: -0.9
  - Age dependency (young): -3.8
  - NFA: -3.2
- Jordan (Norm -7.9)
  - Fiscal balance: -2.7
  - Age dependency (young): -2.6
  - NFA: -6.0
- Panama (Norm -5.8)
  - Fiscal balance: 0.1
  - Age dependency (young): -2.4
  - NFA: -5.8

### Confidence intervals, deviations, and out-of-equilibrium countries
- Confidence bands constructed from standard errors of fitted values; a four standard deviation band contains over 95 percent of possible observed sample estimates under normality and unbiasedness assumptions.
- Underlying current account balances projected for 2014 use IMF‘s October 2009 WEO forecasts.
- Rule for potential disequilibrium: underlying current account more than two standard deviations away from estimated norm may warrant a real exchange rate adjustment.
- Countries with underlying balances more than two standard deviations away from the norm: Algeria, Chile, Ecuador, Korea, Indonesia, Philippines, South Africa, and Thailand.
- Caution for natural-resource-rich countries (Thomas et al. (2008)): approach may be inappropriate because it may not capture the stock of wealth from natural resources; results for Algeria, Ecuador, and Chile should be interpreted with caution.
- If the Asian crisis dummy effect fades in the medium term, underlying balances for Korea, Indonesia, Philippines, and Thailand would not be statistically different from norms, but Malaysia would fall outside the 95 percent confidence interval.

### Table 4 (Norms and Confidence Intervals) — selected baseline entries (Norm, SD, 2014 Projection)
- Colombia: Norm -0.4, SD 1.3, 2014 Proj. -2.1 (Included in 95 percent confidence interval)
- El Salvador*: -2.7, 1.3, -2.7
- Brazil: -0.5, 1.3, -0.8
- India: -2.6, 1.4, -1.8
- Turkey: -4.9, 1.4, -3.1
- Chile: 1.3, 1.4, -2.2
- Ecuador*: 0.2, 1.4, -3.2
- South Africa: -3.4, 1.4, -7.5
- Peru: -0.9, 1.5, -1.6
- China: 5.6, 1.5, 8.4
- Malaysia: 8.0, 1.5, 10.8
- Indonesia: 5.4, 1.5, -1.0
- Korea: 8.3, 1.5, 2.1
- Argentina: 1.9, 1.5, 2.0
- Thailand: 6.5, 1.6, 1.6
- Philippines: 5.2, 1.6, 0.9
- Panama*: -5.8, 1.6, -6.0
- Venezuela: 5.6, 1.6, 5.7
- Jordan*: -7.9, 1.6, -7.0
- Algeria: 17.2, 1.8, 28.2
- Lebanon*: -9.2, 2.0, -9.3
- Uruguay*: -0.1, 2.8, -0.1
- Legend in table: * Not included in CGER sample. 1/ Included in the 95 percent confidence interval. 2/ Included in the 99 percent confidence interval. 3/ Statistically different from norm at 99 confidence level.

### Conclusions and implications
- The fundamental determinants of equilibrium current account balances estimated for 33 emerging market economies are similar to those identified by the CGER, but:
  - Fiscal balance has a considerably stronger impact on current account norms for emerging markets in this sample.
  - Evidence supports that the Asian crisis caused a permanent shift toward higher savings in affected countries (Indonesia, Korea, Malaysia, Philippines, Thailand).
- For eleven smaller emerging market economies not included in the CGER sample, almost all have underlying current account balances not statistically different from their norms, suggesting policies broadly consistent with external and domestic balance in the medium term.
- Caveats and recommended next steps:
  - Estimates are subject to significant uncertainty; goodness of fit is less than 40 percent, reflecting large variation in current account balances across countries and time.
  - The common specification may not capture heterogeneity across diverse countries.
  - Results for natural-resource-rich countries should be interpreted with caution; alternative methodologies (e.g., those addressing resource wealth stocks) are warranted.
  - The approach should be complemented with other methods to arrive at a more rigorous assessment of policy appropriateness.

*Source: IMF staff estimates and analysis in the provided content unit.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp1043.pdf_