## _wp10235 - Appendix Tables

## Source details

**Canonical URL:** [_wp10235 - Appendix Tables](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10235.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10235.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10235.pdf.json)

---

### Introduction and scope
- Revisits the Lucas (1990) paradox by quantifying empirically the relevance of restrictions on international capital flows in shaping patterns of capital movements at various stages of economic development.
- Focuses on the impact over time of capital account liberalization on capital flows across countries at different levels of income per capita.
- Uses an updated index of capital account openness from Quinn (1997); de jure index normalized between 0 and 1 (0 = fully closed, 1 = fully open), constructed from IMF AREAER.
- Dataset: 109 countries with populations above one million; period 1980-2006; main analysis uses panel of non-overlapping five-year averages over 1982-2006.

### Key empirical observations and time patterns
- Many developing countries had significant capital account restrictions at the time of Lucas (1990); progressive liberalization across income groups since then.
- High income countries liberalized in the 1980s; by early 2000s cross-border capital flowed freely among advanced economies.
- Emerging markets liberalized with a lag (many restrictions removed in the early 1990s); lower income countries liberalized mostly in the second half of the 1990s with some moderate restrictions remaining.
- Time series properties matter: for financial openness and the current account roughly 30% of the variance is across time (remainder within-country variation).

### Main empirical findings
- Neoclassical prediction not verified in cross-section during the 1980s when many countries had restrictions.
- Post early-1990s pattern:
  - Among countries with an open capital account, richer countries tend to experience net capital outflows, while poorer countries tend to experience net capital inflows (conditional on controls).
  - In countries with closed capital accounts, there is no systematic relationship between level of development and net capital flows.
- Interaction interpretation:
  - Key regressors: log initial GDP per capita relative to U.S. (income), capital account openness index (Q), interaction income × Q.
  - If β3 (income×openness) significantly positive, financial openness amplifies the relation between income and net flows in the neoclassical direction.

### Composition of flows and responsiveness by flow type
- Flow types that respond as predicted when capital accounts are open:
  - Foreign direct investment (FDI)
  - Portfolio equity investment (PE)
  - Other private investments (e.g., loans to private sector)
- Flow types that do not show a relation with income levels even with open capital accounts:
  - Portfolio debt
  - Other investments vis-à-vis the public sector
- Cross-section illustrative magnitudes (1980–2006, fully opened capital account):
  - Difference in predicted net FDI inflows between country at 10 percent vs. 50 percent of US income: 4.5 percent of GDP.
  - For a lower middle income country at 10 percent of US level after opening: predicted net inflows of FDI ≈ 5 percent of GDP; PE ≈ 0.03 percent of GDP; other private sector investment ≈ 2 percent of GDP.
  - For a high income country at 90 percent of US level after opening: predicted outflows of FDI ≈ 0.9 percent of GDP; PE ≈ 1.6 percent of GDP; other private sector investment ≈ 4.2 percent of GDP.

### Data, variables, and empirical approach
- Dependent variable: current account balance relative to GDP (treats errors and omissions as unreported capital flows and includes changes in reserve assets). Alternatives used: current account plus concessional loans; current account minus reserves; balance-of-payments flow types.
- Main openness measures:
  - Quinn (1997) index, updated to 2006; normalized 0–1.
  - Chinn and Ito index used as robustness alternative; normalized 0–1.
- Relative income: PPP GDP per capita relative to U.S., constant 2005 international Dollars; index value of 100 for the U.S.; log of GDP per capita relative to the U.S. used.
- Controls (“standard”): fiscal balance, old age dependency ratio, population growth, initial net foreign asset position, oil trade balance, real per capita GDP growth; terms of trade included in fixed-effect panels.
- Estimation strategies:
  - Cross-section OLS on full-sample averages.
  - Repeated cross-sections: 1980–1992 and 1993–2006.
  - Preferred panel: non-overlapping five-year averages, country fixed effects.
  - Spline specification: dummy for sufficiently open capital accounts chosen via spline search maximizing within R2; spline yields 72nd percentile threshold corresponding to openness index value 0.8125.

### Empirical results — cross-section and panel (selected quantitative findings)
- Cross-section (1980–2006, conditional):
  - Removal of capital account restrictions implied inflows for countries with real income per capita below 65 percent of US level (from one cross-sectional specification).
  - For the 1993–2006 cross-section: threshold example where removal of restrictions leads to inflows for countries with income per capita below 41 percent of US level (in one specification).
  - Example: lower middle income country at 10 percent of US PPP income (China, Thailand, Indonesia in 2000) with initially closed capital account: complete removal would result in annual net capital inflow of 4.5 percent of GDP (cross-sectional calculation).
  - Example: high income country at 90 percent of US level: complete opening would yield net capital outflow of 2.5 percent of GDP (cross-sectional calculation).
- Panel (five-year averages, fixed effects):
  - Within-country coefficient examples:
    - Middle income country at 10 percent of US level: net capital inflows of about 2.1 percent of GDP annually following complete opening.
    - Advanced country at 90 percent of US level: capital outflows of 5 percent of GDP after complete opening (panel within-country estimate).
  - Relationship becomes significantly positive for index of financial openness above 0.6.
  - F-test evidence: F-test values reported (e.g., 24 without fixed effects, 11 with fixed effects) reject null of no correlation for fully opened capital accounts.

### Robustness checks (selected)
- Financial crises: excluding crisis effects does not bias main findings; crisis countries had, on average, higher trade balances over the period.
- Aid: controlling for concessional loans to GDP and grants does not alter main results.
- Domestic financial development: controls for private credit to GDP, de jure domestic financial reform index, and credit growth do not overturn findings.
- Institutions and human capital: controlling for property rights (ICRG) shows better institutions associated with lower current account, but main results persist; years of schooling does not materially affect main results.
- Reserve accumulation: positively associated with current account; coefficient smaller for opened capital accounts. Controlling for reserves does not materially change coefficients of interest.
- Alternative measures and samples: netting out reserve accumulation, scaling by population, using different flow definitions, excluding transition economies, adding time fixed effects, and using Chinn-Ito index all leave core conclusions intact.
- Spline robustness: spline dummy for open countries (72nd percentile / index = 0.8125) produces consistent interaction results; examples from Tables A3 and A4 show dummy for open countries coefficients negative and interaction with Log Initial GDP positive and significant in many specifications.

### Key regression coefficients and summary statistics (selected exact values extracted from appendix tables)
- Table 4 (five-year panel, sample: Observations 421; Countries 102; Country Fixed Effects: YES; R-squared overall: 0.346; R-squared within: 0.291; F-Test: 12.29; p-value: 0.001)
  - Index of Initial Capital Account Openness (Quinn): coefficients include -0.0990***, -0.0913**, -0.1362***, -0.0824**, -0.0666*
  - Log Initial GDP (PPP per capita): coefficients include 0.0331***, 0.0338***, 0.0440***, 0.0313**, 0.0217*
  - Net Grants to GDP: 0.3336***
  - Concessional Loans to GDP: -0.6202***
  - Growth in Private Credit to GDP: -0.1002***
  - Institutions (ICRG): -0.0827**
  - Reserve Accummulation to GDP: 0.3794*** and 0.8490*** (in different specifications)
- Table 5 (robustness alternatives; common Observations 421; Countries 102 in many columns)
  - Index of Initial Capital Account Openness: examples -0.1093*** (col 1), -0.1195*** (col 2), -0.1015*** (col 3), -0.0985** (col 6)
  - Log Initial GDP: examples 0.0392***; 0.0386***; 0.0349***; 0.0366***; 0.0247***
  - F-Test examples: 15.39 (p-value 0.000) in column (1); 9.275 (p-value 0.003) in column (2)
- Tables 6 and 7 (capital flows by type):
  - Table 6 (cross-sectional averages, 1980–2006): Index coefficients (selected):
    - FDI Outflows: -0.1162***; interaction LogInitialGDP×Index for FDI: 0.0279***
    - Portfolio Equity Outflows: -0.0165**; interaction 0.0073**
    - Private Other Investment Outflows: -0.0879***; interaction 0.0289***
  - Table 7 (panel, five-year averages, country fixed effects):
    - Index coefficients (selected): FDI -0.0416***; Portfolio Equity -0.0220**; Other investment for banks -0.0433**
    - Interaction (LogInitialGDP × Index): FDI 0.0083*; Portfolio Equity 0.0105***; Other investment for banks 0.0146**

### Appendix summary statistics and correlations (selected exact figures)
- Table A1 (Summary Statistics):
  - Current Account to GDP (overall mean) -0.0222; Std. Dev. 0.0531; Min -0.2423; Max 0.2061; N = 462
  - Private Credit to GDP ratio (overall mean) 0.4687; Std. Dev. 0.4183; Min 0.0000; Max 2.0996; N = 410
  - Reserve Accummulation to GDP (overall mean) 0.0120; Std. Dev. 0.0197; Min -0.0835; Max 0.1468; N = 459
  - Financial Reform (DF) (overall mean) 0.5885; Std. Dev. 0.2706; Min 0.0000; Max 1.0000; N = 384
  - Institutions (ICRG) (overall mean) 0.6511; Std. Dev. 0.1500; Min 0.2028; Max 0.9542; N = 439
- Table A2 (pairwise correlations, examples):
  - Log Initial GDP vs Index of Initial Capital Account Openness: 0.5176 (p-value 0)
  - Index of Initial Capital Account Openness vs Per capita real GDP growth: 0.123 (p-value 0.0077)
  - Initial NFA to GDP vs Per capita real GDP growth: 0.1021 (p-value 0.0296)

### Policy-relevant implications and conclusions
- Accounting for capital account openness reconciles observed capital flow patterns with neoclassical theory:
  - In financially open countries, capital tends to flow from richer to poorer countries (net inflows into poorer, net outflows from richer), conditional on standard controls.
  - In financially closed countries, net capital inflows are not systematically correlated with the level of development.
- Types of flows driving results: predominantly FDI, portfolio equity investment, and (to some extent) loans to the private sector; portfolio debt and public-sector flows show no systematic relation with development level.
- Policy insight: policy-induced capital account restrictions have had a statistically and economically large effect on the global allocation of capital over the past three decades; debates on global imbalances and capital allocation should incorporate capital account policies alongside institutions, human capital, financial imperfections, and sovereign risk.

*Source: _wp10235 - Appendix Tables*

### Appendix Tables ...................................................................................................... 4

### _wp10235 - Appendix Tables

### Introduction and scope
- Revisits the Lucas (1990) paradox by quantifying empirically the relevance of restrictions on international capital flows in shaping patterns of capital movements at various stages of economic development.
- Focuses on the impact over time of capital account liberalization on capital flows across countries at different levels of income per capita, rather than only long-term cross-sectional determinants.
- Notes that policies related to capital account openness have dramatically evolved during the past 30 years.
- Uses an updated index of capital account openness from Quinn (1997); appendix contains further details on the measure.

### Key empirical observations and time patterns
- At the time of Lucas (1990), many developing countries still had significant capital account restrictions; since then countries across all income groups have progressively liberalized capital movements.
- High income countries initiated liberalization in the 1980s; by the early 2000s cross-border capital was flowing freely among advanced economies.
- Emerging markets followed with a lag; many restrictions were removed in the early 1990s (examples in text include Korea and Mexico).
- Liberalization in lower income countries started mostly in the second half of the 1990s, with some moderate restrictions remaining in place thereafter.
- The time series properties of the data are crucial when exploring the role of capital account openness.

### Main empirical findings
- The prediction of the standard neoclassical model is not verified in the cross-section of countries during the 1980s when many countries had capital account restrictions.
- After the early 1990s, observed patterns change: poorer (respectively richer) countries with open capital accounts tended to experience net capital inflows (respectively outflows), conditional on a set of fundamentals.
- More generally, across the whole sample:
  - Among countries with an open capital account, richer countries tend to experience net capital outflows, while poorer countries tend to experience net capital inflows.
  - In countries with closed capital accounts, there appears to be no systematic relationship between the level of economic development and net capital flows.
- Interpretation: capital account restrictions were effective in constraining capital flows when in place; liberalization leads rich countries to net capital outflows and poor countries to net capital inflows.

### Composition of flows and responsiveness by flow type
- Evidence consistent with the hypothesis that capital flows more responsive to the marginal product of physical capital flow from rich to poor countries when the capital account is open.
- Flow types that respond according to neoclassical model predictions when capital accounts are open:
  - Foreign direct investment (FDI)
  - Portfolio equity investment
  - Other private investments (e.g., loans) in the private sector also tend to flow “downhill” in the absence of capital account restrictions.
- Flow types that do not show a relation with income levels even with open capital accounts:
  - Portfolio debt
  - Other investments vis-à-vis the public sector

### Data coverage and supporting material (as listed)
- Figures and tables in the document include:
  - Figure 1: Evolution of Capital Account Openness by Income Group
  - Figure 2: Current Account to GDP and Capital Account Openness
  - Figures 3a–3b: Net and Gross Capital Inflows by Type of Flows and by Income Groups, 1980-2006
  - Figures 4–5: Conditional correlation plot and effect of openness on the marginal effect of income
  - Tables 1–7 covering Current Account and Capital Account Openness (1980-2006), repeated cross-sections, panel estimates, robustness tests, and capital account openness by types of capital flows
- Appendix Tables span pages 41–44; References listed at page 22.

*Source: _wp10235 - Appendix Tables; _wp10235 - Appendix Tables ...................................................................................................... 4*

### Section III presents the data and simple stylized facts, and outlines our empirical strategy.

### _wp10235 - Section III presents the data and simple stylized facts, and outlines our empirical strategy.

### Literature: framing and hypotheses
- Existing literature offers multiple explanations for the Lucas paradox: differences in human capital, sovereign default risk, capacity to use technologies, and institutional quality.
- Institutional quality and social infrastructure are emphasized as first-order determinants of investment and total factor productivity.
- Portfolio diversification vs. development finance: Obstfeld and Taylor (2005) argue gross flows were large among advanced economies while net flows to poor countries remained small, consistent with portfolio diversification motives.
- Kalemli-Ozcan et al. (2008) show the standard model performs well within the US (no border frictions), implying frictions at national borders matter.
- Caselli and Feyrer (2007) highlight measurement issues for the marginal product of capital (MPK); micro evidence shows within-country firm-level MPK heterogeneity.
- Financial frictions and allocation puzzles: Gourinchas and Jeanne (2009) find capital flows more to countries that invest and grow less; Verdier (2008) and other studies link borrowing constraints, precautionary savings, and reform processes to capital flow patterns.
- Empirical ambiguity on effectiveness of capital controls: controls can change composition more than aggregate volume; in some cases (Chile, Colombia) controls tilted inflow composition toward less volatile flows.
- Paper’s contribution: empirical evidence that removal of capital controls over the past three decades affected the global allocation of capital.

### Data and empirical approach
- Dataset scope:
  - 109 countries with populations above one million.
  - Period: 1980-2006.
  - Main analysis: panel of non-overlapping five-year averages over 1982-2006.
  - Most countries have 5 time observations in the panel.
- Key variables and definitions:
  - Dependent variable: current account balance relative to GDP (treats errors and omissions as unreported capital flows and includes changes in reserve assets). Alternatives used: current account plus concessional loans; current account minus reserves; balance-of-payments flow types (FDI, portfolio equity, portfolio debt, other official investment, other private investment).
  - Main measure of capital account openness: Quinn (1997) index, updated to 2006; de jure index normalized between 0 and 1 (0 = fully closed, 1 = fully open), constructed from IMF AREAER.
  - Relative income: log of GDP per capita relative to the U.S. (in PPP), with initial values used in panel (value for year preceding the 5-year average).
  - Controls (“standard” controls): fiscal balance, old age dependency ratio, population growth, initial net foreign asset position, oil trade balance, real per capita GDP growth; terms of trade index included in fixed-effect panel specifications.
- Stylized data facts:
  - Time variation share: for financial openness and the current account roughly 30% of the variance is across time (remainder within-country time variation); (log) initial income variation is mostly across countries.
  - Grouping by openness: for each five-year period, countries with openness above (below) the full-period median are classified as open (closed).
  - Open-capital-account group patterns (1980-2006): advanced countries on average experienced net capital outflows; other income groups experienced net capital inflows. Differences: upper middle vs. high income ≈ 3 percentage points of GDP; lower middle vs. upper middle ≈ 2 percentage points of GDP.
  - Closed-capital-account group patterns: did not match theoretical direction; e.g., upper middle income countries experienced small net capital outflows on average while advanced countries experienced net capital inflows.
  - Low income countries: little stark difference between open vs. closed in current accounts, but foreign aid complicates interpretation.
- Empirical specification (cross-section and panel):
  - Baseline estimating equation (symbols preserved from source):
    - Dependent variable: net capital outflows (relative to GDP), denoted in source as the dependent variable symbol.
    - Key regressors: log initial GDP per capita relative to U.S. (income), capital account openness index (Q), interaction income × Q, control vector X, error term.
  - Interpretation of coefficients:
    - β1 (income) and β3 (income×openness) are main coefficients of interest.
    - If β3 significantly positive, financial openness amplifies the relation between income and net flows in the direction predicted by neoclassical theory.
  - Estimation strategies:
    - Cross-section OLS on full-sample averages.
    - Repeated cross-sections split into 1980-1992 and 1993-2006.
    - Preferred panel: non-overlapping five-year averages, country fixed effects included to account for slow-moving unobservables.
    - Spline specification: create dummy for sufficiently open capital accounts by choosing an optimal percentile via spline search maximizing within R2; dummy replaces continuous openness index in some specifications.

### Empirical results (cross-section and panel)
- Cross-section (1980-2006):
  - Unconditional: strongly positive and significant correlation between initial GDP per capita and average current account to GDP — poorer countries initially experienced larger net capital inflows.
  - Conditional on standard controls and capital account openness (and interaction):
    - The direct income coefficient often becomes insignificant once interaction and openness included.
    - Total effect for countries with open capital accounts (sum of coefficients) is positive and significant: prediction of neoclassical theory confirmed only for countries with open capital accounts.
    - Removal of capital account restrictions implied to result in inflows for countries with real income per capita below 65 percent of US level (from one cross-sectional specification).
- Repeated cross-sections (period split):
  - 1980-1992: no clear relationship between initial development and current account; capital account openness coefficient sometimes opposite theoretical prediction. Many developing and emerging markets had strong restrictions in this period, impeding capital flows from advanced countries.
  - 1993-2006: robust positive conditional correlation between initial development and current account; relationship depends on openness and is stronger for open countries. F-tests reject null of no correlation for fully opened capital account.
  - Quantitative examples (cross-section, 1993-2006):
    - Threshold: removal of restrictions leads to inflows for countries with income per capita below 41 percent of the US level (in one specification).
    - For a lower middle income country at 10 percent of US PPP income (example countries: China, Thailand, Indonesia in 2000) with initially closed capital account, a complete removal of restrictions would result in annual net capital inflow of 4.5 percent of GDP.
    - For a high income country at 90 percent of US level, complete opening would yield net capital outflow of 2.5 percent of GDP (from that cross-sectional calculation).
- Panel analysis (five-year averages, fixed effects):
  - Controlling for standard determinants, capital account openness is negatively correlated with the current account for poorer countries; interaction term shows effect weakens with development.
  - Example quantitative panel results:
    - Within-country coefficient implies: a middle income country at 10 percent of US level would experience net capital inflows of about 2.1 percent of GDP annually following a complete opening of the capital account.
    - An advanced country at 90 percent of US level would experience capital outflows of 5 percent of GDP after a complete opening (panel within-country estimate).
  - Prediction of neoclassical model:
    - For countries with strong restrictions (index ≈ 0): no significant positive correlation between initial development and current account.
    - For countries with few restrictions (index ≈ 1): very strongly positive and significant correlation (F-test value: 24 without fixed effects, 11 with fixed effects).
    - Relationship becomes significantly positive for index of financial openness above 0.6.
- Robustness checks:
  - Financial crises: excluding crisis effects does not bias main findings; crisis countries had, on average, higher trade balances over the period.
  - Aid flows: controlling for concessional loans to GDP and grants does not alter main results.
  - Domestic financial development and reforms: controls for private credit to GDP, de jure index of domestic financial reforms (Abiad et al., 2009), and domestic credit growth do not overturn findings.
  - Institutions and human capital: controlling for property rights (ICRG index) shows better institutions associated with lower current account, but main results persist; years of schooling does not materially affect main results.
  - Reserve accumulation: positively associated with current account; coefficient smaller for opened capital accounts. Controlling for reserves does not change coefficients of interest materially.
  - Alternative measures of net capital inflows: netting out reserve accumulation, scaling by population, or using different flow definitions (balance of payments items) do not change main conclusions.
  - Excluding transition economies: results robust to dropping transition countries.
  - Time fixed effects and alternative openness index: adding time fixed effects does not modify coefficients; results robust to using Chinn-Ito index instead of Quinn index.
  - Functional form / threshold effects: spline dummy for “fully opened capital accounts” interacted with initial development yields similar findings (see appendix tables A3 and A4 referenced in source).
- Which flows drive the pattern from rich to poor?
  - Cross-sectional and panel results show net FDI, net portfolio equity (PE), and net other investment vis-à-vis the private sector flow from richer to poorer countries when capital accounts are open.
  - Size illustrations (cross-section long period 1980-2006, for fully opened capital account):
    - Difference in predicted net FDI inflows between country at 10 percent vs. 50 percent of US income: 4.5 percent of GDP.
    - For a lower middle income country at 10 percent of US level after opening: predicted net inflows of FDI ≈ 5 percent of GDP; PE ≈ 0.03 percent of GDP; other private sector investment ≈ 2 percent of GDP.
    - For a high income country at 90 percent of US level after opening: predicted outflows of FDI ≈ 0.9 percent of GDP; PE ≈ 1.6 percent of GDP; other private sector investment ≈ 4.2 percent of GDP.
  - Portfolio debt investments and net other investments vis-à-vis government/central bank do not systematically flow from richer to poorer countries; such public-sector flows are less sensitive to capital controls.

### Conclusion: implications and main takeaways
- Main empirical conclusion:
  - Accounting for capital account openness reconciles observed capital flow patterns with neoclassical theory: in financially open countries, capital tends to flow from richer to poorer countries (net inflows into poorer, net outflows from richer), conditional on standard controls.
  - In financially closed countries, net capital inflows are not systematically correlated with the level of development.
- Types of flows driving results:
  - Evidence predominantly driven by FDI, portfolio equity investment, and (to some extent) loans to the private sector.
  - Portfolio debt and public-sector loans/bonds are not systematically correlated with development level, irrespective of openness.
- Policy-relevant insight:
  - Policy-induced capital account restrictions have had a statistically and economically large effect on the global allocation of capital over the past three decades.
  - Debates on global imbalances and capital allocation should incorporate the role of capital account policies alongside institutions, human capital, financial imperfections, and sovereign risk.

*Source: IMF working paper content — Section III, data and empirical approach; results and conclusions summarized from the supplied text.*

### REFERENCES

### _wp10235 - REFERENCES

### Financial reforms, financial development, and openness
- Abiad,  Abdul,  Enrica  Detragiache,  and  Thierry  Tressel  (2008)  “A  new  database  of  financial  reforms,” IMF Working Paper WP/08/266.
- Abiad,  Abdul,  Daniel  Leigh,  and  Ashoka  Mody  (2007)  “International  Finance  and  Income  Convergence: Europe is Different,” IMF Working Paper WP/07/64.
- Chinn,  Menzie  and  Hiro  Ito  (2008)  “A  New  Measure  of  Financial  Openness,”  Journal  of  Comparative Policy Analysis, 10 (3), p. 309 - 322.
- Tressel,  Thierry  and  Enrica  Detragiache  (2008)  “Do  Financial  Sector  Reforms  Lead  to  Financial  Development?  Evidence  from  a  new  Dataset,”  IMF  Working  Paper  No.  08/265.
- Kose,  M.  Ayhan,  Eswar  S.  Prasad,  Kenneth  Rogoff,  and  Shang-Jin  Wei  (2006)  "Financial  Globalization: A Reappraisal," IMF Working Paper No. 06/189.
- Lane, Philip R. and Gian Maria Milesi-Ferretti (2008) "The Drivers of Financial Globalization," American Economic Review 98(2), 327-32.
- Gourinchas,   Pierre-Olivier   and   Olivier   Jeanne   (2009)   "Capital   Flows   to   Developing   Countries: The Allocation Puzzle," Peterson Institute Working Paper Series No. 09-12.
- Gruber,  Joseph  W.  and  Steven  B.  Kamin  (2008)  "Do  Differences  in  Financial  Development  Explain  the  Global  Pattern  of  Current  Account  Imbalances?"  Board  of  Governors  of  the  Federal  Reserve  System,  International  Finance  Discussion  Papers  No.  923, March.

### Capital flows, capital controls, and current account dynamics
- Caballero,   Ricardo   J.,   Emmanuel   Farhi,   and   Pierre-Olivier   Gourinchas   (2008)   "An   Equilibrium  Model  of  `Global  Imbalances’  and  Low  Interest  Rates," American Economic Review 98(1), 358-393.
- Lucas, Robert E. (1990) “Why doesn't Capital Flow from Rich to Poor Countries?" American Economic Review 80(2), 92-96.
- Alfaro, Laura, Sebnem Kalemli-Ozcan, and Vadym Volosovych (2008) "Why doesn't capital flow  from  rich  to  poor  countries?  An  empirical  investigation,"  Review  of  Economics  and Statistics 90(2), 347-368.
- Reinhart,  Carmen  and  Kenneth  Rogoff  (2004)  „Serial  Default  and  the  `Paradox ́  of  Rich  to  Poor Capital Flows,” American Economic Review Papers and Proceedings 94(2), 53-58.
- De  Gregorio,  José,  Sebastian  Edwards,  and  Rodrigo  Valdes  (2000)  “Controls  on  Capital  Inflows: Do They Work?” Journal of Development Economics 63(1), 59-83.
- Edwards,  Sebastian  (1999)  “How  Effective  Are  Capital  Controls?”  Journal  of  Economic  Perspectives 13(4), 65–84.
- Edwards,  Sebastian  and  Roberto  Rigobon  (2009)  “Capital  Controls  on  Inflows,  Exchange  Rate Volatility and External Vulnerability,” Journal of International Economics 78(2), 256–67.
- Cardenas, Mauricio and Felipe Barrera (1997) “On the Effectiveness of Capital Controls: The Experience of Colombia During the 1990s” Journal of Development Economics 54(1), 27-57.
- Edison,  Hali  and  Carmen  Reinhart  (2001)  “Stopping  Hot  Money,”  Journal  of  Development  Economics 66(2), 533–53.
- Ostry, Jonathan, Ghosh, Atish R., Habermeier, Karl Friedrich, Chamon, Marcos, Qureshi, Mahvash Saeed and Dennis Reinhardt (2010) “Capital Inflows: The Role of Controls”, IMF Staff Position Note No. 2010/04.
- Card-supplement: Gruber, Joseph W. and Steven B. Kamin (2007) “Explaining the Global Pattern of Current Account Imbalances,” Journal of International Money and Finance 26, 500-522.
- Chinn,  Menzie.  and  E.  Prasad  (2003)  “Medium-term  determinants  of  current  accounts  in  industrial    and    developing    countries:    an    empirical    exploration,”    Journal    of    International Economics, 59, 47-76.
- Blanchard, Oliver and Gian Maria Milesi-Ferretti (2009) “Global Imbalances: In Midstream?” IMF Staff Position Note No. 09/29.
- Blanchard, Oliver and Francesco Giavazzi (2002) “Current Account Deficits in the Euro Area: The  End  of  the  Feldstein-Horioka  Puzzle?”  Brookings  Papers  on  Economic  Activity 33(2), 147-210.

### Growth, productivity, firms, and allocation
- Hall, Robert E. and Charles Jones (1999) “Why Do Some Countries Produce So Much More Output per Worker than Others?” The Quarterly Journal of Economics 114(1), 83-116.
- Hsieh,  Chang-Tai  and  Peter  J.  Klenow  (2009)  "Misallocation  and  Manufacturing  TFP  in  China and India," The Quarterly Journal of Economics 124(4), 1403-1448.
- Restuccia,   Diego   and   Richard   Rogerson   (2008)   "Policy   Distortions   and   Aggregate   Productivity  with  Heterogeneous  Plants,"  Review  of  Economic  Dynamics  11(4),  707-720.
- Alfaro,  Laura,  Charlton,  Andrew,  and  Fabio  Kanczuk  (2007),  “Plant-Size  Distribution  and  Cross-Country Income Differences”, Harvard Business School.
- Sandri,  Damiano  (2010)  "Growth  and  Capital  Flows  with  Risky  Entrepreneurship,"  IMF Working Papers No. 10/37.
- Song,  Zheng  M.,  Kjetil  Storesletten,  and  Fabrizio  Zilibotti  (2009)  “Growing  like  China,”  CEPR Discussion Paper No. 7149.
- Buera, Francisco J. and Yongseok Shin (2010) "Productivity Growth and Capital Flows: The Dynamics of Reforms," mimeo.
- Mendoza,   Enrique,   Vincenzo   Quadrini,   and   José-Victor   Ríos-Rull   (2008)   “Financial   Integration,  Financial  Development  and  Global  Imbalances”,  forthcoming,  Journal  of  Political Economy.

### Institutions, political economy, and external structure
- Acemoglu, Daron and Simon Johnson (2005) “Unbundling Institutions,” Journal of Political Economy 113(5), 949-995.
- Faria,  Andre  and  Paolo  Mauro  (2009)  "Institutions  and  the  external  capital  structure  of  countries," Journal of International Money and Finance 28(3), 367-391.
- Quinn,  Dennis  (1997)  “The  Correlates  of  Change  in  International  Financial  Regulation,”  American Political Science Review 91(3), 531-551.
- Tornell,  Aarón  and  Andrés  Velasco  (1992)  “Why  Does  Capital  Flow  from  Poor  to  Rich  Countries? The Tragedy of the Commons and Economic Growth,” Journal of Political Economy 100, 1208-1231.
- Verdier, Geneviève (2008). “What drives long-term capital flows? A theoretical and empirical investigation,” Journal of International Economics 74, 120-142.
- Heston,  Alan,  Robert  Summers,  and  Bettina  Aten  (2006)  "Penn  World  Table  Version  6.3,"  Center   for   International   Comparisons   of   Production,   Income   and   Prices   at   the   University of Pennsylvania.

### Crises, macro policy, aid, and external balance in low-income countries
- Demirgüç-Kunt, Asli and Enrica Detragiache (1998) “The Determinants of Banking Crises – Evidence from Developing and Developed Countries,” IMF Staff Papers 45(1).
- Eichengreen, Barry (2003) Capital Flows and Crisis. The MIT Press, Cambridge.
- International  Monetary  Fund,  Independent  Evaluation  Office  (2003),  "The  IMF  and  Recent  Capital Account Crises - Indonesia, Korea, Brazil," International Monetary Fund.
- Prati, Alessandro and Thierry Tressel (2006) “Aid volatility and Dutch Disease: Is there a role for macroeconomic policies?” IMF Working Paper No. 06/145.
- Christiansen,  Lone,  Alessandro  Prati,  Luca  Antonio  Ricci,  and  Thierry  Tressel  (2009)  ”External  Balance  in  Low  Income  Countries,”  in  Reichlin,  L.  and  West,  K.,  eds.  (2010) NBER International Seminar on Macroeconomics 2009, 265-322.
- Roodman, David (2006) "An index of Donor performance," Center for Global Development. Working Paper 67 November 2006 edition.
- Card entries on aid and banks: Banking Crisis; Aid (Concessional Loans to GDP) noted in empirical tables in the main text (see appendix for definitions).

### Empirical methods, datasets, and measurement
- Barro,  Robert  J.  and  Jong-Wha  Lee  (2000)  “International  Data  on  Educational  Attainment:  Updates  and  Implications,”  The  Center  for  International  Development  at  Harvard  University Working Paper 42.
- Lane, Philip R. and Gian Maria Milesi-Ferretti (2007) “The External Wealth of Nations Mark II:  Revised  and  Extended  Estimates  of  Foreign  Assets  and  Liabilities,  1970-2004,”  Journal of International Economics 73(2), 223-250.
- Kalemli-Ozcan, Sebnem, Ariell Reshef, Bent E. Sorensen and Oved Yosha (2008) „Why does Capital Flow to Rich States?” Review of Economics and Statistics (forthcoming).
- Coeurdacier, Nicolas and Philippe Martin (2009) “The geography of asset trade and the euro: Insiders  and  outsiders,”  Journal  of  the  Japanese  and  International  Economies  23(2),  90-113.
- Verdier, Geneviève (2008). “What drives long-term capital flows? A theoretical and empirical investigation,” Journal of International Economics 74, 120-142.

*Source: _wp10235 - REFERENCES*

### appendix for the precise definition of each variable. Standard errors are robust to heteroskedasticity. The test statist

### _wp10235 - appendix for the precise definition of each variable. Standard errors are robust to heteroskedasticity. The test statist

### Robustness: Key regression findings (Tables 4 and 5)
- Table 4 (Adding Control Variables): dependent variable = Current Account to GDP ratio; data averaged over 5 year periods covering 1982-2006; "Initial" refers to the year before the respective 5 year period.
  - Index of Initial Capital Account Openness (Quinn) coefficients (selected columns):
    - -0.0990*** (column summary first reported)
    - -0.0913** 
    - -0.1362*** 
    - -0.0824** 
    - -0.0666* 
  - Log Initial GDP (PPP per capita) coefficients (selected):
    - 0.0331*** 
    - 0.0338*** 
    - 0.0440*** 
    - 0.0313** 
    - 0.0217*
  - Selected control coefficients and significance:
    - Net Grants to GDP: 0.3336*** (one specification)
    - Concessional Loans to GDP: -0.6202*** (one specification)
    - Growth in Private Credit to GDP: -0.1002*** (one specification)
    - Institutions (ICRG): -0.0827** (one specification)
    - Reserve Accummulation to GDP: 0.3794*** and 0.8490*** (in different specifications)
  - Sample and model statistics (selected columns):
    - Observations: 421; Countries: 102; Country Fixed Effects: YES; R-squared (overall): 0.346; R-squared (within): 0.291
    - F-Test: 12.29; p-value: 0.001
  - Note on reported test: "The test statistic given in the last two lines refers to a F test for coeff[Log Initial GDP] = 0 (columns 1, 2, 5 and 6) or for coeff[Log Initial GDP] + coeff[Log(InitialGDP)xCapital Account Openness]= 0 conditional on the capital account openness index equal to one (columns 3,4, 7 and 8)."

- Table 5 (Additional Robustness Tests): multiple alternative specifications; dependent and explanatory variables averaged over 5 year periods covering 1982-2006; alternative dependent variables and indices in specified columns.
  - Index of Initial Capital Account Openness (selected columns):
    - -0.1093*** (column 1)
    - -0.1195*** (column 2)
    - -0.1015*** (column 3)
    - -0.0985** (column 6)
    - -0.6651** and -1.0161* (columns where alternative dependent variable used; large-magnitude estimates reported)
  - Log Initial GDP (PPP per capita) (selected):
    - 0.0392***; 0.0386***; 0.0349***; 0.0366***; 0.0247***
  - Model setups described in notes:
    - Columns (1)-(2): dependent variable = sum of concessional loans to GDP and the Current Account to GDP ratio.
    - Columns (3)-(4): subtract Reserve Accumulation to GDP from the Current Account to GDP ratio.
    - Columns (5)-(6): exclude transition economies.
    - Columns (7)-(8): include time fixed effects.
    - Columns (9)-(10): dependent variable = Current Account to Population ratio.
    - Columns (11)-(12): use the Chinn and Ito index instead of the Quinn index.
  - Sample and model statistics (selected):
    - Observations: 421 (common in many columns); Countries: 102; Country Fixed Effects alternates between NO and YES across columns.
    - Example F-Tests and p-values: F-Test 15.39 (p-value 0.000) in column (1); F-Test 9.275 (p-value 0.003) in column (2); other columns report F-Test and p-value pairs (e.g., 31.38 / 0.000; 7.003 / 0.009).

### Capital account openness and types of capital flows (Tables 6 and 7)
- Table 6 (cross-sectional averages, 1980-2006): dependent variables are various outflow types (ratios to GDP).
  - Index of Capital Account Openness coefficients (selected dependent flows):
    - FDI Outflows: -0.1162*** (column 1)
    - Portfolio Equity Outflows: -0.0165** (column 2)
    - Portfolio Debt Outflows: -0.0048 (column 3)
    - Private Other Investment Outflows (Other Sectors and Banks): -0.0879*** (column 4)
    - Other Investment Outflows for Other Sectors: -0.0368** (column 5)
    - Other Investment Outflows for Banks: -0.0407 (column 6)
    - Official Other Investment Outflows: 0.0173 (column 7)
  - Log Initial GDP (PPP per capita) x Index interaction (selected):
    - FDI: 0.0279*** 
    - Portfolio Equity: 0.0073** 
    - Private Other Investment (column 4): 0.0289***
  - Sample and fit: Observations/Countries: 105/101 (varies by flow); R-squared examples: 0.497 (FDI), 0.405 (Portfolio Equity), 0.228 (Portfolio Debt).
  - F-Test and p-value examples: F-Test 16.02 (p-value 0.000) for FDI; F-Test 5.998 (p-value 0.016) for Portfolio Equity; other flows show lower significance.

- Table 7 (panel estimates, 5-year averages covering 1982-2006 with country fixed effects):
  - Index of Initial Capital Account Openness coefficients (selected):
    - FDI: -0.0416*** (column 1)
    - Portfolio Equity: -0.0220** (column 2)
    - Other Investment Outflows for Other Sectors: -0.0280 (column 4)
    - Other Investment Outflows for Banks: -0.0433** (column 5)
  - Interaction (Log Initial GDP x Index) (selected):
    - FDI: 0.0083* 
    - Portfolio Equity: 0.0105*** 
    - Other Investment Outflows for Other Sectors: 0.0084
    - Other Investment Outflows for Banks: 0.0146**
  - Sample and fit:
    - Observations range (examples): 376 (FDI), 336 (Portfolio Equity), 345 (Portfolio Debt).
    - Countries range (examples): 103, 99, 98.
    - Country Fixed Effects: YES for all columns reported.
    - R-squared (overall) and (within): e.g., FDI overall 0.173, within 0.222; Portfolio Equity overall 0.083, within 0.083.

### Data definitions and construction (Appendix VI.A)
- Coverage and sample construction:
  - Database includes data since 1980 for all countries with a population larger than one million. 146 countries included in the database.
  - No Quinn index data for 37 countries; largest sample used in paper: 109 countries.
  - Observations pre-screened for panel data: exclude observations where dependent variable and baseline controls deviate by more than 3 standard deviations from sample mean. No country dropped entirely due to pre-screening for the panel dataset.
  - Repeated cross-section: exclude observations with less than 5 years of data to maintain comparability with panel specification.

- Key variable sources and definitions (verbatim terminology preserved):
  - Current Account/GDP: based on IFS spliced with data from WEO.
  - Capital Flows/GDP: Net and gross outflows of FDI, Portfolio Equity, Portfolio Debt and Other Investment from IFS (BoP statistics).
  - Reserve Accumulation/GDP: measured by the negative of reserve flows taken from IFS (BOP statistics).
  - Index of Initial Capital Account Openness (Quinn): normalized between 0 and 1 (1 for fully open countries); computed by Dennis Quinn (1997, updated to 2006) based on AREAER.
  - Index of Initial Capital Account Openness (Chinn and Ito): alternative measure, data updated in July 2010; normalized between 0 and 1; based on AREAER.
  - Relative GDP per capita: PPP income per capita relative to the U.S., constant 2005 international Dollars; index value of 100 for the U.S.; data from PWT 6.3.
  - Old-age dependency ratio: share of people older than 64 relative to working age population (15-64); based on UN data, annualized by World Bank.
  - Years of Schooling: average years of schooling in population aged 25+ from Barro and Lee (2000).
  - Net foreign assets/GDP: from Lane and Milesi-Ferretti (2007); if missing use cumulative current account.
  - Aid (net grants/GDP and concessional loans/GDP): Roodman (2006) measure based on ODA; concessional loans = foreign aid minus net grants.
  - Oil trade balance/GDP: from WEO.
  - General Government Balance (GGB)/GDP: from WEO; use central government balance where general not available.
  - Real per capita GDP growth: from PWT 6.3.
  - Income Groups: aggregate based on World Bank income group classification (as of 2006).
  - Terms of Trade: natural logarithm of terms of trade of goods and services; data from WEO.
  - Financial Reform Index: domestic financial reform index coded between 0 and 1 from Abiad et. al (2008).
  - Private Credit/GDP: Private Credit by Deposit Money Banks and Other Financial Institutions relative to GDP from World Bank.
  - Growth in Private Credit/GDP: growth rate of Private Credit/GDP.
  - Banking Crisis: dummy based on updated Demirgüç-Kunt and Detragiache (1998); value 1 in crisis years, 0 otherwise.
  - Institutions (ICRG): composite index from International Country Risk Guide; weighted average of listed political risk sub-components.
  - Notation references: INS, WEO, IFS, BOP, PWT 6.3.

### Sample list and sizing (Appendix VI.B)
- Biggest sample for unconditional regressions: 109 Countries (list of country names provided in original appendix; sample includes Albania (ALB) through Zimbabwe (ZWE) as enumerated).

### Spline search method for splitting closed vs open groups (Appendix VI.C)
- Objective: split countries into financially open and closed groups using spline search to maximize explanatory power (within R-squared) of baseline regression of current account/GDP on standard controls and interaction with initial income.
- Procedure:
  - Define #ܲܮܣܥ
௜௧ equal to 1 if country's level of openness >= the #th percentile of openness across countries in a given panel window; otherwise 0.
  - Choose # between 20 and 80 to ensure sufficiently large groups.
  - Select # that maximizes within R-squared of baseline regression.
- Result:
  - Method yields the 72nd percentile for baseline specifications (5 year averages with financial openness in initial terms), regardless of fixed effects inclusion.
  - The 72nd percentile corresponds to an index value of financial openness of 0.8125.
  - This threshold splits observations into closed and open groups with markedly different effects of initial income on current account balance.

*Source: IMF staff appendix and tables as provided in the content unit.*

### Appendix Tables

### _wp10235 - Appendix Tables

### Summary Statistics (Table A1)
- Current Account to GDP (overall mean) -0.0222; Std. Dev. 0.0531; Min -0.2423; Max 0.2061; N = 462
  - between 0.0494; Min -0.1874; Max 0.1037; n = 109
  - within 0.0306; Min -0.1505; Max 0.1197; T = 4.23853
- Fiscal balance to GDP (overall mean) -0.0013; Std. Dev. 0.0351; Min -0.1286; Max 0.1144; N = 444
  - between 0.0260; Min -0.0597; Max 0.0859; n = 106
  - within 0.0240; Min -0.0936; Max 0.0803; T = 4.18868
- Old age dependency ratio # (overall mean) -0.0732; Std. Dev. 0.0583; Min -0.1910; Max 0.0732; N = 448
  - between 0.0577; Min -0.1627; Max 0.0550; n = 106
  - within 0.0104; Min -0.1106; Max -0.0237; T = 4.22642
- Population growth # (overall mean) 0.0083; Std. Dev. 0.0101; Min -0.0174; Max 0.0436; N = 448
  - between 0.0101; Min -0.0163; Max 0.0288; n = 106
  - within 0.0037; Min -0.0125; Max 0.0267; T = 4.22642
- Initial NFA to GDP (overall mean) -0.3933; Std. Dev. 0.5266; Min -3.5789; Max 1.6825; N = 454
  - between 0.4955; Min -2.2570; Max 0.9833; n = 109
  - within 0.2524; Min -1.7153; Max 1.0761; T = 4.16514
- Oil trade balance to GDP (overall mean) 0.0035; Std. Dev. 0.0813; Min -0.1367; Max 0.3891; N = 452
  - between 0.0905; Min -0.0916; Max 0.3481; n = 108
  - within 0.0194; Min -0.0786; Max 0.1434; T = 4.18519
- Per capita real GDP growth (overall mean) 0.0203; Std. Dev. 0.0286; Min -0.0864; Max 0.1122; N = 468
  - between 0.0231; Min -0.0296; Max 0.1004; n = 109
  - within 0.0207; Min -0.0801; Max 0.0938; T = 4.29358
- Terms of Trade (overall mean) 4.6118; Std. Dev. 0.1646; Min 3.9912; Max 5.2514; N = 463
  - between 0.1261; Min 4.1101; Max 5.0489; n = 108
  - within 0.1263; Min 4.1706; Max 5.1676; T = 4.28704
- Net Grants to GDP (overall mean) 0.0228; Std. Dev. 0.0410; Min 0.0000; Max 0.3243; N = 460
  - between 0.0388; Min 0.0000; Max 0.2142; n = 106
  - within 0.0173; Min -0.0582; Max 0.1330; T = 4.33962
- Concessional Loans to GDP (overall mean) 0.0099; Std. Dev. 0.0196; Min -0.0062; Max 0.1479; N = 460
  - between 0.0180; Min -0.0008; Max 0.0846; n = 106
  - within 0.0102; Min -0.0378; Max 0.0837; T = 4.33962
- Private Credit to GDP ratio (overall mean) 0.4687; Std. Dev. 0.4183; Min 0.0000; Max 2.0996; N = 410
  - between 0.3831; Min 0.0321; Max 1.6609; n = 103
  - within 0.1515; Min -0.1307; Max 1.3309; T = 3.98058
- Growth in Private Credit to GDP (overall mean) 0.0281; Std. Dev. 0.0908; Min -0.2670; Max 0.4439; N = 405
  - between 0.0582; Min -0.0670; Max 0.2664; n = 103
  - within 0.0777; Min -0.2200; Max 0.3602; T = 3.93204
- Political Constraints (overall mean) 4.9620; Std. Dev. 2.0862; Min 0.0000; Max 7.0000; N = 463
  - between 1.9317; Min 1.0000; Max 7.0000; n = 108
  - within 0.9488; Min 0.6020; Max 7.9620; T = 4.28704
- Reserve Accummulation to GDP (overall mean) 0.0120; Std. Dev. 0.0197; Min -0.0835; Max 0.1468; N = 459
  - between 0.0139; Min -0.0069; Max 0.0718; n = 109
  - within 0.0147; Min -0.0729; Max 0.1078; T = 4.21101
- Financial Reform (DF) (overall mean) 0.5885; Std. Dev. 0.2706; Min 0.0000; Max 1.0000; N = 384
  - between 0.1952; Min 0.1956; Max 0.9611; n = 89
  - within 0.1879; Min 0.1085; Max 1.0085; T = 4.31461
- Institutions (ICRG) (overall mean) 0.6511; Std. Dev. 0.1500; Min 0.2028; Max 0.9542; N = 439
  - between 0.1282; Min 0.3099; Max 0.9061; n = 102
  - within 0.0698; Min 0.4405; Max 0.8451; T = 4.30392
- Note: # indicates deviation from trading partners. "Initial" refers to the year before the respective 5 year period.

### Pairwise correlations (Table A2)
- Correlations (with P values in parenthesis)
  - Terms of Trade vs Log Initial GDP (PPP per capita), relative to US: -0.0895 (0.0544)
  - Terms of Trade vs Index of Initial Capital Account Openness: -0.0297 (0.5244)
  - Terms of Trade vs Fiscal balance to GDP: -0.0037 (0.938)
  - Terms of Trade vs Old age dependency ratio: -0.0224 (0.6365)
  - Terms of Trade vs Population growth: 0.0575 (0.2254)
  - Terms of Trade vs Initial NFA to GDP: -0.1766 (0.0002)
  - Terms of Trade vs Oil trade balance to GDP: -0.1422 (0.0025)
  - Terms of Trade vs Per capita real GDP growth: -0.0235 (0.6139)
- Other selected correlations (with P values)
  - Log Initial GDP vs Index of Initial Capital Account Openness: 0.5176 (0)
  - Log Initial GDP vs Old age dependency ratio: 0.7062 (0)
  - Log Initial GDP vs Population growth: -0.5729 (0)
  - Index of Initial Capital Account Openness vs Per capita real GDP growth: 0.123 (0.0077)
  - Initial NFA to GDP vs Per capita real GDP growth: 0.1021 (0.0296)
  - Oil trade balance to GDP vs Fiscal balance to GDP: 0.2021 (0)

### Robustness: Spline Specification (Table A3)
- Dependent variable: Current Account to GDP ratio. Data averaged over 5 year periods covering 1982-2006. "Initial" refers to year before 5-year period. Robust standard errors in parentheses. Significance: *** p<0.01, ** p<0.05, * p<0.1.
- Selected coefficient estimates and statistics across columns (1)–(8):
  - Log Initial GDP (PPP per capita):
    - Column (1): 0.0201*** (0.0023)
    - Column (3): 0.0142*** (0.0039)
    - Column (5): 0.0145*** (0.0025)
    - Column (7): 0.0100** (0.0040)
  - Dummy for Open Countries:
    - Column (2): 0.0016 (0.0050)
    - Column (4): -0.0791*** (0.0229)
    - Column (6): -0.0594*** (0.0189)
    - Column (8): -0.0672*** (0.0193)
  - Interaction term: Log Initial GDP * Dummy for Open Countries reported in some columns (coefficients shown in table header)
  - Standard Control Variables: columns indicate combinations of YES/NO
  - Observations: 462 (cols 1–2), 427 (cols 3–4), 462 (cols 5–6), 427 (cols 7–8)
  - Countries: 109 (cols 1–2), 105 (cols 3–4), 109 (cols 5–6), 105 (cols 7–8)
  - Country Fixed Effects: NO (odd columns), YES (even columns)
  - R-squared (overall): 0.161 (col 1), 0.161 (col 2), 0.416 (col 3), 0.316 (col 4), 0.196 (col 5), 0.184 (col 6), 0.436 (col 7), 0.337 (col 8)
  - R-squared (within): 0.009 (col 1), 0.234 (col 2), 0.036 (col 3), 0.270 (col 4)
  - F-Test and p-values (test for coeff[Log Initial GDP]=0 or coeff[Log Initial GDP]+coeff[Log(InitialGDP)xDummy for Open Countries]=0):
    - Examples: F-Test 78.59, p-value 0.000 (col 1); F-Test 1.24, p-value 0.266 (col 2); F-Test 13.49, p-value 0.000 (col 3); F-Test 6.29, p-value 0.014 (col 4).

- Note: Dummy for Open Countries = 1 if initial openness above the 72th Percentile across all countries for a given 5 year period (spline search procedure).

### Robustness: Spline Specification (continued) (Table A4)
- Selected coefficient estimates and statistics across columns (1)–(10):
  - Log Initial GDP (PPP per capita):
    - Column (1): 0.0346* (0.0154)
    - Column (2): 0.0505** (0.0208)
    - Column (3): 0.0246 (0.0161)
    - Column (4): 0.0346* (0.0190)
    - Column (5): 0.0287* (0.0164)
    - Column (6): 0.0573** (0.0218)
    - Column (7): 0.0318* (0.0165)
    - Column (8): 0.0322* (0.0167)
    - Column (9): 0.0465* (0.0259)
    - Column (10): 0.0343 (0.0212)
  - Dummy for Open Countries:
    - Column (1): -0.0660*** (0.0180)
    - Column (2): -0.0538*** (0.0186)
    - Column (3): -0.0350* (0.0204)
    - Column (4): -0.0536*** (0.0187)
    - Column (5): -0.0571*** (0.0199)
    - Column (6): -0.0629** (0.0242)
    - Column (7): -0.0672*** (0.0179)
    - Column (8): -0.0638*** (0.0201)
    - Column (9): -0.0325 (0.0205)
    - Column (10): -0.0516*** (0.0193)
  - Net Grants to GDP:
    - Column (1): 0.2783*** (0.1035)
    - Column (2): -0.2651* (0.1501)
    - Column (3): 0.1715 (0.1363)
  - Concessional Loans to GDP:
    - Column (1): -0.5897*** (0.1562)
    - Column (2): -0.1700 (0.2182)
    - Column (3): -0.4583*** (0.1291)
  - Private Credit to GDP ratio:
    - Column (1): -0.0249** (0.0103)
    - Column (2): -0.0185 (0.0113)
    - Column (3): -0.0155 (0.0107)
  - Financial Reform Index:
    - Column (7): 0.0147 (0.0111)
    - Column (8): 0.0425** (0.0167)
  - Growth in Private Credit to GDP:
    - Column (4): -0.0970*** (0.0231)
    - Column (5): -0.0860*** (0.0235)
    - Column (6): -0.0739*** (0.0248)
  - Institutions (ICRG):
    - Column (1): -0.0689** (0.0325)
    - Column (2): -0.0894*** (0.0317)
    - Column (3): -0.0290 (0.0324)
  - Years of Schooling:
    - Column (7): 0.0036 (0.0057)
  - Reserve Accummulation to GDP:
    - Column (1): 0.3784*** (0.1360)
    - Column (2): 0.3959*** (0.1358)
    - Column (7): 0.2656 (0.1807)
    - Column (8): 0.3475** (0.1552)
    - Interaction Reserve Accummulation to GDP * Dummy for Open Countries:
      - Column (7): -0.1204 (0.3402)
  - Standard Control Variables: YES in all columns (1)–(10)
  - Observations: vary by column (e.g., 421, 385, 361, 381, 403, 288, 425, 425, 319, 360)
  - Countries: vary (e.g., 102, 99, 98, 99, 98, 83, 105, 105, 80, 92)
  - Country Fixed Effects: YES for all columns
  - R-squared (overall): range 0.271 to 0.392 across columns
  - R-squared (within): range 0.185 to 0.383 across columns
  - F-Test and p-values (test for coeff[Log Initial GDP] + coeff[Log(InitialGDP)xDummy for Open Countries]=0 for fully opened capital account)
    - Examples: F-Test 11.09, p-value 0.001 (col 1); F-Test 8.85, p-value 0.004 (col 2); F-Test 4.27, p-value 0.042 (col 9); F-Test 4.99, p-value 0.028 (col 10)

- Note: The Dummy for Open Countries takes the value of 1 if a country's level of initial openness (over a 5 year period) is above the 72th Percentile of openness across all countries for a given 5 year period (spline search procedure). The dependent and explanatory variables are averaged over 5 year periods covering 1982-2006 (except when stated otherwise). Standard errors are robust to heteroskedasticity.

*Source: _wp10235 - Appendix Tables*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2010/_wp10235.pdf_