## 1. Private Capital Flows Volatility

## Source details

**Canonical URL:** [1. Private Capital Flows Volatility](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09114.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09114.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09114.pdf.json)

---

### Introduction: purpose and approach
- Objective: identify factors behind the composition and distribution of recent private capital flows to sub-Saharan Africa (SSA), and begin investigating their possible contribution to growth.
- Dataset: unique panel for 44 sub-Saharan African countries (Zimbabwe excluded for analysis due to data issues), annual data on private capital inflows (broken down by FDI, portfolio, private debt) for 2000–2007.
- Empirical strategy:
  - Instrumental variables approach to study determinants of private capital flows (addressing endogeneity).
  - Two-step approach (following Rodrik and Subramanian (2009)) to investigate whether capital flows have a first-order impact on variables known to influence growth (as a necessary condition for second-order impacts on growth).

### Recent trends in financial globalization in SSA (2000–2007)
- Magnitude and composition of flows:
  - Private capital flows to SSA rose almost 6-fold since 2000, from about $15 billion to about $84 billion in 2007.
  - In 2007, private capital flows represented 10 percent of the region’s GDP.
  - In 2007, recorded net capital flows to SSA amounted to about US$76 billion.
  - In 2007 private inflows were about twice the volume of official development assistance (about US$40 billion).
- Growth rates:
  - On average between 2001 and 2007 total gross inflows to the region increased by 32 percent annually—the rate accelerating to 44 percent after 2005.
- Volatility (sample averages reported):
  - Total inflow: Mean 19.8, Standard Deviation 4.5, Volatility 4.4 (2000-03 vs 2004-07 shown in table header).
  - FDI: Mean 12.4, Standard Deviation 5.0, Volatility 2.5.
  - Portfolio: Mean 1.3, Standard Deviation 0.7, Volatility 1.9.
  - Private debt: Mean 6.1, Standard Deviation 0.5, Volatility 13.2.
- Volatility dynamics:
  - Overall volatility of capital flows fell in later years of the sample, but volatility of FDI increased (partly driven by large movements in South Africa in 2005-2006).
- Push and pull factors:
  - Global “push” factors: abundant global liquidity; world real GDP growth and rising oil and non-fuel commodity prices.
  - Domestic “pull” factors: improved macroeconomic performance across much of the continent (real GDP growth and rising per capita GDP), lower inflation, improved central government balances.
- Distributional patterns and concentration:
  - FDI shares (2000–03): South Africa 27.3, Nigeria 16.8, Angola 16.5, Equatorial Guinea 5.8, Chad 5.0, Other SSA 28.5.
  - FDI shares (2004–07): Nigeria 34.3, South Africa 15.5, Equatorial Guinea 10.2, Tanzania 3.2, Congo, Republic of, 2.9, Other SSA 33.8.
  - Portfolio and debt shares (2000–03): South Africa 23.1, Kenya 7.4, Ethiopia 6.8, Ghana 6.8, Congo, Dem. Rep. of 6.3, Other SSA 49.6.
  - Portfolio and debt shares (2004–07): South Africa 72.8, Nigeria 5.6, Ghana 2.2, Angola 2.0, Kenya 1.7, Other SSA 15.7.
  - Top recipients (averages 2000–07) include Angola, Kenya, Chad, Tanzania, South Africa, Nigeria, Equatorial Guinea, Ghana, Congo, Dem. Rep. of, Ethiopia (figures in source).
- External vulnerability:
  - For low-income SSA countries, strong correlation between capital inflows and current account balances (inflows increase as current account improves and decrease as it deteriorates), magnifying swings in overall balance of payments.
  - Surge in private capital inflows contributed to a steady increase in international reserves for SSA countries.

### Key empirical finding on determinants of private capital flows
- Main result: capital market development, rather than basic property rights institutions, is more closely associated with higher private capital inflows to SSA—reflecting the rising importance of portfolio and debt flows relative to FDI.
- Financial market structure:
  - Capital market development index: sum of four dummy variables (exists and well-developed = 1) for treasury bills, treasury bonds, corporate bond markets, and equities markets; values 0, 1, 2, 3, 4.
  - Only 10 countries (21 percent of sample) have complete capital markets: Botswana, Ghana, Kenya, Namibia, Nigeria, South Africa, Swaziland, Tanzania, Uganda, Zambia.
  - Countries with more complete capital markets are among the largest recipients of private capital inflows, alongside several oil exporters (Angola, Congo, Chad, Equatorial Guinea).
- Other significant determinants:
  - Market size dummy (South Africa or Nigeria) significant, particularly for portfolio inflows.
  - Oil-exporter dummy significant, particularly for FDI flows.
- Growth correlation caveat:
  - Strong positive correlations found between private capital inflows and real GDP growth in recipient countries; causality not established due to limited time period and endogeneity.

### Estimation strategy and instruments
- Reduced-form linear equation estimated:
  - K_i = μ + α F_i + β P_i + γ χ_i + u_i
    - K_i: log of average inflows to country i.
    - F_i: measure of financial institutions.
    - P_i: measure of property rights.
    - χ_i: set of exogenous control variables.
- Endogeneity handled via two-stage least squares (2SLS):
  - First-stage equations:
    - F_i = ν_1 + δ_11 L_i + χ_1 C_i + ε_1i
    - P_i = ν_2 + δ_22 E_i + χ_2 C_i + ε_2i
    - L_i and E_i are instruments for financial institutions and property rights institutions respectively.
  - Exclusion restriction: Cov(u_i, L_i) = Cov(u_i, E_i) = 0.
- Instruments:
  - Financial market development instrument: British legal origin (from La Porta et al., 1999).
  - Property rights instruments: ethno-linguistic fragmentation and latitude (from La Porta et al., 1999).
  - Robustness: population density in 1500 used in robustness checks yields consistent results (not shown).

### Empirical variables and controls (selection)
- Core explanatory:
  - Financial Institutions index (0–4).
  - Rule of law (scale -2.5 to 2.5).
- Controls and expectations:
  - Oil-exporter dummy (Angola, Cameroon, Chad, Congo, Rep. of, Equatorial Guinea, Gabon and Nigeria).
  - Market-size dummy (South Africa or Nigeria).
  - Secondary school enrollment (%).
  - Chinn-Ito capital account openness (average 1999–2005).
  - Trade openness (% of GDP).
  - Real GDP growth (percent).

### Baseline OLS estimation results (Table 3) — selected exact coefficients and stats
- Financial Institutions: 0.50*** 0.56*** 0.52*** 0.55*** 0.56*** 0.58** 0.56*** 0.57*** 0.67*** 0.63*** 0.55*** 0.57*** 0.55***
- Rule of Law: -0.21 -0.33 -0.41 -0.27 -0.28 -0.33 -0.37 -0.03 -0.10 -0.20 -0.18 -0.52 -0.58
- South Africa and Nigeria: 1.85 1.87* 1.98 1.89 1.79 1.58 1.35 1.37 2.11 2.12
- Oil Exporters: 1.61*** 1.01* 1.62*** 1.68*** 1.60*** 1.64*** 2.67*** 1.94*** 0.51 0.12
- Real GDP growth: 0.11*** 0.14*** 0.07
- Trade Openness: -0.01** (t-stat reported as -2.02)
- Constants (examples): 4.47*** 5.22*** 3.79*** 3.35*** 3.92*** 4.13*** 3.74*** 4.41*** 3.42*** 4.42*** 2.68*** 2.16*** 3.45*** 4.30 2.88*** 2.60***
- Observations: 43 43 43 43 42 43 43 43 43 43 43 43 43 43 43 43 43 43 43
- R-squared examples: 0.26 0.01 0.57 0.62 0.57 0.57 0.58 0.59 0.22 0.00 0.55 0.59 0.25 0.00 0.42 0.44
- F-Statistic examples: 11.96 0.31 12.00 15.81 8.78 9.46 9.41 15.32 11.97 0.01 12.89 14.06 11.76 0.21 5.45 4.93
- Note: ***, **, * indicate significance at 1, 5, and 10 percent. t-Statistics are reported in parentheses. Robust standard errors.

### Instrumental variables (2SLS) results — first and second stage highlights
- First-stage:
  - British legal origin: strong positive and significant impact on financial development, but not on the rule of law.
  - Ethnolinguistic fractionalization / latitude: strong and mostly significant impact on property rights (greater fragmentation and closeness to the Equator negative).
  - Market size: positive and significant impact on capital market development; market size negatively related to rule of law (attributed to low ratings of Nigeria and South Africa).
  - Oil exporting status: negatively and significantly related to rule of law.
- Second-stage:
  - Impact of financial institutions weaker in 2SLS (smaller coefficients) but remains strong and significant for total inflows and FDI inflows, and the share of FDI in total inflows.
  - Property rights variable: never significant and continues to have the "wrong" sign in most cases, except for FDI inflows and the share of FDI (column 2).
  - Oil resources: strong determinant of FDI inflows, not of other flows.
  - Market size: matters for portfolio and debt inflows.
- Robustness: excluding South Africa and Nigeria or excluding oil-rich countries yields broadly similar results; coefficient on financial institutions larger particularly in the FDI regressions.
- Interpretation and caution:
  - In non-oil SSA countries excluding South Africa, "FDI inflows are significantly larger, the more financial markets are developed."
  - Property rights protection remains insignificant; possible explanations include difficulty of expropriating FDI or negotiated protections for foreign investors.
  - Given the small sample size, results must be interpreted with caution.
- Conclusion from IV results: support for unbundling institutions and measuring them more precisely; "financial institutions as proxied by capital markets development matter more than basic property rights institutions."

### Do private capital flows have an impact on growth in SSA? — Correlation analysis
- Data: cross-section sample of average flows for the 2000-07 period; use of net flows as well; normalization by population or GDP.
- Scatter-plot evidence: generally positive correlation between growth and capital flows; relationship more pronounced for FDI.
- Table 5 correlations (preserved values):
  - Real GDP Growth with Net Total Flows/GDP: 0.38*
  - Real GDP Growth with FDI Flows/GDP: 0.66*
  - Debt Flows/GDP correlations include: -0.02 0.72* -0.07 1.00
  - Portfolio Flows/GDP correlations include: -0.06 0.31* -0.10 0.02 1.00
  - Per capita correlations: Real GDP Growth with Net Total Flows per capita: 0.37*; with FDI Flows per capita: 0.44*; with Debt Flows per capita: 0.08; with Portfolio Flows per capita: -0.07
- Note: * indicates significant at 10 percent.

### Channels of transmission tested (Table 6) — overvaluation, savings & investment, institutions
- Method: two-stage bivariate regressions where first stage regresses channel variable on capital flows; second stage regresses growth on predicted channel value.
- Table 6 reported coefficients (preserved verbatim):
  - Overvaluation: Net flows/GDP -0.004; Net flows per capita -0.00003
  - Investment: Net flows/GDP 0.164; Net flows per capita 0.012 ***
  - Domestic Savings: Net flows/GDP 0.337; Net flows per capita 0.021 **
  - Institutions: Net flows/GDP -0.214; Net flows per capita -0.0042
- Interpretation:
  - In most instances capital flows are not significantly correlated with the channel variables.
  - Exception: when capital flows are scaled to population, both savings and investment are positively and significantly related to capital flows.
  - Implication: for SSA countries, the impact of private capital flows on growth could be partly intermediated through net capital flows' impact on domestic savings and investment.
  - Unexpected result: institutional variable is negatively related to growth (holds even when oil-exporters are excluded).

### Multivariate growth regressions (Table 7) — selected exact coefficients and stats
- Model: cross-country OLS using averages for 2000-07; g_i = α + β·K_i + η·X_i + u_i.
- Selected results (preserved verbatim):
  - Initial per capita GDP (log): -1.53*** -1.86*** -1.67*** -2.16***.
  - Overvaluation: -0.82 -1.53 -0.68 -1.25.
  - Domestic savings: 0.09*** 0.10*** 0.09*** 0.10*** (t-statistics: 3.39, 3.59, 3.61, 3.62).
  - Investment: 0.18*** 0.16*** 0.15*** 0.17*** (t-statistics: 3.33, 2.93, 2.58, 2.97).
  - Contract enforcement: 0.11 0.09 0.09 0.08.
  - Net total flows/GDP: 0.11* (t-statistic shown as -1.94 in table layout).
  - Net total flows per capita: 0.003** (t-statistic 2.05).
  - Total inflows/GDP: 0.18*** (t-statistic 2.27).
  - Total inflows per capita: 0.003 (t-statistic 1.47).
  - Adjusted R-squared: 0.51 0.52 0.53 0.57.
  - Observations: 42 42 42 42.
- Interpretation:
  - Capital flows are positively and significantly associated with growth in most specifications.
  - Initial per capita GDP negative and significant: evidence of convergence.
  - Domestic savings and investment significant with expected signs.
  - Contract enforcement (proxy for institutions) not significant.
  - Overvaluation has expected sign but is never significant.
  - Results similar when using net FDI instead of total net private capital, with even stronger association for FDI.
- Caveats: short time period since inflows began; many inflows occurred late in 2000-07; OLS limitations and endogeneity concerns remain.

### Conclusion and policy recommendations
- Near-term context: global financial crisis and recession have caused declines in private capital flows; investors likely to remain more risk-averse for some time; donor flows to low-income countries likely to be adversely impacted.
- Empirical conclusion: results "were inconclusive regarding the impact of private capital flows on growth in SSA countries"; additional observations and research needed.
- Policy implications and recommendations:
  - SSA countries likely to remain dependent on financial inflows to finance current account deficits associated with development-oriented investments.
  - Policymakers in SSA should pay increasing attention to attracting private capital.
  - Key policy recommendation: "strengthen domestic capital markets" — develop deep and well-functioning money, foreign exchange, equity and (where relevant) derivative markets to attract private capital.
  - Simultaneous priority: "put in place strong monitoring, regulatory and supervisory frameworks, to avoid a buildup of balance-sheet vulnerabilities."

### Data, definitions, and sample moments (selected highlights)
- Data coverage and timing:
  - Data based on detailed balance of payments data obtained from AFR country desk economists in late 2007 for 44 sub-Saharan African countries.
  - Data for 2000-2006 are actual; data for 2007 are preliminary estimates.
  - Period covers 2000-2007; covers 43/44 SSA countries (text alternates between 43 and 44).
- Balance of payments recording convention used:
  - Capital inflows = sum of increases in liabilities to non-residents and of decreases in net foreign assets held by residents in a year.
  - Net flows of capital = inflows (positive sign) + outflows (negative sign).
- Appendix II selected variable definitions and sources (preserved verbatim for key entries):
  - Net total flows: FDI, portfolio and debt net flows. Units: Billions of US$. Source: IMF, African Department database.
  - Net FDI: Net foreign direct investment flows (FDI in reporting economy, net of direct investment abroad). Units: Billions of US$. Source: IMF, African Department database.
  - Total inflows: FDI, portfolio and debt inflows. Units: Billions of US$. Source: IMF, African Department database.
  - Financial Institutions index: sum of four dummy variables (Treasury bills, treasury bonds, corporate bonds, equity markets). Values: 0, 1, 2, 3, 4. Source: Lukonga (forthcoming) and authors' calculations.
  - Rule of law: scale -2.5 to 2.5. Source: http://info.worldbank.org/governance/wgi/index.asp
  - South Africa and Nigeria: dummy 0 or 1 (1 for South Africa and Nigeria). Source: Authors' calculations.
  - Oil exporters: dummy 0 or 1 (1 for Angola, Cameroon, Chad, Congo, Rep. of, Equatorial Guinea, Gabon and Nigeria). Source: IMF, African Department database.
  - Real GDP growth: Percent. Source: IMF, World Economic Outlook.
  - Secondary enrollment: Percent. Source: World Bank, World Development Indicators.
  - Capital account openness: Index (Chinn-Ito). Source: Chinn-Ito, forthcoming.
  - Trade openness: Percent of GDP. Source: IMF, World Economic Outlook.
  - Overvaluation: Deviation from "equilibrium" between PPP and market rate, normalized so 1 is the world. Source: IMF, African Department database.
  - Population: Millions. Source: IMF, World Economic Outlook.
- Appendix III dependent-variable sample moments (selected):
  - Total inflows: Observations 436; Minimum 6.3; Maximum 16,210.6; Mean 826.5; Median 233.9; Standard Deviation 2,541.0.
  - FDI inflows: Observations 430; Minimum 0.0; Maximum 4,237.1; Mean 359.8; Median 90.9; Standard Deviation 802.7.
  - Portfolio and debt inflows: Observations 430; Minimum 0.0; Maximum 13,072.4; Mean 466.6; Median 102.1; Standard Deviation 1,978.6.
  - Total inflows per capita: Observations 433; Minimum 5.3; Maximum 1,625.0; Mean 129.1; Median 25.5; Standard Deviation 307.8.
  - Total inflows to GDP: Observations 431; Minimum 0.1; Maximum 30.7; Mean 8.1; Median 6.2; Standard Deviation 6.5.
- Appendix IV growth-related sample moments (selected):
  - Real GDP growth: Observations 43; Mean 4.8; Standard deviation 3.6; Min -0.4; Max 20.1.
  - Initial per capita GDP: Observations 43; Mean 6.1; Standard deviation 1.1; Min 4.5; Max 9.0.
  - Net total flows/GDP: Observations 43; Mean 3.7; Standard deviation 7.4; Min -10.9; Max 27.1.
  - Net total flows per capita: Observations 43; Mean 54.7; Standard deviation 278.2; Min -500.6; Max 1,232.3.
  - Total inflows/GDP: Observations 43; Mean 7.7; Standard deviation 5.7; Min 1.0; Max 28.2.
  - Total inflows per capita: Observations 43; Mean 131.1; Standard deviation 313.9; Min 3.5; Max 1,645.0.
  - Domestic savings: Observations 42; Mean 12.2; Standard deviation 20.0; Min -24.7; Max 81.3.
  - Investment: Observations 42; Mean 21.6; Standard deviation 8.7; Min 8.7; Max 46.9.

*Source: _wp09114 (IMF Working Paper content unit).*

### 1. Private Capital Flows Volatility.....................................................................................

### 1. Private Capital Flows Volatility

### Major Sections (by heading and page number)
- 1. Private Capital Flows Volatility.....................................................................................8
- 2. Capital Market Development in Sub-Saharan Africa ..................................................10
- 3. Baseline OLS Estimation Results. ...............................................................................18
- 4. Instrumental Variables Estimation Results..................................................................19
- 5. Correlations: Net Capital Flows and Real Growth ......................................................23
- 6. Channels of Transmission Between Capital Flows and Growth .................................24
- 7. Estimation Results: Linear Growth Model with Capital Flows...................................26

### Figures (listed)
- 1. Private Capital Inflows to Sub-Saharan Africa, 2000-07 ..............................................5
- 2.         Global         Liquidity.............................................................................................................6
- 3.         Domestic ........................................................................................................................7
- 4. Recipients of Private Capital Flows...............................................................................8
- 5. Top Ten Recipients of Private Capital...........................................................................8
- 6.         External         Vulnerability....................................................................................................9
- 7. Sub-Saharan Africa: Capital Flows and Real GDP Growth ........................................22

### Appendices (listed)
- I. Capital Flows Data.......................................................................................................28
- II. Variables Definitions and Sources...............................................................................30
- III. Determinants of Capital Flows: Descriptive Statistics and Correlations.....................32
- IV. Private Capital Flows and Growth ...............................................................................34

*Source: _wp09114 - 1. Private Capital Flows Volatility (IMF PDF).*

### References..............................................................................................................

### _wp09114 - References..............................................................................................................

### Introduction: purpose and approach
- Objective: identify factors behind the composition and distribution of recent private capital flows to sub-Saharan Africa (SSA), and begin investigating their possible contribution to growth.
- Dataset: unique panel for 44 sub-Saharan African countries (Zimbabwe excluded for analysis due to data issues), annual data on private capital inflows (broken down by FDI, portfolio, private debt) for 2000–2007.
- Empirical strategy:
  - Instrumental variables approach to study determinants of private capital flows (addressing endogeneity).
  - Two-step approach (following Rodrik and Subramanian (2009)) to investigate whether capital flows have a first-order impact on variables known to influence growth (as a necessary condition for second-order impacts on growth).

### Recent trends in financial globalization in SSA (2000–2007)
- Magnitude and composition of flows:
  - Private capital flows to SSA rose almost 6-fold since 2000, from about $15 billion to about $84 billion in 2007.
  - In 2007, private capital flows represented 10 percent of the region’s GDP.
  - In 2007, recorded net capital flows to SSA amounted to about US$76 billion.
  - In 2007 private inflows were about twice the volume of official development assistance (about US$40 billion).
- Growth rates:
  - On average between 2001 and 2007 total gross inflows to the region increased by 32 percent annually—the rate accelerating to 44 percent after 2005.
- Volatility:
  - Table 1 (sample averages reported):
    - Total inflow: Mean 19.8, Standard Deviation 4.5, Volatility 4.4 (2000-03 vs 2004-07 shown in table header).
    - FDI: Mean 12.4, Standard Deviation 5.0, Volatility 2.5.
    - Portfolio: Mean 1.3, Standard Deviation 0.7, Volatility 1.9.
    - Private debt: Mean 6.1, Standard Deviation 0.5, Volatility 13.2 (note: table format in source; ensure preservation of numbers).
  - Overall volatility of capital flows fell in later years of the sample, but volatility of FDI increased (partly driven by large movements in South Africa in 2005-2006).
- Push and pull factors:
  - Global “push” factors: abundant global liquidity; world real GDP growth and rising oil and non-fuel commodity prices (figures shown for 2000–2007).
  - Domestic “pull” factors: improved macroeconomic performance across much of the continent (real GDP growth and rising per capita GDP), lower inflation, improved central government balances (figures shown for 2000–2007).
- Distributional patterns and concentration:
  - FDI shares (2000–03): South Africa 27.3, Nigeria 16.8, Angola 16.5, Equatorial Guinea 5.8, Chad 5.0, Other SSA 28.5.
  - FDI shares (2004–07): Nigeria 34.3, South Africa 15.5, Equatorial Guinea 10.2, Tanzania 3.2, Congo, Republic of, 2.9, Other SSA 33.8.
  - Portfolio and debt shares (2000–03): South Africa 23.1, Kenya 7.4, Ethiopia 6.8, Ghana 6.8, Congo, Dem. Rep. of 6.3, Other SSA 49.6.
  - Portfolio and debt shares (2004–07): South Africa 72.8, Nigeria 5.6, Ghana 2.2, Angola 2.0, Kenya 1.7, Other SSA 15.7.
  - Top recipient patterns (averages 2000–07) highlighted country rankings by total inflows, inflows-to-GDP, and inflows-to-population ratios (figures show country lists including Angola, Kenya, Chad, Tanzania, South Africa, Nigeria, Equatorial Guinea, Ghana, Congo, Dem. Rep. of, Ethiopia).
- External vulnerability:
  - For low-income SSA countries, strong correlation between capital inflows and current account balances (inflows increase as current account improves and decrease as it deteriorates), magnifying swings in overall balance of payments (Figure 6a).
  - For oil producers, similar patterns depicted (Figure 6b).
  - Surge in private capital inflows contributed to a steady increase in international reserves for SSA countries (Figures 6c and 6d).

### Key empirical findings on determinants of private capital flows
- Main empirical result: capital market development, rather than basic property rights institutions, is more closely associated with higher private capital inflows to SSA—reflecting the rising importance of portfolio and debt flows relative to FDI.
- Financial market structure matters:
  - Constructed index of capital market development: sum of four dummy variables (exists and well-developed = 1) for treasury bills, treasury bonds, corporate bond markets, and equities markets; covers all SSA countries except Somalia.
  - Only 10 countries (21 percent of sample) have complete capital markets: Botswana, Ghana, Kenya, Namibia, Nigeria, South Africa, Swaziland, Tanzania, Uganda, Zambia.
  - Countries with more complete capital markets are among the largest recipients of private capital inflows, along with several oil exporters (Angola, Congo, Chad, Equatorial Guinea).
- Other determinants highlighted:
  - Market size is significant, particularly for portfolio inflows (dummy for South Africa or Nigeria).
  - Oil-exporting status is significant, particularly for FDI flows (oil-exporter dummy = 1 if country is oil exporter).
- Growth correlation:
  - Strong positive correlations found between private capital inflows and real GDP growth in recipient countries.
  - Causality not established: limited time period and endogeneity prevent proving that inflows cause increased growth.

### Estimation strategy and model specification
- Reduced-form linear equation estimated (notation preserved):
  - K_i = μ + α F_i + β P_i + γ χ_i + u_i
    - K_i: log of average inflows to country i.
    - F_i: measure of financial institutions.
    - P_i: measure of property rights.
    - χ_i: set of exogenous control variables.
    - u_i: independent error term.
  - α and β of particular interest.
- Endogeneity handled via two-stage least squares (2SLS):
  - First-stage equations (notation preserved):
    - F_i = ν_1 + δ_11 L_i + χ_1 C_i + ε_1i
    - P_i = ν_2 + δ_22 E_i + χ_2 C_i + ε_2i
    - L_i and E_i are instruments for financial institutions and property rights institutions respectively.
  - Exclusion restriction: Cov(u_i, L_i) = Cov(u_i, E_i) = 0.
- Instruments and justification:
  - Instrument for financial market development: British legal origin (from La Porta et al., 1999); common law vs. civil law variation shown to be associated with financial market development.
  - Instruments for property rights: ethno-linguistic fragmentation and latitude (from La Porta et al., 1999) — chosen for better country coverage in SSA versus settler mortality/population density in 1500.
  - Robustness note: population density in 1500 used in robustness checks yields consistent results (not shown).

### Empirical analysis — choice of variables and controls
- Core explanatory variables:
  - Financial institutions: simple index of capital market development (Lukonga, forthcoming).
  - Property rights institutions: rule of law score from World Governance Indicators (Kaufmann, Kray and Mastruzzi, 2008).
- Control variables and expected effects:
  - Oil-exporter dummy: expected to attract more FDI.
  - Market-size dummy (South Africa or Nigeria): expected to attract more capital flows.
  - Human capital: secondary school enrollment share (WDI) — expected positive effect.
  - De jure capital account openness: Chinn-Ito measure (average over 1999–2005) — expected positive effect.
  - Trade openness: sum of exports and imports over GDP — expected positive effect (capital mobility often develops alongside trade openness).
  - Macroeconomic performance: real GDP growth rate — expected positive correlation with inflows but causality difficult due to feedback from inflows to macro outcomes.
- Empirical concerns addressed:
  - Endogeneity of financial and property rights measures tackled by instruments and 2SLS.
  - Measurement issues noted for conventional proxies of financial development (e.g., money and credit to GDP ratios) and rationale for using the constructed capital market index.

### Literature context and theoretical considerations
- Lucas Paradox: literature attempts to explain why capital does not flow from rich to poor as predicted by neoclassical theory.
- Two broad explanations:
  - Country-specific fundamentals (missing factors such as human capital, poor policies, weak institutions) lowering marginal product of capital in developing countries.
  - International capital market imperfections (asymmetric information, sovereign risk, capital controls) limiting flows to poorer countries.
- Institutional distinctions:
  - Property rights (basic institutions) versus contracting/financial market institutions can have distinct effects on capital flows and economic outcomes (references to Acemoglu and Johnson, Alfaro et al., Faria and Mauro, Ju and Wei).
  - Theoretical models suggest inefficient financial systems can be bypassed by FDI, but weak property rights cannot be bypassed.
- Contribution: first study (to authors’ knowledge) to analyze determinants and growth association of private capital flows for whole SSA in the context of the 2000–2007 surge, with disaggregation by FDI, portfolio, and private debt.

*Source: _wp09114 - References..............................................................................................................*

### Appendix Table III.2 shows pairwise correlations between different measures of capital flows,

### _wp09114 - Appendix Table III.2 shows pairwise correlations between different measures of capital flows,

### Scaling and measurement
- Appendix Table III.2 shows pairwise correlations between different measures of capital flows.
- The choice of scaling variable affects the pairwise correlations of the inflows with the independent variables.
- The picture of the allocation of private capital inflows across sub-Saharan countries differs depending on whether one looks at:
  - overall volume of inflows, or
  - the size of the inflows relative to a country’s population or GDP.
- For the remainder of the section the variable of interest is:
  - the log of private capital inflows and their components, FDI on the one hand and portfolio and debt flows on the other, averaged over 2000-2007 for each sub-Saharan African country excluding Somalia and Zimbabwe.

### Determinants of private capital flows (Appendix Table III.2, columns 1-4)
- Overall inflows as well as portfolio and debt inflows are correlated with:
  - the index of financial sector development,
  - secondary school enrollment,
  - the market size dummy variable,
  - the log of initial GDP per capita.
- The measure of trade openness is never significantly correlated with private inflows or their components.
- The Chinn-Ito capital openness variable is never significantly correlated with private inflows or their components.

### FDI-specific correlations
- Determinants of FDI differ from those of portfolio and debt inflows:
  - GDP per capita and secondary school enrollment are not significantly correlated with FDI flows.
  - The proxy for property rights is correlated with FDI flows.
  - The dummy for oil producers is correlated with FDI flows.

### Capital account openness and institutional measures
- Capital account openness (Chinn-Ito) is correlated with:
  - the degree of financial sector development, as in Chinn and Ito (2005),
  - secondary school enrollment.
- Footnote on the Chinn-Ito measure:
  - The Chinn-Ito measure attempts to capture the intensity of the capital controls by incorporating all the restrictions in the index using a principal components approach.
  - Its advantage over other indices is its near universal country and extended time coverage: 181 countries between 1970 and 2005.

*Source: Appendix Table III.2 and accompanying text from the provided IMF content unit.*

### Appendix Table III.3 illustrates the challenges associated with finding truly exogenous

### _wp09114 - Appendix Table III.3 illustrates the challenges associated with finding truly exogenous

### Endogeneity, correlated regressors, and instrument validity
- Appendix Table III.3: several potential explanatory variables are correlated with each other; financial market development is highly correlated with proxies for property rights institutions; educational achievement is significantly correlated with institutional quality variables and economic development.
- Conclusion: strong endogeneity implies educational achievement and log of GDP per capita in 2000 are "probably not a useful explanatory variable" for disentangling the impact of institutions per se on capital inflows.
- Instruments examined: British legal origin, population density in 1500, and ethnolinguistic fragmentation. Pairwise correlations in Appendix Table III.4 are "encouraging for the instrumental variables approach."
- Finding on instruments: potential instruments for property rights institutions are uncorrelated with inflows variables but significantly correlated with property rights proxies. British legal origin is significantly correlated with total inflows in addition to financial institutions, but "it is not correlated with property rights nor its potential instruments."

### Baseline OLS estimation results (Table 3)
- Financial market development index: coefficient positive and significant and "robust across all specifications."
- For portfolio and debt flows: financial market development is "even the only variable that is significant."
- Rule of law (property rights institutions): "never significant and has the wrong sign."
- Controls: coefficients on country size and oil exporters are large and generally significant; oil exporters receive more FDI but not more portfolio and debt inflows.
- Real GDP growth: "the only other variable that is significant for total inflows and FDI," but "real GDP growth is not exogenous."
- Variables never significant in OLS: log of real GDP per capita in 2000, secondary enrollment, capital account and trade openness.
- Robustness: results similar when breaking sample into two sub-periods; robust to excluding South Africa and Nigeria and to excluding all oil exporters; "impact of financial market development on inflows is even larger in both cases."
- Selected exact coefficients and statistics as reported in Table 3 (preserved verbatim):
  - Financial Institutions: 0.50*** 0.56*** 0.52*** 0.55*** 0.56*** 0.58** 0.56*** 0.57*** 0.67*** 0.63*** 0.55*** 0.57*** 0.55***
  - Rule of Law: -0.21 -0.33 -0.41 -0.27 -0.28 -0.33 -0.37 -0.03 -0.10 -0.20 -0.18 -0.52 -0.58
  - South Africa and Nigeria: 1.85 1.87* 1.98 1.89 1.79 1.58 1.35 1.37 2.11 2.12
  - Oil Exporters: 1.61*** 1.01* 1.62*** 1.68*** 1.60*** 1.64*** 2.67*** 1.94*** 0.51 0.12
  - Real GDP growth: 0.11*** 0.14*** 0.07
  - Trade Openness: -0.01** (-2.02 t-stat reported)
  - Constant examples: 4.47*** 5.22*** 3.79*** 3.35*** 3.92*** 4.13*** 3.74*** 4.41*** 3.42*** 4.42*** 2.68*** 2.16*** 3.45*** 4.30 2.88*** 2.60***
  - Observations: 43 43 43 43 42 43 43 43 43 43 43 43 43 43 43 43 43 43 43
  - R-squared examples: 0.26 0.01 0.57 0.62 0.57 0.57 0.58 0.59 0.22 0.00 0.55 0.59 0.25 0.00 0.42 0.44
  - F-Statistic examples: 11.96 0.31 12.00 15.81 8.78 9.46 9.41 15.32 11.97 0.01 12.89 14.06 11.76 0.21 5.45 4.93
- Note: ***, **, and * indicate significant at 1, 5, and 10 percent. t-Statistics are reported in parentheses. Robust standard errors.

### Instrumental variables (2SLS) results (Table 4) — first stage and second stage
- First-stage highlights:
  - British legal origin: "strong positive and significant impact on financial development, but not on the rule of law."
  - Instrument for property rights (ethnolinguistic fractionalization or latitude): "stronger and mostly significant impact on property rights" (greater ethnolinguistic fragmentation has a negative impact on rule of law; closeness to the Equator also negative).
  - Market size: positive and significant impact on capital market development; market size is negatively related to rule of law (attributed to low ratings of Nigeria and South Africa).
  - Oil exporting status: negatively and significantly related to rule of law.
- Second-stage highlights and comparisons with OLS:
  - Impact of financial institutions is overall weaker in 2SLS (smaller coefficients) but "remains strong and significant for total inflows and FDI inflows, as well as the share of FDI in total inflows."
  - Property rights variable: "never significant and continues to have the 'wrong' sign in most cases, except for FDI inflows and the share of FDI (column 2)."
  - Oil resources: "a strong determinant of FDI inflows, but not other types of flows."
  - Market size: "seems to matter for portfolio and debt inflows."
- Robustness: excluding South Africa and Nigeria or excluding oil-rich countries yields broadly similar results; "coefficient on financial institutions larger particularly in the FDI regressions."
- Interpretation: in non-oil SSA countries excluding South Africa, "FDI inflows are significantly larger, the more financial markets are developed." Property rights protection remains insignificant; hypothesized explanations include difficulty of expropriating FDI or negotiated protections for foreign investors.
- Caution: "Given the small sample size, the results must be interpreted with caution."
- Conclusion from IV results: support for unbundling institutions and measuring them more precisely; "financial institutions as proxied by capital markets development matter more than basic property rights institutions."

### Do private capital flows have an impact on growth in SSA? — Correlation analysis
- Data: cross-section sample of average flows for the 2000-07 period; use of net flows as well; normalization by population or GDP.
- Scatter-plot evidence: "generally positive correlation between growth and capital flows"; relationship more pronounced for FDI.
- Table 5 correlations (preserved values):
  - Real GDP Growth with Net Total Flows/GDP: 0.38* (1.00 / 0.38* shown)
  - Real GDP Growth with FDI Flows/GDP: 0.66*
  - Debt Flows/GDP correlation entries: -0.02 0.72* -0.07 1.00
  - Portfolio Flows/GDP correlations: -0.06 0.31* -0.10 0.02 1.00
  - Per capita correlations: Real GDP Growth with Net Total Flows per capita: 0.37*; with FDI Flows per capita: 0.44*; with Debt Flows per capita: 0.08; with Portfolio Flows per capita: -0.07
- Note: * indicates significant at 10 percent.

### Channels of transmission tested (Table 6) — overvaluation, savings & investment, institutions
- Methodology: two-stage bivariate regressions where first stage regresses channel variable on capital flows; second stage regresses growth on predicted channel value.
- Table 6 reported coefficients (preserved verbatim):
  - Overvaluation: Net flows/GDP -0.004; Net flows per capita -0.00003
  - Investment: Net flows/GDP 0.164; Net flows per capita 0.012 ***
  - Domestic Savings: Net flows/GDP 0.337; Net flows per capita 0.021 **
  - Institutions: Net flows/GDP -0.214; Net flows per capita -0.0042
- Interpretation:
  - In most instances capital flows are not significantly correlated with the channel variables.
  - Exception: when capital flows are scaled to population, both savings and investment are positively and significantly related to capital flows.
  - Implication: for SSA countries, the impact of private capital flows on growth could be partly intermediated through net capital flows' impact on domestic savings and investment.
  - Unexpected result: institutional variable is negatively related to growth (holds even when oil-exporters are excluded).

### Multivariate growth regressions (Table 7)
- Model: cross-country OLS using averages for 2000-07; g_i = α + β·K_i + η·X_i + u_i; controls guided by channels analysis.
- Controls and results (preserved verbatim from Table 7):
  - Initial per capita GDP (log): -1.53*** -1.86*** -1.67*** -2.16*** (t-statistics reported)
  - Overvaluation: -0.82 -1.53 -0.68 -1.25
  - Domestic savings: 0.09*** 0.10*** 0.09*** 0.10*** (t-statistics reported: 3.39, 3.59, 3.61, 3.62)
  - Investment: 0.18*** 0.16*** 0.15*** 0.17*** (t-statistics reported: 3.33, 2.93, 2.58, 2.97)
  - Contract enforcement: 0.11 0.09 0.09 0.08
  - Net total flows/GDP: 0.11* (net total flows/GDP t-statistic shown as -1.94 in the table layout)
  - Net total flows per capita: 0.003** (t-statistic 2.05)
  - Total inflows/GDP: 0.18*** (t-statistic 2.27)
  - Total inflows per capita: 0.003 (t-statistic 1.47)
  - Constant examples: -1.06 4.12 0.54 6.29
  - Adjusted R-squared: 0.51 0.52 0.53 0.57
  - Number of observations: 42 42 42 42
- Interpretation:
  - Capital flows are positively and significantly associated with growth in most specifications.
  - Initial per capita GDP negative and significant: supports convergence.
  - Domestic savings and investment significant with expected signs.
  - Contract enforcement (proxy for institutions) not significant.
  - Overvaluation has expected sign but is never significant.
  - Results similar when using net FDI instead of total net private capital, with even stronger association for FDI.
- Caveats: short time period since inflows began; many inflows occurred late in 2000-07; ordinary least squares limitations and endogeneity concerns remain.

### Conclusion and policy recommendations
- Near-term context: global financial crisis and recession have caused declines in private capital flows; investors likely to remain more risk-averse for some time; donor flows to low-income countries likely to be adversely impacted.
- Empirical conclusion: results "were inconclusive regarding the impact of private capital flows on growth in SSA countries"; additional observations and research needed.
- Policy implications emphasized in the text:
  - SSA countries likely to remain dependent on financial inflows to finance current account deficits associated with development-oriented investments.
  - Policymakers in SSA should pay increasing attention to attracting private capital.
  - Key policy recommendation: "strengthen domestic capital markets" — develop deep and well-functioning money, foreign exchange, equity and (where relevant) derivative markets to attract private capital.
  - Simultaneous priority: "put in place strong monitoring, regulatory and supervisory frameworks, to avoid a buildup of balance-sheet vulnerabilities."

*Source: Authors' calculations.*

### Appendix I. Capital Flows Data

### Appendix I. Capital Flows Data

### Data sources, coverage, and period
- Data based on detailed balance of payments data obtained from AFR country desk economists in late 2007 for 44 sub-Saharan African countries.
- Data for 2000-2006 are actual; data for 2007 are preliminary estimates.
- Covers 43 sub-Saharan African countries (text alternately references 43 and 44 in different places).
- Period covers 2000-2007, the episode of private capital inflows to sub-Saharan Africa which "took off around 2005-2006".
- Starting point: World Economic Outlook (WEO) database; desk economists refined WEO figures and disaggregated components normally not used in aggregated WEO analysis.

### Balance of payments recording and definition of capital inflows (Appendix Box I.1)
- Financial-account perspective: changes in liabilities are positive (inflows) and changes in assets are negative (outflows). In practice both are reported net of disinvestment and can have any sign.
- BOP entries:
  - Increase (decrease) in liabilities to nonresidents entered as positive (negative).
  - Increase (decrease) in net foreign assets held by residents entered as negative (positive).
- Recording convention used in this study:
  - Capital inflows = sum of increases in liabilities to non-residents and of decreases in net foreign assets held by residents in a year.
  - Capital outflows = sum of decreases in liabilities to non-residents and of increases in net foreign assets held by residents in a year.
  - Net flows of capital = inflows (positive sign) + outflows (negative sign).
- Purpose: capture all transactions that give rise to an inflow of capital on both liability and asset sides and avoid negative "inflows".

### Distinctions and classification of private flows
- Data distinguishes between:
  - FDI,
  - portfolio (debt and equity combined),
  - private debt-creating flows,
  - and between inflows and outflows.
- Portfolio debt and equity are grouped together (each category individually too small to stand alone).
- FDI is kept separate from portfolio equity (unlike some other studies that aggregate FDI and portfolio equity).
- Rationale: interest in determinants of various types of flows; portfolio debt and equity require bond and equity market infrastructure, while straight FDI does not.

### Comparisons with other data sources and rationale for desk-based refinement
- Other sources with broad coverage: IMF WEO, IFS, Balance of payments data, World Bank Global Development Finance (GDF), Lane-Milesi-Ferretti (updated only to 2005 and subset of SSA).
- Current account data used in some studies (Prasad et al., 2007; Gourinchas and Jeanne, 2007) provide wide coverage but cannot disaggregate official vs private or transfers vs different capital flow types.
- WEO and GDF provide comprehensive coverage with few missing observations, but rely on Fund/Bank staff estimates to fill gaps (Dorsey et al., 2008).
- IFS and Balance of Payments Statistics rely on official reporting without staff estimates; Balance of Payments Statistics yearly publication shows a complete set of annual data for just 18 out of 44 countries in the region in the most recent database (mostly middle-income, two oil-exporters, and the largest low-income economies).
- Country desk economists' work: ensure series availability and consistency of components with main aggregates; checked consistency across published databases and found discrepancies mostly in the last few years of the sample and few revisions to historical numbers in the early 2000s for countries with data in various publications.
- Approach: did not create a separate database but refined available information to cover as many countries as possible, including countries reporting zeros (treated as information in the dependent variable).

### Key dataset features emphasized
- Timeliness: includes recent episode through 2007 (preliminary).
- Coverage: 43/44 SSA countries (text notes both figures).
- Detail: disaggregation by flow type (FDI; portfolio & debt combined) and direction (inflows/outflows).
- Inclusion of zero-flow observations as valid dependent-variable information.

---

### Appendix II: Variables — definitions and primary sources (selected highlights)
- Net total flows: FDI, portfolio and debt net flows. Units: Billions of US$. Source: IMF, African Department database.
- Net FDI: Net foreign direct investment flows (FDI in reporting economy, net of direct investment abroad). Units: Billions of US$. Source: IMF, African Department database.
- Net portfolio and debt flows: Net private portfolio investment and other private flows; includes equity and debt securities and all other net private debt-creating flows. Units: Billions of US$. Source: IMF, African Department database.
- Total inflows: FDI, portfolio and debt inflows. Units: Billions of US$. Source: IMF, African Department database.
- FDI Inflows: Foreign direct investment in the reporting economy. Units: Billions of US$. Source: IMF, African Department database.
- Portfolio and debt inflows: Portfolio investment liabilities (equity and debt securities liabilities) plus other private capital inflows. Includes all other private debt-creating flows (Liabilities, excludes monetary authorities and government). Units: Billions of US$. Source: IMF, African Department database.
- CPIA: Country Policy and Institutional Assessment, composite index (1 to 7). Source: World Bank.
- Financial Institutions index: sum of four dummy variables (Treasury bills, treasury bonds, corporate bonds, equity markets). Values: 0, 1, 2, 3, 4. Source: Lukonga (forthcoming) and authors' calculations.
- Rule of law: scale -2.5 to 2.5. Source: http://info.worldbank.org/governance/wgi/index.asp
- South Africa and Nigeria: dummy 0 or 1 (1 for South Africa and Nigeria). Source: Authors' calculations.
- Oil exporters: dummy 0 or 1 (1 for Angola, Cameroon, Chad, Congo, Rep. of, Equatorial Guinea, Gabon and Nigeria). Source: IMF, African Department database.
- Real GDP growth: Percent. Source: IMF, World Economic Outlook.
- Secondary enrollment: Percent. Source: World Bank, World Development Indicators.
- 2000 per capita GDP: Constant 2000 US dollars. Source: IMF, World Economic Outlook.
- Capital account openness: Index (Chinn-Ito). Source: Chinn-Ito, forthcoming.
- Trade openness: Percent of GDP. Source: IMF, World Economic Outlook.
- Overvaluation: Deviation from "equilibrium" between PPP and market rate, normalized so 1 is the world. Source: IMF, African Department database.
- GDP in dollars: Billions of US$. Source: IMF, World Economic Outlook.
- Inflation: Percent (Headline CPI, average annual percent change). Source: IMF, World Economic Outlook.
- Government balance: Percent of GDP (general government balance, excluding grants). Source: IMF, World Economic Outlook.
- Credit: Total credit to the private sector in percent of GDP. Source: IMF, International Financial Statistics.
- Current account: Percent of GDP (excluding grants). Source: IMF, African Department database.
- REER: Percent (real effective exchange rate, weighted index). Source: IMF, Information Notice System.
- U.S. Tbill rate: Percent. Source: IMF, International Financial Statistics.
- Terms of trade: Percent. Source: IMF, World Economic Outlook.
- Oil prices: Percent (simple average of Texas, Dubai and Brent prices, deflated by manufacturing unit cost index). Source: IMF, World Economic Outlook.
- Reserves: Billions of US$. Source: IMF, World Economic Outlook.
- Population: Millions. Source: IMF, World Economic Outlook.
- Contract enforcement: Number of procedures. Source: http://www.doingbusiness.org/
- Constraint on the executive: Index 1 to 7. Source: Freedom House / Polity IV.
- ICRG ranking: 0 to 100. Source: www.prsgroup.com

(Note: Appendix II in the source provides a comprehensive list of variables, descriptions, units, and sources; above are selected highlights preserved exactly as in the source.)

---

### Appendix III: Descriptive statistics, correlations, and sample moments (high-level highlights)
- Dependent-variable sample moments (Appendix Table III.1):
  - Total inflows: Observations 436; Minimum 6.3; Maximum 16,210.6; Mean 826.5; Median 233.9; Standard Deviation 2,541.0.
  - FDI inflows: Observations 430; Minimum 0.0; Maximum 4,237.1; Mean 359.8; Median 90.9; Standard Deviation 802.7.
  - Portfolio and debt inflows: Observations 430; Minimum 0.0; Maximum 13,072.4; Mean 466.6; Median 102.1; Standard Deviation 1,978.6.
  - Total inflows per capita: Observations 433; Minimum 5.3; Maximum 1,625.0; Mean 129.1; Median 25.5; Standard Deviation 307.8.
  - FDI inflows per capita: Observations 430; Minimum 0.0; Maximum 1,165.3; Mean 82.6; Median 10.5; Standard Deviation 232.5.
  - Portfolio and debt inflows p.c.: Observations 430; Minimum 0.0; Maximum 590.6; Mean 46.5; Median 14.7; Standard Deviation 103.8.
  - Total inflows to GDP: Observations 431; Minimum 0.1; Maximum 30.7; Mean 8.1; Median 6.2; Standard Deviation 6.5.
  - FDI inflows to GDP: Observations 430; Minimum 0.0; Maximum 29.0; Mean 4.2; Median 2.4; Standard Deviation 5.3.
  - Portfolio and debt inflows to GDP: Observations 430; Minimum 0.0; Maximum 18.7; Mean 3.9; Median 3.2; Standard Deviation 3.6.
- Explanatory-variable sample moments (selected):
  - Financial institutions: Observations 430; Min 0; Max 4; Mean 1.8; Median 1.0; Std Dev 1.6.
  - Rule of law: Observations 43; Min -1.8; Max 0.8; Mean -0.7; Median -0.7; Std Dev 0.6 (table structure shows varying observation counts across variables).
  - Trade openness: Observations 433; Min 6.9; Max 155.9; Mean 81.8; Median 78.9; Std Dev 30.4.
  - Secondary enrollment: Observations 428; Min 2.2; Max 107.9; Mean 34.2; Median 28.3; Std Dev 22.6.
- Appendix III tables present pairwise correlation coefficients and significance levels between capital-flow measures and a range of independent variables (financial institutions, rule of law, market size, oil exporter dummy, real GDP growth, secondary enrollment, log 2000 GDP per capita, capital account openness, trade openness, UK legal origin, ethnolinguistic fragmentation, latitude, population density). (Numerical matrices and pairwise significance levels are provided in the source tables.)

### Correlation and multicollinearity diagnostics (summary from Appendix Tables)
- Financial Institutions shows positive correlations with several institutional measures (e.g., CPIA, Property rights, Legal rights) and moderate correlations with trade and capital openness measures (see Appendix Table III.3 and III.4 for full matrices).
- Several institutional and instrument measures are strongly correlated with each other (e.g., CPIA and Property rights; Rule of Law and CPIA), indicating potential collinearity across institutional regressors (pairwise coefficients and significance levels reported in source).

---

### Appendix IV: Private capital flows and growth — sample moments
- Real GDP growth: Observations 43; Mean 4.8; Standard deviation 3.6; Min -0.4; Max 20.1.
- Initial per capita GDP: Observations 43; Mean 6.1; Standard deviation 1.1; Min 4.5; Max 9.0.
- Net total flows/GDP: Observations 43; Mean 3.7; Standard deviation 7.4; Min -10.9; Max 27.1.
- Net total flows per capita: Observations 43; Mean 54.7; Standard deviation 278.2; Min -500.6; Max 1,232.3.
- Total inflows/GDP: Observations 43; Mean 7.7; Standard deviation 5.7; Min 1.0; Max 28.2.
- Total inflows per capita: Observations 43; Mean 131.1; Standard deviation 313.9; Min 3.5; Max 1,645.0.
- Overvaluation: Observations 43; Mean 1.1; Standard deviation 0.3; Min 0.7; Max 1.7.
- Domestic savings: Observations 42; Mean 12.2; Standard deviation 20.0; Min -24.7; Max 81.3.
- Investment: Observations 42; Mean 21.6; Standard deviation 8.7; Min 8.7; Max 46.9.
- Contract enforcement: Observations 43; Mean 101.4; Standard deviation 6.6; Min 82.6; Max 113.3.

---

*Source: _wp09114 - Appendix I. Capital Flows Data (IMF authors' calculations and AFR country desk data as presented in the supplied content).*

### Appendix Table IV.1 Private Capital Flows and Growth: Summary Statistics for Multivariate

### Appendix Table IV.1 Private Capital Flows and Growth: Summary Statistics for Multivariate Regression

### Pairwise correlations (variables and correlations)
- Variables included: Real GDP growth; Initial per capita GDP; Netflows/GDP; Netflows per capita; Inflows/GDP; Inflows per capita; Overvaluation; Domestic savings; Investment; Contract enforcement.

### Key correlation coefficients (pairwise, as reported)
- Real GDP growth with:
  - Real GDP growth: 1.00
  - Initial per capita GDP: 0.00
  - Netflows/GDP: 0.38  *
  - Netflows per capita: 0.37  *
  - Inflows/GDP: 0.49  *
  - Inflows per capita: 0.28  *
  - Overvaluation: -0.32  *
  - Domestic savings: 0.48  *
  - Investment: 0.45  *
  - Contract enforcement: 0.18
- Initial per capita GDP with:
  - Initial per capita GDP: 1.00
  - Netflows/GDP: -0.13
  - Netflows per capita: 0.28  *
  - Inflows/GDP: 0.16
  - Inflows per capita: 0.63  *
  - Overvaluation: -0.21
  - Domestic savings: 0.58  *
  - Investment: 0.45  *
  - Contract enforcement: -0.07
- Netflows/GDP with:
  - Netflows/GDP: 1.00
  - Netflows per capita: 0.67  *
  - Inflows/GDP: 0.76  *
  - Inflows per capita: 0.42
  - Overvaluation: -0.11
  - Domestic savings: 0.13
  - Investment: 0.14  *
  - Contract enforcement: -0.24
- Netflows per capita with:
  - Netflows per capita: 1.00
  - Inflows/GDP: 0.66  *
  - Inflows per capita: 0.87  *
  - Overvaluation: -0.03
  - Domestic savings: 0.29  *
  - Investment: 0.39  *
  - Contract enforcement: -0.18
- Inflows/GDP with:
  - Inflows/GDP: 1.00
  - Inflows per capita: 0.61  *
  - Overvaluation: -0.24
  - Domestic savings: 0.28  *
  - Investment: 0.45  *
  - Contract enforcement: -0.17
- Inflows per capita with:
  - Inflows per capita: 1.00
  - Overvaluation: -0.16
  - Domestic savings: 0.48  *
  - Investment: 0.51  *
  - Contract enforcement: -0.13
- Overvaluation with:
  - Overvaluation: 1.00
  - Domestic savings: -0.36  *
  - Investment: -0.26  *
  - Contract enforcement: -0.15
- Domestic savings with:
  - Domestic savings: 1.00
  - Investment: 0.42  *
  - Contract enforcement: 0.10
- Investment with:
  - Investment: 1.00
  - Contract enforcement: -0.12
- Contract enforcement with:
  - Contract enforcement: 1.00

### Statistical significance note
- Note: * indicates significant at 10 percent.

*Source: Authors' calculations.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09114.pdf_