## 1. Net FDI inflows to Africa

## Source details

**Canonical URL:** [1. Net FDI inflows to Africa](https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025189-source-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2025/english/wpiea2025189-source-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2025/english/wpiea2025189-source-pdf.pdf.json)

---

### Introduction
- Objective: Assess whether Compact with Africa (CwA) membership has helped members attract foreign direct investment (FDI).
- CwA overview:
  - Launched in 2017 under Germany’s G20 Presidency to promote private investment in Africa through country-specific reform compacts, policy dialogue, and coordination with G20 partners and international financial institutions.
  - Initiative is demand-driven and open to all African countries.
  - Fourteen African countries have joined the initiative: Benin, Burkina Faso, Côte d’Ivoire, Democratic Republic of the Congo (DRC), Egypt, Ethiopia, Ghana, Guinea, Morocco, Rwanda, Senegal, Togo, Tunisia, and Zambia.
  - Platform provides technical assistance, structured policy dialogue, and peer engagement to support tailored, investment-focused reforms.
- Rationale for focusing on FDI:
  - FDI provides capital, technology, expertise, jobs, infrastructure improvements, and government revenues; strengthens trade linkages and foreign exchange reserves; supports diversification for developing countries.

### Stylized facts on FDI to Africa (authors’ calculations, UNCTAD)
- Trend and magnitude:
  - Inflows of FDI to Africa have increased substantially, growing by approximately factor of eight in real terms from the 1990s to the present.
  - The fastest rate of growth in FDI inflows occurred in the 1990s; growth continued thereafter at a slower pace.
  - Africa’s share in total world (net) FDI inflows remains low at around 3.5 percent.
- CwA members versus non-members:
  - FDI inflows to CwA member countries increased relative to the pre-CwA period.
  - CwA members have experienced a steeper increase in inward FDI and received higher FDI on average relative to non-CwA countries; however, this difference can also be observed in the pre-CwA period.
  - Excluding Egypt (largest economy and primary recipient of FDI in the CwA group) reduces the difference between the two groups.
  - From 2017 to the present, CwA countries have experienced somewhat larger and less volatile FDI inflows on average compared to others.

### Research question and empirical approach
- Core question: Has the CwA initiative helped its members attract investment?
- Sample and period:
  - Panel of 42 countries in Africa over the period 1990-2023.
  - Dataset includes 10 members of the CwA group in the empirical sample used for the main analysis.
- Empirical strategy:
  - Use entropy balancing methodology to improve covariate balance between treated (CwA) and control (non-CwA) countries.
  - Complement entropy balancing with country fixed effects and time fixed effects in a panel regression.
  - Treatment definition: participation in CwA from 2017 onwards (block treatment assignment in 2017 for all members except DRC and Zambia).
  - Outcome: FDI inflows (FDI per capita and variants; main regressions use FDI inflows as percent of GDP in the regression step).
- Formal estimand:
  - Average treatment effect on treated (τ):
    - τ = E[FDI(1) | P = 1] − E[FDI(0) | P = 1]
    - Rewritten conditional on covariates X = x:
      - τ = E[FDI(1) | P = 1, X = x] − E[FDI(0) | P = 0, X = x]

### Methodology (identification and implementation)
- Identification challenges:
  - Endogeneity of CwA participation and absence of a direct counterfactual.
  - Lack of parallel pre-treatment trends undermines a straight difference-in-differences (DiD) approach.
  - Participation is demand-driven; Regression Discontinuity Design or Instrumental Variables are not feasible.
- Chosen strategy: entropy balancing (Hainmueller, 2012)
  - Rationale:
    - Guarantees high degree of balance on selected covariates.
    - Avoids randomness and data loss from conventional matching by reweighting control observations rather than discarding them.
    - Non-parametric—does not require modeling selection into CwA.
    - Suitable for moderate number of treated units and limited pre-treatment periods.
  - Implementational details:
    - Two-step approach:
      1. Compute weights for control units to satisfy balance constraints (equal covariate means across treated and control groups).
      2. Use weights in a regression where dependent variable is FDI inflows as percent of GDP and main explanatory variable is CwA dummy (1 for CwA members from 2017 onwards; 0 otherwise), controlling for country and time fixed effects and selected controls.
  - Remaining concerns addressed:
    - Country fixed effects absorb time-invariant unobserved differences.
    - Time fixed effects control for global or region-wide shocks in each year.
    - Authors acknowledge that covariates used in balancing plus fixed effects must sufficiently capture selection into CwA and confounders.

### Data (variables, sample, controls)
- Sample composition:
  - 42 countries in Africa over 1990-2023.
  - Empirical sample for main analysis includes 10 CwA members: Benin, Burkina Faso, Egypt, Ghana, Guinea, Morocco, Rwanda, Senegal, Togo and Tunisia (CwA dummy = 1 for 2017-2023; otherwise 0).
  - Côte d’Ivoire and Ethiopia were excluded because of missing data for some covariates.
  - DRC and Zambia are included in the control group.
- Outcome variables:
  - Main dependent variable: Net FDI inflows as percent of GDP.
  - Other outcomes used in robustness: Net FDI per Capita; Net FDI inflows as % of Gross Fixed Capital Formation (GFCF).
  - Main data source for inward net FDI flows: UNCTAD database.
- Control variables (selected):
  - Macroeconomic factors:
    - Return on capital: inverse of GDP per capita (primary proxy); robustness checks use real interest rate as proxy.
    - Inflation: CPI inflation.
    - Natural resource dependence: natural resource rent.
    - GDP: PPP weighted GDP in international U.S. dollars to capture market size.
  - Economic and political institutions:
    - Trade openness: sum of exports and imports of goods and services as a share of GDP.
  - Additional covariates included in balancing/controls:
    - Urbanization rate; credit to private sector (% of GDP); regulatory quality; political stability; control of corruption; government effectiveness; access to electricity (% of population); life expectancy.
- Data sources:
  - World Bank Development Indicators and World Governance Indicators; Authors’ Calculations.

### Balance diagnostics
- Re-weighting outcomes:
  - Visual inspection (Figures 6 and 7) indicates re-weighting improved balance across most covariates.
- Standardized Mean Difference (SMD) test (Table 1) after re-weighting:
  - Urbanization 0.02
  - Life Expectancy 0.09
  - Inflation 0.07
  - Trade Openness 0.01
  - Electricity Access 0.05
  - Credit to Private Sector 0.05
  - Natural Resource Rent 0.15
  - Size of Economy 0.11
  - Regulatory Quality 0.04
  - Political Stability 0.01
  - Control of Corruption 0.03
  - Government Effectiveness 0.09
- Interpretation:
  - An SMD close to 0.1 and lower suggests good balance; most covariates meet this criterion after re-weighting.

### Main results
- Main regression (Dependent Variable: Net FDI Inflows as % of GDP) (Table 2):
  - CwA Participation: 1.613* (0.628)
  - RoI: 1.076 (1.424)
  - Num.Obs.: 748
  - RMSE: 3.92
  - Std.Errors by: Country; FE: Year; FE: Country: Yes
  - Note: * p < 0.05
- Interpretation:
  - Coefficient on CwA dummy is positive and statistically significant in main specification, suggesting CwA participation is associated with an improvement in net FDI inflows as % of GDP conditional on identifying assumptions.

### Robustness checks and sensitivity
- Alternate dependent variables and specifications (Table 3):
  - FDI Inflows Per Capita:
    - CwA Participation: 20.479 (16.044)
    - RoI: 3.911 (60.903)
    - Num.Obs.: 748; RMSE: 193.65
  - FDI Inflow (as % of GFCF):
    - CwA Participation: 4.936* (2.069)
    - RoI: 3.356 (4.685)
    - Num.Obs.: 748; RMSE: 14.56
  - Note: * p < 0.05
- Lagged dependent variable included (Table 4):
  - Dependent Variable columns: as % of GDP | Per Capita | as % of GFCF
  - CwA Participation: 0.725+ (0.413) | 4.662 (11.767) | 2.868 (1.764)
  - RoI: 0.120 (0.896) | -6.171 (35.726) | 1.931 (3.745)
  - Lagged Dependent Variable: 0.373** (0.121) | 0.543*** (0.049) | 0.161 (0.162)
  - Num.Obs.: 706; RMSE: 3.271 | 50.561 | 3.51
  - Note: + p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001
- Real interest rate as RoI proxy (Tables 5 and 6):
  - Table 5 (Dependent Variable: Net FDI Inflows as % of GDP / Per Capita / as % of GFCF):
    - CwA Participation: 0.010 (0.489) | 2.648 (7.149) | 1.029 (2.321)
    - Real Interest Rate: -0.012 (0.022) | -0.097 (0.388) | -0.078 (0.107)
    - Lagged Dependent Variable: 0.490*** (0.086) | 0.282*** (0.064) | 0.327** (0.112)
    - Num.Obs.: 453; RMSE: 3.741 | 24.791 | 4.36
    - Note: + p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001
  - Table 6 (Alternative sample):
    - CwA Participation: 1.241 (0.868) | 11.891 (9.749) | 3.848 (3.136)
    - Real Interest Rate: -0.012 (0.022) | -0.037 (0.398) | -0.063 (0.116)
    - Num.Obs.: 484; RMSE: 4.64 | 163.20 | 15.49
- Overall robustness conclusion:
  - Alternative specifications produce positive estimated coefficients on the CwA dummy in all models, but statistical significance is not consistent across specifications.
  - Authors conclude they are not able to establish a robust causal link between CwA participation and FDI inflows.

### Discussion and policy implications
- Reasons for non-robust significance:
  - Limited time since CwA launch; delays in full implementation of reforms.
  - Perturbations from COVID-19 and geopolitical conflicts may affect FDI flows.
- Synthesis of empirical literature on FDI drivers:
  - Stronger institutions, reliable infrastructure, and a productive workforce create a favorable environment for FDI.
  - Infrastructure reduces operational costs and increases productivity.
  - Human development (health and education) supports labor productivity and the investment climate; study uses life expectancy as a health proxy due to limited education data.
  - Stable legal frameworks, efficient regulatory bodies, corruption control, and stable economic policies reduce investor uncertainty.
- Policy recommendations for CwA members and G20:
  - Leverage CwA membership for peer-to-peer learning: sharing best practices, capacity building, policy dialogue, data sharing.
  - Collaborate on cross-border infrastructure projects, trade agreements, and investment initiatives to enhance regional competitiveness.
  - Use CwA to foster private sector development and public–private collaboration.
  - Establish strong coordination among G-20, participating African countries, and multilateral partners; create efficient communication channels.
  - Create robust monitoring and evaluation frameworks with consistent performance indicators to track progress and inform adjustments.
  - Maintain regular discussion forums for information exchange, experience sharing, and mutual learning.

### Conclusion
- Empirical takeaway:
  - CwA member countries experienced an increase in FDI inflows during the period of the CwA initiative.
  - Main specification shows a positive and statistically significant association between CwA participation and net FDI inflows as % of GDP, but alternative specifications do not consistently confirm statistical significance.
  - Authors cannot formally establish a statistically definitive causal effect of CwA on FDI inflows given current data and robustness checks.
- Forward-looking implications:
  - Continued structural reforms to strengthen institutions and governance, improve infrastructure, develop human capital, achieve macroeconomic stability, and improve business environment remain critical.
  - CwA has potential as a knowledge-sharing platform and a framework for tracking progress, comparing performance, and facilitating reform implementation.
  - Success depends on structured coordination, monitoring mechanisms, and sustained implementation over time.

### Annex — Key summary statistics (1990–2023, 42 countries)
- Dependent variables:
  - Net FDI/GDP (in percent): Mean 3.627; Std.Dev 4.994; Min -6.370; Max 40.882
  - Net FDI per Capita (in US dollars): Mean 110.948; Std.Dev 292.435; Min -731.997; Max 2625.599
  - Net FDI /GFCF (in percent): Mean 15.529; Std.Dev 17.469; Min -68.211; Max 95.641
- Macroeconomic factors:
  - Return on Capital (Real Interest Rate) (in percent): Mean -7.216; Std.Dev 1.097; Min -9.810; Max -4.830
  - Inflation (in percent): Mean 7.599; Std.Dev 27.482; Min -16.860; Max 557.202
  - Natural Resource Rent (in percent of GDP): Mean 10.810; Std.Dev 10.844; Min 0.002; Max 66.060
  - GDP per Capita (in current US dollars): Mean 5291.666; Std.Dev 6016.051; Min 291.989; Max 34219.060
  - Nominal GDP (PPP weighted in billion US dollars): Mean 76.479; Std.Dev 182.935; Min 0.274; Max 2121.760
- Economic and political institutions:
  - Trade Openness (in percent of GDP): Mean 70.578; Std.Dev 40.101; Min 4.128; Max 347.997
  - Urbanization: Mean 43.445; Std.Dev 17.924; Min 8.246; Max 89.741
  - Credit to Private Sector (in percent of GDP): Mean 24.233; Std.Dev 25.845; Min 0.003; Max 142.422
  - Regulatory Quality: Mean 0.389; Std.Dev 0.118; Min 0.100; Max 0.739
  - Political Stability: Mean 0.420; Std.Dev 0.164; Min -0.040; Max 0.757
  - Control of Corruption: Mean 0.390; Std.Dev 0.128; Min 0.155; Max 0.749
  - Government Effectiveness: Mean 0.373; Std.Dev 0.127; Min 0.070; Max 0.730
- Development factors:
  - Electricity Access (in percent of total population): Mean 47.752; Std.Dev 31.713; Min 1.300; Max 100.000
  - Life Expectancy (in years): Mean 61.229; Std.Dev 7.347; Min 41.681; Max 76.593

*Source: IMF Working Papers — Economic Impact of G20 Compact with Africa Initiative on Member Countries (excerpted content from wpiea2025189-source-pdf).*

### 1. Net FDI inflows to Africa ...........................................................................................

### 1. Net FDI inflows to Africa

### Introduction
- Objective: Assess whether Compact with Africa (CwA) membership has helped members attract foreign direct investment (FDI).
- CwA overview:
  - Launched in 2017 under Germany’s G20 Presidency to promote private investment in Africa through country-specific reform compacts, policy dialogue, and coordination with G20 partners and international financial institutions.
  - Initiative is demand-driven and open to all African countries.
  - Fourteen African countries have joined the initiative: Benin, Burkina Faso, Côte d’Ivoire, Democratic Republic of the Congo (DRC), Egypt, Ethiopia, Ghana, Guinea, Morocco, Rwanda, Senegal, Togo, Tunisia, and Zambia.
  - Platform provides technical assistance, structured policy dialogue, and peer engagement to support tailored, investment-focused reforms.
- Rationale for focusing on FDI:
  - FDI provides capital, technology, expertise, jobs, infrastructure improvements, and government revenues; strengthens trade linkages and foreign exchange reserves; supports diversification for developing countries.

### Stylized facts on FDI to Africa (from authors’ calculations, UNCTAD)
- Trend and magnitude:
  - Inflows of FDI to Africa have increased substantially, growing by approximately factor of eight in real terms from the 1990s to the present.
  - The fastest rate of growth in FDI inflows occurred in the 1990s; growth continued thereafter at a slower pace.
  - Africa’s share in total world (net) FDI inflows remains low at around 3.5 percent.
- CwA members versus non-members:
  - FDI inflows to CwA member countries increased relative to the pre-CwA period.
  - CwA members have experienced a steeper increase in inward FDI and received higher FDI on average relative to non-CwA countries; however, this difference can also be observed in the pre-CwA period.
  - Excluding Egypt (largest economy and primary recipient of FDI in the CwA group) reduces the difference between the two groups.
  - From 2017 to the present, CwA countries have experienced somewhat larger and less volatile FDI inflows on average compared to others.

### Research question and empirical approach
- Core question: Has the CwA initiative helped its members attract investment?
- Sample and period:
  - Panel of 42 countries in Africa over the period 1990-2023.
  - Dataset includes 10 members of the CwA group in the empirical sample used for the main analysis (see Data section).
- Empirical strategy:
  - Use entropy balancing methodology to address identification by improving covariate balance between treated (CwA) and control (non-CwA) countries.
  - Complement entropy balancing with country fixed effects and time fixed effects in a panel regression.
  - Treatment definition: participation in CwA from 2017 onwards.
  - Outcome: FDI inflows (FDI per capita and variants; main regressions use FDI inflows as percent of GDP in the regression step).

### Key introductory conclusions stated by authors
- Entropy balancing allows comparison of member countries to non-members that are as similar as possible in observable characteristics that explain FDI inflows, while controlling for country and time-specific trends.
- Limitations noted:
  - Entropy balancing does not account for time-invariant unobservable confounders and common shocks across all of Africa not captured by balancing variables; country and time fixed effects are used to mitigate this concern.
  - Relatively smaller size of the treatment group may affect stability of covariate moments and robustness of inference.

### Figures and data sources referenced
- Figures described (no image files provided here):
  - Figure 1. Net FDI inflows to Africa (Billion US$, at constant prices)
  - Figure 2. Net FDI inflows, by region (Billion US$, at constant prices)
  - Figure 3. FDI inflows to CwA members (Billion US$, at constant prices)
  - Figure 4. FDI inflows to CwA and non-CwA (Billion US$, at constant prices, average)
- Source cited for FDI data: UNCTAD, Authors’ Calculations

### Policy relevance stated
- Empirical results and policy discussion aim to help policymakers in receiving and investing countries to:
  - Better understand the effects of CwA.
  - Analyze key determinants of FDI in Africa.
  - Refine measures and strategies to enhance private sector participation and FDI in the region.

---

### II. Literature Review (key points)
- Role of FDI:
  - FDI has become one of the largest sources of external financing in developing countries, often exceeding official development assistance (World Bank, 2018).
  - Benefits include capital formation, productivity gains, technology transfer, human capital development, improved corporate governance, stronger competition, and contribution to SDGs (Aust et al., 2020; Borensztein et al., 1998; Emako et al., 2023).
  - Benefits depend on host country’s absorptive capacity (education, health, technological development, infrastructure, openness, regulatory frameworks).
- Determinants of FDI (literature consensus):
  - Market size, trade openness, infrastructure quality, labor cost, macroeconomic stability, human capital, growth prospects, political stability, legal protection, transparency, and dependence on natural resources (Islam and Beloucif, 2023; World Bank Global Investment Competitiveness Survey; Asiedu, 2006).
  - Urbanization can attract FDI (Bhattacharya et al., 2021).
- Prior evaluations of CwA:
  - Duodu et al. (2022): CwA membership promoted FDI and pollution (DiD and triple-difference, 2005-2019).
  - Fleuriet and Vertier (2024): Positive and statistically significant impact on GDP per capita; mixed or insignificant impacts on other variables (synthetic differences, 2003-2021).
  - Gbadegesin and Yameogo (2024): CwA associated with increased FDI and export growth but no clear translation into gross capital formation and GDP per capita (targeted maximum likelihood estimation).

---

### III. Methodology (detailed)
- Identification challenges:
  - Endogeneity of CwA participation and absence of a direct counterfactual.
  - Lack of parallel pre-treatment trends undermines a straight difference-in-differences (DiD) approach.
  - Participation is demand-driven; Regression Discontinuity Design or Instrumental Variables are not feasible.
- Chosen strategy: entropy balancing (Hainmueller, 2012)
  - Rationale:
    - Guarantees high degree of balance on selected covariates.
    - Avoids randomness and data loss from conventional matching by reweighting control observations rather than discarding them.
    - Non-parametric—does not require modeling selection into CwA.
    - Suitable for moderate number of treated units and limited pre-treatment periods.
  - Implementational details:
    - Two-step approach:
      1. Compute weights for control units to satisfy balance constraints (equal covariate means across treated and control groups).
      2. Use weights in a regression where dependent variable is FDI inflows as percent of GDP and main explanatory variable is CwA dummy (1 for CwA members from 2017 onwards; 0 otherwise), controlling for country and time fixed effects and selected controls.
- Formal estimand:
  - Average treatment effect on treated (τ):
    - τ = E[FDI(1) | P = 1] − E[FDI(0) | P = 1]
    - Rewritten conditional on covariates X = x:
      - τ = E[FDI(1) | P = 1, X = x] − E[FDI(0) | P = 0, X = x]
- Addressing remaining concerns:
  - Country fixed effects absorb time-invariant unobserved differences (e.g., geographic endowment).
  - Time fixed effects control for global or region-wide shocks in each year.
  - Authors acknowledge remaining identifying assumption: covariates used in balancing plus fixed effects sufficiently capture selection into CwA and confounders.

---

### IV. Data (variables, sample, and controls)
- Sample:
  - 42 countries in Africa over 1990-2023.
  - Empirical sample includes 10 CwA members for the main analysis: Benin, Burkina Faso, Egypt, Ghana, Guinea, Morocco, Rwanda, Senegal, Togo and Tunisia for the years 2017-2023; otherwise the CwA dummy is 0.
  - Note in the text: Burkina-Faso and Togo joined CwA in 2018, the DRC joined in 2023 and Zambia in 2025; all other countries joined in (text truncated).
- Outcome variables:
  - Main dependent variable referenced in regression step: FDI inflows as percent of GDP.
  - Main data source for inward net FDI flows: UNCTAD database.
  - Also referenced earlier: FDI per capita used as a dependent variable in some parts of the analysis.
- Control variables (selected based on FDI literature) grouped into three areas:
  (i) Macroeconomic factors
    - Return on capital: proxied by inverse of GDP per capita (following Asiedu, 2002); implies poorer countries have higher marginal product of capital. Robustness checks use the real interest rate as a proxy for return on capital.
    - Inflation: CPI inflation as measure of macroeconomic stability.
    - Natural resource dependence: natural resource rent.
    - GDP: PPP weighted GDP in [international] U.S. dollars to capture market size.
  (ii) Economic and political institutions
    - Trade openness: sum of exports and imports of goods and services as a share of GDP.
  (iii) (Additional controls referenced in methodology and annex but not fully listed in the supplied excerpt.)

---

*Source: IMF Working Papers — Economic Impact of G20 Compact with Africa Initiative on Member Countries (excerpted content from wpiea2025189-source-pdf).*

### 2017. We assume block treatment assignment in 2017 for all members, except the DRC and Zambia, which we include in the

### wpiea2025189-source-pdf - 2017. We assume block treatment assignment in 2017 for all members, except the DRC and Zambia, which we include in the

### Data and empirical strategy
- Sample composition and exclusions:
  - Final model includes ten CwA countries; Côte d’Ivoire and Ethiopia were excluded because of missing data for some of the covariates.
  - DRC and Zambia are included in the control group (block treatment assignment assumed in 2017 for all members except DRC and Zambia).
- Data sources:
  - World Bank Development Indicators and World Governance Indicators; Authors’ Calculations.
- Covariates and measures:
  - Demographic: urbanization rate.
  - Financial sector development: credit to private sector in percent of GDP.
  - Institutional quality: regulatory quality, political stability, control of corruption, government effectiveness.
  - Physical infrastructure: access to electricity in percent of population.
  - Human capital: life expectancy.
  - Return on Investment (RoI) proxies: inverse of GDP per capita and real interest rate.
- Empirical approach:
  - Step 1: Entropy balancing to re-weight control (non-CwA) observations to match treated (CwA) means across covariates.
  - Step 2: Two-way fixed effects regressions using entropy-balancing weights. Main estimating equation:
    - FDI_it = CwA_it + RoI_it + α_i + γ_t + ε_it
  - RoI excluded from balancing due to multicollinearity with size of economy; RoI included as control in regressions.
  - Outcome variables analyzed: Net FDI inflows as % of GDP (main), Net FDI inflows per capita, Net FDI inflows as % of Gross Fixed Capital Formation (GFCF).
- Acknowledged limitations:
  - Lucas Paradox noted; MPK estimates exist (e.g., Lowe et al. (2019)) but authors rely on simpler proxies (inverse of GDP per capita, real interest rate) given data limitations and structural assumptions in MPK estimation.
  - Time span since CwA launch may be short; global shocks (COVID-19, geopolitical conflicts) may perturb FDI flows.

### Balance diagnostics
- Visual inspection:
  - Figures 6 and 7 indicate re-weighting improved balance across most covariates.
- Standardized Mean Difference (SMD) test (Table 1):
  - Urbanization 0.02
  - Life Expectancy 0.09
  - Inflation 0.07
  - Trade Openness 0.01
  - Electricity Access 0.05
  - Credit to Private Sector 0.05
  - Natural Resource Rent 0.15
  - Size of Economy 0.11
  - Regulatory Quality 0.04
  - Political Stability 0.01
  - Control of Corruption 0.03
  - Government Effectiveness 0.09
- Interpretation:
  - An SMD close to 0.1 and lower suggests good balance; most covariates meet this criterion after re-weighting.

### Main results
- Main regression (Dependent Variable: Net FDI Inflows as % of GDP) (Table 2):
  - CwA Participation: 1.613* (0.628)
  - RoI: 1.076 (1.424)
  - Num.Obs.: 748
  - RMSE: 3.92
  - Std.Errors by: Country; FE: Year; FE: Country: Yes
  - Note: * p < 0.05
- Interpretation:
  - Coefficient on CwA dummy is positive and statistically significant in main specification, suggesting CwA participation is associated with an improvement in net FDI inflows as % of GDP conditional on identifying assumptions.

### Robustness checks and sensitivity
- Alternate dependent variables and specifications:
  - FDI Inflows Per Capita and FDI Inflow (as % of GFCF) (Table 3):
    - CwA Participation: 20.479 (16.044) for Per Capita; 4.936* (2.069) for % of GFCF
    - RoI: 3.911 (60.903) for Per Capita; 3.356 (4.685) for % of GFCF
    - Num.Obs.: 748; RMSE: 193.65 (Per Capita), 14.56 (% GFCF)
    - Note: * p < 0.05
  - Lagged dependent variable included (Table 4):
    - Dependent Variable columns: as % of GDP | Per Capita | as % of GFCF
    - CwA Participation: 0.725+ (0.413) | 4.662 (11.767) | 2.868 (1.764)
    - RoI: 0.120 (0.896) | -6.171 (35.726) | 1.931 (3.745)
    - Lagged Dependent Variable: 0.373** (0.121) | 0.543*** (0.049) | 0.161 (0.162)
    - Num.Obs.: 706; RMSE: 3.271 | 50.561 | 3.51
    - Note: + p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001
  - Real interest rate as RoI proxy (Tables 5 and 6):
    - Table 5 (Dependent Variable: Net FDI Inflows as % of GDP / Per Capita / as % of GFCF):
      - CwA Participation: 0.010 (0.489) | 2.648 (7.149) | 1.029 (2.321)
      - Real Interest Rate: -0.012 (0.022) | -0.097 (0.388) | -0.078 (0.107)
      - Lagged Dependent Variable: 0.490*** (0.086) | 0.282*** (0.064) | 0.327** (0.112)
      - Num.Obs.: 453; RMSE: 3.741 | 24.791 | 4.36
      - Note: + p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001
    - Table 6 (Dependent Variable: Net FDI Inflows as % of GDP / Per Capita / as % of GFCF with alternative sample):
      - CwA Participation: 1.241 (0.868) | 11.891 (9.749) | 3.848 (3.136)
      - Real Interest Rate: -0.012 (0.022) | -0.037 (0.398) | -0.063 (0.116)
      - Num.Obs.: 484; RMSE: 4.64 | 163.20 | 15.49
- Overall robustness conclusion:
  - Alternative specifications produce positive estimated coefficients on the CwA dummy in all models, but statistical significance is not consistent across specifications. Authors conclude they are not able to establish a robust causal link between CwA participation and FDI inflows.

### Discussion and policy implications
- Interpretation of non-robust significance:
  - Lack of consistent statistical significance may reflect limited time since CwA launch, delays in full implementation of reforms, and perturbations from COVID-19 and geopolitical conflicts.
- Empirical literature synthesis on FDI drivers:
  - Stronger institutions, reliable infrastructure, and productive workforce create favorable environment for FDI.
  - Infrastructure reduces operational costs and increases productivity.
  - Human development (health and education) supports labor productivity and investment climate; study uses health (life expectancy) as proxy due to limited education data.
  - Stable legal frameworks, efficient regulatory bodies, corruption control, and stable economic policies reduce investor uncertainty.
- Policy recommendations for CwA members and G20:
  - Leverage CwA membership for peer-to-peer learning: sharing best practices, capacity building, policy dialogue, data sharing.
  - Collaborate on cross-border infrastructure projects, trade agreements, and investment initiatives to enhance regional competitiveness.
  - Use CwA to foster private sector development and public–private collaboration.
  - Establish strong coordination among G-20, participating African countries, and multilateral partners; create efficient communication channels.
  - Create robust monitoring and evaluation frameworks with consistent performance indicators to track progress and inform adjustments.
  - Maintain regular discussion forums for information exchange, experience sharing, and mutual learning.

### Conclusion
- Empirical takeaway:
  - CwA member countries experienced an increase in FDI inflows during the period of the CwA initiative.
  - Main specification shows a positive and statistically significant association between CwA participation and net FDI inflows as % of GDP, but alternative specifications do not consistently confirm statistical significance.
  - Authors cannot formally establish a statistically definitive causal effect of CwA on FDI inflows given current data and robustness checks.
- Forward-looking implications:
  - Continued structural reforms to strengthen institutions and governance, improve infrastructure, develop human capital, achieve macroeconomic stability, and improve business environment remain critical.
  - CwA has potential as a knowledge-sharing platform and a framework for tracking progress, comparing performance, and facilitating reform implementation.
  - Success depends on structured coordination, monitoring mechanisms, and sustained implementation over time.

### Annex — Key summary statistics (1990–2023, 42 countries) (Table 1)
- Dependent variables:
  - Net FDI/GDP (in percent): Mean 3.627; Std.Dev 4.994; Min -6.370; Max 40.882
  - Net FDI per Capita (in US dollars): Mean 110.948; Std.Dev 292.435; Min -731.997; Max 2625.599
  - Net FDI /GFCF (in percent): Mean 15.529; Std.Dev 17.469; Min -68.211; Max 95.641
- Macroeconomic factors:
  - Return on Capital (Real Interest Rate) (in percent): Mean -7.216; Std.Dev 1.097; Min -9.810; Max -4.830
  - Inflation (in percent): Mean 7.599; Std.Dev 27.482; Min -16.860; Max 557.202
  - Natural Resource Rent (in percent of GDP): Mean 10.810; Std.Dev 10.844; Min 0.002; Max 66.060
  - GDP per Capita (in current US dollars): Mean 5291.666; Std.Dev 6016.051; Min 291.989; Max 34219.060
  - Nominal GDP (PPP weighted in billion US dollars): Mean 76.479; Std.Dev 182.935; Min 0.274; Max 2121.760
- Economic and political institutions:
  - Trade Openness (in percent of GDP): Mean 70.578; Std.Dev 40.101; Min 4.128; Max 347.997
  - Urbanization: Mean 43.445; Std.Dev 17.924; Min 8.246; Max 89.741
  - Credit to Private Sector (in percent of GDP): Mean 24.233; Std.Dev 25.845; Min 0.003; Max 142.422
  - Regulatory Quality: Mean 0.389; Std.Dev 0.118; Min 0.100; Max 0.739
  - Political Stability: Mean 0.420; Std.Dev 0.164; Min -0.040; Max 0.757
  - Control of Corruption: Mean 0.390; Std.Dev 0.128; Min 0.155; Max 0.749
  - Government Effectiveness: Mean 0.373; Std.Dev 0.127; Min 0.070; Max 0.730
- Development factors:
  - Electricity Access (in percent of total population): Mean 47.752; Std.Dev 31.713; Min 1.300; Max 100.000
  - Life Expectancy (in years): Mean 61.229; Std.Dev 7.347; Min 41.681; Max 76.593

### Annex — Countries included in the model (Table 2)
- CwA members:
  - Benin, Burkina Faso, Egypt, Ghana, Guinea, Morocco, Rwanda, Senegal, Togo, Tunisia, Equatorial Guinea, Eswatini, Gambia, Kenya, Lesotho, Libya, Madagascar, Mali, Mauritania, Mauritius, Mozambique, Namibia, Niger, Seychelles, Sierra Leone, South Africa, South Sudan, Uganda, Zimbabwe
- Non-CwA members:
  - Algeria, Angola, Botswana, Burundi, Cabo Verde, Cameroon, Central African Republic, Chad, Comoros, Djibouti, Equatorial Guinea, Eswatini, Gabon, Gambia, Guinea-Bissau, Kenya, Lesotho, Libya, Madagascar, Mali, Mauritania, Mauritius, Mozambique, Namibia, Niger, Seychelles, Sierra Leone, South Africa, South Sudan, Uganda, Zambia, Zimbabwe

*IMF WORKING PAPERS Economic Impact of G20 Compact with Africa Initiative on Member Countries — INTERNATIONAL MONETARY FUND*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025189-source-pdf.pdf_