## wp18102

## Source details

**Canonical URL:** [wp18102](https://www.imf.org/-/media/files/publications/wp/2018/wp18102.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2018/wp18102.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2018/wp18102.pdf.json)

---

### Abstract and scope
- Remittances amounted to over $400billion in the year 2015 and are second only to foreign direct investment in size among international financial flows.
- During 2015, around 30 countries received remittance transfers worth more than five percent of GDP, and many more countries received remittances worth more than one percent of GDP.
- The paper presents cross-country evidence on remittances' impact on labor market outcomes, comparing effects to those of foreign direct investment (FDI) and official development aid (ODA).
- JEL Classification Numbers: D33, E24, E26, F24, J21, J23
- Keywords: Remittances, fragile countries, low income countries, middle income countries, Dutch Disease, labor markets, inclusive growth

### Main contributions of the paper
- Comprehensive cross-country analysis of multiple labor-market outcomes (unemployment, labor force participation, wage growth, inequality) rather than single-outcome, single-country household studies.
- Estimation of remittance effects on sectoral employment across 14 sectors using ILO Global Wage Database, ILO Global Employment Trends, and Sectoral Employment databases.
- Measurement of heterogeneity in effects across geographic regions, country income levels, and degree of fragility.
- Robust evidence reconciling mixed prior findings by showing simultaneous impacts on labor supply and labor demand.

### Key empirical findings — aggregate effects
- Remittances have significant impacts on both labor supply and labor demand; effects are larger and more statistically significant than those of FDI or ODA.
- Labor supply effects:
  - Remittances reduce labor force participation.
  - Remittances increase informality in the labor market.
  - Male and female labor supply show significantly different sensitivities to remittances.
- Labor demand effects:
  - Remittances reduce overall unemployment.
  - Employment increases predominantly in lower-wage, lower-productivity nontradables industries and declines in high-productivity, high-wage tradables sectors.
- Distributional and macro labor outcomes:
  - Larger remittances are associated with a decline in measured inequality.
  - Average wage and productivity growth decline; productivity declines more strongly than wages, leading to an increase in the labor income share.
- Fragile states:
  - In fragile states, remittances increase wage growth and do not depress labor force participation — opposite to effects in more stable states.
- Regional variation:
  - The impact of remittances on labor-market informality is greater in regions where informality is lower.
- Interpretation:
  - Evidence is consistent with a Dutch Disease narrative: remittance inflows shift relative prices toward nontradables (less productive labor), raising nontradables employment and reducing tradables employment, explaining simultaneous declines in unemployment and wage/productivity growth.

### Policy implications and development strategy
- Remittances are not an unmitigated boon; they have complex and heterogeneous effects via multiple channels on recipients' behavior and broader labor markets.
- Countries receiving significant remittance flows need to integrate remittance-aware strategies into overall development and inclusive growth plans.
- Reforms aimed at fostering inclusive growth should account for remittances' role in:
  - labor force participation dynamics,
  - sectoral composition of employment (tradables vs nontradables),
  - informality and wage-productivity dynamics,
  - differing effects across gender and fragility status.
- Suggested policy-relevant areas for investigation: investments in infrastructure, reforms to improve the ease of doing business, changes in tax policy.

### Data sources and methodology
- Remittances: "Personal Transfers" from IMF Balance of Payments Statistics.
- FDI: IMF Balance of Payments Statistics.
- ODA: OECD International Development Statistics database.
- Employment: ILO Global Employment Trends database; covers 177 countries between 1991 and 2015; balanced panel with approximately half observations imputed.
- Wages: ILO Global Wage Database augmented by ILO Wage Projection database; panel of 112 countries for 1995–2014.
- Inequality: Gini coefficients from the Standardized World Income Inequality Database (SWIID); focus on market income inequality.
- Estimation approach:
  - Dynamic model Y_it = β·Y_it−1 + γ·Remittances_it + δ·X_it + ε_it.
  - Arellano-Bond (system) GMM estimator used; Sargan test to choose lag structure; β = 0 assumed for sectoral employment shares (fixed-effects OLS used).
  - Quantile regressions used to study distributional heterogeneity.

### Labor demand — unemployment (selected results)
- Unemployment declines significantly and strongly with a rise in the share of remittances.
- Remittances coefficients (selected, remittances share in % of GDP):
  - -4.114** (1.818), -4.375*** (1.657), -6.695** (2.773), -6.954*** (2.554), -4.234** (1.877), -11.62** (5.278), -6.866** (2.769)
- Unemployment (lagged) coefficients (selected):
  - 0.503*** (0.137), 0.404*** (0.0397), 0.426*** (0.0474), 0.420*** (0.118), 0.494*** (0.0559), 0.464*** (0.136)
- Sample and diagnostics (selected):
  - Observations vary by specification: 2,284; 2,153; 1,826; 1,804; 1,725; 1,929; 1,361; 1,331.
  - Number of countries vary: 139; 132; 120; 120; 115; 117; 124; 118.
  - Number of instruments vary: 69; 69; 143; 131; 131; 121; 96; 5.

### Labor supply — labor-force participation (selected results)
- Remittances enter negatively and robustly across dynamic specifications and samples.
- Remittances (as a share of GDP) coefficients (selected):
  - -6.220*** (2.171), -4.302*** (1.374), -3.191** (1.429), -2.678** (1.350), -2.358** (1.085), -2.292** (1.164), -6.221*** (1.393), -7.055*** (1.792), -3.817*** (1.232)
- Labor-force participation (lagged) coefficients (selected):
  - 0.737*** (0.0961), 0.789*** (0.0628), 0.892*** (0.0474), 0.900*** (0.0476), 0.926*** (0.0347), 0.934*** (0.0346)
- Official aid coefficients (selected): 0.0426*** (0.0153), 0.0296** (0.0148), 0.0314** (0.0130), 0.0276** (0.0127).
- FDI generally not significant for labor-force participation in most specifications.

### Gender-differentiated effects
- Remittances depress labor-force participation more for women than for men.
- Male participation: remittances (as share of GDP) coefficients across nine specifications include: 0.854*, 0.535, -1.493***, -2.653***, -0.734*, -0.839**, -1.567*, -2.013**, -0.529.
- Female participation: remittances (as share of GDP) coefficients across nine specifications include: -4.527, -7.915***, -5.322**, -3.618**, -1.607, -1.619, -7.807***, -6.251***, -4.552**.
- Participation-gap coefficient on remittances significant only in simplest specifications (results available from authors).

### Wages, labor-income share, and inequality
- Wages:
  - Lagged remittances (in % of GDP) exert downward pressure on aggregate wage growth; coefficients in Table 5 include: -30.82**, -35.15***, -34.16**, -29.74**, -31.15**.
  - Other wage-related coefficients (selected): real wage growth (lagged) up to 0.252***; investment share positive and significant.
- Labor-income share:
  - Remittances associated with a positive and significant effect on changes in the labor-income share; remittances coefficients include: 6.205**, 7.681**, 7.047**, 7.656**, 7.104**.
  - Interpretation: productivity growth slows more than wage growth, raising labor-income share.
- Inequality:
  - Remittances enter significantly negatively in all specifications for market inequality.
  - Remittances coefficients (share of GDP): -3.284***, -2.135**, -3.452***, -3.394***, -1.824**, -1.961**, -1.801**, -1.561**.
  - Market inequality (lagged) coefficients: 0.815***, 0.865***, 0.836***, 0.835***, 0.878***, 0.858***, 0.873***, 0.853***.

### Informality and vulnerable employment
- Vulnerable employment proxy (own-account + contributing family workers) used as informality measure.
- Dynamic GMM estimates (Table 8) show a positive impact of remittances on the share of vulnerable employment; remittances coefficients (selected across columns) include:
  - 10.21*** (1.904), 11.73*** (2.183), 31.67** (15.72), 12.01*** (2.208), 10.41*** (2.327), 11.35*** (2.446), 7.352*** (2.258).
- Informal employment (lagged) coefficients (selected): 0.131 (0.0844), 0.235*** (0.0800), 0.412** (0.206), 0.235*** (0.0805), 0.276*** (0.0794), 0.120 (0.122), 0.377*** (0.0780).
- Results robust to series of controls except when government debt is included (reduces observations).

### Sectoral shifts and structural implications
- Sectoral employment estimation across 14 sectors (levels, fixed-effects OLS) finds:
  - Employment flows out of agriculture and into service-oriented sectors.
  - No discernible push toward manufacturing job creation; manufacturing employment declines especially where initial manufacturing shares are high.
  - Sectors with increased employment: construction, real estate, transportation, accommodation and restaurants, some utilities, education, health.
- Quantile patterns:
  - Construction employment increases in countries with intermediate or low starting shares of construction employment.
  - Transportation benefits broadly across quantiles.
- Behavioral links:
  - Consistent with literature showing remittances reduce farmers' effort; agriculture employment declines.
  - Findings align with a Dutch-disease channel shifting resources to nontradables and lower-productivity services.

### Regional heterogeneity (selected patterns)
- Labor force participation:
  - Negative impact in all regions, but absolute size differs substantially by region.
- Informal employment:
  - Remittances increase informality more strongly in MENA, Sub-Saharan Africa, and Central and Western Asia.
  - Weaker impact in Eastern Europe, Southern Asia, Latin America and the Caribbean.
  - MENA average informality: 16 percent of total employment; Latin America and the Caribbean: 35 percent.
- Wage growth:
  - Positive effects visible among countries in South Asia and Sub-Saharan Africa and in relatively more affluent countries in Eastern Europe.
  - Negative effects dominate in Central and Western Asia and South-East Asia and the Pacific.

### Fragility and remittance effects
- Fragility classification: Fund for Peace Fragile States Index; countries in highest quintile of the Index between 2006 and 2017 labeled fragile; average score used for 2006–2017.
- Labor force participation by fragility:
  - Adverse effects of remittances on participation particularly strong for intermediate levels of fragility; effect declines for very high fragility but remains significant; effect disappears for most stable countries.
  - Number of countries by fragility quintile: 4, 25, 30, 29, 28 (from lowest to highest quintiles).
- Wage growth and inequality by fragility:
  - Impact of remittances on inequality does not seem to depend on degree of fragility (differences not statistically significant).
  - Real wage growth: remittances raise real wages in fragile countries but not in stable ones.
  - Explanation: fragile states tend to have lower trade openness and export shares, implying smaller tradables sectors and weaker Dutch-disease channels.

### Interpretation challenges and implied mechanisms
- Apparent tensions:
  - Remittances reduce unemployment while being associated with lower wage and productivity growth.
- Reconciliation via segmented labor-market and composition effects:
  - Remittances shift employment toward lower-productivity/lower-wage nontradables; if nontradables employment gains exceed tradables employment losses, unemployment falls while aggregate productivity and wage growth decline.
- Institutional considerations:
  - Remittances suspected of reducing the quality of institutions; given positive relation between governance and growth, this interaction requires further investigation.

### Conclusion and research needs
- Stylized facts:
  - Remittances reduce labor force participation and increase informality.
  - Remittances reduce overall unemployment.
  - Remittance effects on labor-market outcomes are larger and more significant than those of FDI and ODA.
  - Remittances lower measured income inequality but are associated with lower wage growth and higher labor-income shares, implying lower productivity growth.
- Further research required to:
  - Verify the stylized facts.
  - Discover other stylized facts relevant to labor-market effects of remittances.
  - Quantify price distortions and negative externalities remittances impose on formal labor markets and the tradables sector.
  - Determine which policy approaches most effectively mitigate negative impacts (infrastructure, ease of doing business, tax policy).
- Policy implication summary:
  - Because remittances may impair the tradables sector and increase informality, development policies should focus on improving competition and reducing informality.

### Appendix — selected summary statistics (exact figures)
- Working-age population (in%): #Observations 3,799; Mean 69.11; Std. Dev. 11.64; Min 3.0; Max 100.0
- Labor force participation rate (in%): #Observations 4,080; Mean 63.3; Std. Dev. 10.4; Min 9.9; Max 90.2
- Unemployment rate (in%): #Observations 4,080; Mean 9.3; Std. Dev. 6.8; Min 0.14; Max 40.2
- Real wage growth (p.a., in%): #Observations 2,315; Mean 2.3; Std. Dev. 6.1; Min -30.6; Max 47.0
- Share of informal employment (in%): #Observations 4,080; Mean 40.8; Std. Dev. 28.1; Min 0.3; Max 96.0
- Employment shares (in% of total employment):
  - Agriculture: #Observations 4,080; Mean 30.2; Std. Dev. 24.9; Min 0.2; Max 90.9
  - Manufacturing: #Observations 4,080; Mean 13.1; Std. Dev. 7.5; Min 0.2; Max 43.0
  - Construction: #Observations 4,080; Mean 6.3; Std. Dev. 4.2; Min 0.1; Max 46.7
  - Accommodation and restaurants: #Observations 4,080; Mean 3.2; Std. Dev. 2.6; Min 0.1; Max 20.3
  - Health and social work: #Observations 4,080; Mean 4.6; Std. Dev. 3.9; Min 0.1; Max 21.6
- Market GINI: #Observations 2,825; Mean 45.0; Std. Dev. 7.6; Min 18.0; Max 72.8
- Remittances (in% of GDP): #Observations 2,343; Mean 3.3; Std. Dev. 5.4; Min 0.04; Max 49.4
- FDI (in% of GDP): #Observations 3,227; Mean 5.6; Std. Dev. 30.7; Min -46.2; Max 726.8
- Official development aid (in% of GDP): #Observations 2,976; Mean 6.0; Std. Dev. 8.7; Min -0.7; Max 115.4
- Real GDP growth (p.a., in%): #Observations 4,080; Mean 3.8; Std. Dev. 6.4; Min -52.6; Max 147.7
- Investment share (in% of GDP): #Observations 3,602; Mean 23.0; Std. Dev. 11.0; Min 0.1; Max 227.5
- Trade openness (in% of GDP): #Observations 3,725; Mean 84.2; Std. Dev. 52.9; Min 4.9; Max 809.2
- real GDP per capita (in Int$ PPP): #Observations 3,736; Mean 9,158.8; Std. Dev. 14,457.9; Min 30.9; Max 112,429.4

*Prepared by Ralph Chami, Ekkehard Ernst, Connel Fullenkamp, and Anne Oeking; IMF Working Paper WP/18/102 (May 2018).*

### Section 1

### Are Remittances Good for Labor Markets in LICs, MICs and Fragile States? Evidence from Cross-Country Data

### Abstract and scope
- Remittances amounted to over $400billion in the year 2015 and are second only to foreign direct investment in size among international financial flows.
- During 2015, around 30 countries received remittance transfers worth more than five percent of GDP, and many more countries received remittances worth more than one percent of GDP.
- The paper presents cross-country evidence on remittances' impact on labor market outcomes, comparing effects to those of foreign direct investment (FDI) and official development aid (ODA).
- JEL Classification Numbers: D33, E24, E26, F24, J21, J23
- Keywords: Remittances, fragile countries, low income countries, middle income countries, Dutch Disease, labor markets, inclusive growth

### Main contributions of the paper
- Comprehensive cross-country analysis of multiple labor-market outcomes (unemployment, labor force participation, wage growth, inequality) rather than single-outcome, single-country household studies.
- Estimation of remittance effects on sectoral employment across 14 sectors using ILO Global Wage Database, ILO Global Employment Trends, and Sectoral Employment databases.
- Measurement of heterogeneity in effects across geographic regions, country income levels, and degree of fragility.
- Robust evidence reconciling mixed prior findings by showing simultaneous impacts on labor supply and labor demand.

### Key empirical findings
- Remittances have significant impacts on both labor supply and labor demand; effects are larger and more statistically significant than those of FDI or ODA.
- Labor supply effects:
  - Remittances reduce labor force participation.
  - Remittances increase informality in the labor market.
  - Male and female labor supply show significantly different sensitivities to remittances.
- Labor demand effects:
  - Remittances reduce overall unemployment.
  - Employment increases predominantly in lower-wage, lower-productivity nontradables industries and declines in high-productivity, high-wage tradables sectors.
- Distributional and macro labor outcomes:
  - Larger remittances are associated with a decline in measured inequality.
  - Average wage and productivity growth decline; productivity declines more strongly than wages, leading to an increase in the labor income share.
- Fragile states:
  - In fragile states, remittances impose a positive externality: remittances increase wage growth and do not depress labor force participation — opposite to effects in more stable states.
- Regional variation:
  - The impact of remittances on labor-market informality is greater in regions where informality is lower.
- Interpretation:
  - Evidence is consistent with a Dutch Disease narrative: remittance inflows shift relative prices toward nontradables (less productive labor), raising nontradables employment and reducing tradables employment, explaining simultaneous declines in unemployment and wage/productivity growth.

### Implications for policy and development strategy
- Remittances are not an unmitigated boon; they have complex and heterogeneous effects via multiple channels on recipients' behavior and broader labor markets.
- Countries receiving significant remittance flows need to integrate remittance-aware strategies into overall development and inclusive growth plans.
- Reforms aimed at fostering inclusive growth should account for remittances' role in:
  - labor force participation dynamics,
  - sectoral composition of employment (tradables vs nontradables),
  - informality and wage-productivity dynamics,
  - differing effects across gender and fragility status.

### Literature positioning
- Complements and expands migration and remittance literatures that typically use household-level data or focus on single outcomes.
- Relates to findings that migration can increase wages by reducing labor supply (e.g., findings for Lithuania, post-Soviet era, Poland) and to the literature on brain drain and differential skill outflows.
- Places its cross-country aggregate findings alongside single-country household studies that generally find remittances reduce household labor force participation.
- Notes some contrasting findings in the literature (e.g., Posso (2012) finding increased aggregate male labor force participation in a 25-year panel), and situates its multi-outcome, cross-country results as reconciling mixed evidence through sectoral composition effects.

*Prepared by Ralph Chami, Ekkehard Ernst, Connel Fullenkamp, and Anne Oeking; IMF Working Paper WP/18/102 (May 2018).*

### Section 2

### Section 2

### Labor-market channels and behavioral responses
- Migration and remittances affect occupational choice and the broad types of work (formal and informal employment, self-employment, unpaid work such as caring for family members or contributing labor on a family farm).
- Empirical findings cited:
  - Amuedo-Dorantes and Pozo (2006a): remittances reduce hours men spend in informal work and self-employment, and increase hours spent in informal work.
  - Binzel and Assaad (2011): remittances reduce paid work outside the home that women perform in Egypt.
  - Görlich et al. (2007): remittances cause women in Moldova to decrease paid work in favor of unpaid household work.
  - Cabegin (2006): migration lowers female labor force participation and increases household work.
  - Ivlevs (2016): remittances and emigration increase the share of informal employment in a sample of six transition economies (Kazakhstan, the FYR Macedonia, Moldova, Serbia, Tajikistan and Ukraine) using the Social Exclusion Survey conducted in 2009.
- Self-employment effects are mixed in the literature:
  - Amuedo-Dorantes and Pozo (2006b) and Demirgüç-Kunt et al. (2011): remittance receipt lowers likelihood of business ownership (Dominican Republic; Bosnia and Herzegovina).
  - Funkhouser (1992) and Stanley (2015): remittances increase likelihood of self-employment.
  - Edwards and Rodríguez-Oreggia (2009): remittances increase labor force participation for some women in Mexico, possibly via self-employment.
- Theory: remittances may loosen financing constraints enabling entrepreneurship, or may reduce necessity to become self-employed where employment opportunities are limited.
- Remittances may change effort on the job, especially in agriculture:
  - Khanal et al. (2015): remittance receipt among rural families in Nepal increases amount of land abandoned (left permanently fallow).
  - Damon (2010): remittance receipt increases land devoted to subsistence crops and reduces land devoted to cash crops.
- Education and child labor:
  - Studies finding remittances reduce child wage labor and increase spending on education include Edwards and Ureta (2003), Yang (2008), Calero et al. (2009), Acosta (2011), and Alcaraz et al. (2012).
  - "Brain gain" demonstration effects from migration of highly skilled workers may induce older students to obtain more education to migrate, but most research suggests this benefit is outweighed by brain drain effects.

### Aggregate patterns from existing literature (summary)
- Consistent story on aggregate labor-market effects of remittances:
  - Wages and productivity tend to increase as a result of emigration and remittances, consistent with a decrease in overall labor supply due to emigration.
  - Remittances decrease labor-force participation among recipients, which (ceteris paribus) tends to reduce labor supply and contribute to increased wages.
  - Remittances induce shifts away from formal employment toward informal and unpaid work, reducing supply of labor in formal employment and placing upward pressure on wages.
- Ambiguities and mixed findings:
  - Whether remittances reduce aggregate labor-force participation is unclear (Posso (2012) suggests possible increases).
  - Evidence on effects of emigration and remittances on self-employment, entrepreneurship, and the unemployment rate is mixed.
  - Effects of remittances on child labor and education are apparent, but these likely do not affect the labor market for adult workers.

### Transition to aggregated empirical analysis
- Given limited cross-country aggregated analysis in the literature, the paper turns to empirical exercises to confirm or contradict micro-evidence and to produce internally consistent stylized facts across labor-market outcomes.

### 3 Empirical assessment — objectives and outcomes analyzed
- Objective: compile a broad set of stylized facts regarding labor-market effects of remittances to suggest a consistent theoretical framework.
- Outcomes estimated:
  - Labor demand, proxied by unemployment.
  - Labor supply, measured by labor-force participation.
  - Wages.
  - Inequality.
  - Sectoral shifts in employment.

### 3.1 Data and methodology — data sources and choices
- Remittances:
  - Taken from category "Personal Transfers" in the IMF Balance of Payments Statistics; refers to current transfers by non-resident households to resident households.
  - Definition captures regular and unrequited private transfers from residents in one country to another.
- Foreign direct investment: IMF Balance of Payments Statistics.
- Official development assistance: OECD International Development Statistics database.
- Employment: ILO Global Employment Trends database; covers 177 countries between 1991 and 2015; contains labor-force participation, unemployment, employment, sectoral employment and self-employment.
  - Balanced panel with approximately half observations imputed using statistical estimates based on Okun's-law relations between employment and GDP growth.
  - Estimations carried out using both full database (real and imputed) and only real observations.
- Wages: ILO Global Wage Database augmented by ILO Wage Projection database; panel of 112 countries for period 1995 to 2014; series chosen to cover wide range of sectors and regions within a country.
- Inequality: Gini coefficients from the Standardized World Income Inequality Database (SWIID); focus on market income inequality to avoid bias from cross-country differences in redistribution.
- Sample notes: also present estimates on smaller sample containing only non-OECD countries to avoid bias from advanced-economy institutional specifics.
- Summary statistics: referenced in Appendix Table 9 (not reproduced here).

Methodology
- Labor-market indicators are persistent at annual frequency; dynamic adjustment models used where possible.
- Estimated dynamic model:
  - Y_it = β·Y_it−1 + γ·Remittances_it + δ·X_it + ε_it
    - Y_it: relevant labor-market indicator.
    - Remittances_it: remittances as a share of GDP.
    - X_it: vector of control variables (vary by dependent variable; include level of development, GDP growth, investment share of GDP, demographic variables such as share of working-age population).
  - When wages are dependent variables, both wage curves and wage-inflation curves estimated, adding unemployment to independent variables.
- Estimator:
  - Arellano-Bond (system) GMM estimator using lagged dependent variables to address endogeneity.
  - Sargan test used to identify lag structure; shortest possible lag chosen to limit instruments.
  - Overidentifying restrictions validity typically led to a small number of instruments.
  - GMM specification allows low-level autocorrelation in the error term.
- Sectoral employment shares:
  - Dynamic treatment infeasible; assumed β = 0 in equation (1) and estimated using standard fixed-effects OLS.
- Quantile regressions performed to study distributional heterogeneity; reported graphically with confidence intervals (not reproduced here).

### 3.2 Labor demand: Unemployment and remittances — key empirical findings
- Labor demand measured using ILO global unemployment estimates; robustness checked using only real observations.
- Dynamic Okun's-curve specifications augmented with capital and income flows; also present results limiting sample to non-OECD countries.
- Results (summary of Table 1):
  - Unemployment declines significantly and strongly with a rise in the share of remittances, whether contemporaneous or lagged (see specification (2)).
  - Estimated remittances coefficients are smaller when sample limited to non-OECD countries (around 1/3 smaller; see specification (6)).
  - Using only real observations (specification (7)) yields an estimated coefficient almost twice as large as in other specifications, suggesting larger database with imputed data may understate true effect.
  - Remittances have both larger and more consistently significant impact on unemployment than either FDI or ODA when all three included.
- Selected exact coefficient estimates from Table 1 (columns labelled by specification):
  - Unemployment (lagged): 0.503*** (0.137), 0.404*** (0.0397), 0.426*** (0.0474), 0.420*** (0.118), 0.494*** (0.0559), 0.464*** (0.136), 0.306*** (0.0660), 0.242*** (0.0554)
  - GDP growth: -0.0451*** (0.0170), -0.0264*** (0.00915), -0.0263** (0.0125), -0.0238** (0.0116), -0.0295* (0.0171), -0.118*** (0.0201)
  - Remittances share (in % of GDP): -4.114** (1.818), -4.375*** (1.657), -6.695** (2.773), -6.954*** (2.554), -4.234** (1.877), -11.62** (5.278), -6.866** (2.769)
  - Investment share (in % of GDP): -0.133*** (0.0224), -0.171*** (0.0304)
  - Remittances share (lagged, in % of GDP): -5.355** (2.666)
  - FDI share (in % of GDP): -1.233 (0.901), -0.791 (2.396), -0.764 (1.104)
  - Official aid share (lagged, in % of GDP): 0.00645 (0.00806), -0.00976 (0.00960), -0.0136 (0.00949)
  - GDP per capita: -2.23e-05 (6.71e-05), 4.10e-05 (5.22e-05)
  - Constant terms example: 4.632*** (1.252), 8.620*** (0.692), 5.833*** (0.586), 6.174*** (1.410), 5.327*** (0.735), 5.039*** (1.303), 7.183*** (0.633), 10.12*** (0.786)
  - Observations: 2,284; 2,153; 1,826; 1,804; 1,725; 1,929; 1,361; 1,331 (varies by specification)
  - Number of countries: 139; 132; 120; 120; 115; 117; 124; 118
  - Number of instruments: 69; 69; 143; 131; 131; 121; 96; 5 (varies by specification)
- Interpretation:
  - Results can be consistent with either a fall in labor-force participation or an increase in labor demand; further analysis of supply effects is required.

### 3.3 Labor supply: Labor-force participation and remittances — key empirical findings
- Estimation focuses on labor-force participation rate; in the absence of good standard controls, only lagged dependent variable and size of working-age population used in some specifications.
- Remittances enter negatively and robustly across different specifications, including when sample limited to non-OECD countries (see specification (6)).
- Additional controls (time dummies, investment ratios, GDP per capita relative to the United States, trade openness, real wages growth or levels) do not change the negative sign of remittances on labor-force participation.
- In contrast, official aid enters positively, suggesting unconditional transfers depress labor supply; FDI does not seem to affect labor-force participation in most specifications.
- Selected exact coefficient estimates from Table 2:
  - Labor-force participation (lagged): 0.737*** (0.0961), 0.789*** (0.0628), 0.892*** (0.0474), 0.900*** (0.0476), 0.926*** (0.0347), 0.934*** (0.0346), 0.801*** (0.0585), 0.776*** (0.0589), 0.855*** (0.0444)
  - Working-age population (in % of total population): 0.00198* (0.00119), 0.00340* (0.00178), 0.00232 (0.00232), 0.00276* (0.00151), 0.00274* (0.00157), -0.00429*** (0.00160), 0.00171 (0.00215), -0.00165 (0.00164)
  - Remittances (as a share of GDP): -6.220*** (2.171), -4.302*** (1.374), -3.191** (1.429), -2.678** (1.350), -2.358** (1.085), -2.292** (1.164), -6.221*** (1.393), -7.055*** (1.792), -3.817*** (1.232)
  - Official aid (as a share of GDP): 0.0426*** (0.0153), 0.0296** (0.0148), 0.0314** (0.0130), 0.0276** (0.0127)
  - FDI (as a share of GDP): -1.061* (0.608), -0.576 (0.754), -0.902 (0.566), -0.926 (0.606)
  - Liquid liabilities (as a share of GDP): -0.00951** (0.00376)
  - Total investment (as a share of GDP): 0.0106** (0.00415), -0.00377 (0.00459), 0.00976** (0.00446)
  - Real wage growth: 0.0107*** (0.00338), 0.00657** (0.00310)
  - GDP per capita (relative to US): 1.66e-05*** (4.35e-06), 2.40e-05*** (6.01e-06), 1.16e-05*** (3.97e-06)
  - Log of real wage levels: -0.595*** (0.194)
  - Trade openness (in percent of GDP): -0.00503*** (0.00189)
  - Constant examples: 16.71*** (6.099), 12.00*** (3.833), 4.506 (2.967), 5.171* (3.084), 2.884 (1.932), 2.402 (1.921), 15.49*** (4.002), 18.24*** (4.879), 10.47*** (3.287)
  - Observations: 2,284; 2,278; 1,820; 1,450; 1,820; 1,741; 1,484; 1,528; 1,480 (varies by specification)
  - Number of countries: 139; 139; 120; 107; 120; 115; 92; 92; 92
  - Number of instruments: 478; 183; 611; 104; 104; 103; 131; 131; 131
  - Time dummies: included in Equations 5–9; Sargan test statistics reported (e.g., 44.638, 6.288, 6.186, 9.40*, 75.306, 8.177, 8.581, 120.711, 8.6)
- Interpretation:
  - Robust negative association between remittances (share of GDP) and labor-force participation rates across dynamic specifications and samples, including non-OECD countries.
  - Official aid shows positive association with participation, indicating different behavioral responses to unconditional official transfers versus private remittances.
  - FDI generally not significant for labor-force participation.

*Source: wp18102 - Section 2 (IMF working paper, content excerpt).*

### Section 3

### wp18102 - Section 3

### Gender-differentiated effects on labor-force participation (Tables 3 and 4)
- Remittances tend to depress labor-force participation more for women than for men; estimated elasticities of female participation with respect to remittances are larger and more statistically significant than for men (see equations (7) and (8) in Tables 3 and 4).
- Disaggregated coefficients on remittances (as a share of GDP) from Table 3 (Male labor-force participation rate, nine specifications):
  - 0.854*, 0.535, -1.493***, -2.653***, -0.734*, -0.839**, -1.567*, -2.013**, -0.529
- Disaggregated coefficients on remittances (as a share of GDP) from Table 4 (Female labor-force participation rate, nine specifications):
  - -4.527, -7.915***, -5.322**, -3.618**, -1.607, -1.619, -7.807***, -6.251***, -4.552**
- When analyzing the participation gap (percentage-point difference in male and female labor-force participation rates), the corresponding coefficient on remittances was significant only in the simplest specifications (results available from authors upon request).
- Finding consistent with other labor-supply studies: female participation is more elastic to policy and economic variables (e.g., taxes, wages) than male participation.

### Comparative impact of remittances, FDI, and ODA on labor-market outcomes
- Text finding: Remittances tend to have both larger (in absolute value) and more consistently significant impacts on labor-force participation rates than FDI and ODA.
- Implication: Remittances may have larger impacts than FDI or ODA on other labor-market outcomes (e.g., unemployment rate); policymakers may need to adjust the focus and execution of development and labor-market policies to take remittances' effects into account.

### Heterogeneous effects across the distribution of participation (Figure 1 — quantile regressions)
- Quantile regressions show the marginal effects of income and capital flows vary across quantiles of labor-force participation:
  - For remittances and FDI, effects are largest (and significant) only for either very large or very low rates of labor-force participation.
  - For ODA and changes in liquid liabilities, effects are more homogeneous across quantiles.
- Regional patterns at extremes of the participation distribution:
  - Countries with high participation and high remittance inflows: Sub-Saharan Africa and South and South-East Asia (examples: Burkina Faso, Nepal, Cambodia).
  - Countries with low participation and high remittance inflows: Middle East and North Africa (examples: Jordan, Lebanon, Egypt).
- Within these groups:
  - High shares of remittances-to-GDP depress labor supply significantly (Figure 1, Panel A).
  - In the same group of countries, FDI leads to a significant increase in labor supply (Figure 1, Panel B).
  - Official aid appears to further increase labor supply only in countries with already high levels of labor-force participation (Figure 1, Panel C).

### Wages and remittances (Section 3.4.1; Table 5)
- Approach: Estimate a wage Phillips curve augmented with lagged remittance shares; lagged remittances used to mitigate reverse causality.
- Across specifications, the share of remittances to GDP exerts downward pressure on aggregate wage growth.
- Interpreted result: Remittances might lower the pressure from workers to claim a higher share of the joint matching rent; composition effects may also explain negative estimated coefficient.
- Coefficients on lagged remittances (in % of GDP) from Table 5 across specifications:
  - -30.82**, -35.15***, -34.16**, -29.74**, -31.15**
- Other notable coefficients from Table 5 (selected):
  - Real wage growth (lagged): 0.1150, 0.2500, 0.08610, 0.223**, 0.252***, 0.246***, 0.187*, 0.186*
  - Investment share (in % of GDP): 0.171***, 0.199***, 0.220***, 0.222***
  - Change in LFPR (ΔLFPR_t): -0.328**, -0.516**, -0.466**, -0.348
  - Unemployment rate: -0.280**, -0.228*, -0.398***, -0.386**, -0.381**, -0.301**, -0.330*

### Labor-income share and remittances (Section 3.4.2; Table 6)
- Remittances are associated with a positive and significant effect on changes in the labor-income share across all specifications, including the sub-sample of non-OECD countries.
- Interpretation: If productivity growth slows more than wage growth, the labor-income share would increase; public sector responses (e.g., higher public-sector wages when governments tax remittances) may also raise the labor-income share.
- Coefficients on remittances (in % of GDP), Table 6:
  - 6.205**, 7.681**, 7.047**, 7.656**, 7.104**

### Inequality effects (Section 3.4.3; Table 7)
- Using the Standardized World Income Inequality Database (GINI estimates), remittances as a share of GDP enter significantly negatively in all specifications for market inequality.
- The magnitude of negative coefficients on remittances changes moderately across specifications and increases in absolute terms when more controls are added (specification (3)).
- Coefficients on remittances (share of GDP, in %), Table 7:
  - -3.284***, -2.135**, -3.452***, -3.394***, -1.824**, -1.961**, -1.801**, -1.561**
- Other selected coefficients from Table 7:
  - Market inequality (lagged): 0.815***, 0.865***, 0.836***, 0.835***, 0.878***, 0.858***, 0.873***, 0.853***
  - Export share (in %): -6.033***, -6.365***, -3.219***, -3.383***, -3.117*, -2.419
  - GDP growth (p.a., in %): -0.0212***, -0.0240***, -0.0122**, -0.0152**, -0.00537, -0.00601
  - Unemployment rate (%): 0.0290***, 0.0195**, 0.0227***, 0.0221***, 0.0234***, 0.0245***, 0.0165***, 0.0153***

### Informality and remittances (Section 3.4.4; reference to Table 8)
- Hypothesis: A rise in the reservation wage induced by remittances could increase informal employment.
- Using Global Employment Trends data, vulnerable employment is defined as the sum of own-account and contributing family workers and serves as a proxy for informality.
- Results (Table 8 referenced): A positive impact of remittances on the share of vulnerable employment is detected; the effect is robust to a series of control variables except when government debt is included (which reduces the number of observations).
- Text summary: In all other cases the effect is significantly positive.

*Source: wp18102 - Section 3*

### Section 4

### Section 4 — Remittances and Labor Markets (wp18102 - Section 4)

### Regression evidence: Table 8 — Remittances and informal employment
- Estimation method: Dynamic estimates using Arellano-Bond GMM estimator, assuming weak exogeneity for income flows. Year dummies included where indicated. Standard errors in parentheses. ***p<0.01, **p<0.05, *p<0.1.
- Informal employment (lagged) coefficients:
  - (1) Baseline: 0.131 (0.0844)
  - (2) GDP per capita + time dummies: 0.235*** (0.0800)
  - (3) GDP per capita, GDP per capita (in logs), Full controls: 0.412** (0.206)
  - (4) Baseline + time dummies: 0.235*** (0.0805)
  - (5) + investment + government debt + inflation: 0.276*** (0.0794)
  - (6) Baseline + Full controls + time dummies: 0.120 (0.122)
  - (7) + investment: 0.377*** (0.0780)
- GDP per capita (in logs) coefficients:
  - (1) -5.173*** (0.493)
  - (2) -3.673*** (0.451)
  - (3) -1.870*** (0.644)
  - (4) -3.677*** (0.458)
  - (5) -3.391*** (0.456)
  - (6) -5.049*** (0.644)
  - (7) -3.338*** (0.481)
- Investment ratio (in % of GDP) coefficients:
  - -0.117*** (0.0133)
  - -0.0895** (0.0374)
  - -0.116*** (0.0131)
  - -0.145*** (0.0175)
  - -0.119*** (0.0140)
- Remittances (in % of GDP) coefficients:
  - (1) 10.21*** (1.904)
  - (2) 11.73*** (2.183)
  - (3) 31.67** (15.72)
  - (4) 12.01*** (2.208)
  - (5) 10.41*** (2.327)
  - (6) 11.35*** (2.446)
  - (7) 7.352*** (2.258)
- Additional controls (where included):
  - Government debt (in % of GDP): 0.0175*** (0.00521)
  - Inflation rate: 4.25e-4*** (1.35e-4); 3.93e-4*** (1.38e-4); 3.02e-4** (1.42e-4)
  - Import ratio (in % of GDP): 2.24e-2** (8.71e-3); 1.90e-2** (8.69e-3)
- Constant terms:
  - 82.16*** (7.620); 67.40*** (6.949); 38.13*** (10.77); 67.44*** (7.020); 62.88*** (6.896); 81.59*** (10.67); 57.44*** (7.296)
- Sample and diagnostics:
  - Observations: 2,257; 2,176; 968; 2,165; 2,152; 2,257; 2,152
  - Number of countries: 139; 132; 73; 132; 131; 139; 131
  - Number of instruments: 39; 40; 20; 41; 42; 52; 63
  - Sargan test statistics: 37.62; 44.11; 19.04; 43.66; 44.35; 33.02; 46.05
  - Time dummies: No for columns 1–5; Yes for columns 6–7

### Interpretation of micro and aggregate labor-market effects
- Remittances have significant effects on the composition of the labor market in receiving countries.
- Apparent tensions in results:
  - Remittances reduce unemployment but are associated with negative impacts on wage and productivity growth.
  - Reconciliation via segmented labor-market models: remittances can shift employment from higher-productivity/high-wage (tradables, high-skilled) to lower-productivity/low-wage (nontradables, low-skilled) sectors.
  - If employment gains in nontradables exceed losses in tradables, overall unemployment can fall while measured wage and productivity growth decline.
- The estimated increase in informal employment associated with remittance inflows supports this composition-effect conjecture.

### Sectoral shifts (Section 3.5)
- Data and method:
  - ILO Sectoral Employment database derived from the ILO's Trends Econometric Models (ILO, 2010).
  - For each of 14 sectors, an employment-demand equation estimated in levels without a lagged dependent variable; each sector estimated independently.
- Main sectoral findings (Figure 2 summary):
  - Remittance inflows are associated with significant sectoral employment shifts: employment flows out of agriculture and into service-oriented sectors.
  - No discernible push toward manufacturing job creation; instead, low-productive services such as accommodation and construction, and some higher value-added services in transportation and utilities, benefit.
  - The structural transformation induced by remittances calls into question the overall impact on aggregate productivity.
- Quantile-specific results (Figure 3 summary):
  - Manufacturing employment declines especially in economies with a relatively high initial share of manufacturing employment.
  - Construction employment increases in countries with intermediate or low starting shares of construction employment.
  - Financial services show a similar but smaller pattern.
  - Transportation services benefit broadly across quantiles, suggesting remittances finance small-scale transport services as alternative incomes.
- Behavioral links:
  - Consistent with literature finding remittances reduce farmers' effort—agricultural employment declines.
  - Negative impact on manufacturing employment may reflect Dutch-disease effects, depending on tradability of local manufacturing outputs.
  - Sectors with increased employment (real estate, construction, utilities, education, health) correspond to typical household spending patterns from remittances (home purchases, home improvements, education, healthcare).

### Distributional and compositional implications
- Remittances exhibit strong compositional effects:
  - Consistently reduce measured income inequality.
  - Associated with better economic performance across income levels, with relatively greater increases for lower-income earners than for higher-income earners.
  - Simultaneously, some higher-income earners fare worse, helping explain why remittances reduce poverty and inequality but have not been found to enhance economic growth.

### Regional variation (Section 3.6)
- Approach:
  - Two approaches: (i) remove each region one at a time to estimate difference from full sample; (ii) interact regional dummies with remittances/GDP. Figure 4 uses the first approach.
- Labor force participation:
  - Significant cross-regional divergence in the (absolute) impact of remittances on labor force participation; in all regions the effective impact remains negative, but absolute size differs.
- Informal employment (Figure 5):
  - Remittances increase informality more strongly in MENA, Sub-Saharan Africa, and Central and Western Asia.
  - Weaker impact on informality in Eastern Europe, Southern Asia, Latin America and the Caribbean.
  - Note: MENA countries have on average a lower informality rate than Latin America and the Caribbean: 16 percent of total employment vs. 35 percent.
- Wage growth (Figure 6):
  - Interaction approach shows significant regional variation in remittances' impact on wage growth.
  - Positive effects visible among (poorer) countries in South Asia and Sub-Saharan Africa and in relatively more affluent countries in Eastern Europe.
  - Negative effects dominate in Central and Western Asia and South-East Asia and the Pacific.
  - Sample sizes limit further regional inference.

### Fragility and remittance effects (Section 3.7)
- Fragility measure:
  - Country fragility classified using the Fund for Peace Fragile States Index; countries in the highest quintile of the Index between 2006 and 2017 labeled fragile. An average score used for 2006–2017. Note: "The score changes very little over the indicated time period and exhibits an auto-correlation of 85%."
- Labor force participation by fragility quintile (Figure 7):
  - Adverse effects of remittances on labor force participation are particularly strong for intermediate levels of fragility (non-linear relationship).
  - For very high fragility the impact declines but remains statistically significant; for the most stable countries the effect disappears.
  - Number of countries by fragility quintile: 4, 25, 30, 29, 28 (from lowest to highest quintiles).
- Wage growth and inequality by fragility (Figure 8):
  - Interaction of remittances/GDP with a fragile-country dummy — coefficients measured in natural logarithms.
  - Impact of remittances on inequality does not seem to depend on degree of country fragility; differences in remittances coefficients for fragile vs. stable countries are not statistically significant.
  - Real wage growth reacts differently: remittances raise real wages in fragile countries but not in stable ones.
  - Possible explanation: fragile states have lower trade openness and export shares (Figure 9), implying smaller tradables sectors and weaker Dutch-disease channels; remittances therefore can lift real wages rather than merely shifting resources to nontradables.
- Openness and fragility (Figure 9):
  - As fragility increases, export share as a percent of GDP declines rapidly.
  - Trade openness measured as sum of exports and imports over GDP; export share measures nominal export value over nominal GDP.

### Conclusion (Section 4)
- Approach and contribution:
  - Cross-country aggregated data used to estimate remittances' labor-market effects, yielding stylized facts complementary to household-survey studies.
- Stylized facts from the study:
  - Remittances have strong impacts on both labor supply and labor demand in recipient countries.
    - Supply-side: remittances reduce labor force participation and increase informality of the labor market.
    - Demand-side: remittances reduce overall unemployment.
  - The size and significance of these remittance effects on labor-market outcomes is greater than those of FDI and ODA on the same outcomes.
  - Narrative complexity:
    - Although remittances reduce unemployment rates, they are also associated with lower wage growth and higher labor shares of income, implying lower productivity growth.

### Section 5

### Section 5

### Remittances and labor-market effects — key findings
- Remittances also lower measured income inequality.
- A consistent story explaining these findings can be framed by the Dutch disease narrative, in which remittances benefit lower-wage, lower-productivity nontradables industries at the expense of the high-productivity, high-wage tradables industries.
- Sectoral estimation of employment tends to support this interpretation:
  - Manufacturing employment is negatively related to remittances.
  - Construction, real estate, and transportation are positively related to remittances.
- Cross-country variation highlights complexity in the pathways through which remittances affect labor markets:
  - Substantial regional variation in remittances' effects on labor force participation, informality, and wage growth is documented.
  - Distinctive differences exist in the effect of remittances on fragile states relative to more stable states.
  - In fragile states, remittances have a positive impact on wage growth, but in the less-productive sector.
- The findings raise the possibility that remittances help stabilize incomes in fragile states at the expense of lower economic growth — a proposition that merits further research.

### Interpretation challenges and implied mechanisms
- The interpretation of aggregated data implies that remittances lead to disproportionately large increases in employment in the nontradables sector, or in the less-productive or lower-skill segment of the labor market.
- There is mounting evidence that remittances are suspected of reducing the quality of institutions in recipient countries; given the positive relation between governance and growth, this interaction requires further investigation.

### Research needs and future tasks
- Further research is needed to:
  - Verify the stylized facts presented in the paper.
  - Discover other stylized facts that can shed light on the labor-market effects of remittances.
  - Test the implications of the interpretation of aggregated data.
  - Quantify the price distortions and negative externalities that remittances impose on formal labor markets and the tradables sector.
  - Determine which policy approaches are likely to be most effective at mitigating the negative impacts of remittances.
- Specific policy-relevant areas suggested for investigation include:
  - Investments in infrastructure.
  - Reforms to improve the ease of doing business.
  - Changes in tax policy.
- Because remittances may impair the tradables sector, researchers and policymakers need to push harder to find development policies that improve competition and reduce informality.

### Appendix — Summary statistics (selected key figures)
- Working-age population (in%): #Observations 3,799; Mean 69.11; Std. Dev. 11.64; Min 3.0; Max 100.0
- Labor force participation rate (in%): #Observations 4,080; Mean 63.3; Std. Dev. 10.4; Min 9.9; Max 90.2
- Unemployment rate (in%): #Observations 4,080; Mean 9.3; Std. Dev. 6.8; Min 0.14; Max 40.2
- Real wage growth (p.a., in%): #Observations 2,315; Mean 2.3; Std. Dev. 6.1; Min -30.6; Max 47.0
- Share of informal employment (in%): #Observations 4,080; Mean 40.8; Std. Dev. 28.1; Min 0.3; Max 96.0
- Employment shares (in% of total employment):
  - Agriculture: #Observations 4,080; Mean 30.2; Std. Dev. 24.9; Min 0.2; Max 90.9
  - Manufacturing: #Observations 4,080; Mean 13.1; Std. Dev. 7.5; Min 0.2; Max 43.0
  - Construction: #Observations 4,080; Mean 6.3; Std. Dev. 4.2; Min 0.1; Max 46.7
  - Accommodation and restaurants: #Observations 4,080; Mean 3.2; Std. Dev. 2.6; Min 0.1; Max 20.3
  - Health and social work: #Observations 4,080; Mean 4.6; Std. Dev. 3.9; Min 0.1; Max 21.6
- Market GINI: #Observations 2,825; Mean 45.0; Std. Dev. 7.6; Min 18.0; Max 72.8
- Remittances (in% of GDP): #Observations 2,343; Mean 3.3; Std. Dev. 5.4; Min 0.04; Max 49.4
- FDI (in% of GDP): #Observations 3,227; Mean 5.6; Std. Dev. 30.7; Min -46.2; Max 726.8
- Official development aid (in% of GDP): #Observations 2,976; Mean 6.0; Std. Dev. 8.7; Min -0.7; Max 115.4
- Real GDP growth (p.a., in%): #Observations 4,080; Mean 3.8; Std. Dev. 6.4; Min -52.6; Max 147.7
- Investment share (in% of GDP): #Observations 3,602; Mean 23.0; Std. Dev. 11.0; Min 0.1; Max 227.5
- Trade openness (in% of GDP): #Observations 3,725; Mean 84.2; Std. Dev. 52.9; Min 4.9; Max 809.2
- real GDP per capita (in Int$ PPP): #Observations 3,736; Mean 9,158.8; Std. Dev. 14,457.9; Min 30.9; Max 112,429.4

### Appendix — Regional country coverage (regions listed)
- Central and Western Asia: Armenia, Azerbaijan, Cyprus, Georgia, Israel, Kazakhstan, Kyrgyzstan, Tajikistan, Turkey, Turkmenistan, Uzbekistan
- Eastern Asia: China, Hong Kong SAR, Japan, South Korea, Mongolia, Taiwan POC
- Eastern Europe: Belarus, Bulgaria, Czech Republic, Hungary, Moldova, Poland, Romania, Russian Federation, Slovakia, Ukraine
- Latin America and Caribbean: Argentina, Bahamas, Barbados, Belize, Bolivia, Brazil, Chile, Colombia, Costa Rica, Dominican Republic, Ecuador, El Salvador, Guatemala, Guyana, Haiti, Honduras, Jamaica, Mexico, Nicaragua, Panama, Paraguay, Peru, Suriname, Trinidad and Tobago, Uruguay, Venezuela
- Middle East: Bahrain, Iraq, Jordan, Kuwait, Lebanon, Oman, Qatar, Saudi Arabia, Syrian Arab Republic, United Arab Emirates, Yemen
- Northern Africa: Algeria, Egypt, Libya, Morocco, Sudan, Tunisia
- Northern America: Canada, United States
- Northern, Southern and Western Europe: Albania, Austria, Belgium, Bosnia and Herzegovina, Croatia, Denmark, Estonia, Finland, France, Germany, Greece, Iceland, Ireland, Italy, Latvia, Lithuania, Luxembourg, Macedonia, Malta, Montenegro, Netherlands, Norway, Portugal, Serbia, Slovenia, Spain, Sweden, Switzerland, United Kingdom
- South-Eastern Asia and the Pacific: Australia, Brunei Darussalam, Cambodia, Fiji, Indonesia, Lao PDR, Malaysia, Myanmar, New Zealand, Papua New Guinea, Philippines, Singapore, Solomon Islands, Thailand, Timor-Leste, Vietnam
- Southern Asia: Afghanistan, Bangladesh, Bhutan, India, Iran, Maldives, Nepal, Pakistan, Sri Lanka
- Sub-Saharan Africa: Angola, Benin, Botswana, Burkina Faso, Burundi, Cameroon, Cape Verde, Central African Republic, Chad, Comoros, Congo, Democratic Republic of the Congo, Côte d'Ivoire, Equatorial Guinea, Eritrea, Ethiopia, Gabon, Gambia, Ghana, Guinea, Guinea-Bissau, Kenya, Lesotho, Liberia, Madagascar, Malawi, Mali, Mauritania, Mauritius, Mozambique, Namibia, Niger, Nigeria, Rwanda, Senegal, Sierra Leone, Somalia, South Africa, Swaziland, Tanzania, Togo, Uganda, Zambia, Zimbabwe

*Source: wp18102 - Section 5*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2018/wp18102.pdf_