## Introduction — The Propensity to Remit: Macro and Micro Factors Driving Remittances to Central America and the Caribbean

## Source details

**Canonical URL:** [Introduction — The Propensity to Remit: Macro and Micro Factors Driving Remittances to Central America and the Caribbean](https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022203-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2022/english/wpiea2022203-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2022/english/wpiea2022203-print-pdf.pdf.json)

---

### Overview and motivation
- Remittance inflows to Central America and the Caribbean (CAC) "surprised on the upside during 2020 and 2021."
- Forecasts based on sending-country GDP-growth elasticities had projected a sharp drop in remittances in 2020 of "about 20-40 percent."
- After an initial dip, worker remittances "recovered mid-year and surpassed 2019 levels in many countries in the region."
- Remittances to low and middle-income countries amounted to "$550 and $540 billion" globally in 2019 and 2020 respectively (World Bank).
- Remittances exceeded total inflows from foreign direct investment and other external current account inflows and are large as a share of GDP, "exceeding 5 (10) percent of GDP in 59 (34) countries in 2020" (World Bank).
- The number of migrants from CAC represent "at least 10 percent of the population of the region and are as high as 20 percent for the Caribbean" (Beaton et al (2017), UNDP).

### Major macro findings and elasticities
- Remittance responsiveness to sending- and receiving-country conditions:
  - "Remittance flows are responsive to economic conditions in the sending as well as receiving countries."
  - Empirical regression results (annual data 2001-2019):
    - A 1.0 percentage point increase in US GDP growth translates into a 16 percentage points increase in remittances as a percentage of GDP (US GDP growth coefficient reported as 16.68***, 16.81***, 17.31***, 18.12*** across specifications).
    - A 1.0 percentage point increase in GDP growth in the receiving country translates into a 4 percentage points decrease in remittances as a percentage of GDP (GDP growth in receiving country coefficients: -3.636***, -3.689***, -3.661***, -3.228**).
    - The 1-year lag Remitt %GDP coefficient is 0.831***, 0.835***, 0.832***, 0.823*** across specifications.
  - Standard-deviation interpretation:
    - One standard deviation increase in GDP growth in the U.S. corresponds to a 16 percentage points increase in remittances as a percentage of GDP.
    - One standard deviation increase in GDP growth in the receiving country corresponds to a 28 percentage points decrease in remittances as a percentage of GDP.
  - Bilateral exchange rate and inflation in both countries do not show robust effects in these specifications.

### Role of labor-market indicators and income
- Labor-market indicators add predictive power beyond GDP elasticities:
  - "Labor market conditions are particularly important determinants of remittance flows."
  - Unemployment differentials matter:
    - Diff in unemployment rate changes between US foreign-born and receiving country coefficients: -0.0783**, -0.0775*, -0.0557* across specifications.
    - Diff in GDP growth between US and receiving country coefficients: 5.177*, 4.132* in some specifications.
    - 1-year lag remittances %GDP coefficients range 0.782*** to 0.825*** across Table 3 specifications.
  - Family income and US GDP growth jointly matter:
    - US GDP growth coefficient reported as 18.12*** in a specification including family income growth.
    - Family income (first- and second-generation growth) coefficient reported as 5.812** in one specification and 2.992 (not significant) in another.
- Interpretation: unemployment-rate differentials serve as a proxy for cross-border resource sharing; as receiving-country labor markets deteriorate relative to the U.S., remitters have greater propensity to send funds.

### COVID-19 versus the GFC — observed behavior and interpretation
- Puzzle: During COVID-19 (2020) remittances increased more than predicted by GDP-based elasticities.
- Key U.S. labor-market and policy context in 2020:
  - U.S. average household income increased by 7 percent in real terms around July 2020 for both US-born and immigrants.
  - Immigrants were less likely than US-born persons to work from home, more likely to be unable to work due to COVID-19, less likely to receive pay for hours not worked, and more likely to be prevented from looking for work.
  - The increase in household income is partly attributed to large and quasi-universal income support measures implemented in the U.S., and possibly to shifts from informal to formal employment.
- Comparative dynamics:
  - GDP differentials alone do not explain the unexpected jump in remittance flows in 2020-2021.
  - The unemployment differential with respect to the U.S. better explains remittance patterns during the pandemic and captures dynamics in both the GFC (2008-09) and COVID (2020) more closely than GDP differentials alone.

### Micro (household) determinants — CPS evidence (2000–2021; remittances module 2008)
- Data features:
  - CPS sample size is "about 60,000 occupied households in each month."
  - CPS covers civilian noninstitutional population "16 years of age and older."
  - Rotation design: interviewed for 4 successive months, 8-month pause, then another 4 months.
  - CPS collection changes during COVID: March 2020 to June 2020 interviews were "exclusively by telephone;" response rates declined "to 65 percent in June compared to 82 percent in February;" by October 2020 total response rate had recovered to 80.3 percent.
- Likelihood to remit (logit) — full immigrant sample (2008 remittances module):
  - Education of head of household is inversely related to likelihood to remit: education d1–d4 coefficients negative and statistically significant (e.g., Education level d4: -0.815*** in one specification).
  - Household size increases likelihood to remit (household size coefficient 0.0865**).
  - Men are more likely to remit than women (Sex coefficient -0.213** where female = 1).
  - Household income is not statistically significant in the full-sample logit once education is included.
- Amount remitted (conditional on remitting) — full immigrant sample:
  - Family income elasticity of remittances estimated around 0.25-0.30 (Family income coefficients 0.326***, 0.264***).
  - Higher education (tertiary and postgraduate) associated with larger remitted amounts conditional on remitting (e.g., Education level d3: 0.369***; d4: 0.723***).
  - Men remit higher amounts than women (Sex coefficients -0.214*, -0.257* where female = 1).
  - Household size has positive effect on amount remitted (Household size 0.0613* in a weighted specification).

### Findings for CAC migrants (Haiti, Caribbean, Central America, including Mexico)
- Likelihood to remit (CAC subsample):
  - Household income has a statistically significant positive impact on likelihood to remit (Family income coefficients 0.234**, 0.231*).
  - Education plays a lesser and less consistent role for CAC migrants.
  - Men are more likely to remit than women (Sex coefficients -0.294*, -0.315*).
  - Household size and metropolitan area indicators show positive effects in some specifications.
- Amount remitted (conditional on remitting) — CAC subsample:
  - Family income elasticity larger than full sample: coefficients 0.401*** and 0.367*** indicating an estimated elasticity around 0.35-0.40.
  - Younger immigrants remit relatively more (Age coefficients -0.0813**, -0.0831**; Age squared positive and significant).
  - Some evidence of lower remitted amounts for highest education groups in certain specifications (Education level d4 coefficients negative and significant in some specs: -0.972*,-0.959*).

### Mechanisms and modeling implications
- Remittances reflect:
  - Senders’ ability to remit (sending-country economic and labor-market conditions, household income).
  - Recipients’ needs (receiving-country economic conditions).
  - Cost and methods of transfer (transactions costs; switching from informal/cash to formal/digital channels could affect recorded remittances).
- Tests for switching from informal to formal channels:
  - Using island status as a proxy for likelihood of switching from informal to formal methods showed no systematic relationship with model residuals, implying the employment-differential channel is robust to this potential switching effect.
- Policy-relevant implication: labor-market indicators, particularly unemployment-rate differentials for immigrants in the U.S., provide important explanatory power beyond GDP elasticities for forecasting remittance flows.

### Conclusions and policy considerations
- New estimates of elasticities to growth and employment rates in both sending and receiving countries strengthen forecasting toolkits for remittances.
- Remittances rise more when U.S. economic growth is stronger and when receiving-country growth is weaker; remittance flows are more responsive to sending-country economic conditions than to those in the receiving country.
- Labor market conditions, specifically unemployment differentials, are critical and support a resource-sharing model across borders.
- Micro-data suggest:
  - Conditional on sending, the higher the household income of immigrants, the larger the amount sent.
  - Estimated elasticity of remitted amount with respect to household income is about 0.35-0.40 for immigrants from CAC.
  - For the full immigrant sample, higher education of head of household is associated with lower likelihood to remit; household income not statistically significant in the logit model—possible interpretation: more-educated immigrant households may have more family present in the host country or recipients with higher status, reducing reliance on transfers.
- Policy implications:
  - Estimated elasticities and labor-market indicators can inform forecasters and policymakers about remittance flows.
  - Micro-data results could inform U.S. domestic policies on education, labor market, and reducing inequality and gender gaps, with knock-on effects on immigrant income and remittances.
  - Domestic policy responses to shocks (e.g., COVID income supports) influenced immigrant income and employment and therefore remittance flows.

### Key micro-statistics (Appendix Table 1, 2015-2019)
- USA: 52% female; Average age 47; 30% college degree; Employment rate 59%; Unemployment rate 4%; Participation rate 62%.
- ATG: 61% female; Average age 43; 31% college degree; Employment rate 57%; Unemployment rate 7%; Participation rate 61%; US state where majority lives: New York.
- BHS: 51% female; Average age 42; 33% college degree; Employment rate 62%; Unemployment rate 7%; Participation rate 67%; US state where majority lives: Florida.
- GTM: 43% female; Average age 39; 10% college degree; Employment rate 70%; Unemployment rate 4%; Participation rate 73%; US state where majority lives: California.
- HND: 53% female; Average age 38; 10% college degree; Employment rate 66%; Unemployment rate 5%; Participation rate 70%; US state where majority lives: Texas.
- MEX: 48% female; Average age 43; 9% college degree; Employment rate 64%; Unemployment rate 5%; Participation rate 67%; US state where majority lives: California.
- SLV: 49% female; Average age 42; 8% college degree; Employment rate 70%; Unemployment rate 4%; Participation rate 73%; US state where majority lives: California.

### Selected 2019 vs 2020 indicators (Appendix Table 2)
- USA: Employment Rate 2019 60% / 2020 56%; Unemployment Rate 2019 4% / 2020 8%; Participation Rate 2019 62% / 2020 60%.
- BHS: Employment Rate 2019 69% / 2020 56%; Unemployment Rate 2019 5% / 2020 27%; Participation Rate 2019 72% / 2020 77%.
- CRI: Employment Rate 2019 62% / 2020 73%; Unemployment Rate 2019 3% / 2020 8%; Participation Rate 2019 64% / 2020 79%.
- GTM: Employment Rate 2019 69% / 2020 63%; Unemployment Rate 2019 3% / 2020 10%; Participation Rate 2019 71% / 2020 70%.
- NIC: Employment Rate 2019 64% / 2020 55%; Unemployment Rate 2019 5% / 2020 7%; Participation Rate 2019 67% / 2020 59%.

*Source: IMF Working Paper — The Propensity to Remit: Macro and Micro Factors Driving Remittances to Central America and the Caribbean (WP/2022/203).*

### Introduction ...........................................................................................................

### Introduction

### Overview and motivation
- Remittance inflows to Central America and the Caribbean (CAC) "surprised on the upside during 2020 and 2021."
- Forecasts based on sending-country GDP-growth elasticities had projected a sharp drop in remittances in 2020 of "about 20-40 percent."
- After an initial dip, worker remittances "recovered mid-year and surpassed 2019 levels in many countries in the region."
- Remittances to low and middle-income countries amounted to "$550 and $540 billion" globally in 2019 and 2020 respectively (World Bank).
- Remittances exceeded total inflows from foreign direct investment and other external current account inflows and are large as a share of GDP, "exceeding 5 (10) percent of GDP in 59 (34) countries in 2020" (World Bank).

### Importance of remittances for CAC
- The number of migrants from CAC represent "at least 10 percent of the population of the region and are as high as 20 percent for the Caribbean" (Beaton et al (2017), UNDP).
- Private transfers are "equivalent to about 8 percent of GDP on average," while in some sample countries, "like El Salvador and Honduras, remittances account for as much as 24 percent of GDP" (World Bank, 2020).
- Large remittance flows affect the balance of payments, domestic liquidity, the labor market, and the financial sector, underscoring the value of better prediction for policymakers (financing needs, liquidity monitoring, product markets, financial sector stability).

### Key research questions and findings
- Aim: Re-evaluate macroeconomic determinants of remittances in sending and receiving countries and investigate characteristics of immigrant households in the U.S. who remit, using micro data from the U.S. Current Population Survey (CPS).
- Main findings summarized in the Introduction:
  - "Remittance flows are responsive to economic conditions in the sending as well as receiving countries."
  - "Labor market conditions are particularly important determinants of remittance flows, explaining the unexpected jump in remittance flows in 2020-2021 and providing stronger predictive power when combined with income variables."
  - Microdata analysis finds "higher levels of income and education are associated with stronger remittance flows." This reflects a "resource sharing model" though these relationships "did not hold for the sample of all immigrants in the U.S."
- The stronger-than-expected remittance volumes during COVID "might be explained by unusual factors, including the income support provided in the U.S. during the COVID pandemic."
- The paper finds that a shift from informal methods of sending remittances (cash) to formal methods (digital and recorded in the balance of payments) "is not found to play a role here."

### Literature surveyed (high-level)
- Three broad strands: developmental impact, cyclical implications, macro determinants.
- Developmental impact: remittances associated with higher personal consumption and lower poverty, potential Dutch disease risks, impacts on labor supply and possible "remittance trap" (Chami et al, 2018).
- Cyclical role: mixed evidence — remittances can be countercyclical and stabilizing (Bettin et al, 2017; Frankel, 2011), but dependence can increase vulnerability to source-country shocks (Barajas et al, 2012; IMF, 2012).
- Determinants: migration stocks, bilateral gravity-type factors (source and destination size, distance), institutional and geographic features (common language, colonial ties, borders, island status).
- Reported elasticity estimates of remittances to sending-country GDP "varying between 0.25 – 3.9, and reaching the 5 – 10 range" (Poghosyan, 2020).
- Prior elasticity-based forecasts "proved less reliable when forecasting remittances in 2020 and 2021."

### Data and empirical approach (summary of data sources and methods)
- Focus: first- and second-generation immigrants in the U.S. from CAC.
- Microdata: Current Population Survey (CPS) from years "2000 to 2021"; Section VII uses the remittances supplement for the year "2008."
- Labor market statistics: combined with U.S. labor force survey data over "2000 to 2020" for unemployment rates and participation rates by demographic group.
- CPS features:
  - Sample size is "about 60,000 occupied households in each month."
  - Covers civilian noninstitutional population "16 years of age and older."
  - Rotation design: interviewed for 4 successive months, 8-month pause, then another 4 months before exiting the sample.
- CPS collection changes during COVID:
  - From March 2020 to June 2020 interviews were "exclusively by telephone."
  - In-person data collection restarted in some localities in July 2020 and "personal visits resumed across the country" in September 2020.
  - Response rates declined "to 65 percent in June compared to 82 percent in February."

### Paper structure (what follows in the study)
- Section II: survey of related literature.
- Sections III and IV: present data and stylized facts about remittance flows to CAC and characteristics of migrants from CAC in the U.S.
- Section V: source-country factors affecting remittance flows; variance decomposition analysis to estimate share of remittance variation explained by U.S. economic conditions, notably during COVID.
- Section VI: estimates the impact of COVID on employment, income, and likelihood to remit for migrants from each country in the sample; uses microsimulations; compares COVID shock with the Global Financial Crisis (GFC) and assesses the role of the U.S. policy response.
- Section VII: applies a discrete choice model using individual-level features to estimate likelihood to send cash (remit).
- Section VIII: summarizes findings and highlights research and policy considerations and further research areas.

*Source: IMF Working Paper — Introduction (wpiea2022203-print-pdf).*

### 2020.  By October 2020, the total response rate had recovered to 80.3 percent, close to the previous annual

### The Propensity to Remit: Macro and Micro Factors Driving Remittances to Central America and the Caribbean

### Remittance flows and macro trends
- Remittance flows have increased steadily over the last 20 years and constitute an important share of GDP in CAC.
- In 2019, remittances exceeded 20 percent of GDP in Haiti and Honduras; in Panama, Costa Rica, and Anguilla remittances account for 1-2 percent of GDP annually.
- By October 2020, the total response rate had recovered to 80.3 percent, close to the previous annual average.
- Empirical regression results (annual data 2001-2019) indicate:
  - A 1.0 percentage point increase in US GDP growth translates into a 16 percentage points increase in remittances as a percentage of GDP (US GDP growth coefficient reported as 16.68***, 16.81***, 17.31***, 18.12*** across specifications).
  - A 1.0 percentage point increase in GDP growth in the receiving country translates into a 4 percentage points decrease in remittances as a percentage of GDP (GDP growth in receiving country coefficients: -3.636***, -3.689***, -3.661***, -3.228**).
  - The 1-year lag Remitt %GDP coefficient is 0.831***, 0.835***, 0.832***, 0.823*** across specifications.
  - Bilateral exchange rate and inflation in both countries do not show robust effects on the decision to send funds in these specifications.
- Standard-deviation interpretation reported:
  - One standard deviation increase in GDP growth in the U.S. corresponds to a 16 percentage points increase in remittances as a percentage of GDP.
  - One standard deviation increase in GDP growth in the receiving country corresponds to a 28 percentage points decrease in remittances as a percentage of GDP.

### Role of sending-country (U.S.) labor market and income
- Remittances are positively correlated with household income among immigrants; household income in turn correlates with U.S. economic conditions.
- Table results show family income and US GDP growth jointly matter:
  - US GDP growth coefficient reported as 18.12*** in a specification including family income growth.
  - Family income (first- and second-generation growth) coefficient reported as 5.812** in one specification and 2.992 (not significant) in another.
- Unemployment differentials matter:
  - Diff in unemployment rate changes between US foreign-born and receiving country coefficients: -0.0783**, -0.0775*, -0.0557* across specifications.
  - Diff in GDP growth between US and receiving country coefficients: 5.177*, 4.132* in some specifications.
  - 1-year lag remittances %GDP coefficients range 0.782*** to 0.825*** across Table 3 specifications.
  - Results highlight the importance of the unemployment differential between immigrants in the U.S. and the receiving country as a proxy for resource sharing across borders.

### COVID-19 versus the GFC: observed remittance behavior
- During COVID-19 (2020) remittances increased more than predicted by GDP-based elasticities; this created a “puzzle” relative to expectations from GDP regressions.
- Key labor-market and policy context in the U.S. during COVID-19:
  - U.S. average household income increased by 7 percent in real terms around July 2020 for both US-born and immigrants.
  - Immigrants were less likely than US-born persons to work from home, more likely to be unable to work due to COVID-19, less likely to receive pay for hours not worked, and more likely to be prevented from looking for work.
  - The increase in household income is partly attributed to large and quasi-universal income support measures implemented in the U.S., and possibly to shifts from informal to formal employment.
- Comparisons with the GFC:
  - Relationship between remittance changes and GDP growth differentials showed that remittances rose more during COVID-19 than predicted by GDP differentials alone.
  - Relationship between remittance changes and unemployment-rate differentials captured dynamics in both the GFC (2008-09) and COVID (2020) more closely than GDP differentials alone.

### Individual and household determinants of propensity to remit
- Data source: 2008 CPS remittances module (sample exclusions noted in text).
- Likelihood to remit (logit) — full immigrant sample:
  - Education of head of household is inversely related to the likelihood to remit: higher education levels correspond to lower probability of sending remittances (education d1–d4 coefficients negative and statistically significant; e.g., Education level d4: -0.815*** in one specification).
  - Household size increases likelihood to remit (household size coefficient 0.0865** in one specification).
  - Men are more likely to remit than women (Sex coefficient -0.213** where female = 1).
  - Household income is not statistically significant in the full-sample logit once education is included.
- Amount remitted (conditional on remitting) — full immigrant sample:
  - Family income elasticity of remittances estimated around 0.25-0.30 (Family income coefficients 0.326***, 0.264*** in Table 4b).
  - Higher education (tertiary and postgraduate) is associated with larger amounts remitted conditional on remitting (e.g., Education level d3: 0.369***; d4: 0.723*** in some specs).
  - Men remit higher amounts than women (Sex coefficients -0.214*, -0.257* where female = 1).
  - Household size has some positive effect on amount remitted (Household size 0.0613* in a weighted specification).

### Findings specific to migrants from Haiti, the Caribbean, and Central America (including Mexico)
- Likelihood to remit (CAC subsample):
  - Unlike the full sample, household income has a statistically significant positive impact on the likelihood to remit (Family income coefficients 0.234**, 0.231*).
  - Education plays a lesser role and is not consistently statistically significant in the CAC sub-sample.
  - Men are more likely to remit than women (Sex coefficients -0.294*, -0.315*).
  - Household size and metropolitan area indicators show positive effects in some specifications.
- Amount remitted (conditional on remitting) — CAC subsample:
  - Family income elasticity larger than full sample: coefficients 0.401*** and 0.367*** indicating an estimated elasticity around 0.35-0.40.
  - Younger immigrants remit relatively more (Age coefficients -0.0813**, -0.0831**; Age squared positive and significant).
  - Some evidence of lower remitted amounts for highest education groups in certain specifications (Education level d4 coefficients negative and significant in some specs: -0.972*,-0.959*).

### Mechanisms, interpretation, and modeling implications
- Remittances depend on:
  - Senders’ ability to remit (sending-country economic and labor-market conditions, household income).
  - Recipients’ needs (receiving-country economic conditions).
  - Cost and methods of transfer (transactions costs, switching from informal/cash to formal/digital channels potentially affecting recorded remittances).
- Labor-market indicators, particularly unemployment-rate differentials for immigrants in the U.S., provide important explanatory power beyond GDP elasticities for forecasting remittance flows.
- The data and regressions suggest a resource-sharing model: worsening receiving-country labor-market conditions relative to the U.S. increase remittance flows.
- Using island status as a proxy for likelihood of switching from informal to formal transfer methods was tested; no systematic relationship between island status and model residuals was found, implying the employment-differential channel is robust to this potential switching effect.

*Source: IMF Working Paper — The Propensity to Remit: Macro and Micro Factors Driving Remittances to Central America and the Caribbean*

### Conclusions and Policy Considerations

### Conclusions and Policy Considerations

### Macro determinants of remittances
- New estimates of elasticities of remittances with respect to growth and employment rates in both sending and receiving countries are presented.
- Remittance flows are responsive to economic conditions in the sending country as well as receiving countries.
- Remittances rise more when U.S. economic growth is stronger and when the receiving country growth is weaker.
- The difference in the coefficients of the growth rates in the sending and receiving countries suggests that remittance flows are more responsive to sending country economic conditions than to those in the receiving country.
- Labor market conditions are particularly important determinants of remittance flows.
  - The unemployment rate in the sending country relative to the rate in the receiving country is the more important factor.
  - The unemployment differential is a rough proxy for the difference in income between family members split across countries and confirms a resource sharing model in which family members living in two different countries share resources.
  - As labor market conditions in the receiving country deteriorate relative to the U.S., remitters have a greater propensity to send funds.
  - When combined with growth variables, labor market conditions provide stronger predictive power than when relying exclusively on income variables.

### Global shocks and remittance patterns
- Section VI compares the impact of two global shocks on remittance patterns and finds:
  - GDP differentials alone do not explain well the unexpected jump in remittance flows in 2020-2021.
  - The unemployment differential with respect to the U.S. better explains remittance patterns during the pandemic.
  - These results support findings on the predictive power of labor market conditions and the role of the family resource sharing model.

### Micro (household) evidence from U.S. immigrant data
- Conditional on sending remittances:
  - The higher the income of the household of all immigrants, the larger the amount sent.
  - The estimated elasticity of remitted amount with respect to household income is larger for immigrants from CAC at about 0.35-0.40.
  - Education and income are positively correlated for all immigrants; more tertiary education has a positive effect on the amount remitted for the full sample―conditional on sending remittances―reinforcing the income channel.
  - Education plays a lesser role in determining remittance flows for immigrants from the CAC, possibly because immigrants from the CAC region have lower levels of education than the average of all immigrants.
- For the full sample of all immigrants in the U.S., not conditional on sending remittances:
  - Higher education of the “head of household” was associated with a lower likelihood that the household remits.
  - The household income variable was not statistically significant in the logit model.
  - Possible interpretation: immigrant households with higher education in the U.S. may already have more or most family members present in the host country, and/or family members still in the receiving country may have relatively higher education and economic status, reducing reliance on diaspora transfers for income support or insurance against income shocks.

### Implications for forecasters and policymakers
- Estimated elasticities to key indicators in both sending and receiving countries, including labor market conditions, can strengthen the toolkit for forecasters and policymakers to anticipate remittance flows.
- Micro-data results could inform U.S. domestic policies directed at education, the labor market, or at reducing inequality and gender gaps, which may affect immigrant income and employment.
- The analysis comparing remittances during the COVID pandemic and GFC shocks suggests that domestic policies influenced immigrant income and employment and therefore remittance flows.

### Key micro-statistics (summary from Appendix Table 1, 2015-2019)
- USA: 52% female; Average age 47; 30% college degree; Employment rate 59%; Unemployment rate 4%; Participation rate 62%.
- ATG: 61% female; Average age 43; 31% college degree; Employment rate 57%; Unemployment rate 7%; Participation rate 61%; US state where majority lives: New York.
- BHS: 51% female; Average age 42; 33% college degree; Employment rate 62%; Unemployment rate 7%; Participation rate 67%; US state where majority lives: Florida.
- GTM: 43% female; Average age 39; 10% college degree; Employment rate 70%; Unemployment rate 4%; Participation rate 73%; US state where majority lives: California.
- HND: 53% female; Average age 38; 10% college degree; Employment rate 66%; Unemployment rate 5%; Participation rate 70%; US state where majority lives: Texas.
- MEX: 48% female; Average age 43; 9% college degree; Employment rate 64%; Unemployment rate 5%; Participation rate 67%; US state where majority lives: California.
- SLV: 49% female; Average age 42; 8% college degree; Employment rate 70%; Unemployment rate 4%; Participation rate 73%; US state where majority lives: California.

### Comparison 2019 vs 2020 (selected indicators from Appendix Table 2)
- USA: Employment Rate 2019 60% / 2020 56%; Unemployment Rate 2019 4% / 2020 8%; Participation Rate 2019 62% / 2020 60%.
- BHS: Employment Rate 2019 69% / 2020 56%; Unemployment Rate 2019 5% / 2020 27%; Participation Rate 2019 72% / 2020 77%.
- CRI: Employment Rate 2019 62% / 2020 73%; Unemployment Rate 2019 3% / 2020 8%; Participation Rate 2019 64% / 2020 79%.
- GTM: Employment Rate 2019 69% / 2020 63%; Unemployment Rate 2019 3% / 2020 10%; Participation Rate 2019 71% / 2020 70%.
- NIC: Employment Rate 2019 64% / 2020 55%; Unemployment Rate 2019 5% / 2020 7%; Participation Rate 2019 67% / 2020 59%.

*IMF Working Paper: The Propensity to Remit: Macro and Micro Factors Driving Remittances to Central America and the Caribbean (WP/2022/203).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022203-print-pdf.pdf_