## _wp0944

## Source details

**Canonical URL:** [_wp0944](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp0944.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp0944.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp0944.pdf.json)

---

### Mechanisms linking exchange rates and wages
- Exchange rate depreciation can reduce real wages by making imported goods more expensive and increasing the consumer price index at unchanged nominal wages.
- Depreciation can enhance competitiveness, expand local production, raise labor demand, and increase real wages (Campa and Goldberg, 2001; Goldberg and Tracey, 2003).
- Depreciation can raise costs of imported capital and intermediate goods, and technology (often complements to domestic labor), reducing labor demand and real wages (Robertson, 2003).
- Exchange rate fluctuations may influence inflation expectations and enter wage-setting mechanisms; in some countries wages are indexed to foreign currency.
- A distinct channel: depreciation raises the value in domestic currency of wages earned abroad and raises reservation wages at home; if workers can migrate (or threaten to migrate) this labor supply response can affect domestic wages.

### Identification strategy and empirical approach
- New empirical strategy exploits cross-country variation in the degree of integration between domestic and international labor markets to isolate the labor supply channel.
- The effect of a devaluation on wages is larger if:
  - the cost of migrating abroad is low;
  - workers have information about outside options;
  - it is easy to transfer remittances or consume abroad and remit home.
- Countries with migration histories have lower costs of moving and better information via networks.
- To control for labor demand shifts, empirical specifications include exports and imports; identification assumes shifts in labor demand in response to depreciation are uncorrelated with degree of labor market integration after controlling for exports and imports.
- Empirical estimation includes country fixed effects and year fixed effects; explanatory variables lagged by one year.

### Measures of labor market integration (proxies used)
- Size of networks of nationals living abroad (past migration stocks).
- Remittances to GDP.
- Common language with destination countries.
- Historical ties between origin and destination countries.
- Geographical proximity to destination countries.
- Migration-weighted exchange rates (alternative exchange rate definition).
- Composite “super-integration” index: common official language (at least 10 percent of population shares an official language with any of top-five destinations), common border with any of top-five destinations, and colonial linkage to any of top-five destinations (a country is defined as integrated only if all three conditions are satisfied).

### Theoretical framework (sending and receiving countries)
- Two-country model (Home H and Foreign F); labor demand depends on nominal wage w, domestic price index P, foreign price index *P, exports E and imports M, nominal exchange rate e; exchange rate affects labor demand only through trade.
- Labor supply introduces I (degree of labor market integration) and foreign wage w*; fraction β denotes the share of foreign wage consumed at home (remittances/return migration).
- Key reduced-form implication: ln(w/P) = a0 + a1 ln e + a2 ln E + a3 ln M + a4 I + aI * ln w* + ... — the higher I, the larger impact of a change in the real exchange rate on real wages in sending countries.
- For receiving countries (U.S. focus): ln(w_USi) = - ln(P_US) + ln(e_{iUS}) * fI + x_US + x_i + ... ; higher I_{i,US} implies a given origin-country real depreciation leads to a larger drop in immigrant wages in the U.S.

### Empirical specification and data
- Estimation equation compact form: ln(w_it/P_it) - ln(P*_it/P_it) = β * I_it-1 * ee_it-1 + θ * X_it-1 + s_i + v_t + ε_it.
- Sample: 66 countries over 1981–2005 (dependent variable: average real wage earned in manufacturing per hour).
- Main nominal wage source: Labor Statistics database (International Labor Organization); deflated using CPI from International Financial Statistics (IMF).
- Alternative wage sources: International Financial Statistics (IMF, line 65) and Freeman and Oostendorp database.
- U.S. immigrant wages: IPUMS-CPS (1994–2005), sample restricted to foreign-born individuals aged 18–64 in civilian labor force; average wage constructed using sampling weights; wage adjusted for inflation using CPI in origin countries.
- Migration data: International Migration Statistics (IMS) dataset for OECD countries (OECD, 2006) via SourceOECD; main analysis uses immigrants defined by nationality; emigration rates = migrant stock normalized by origin-country population; one-year lagged emigration rates used as principal measure of labor market integration.

### Key empirical findings and magnitudes (sending countries)
- Main result: "the elasticity of domestic wages to real exchange rate is 0.1 after one year in countries with high barriers to external labor mobility while it is four times higher (about 0.4 percent) in countries with low barriers to mobility, where labor market integration is defined in terms of past migration rates."
- Baseline pass-through: “28 percent of a devaluation feeds into wages within one year.”
- Selected coefficient estimates (preserved exactly):
  - Table 1: Ln real exchange rate t-1 coefficient = 0.312*** (0.102); 0.336** (0.139) when controlling for imports/exports.
  - Table 2 interaction: Ln real exchange rate t-1 * ln emigration rate t-1 coefficients range from 0.044*** to 0.092*** across specifications (standard errors reported in source).
  - Table 3 migration-weighted: Ln migration-wt exchange rate t-1 = 0.107*** (0.035); interaction = 0.022*** (0.007).
  - Table 4 composite measure interaction = 0.416*** (0.150) and 0.384*** (0.130).
- Interaction term between ln(real exchange rate) t-1 and ln(emigration rate) t-1 is consistently positive and statistically significant (often at 1 percent).
- Ln (imports/GDP) t-1 is negative and often significant (examples: -0.434***; -0.421***).
- Sample sizes and fit examples:
  - Table 1 Observations = 801740; Number of countries = 6666; R-squared = 0.97 0.97.
  - Table 2 Observations vary by column: 740, 710, 574, 419, 393, 393; Number of countries: 66, 66, 58, 47, 44, 44; R-squared ≈ 0.97–0.98.

### Complementary evidence: receiving countries, migration, and remittances
- Receiving-country (U.S.) immigrant wage results (Table 10 and A7):
  - Interaction Ln real exchange rate t-1 * ln emigration rate to the US t-1 is negative and statistically significant in several specifications (e.g., -0.106**, -0.111**, -0.158*), implying a devaluation in origin country depresses wages of immigrants in the U.S. more when origin countries are more integrated with the U.S.
  - Example magnitudes: for a country at the 99th percentile of emigration rates, a 1 percent depreciation vis-à-vis the U.S. dollar is associated with a 0.03 percent increase in real wages of immigrants from that country in the U.S.; for a country at the 1st percentile, real wages increase by ¾ percent (source text).
  - Table A7 column [1] example: Ln real exchange rate t-1 = 0.555 (0.477); Ln real exchange rate t-1 * ln emigration rate to the US t-1 = -0.116 (0.107); Ln emigration rate to the US t-1 = -0.626** (0.305); Observations = 264; Number of countries = 47; R-squared = 0.36.
- Exchange rates and migration/remittances (Table 11):
  - Selected coefficients for Ln real exchange rate t-1 on migration/remittances: 0.364***, 1.161***, 0.661***, 0.118, 0.245, 0.372** (across columns).
  - Key substantive summary: "A 1 percent depreciation of the real exchange rate is associated with a little over ½ percent increase in the emigration rate; and a more than ⅓ percent increase in remittances/GDP."
  - These results support the labor supply channel linking exchange rates to wages.

### Robustness and alternative measures
- Results robust to:
  - Inclusion of country- and time-fixed effects.
  - Country-time varying controls: trade flows, crisis episodes (negative per capita real GDP growth), unemployment, FDI, measures of labor-market institutions, foreign wages and prices.
  - Alternative exchange rate definitions: migration-weighted real exchange rate (correlation with U.S.-vis real exchange rate ≈ 0.98), real trade-weighted exchange rate.
  - Alternative integration measures: remittances/GDP, past emigration stocks, composite "super-integration" index.
  - Different samples: developing vs developed countries (evidence of stronger channel in developing countries); migrant definition by nationality vs foreign-born (correlation between datasets = 0.95).
- Composition of trade and capital intensity:
  - Interaction between measures of labor market integration and trade is "insignificant in most specifications."
  - Controlling for capital-intensive composition of exports/imports does not overturn main interaction result (Table 7).
- Alternative wage data and skill breakdown:
  - IFS and Freeman-Oostendorp wage series generally corroborate interaction effect though sample sizes are limited; Table 9 shows interaction positive and significant for both skilled and unskilled wages (e.g., Ln real exchange rate t-1 * ln emigration rate t-1 = 0.514**, 0.556**, 0.496**, 0.534** across columns).
- Panel unit root tests (Table A4) suggest little evidence of unit roots for Ln(real wage) and Ln(real exchange rate) (Levin-Lin ADF-stat values: -1.91261 and -2.67213; Im, Pesaran & Shin ADF-stat values: -3.86461 and -6.07111).

### Key statistics and sample summary (selected, preserved exactly)
- Migration and exchange rate correlations:
  - Correlation between two migration datasets = 0.95.
  - Correlation between real exchange rate vis-à-vis dollar and migration-weighted measure = 0.98.
- Migration and remittance magnitudes and trends:
  - 75 million (2.3 percent of total population) lived and worked outside their country of birth in 1965.
  - 175 million—that is 2.9 percent of the world population—in 2001.
  - On average over 1980–2005, about 36 percent of all the migrants to the OECD end up in the U.S.
  - Cost of sending remittances to Mexico declined from 15 percent of the amount sent to about 5 percent between 1990 and 2003.
- Summary statistics (Table A3 excerpts):
  - Real wage per hour (local currency units): Obs 740 Mean 2,839.66 Std. Dev. 60.00 Min 1 Max 163.08
  - Lag real exchange rate (local currency units per US$): Obs 740 Mean 2.07 Std. Dev. 13.80 Min 0.00 Max 264.29
  - Lag emigration rate to the OECD: Obs 740 Mean 2.20 Std. Dev. 3.27 Min 0.00 Max 21.14
  - Lag Remittances/GDP: Obs 333 Mean 1.28 Std. Dev. 2.11 Min 0.04 Max 13.40
  - Real wage per hour of immigrants in the United States: Obs 546 Mean 5.72 Std. Dev. 3.32 Min 0.31 Max 25.00

### Interpretation and policy implications
- Core interpretation: the effect of a devaluation on wages depends on the degree of integration between domestic and international labor markets; higher labor market integration amplifies the wage response to real exchange rate changes via a labor supply (migration/reservation wage) channel.
- Policy implications highlighted:
  - Labor mobility should be included in macro models of external adjustment; omission is a grave limitation.
  - Including labor mobility affects welfare calculations for the effects of devaluation and the optimal speed of adjustment in response to external shocks.
  - Policymakers in sending countries should consider labor mobility linkages when assessing domestic wage impacts of exchange rate shocks.
  - Migration policies and facilitation of integration can influence macro-labor outcomes in origin countries.
  - Heterogeneous effects imply stronger attention to exchange rate–migration interplay in developing-country policy design.

### Conclusions and suggested avenues for future work
- Main empirical conclusion: "The effect of a devaluation on wages depends on the degree of integration" between domestic and international labor markets; results robust across multiple controls, exchange rate measures, integration concepts, migrant definitions, and samples.
- Recommended future research:
  - Define more nuanced, skill-specific measures of labor market integration.
  - Study specific information channels through which markets become more integrated.
  - Theoretical macroeconomic models of devaluation and crisis should explicitly consider devaluation effects on labor markets when labor market integration is important.

*Source: _wp0944 — IMF working paper excerpt.*

### 1. Effect of Exchange Rates on Wages ............................................................................14

### 1. Effect of Exchange Rates on Wages

### Mechanisms linking exchange rates and wages
- Exchange rate depreciation can reduce real wages by making imported goods more expensive and increasing the consumer price index at unchanged nominal wages.
- Depreciation can enhance competitiveness, expand local production, raise labor demand, and increase real wages (Campa and Goldberg, 2001; Goldberg and Tracey, 2003).
- Depreciation can raise costs of imported capital and intermediate goods, and technology (often complements to domestic labor), reducing labor demand and real wages (Robertson, 2003).
- Exchange rate fluctuations may influence inflation expectations and enter wage-setting mechanisms; in some countries wages are indexed to foreign currency.
- A distinct channel: depreciation raises the value in domestic currency of wages earned abroad and raises reservation wages at home; if workers can migrate (or threaten to migrate) this labor supply response can affect domestic wages.

### Identification strategy and empirical approach
- New empirical strategy exploits cross-country variation in the degree of integration between domestic and international labor markets to isolate the labor supply channel.
- The effect of a devaluation on wages is larger if:
  - the cost of migrating abroad is low;
  - workers have information about outside options;
  - it is easy to transfer remittances or consume abroad and remit home.
- Countries with migration histories have lower costs of moving and better information via networks.
- The pass-through of exchange rates to wages depends on these conditions; higher labor market integration implies a larger wage response to a given depreciation.
- To control for labor demand shifts, the empirical specifications include exports and imports; the identification assumes shifts in labor demand in response to depreciation are uncorrelated with degree of labor market integration after controlling for exports and imports.

### Measures of labor market integration (proxies used)
- Size of networks of nationals living abroad (past migration stocks).
- Remittances to GDP.
- Common language with destination countries.
- Historical ties between origin and destination countries.
- Geographical proximity to destination countries.
- Migration-weighted exchange rates (alternative exchange rate definition).

### Key empirical findings and magnitudes
- Main result: "the elasticity of domestic wages to real exchange rate is 0.1 after one year in countries with high barriers to external labor mobility while it is four times higher (about 0.4 percent) in countries with low barriers to mobility, where labor market integration is defined in terms of past migration rates."
- Definition of high and low barriers: "Countries with high and low barriers to labor mobility are defined respectively by emigration rates being less than the 10th percentile and greater than the 90th percentile in the sample."
- Migration and remittance statistics and trends cited:
  - 75 million (2.3 percent of total population) lived and worked outside their country of birth in 1965.
  - 175 million—that is 2.9 percent of the world population—in 2001.
  - On average over 1980–2005, about 36 percent of all the migrants to the OECD end up in the U.S.
  - Cost of sending remittances to Mexico declined from 15 percent of the amount sent to about 5 percent between 1990 and 2003.
- Evidence on rapid mobility and exchange-rate sensitivity:
  - Hanson and Spilimbergo (1999): a devaluation of the Mexican peso by 10 percent vis-à-vis the U.S. dollar increases, ceteris paribus, border apprehensions by 6 to 8 percent.
  - Yang (2008): a 10 percent increase in the peso/$ exchange rate increases remittance receipts in pesos by 6 percent.
- Robustness of results:
  - Estimates are robust to inclusion of country- and time-fixed effects.
  - Robust to country-time varying controls such as trade flows, measures of crisis in origin country of migrants, unemployment, FDI, measures of labor-market institutions, and foreign wages and prices.
  - Robust to alternative exchange rate definitions (e.g., migration weighted exchange rate), alternative measures of integration (e.g., remittances to GDP and past emigration stocks), and different samples (developing vs. developed).

### Complementary evidence: wages in recipient countries and migration outcomes
- Hypothesis tested: a given depreciation in the origin country causes a larger decline (or smaller increase) in real wages of immigrants in the U.S., the more integrated the origin country’s labor market is with the United States.
- Finding: strong empirical evidence that devaluation in origin country exerts downward pressure on wages of immigrants in the U.S. when origin countries are more integrated with the U.S. labor market.
- Direct relationship between exchange rates and emigration/remittances:
  - Exchange rate depreciations are significantly associated with higher emigration rates and higher remittances to GDP, after controlling for push and pull factors.
- Complementary pieces of evidence on wages in sending and receiving countries, and on emigration and remittances, consistently point to the importance of labor market integration for the labor market effects of devaluation.

### Policy implications
- Labor mobility should be included in macro models of external adjustment; omission is a grave limitation.
- Including labor mobility affects:
  - Welfare calculations for the effects of devaluation.
  - The optimal speed of adjustment in response to external shocks.
- The paper’s results suggest that the welfare and distributional consequences of exchange rate policy depend on degree of labor market integration and the potential for migration responses.

### Organization of the paper (as presented)
- Section 2: theoretical framework analyzing effects of labor mobility on wages.
- Section 3: empirical implementation of the reduced form of the framework.
- Section 4: data with focus on measures of labor market integration.
- Section 5: empirical evidence on effect of migration in sending countries.
- Section 6: mirror evidence in the U.S. (main recipient country).
- Section 7: evidence on how migration rates and remittances respond to exchange rate movements.

*Source: _wp0944 - 1. Effect of Exchange Rates on Wages (IMF working paper content).*

### Section 8 concludes.

### _wp0944 - Section 8 concludes.

### Exchange rates and labor mobility: theoretical framework
- Paper develops a simple model of labor demand and supply for two countries: Home (H) and Foreign (F). Variables for F denoted with an asterisk (*). Time subscript t dropped for simplicity; empirical section allows lagged responses.
- Labor demand specification (equation (1)):
  - dL = wePeP LEMD PPP α− = ()()()  (expressed in original notation)
  - Definitions: dL is domestic labor demand; w is the nominal wage; P is the domestic price index; *P is the foreign price index; E and M are exports and imports; e is the nominal exchange rate (home currency units per unit of foreign currency).
  - E and M capture channels through which labor demand is exposed to goods market integration.
  - Assumption: exchange rate affects labor demand only through trade (consistent with prior literature, e.g., Campa and Goldberg, 2001).
  - If multiple foreign labor markets exist, *eP/P should become the real exchange rate defined as ∏j (wj e j Pj*/Pj) with weights wj calculated from trade flows.

### Labor supply in an integrated world
- Labor supply specification (equation (2)):
  - sL = [I w* w PPP φ γ β β β − − 1]  (expressed in original notation)
  - sL is domestic labor supply; w* is foreign wage; S captures country-specific historical determinants of labor participation (demographics, female labor force participation).
  - Iw* is introduced as a measure of the degree of labor market integration. If I = 0 (closed to international labor markets), labor supply depends only on domestic wages.
  - β denotes the fraction of wages a migrant consumes at home (remittances or return migration). If β = 1, foreign wages are deflated only by domestic prices.
  - Assumptions: 0 > φ (increase in the real exchange rate reduces domestic labor supply due to emigration). Higher labor market integration (higher I) magnifies the impact of a given real exchange rate increase on labor supply.
- Equilibrium condition (equation (3)): d = sL (expressed as ds LL =).
- Reduced-form log specification for real wage (equation (4)):
  - ln wP/P = a0 + a1 ln e + a2 ln E + a3 ln M + a4 I + aI * ln w* + ... (expressed in original compact notation)
  - Parameters a1..a4 defined in terms of structural parameters (φ, γ, β, α, X, etc.) as shown in the source.
  - Implication: the higher the degree of labor market integration, the larger the impact of a change in the real exchange rate on real wages in sending countries. Equation (4) is the reduced form for real wages in sending countries.

### Effects on receiving countries
- Framework applied to recipient countries (focus restricted to the U.S. due to data):
  - Reduced-form for wages of immigrants in receiving countries (equation (5)):
    - ln(w_USi) = - ln(P_US) + ln(e_{iUS}) * fI + x_US + x_i + ...  (expressed in original notation)
  - Subscript i indicates origin country of migrants.
  - I_{i,US} is a measure of labor market integration with the U.S.; I_{i,US} is positive and increasing in fI_{i,US}.
  - w_USi is the wage migrants from country i earn in the U.S.; P_US is the U.S. price index.
  - x_US and x_i are control variables in the U.S. and origin country respectively.
- Implications tested in Section 6:
  - (i) A devaluation in origin country i depresses the wage of immigrants from i in the U.S.
  - (ii) Higher degree of labor market integration with the U.S. implies a given exchange rate depreciation leads to a larger decline in immigrants’ wages.

### Empirical specification
- Adjustments made to the theoretical model for estimation:
  - Country-specific fixed effects included to account for time-invariant factors affecting wages (e.g., institutional factors).
  - Year-fixed effects included to control for world-wide factors that could generate spurious correlations.
  - Labor market integration variables and exchange rate included separately and interacted to test direct and interaction effects.
  - Foreign wages dropped from benchmark specifications because time dummies control for time-varying factors common across source countries (including recipient labor market conditions).
  - Explanatory variables lagged by one year to account for slow labor market reaction when migration is involved.
  - Model estimated in levels (panel unit root tests per Pedroni (2001) show little evidence of unit roots; see Table A4 in source).
- Estimation equation (6) (expressed in original notation):
  - ln(w_it/P_it) - ln(P*_it/P_it) = β * I_it-1 * ee_it-1 + θ * X_it-1 + s_i + v_t + ε_it  (compact representation of the source notation)
  - i denotes origin country; t denotes year; X_{it-1} are lagged controls; s_i and v_t are country and year fixed effects respectively.

### Data
- Sample: 66 countries over the period 1981–2005.
- Dependent variable: average real wage earned in manufacturing per hour.
- Main nominal wage source: Labor Statistics database (International Labor Organization).
  - Nominal wages provided in local currency; deflated using consumer price index (CPI) from International Financial Statistics (IMF, various years).
  - In most countries statistics refer to “wages and salaries” (direct wages, bonuses, gratuities); in some countries refer to “earnings” (broader compensation including paid leave, pension, insurance).
  - Conversion to hourly wages by dividing total wage payments by total number of hours worked (hours data from ILO).
- Two alternative wage data sources used for robustness checks:
  - International Financial Statistics (IMF, various years, line 65): wage indices with 2000=100; represent wage rates or earnings per worker per specified time period, typically manufacturing. Data include payments in kind and family allowances; cover salaried employees and wage earners.
  - Freeman and Oostendorp database of Occupational Wages around the World: based on October ILO inquiry; standardized and disaggregated by occupations. Coverage limited; analyses using these cover at most 30 countries.

### Wages in the U.S. by nationality of immigrants
- Data source: Integrated Public Use Microdata Series - Current Population Survey (IPUMS-CPS) for years 1994–2005.
  - IPUMS-CPS based on the March Annual Demographic File and Income Supplement to the Current Population Survey (CPS).
  - Sample restricted to foreign-born individuals aged 18–64 who participate in the civilian labor force.
  - Individual data averaged to construct mean hourly wage for immigrants in the U.S. from various origin countries.
  - Average wage constructed using sampling weights as recommended by IPUMS-CPS.
  - Wage adjusted for inflation using CPI in the origin countries.
  - Note: beginning 1994, the CPS included a question on country of birth of individuals.

### Migration data
- Source: International Migration Statistics (IMS) dataset for OECD countries (OECD, 2006), available through SourceOECD.
  - Immigrants in the OECD defined by nationality and/or country of birth.
  - Main analysis uses immigrants defined by nationality due to broader coverage (see appendix for migration data details).
  - Except Australia, Canada, Mexico and New Zealand, all other OECD countries record migrants by nationality.
  - IMS provides information on migrants in the U.S. by nationality for 1990; information on immigrants by birth available for 1980, 1990 and from 1995 onward.

*Source: _wp0944 - Section 8 concludes.*

### 2001. The correlation between the two sets of data is very high (0.95), and the results in the

### _wp0944 - 2001. The correlation between the two sets of data is very high (0.95), and the results in the

### Data and variable construction
- Migration data:
  - Bilateral stocks of migrants aggregated across destination countries to obtain emigrant stock in the OECD for each origin country.
  - Emigration rates = migrant stock normalized by origin-country population.
  - One-year lagged emigration rates used as principal measure of labor market integration.
  - Alternative migrant definition tested based on country of birth; correlation between the two datasets is 0.95.
  - On average, about 36 percent of all the migrants to the OECD end up in the U.S., relative to 22 percent in Europe.
  - Top-five destination-country information reported in Table A1 (year in period 1981–2005 with maximum number of destination countries).
  - Note: data on stock of immigrants defined by nationality is available for 185 source countries for 26 years vis-à-vis 179 countries for 16 years for the data on immigrants defined by country of birth.

- Exchange rates:
  - Data from International Financial Statistics (IMF, various issues).
  - Exchange rates expressed in nominal currency per U.S. dollar and deflated by CPI.
  - Migration-weighted exchange rate constructed by weighting bilateral exchange rates with share of migrants in destination countries; also deflated by CPI.
  - Correlation between real exchange rate vis-à-vis dollar and migration-weighted measure is high (0.98).
  - Primary measure used: real exchange rate vis-à-vis the U.S. dollar (considered relatively more exogenous to wages).
  - Also used: migration-weighted real exchange rates and real trade-weighted exchange rates.

- Remittances and composite measures:
  - Worker remittances defined as monetary transfers sent by workers abroad for more than one year; recorded under “current transfers” in IMF Balance of Payments Statistics Yearbook; normalized by GDP.
  - Composite “super-integration” index: common official language (at least 10 percent of population shares an official language with any of top-five destinations), common border with any of top-five destinations, and colonial linkage to any of top-five destinations. A country is defined as integrated only if all three conditions are satisfied.
  - CEPII dataset used for language, border, and colonial linkage indicators.

- Control variables:
  - Trade data: UN COMTRADE via WITS; exports and imports (U.S. dollars) deflated by GDP in U.S. dollars from IMF.
  - Unemployment rate and FDI (ratio of GDP): International Financial Statistics (IMF).
  - Episodes of crisis: defined as negative growth in real GDP per capita from World Development Indicators.
  - Tax wedge (labor market institutions): IMF database on structural reforms (IMF, 2008).

### Main empirical specification and identification
- Main regression framework based on equation (6) (see paper); dependent variable ln(real wage).
- Principal explanatory variable: ln(real exchange rate) t-1 and its interaction with ln(emigration rate) t-1 (or other integration measures).
- Fixed effects: country fixed effects and year fixed effects included in many specifications.
- Robust standard errors reported in parentheses.

### Key empirical findings (sending countries)
- Strong and significant correlation between real wages and real exchange rate:
  - Table 1: Ln real exchange rate t-1 coefficient = 0.312*** (0.102) in [1]; 0.336** (0.139) in [2] when controlling for imports/exports.
  - Observations: 801740 in Table 1; Number of countries: 6666; R-squared: 0.97 0.97.
- Pass-through magnitude:
  - “The pass-through from real exchange rate to wages is large: 28 percent of a devaluation feeds into wages within one year.”
- Interaction with labor market integration (Table 2 main results):
  - Ln real exchange rate t-1 coefficients across columns [1]–[6]: 0.307**, 0.142, 0.275, 0.340, 0.207, 0.19 (standard errors shown in table).
  - Interaction term Ln real exchange rate t-1 * ln emigration rate t-1: coefficients 0.044***, 0.044***, 0.050**, 0.070***, 0.093***, 0.092*** (standard errors in parentheses).
  - Ln emigration rate t-1 alone: 0.120***, 0.137***, 0.146***, 0.187***, 0.239***, 0.234***.
  - Ln (imports/GDP) t-1 is negative and often significant (e.g., -0.434*** in Table 1; -0.421*** in Table 2 column [1]).
  - Table 2 summary statistics: Observations vary by column: 740, 710, 574, 419, 393, 393; Number of countries: 66, 66, 58, 47, 44, 44; R-squared around 0.97–0.98.

- Robustness to controls and endogeneity concerns:
  - Controlling for crisis episodes (dummy for crisis defined as negative growth in real GDP per capita) does not overturn main result; interaction effect remains.
  - Additional controls included across specifications: unemployment rate, tax wedge, Ln (FDI/GDP), Ln average OECD wage, Ln average OECD price.
  - Interaction between exchange rates and emigration remains positive and significant across specifications [3]–[6].

### Robustness and alternative measures
- Alternative exchange rate measures (Table 3):
  - Migration-weighted real exchange rate t-1 coefficient examples: 0.107*** (0.035), 0.075** (0.033), 0.206*** (0.072), 0.204*** (0.067).
  - Interaction migration-wt * ln emigration rate t-1: 0.022***, 0.018**, 0.053***, 0.051***.
  - Real trade-weighted exchange rate t-1 coefficients and interactions also show positive and often significant interaction effects (e.g., Ln real trade-wt exchange rate t-1 * ln emigration rate t-1 = 0.061***, 0.068***, 0.082**, 0.080**).
  - Correlation between U.S.-vis real exchange rate and migration-weighted measure ≈ 0.98.

- Alternative measures of labor market integration (Table 4):
  - Using emigration stocks:
    - Ln real exchange rate t-1 * Ln emigration stocks t-1 = 0.084** and 0.083*** in columns shown.
    - Ln emigration rate t-1 = 0.213*** and 0.209*** in respective columns.
  - Using remittances/GDP:
    - Ln real exchange rate t-1 * ln (remittances / GDP) t-1 = 0.093*** and 0.087***.
    - Ln remittances to GDP t-1 = 0.058 and 0.071*.
  - Using composite emigration measure (super-integration):
    - Ln real exchange rate t-1 * Ln composite emigration measure t-1 = 0.416*** and 0.384***.
  - Conclusion: interaction coefficients always statistically significant (at the 1 percent level) and positive when using emigration stocks, remittances, or composite index.

- Heterogeneity: developing vs developed countries
  - Table 5 (described): inclusion of triple interaction (real exchange rate × migration rates × dummy for developing country) yields a positive and statistically significant interaction (at 5 percent level) in some columns, indicating the exchange-rate-to-wage channel via labor supply is stronger for developing countries.

- Alternative migrant definition—foreign-born (Table 6 summary):
  - Using foreign-born definition for OECD migrants: interaction between exchange rate and integration positive and significant at 10 percent in Columns [1]–[2]; turns statistically insignificant in Columns [3]–[4], driven by sample change rather than additional controls.

- Composition of trade (Table 7 summary):
  - Addressed via interactions of exports/imports with dummies for shares of capital-intensive exports/imports (NBER-CES Manufacturing Industry Database averaged across 4-digit SIC products 1986–97; capital-intensive defined by top 100 products).
  - Controlling for composition of trade does not alter main result: interaction between real exchange rate and labor market integration remains positive and statistically significant with similar magnitude to baseline.

### Interpretation and substantive analysis
- Core interpretation:
  - Exchange rate depreciations (real devaluations) raise real wages in origin (sending) countries through a labor supply channel when labor markets are integrated with destination countries.
  - Magnitude: baseline reported pass-through from real exchange rate to wages is about 28 percent within one year; interaction terms imply larger effects when emigration/integration is higher.
- Complementary findings:
  - Higher lagged emigration rates are associated with higher wages in origin countries (consistent with country studies).
  - Results are robust across alternative exchange rate measures, alternative integration proxies (stocks, remittances, super-integration), migrant definitions (nationality vs foreign-born, with some sensitivity due to sample coverage), and controls for trade composition.
- Mechanisms tested/considered:
  - Omitted variable concerns addressed via crisis control, unemployment, tax wedge, FDI, OECD wages/prices.
  - Potential endogeneity of migration-weighted exchange rate acknowledged; hence primary reliance on U.S.-vis real exchange rate.

### Key statistics and coefficients (selected, preserved exactly as reported)
- Correlations:
  - Correlation between two migration datasets = 0.95.
  - Correlation between real exchange rate vis-à-vis dollar and migration-weighted measure = 0.98.
- Pass-through statement:
  - “28 percent of a devaluation feeds into wages within one year.”
- Selected coefficient examples:
  - Table 1: Ln real exchange rate t-1 = 0.312*** (0.102); 0.336** (0.139).
  - Table 2 interaction: Ln real exchange rate t-1 * ln emigration rate t-1 = 0.044*** (0.014) ... up to 0.092*** (0.033) across columns.
  - Table 3 migration-weighted: Ln migration-wt exchange rate t-1 = 0.107*** (0.035); interaction = 0.022*** (0.007).
  - Table 4 composite measure interaction = 0.416*** (0.150) and 0.384*** (0.130).
- Sample sizes and fit:
  - Table 1 Observations = 801740; Number of countries = 6666; R-squared = 0.970.97.
  - Table 2 Observations range: 740, 710, 574, 419, 393, 393; Number of countries range: 66, 66, 58, 47, 44, 44; R-squared ~ 0.97–0.98.
  - Table 3 Observations by column examples: 740, 710, 393, 393, 719, 689, 393, 393; R-squared ~ 0.97–0.98.
  - Table 4 Observations examples: 393, 393, 333, 333, 441, 441; R-squared = 0.97 in reported columns.

### Policy-relevant implications (inferred from empirical results)
- Exchange rate policy and labor mobility:
  - Real depreciations can raise real wages in origin countries notably when labor markets are integrated; policymakers in sending countries should consider labor mobility linkages when assessing the domestic wage impact of exchange rate shocks.
- Migration and wage outcomes:
  - Higher emigration rates and stronger integration with destination-country labor markets amplify the wage response to exchange rate movements; migration policies and integration facilitation can therefore influence macro-labor outcomes in origin countries.
- Heterogeneous effects:
  - Effects stronger in developing countries (evidence of positive and significant triple interaction), implying targeted attention to exchange rate and migration interplay in developing-country policy design.
- Robustness to trade composition:
  - Results hold after accounting for capital intensity of trade; trade policy alone does not obviate the role of labor market integration in wage responses to exchange rates.

*Source: _wp0944 (PDF chapter/section) — empirical analysis and tables on exchange rates, migration measures, and effects on real wages.*

### 2. The interaction between capital intensity and trade is statistically insignificant in most of

### _wp0944 - 2. The interaction between capital intensity and trade is statistically insignificant in most of

### Main empirical finding on interaction between capital intensity and trade
- The paper reports that "The estimated coefficients on the interaction between measures of labor market integration and trade are insignificant in most specifications."
- A robustness check using exports and imports with dummies for the share of K-intensive exports and imports in overall being larger than 50 percent yields unchanged results (available upon request).

### Effect of exchange rates on origin-country wages (selected estimates)
- Table 5 (Developing countries) and Table 6 (Migrants defined by foreign-born) present panel regressions with dependent variable ln(real wage). Key coefficient patterns:
  - Ln real exchange rate t-1 coefficients reported across specifications: 0.307**, 0.137, 0.048, 0.134 (Table 5); 0.360**, 0.198, 0.091, 0.240 (Table 6).
  - Interaction Ln real exchange rate t-1 * Ln emigration rate t-1: 0.041***, 0.033***, 0.050**, 0.052** (Table 5); 0.043*, 0.042*, 0.032, 0.024 (Table 6).
  - Ln emigration rate t-1 coefficients: 0.119***, 0.134***, 0.204***, 0.198*** (Table 5); 0.009, 0.004, 0.005, 0.026 (Table 6).
  - Ln (imports/GDP) t-1 often negative and significant: -0.416***, -0.264***, -0.318, -0.251 (Table 5); -0.447***, -0.321***, -0.660**, -0.541** (Table 6).
- Sample sizes and fit (selected):
  - Observations: 740, 710, 393, 393 (Table 5); 623, 602, 338, 338 (Table 6).
  - Number of countries: 66, 64, 44, 44 (Table 5); 66, 64, 43, 43 (Table 6).
  - R-squared values around 0.97–0.98 in these tables.

### Interaction with labor market integration and composition of trade (Table 7)
- Table 7 reports specifications controlling for composition of trade (share of capital-intensive exports/imports). Selected reported coefficients:
  - Ln real exchange rate t-1: 0.312**, 0.137, 0.27, 0.362, 0.169, 0.183 (columns [1]–[6]).
  - Ln real exchange rate t-1 * ln emigration rate t-1: 0.038**, 0.042***, 0.046**, 0.065**, 0.087**, 0.085** (columns [1]–[6]).
  - Ln (imports/GDP) t-1 * Share of capital-intensive imports: 1.141***, 0.469, 0.359, 0.974**, 0.727, 0.779 (columns [1]–[6]).
  - Ln (exports/GDP) t-1 * Share of capital-intensive exports: 0.425, 0.607, 0.783, -0.121, 0.407, 0.404 (columns [1]–[6]).
- Observations by column: 740, 710, 574, 419, 393, 393. Number of countries: 66, 66, 58, 47, 44, 44. R-squared: 0.97–0.98.

### Alternative wage data sources and skill breakdown (Tables 8 and 9)
- Alternative wage series used: IFS and Freeman-Oostendorp.
  - Table 8 (IFS and Freeman-Oostendorp): Ln real exchange rate t-1 coefficients vary by specification, with some large positive coefficients in certain specifications (e.g., 5.784***, 4.775*** in columns [3]–[4] for one dataset).
  - The interaction Ln real exchange rate t-1 * ln emigration rate t-1 is estimated positive and statistically significant at 10 percent for IFS wages and strongly significant at 5 percent for Freeman-Oostendorp wages, though the number of observations is limited relative to preferred specifications.
- Skilled vs. unskilled wages (Table 9):
  - Ln real exchange rate t-1 coefficients: 5.725***, 4.751***, 5.780***, 4.776*** (columns [1]–[4]).
  - Ln real exchange rate t-1 * ln emigration rate t-1: 0.514**, 0.556**, 0.496**, 0.534** (columns [1]–[4]).
  - The interaction between labor market integration and real exchange rates is positively and statistically significant at the 5 percent level on both skilled and unskilled wages. Estimated magnitude is higher for unskilled wages, but the difference is not statistically significant.
  - Observations: 149 in Table 9 columns; Number of countries: 30; R-squared around 0.90–0.92.

### Receiving-country results — U.S. immigrant wages (Table 10)
- Specification for immigrants’ real wages in the U.S. uses interaction between exchange rate movements and emigration rates to the U.S.
- Selected coefficients (Table 10):
  - Ln real exchange rate t-1: 0.376, 0.361, 0.914**, 0.751* (columns [1]–[4]).
  - Ln real exchange rate t-1 * ln emigration rate to the US t-1: -0.106**, -0.111**, -0.158*, -0.098 (columns [1]–[4]).
  - Ln emigration rate to the US t-1: -0.226, -0.213, -0.603**, -0.582* (columns [1]–[4]).
  - Observations: 546, 537, 289, 264. Number of countries: 74, 73, 47, 47. R-squared: 0.32, 0.32, 0.34, 0.36.
- Interpretation: The interaction between exchange rates and labor market integration with the U.S. is negative and statistically significant in columns [1], [2] and [3], implying that a given exchange rate devaluation leads to a larger decline (or smaller increase) in real wages of migrants in the U.S. from countries that are more integrated with the U.S.
  - Example magnitudes reported: For a country at the 99th percentile of emigration rates, a 1 percent depreciation of the real exchange rate vis-à-vis the U.S. dollar is associated with a 0.03 percent increase in real wages of immigrants from that country in the U.S.; for a country at the 1st percentile of emigration rates, real wages increase by ¾ percent.
  - Note: estimates turn statistically insignificant in Column [4], and this change is driven by a change in the sample rather than adding additional controls (Table A7).

### Exchange rates and migration/remittances (Table 11)
- Table 11 analyzes the effect of exchange rates on emigration rates and remittances/GDP, controlling for push and pull factors.
- Selected coefficients:
  - Ln real exchange rate t-1 (dependent variable ln(emigration rate) or ln(remittances / GDP)): 0.364***, 1.161***, 0.661***, 0.118, 0.245, 0.372** (columns [1]–[6]).
  - Ln real wage t-1 coefficients: 0.213, 0.211, -0.143*, -0.131* (columns where reported).
  - Observations: 22968, 24824, 22482, 23280, 5805 (as reported across columns). Number of countries vary (reported values include 16, 66, 66, 14, 46, 36).
  - R-squared: 0.84, 0.83, 0.80, 0.77, 0.82, 0.82 (across columns).
- Key substantive magnitudes and interpretation:
  - "A 1 percent depreciation of the real exchange rate is associated with a little over ½ percent increase in the emigration rate; and a more than ⅓ percent increase in remittances/GDP."
  - These results provide evidence that exchange rate depreciations are associated with higher emigration rates and higher remittances/GDP, supporting the labor supply channel linking exchange rates to wages.

### Robustness, data coverage, and limitations
- Different definitions of migrants (e.g., foreign-born) affect the number of observations; the number of observations decreases dramatically when using the foreign-born definition.
- Alternative wage series and skill-disaggregated wages (Freeman-Oostendorp, IFS) generally corroborate the main interaction effect, but sample sizes are limited for these alternative data sources and results should be interpreted with caution.
- The paper notes that high-frequency migration data are noisy, making measurement of immediate migration responses to devaluation difficult.
- Country and year fixed effects are included in the majority of specifications; reported R-squared values are generally high (around 0.97) for origin-country wage regressions and lower for U.S. immigrant wage regressions (around 0.32–0.36).

### Conclusions and implications
- The paper develops a framework including reservation wages abroad and tests implications across four margins: wages in sending countries, wages of foreign-born individuals in the U.S., migration rates, and remittances.
- Main empirical conclusion: "The effect of a devaluation on wages depends on the degree of integration" between domestic and international labor markets; results are robust to multiple controls, alternative exchange rate measures, multiple concepts of labor market integration, different definitions of migrants, and different country samples.
- Policy and research implications:
  - Future empirical work should focus on defining more nuanced measures of labor market integration, including skill-specific labor market integration indices.
  - Future research should study specific information channels through which markets become more integrated.
  - Theoretical macroeconomic models of devaluation and crisis should explicitly consider the effect of devaluation on labor markets when labor market integration is important, as this could affect welfare analysis and policy dialogue.

*Italic: Content derived from the supplied IMF working paper excerpt.*

### REFERENCES

### _wp0944 - REFERENCES

### References cited
- Braga, Michela, 2008, “Dreaming Another Life: Role of Foreign Media in Migration Decision: Evidence from Albania,” mimeo, Bocconi University.
- Borjas, George, and Eric O’N. Fisher, 2001, “Dollarization and the Mexican Labor Market,” Journal of Money, Credit and Banking, Volume 33(2), May.
- Campa, Jose Manuel, and Linda S. Goldberg, 2001, “Employment Versus Wage Adjustment And The U.S. Dollar,” The Review of Economics and Statistics, Vol. 83(3), August, pp. 477–89.
- Carrington, William J, Enrica Detragiache, and Tara Vishwanath, 1996, “Migration with Endogenous Moving Costs,” American Economic Review, Vol. 86(4), pp. 909–30.
- Davis, Donald R., and Prachi Mishra, 2007, “Stolper-Samuelson is Dead and Other Crimes of Both Theory and Data” in Ann Harrison eds. Globalization and Poverty: University of Chicago Press and the National Bureau of Economic Research, March.
- Goldberg, Linda, and Joseph Tracy, 2003, “Exchange Rates and Wages,” mimeo, Federal Reserve Bank of New York.
- Hanson, H. Gordon, and Antonio Spilimbergo, 1999, “Illegal Immigration, Border Enforcement, and Relative Wages: Evidence from Apprehensions at the U.S.–Mexico Border,” The American Economic Review, Vol. 89 No. 5, pp. 1337–57.
- Hanson, H. Gordon, Raymond Robertson, and Antonio Spilimbergo, 2002, “Does Border Enforcement Protect U.S. Workers From Illegal Immigration?,” The Review of Economics and Statistics, Vol. 84(1), pp. 73–92.
- Hassett, Kevin, and Aparna Mathur, 2008, “Taxes and Wages,” American Enterprise Institute, mimeo.
- International Monetary Fund, International Financial Statistics. Various years.
- International Monetary Fund, 2005, World Economic Outlook, April.
- Massey, D.S., Espinoza, K.E., 1997, “What’s driving Mexico–US migration? A theoretical, empirical and policy analysis,” American Journal of Sociology, Vol.102 (4), pp. 939–99.
- Mishra, Prachi, 2007, “Emigration and Wages in Source Countries: Evidence from Mexico,” Journal of Development Economics, No. 82, pp. 180–99.
- Munshi, Kaivan, 2003, “Networks in the Modern Economy: Mexican Migrants in the U.S. Labor Market,” Quarterly Journal of Economics Vol. 118 No.2, pp. 549–97.
- OECD, 2006, “Immigrants and expatriates: Total population by nationality and country of birth,” Vol. 2006 release 01.
- Pedroni, Peter, 2001, “Purchasing Power Parity Tests in Cointegrated Panels,” The Review of Economics and Statistics, Vol. 83, No. 4, pp. 727.
- Robertson, Raymond, 2003, “Exchange Rates and Relative Wages: Evidence from Mexico,” North American Journal of Economics and Finance, Vol. 14, pp. 25–48.
- Yang, Dean, 2006, “Why Do Migrants Return to Poor Countries? Evidence from Philippine Migrants’ Responses to Exchange Rate Shocks,” Review of Economics and Statistics, Vol. 88, No.4, November 2006, pp. 715–35.
- Yang, Dean, 2008, “International Migration, Remittances, and Household Investment: Evidence from Philippine Migrants’ Exchange Rate Shocks,” The Economic Journal, Vol. 118, April 2008, pp. 591–630.

### Appendix — Wage data: definitions and coverage
- Source: ILO’s Key Indicators of the Labor Market (KILM).
- The ILO reports average earnings per worker or average wage rates; series may cover:
  - wage earners (manual or production workers),
  - salaried employees (non-manual workers),
  - all employees (wage earners and salaried employees).
- Coverage: workers of both sexes, irrespective of age.
- Concept of earnings:
  - Relates to remuneration in cash and in kind paid at regular intervals for time worked or work done, and remuneration for time not worked (annual vacation, other paid leave or holidays).
  - Excludes employers’ contributions to social security and pension schemes and benefits under these schemes, except where countries report such payments.
  - Excludes severance and termination pay.
  - Statistics refer to employees’ gross remuneration (total before deductions for taxes, employee social security/pension contributions, life insurance premiums, union dues, and other obligations).
- Specifically, earnings include:
  - (a) Direct wages and salaries for time worked or work done: (i) straight time pay of time-rated workers; (ii) incentive pay of time-rated workers; (iii) earnings of piece workers (excluding overtime premiums); (iv) premium pay for overtime, shift, night and holiday work; (v) commissions; premiums for seniority, special skills, geographical zone differentials, responsibility premiums, dirt/danger/discomfort allowances, guaranteed wage systems, cost-of-living allowances and other regular allowances.
  - (b) Remuneration for time not worked: payments in respect of public holidays, annual vacations and other paid time off.
  - (c) Bonuses and gratuities: seasonal and end-of-year bonuses, vacation supplements, profit-sharing bonuses.
  - Distinction: cash earnings vs payments in kind.
- Wage rates:
  - Include basic wages, cost-of-living allowances and other guaranteed/regular allowances.
  - Exclude overtime payments, bonuses and gratuities, family allowances and other social security payments made by employers.
  - Ex gratia payments in kind supplementary to normal wage rates are excluded.
- Cross-country differences arise because of:
  - (1) whether the reported statistic is wages or earnings;
  - (2) whether it covers employees, wage earners, or salaried employees;
  - (3) whether it includes employer social security contributions.
- Countries noted to include employer social security contributions in earnings data: Chile, Turkey, Colombia, Ecuador, Kenya, Kyrgyzstan, Mexico, Malaysia, Panama, Ukraine.
- Industrial classification changes: switchover from ISIC Revision 2 to Revision 3 in many countries since the beginning of the 1990s.
- Empirical approach: country fixed effects are included to allow for coverage differences in panel regressions.

### Appendix — Data on Migration: sources and definitions
- Principle sources of OECD migration statistics: population registers, residence or work permits, censuses, surveys; other sources include special surveys, counts at border crossings, analysis of landing cards.
- Immigrant population defined either as:
  - foreign nationals (European countries, Japan, Korea), or
  - foreign-born (Australia, Canada, New Zealand, United States).
- Foreign-born: represents first-generation migrants; may include foreign and national citizens; stock influenced by migration flows and mortality.
- Foreign nationals: may represent second and higher generations as well as first-generation; size/composition affected by migration history, natural increase, naturalizations.
- Measurement of “illegal immigrants”: data do not include explicit estimates; some stock data partially incorporate illegal migration (individuals may remain on population registers after permits expire).

### Appendix — Theoretical derivation (relationship between U.S. immigrant wages and labor market integration)
- Setting: two countries, the U.S. and origin country i; homogeneous labor; U.S. labor market segmented by nationality of immigrants.
- Labor demand for immigrants from country i in the U.S.:
  - Equation (A1): , () US dUS dUS i i US w L P β− = , X
  - Interpretation: is labor demand for immigrants from country i in the U.S.; is nominal wage of immigrants from country i in the U.S.; is the price index in the U.S.; is a composite term capturing other factors (income in the U.S.) affecting labor demand.
- Labor supply of immigrants from country i in the U.S.:
  - Equation (A2): , , (*) iUS US I , , sUS s is US i ii i w LeX P δ =
  - Where is labor supply of immigrants in the U.S. from origin country i; iP is domestic price in origin country i; i e is nominal exchange rate in local currency units per US$ in origin country i.
  - sUS X and ,si X are composite terms reflecting other factors in the U.S. and origin country affecting labor supply.
  - Key innovation: introduce IUS,i; assume 0>δ, i.e., ceteris paribus, an increase in the real exchange rate in the origin country increases labor supply of immigrants in the U.S.; higher labor market integration with the U.S. magnifies this effect.
- Equilibrium under segmented labor markets:
  - Equation (A3): , dUS sUS i , i LL=
  - Log-linearized form (A4): , ln() * ln US USi ii iUS US i we fIx PP =−+ +x
  - Interpretation: fI is defined as ' , , 0 , ()0; () iUS iUSiUS iUS I f I I δ β δ => + > US ; x and i x are control variables in the U.S. and origin country respectively.
  - Implication: higher degree of labor market integration of origin country with the U.S. implies a given change in real exchange rate leads to a larger drop in immigrant wages in the U.S.

### Table A1 (Countries and top destination shares, 1981-2005)
- Table lists for each origin country up to five top destination countries with shares in parentheses (selected excerpts preserved exactly as shown):
  - Australia: US  (45.55259), UK  (36.86518), Japan  (5.604491), Germany  (4.340753), Ireland  (1.774444)
  - Austria: Germany  (60.98495), US  (22.04737), Switzerland  (9.658471), Italy  (2.503857), Netherlands  (1.114762)
  - Bolivia: US  (95.54259), Italy  (2.607727), Sweden  (1.54805), Netherlands  (.3016379)
  - Brazil: Japan  (49.30959), US  (33.36343), Portugal  (10.16177), Italy  (3.764487), Spain  (2.194695)
  - Chile: US  (60.81044), Australia  (19.75738), Sweden  (7.607743), Spain  (5.304509), Italy  (2.291702)
  - China: US  (44.40633), Japan  (17.48408), Canada  (15.26432), Australia  (7.200474), Korea  (3.373995)
  - Mexico: US  (99.85799), Germany  (.0727241), Italy  (.0508816), Netherlands  (.008897), Sweden  (.0067854)
  - Pakistan: US  (55.37356), UK  (20.03692), Germany  (7.90257), Italy  (4.555847), Spain  (3.352581)
  - Philippines: US  (67.11703), Canada  (11.71443), Japan  (7.887843), Australia  (5.649026), Italy  (3.409102)
  - Russia: US  (67.92616), Germany  (18.91619), Finland  (3.053924), Greece  (2.720376), Italy  (1.681843)
  - Ukraine: US  (46.41976), Germany  (21.16271), Portugal  (11.50146), Czech Rep  (10.78997), Italy  (2.700366)
  - Zimbabwe: UK  (84.85722), New Zealand  (12.25763), Greece  (1.900802), Netherlands  (.6067292), Italy  (.3776146)
- Source noted: Q:\DATA\mobility\data\summarystatstable.dta
- Table title: Table A1. Countries in the Sample and Top Destination Countries, 1981-2005 (share of migrants in parentheses)
- Note: For each country, the share corresponds to the year in 1981-2005 with the maximum number of destination countries. Migrants in the OECD defined by nationality or country of birth.

### Table A2 — Years in the sample (1981-2005)
- Year / Number of observations:
  - 1981 22
  - 1982 24
  - 1983 22
  - 1984 22
  - 1985 20
  - 1986 25
  - 1987 24
  - 1988 24
  - 1989 25
  - 1990 25
  - 1991 27
  - 1992 29
  - 1993 29
  - 1994 28
  - 1995 29
  - 1996 38
  - 1997 48
  - 1998 48
  - 1999 49
  - 2000 50
  - 2001 39
  - 2002 40
  - 2003 26
  - 2004 22
  - 2005 5
  - Total 740
- Table title: Table A2. Years in the Sample. 1981-2005

### Table A3 — Summary statistics (selected variables)
- Real wage per hour (local currency units): Obs 740 Mean 2,839.66 Std. Dev. 60.00 Min 1 Max 163.08
- Lag real exchange rate (local currency units per US$): Obs 740 Mean 2.07 Std. Dev. 13.80 Min 0.00 Max 264.29
- Lag emigration rate to the OECD: Obs 740 Mean 2.20 Std. Dev. 3.27 Min 0.00 Max 21.14
- Lag exports / GDP: Obs 740 Mean 26.12 Std. Dev. 18.67 Min 3.71 Max 184.31
- Lag imports/GDP: Obs 740 Mean 30.02 Std. Dev. 19.06 Min 5.24 Max 160.88
- Lag crisis (per capita real GDP growth <0): Obs 710 Mean 0.06 Std. Dev. 0.24 Min 0.00 Max 1.00
- Lag unemployment rate: Obs 5748 Mean 0.24 Std. Dev. 4.94 Min 0.50 Max 25.20
- Lag tax wedge: Obs 4194 Mean 2.01 Std. Dev. 13.89 Min 6.93 Max 82.94
- Lag FDI/GDP: Obs 393 Mean 3.00 Std. Dev. 3.87 Min 0.00 Max 45.15
- Average nominal wage per hour in the OECD (local currency units): Obs 393 Mean 1201.75 Std. Dev. 313.71 Min 364.51 Max 1656.60
- Average CPI in the OECD: Obs 393 Mean 80.98 Std. Dev. 17.84 Min 40.98 Max 118.44
- Lag real migration-weighted exchange rate: Obs 740 Mean 1.18 Std. Dev. 10.56 Min 0.00 Max 242.14
- Lag real trade-weighted exchange rate: Obs 719 Mean 97.66 Std. Dev. 944.16 Min 0.51 Max 4874.69
- Lag Remittances/GDP: Obs 333 Mean 1.28 Std. Dev. 2.11 Min 0.04 Max 13.40
- Lag stock of migrants in the OECD (in '000): Obs 393 Mean 390.87 Std. Dev. 554.52 Min 0.16 Max 5895.74
- Real wage per hour of immigrants in the United States: Obs 546 Mean 5.72 Std. Dev. 3.32 Min 0.31 Max 25.00
- Lag emigration rate to the US: Obs 546 Mean 2.76 Std. Dev. 5.82 Min 0.02 Max 59.52
- Real wage from IFS (index number): Obs 2119 Mean 1.18 Std. Dev. 30.44 Min 6.45 Max 274.46
- Real wage per month; Freeman-Oostendorp (local currency units): Obs 1462 Mean 369.82 Std. Dev. 12814.97 Min 4.26 Max 132242.60
- Share of capital-intensive exports in overall (in percent): Obs 740 Mean 45.17 Std. Dev. 14.86 Min 12.14 Max 86.73
- Share of capital-intensive imports in overall (in percent): Obs 740 Mean 44.4 Std. Dev. 17.20 Min 31.59 Max 69.52
- Low-skill real wage per month; Freeman-Oostendorp database (local currency units): Obs 1461 Mean 938.79 Std. Dev. 9986.17 Min 3.94 Max 85986.69
- High-skill real wage per month; Freeman-Oostendorp database (local currency units): Obs 1463 Mean 3009.25 Std. Dev. 17340.56 Min 4.85 Max 189400.30
- Table title: Table A3. Summary Statistics

### Table A4 — Panel unit root tests (selected)
- Variables tested: Ln (real wage), Ln (real exchange rate)
- Levin-Lin ADF-stat:
  - Ln (real wage): -1.91261
  - Ln (real exchange rate): -2.67213
- Im, Pesaran & Shin ADF-stat:
  - Ln (real wage): -3.86461
  - Ln (real exchange rate): -6.07111
- Number of countries: 69
- Number of periods: 25
- Notes: Missing values for intermediate years were interpolated to apply unit root tests. All reported values are distributed N(0,1) under null of unit root or no cointegration. Large negative values imply rejection of the unit root; 5% critical value is -1.64.
- Table title: Table A4. Panel Unit Root Test

### Table A5 — Effect of exchange rates on wages: Migrants defined by foreign-born (sample-selection)
- Dependent variable: ln(real wage)
- Selected coefficient estimates (columns [1] and [2] formatted as in source):
  - Ln real exchange rate t-1: 0.159 (column [1]); -0.020 (column [2])
    - Standard errors: (0.322) and (0.114)
  - Ln real exchange rate t-1 * ln emigration rate t-1: 0.015 (column [1]); 0.009 (column [2])
    - Standard errors: (0.027) and (0.027)
  - Ln emigration rate t-1: -0.039 (column [1]); 0.025 (column [2])
    - Standard errors: (0.048) and (0.037)
  - Ln (exports/GDP) t-1: 0.256 (column [1]); 0.409** (column [2])
    - Standard errors: (0.263) and (0.190)
  - Ln (imports/GDP) t-1: -0.549** (column [1]); -0.476** (column [2])
    - Standard errors: (0.231) and (0.223)
- Country fixed effects: Y Y
- Year fixed effects: Y Y
- Observations: 338 338
- Number of countries: 43 43
- R-squared: 0.96 0.96
- Notes: * significant at 10%; ** significant at 5%, *** significant at 1%. Robust standard errors in parentheses. All variables refer to the origin country of migrants.
- Table title: Table A5. Effect of Exchange Rates on Wages: Migrants defined by Foreign-Born: Sample-Selection

### Table A7 — Effect of exchange rates on U.S. immigrant wages (interactions, sample selection)
- Dependent variable: ln (real wage of immigrants in the U.S.)
- Reported coefficients (column [1]):
  - Ln real exchange rate t-1: 0.555
    - Standard error: (0.477)
  - Ln real exchange rate t-1 * ln emigration rate to the US t-1: -0.116
    - Standard error: (0.107)
  - Ln emigration rate to the US t-1: -0.626**
    - Standard error: (0.305)
  - Ln (exports/GDP) t-1: -0.197
    - Standard error: (0.228)
  - Ln (imports/GDP) t-1: 0.036
    - Standard error: (0.345)
- Country fixed effects: Y
- Year fixed effects: Y
- Observations: 264
- Number of countries: 47
- R-squared: 0.36
- Notes: * significant at 10%; ** significant at 5%, *** significant at 1%. Robust standard errors in parentheses. All explanatory variables refer to the origin country of migrants.
- Table title: Table A7. Effect of Exchange Rates on U.S. Immigrant Wages-Interactions: Sample Selection

### Freeman-Oostendorp occupations list (selected)
- The document reproduces extensive occupation-by-skill categorizations from the Freeman-Oostendorp Occupational Wages Around the World Database. Examples preserved exactly as listed:
  - Occupation — Skill: Farm supervisor — Skilled; Field crop farm worker — Unskilled; Plantation supervisor — Skilled; Plantation worker — Unskilled; Forestry worker — Unskilled; Logger — Unskilled; Deep-sea fisherman — Unskilled; Coalmining engineer — Skilled; Miner — Skilled; Underground helper, loader — Unskilled; Petroleum and natural gas engineer — Skilled; Petroleum and natural gas extraction technician — Skilled; Supervisor or general foreman — Skilled; Derrickman — Unskilled; Quarryman — Unskilled; Butcher — Unskilled; Packer — Unskilled; Dairy product processor — Unskilled; Grain miller — Unskilled; Baker (ovenman) — Unskilled; Thread and yarn spinner — Unskilled; Loom fixer, tuner — Unskilled; Cloth weaver (machine) — Unskilled; Labourer — Unskilled; Garment cutter — Unskilled; Sewing-machine operator — Unskilled; Tanner — Unskilled; Leather goods maker — Unskilled; Clicker cutter (machine) — Unskilled; Laster — Unskilled; Shoe sewer (machine) — Unskilled; Sawmill sawyer — Unskilled; Veneer cutter — Unskilled; Plywood press operator — Unskilled; Furniture upholsterer — Unskilled; Cabinetmaker — Unskilled; Wooden furniture finisher — Unskilled; Wood grinder — Unskilled; Paper-making-machine operator (wet end) — Unskilled; Journalist — Skilled; Stenographer-typist — Skilled; Office clerk — Skilled; Hand compositor — Skilled; Machine compositor — Skilled; Printing pressman — Skilled; Bookbinder (machine) — Skilled; Chemical engineer — Skilled; Chemistry technician — Skilled; Mixing- and blending-machine operator — Unskilled; Occupational health nurse — Skilled; Metalworking machine setter — Unskilled; Welder — Unskilled; Machinist and many others.
- Source cited for occupations: http://www.nber.org/oww/
- Table title: Table A6. List of Occupations: Freeman-Oostendorp Occupational Wages Around the World Database

*Source: _wp0944 - REFERENCES (PDF).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp0944.pdf_