Economics of immigration
The current refugee crisis sparked a heated debate about the effects of immigration on the labour market. Here is what the economists have to say about it. This is a small - and in no way comprehensive ! - literature review of the topic that I made five years ago back to my student years.
Theoretical issues in measuring the impact of immigration
1st model: closed economy, one industry, one output good, wages determined by production technology and supply of factors - labor and capital, immigrant and native workers are perfect substitutes; labor supply perfectly inelastic. Adjustment to immigration is through wages.
- Immigrants affect wages and employment only if their skill distribution (the degree of substitution) is different from that of natives. If skill distribution is the same for both groups, immigration will only increase the scale of the economy through an increase in output with no effect on wages and employment of natives. If immigrants are, say, all unskilled - there would be a decrease in wages of unskilled and a subsequent increase in demand until all unskilled are employed but at a lower equilibrium wage. At the opposite, relative scarcity of skilled drives their wages up.
- If capital and skilled labor are complements, and unskilled labor is a substitute for the other two factors, immigration of unskilled workers lowers their wage of the unskilled. Lower unskilled wage induces firms to substitute away from capital and skilled labor to unskilled labor. But, the optimal output is now higher which induces firms to use more of all inputs. The final distribution of gains and losses is ambiguous.
- Immigration has no effect on average wages if capital is elastic in supply. If not, capital and skilled realize gains, unskilled lose, and average wages might decrease.
ð Crucial role of speed adjustment: the faster capital adjusts, the smaller is the effect on average wages in the economy.
ð Empirically: Ottaviano and Peri (2006) find a convergence rate of 10%; this means that instead of reducing the capital/labor ratio by 11% and so average real wages by 3,6%, immigration in the US in 1990-2004 reduced the capital/labor ratio by 3,4% and so average real wages only by 1,1%.
- If labor supply is not inelastic, immigration may cause voluntary unemployment among workers whose wages fall.
2nd model: small open economy. One good produced with intensive use of unskilled labor, the other - with intensive use of skilled labor. Output prices are determined on the world markets.
- Adjustment through output mix: increase in labor supply of the unskilled increases production of goods which use intensively the unskilled labor. Additional supply of unskilled drives wages down (and so profits up) in the sector using intensively the unskilled labor: new firms enter the industry, expanding the output, demand for unskilled grows until their wages are back on the pre-immigration level.
- Adjustment through technology: immigration increases relative supply of unskilled workers, and firms select from available technologies, the one which is more intensive in the use of unskilled labor. The economy absorbs additional workers without change in output mix or wage structure.
ð Empirically: 2/3 of the absorption of additional workers is done through endogenous changes in production technology.
To test for the impact of immigration on wages, we can use several methods:
1. Factor proportions approach. It proceeds by three steps:
- Estimate the amount and the skill distribution of immigrated labor
- Calculate the % change in the ratio (high skilled)/(low skilled) labor supply induced by this inflow
- Assess the effect of changes in the skilled endowments on earnings differentials by education.
2. Computable General Equilibrium Analysis: a set of equations describes the economy, calibration or estimation fixes the parameters; the model is solved for an equilibrium associated with an immigration inflow. A comparison between equilibria with and without immigration shock shows the impact of immigration on the income distribution.
3. Spatial correlation cross-section analysis: split the labor market by regions and study how wages change in cities or regions where the immigrant density increased compared to cities and regions where this density stayed the same. If wages increased relatively less in a region where immigrant density increased that would hint at the negative effect of immigration of wages.
1st problem: endogeneity. Immigrants may choose to settle in regions where wages or wage growth are higher. The estimate for the impact of immigration on wages would be downward biased or even show that greater immigrant density leads to higher wages. Ways out:
- Differencing: if immigrants choose their location based upon the level of the wage, but not on foreseen rises in it, the endogeneity issue may be waived by using data from two or more time periods.
- Natural experiments: eliminate the bias due to immigrants’ choice of locations based on the labor market conditions by looking at the cases where timing and location of immigration are exogenous (and so immigrants did not choose the location).
- Instrumental variables: example - historical settlement of immigrants (immigrants tend to settle where other immigrants had already settled, and past settlement decisions are likely to be unrelated to contemporary wage shocks). It the IV is also uncorrelated with omitted information on trade flows and the mobility of native labor and capital, the bias due to their omission is also remedied.
2nd problem: natives may move out in response to immigrant workers moving in. Then the wage impact of immigrant inflow would be dispersed throughout the economy. This would produce a downward bias in the estimate.
ð Borjas (2005): native migration can account for 40-60% of the difference in the measured wage impact of immigration between local and national labor market levels.
3rd problem: with free trade, capital mobility or labor mobility within the host country, wages are equalized by flows in goods or factors. So even if immigrants affect native wages at the national level, differences in immigrant density across the country don’t result in cross-section wage differences. Due to factor price equalization, cross-section studies can be biased toward zero.
ð Empirically: if there are mobility costs, wage differentials may persist in the steady state. Yet even in this case, relative wages across regions are less sensitive to immigration than the national wage, if domestic regions are more open to each other than the national economy is to other countries. Blanchard and Katz (1992): for the US, wage differentials caused by a demand shock of 1 % to employment growth in a particular state may persist for up to 10 years, unemployment and participation differentials for about 6 years. Decressin and Fatas (1995): for Europe, unemployment and participation are perturbed for about 4 years by region-specific shocks. These studies suggest that the bias toward zero due to factor price equalization may be small.
4. Skill/Occupation Cross-Section Analysis: slice the labor market by education/experience and use the fact that immigration affects different skill groups to a different extent.
1st problem: those individuals that immigrate belong to skill groups which do very well.
2nd problem: relies on the assumption that immigrants and natives are perfect substitutes.
5. Production Theory Approach: estimates a production function to compute elasticities of substitution between immigrants and natives, which allows computation of the effect of immigrants on other factor prices. Greenwood et al. (1996, 1997) note that some of the studies estimate production functions that violate the curvature conditions. Yet these conditions are part of the theoretical framework, and must be met for the estimates to make any economic sense.
6. Aggregate Time-Series Analysis: use of time-series data at a national level avoids any bias toward zero due to factor price equalization and endogenous regional choice of immigrants, but introduces a timing bias (toward zero): immigrants tend to come to a country when labor market outcomes are favorable.
Results:
1. Factor proportions approach
- Borjas et al. (1997) calculate the contribution of immigrants to the increasing wage gap in 1980-1995. The percentage wage differential skilled/unskilled natives increased from 30.1% to 41%. In the same period immigrants increased the relative supply of high school dropouts by 14.9%. 1/σ is around −0.322 for 1963-1987. This implies that immigrants reduced the relative wage of high school dropouts by −0.322 × 0.149 or 4.8 percentage points. Immigration accounts for about 44% of the widening wage gap between high school dropouts and high school graduates.
- Jaeger (1995): Immigration explains about 2.9 percentage points of the 13.4 percentage-point increase in the native dropout-college differential, but only 1.6 percentage points of a 12 percentage point increase in the native high school-college premium. Immigration in the 1980s accounted for roughly 1/3 of the decline in real wages for high school dropouts. The effects on the wage levels of other skill groups were comparatively smaller.
- Borjas (2003) increases the number of labor aggregates using a three-level CES technology. The bottom level combines similarly educated workers with different levels of work experience into labor supply for each education group. The second stage aggregates workers across education groups into the national workforce. Finally, the upper level combines labor with capital. He uses data for 4 education groups and 8 experience levels in 1960, 1970, 1980, 1990 and 2000 to estimate elasticities of substitution for each stage of the CES technology. With these estimates he calculates the wage impact of the immigrant influx that entered the USA between 1980 and 2000. Results show a wage decrease for the average native worker by 3.2%. Workers at the bottom and top of the education distribution are most affected with wage decreases of 8.9% and 4.9%, respectively.
2. Computable General Equilibrium Analysis
- Müller (1997): measures the impact of immigration in Switzerland and tests the sensitivity of the results for different modeling hypotheses on labor market segmentation, capital mobility and terms of trade. In general immigration has a positive but small effect on native welfare.
- Barrett et al. (2005): simulate the impact of immigrants for in Ireland in 1993-2003. Although the immigrants have higher levels of education relative to the domestic populations, they are not all employed in occupations that match their educational levels. Immigrants increased GNP by 3% but worsened the position of the low skilled who face lower wages or higher unemployment rates. The impact of immigrants would be more favorable (GNP 3% higher and earnings inequality smaller) in absence of occupational gap (if immigrants would have access to the same occupations as natives).
- Weyerbrock (1995): studies the effects of immigration into the EU. Unemployment increase or wage decreases are small even with huge migration flows. The more flexibly wages can react, the smaller possible negative effects will be.
- Boeri and Brücker (2005): analyze costs and benefits of East-West migration in the enlarged EU. When labor markets clear, immigration of 1% of the population increases GDP of the EU by around 0.3%. However, simulations with wage rigidities show that although the EU as a whole gains from migration, natives in receiving countries lose.
- Keuschnigg and Kohler (2002), Heijdra et al. (2002) and Brücker and Kohlhaas (2004): study the impact of EU enlargement on the labor markets of Austria and Germany. Immigration of 1% of the labor force produces a decline in wages of around 0.5% and an output increase.
- Sarris and Zografakis (1999): assess the impact of illegal immigration on the Greek economy. Illegal immigrants decrease real disposable income of households headed by an unskilled person (37% of the Greek economy) and benefit to all other households.
3. Spatial Cross-Section Analysis
- Altonji and Card (1991): use the stock of immigrants in 1970 as an IV for the change in the share of the foreign-born from 1970 to 1980. A 1 percentage point increase in the % of foreign-born in a city lowers wages of unskilled natives by 1.2%, at most. Had the US increased by 10% the number of immigrants in the 1980s, U.S. wages would have been 0.4% lower. If there were no illegal immigrants in California, the wage level would be 2.0-5.5% higher.
- LaLonde and Topel (1991): if immigrants affect each other more than they affect natives, any effect found is an upper bound on the impact of immigrants on native groups. The largest effect found is the impact of male immigrants who have been in the US for five or fewer years on their own group: increasing their share in the labor force by 10% reduces their wages by about 0.3%. If young blacks and Hispanics are included, their wages are found to be insignificantly affected by the immigrants.
- Goldin (1994): 1 percentage point increase in the share of the foreign-born lowers wages by 1-1.6%. Some of the results may suffer from the composition problem: if immigrants earn less than natives, cities with higher shares of immigrants will have lower average wages, even if immigrants have non-negative impact on the native wage.
- Pischke and Velling (1994): apply the technique developed in Altonji and Card to the study of the West German labor market for 1985 and 1989. No significant detrimental effects of immigrants on employment, unemployment or wages. - Ottaviano and Peri (2006): slice the market by education/age, natives, immigrants are imperfect substitutes even within the same education/age group; in 1990-2004, immigration increased the average wage of US native workers; large positive effect for highly educated natives, slight negative effect for low educated natives.
- Manacorda, Manning, Wadsworth (2006): cut the market by education/age; immigrants are imperfect substitutes even within the same education/age group; natives and immigrants are imperfect substitutes within the same group (estimated elasticity of substitution of 6); immigration in the UK had a small impact on natives’ wage distribution, impacting mostly wages of earlier immigrants.
- Dustman, Frattini, Preston (2008): immigrants compete with natives that are in the same percentile in the wage distribution rather than in skill groups;
ð Empirically, in the UK, distribution of immigrants by wage distribution is skewed towards the low and, while their skill distribution is skewed towards upper end! Then, pre-allocation based on observed skills would allocate immigrants to compete with native in skill groups where in reality they do not compete.
Immigration has a positive effect on average wages. Explanations: 1. Immigration surplus; 2) Wages initially deviate from their marginal products, immigration equalizes differences in marginal product, the surplus is captured by natives; 3. Higher skilled immigrants are more productive than natives, but paid the same wage. This generates a surplus captured by natives. Only a combination of these three reasons can explain positive wage effects of the observed magnitude.
Dustmann, Fabbri and Preston (2005) [1] explore the impact of immigration on British participation in the labor market, unemployment and wages. Theoretical predictions rely strongly on the relative skill composition of immigrant and native populations: there are no effects of immigration to be expected on labor market outcomes of residents if immigration does not affect the skill composition of the resident labor force, and if capital supply is perfectly elastic. They compare the skill and occupation composition for immigrants and natives and find that there is a remarkable similarity in education and skill distribution for natives, immigrants and recent immigrants. In the second step, they estimate the impact of movement in the immigrant population share on economic outcomes on interest (unemployment, wages, and participation) by performing spatial correlations across regions approach. They run three procedures: OLS estimation, estimation in differences and instrumental variables approach. They use measures of historic settlement patterns as instruments for immigrant inflows. The results show a positive and significant impact of immigration on wages for OLS (which is likely to be driven by the endogeneity bias). The estimates for the impact of immigration on wages in the differences estimation and IV frameworks are not statistically significant. Furthermore, results for wages are likely to be affected by a small sample size (data is only available starting from 1992). The effects of immigration on wages (obtained from the IV procedure) by different education group are never statistically significant for any group.
4. Production Theory Approach
- Borjas (1983, 1986b): immigrants and native-born male labourers are complements.
- Borjas (1987), Grossman (1982), Kohli (1999): small negative effects of immigrants on the earnings of natives although the numerical impact of this competition is trivial. Borjas (1987) finds that a 1% increase in the number of white immigrants reduces the earnings of white immigrants by 1%.
- Akbari and Devoretz (1992): foreign-born are neither substitutes nor complements to natives. Competition of new immigrants hits immigrants themselves most.
- Grossman (1982): a 10% increase in the number of employed immigrants reduces native wages by 1%. In this model natives are not distinguished by skill.
- Greenwood et al. (1996, 1997): estimate normalized quadratic functions that allow curvature conditions to be imposed. Negative effects on wages of native workers are small; however, wages of previous immigrants fall considerably.
5. Aggregate Time-Series Analysis
- Borjas, Freeman, and Katz (1992): immigration accounted for one quarter of the 10% decline in the relative earnings of high school dropouts from 1980-1988, a period when immigrants as a proportion of the labor force rose from 6.9% to 9.3%. This means a 1 percentage point increase in the share of immigrants reduces the absolute wage of dropouts by at most 1.2%.
- Borjas and Ramey (1993): 1 percentage point increase in fraction foreign-born reduces the wage of high school dropouts relative to college graduates by 0.6%.
ð Aggregate Time-Series Analysis may overstate immigration's impact on native wages because of the composition problem. It also constrains the effect on relative wages of increasing the supply of unskilled workers to be the same, whether the increase is due to natives or immigrants, so if unskilled natives are better substitutes for each other than are unskilled immigrants, the elasticity estimated will be larger than the true elasticity for unskilled immigrants. If foreign-born high school dropouts earn 20% less than native dropouts, the increase in their number from 1980- 1988 reduces the average dropout wage by 1.5% just due to the composition effect, which is large compared to the total impact of 2.5% ascribed to immigration in Borjas, Freeman, and Katz (1992). So it sets an upper bound on the negative effect of immigration on wages.
6. Natural Experiments
- Card (1990): examines the impact of the Cubans who increased Miami's population by about 7% in May 1980. The timing and the arrival location were exogenous. He considers wages, employment rates and unemployment rates for several groups of unskilled natives by ethnicity. Only Cubans appear to have been negatively affected, and the observed pattern matches the idea that this decline was a compositional effect.
- Hunt (1992): studies the repatriation to France of the “pieds-noirs” (skilled labor) in 1962, which increased the French labor force by 1.6%. Its timing was exogenous, and the location of the repatriates largely determined by climate and proximity to port of arrival. In a differenced cross-section analysis for the years 1962 and 1968, Hunt uses as IV’s the temperature of the region and the stock of pre-1962 repatriates. 1 percentage point increase in the repatriate share of the labor force reduced the wage of a region by at most 0.8% (zero in some specifications).
- Friedberg (2001): for Israel, identifies different labor markets by occupation and uses individual and aggregate data; no statistically significant impact of immigration on natives’ wages.
[1] Dustman, C., F. Fabbri and I. Preston (2005) The impact of immigration on the British Labour Market (joint with Francesca Fabbri and Ian Preston), The Economic Journal , 115(November), F324-F341
Theoretical issues in measuring the impact of immigration
1st model: closed economy, one industry, one output good, wages determined by production technology and supply of factors - labor and capital, immigrant and native workers are perfect substitutes; labor supply perfectly inelastic. Adjustment to immigration is through wages.
- Immigrants affect wages and employment only if their skill distribution (the degree of substitution) is different from that of natives. If skill distribution is the same for both groups, immigration will only increase the scale of the economy through an increase in output with no effect on wages and employment of natives. If immigrants are, say, all unskilled - there would be a decrease in wages of unskilled and a subsequent increase in demand until all unskilled are employed but at a lower equilibrium wage. At the opposite, relative scarcity of skilled drives their wages up.
- If capital and skilled labor are complements, and unskilled labor is a substitute for the other two factors, immigration of unskilled workers lowers their wage of the unskilled. Lower unskilled wage induces firms to substitute away from capital and skilled labor to unskilled labor. But, the optimal output is now higher which induces firms to use more of all inputs. The final distribution of gains and losses is ambiguous.
- Immigration has no effect on average wages if capital is elastic in supply. If not, capital and skilled realize gains, unskilled lose, and average wages might decrease.
ð Crucial role of speed adjustment: the faster capital adjusts, the smaller is the effect on average wages in the economy.
ð Empirically: Ottaviano and Peri (2006) find a convergence rate of 10%; this means that instead of reducing the capital/labor ratio by 11% and so average real wages by 3,6%, immigration in the US in 1990-2004 reduced the capital/labor ratio by 3,4% and so average real wages only by 1,1%.
- If labor supply is not inelastic, immigration may cause voluntary unemployment among workers whose wages fall.
2nd model: small open economy. One good produced with intensive use of unskilled labor, the other - with intensive use of skilled labor. Output prices are determined on the world markets.
- Adjustment through output mix: increase in labor supply of the unskilled increases production of goods which use intensively the unskilled labor. Additional supply of unskilled drives wages down (and so profits up) in the sector using intensively the unskilled labor: new firms enter the industry, expanding the output, demand for unskilled grows until their wages are back on the pre-immigration level.
- Adjustment through technology: immigration increases relative supply of unskilled workers, and firms select from available technologies, the one which is more intensive in the use of unskilled labor. The economy absorbs additional workers without change in output mix or wage structure.
ð Empirically: 2/3 of the absorption of additional workers is done through endogenous changes in production technology.
To test for the impact of immigration on wages, we can use several methods:
1. Factor proportions approach. It proceeds by three steps:
- Estimate the amount and the skill distribution of immigrated labor
- Calculate the % change in the ratio (high skilled)/(low skilled) labor supply induced by this inflow
- Assess the effect of changes in the skilled endowments on earnings differentials by education.
2. Computable General Equilibrium Analysis: a set of equations describes the economy, calibration or estimation fixes the parameters; the model is solved for an equilibrium associated with an immigration inflow. A comparison between equilibria with and without immigration shock shows the impact of immigration on the income distribution.
3. Spatial correlation cross-section analysis: split the labor market by regions and study how wages change in cities or regions where the immigrant density increased compared to cities and regions where this density stayed the same. If wages increased relatively less in a region where immigrant density increased that would hint at the negative effect of immigration of wages.
1st problem: endogeneity. Immigrants may choose to settle in regions where wages or wage growth are higher. The estimate for the impact of immigration on wages would be downward biased or even show that greater immigrant density leads to higher wages. Ways out:
- Differencing: if immigrants choose their location based upon the level of the wage, but not on foreseen rises in it, the endogeneity issue may be waived by using data from two or more time periods.
- Natural experiments: eliminate the bias due to immigrants’ choice of locations based on the labor market conditions by looking at the cases where timing and location of immigration are exogenous (and so immigrants did not choose the location).
- Instrumental variables: example - historical settlement of immigrants (immigrants tend to settle where other immigrants had already settled, and past settlement decisions are likely to be unrelated to contemporary wage shocks). It the IV is also uncorrelated with omitted information on trade flows and the mobility of native labor and capital, the bias due to their omission is also remedied.
2nd problem: natives may move out in response to immigrant workers moving in. Then the wage impact of immigrant inflow would be dispersed throughout the economy. This would produce a downward bias in the estimate.
ð Borjas (2005): native migration can account for 40-60% of the difference in the measured wage impact of immigration between local and national labor market levels.
3rd problem: with free trade, capital mobility or labor mobility within the host country, wages are equalized by flows in goods or factors. So even if immigrants affect native wages at the national level, differences in immigrant density across the country don’t result in cross-section wage differences. Due to factor price equalization, cross-section studies can be biased toward zero.
ð Empirically: if there are mobility costs, wage differentials may persist in the steady state. Yet even in this case, relative wages across regions are less sensitive to immigration than the national wage, if domestic regions are more open to each other than the national economy is to other countries. Blanchard and Katz (1992): for the US, wage differentials caused by a demand shock of 1 % to employment growth in a particular state may persist for up to 10 years, unemployment and participation differentials for about 6 years. Decressin and Fatas (1995): for Europe, unemployment and participation are perturbed for about 4 years by region-specific shocks. These studies suggest that the bias toward zero due to factor price equalization may be small.
4. Skill/Occupation Cross-Section Analysis: slice the labor market by education/experience and use the fact that immigration affects different skill groups to a different extent.
1st problem: those individuals that immigrate belong to skill groups which do very well.
2nd problem: relies on the assumption that immigrants and natives are perfect substitutes.
5. Production Theory Approach: estimates a production function to compute elasticities of substitution between immigrants and natives, which allows computation of the effect of immigrants on other factor prices. Greenwood et al. (1996, 1997) note that some of the studies estimate production functions that violate the curvature conditions. Yet these conditions are part of the theoretical framework, and must be met for the estimates to make any economic sense.
6. Aggregate Time-Series Analysis: use of time-series data at a national level avoids any bias toward zero due to factor price equalization and endogenous regional choice of immigrants, but introduces a timing bias (toward zero): immigrants tend to come to a country when labor market outcomes are favorable.
Results:
1. Factor proportions approach
- Borjas et al. (1997) calculate the contribution of immigrants to the increasing wage gap in 1980-1995. The percentage wage differential skilled/unskilled natives increased from 30.1% to 41%. In the same period immigrants increased the relative supply of high school dropouts by 14.9%. 1/σ is around −0.322 for 1963-1987. This implies that immigrants reduced the relative wage of high school dropouts by −0.322 × 0.149 or 4.8 percentage points. Immigration accounts for about 44% of the widening wage gap between high school dropouts and high school graduates.
- Jaeger (1995): Immigration explains about 2.9 percentage points of the 13.4 percentage-point increase in the native dropout-college differential, but only 1.6 percentage points of a 12 percentage point increase in the native high school-college premium. Immigration in the 1980s accounted for roughly 1/3 of the decline in real wages for high school dropouts. The effects on the wage levels of other skill groups were comparatively smaller.
- Borjas (2003) increases the number of labor aggregates using a three-level CES technology. The bottom level combines similarly educated workers with different levels of work experience into labor supply for each education group. The second stage aggregates workers across education groups into the national workforce. Finally, the upper level combines labor with capital. He uses data for 4 education groups and 8 experience levels in 1960, 1970, 1980, 1990 and 2000 to estimate elasticities of substitution for each stage of the CES technology. With these estimates he calculates the wage impact of the immigrant influx that entered the USA between 1980 and 2000. Results show a wage decrease for the average native worker by 3.2%. Workers at the bottom and top of the education distribution are most affected with wage decreases of 8.9% and 4.9%, respectively.
2. Computable General Equilibrium Analysis
- Müller (1997): measures the impact of immigration in Switzerland and tests the sensitivity of the results for different modeling hypotheses on labor market segmentation, capital mobility and terms of trade. In general immigration has a positive but small effect on native welfare.
- Barrett et al. (2005): simulate the impact of immigrants for in Ireland in 1993-2003. Although the immigrants have higher levels of education relative to the domestic populations, they are not all employed in occupations that match their educational levels. Immigrants increased GNP by 3% but worsened the position of the low skilled who face lower wages or higher unemployment rates. The impact of immigrants would be more favorable (GNP 3% higher and earnings inequality smaller) in absence of occupational gap (if immigrants would have access to the same occupations as natives).
- Weyerbrock (1995): studies the effects of immigration into the EU. Unemployment increase or wage decreases are small even with huge migration flows. The more flexibly wages can react, the smaller possible negative effects will be.
- Boeri and Brücker (2005): analyze costs and benefits of East-West migration in the enlarged EU. When labor markets clear, immigration of 1% of the population increases GDP of the EU by around 0.3%. However, simulations with wage rigidities show that although the EU as a whole gains from migration, natives in receiving countries lose.
- Keuschnigg and Kohler (2002), Heijdra et al. (2002) and Brücker and Kohlhaas (2004): study the impact of EU enlargement on the labor markets of Austria and Germany. Immigration of 1% of the labor force produces a decline in wages of around 0.5% and an output increase.
- Sarris and Zografakis (1999): assess the impact of illegal immigration on the Greek economy. Illegal immigrants decrease real disposable income of households headed by an unskilled person (37% of the Greek economy) and benefit to all other households.
3. Spatial Cross-Section Analysis
- Altonji and Card (1991): use the stock of immigrants in 1970 as an IV for the change in the share of the foreign-born from 1970 to 1980. A 1 percentage point increase in the % of foreign-born in a city lowers wages of unskilled natives by 1.2%, at most. Had the US increased by 10% the number of immigrants in the 1980s, U.S. wages would have been 0.4% lower. If there were no illegal immigrants in California, the wage level would be 2.0-5.5% higher.
- LaLonde and Topel (1991): if immigrants affect each other more than they affect natives, any effect found is an upper bound on the impact of immigrants on native groups. The largest effect found is the impact of male immigrants who have been in the US for five or fewer years on their own group: increasing their share in the labor force by 10% reduces their wages by about 0.3%. If young blacks and Hispanics are included, their wages are found to be insignificantly affected by the immigrants.
- Goldin (1994): 1 percentage point increase in the share of the foreign-born lowers wages by 1-1.6%. Some of the results may suffer from the composition problem: if immigrants earn less than natives, cities with higher shares of immigrants will have lower average wages, even if immigrants have non-negative impact on the native wage.
- Pischke and Velling (1994): apply the technique developed in Altonji and Card to the study of the West German labor market for 1985 and 1989. No significant detrimental effects of immigrants on employment, unemployment or wages. - Ottaviano and Peri (2006): slice the market by education/age, natives, immigrants are imperfect substitutes even within the same education/age group; in 1990-2004, immigration increased the average wage of US native workers; large positive effect for highly educated natives, slight negative effect for low educated natives.
- Manacorda, Manning, Wadsworth (2006): cut the market by education/age; immigrants are imperfect substitutes even within the same education/age group; natives and immigrants are imperfect substitutes within the same group (estimated elasticity of substitution of 6); immigration in the UK had a small impact on natives’ wage distribution, impacting mostly wages of earlier immigrants.
- Dustman, Frattini, Preston (2008): immigrants compete with natives that are in the same percentile in the wage distribution rather than in skill groups;
ð Empirically, in the UK, distribution of immigrants by wage distribution is skewed towards the low and, while their skill distribution is skewed towards upper end! Then, pre-allocation based on observed skills would allocate immigrants to compete with native in skill groups where in reality they do not compete.
Immigration has a positive effect on average wages. Explanations: 1. Immigration surplus; 2) Wages initially deviate from their marginal products, immigration equalizes differences in marginal product, the surplus is captured by natives; 3. Higher skilled immigrants are more productive than natives, but paid the same wage. This generates a surplus captured by natives. Only a combination of these three reasons can explain positive wage effects of the observed magnitude.
Dustmann, Fabbri and Preston (2005) [1] explore the impact of immigration on British participation in the labor market, unemployment and wages. Theoretical predictions rely strongly on the relative skill composition of immigrant and native populations: there are no effects of immigration to be expected on labor market outcomes of residents if immigration does not affect the skill composition of the resident labor force, and if capital supply is perfectly elastic. They compare the skill and occupation composition for immigrants and natives and find that there is a remarkable similarity in education and skill distribution for natives, immigrants and recent immigrants. In the second step, they estimate the impact of movement in the immigrant population share on economic outcomes on interest (unemployment, wages, and participation) by performing spatial correlations across regions approach. They run three procedures: OLS estimation, estimation in differences and instrumental variables approach. They use measures of historic settlement patterns as instruments for immigrant inflows. The results show a positive and significant impact of immigration on wages for OLS (which is likely to be driven by the endogeneity bias). The estimates for the impact of immigration on wages in the differences estimation and IV frameworks are not statistically significant. Furthermore, results for wages are likely to be affected by a small sample size (data is only available starting from 1992). The effects of immigration on wages (obtained from the IV procedure) by different education group are never statistically significant for any group.
4. Production Theory Approach
- Borjas (1983, 1986b): immigrants and native-born male labourers are complements.
- Borjas (1987), Grossman (1982), Kohli (1999): small negative effects of immigrants on the earnings of natives although the numerical impact of this competition is trivial. Borjas (1987) finds that a 1% increase in the number of white immigrants reduces the earnings of white immigrants by 1%.
- Akbari and Devoretz (1992): foreign-born are neither substitutes nor complements to natives. Competition of new immigrants hits immigrants themselves most.
- Grossman (1982): a 10% increase in the number of employed immigrants reduces native wages by 1%. In this model natives are not distinguished by skill.
- Greenwood et al. (1996, 1997): estimate normalized quadratic functions that allow curvature conditions to be imposed. Negative effects on wages of native workers are small; however, wages of previous immigrants fall considerably.
5. Aggregate Time-Series Analysis
- Borjas, Freeman, and Katz (1992): immigration accounted for one quarter of the 10% decline in the relative earnings of high school dropouts from 1980-1988, a period when immigrants as a proportion of the labor force rose from 6.9% to 9.3%. This means a 1 percentage point increase in the share of immigrants reduces the absolute wage of dropouts by at most 1.2%.
- Borjas and Ramey (1993): 1 percentage point increase in fraction foreign-born reduces the wage of high school dropouts relative to college graduates by 0.6%.
ð Aggregate Time-Series Analysis may overstate immigration's impact on native wages because of the composition problem. It also constrains the effect on relative wages of increasing the supply of unskilled workers to be the same, whether the increase is due to natives or immigrants, so if unskilled natives are better substitutes for each other than are unskilled immigrants, the elasticity estimated will be larger than the true elasticity for unskilled immigrants. If foreign-born high school dropouts earn 20% less than native dropouts, the increase in their number from 1980- 1988 reduces the average dropout wage by 1.5% just due to the composition effect, which is large compared to the total impact of 2.5% ascribed to immigration in Borjas, Freeman, and Katz (1992). So it sets an upper bound on the negative effect of immigration on wages.
6. Natural Experiments
- Card (1990): examines the impact of the Cubans who increased Miami's population by about 7% in May 1980. The timing and the arrival location were exogenous. He considers wages, employment rates and unemployment rates for several groups of unskilled natives by ethnicity. Only Cubans appear to have been negatively affected, and the observed pattern matches the idea that this decline was a compositional effect.
- Hunt (1992): studies the repatriation to France of the “pieds-noirs” (skilled labor) in 1962, which increased the French labor force by 1.6%. Its timing was exogenous, and the location of the repatriates largely determined by climate and proximity to port of arrival. In a differenced cross-section analysis for the years 1962 and 1968, Hunt uses as IV’s the temperature of the region and the stock of pre-1962 repatriates. 1 percentage point increase in the repatriate share of the labor force reduced the wage of a region by at most 0.8% (zero in some specifications).
- Friedberg (2001): for Israel, identifies different labor markets by occupation and uses individual and aggregate data; no statistically significant impact of immigration on natives’ wages.
[1] Dustman, C., F. Fabbri and I. Preston (2005) The impact of immigration on the British Labour Market (joint with Francesca Fabbri and Ian Preston), The Economic Journal , 115(November), F324-F341