The Labor Market Effects of U.S. Immigration : What is the Latest Evidence? Örn B. Bodvarsson St. Cloud State University St. Cloud, Minnesota & Institute for the Study of Labor (IZA) Bonn, Germany Western Migration Conference London, Ontario April 29, 2011

"By keeping labor supply down, immigration policy tends to keep wages high. Let us underline this basic principle: Limitation of the supply of any grade of labor relative to all other productive factors can be expected to raise its wage rate; an increase in supply will, other things being equal, tend to depress wage rates." (Paul Samuelson, 1964)

Jacques Chirac: “If there were fewer immigrants, there would be less unemployment, fewer tensions in certain towns and neighborhoods, and lower social cost.” Interviewer: “That has never been formally proven” Chirac: “It is easy to imagine, nevertheless” (1984 Interview with Liberation)

“Any sizable increase in the number of immigrants will inevitably lower wages for some American workers” (George Borjas, 2004)

“While immigration may have raised overall income slightly, many of the worst-off native- born Americans are hurt by immigration -- especially immigration from Mexico. Because Mexican immigrants have much less education than the average U.S. worker, they increase the supply of less-skilled labor, driving down the wages of the worst-paid Americans.” (Paul Krugman, 2007)

Story depends on some very simplifying assumptions: Immigration is exogenous Immigrants and natives are perfect substitutes in production Labor supply is perfectly inelastic Capital stock is fixed Product demand is constant The story predicts a short run, partial equilibrium effect of immigration on natives.

Questions: How well has the evidence supported the “popular” story for the USA? What sort of theoretical reappraisal has been done of the story? Is there any evidence to support the theoretical reappraisal? Where do we go from here?

How well has the evidence supported the “popular” story for the USA?

Common empirical measure used to estimate the effect of immigration on native wages: The factor price elasticity of the native- born wage in a particular skill class with respect to the supply of immigrant labor to that class; measures the % change in the wage due to a 1% change in supply

Traditional method used to estimate elasticity: spatial correlation analysis: 1.Cross-section: sample contains many spatial units, e.g. cities Typical empirical specification: Ln(Y ijt ) = a + β(Ln(M jt )) + α(lnX jt ) + ε jt Y ijt is wage paid to native i in a particular skill class in area j during t M jt is the fraction of immigrants in the skill class in j during t X ijt is a vector of controls

Most prominent example of cross- section approach: Card (2001): 175 largest U.S. cities, using data from 1990 U.S. Census; Approximately 2,750,000 observations comprising native- and foreign-born men and women; Main findings: overall factor price elasticity for native born men (women) is ( ) %

2. Unexpected Extraordinary Supply Shock Method Researcher exploits famous natural experiment involving an exogenous immigration shock to an area Many time periods before and after shock are observed Question: Do native-born wages in the treatment area after the shock change relative to outcomes in a counterfactual area? Prominent examples: Card’s (1990) study of Mariel Boatlift and Friedberg (2001) study of Russian Jews to Israel; Card found “Mariels” had benign effect on native-born wages; Friedberg found some evidence of a positive effect.

Using either of these strategies, the evidence generally doesn’t support the popular story! Friedberg and Hunt (1995); “Despite the popular belief that immigrants have a large adverse impact on the wages and employment opportunities of the native-born population, the literature on this question does not provide much support for this conclusion;” Longhi et al (2005); meta-analysis of over 350 estimates of factor price elasticity average about -0.1%; Bodvarsson & Van den Berg (2009): “A comparison of the evidence from all the studies surveyed [ ] makes it clear that, with very few exceptions, there is no strong statistical support for the view held by many members of the public, namely that immigration has an adverse effect on native-born workers in the destination country.”

What is wrong, theoretically, with the “popular” story?: Because it is short-term and a partial equilibrium one: It disallows secondary responses to immigration Disallows analysis of the case where immigration flows to many skill categories at the same time

The long term effects of these responses are that if markets are flexible: Host country native-born wages should eventually return to their pre- migration levels (Leamer-Levensohn (1995) “hypothesis of factor price insensitivity”) Thus, it shouldn’t be surprising to see benign effects of immigration on wages!

What secondary responses? Labor Demand Responses such as stimulation of capital inflows growth of immigrant-employing industries and substitution of immigrant-intensive technologies stimulation of product demand Labor Supply Responses such as * triggering of out-migration “Network effects” Markets must be flexible for these responses to occur

Problems with spatial correlation analysis as well: Usually treat the labor market as a local spatial unit; Because of secondary responses, wage differences between areas will be “arbitraged” away The problem is that typically we can’t control for these secondary responses in the data.

Recent Improvements in empirical methodology have addressed some of these issues: Spatial correlation method has largely been replaced by “Skill-Cell” method (Borjas (2003), latest version of method used in Aydemir & Borjas (2011)) Defines the labor market as a national skill category, thus supposedly avoiding the need to control for secondary adjustment responses occurring in regions.

Evidence from some prominent studies: Borjas (2003):factor price elasticity averages about -0.35; high-school dropouts most adversely affected by immigration; Aydemir & Borjas (2011); previous estimates are upwardly biased due to measurement error in immigrant supply shifts; factor price elasticity is actually around -0.5; Ottaviano and Peri (2005, 2006, 2008, 2010): average native wage increased in response to immigration; usually, workers with at least a high school degree gain; Borjas and Katz (2007): Studied Mexican migration to USA; factor price elasticity averages -0.64; native high school dropouts most adversely affected.

Ottaviano and Peri studies: Build a theoretical model that allows for capital market responses to immigration and which allows for immigration to many different skill categories at the same time; Theoretical model allows for immigration to one skill category to affect productivities and wages of workers in many other skill categories; Use national data for a much longer period ( ).

Main results from 2010 study NATIVE-BORN WORKER GROUP FACTOR PRICE ELASTICITY High school dropouts- 0.10% High school graduates College dropouts College graduates WEIGHTED AVERAGE FOR ALL NATIVE-BORN

Limitations to Ottaviano and Peri studies: U.S. immigration is treated as exogenous; Product market responses are not considered; Native/immigrant differences in labor supply not considered; It is an approach relying on simulations for key estimates

Recent research addressing secondary adjustment processes: * Internal migration responses (Beginning with Borjas (2006)) * Capital market responses (Ottaviano and Peri (2005, 2006, 2008, 2010) * Product demand responses (Hercowitz & Yashiv (2002), Saiz (2003, 2007), Bodvarsson and Van den Berg (2005), Lach (2007), Bodvarsson, et al. (2008)) * Product supply responses (Cortes (2008)) * Technology and industry structure responses (Lewis (2003, 2004, 2005), Card and Lewis (2005))

Are there internal migration responses? Evidence generally mixed, although strong recent evidence by Borjas (2006): For census data, native migration response softens the measured impact of immigration on wages by 40 to 60 percent; Freeman (2006): analysts still “…have reached no consensus about the extent to which internal migration explains the absence of any relation between immigration and wages”

What about the other adjustment processes? Immigration does not increase immigrant- intensive industries, but only increases employment shares of immigrants across all industries; Some evidence that immigration triggers the choice of immigrant-intensive technologies Unskilled immigration appears to lower prices of unskilled-intensive non-tradeable goods in U.S. cities

The strongest secondary adjustment response is consumer demand: Bodvarsson et al. (2008): Re-examination of Mariel Boatlift; Bodvarsson and Van den Berg (2006): study of Hispanic migration to meatpacking industry in Nebraska

Model used in Mariel Boatlift re- examination: Immigration shock triggers an “input substitution effect” (-) and a consumer demand effect (+, - or 0) The “full” effect of the shock on native wage is the sum of the two effects

Empirical findings: Test of retailing sector in Miami before and after Boatlift; counterfactual group of cities identical to Card (1990) The relative sizes of the “input substitution” and “consumer demand” effects are estimated The consumer demand effect was positive and found to more than offset the input substitution effect.

Dawson County, Nebraska case Migrants made up approximately 25% of county’s population in 2000, compared to 3% in 1990 Migrants absorbed by export-driven meatpacking; clean separation between the markets where migrants work and spend

Estimated consumer demand effects? Ceteris paribus, a new immigrant is predicted to raise a county’s: annual retail wage by $0.17 median housing price by $2.

Directions for future research: Need a general equilibrium theory that fully accounts for primary and secondary adjustments Theory should include fiscal effects, remittances, policy responses, and feedback “network effects”. Empirical specifications must be derived directly from theory and carefully account for all forms of simultaneity Conventional panel data may not do.