Immigration Economics I

1
Immigration Economics (I)

• Libertad González
• Universitat Pompeu Fabra
• INSIDE
• October 26, 2006

2
Abstract

• In this session, I will try to provide an
  (incomplete) overview of the recent empirical
  literature on immigration economics, from the
  point of view of a labor economist. The
  presentation will be structured around the three
  main questions that the literature has focused
  on
• 1) Who migrates? (who decides to migrate, and
  how do they decide where to go?)
• 2) What happens to those who migrate? (immigrant
  assimilation, or how well do immigrants fare in
  the labor market relative to natives?), and
• 3) What impact do immigrants have on their
  destination country? (especially on the wage
  structure).
• I will try to lay out the main identification
  strategies, findings and methodological problems.

3
Main Questions

• Who migrates? (self-selection)
• What happens to those who migrate? (assimilation)
• What impact do immigrants have on their
  destination country?
• What impact do immigrants have on their source
  country?

4
Why Have Economists Been so Interested?

• Many rich countries have experienced large
  increases in immigration flows in recent decades.
• In Europe, immigrants account for almost 9 of
  total population in 2005.
• This increasing inflow has been
  disproportionately unskilled.
• In the US, 36 of high school dropouts in the
  labor force were foreign born in 2000.
• This has generated concerns about the effect on
  the wages of low-skill natives.
• Given the recent increases in wage inequality.

5
1. Who Migrates?

• Immigrants are not a randomly selected sample of
  the population in the source country.
• Analysis of the factors that motivate only some
  persons in the source country to migrate to a
  particular host country.
• Borjas, several years (AER 1987, etc).

6
A Simple Two-Country Model

• Two countries (S, H) with different wage
  distributions.
• Migration decisions are motivated by earnings
  differentials.
• The emigration rate is a function of
• Mean income in the host country ().
• Mean income in the source country (-).
• Migration costs (-).

7
The Self-Selection of Immigrants

• Who are those who migrate, relative to the whole
  population in S?
• High or low skill?
• The Roy model.
• It depends on
• The transferability of skills across the two
  countries.
• Wage dispersion in both countries.
• Depending on the values of the parameters, we can
  have positive or negative selection (or neither).

8
Main Message

• It is not obvious that immigrants should be
  positively selected.
• There will be positive selection if
• Skills are positively related across countries,
  AND
• Wage dispersion is higher in the source country.
• The degree of selection does NOT depend on mean
  incomes across countries, or the level of
  migration costs.

9
Testing for Selection

• The implications of the Roy model have been
  tested empirically, by
• Estimating the correlation between the earnings
  of immigrants in H and wage dispersion in S.
• The evidence suggests that immigrants from
  countries with higher rates of return to skills
  have lower earnings in the US.
• Borjas (1987, 1991), Cobb-Clark (1993).
• Bratsberg (1995), Taylor (1987).

10
Selection in Observed Characteristics

• The model can be extended to include both
  observed and unobserved skills.
• The sorting in observed characteristics is diven
  by the returns to those skills in both countries.
• In countries with low returns to education, it is
  the highly educated who want to migrate, and vice
  versa.

11
2. Assimilation

• Many studies have tried to measure the skill
  differential between natives and immigrants at
  the time of entry.
• Beginning with Chiswick (1978) and Carliner
  (1980).
• And how this differential changes with time in
  the host country.
• Key result
• There is a positive correlation between the
  earnings of immigrants and the number of years
  since immigration.
• There has been a lot of debate over the
  interpretation of this correlation.

12
Early Empirical Studies

• The empirical analysis of the relative economic
  performance of immigrants was initially based on
  the following cross-section regression model
• Cross sectional studies have typically found b1lt0
  and b2gt0.

13
The Identification Problem

• This was interpreted as immigrants accumulating
  human capital relative to natives, thus closing
  the gap over time.
• Assimilation.
• This overtaking was interpreted as a selection
  argument.
• However, Borjas (1985) suggested that b2gt0 could
  instead reflect a decline in the relative skills
  across successive immigrant cohorts.

14
Age and Cohort Effects

• Identification raises difficult methodological
  problems.
• It requires longitudinal data or repeated
  cross-sections.
• So we can track cohorts over time.
• But also restrictions on the parameters, such as
• The time effect being the same for immigrants and
  netives.
• The age coefficient is the same for immigrants
  and natives.

15
What is economic assimilation?

• Coceptual disagreement about the definition of
  assimilation.
• The problem is that any definition of assim. must
  define a base group that immigrants assimilate
  to.
• Many studies equate assimilation with the rate of
  wage convergence between immigrants and natives
  (Chiswick 1978).
• Others (Lalonde Topel, 1992) think that
  assimilation occurs if, between two equivalent
  immigrants, the one with a longer time in H earns
  more.

16
Empirical Evidence

• A large literature summarizes the trends in the
  skills and wages of immigrants in the US.
• They usually combine various Census
  cross-sections to identify the age and cohort
  effects.
• We can illustrate the basic results by estimating
  a basic wage regression for the sample of working
  men 25-34, Census 1980, 1990, 2000.

17
Trend in the Relative Wage of Immigrant Men, US
1980-2000
Unadjusted specifications contain only an
intercept in x. Adjusted specifications also
include education, age, age sq., age cubed, and
region dummies.
18
Regression Results

• The regressions suggest that the relative skills
  of immigrants declined across successive
  immigrant cohorts between 1960 and 1990.
• This interpretation, though, requires assumptions
  on the period effects.
• Ass Changes in aggregate economic conditions did
  not affect the relative wage of immigrants.
• This assumption is probably invalid.

19
Tracking a Cohort Over Time
20
Measuring Assimilation
Also included Education, age2, age3, ysm2, ysm3,
region dummies.
21
3. Effects on the Host Country

• The literature measuring how immigrants affect
  the employment opportunities of natives has grown
  rapidly in the past 15 years.
• A number of difficult conceptual and econometric
  problems plague this literature.
• Essentially, people have followed two methods
• Spatial correlations.
• Natural experiments.

22
3.1 Spatial Correlations

• Economic theory suggests that immigration into a
  closed labor market affects the wage structure
  by
• Raising the wage of complementary workers.
• Lowering the wage of substitutes.
• This method tests this prediction by defining the
  labor market along a geographical dimension.
• Such as different states in the US or
  metropolitan areas.

23
Identification

• The idea is to estimate the spatial correlation
  between labor market outcomes in an area and the
  immigrant density.
• This would identify the impact of immigration on
  native outcomes under the following conditions
• If immigrant flows penetrate geographical labor
  markets randomly, and
• If natives do not respond to these supply shocks
  by moving to other areas.

24
Estimation

• The typical study regresses a measure of native
  labor market outcomes in the region (or the
  change) on the relative quantity of immigrants
  (or the change).
• The regression coefficient is then interpreted as
  the impact of immigration on native wages or
  employment.

25
Results and Problems

• OLS usually leads to results not signif. diff.
  from zero, with sizes that vary a lot by period.
• Borjas et al. 1997, Borjas 1999, Friedberg Hunt
  1995.
• Problems
• 1. Immigrant flows in an area may well be
  endogenous.
• 2. Natives may respond by moving.

26
The Endogeneity Problem

• People have addressed it by using instrumental
  variables.
• Altonji Card (1991) instrument the immigrant
  supply with the fraction of the workforce that is
  foreign born at the beginning of the period.
• Assuming that immigrants are attracted by the
  presence of compatriots.
• The IV results pose the same problems as the OLS
  ones.

27
The Skating Rink Problem

• Harder to address.
• Card DiNardo (2000) find no evidence of
  selective out-migration by natives.
• Other studies find contradicting results,
  depending on the methodology.

28
3.2 Natural Experiments

• Some studies exploit exogenous sources of
  variation in migration flows.
• Large enough.
• Unrelated to changes in labor market conditions.
• Card (1990) is probably the most influential
  paper, but many have followed.
• Hunt (1992), Kugler Yuksel (2006), etc.

29
The Mariel Boatlift Experiment

• In April 1980, Castro declared that Cubans
  wishing to move to the US could leave freely from
  the port of Mariel.
• In 5 months, 125,000 Cubans had left, increasing
  Miamis labor force by 7.
• Card then compares the evolution of unemployment
  rates and wages in Miami with those of similar
  cities that did not experience the immigrant
  shock.
• He finds essentially no significant effects.
• Note that this approach does not solve the
  problem of the out-migration of natives.

30
Problems

• Both the spatial correlations and Cards approach
  have been criticized because geographical areas
  such as MAs cannot be treated as closed labor
  markets.
• GE adjustments or spillovers will dissipate the
  effects.
• Labor or product market integration with the rest
  of the country.
• In integrated economies, local factor supply
  changes have no effect on local prices.

31
3.3 GE Experiments

• Friedberg (2001) uses the increase in immigration
  in Israel as a result of emigration restrictions
  being lifted in the Soviet Union.
• This resulted in a 12 increase in Israels
  population in 4 years.
• But No comparison group.
• Idea Use occupations as units of analysis.
• Assumption limited occupational mobility.

32
Problem

• There may be an endogeneity problem.
• More immigrants enter occupations with increasing
  wages.
• She uses self-reported occupation in the Soviet
  Union as an instrument.
• IV estimates are positive but insignificant.

33
Borjas (2003)

• It would be nice to find a GE experiment where
  the categories are more exogenous.
• So we dont have to instrument for them.
• Immigrants cannot easily change their education.
• Enough categories to get identification?
• Borjas (2003) notes that workers of the same
  education but different age or experience are
  unlikely to be perfect substitutes.

34
EducationExperience Cells

• If immigrants were concentrated not only in
  education groups, but also among experience
  groups within education,
• Then we could have many more data points.
• Indeed, he finds the recent increase in
  immigration was concentrated among hgh school
  dropouts and younger workers.
• Then, he analyzes the impact of immigration on
  native earnings in cells defined by decade,
  education, and 5-year potential experience
  groups.
• He finds significant negative effects.

35
Conclusions

• There is still no consensus on the magnitude and
  significance of the effects of immigration on the
  labor market.
• But the literature has definitely evolved and
  made progress on the topic.
• Many questions are still open.
• The recent increases in international migration
  ensure that interest and research on the topic
  will continue.