The Labour Market Integration of Immigrants in OECD Countries

1
The Labour Market Integration of Immigrants in
OECD Countries

• on-going work for OECD's Working Party 1,
  EPCpresented by Sébastien Jean (OECD)
• Workshop on the Economic Integration of
  Immigrants
• OECD, 29 May 2006

2
Motivation

• Integration is a widely shared priority
• But poor labour market outcomes of immigrants,
  compared to natives

3
Unemployment is higher among immigrants than
among natives.
Uimmig 2 x Unatives
Uimmig Unatives
4
even when computed for low skilled males only
Uimmig 2 x Unatives
Uimmig Unatives
5
Motivation (2)

• Immigrants often fare worse than natives (in
  employment and/or wages), for comparable
  characteristics imperfect labour market
  integration
• How does labour market integration compare across
  countries?
• Are immigrants absorbed into employment? Higher
  risk of inactivity and/or unemployment? Wage
  rebate?
• How is integration linked to policy settings on
  product and labour markets?
• How do structural policies interact with
  integration?

6
Key points

• Sizeable cross-country differences in the degree
  and nature of labour market integration of
  immigrants
• Lesser wage gap tends to be associated with
  higher employment gap
• Employment gap more persistent (?)
• Integration can be related to product and labour
  market policies
• Some policy effects seem to be magnified for
  immigrants
• Immigrants tend to suffer disproportionately from
  labour market dualism

7
Outline

• The approach
• Data and implementation
• Results by country and across-countries
• Preliminary conclusions

8
What can be learned from the literature?

• Mainly focused on wages in the US (Chiswick,
  1978, Borjas, 1985, 1995), but also some evidence
  on employment in Europe (Zimmermann Constant
  eds., 2004, ongoing work in OECD-ELS, 2004-05)
• Is there a difference between immigrants and
  natives, once controlled for observable
  characteristics?
• Main findings
• Immigrants earn less than natives in the US (wage
  diff 20 on average), but they catch up over
  time
• In European countries, immigrants display higher
  risk of being unemployed
• Explanations language, unobserved skills
  (self-selection), social capital, legal
  obstacles, imperfect skills transferability,
  discrimination

9
Our approach

• Analyse differences across comparable immigrants
  and natives in activity rates / employment rates
  / wage rates
• Control for human capital and socioeconomic
  characteristics at the individual level
• Step 1- Carry out similar analysis, country by
  country
• Step 2- Jointly study all countries, and relate
  immigrants-natives differences to policies on the
  product and labour market

10
Estimation Framework

• Step 1 (country by country)
• separately for males and females
• X experience, squared experience, marital
  status, educational attainment
• Immig dummies for migration background (
  EU/nonEU if appl.), /- 10 years since migration)
• Step 2 (across countries, with policy variables)
• Pol product and labour market policies

11
The Data

• Individual data (longitudinal household surveys)
  comparable across countries
• EU15 Countries ECHP data
• standardised annual longitudinal survey, European
  Union, common questionnaire
• 7 waves (from 1994 to 2001).
• US PSID, longitudinal household data (1997-2001)
• Australia HILDA, longitudinal household data
  (2001-2003)
• Canada SLID, longitudinal household data
  (1996-2001, not in this presentation)

12
Implementation

• Immigrants defined by country of birth
• In EU countries, treat separately EU15/ non-EU 15
• In Australia, treat separately anglo-saxon
  countries
• Use nationality for Germany
• Duration of stay in the country key variable
• But data limitations do not allow much inference
  about assimilation
• Separate /- 10 years since migration
• Not possible in the US data
• Correct for non-random sample selection into
  activity and into employment based on observables
  and unobservables (Heckman, 1979)

13
Limitations

• Conceptual
• Impossible to control for all factors of
  cross-country differences integration policies,
  immigration motive, immigrants unobserved skills
  (linked to migration policy, country income and
  inequality, geography, history!)
• Return-migration bias
• Statistical
• Limited sample if immigrants weak
  representativeness, clustered on a short period,
  cohort effects (although longitudinal), nb obs
  insufficient to use language and endogamy
  variables
• Attrition bias under-representation of
  recently-arrived immigrants

14
Estimates by country Employment gap versus wage
gap among recently arrived, active immigrants
15
Employment gap vs. wage gap at least 10 years
after migration
16
Cross-country integration differences and
(product and) labour market policies

• Not the only explanation (integration policies
  etc.)
• But product and labour market policies may help
  explain how immigrants fare because
• Different distribution of individual productivity
• Different behaviour (reservation wage, location
  choices)
• Less social capital
• Discriminated against

17
Cross-country differences and policies
18
Illustrative evidence precariousness and
immigrants

• Holds in a regression context higher risk of
  precarious contract among recently arrived
  immigrants
• (conditional on being employed, when individual
  observable characteristics are controlled for)

19
Preliminary conclusions

• Work still in progress
• Immigrants fragile population need to limit
  perverse effects
• Think of integration in a dynamic setting
  differences, but also rhythm of assimilation
• Role of integration policies, targeted policies