## _wp1606

## Source details

**Canonical URL:** [_wp1606](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp1606.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp1606.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp1606.pdf.json)

---

### Introduction and scope
- Immigration to Germany rose after earlier waves in the 1960s and 1990s, increased again since the Global Financial Crisis, and "skyrocketed in 2015, as over one million asylum seekers arrived, nearly twice as many as during previous heights in the 1990s."
- Germany is described as "currently the second most popular migration destination in the world."
- Study uses micro-data from the German Socio-Economic Panel (GSOEP) to estimate determinants of wages, unemployment, and labor force participation for natives and immigrants.
- A newly highlighted driver of wage differences is immigrants’ difficulty in securing jobs with higher "occupational autonomy" (e.g., managerial responsibilities), contributing to skill downgrading.

### Key empirical findings on wages
- Employed immigrants earn "20 percent less than employed natives with similar characteristics after arrival."
- Immigrants’ wages initially catch up "by one percentage point per year," but the process slows and wages never fully converge.
- Immigrants without German writing skills or a German degree have an initial wage gap "as high as 30 percent."
- "Good German writing skills close the gap by 10 percentage points and a German degree by another 5 percentage points."
- Immigrants born in advanced countries have a gap "a third of that of other immigrants."
- "Half of the lower wages of immigrants results from lower levels of job autonomy (given other individual characteristics)."
- Empirical magnitudes and robustness:
  - Baseline Mincer regressions: being an immigrant reduces the wage by 7.3 percent (conditional on experience, education, and controls).
  - Immigrants face an initial wage gap of 21 percent upon arrival; catch-up of roughly 1 percentage point per year; assimilation slows and stops before full convergence.
  - Including occupational autonomy raises R-squared from 35.4 percent to 39.4 percent and halves returns to education and the immigrant wage gap in part.
  - When occupational autonomy is included, direct effects of good German writing skills and a German degree become insignificant, suggesting these operate via access to higher-autonomy jobs.
  - Wage-sample: 224,000 observations; pooled panel 1984–2013.

- Selected exact statistics (2013 snapshot and distributions):
  - Average hourly real wage (2013): natives 15.0 Euros per hour; immigrants 12.7 Euros per hour.
  - Unconditional wage gap: 15.3 percent.
  - Moving up one level of occupational autonomy associated with average wage increase of 20 percent (first-glance) and 13 percentage points in regressions.
  - Raw occupational autonomy means: immigrants 2.2 versus natives 2.8 (job autonomy scale).

### Occupational autonomy and skill downgrading
- Occupational autonomy measure: GSOEP-specific 1–5 scale; one-level step correlated with substantial wage gains.
- Conditional on the kind of job (occupational autonomy) there is no remaining wage gap for many categories—immigrants have different occupational autonomy distribution than natives.
- Distributional findings:
  - Immigrants are on average 30 percent more likely than natives to have a job with low autonomy (conditional on included variables).
  - After 20 years the average marginal effect from being an immigrant is still 23 percent for low-autonomy probability.
  - Immigrants are 9 percentage points less likely to have high-autonomy jobs.
- Positive factors for higher autonomy (average marginal effects):
  - Being born in an advanced country: reduces low-autonomy probability by 6.3 percentage points; increases high-autonomy probability by 1.9 percentage points.
  - Good German writing skills: reduces low-autonomy probability by 12 percentage points; increases high-autonomy probability by 3.6 percentage points.
  - German degree: reduces low-autonomy probability by 10 percentage points; increases high-autonomy probability by 3.2 percentage points.
- Education effects on autonomy (average marginal effects for high education):
  - High education reduces low-autonomy probability by 34 percentage points and increases medium-high autonomy by 30 percentage points and high autonomy by 10 percentage points.
- Sectoral and match findings:
  - For highly educated immigrants: German degree raises probability of a job requiring high education by 0.119 (11.9 percentage points); born in an advanced country raises it by 0.135 (13.5 percentage points); good German writing skills raise it by 0.0935 (9.35 percentage points).
  - Industries with particularly poor outcomes for high-skilled immigrants: Manufacturing and Trade.
  - For medium educated immigrants: years in Germany increase probability of at least medium autonomy by 0.0115 (1.15 percentage points); German degree raises it by 0.145 (14.5 percentage points); good German writing skills raise it by 0.0528 (5.28 percentage points).
  - Worst-performing industries for medium educated: Manufacturing, Transport, Mining, and Construction; Banking and Insurances substantially better relative to Services.

### Unemployment and labor force participation findings
- Initial unemployment and long-run gap:
  - Probability of being unemployed is initially "7 percentage points higher for recently arrived immigrants than for natives of similar characteristics."
  - Over time the unemployment gap narrows but "in the long-run the probability remains 3 percentage points higher."
- Empirical unemployment magnitudes:
  - An immigrant upon arrival has an additional 6.8 percentage points probability of being unemployed relative to a native (estimated coefficient 0.068*** and 0.069*** across specifications).
  - With accumulation of experience the immigrant unemployment effect declines but remains permanently positive and above 3 percentage points.
  - An immigrant from an advanced country with a German degree and who writes German well has the smallest unemployment penalty at 1.5 percentage points; not being born in an advanced country increases the penalty to 3.5 percentage points.
- Participation dynamics:
  - Participation initially lower for immigrants: an immigrant who just arrived is 18 percentage points less likely to participate in the labor market.
  - Convergence over time:
    - After 15 years the difference is down to less than 5 percentage points.
    - After 20 years it is below 1 percentage point.
    - After 22 years it is zero — participation fully converges after 20–22 years.
  - Medium education increases participation by 3.7 percentage points; high education increases participation by 9.7 percentage points.
- Experience and unemployment effects:
  - Every year of experience reduces probability of being unemployed by 0.1 percentage points.
  - Every year of unemployment lowers wages by 3.5 percentage points; differences in time spent unemployed explain less than 10 percent of immigrants’ wage gap.

### Heterogeneity, cohort, and credential effects
- Recent cohorts perform worse:
  - Those who arrived after 2007 have an additional wage gap of 17 percentage points relative to natives.
  - Immigrants arriving after 2007 perform particularly badly in autonomy outcomes; one out of four highly educated recent immigrants has a job with low or low-medium autonomy (versus one out of 14 highly educated natives).
- Language and credential effects:
  - Good German writing skills close wage gap by 10.0 percentage points.
  - German degree closes additional portion of the gap (interaction effects reported across regressions; specific interaction reported as 5.4 percentage points in the excerpt and additional robustness entries).
  - Immigrants who write German well and hold a German degree still need to spend 25 years in Germany to earn the same wage as a comparable native.
- Returns to education:
  - Baseline: medium education ≈ 11 percent wage premium; high education ≈ 39–40 percent.
  - Returns decline substantially when occupational autonomy and match variables are included.
  - Over-education raises wages for natives but not for immigrants; different returns driven by those without a job match.

### Data, empirical strategy, and robustness
- Data source: German Socio-Economic Panel (GSOEP), pooled 1984–2013 with expanded 2013 immigrant sample; original pooled sample 502,000 observations; working wage-sample 224,000 observations; full-sample 350,000 observations after restrictions.
- Key dependent variables: ln(hourly real wage), occupational autonomy (ordered), unemployment dummy, participation dummy.
- Key controls: region, married, full-time, age, years in firm, experience and experience squared, year and industry dummies; immigrant-specific interactions for years since immigration, German degree, born in advanced country, German writing skills.
- Estimation methods: OLS for wage, ordered logit for autonomy, logit for unemployment/participation; standard errors clustered at individual level; all models estimated using individual sample weights.
- Robustness checks include: no sample weights, probit instead of logit, Heckman selection correction, random-group variance estimator. Main substantive results robust to these checks though some significance levels change.

### Policy implications and conclusions
- Immigrants contribute substantially to the economy but face more labor market obstacles than native workers; obstacles are overcome only gradually and never fully for wages and unemployment.
- Highly skilled immigrants experience sizable skill downgrading and persistent wage gaps; addressing downgrading and reducing obstacles to skill transfer is essential to attract and retain highly-skilled immigrants.
- The current large wave of refugees may face bigger obstacles than past immigrants:
  - Preliminary evidence suggests lower education and qualification levels, more likely to come from non-advanced countries, lack German degrees, and have poor German writing skills.
  - Conditional on characteristics, these refugees are less likely to participate and more likely to be unemployed than other immigrants; when employed initially they earn 30 percent less than otherwise similar natives.
- Policy trade-offs: the introduction of the minimum wage may lower the wage gap of new immigrants but could further increase their likelihood of being unemployed (Battisti and Felbermayr, 2015).
- Successful labor market integration of recent newcomers will likely require time, particular efforts, and decisive policy action.

### Appendix — selected summary statistics and model diagnostics (preserved exact figures)
- Representative 2013 snapshot (working population):
  - Natives: Age mean 43.6; Female mean 0.48; Hourly Real Wage mean 15.0 EUR; Medium Education mean 0.57; High Education mean 0.33; Working Experience mean 16.7; Unemployment Experience mean 0.64; Trained for Job mean 0.59; High Edu. Job Match mean 0.22; Autonomy mean 2.75.
  - Immigrants (2013): Age mean 42.4; Female mean 0.48; Hourly Real Wage mean 12.7 EUR; Medium Education mean 0.51; High Education mean 0.26; Working Experience mean 17.1; Unemployment Experience mean 1.09; Trained for Job mean 0.41; High Edu. Job Match mean 0.14; Autonomy mean 2.20; Years in Germany mean 22.0; German Oral Ability Good or Very Good mean 0.78; German Writing Literacy Good or Very Good mean 0.66; German Degree mean 0.39.
- Appendix Mincer robustness selected coefficients (specification notes preserved):
  - Experience: 0.019*** (0.00081); Experience Squared: -0.00041*** (0.000017).
  - Medium Education: 0.045*** (0.0079).
  - High Education: 0.10*** (0.011).
  - HE Job Match: 0.22*** (0.010).
  - Unemployment Exp.: -0.024*** (0.0019).
  - Autonomy (Job Type): 0.11*** (0.0030).
  - Immigrant (indicator): -0.12*** (0.030) | -0.076*** (0.019) (across specifications).
  - Years in Germany: 0.0068*** (0.0024); YiG Squared: -0.00014*** (0.000053).
  - Good German Writing: 0.051*** (0.018).
  - Born in Adv. Country: 0.052*** (0.015).
  - Migrated after 2007: -0.17*** (0.057).
  - Observations (wage equations): 224,272; R-squared: 0.407 and 0.434 in robustness specifications.

*Source: _wp1606 — German Socio-Economic Panel analyses and references as provided in the source content.*

### REFERENCES .............................................................................................................

### _wp1606 - REFERENCES

### Introduction and scope
- Immigration to Germany rose after earlier waves in the 1960s and 1990s, increased again since the Global Financial Crisis, and "skyrocketed in 2015, as over one million asylum seekers arrived, nearly twice as many as during previous heights in the 1990s."
- Germany is described as "currently the second most popular migration destination in the world."
- The study uses micro-data from the German Socio-Economic Panel to estimate determinants of wages, unemployment, and labor force participation for natives and immigrants.
- A newly highlighted driver of wage differences is immigrants’ difficulty in securing jobs with higher "occupational autonomy" (e.g., managerial responsibilities), which contributes to skill downgrading.

### Key empirical findings on wages
- Employed immigrants earn "20 percent less than employed natives with similar characteristics after arrival."
- Immigrants’ wages initially catch up "by one percentage point per year," but the process slows and wages never fully converge.
- Immigrants without German writing skills or a German degree have an initial wage gap "as high as 30 percent."
- "Good German writing skills close the gap by 10 percentage points and a German degree by another 5 percentage points."
- Immigrants born in advanced countries have a gap "a third of that of other immigrants."
- "Half of the lower wages of immigrants results from lower levels of job autonomy (given other individual characteristics)."

### Unemployment and labor force participation findings
- The probability of being unemployed is initially "7 percentage points higher for recently arrived immigrants than for natives of similar characteristics."
- Over time the unemployment gap narrows but "in the long-run the probability remains 3 percentage points higher."
- German language skills and a German degree reduce the unemployment gap.
- The participation rate of immigrants is initially lower but "converges fully after 20 years spent in Germany."

### Data, analysis structure, and sections
- Data source: German Socio-Economic Panel with an expanded immigrant sample.
- Analysis includes descriptive statistics and empirical models estimating wages, unemployment, and participation; occupational autonomy is analyzed as a separate factor.
- Paper structure (sections indicated): II — overview of immigration; III — literature review; IV — data; V — descriptive statistics; VI — empirical model; VII — wages; VIII — occupational autonomy; IX — unemployment and participation; X — robustness checks.
- Figures and tables listed cover topics such as: Fraction of Foreign Born Population, Movements between Germany and the Rest of the World, Net Migration from and to the European Union, Total Permanent Inflows, Conditional Wage Gaps (2013), Downgrading of Immigrants’ Education, Assimilation Path of Immigrant Wages, Average Marginal Effects (education, occupational autonomy, unemployment, participation), and Mincer-type Wage Regression outputs.
- Appendix tables (A1–A9) include summary statistics, educational distribution, downgrading distributions, conditional wages, cross-correlations, and robustness of Mincer-type wage equations.

*Source: _wp1606 - REFERENCES*

### Section XI concludes.

### Section XI concludes.

### II.  IMMIGRATION TO GERMANY—AN OVERVIEW
- In 2013, over 10 million foreign-born persons lived in Germany, about 13 percent of the population.
- 15 million Germans have had at least one foreign-born parent (Gathmann et al., 2014), yielding a fraction of persons with a migration background—as defined by having at least one foreign-born parent—“as high as in Canada.”
- Official migration statistics used: German Federal Statistical Office (Destatis). These figures cover everyone who stays in the country for more than two months and hence include intra-company training, language study periods, seasonal jobs, and asylum applications.
- Historically, only a fraction of incoming immigrants remain permanently: in the past, one third of all immigrants left the country already within one year.
- Example: of all Indians and Chinese nationals who arrived in 2006, less than one quarter was still in Germany in 2012 (Gathmann et al., 2014), despite many arriving with working permits.
- Permanent migration statistics from the OECD are used when discussing reasons for migration and long-term settlement.

### A. Inflows and Outflows from 1950 to 2015
- 1950s: total movements very low; net immigration basically zero.
- Recruitment agreements beginning in 1955 with Turkey and some Southern European countries due to acute labor shortages; starting in 1960 Germany experienced intended net immigration (guest workers), many of whom stayed.
- 1973: recruitment ban on guest workers after oil crisis, with exceptions for highly skilled and seasonal workers; net inflows fluctuated around zero until the late 1980s.
- Early 1990s: second large migration wave after fall of the iron curtain; annual net immigration of more than 750,000 individuals in the early 1990s.
- Late 1990s–2000s: net immigration decreased slowly, turned negative right before the global financial crisis in 2008.
- Post-2008: inflows increasing much faster than outflows and net immigration in 2014 was above 550,000, i.e. 0.6 percent of the population.
- 2015: net immigration “skyrocketed.” Battisti and Felbermayr (2015) expect a net inflow of registered immigrants between 1.4 and 1.5 million. Net immigration could hence reach 2 percent of the population, which would constitute a new record.
- Note: From 1950 to 1990 only Western Germany is considered in the Destatis series.

### B. Recent Inflows: Who are the New Immigrants?
- Post-crisis surge up to 2013 mainly consisted of movements within the EU; movements within the EU surpassed a quarter of a million people in 2012 and net immigration from the EU surpassed 300,000 in 2014.
- Net immigration from within the EU started increasing significantly in 2010 and 2011 due to the financial crisis and the lifting of restrictions on immigration from countries that entered the EU in 2004 (free movement from 2004 EU-enlargement countries).
- Inflows from Bulgaria and Romania (joined EU in 2007) picked up; restrictions were fully lifted only from 2014 onwards.
- Divergence/diversion effect: Bertoli, Brücker, and Moraga (2013) argue that nearly 80 percent of the surge in immigration to Germany after the Global Financial Crisis resulted from diversion effects (attractiveness of alternative destinations controlled for).
- Since 2010, positive and rising net immigration from outside the EU: 150,000 in 2013 and 250,000 in 2014.
  - Reflects increase in asylum applicants: 25,000 in 2008; 130,000 in 2013; over 200,000 in 2014 (over 30 percent of all EU applications).
  - August 2015 federal government estimate: up to 800,000 new asylum applicants (BMI, 2015a). End of October 2015: more than 750,000 refugees registered (BMI, 2015b).
- 2015 prospective flows and EU outlook:
  - Battisti and Felbermayr (2015) expect similar inflows from within the EU as in 2014 and 1.1 million refugees from outside the EU for 2015.
  - European Commission expects an additional inflow of 3 million refugees to the European Union until 2017, implying strong inflows to Germany likely to persist.
- Composition shifts:
  - 2010–2013: immigrants from the EU accounted for roughly two thirds of total net immigration each year.
  - 2014: net immigration from within the EU and from outside the EU were nearly equally strong.
  - 2015: net immigration from outside the EU was much higher than from within the EU.
- Immigrant characteristics are important determinants of labor market integration experiences.

### C. Reasons to Migrate
- Only a fraction of movements are directly employment-related.
  - 2008 special EU Labour Force Survey: only half of all movements within the EU were triggered by employment reasons (OECD, 2013).
  - 2013: residence status of only 12 percent of incoming non-EU immigrants was linked to employment.
- Permanent labor immigration flows from outside the EU/EFTA relative to population are very low compared to other OECD countries (OECD, 2013). Footnote: those who did come were mostly highly-skilled men, mainly from Asia (India and China), and 90 percent moved to Western Germany.
- Other motives: studying, family unification, and an increasingly important motive—application for asylum.
- Immigrants enter under heterogeneous rules and regulations; labor market access varies by category (OECD, 2013).
- Many migrants intend to stay temporarily or have temporary permission; many return after short periods.
- Permanent immigrant inflows (OECD definition): regulated movements considered as settling, enabling cross-country comparisons.

### Permanent Migration Statistics (OECD comparisons)
- Figure 4 (permanent inflows, thousands): 2007–2010 Australia, Canada, and Germany had around 200,000 permanent inflows; from 2010 onward permanent inflows remained constant in Australia and Canada but doubled in Germany.
- U.S. attracts highest permanent inflows with over one million immigrants (comparison year not specified in the excerpt).
- Figure 5 (index, 2007=100): immigration increased in Denmark, Austria, the Netherlands, and Sweden but less than in Germany; Italy and Spain experienced decreasing inflows throughout the period 2007–2013.

### III.  LITERATURE REVIEW — Key Findings Relevant for Analysis
- Classic finding (Chiswick, 1978): immigrants in the U.S. start with much lower wages than natives with similar characteristics and then slowly catch up.
- Adsera and Chiswick (2007): similar convergence pattern for 15 European countries.
- Initial wage penalty often interpreted as devaluation of human capital obtained abroad (Chiswick and Miller, 2009). Over time immigrants invest in human capital, find better jobs, and employers learn about productivity.
- Evidence that immigrants may overtake natives after 10 to 15 years in U.S. and Canada (Chiswick, 1978; Meng, 1987).
- In the U.S.: initially higher unemployment and lower participation for immigrants but fast assimilation (Chiswick, Cohen, and Sachs, 1997).
- In Germany: mixed evidence on wage convergence; no clear evidence of convergence for workers with similar characteristics, suggesting more difficult labor market integration than in U.S. or Canada.
  - Papers finding catching-up: Constant and Massey (2005); Gundel and Peters (2007); Büchel and Frick (2004); Lehmer and Ludsteck (2011, 2014).
  - Papers finding lack of catching-up: Pischke (1992); Dustmann (1993); Schmidt (1997); Bauer et al. (2005).
- Unemployment: Kogan (2005) shows immigrants in Germany are more likely to be unemployed; higher unemployment risk only partially related to lower human capital, also due to unfavorable labor market allocation.
- Skill downgrading debate:
  - Definition 1: immigrants do jobs requiring less qualification than in home country (Akresh, 2008).
  - Definition 2: immigrants perform worse than natives in destination given characteristics (Dustmann et al., 2013 show UK immigrants conditional on education work in jobs with lower socio-economic classification).
- Language skills: German language proficiency reduces immigrant wage penalty (Dustmann and van Soest 2001, 2002; Aldashev, Gernandt, and Thomsen 2009; Zibrowius 2012).

### IV.  DATA
A. The German Socio-Economic Panel (GSOEP)
- Representative household survey maintained at the German Institute for Economic Research (DIW) in Berlin.
- Survey started in 1984 with around 12,000 participants; sample doubled over time to 24,000 individuals in the most recent wave in 2013.
- Panel organized at household level; all participants aged at least 16 are personally interviewed.
- Largest survey of immigrants in Germany and oversamples immigrant population.
- 2013: additional immigrant sample of 5000 persons added; new immigrants sampled using the 2011 census to account for compositional changes since 1995.
- This paper is the first (to the author’s knowledge) to use the 2013 data in an analysis of immigrant labor market performance.

B. Dependent and Control Variables (included in analysis)
- Labor Force Status: working, unemployed, or not part of the labor force.
- Hourly Real Wages: monthly gross and net nominal wages transformed into hourly gross and real wages using actual working hours per week; baseline uses gross wages; net wages also analyzed. Footnote on measurement error issues for hourly wages and self-reported income (Müller, 2009; Falck et al., 2013).
- Education: three categories via ISCED-1997 classification: ISCED-1 and ISCED-2 = “low” (benchmark); ISCED-3 and ISCED-4 = “medium”; ISCED-5 and ISCED-6 = “high”.
- Region: dummy for persons living in Eastern Germany.
- Working and Unemployment Experience: years of full-time work experience and years of unemployment experience.
- Years Spent in Firm: time with current employer in years.
- Job that Requires High Qualification: dummy = 1 if job requires college/university degree, 0 otherwise.
- Autonomy Status of the Job: GSOEP-specific measure of occupational autonomy on a 1–5 scale; correlated with wage—one-level step associated with average wage increase of 20 percent. Categories range from unspecialized manual labor (low) to managers/freelance academics/self-employed with many employees (high). See Hoffmeyer-Zlotnik and Geis (2003).
- Industry: ten categories — Agriculture, Energy, Mining, Manufacturing, Construction, Trade, Transport, Bank and Insurances, Services, Others.
- Being Trained for Job: responses “yes”, “no”, and “no training” (omitted category: not working in an occupation one is trained for).
- Immigrant: dummy = 1 for persons not born in Germany; 0 for persons born in Germany. Footnote: Büchel and Frick (2004) confirm that foreigners born in Germany are generally very similar to natives.
- Additional controls: Age, Gender, Married status, and for workers whether Full-Time Job.
- Immigrant-specific variables:
  - Years Since Immigration: survey year minus year of arrival; includes squared term for non-linear effects.
  - German Degree: dichotomous variable = 1 if immigrant obtained a degree in Germany, 0 otherwise.
  - Born in Advanced Country: grouping based on IMF classification (advanced vs non-advanced).
  - German Writing Skills: dummy = 1 if immigrant has good or very good German writing skills; many missing observations—third category for missing values included.

C. Sample and Data Modifications
- Original pooled sample: 35 vintages (1984–2013) with 502,000 observations.
- Restriction to persons aged 16–70 years reduces sample by 70,000 observations; observations lost only for years person not in age group.
- Additional drops: unknown place of birth (2,000); immigrant with unknown year of arrival (3,000); cannot identify where degree obtained (10,000); in education, retired, or in maternity leave (54,000); unknown education (3,000); individual sample weight zero (7,000). Resulting sample: 350,000 observations.
- For wage analysis further drops:
  - Non-working persons (79,000);
  - Required job training unknown (7,000);
  - Worked less than 10 hours a month, did not report positive wages, or real wages cannot be computed (together 28,000);
  - Time spent in firm and occupational autonomy unknown (together 1,000);
  - Occupational autonomy below one (12,000). Resulting wage-sample: 224,000 observations.
- Wage transformation and trimming:
  - Monthly nominal gross earnings transformed into hourly real wages.
  - Wages below the 5 percent and above the 99 percent percentile replaced with the respective percentile value in that year.
- For experience, training, industry, and language proficiency: dummies for missing values added.
- Descriptive 2013 snapshot (selected statistics):
  - Natives average age: 44.9 years; immigrants average age: 43.7 years.
  - Fraction of women: natives 50 percent; immigrants 51 percent.
  - More natives had medium and high education than immigrants overall; but immigrants who migrated after 2007 had a considerably higher fraction of highly educated than the native population.
  - Unemployment rate: natives 6 percent; immigrants 10 percent.
  - Participation rate: only slightly lower among immigrants than natives.
  - Language and education in 2013: three out of four immigrants able to speak good or very good German; nearly two thirds had good or very good German writing skills.
  - A little over a third of immigrants obtained an educational degree in Germany and on average they had been in Germany already for (text truncated in source).

*Source: _wp1606 - Section XI concludes.*

### 22.5 years.

### _wp1606 - 22.5 years.

### Working population and time unemployed
- Average working experience: 22.5 years.
- Immigrants spent on average more time unemployed: 1.1 years versus 0.6 years.
- Share with a job for which they were trained: immigrants 41 percent versus natives 59 percent.
- Share of less highly skilled immigrants in a job that required high education: immigrants 14 percent versus natives 22 percent.
- Occupational autonomy (mean): immigrants 2.2 versus natives 2.8.

### Raw wage gaps (2013)
- Average hourly real wage (2013): natives 15.0 Euros per hour; immigrants 12.7 Euros per hour.
- Unconditional wage gap: 15.3 percent.
- Raw wage gap is larger for women than for men (table indicates greater gap but specific broken values limited in excerpt).
- Educational distribution notes:
  - For immigrants arriving after 2007: 55.7 percent of those from advanced countries had high education and 28.5 percent of all others had high education.
  - For natives: fraction with high education 22.1 percent.
- Educational distribution of immigrants is more dispersed and has a higher fraction of very low educated than natives.

### Conditional wage gaps and occupational autonomy (first-glance results)
- Even conditional on education, natives earn more than immigrants; gap for very low education is not significant due to very few natives with such low education.
- Wage gaps vanish when conditioning on occupational autonomy:
  - For two out of five autonomy categories immigrants earn more than natives.
  - For one category the gap is basically zero.
  - A significant positive gap only for the medium-high autonomy category (possibly related to German apprenticeship system).
- Comparison implies:
  - Given the kind of job (occupational autonomy), there is no remaining wage gap.
  - Immigrants with the same education do not have jobs of the same occupational autonomy as natives.
- Skill downgrading interpretation:
  - Figure 7 shows natives’ autonomy distribution curve shifted to the right across low, medium, high education panels: higher fraction of natives with medium-high and high autonomy and lower fraction with low and medium-low autonomy.
  - Moving up one level of occupational autonomy increases the wage on average by 20 percent.
  - Therefore, skill downgrading of immigrants can potentially explain part of the wage gap.
- Possible non-discriminatory explanations for downgrading include:
  - Lack of language skills limiting responsibility and management duties.
  - Different quality of educational degrees obtained abroad versus in Germany.
- Significance notes (from footnote):
  - Gaps for education 2 and 5 significant at the 1 percent level.
  - Gap for education 4 significant at the 5 percent level.
  - Gap for education 3 significant at the 10 percent level.
  - Gap for occupational autonomy 4 significant at the 5 percent level.

### Distributional notes for immigrant arrival cohorts
- Table A5 (Appendix) summary: immigrants arriving before 1990 are less disadvantaged than those arriving after 2007, but significant differences remain.
- Recent immigrants (arrived after 2007) perform particularly badly:
  - In this group, one out of four highly educated immigrants has a job with low or low-medium autonomy.
  - For natives, one out of 14 highly educated natives has a job with low or low-medium autonomy.

### Empirical model (specification and estimation)
- General model form used across equations:
  - x_it beta = alpha + gamma Z_it beta + delta W_it beta + controls + error (structure shown in original notation: ݔ௜௧ ߚ = ଵ ᇱ ܧ௜௧ ߚ + ଶ ᇱ ܯ௜௧ ߚ + ଷ ᇱ ܥ௜௧ ߝ + ௜௧)
  - x_it is either ln(hourly real wage), autonomy level, or a dummy for unemployment/participation.
  - Z_it: explanatory variables common to immigrants and natives.
  - W_it: immigrant-specific variables estimated as interaction terms.
  - Controls included (coefficients not reported).
- Estimation methods:
  - Depending on specification: OLS, logit, or ordered logit.
  - Standard errors clustered at individual level to account for repeated observations.
  - All models estimated using individual sample weights. Robustness checks discussed later include: no sample weights, probit instead of logit, Heckman selectivity correction, random-group variance estimator.
- Sample size for wage equations: models estimated with all 224,000 observations.
- Cross-correlations reported in Table A7 and Table A8 (Appendix).

### Mincer-type wage equations — baseline results
- Dependent variable: natural logarithm of hourly real wage before taxes (coefficients interpreted as marginal percent effects).
- Controls always included: region, married status, full-time status, age, years in the firm, year dummies, industry dummies.
- Baseline regression (years of experience, experience squared, medium and high education dummies, immigrant dummy):
  - Wages increase with experience; one additional year raises wages by close to 2 percent at the beginning, but returns diminish with experience and after 20 years the additional benefit is basically zero.
  - Medium education: earn 11 percent more than low education.
  - High education: earn 40 percent more than low education.
  - Female wage gap: 12 percent.
  - Being an immigrant reduces the wage by 7.3 percent (conditional on experience, education, and controls).
  - R-squared of baseline: 34 percent.

### Differentiating among immigrants — assimilation and credential effects
- Immigrant heterogeneity noted: years in Germany range 0 to 60; language skills vary; country of origin and whether immigrant obtained German degree vary.
- Assimilation estimation:
  - Years since arrival and squared term included.
  - Immigrants face an initial wage gap of 21 percent upon arrival.
  - Immigrants catch up by roughly 1 percentage point every year, but assimilation slows and eventually stops before full convergence.
- German language skills and German degrees:
  - Interaction terms included for immigrants with good or very good German writing skills and for those with a German degree to test additional benefit from schooling in Germany beyond language skills.
  - With these interactions:
    - The wage gap increases to over 30 percent for those without any beneficial characteristics.
    - Good German writing skills close the gap by 10.0 percentage points.
    - Having a German degree decreases the gap by an additional (text truncated — interaction effect reported but numeric value not available in excerpt).

*Source: GSOEP 2013, own calculations (excerpted from _wp1606 - 22.5 years.)*

### 5.4 percentage points.

### _wp1606 - 5.4 percentage points.

### Key findings on immigrant-native wage gaps
- Immigrants who write German well and hold a German degree still need to spend 25 years in Germany to earn the same wage as a comparable native.
- Those who arrived after 2007 have an additional wage gap of 17 percentage points relative to natives.
- Including occupational autonomy raises the regression R-squared from 35.4 percent to 39.4 percent.
- When occupational autonomy is included, the returns to education and the immigrant wage gap are halved in part, and the direct effects of good German writing skills and a German degree become insignificant (suggesting these operate via access to higher-autonomy jobs).
- The wage gap for female immigrants is 50 percent larger than the gap for native women.

### Channels behind the wage gap
- Unemployment experience:
  - Every year of unemployment lowers wages by 3.5 percentage points.
  - Differences in time spent in unemployment explain less than 10 percent of the immigrants’ wage gap.
- Job match / training:
  - Being trained for the job raises wages, but inclusion of this variable explains less than 10 percent of the immigrant wage gap.
- Occupational autonomy:
  - Going up one level in the autonomy scale increases the wage on average by 13 percentage points.
  - Some part of the returns to education (about half) reflects that better educated workers obtain jobs with higher autonomy.
  - Some of the immigrants’ wage gap stems from immigrants doing jobs with less autonomy even when observable characteristics are identical.

### Different returns to education
- Medium education yields about 0.11 (11 percent) higher wages in baseline specifications; returns decline when autonomy and other controls are added.
- High education yields about 0.39–0.40 (39–40 percent) higher wages in baseline specifications; returns fall substantially when adding autonomy and match variables.
- A native with a high education job match earns 22 percent more than one without a match.
- Over-education raises wages for natives but not for immigrants — i.e., different returns to high education are driven by those without a job match.

### Determinants of occupational autonomy (ordered logit average marginal effects)
- Immigrant status:
  - Immigrants are on average 30 percent more likely than natives to have a job with low autonomy (conditional on included variables).
  - After 20 years the average marginal effect from being an immigrant is still 23 percent for low-autonomy probability.
  - Immigrants are 9 percentage points less likely to have high-autonomy jobs.
- Positive factors reducing low-autonomy probability and raising high-autonomy probability:
  - Being born in an advanced country: reduces low-autonomy probability by 6.3 percentage points; increases high-autonomy probability by 1.9 percentage points.
  - Good German writing skills: reduces low-autonomy probability by 12 percentage points; increases high-autonomy probability by 3.6 percentage points.
  - German degree: reduces low-autonomy probability by 10 percentage points; increases high-autonomy probability by 3.2 percentage points.
- Education effects (average marginal effects for high education):
  - High education: reduces low-autonomy probability by 34 percentage points; reduces low-medium by 22 percentage points; increases medium by 15 percentage points; increases medium-high by 30 percentage points; increases high by 10 percentage points.
- Notably, immigrants from advanced countries perform better than other immigrants on autonomy outcomes, but being an immigrant nearly offsets the advantages from high education.

### Matches, industries, and heterogeneity by skill level
- For highly educated immigrants:
  - Probability of having a job that requires high education (match): Degree from Germany raises probability by 0.119 (11.9 percentage points); being born in an advanced country raises probability by 0.135 (13.5 percentage points); good German writing skills raise probability by 0.0935 (9.35 percentage points).
  - Experience has a negative marginal effect on match and on high autonomy for highly skilled immigrants.
  - Industries with particularly poor outcomes (over-education / downgrading): Manufacturing and Trade for high-skilled immigrants.
- For medium educated immigrants:
  - Years in Germany increase probability of having at least medium autonomy: 0.0115 (1.15 percentage points).
  - Degree from Germany raises probability of at least medium autonomy by 0.145 (14.5 percentage points).
  - Good German writing skills raise probability of at least medium autonomy by 0.0528 (5.28 percentage points).
  - Industries with worst performance: Manufacturing, Transport, Mining (10 percent significance), and Construction.
  - Banking and Insurances: medium educated immigrants in this sector are over 30 percentage points more likely than those in Services to have at least medium autonomy.

### Unemployment and participation
- Unemployment margins:
  - Every year of experience reduces the probability of being unemployed by 0.1 percentage points.
  - Medium education reduces unemployment probability by 1.3 percentage points.
  - High education reduces unemployment probability by 6.2 percentage points.
  - Being a woman reduces the likelihood of being unemployed by 1.1 percentage points (conditional on controls).
- Immigrant unemployment penalty:
  - An immigrant upon arrival has an additional 6.8 percentage points probability of being unemployed relative to a native (to be read against an average native unemployment rate lower than 6.8 percent over the sample).
  - With accumulation of experience the immigrant unemployment effect declines but remains permanently positive and above 3 percentage points.
  - An immigrant from an advanced country with a German degree and who writes German well has the smallest unemployment penalty at 1.5 percentage points; not being born in an advanced country increases the penalty to 3.5 percentage points.

*Source: _wp1606 - 5.4 percentage points.*

### 5.7 percentage points, and not having a German degree to 6.9 percentage points.

### _wp1606 - 5.7 percentage points, and not having a German degree to 6.9 percentage points.

### Effects of education, origin, and German qualifications
- Medium education increases participation by 3.7 percentage points; high education increases participation by 9.7 percentage points.
- Education effects on unemployment: reported differences include "5.7 percentage points" and "6.9 percentage points" as referenced in the source text.
- Immigrants benefit substantially less from their education than natives; interaction terms indicate smaller returns for immigrants with medium and high education (table reports positive interaction coefficients but smaller magnitudes than for natives).
- A German degree and good German writing skills reduce the probability of unemployment in the unemployment regressions, but neither a German degree nor German writing skills affect the decision to participate.

### Unemployment and participation magnitudes (selected coefficients and findings)
- Experience: participation effect reported as 0.014*** (per year) with Experience Squared = -0.00026*** (nonlinear diminishing returns).
- Unemployment regressions show Experience coefficients as -0.0011*** and -0.0013*** and Experience Squared as 7.7e-06 and 0.000014*** (table columns preserved).
- Immigrant status: increases unemployment probability by 0.068*** and 0.069*** in two specifications; lowers participation probability by -0.18*** and -0.15*** in two specifications.
- Born in advanced country: coefficient reported as -0.020*** in unemployment and 0.038*** in participation (table).
- Years in Germany: participation coefficient 0.014*** and 0.013***; nonlinear term YiG Squared reported as 0.000039*** and 0.000020** (unemployment) and -0.00024*** (participation).
- Medium Education: unemployment coefficients -0.013*** and -0.018***; participation 0.037*** and 0.034***.
- High Education: unemployment coefficients -0.062*** and -0.073***; participation 0.097*** and 0.092***.
- Immigrant with Medium Education interaction reported as 0.0095** and 0.0081 (participation table entry).
- Immigrant with High Education interaction reported as 0.049*** and 0.020* (unemployment/participation entries as in table).
- Female: reduces unemployment slightly (-0.011***,-0.011***) and lowers participation by -0.094*** (both columns).
- Good German Writing Skills: unemployment coefficient -0.022***; German Degree: unemployment coefficient -0.012*** (participation coefficients reported as non-significant in table).
- Unemployment Rate for Immigrant interaction: reported as -0.00024 (unemployment) and -0.0034*** (participation).

- Observations in reported estimations: 351,010 (each column).
- Controls included: region, age, and year dummies. Robust standard errors in parentheses. Significance notation: *** p<0.01, ** p<0.05, * p<0.1.

### Assimilation dynamics and participation convergence
- An immigrant who just arrived is 18 percentage points less likely to participate in the labor market but catches up over time.
- Convergence milestones reported:
  - After 15 years the difference is down to less than 5 percentage points.
  - After 20 years it is below 1 percentage point.
  - After 22 years it is zero.
- Participation fully converges to natives’ rates, in contrast to wages and unemployment where full convergence is not observed.
- The participation gap for immigrants born in advanced countries is 25 percent smaller than that of other immigrants.

### Robustness checks and estimation concerns
- Re-estimations without individual sample weights: wage gap for immigrants becomes somewhat smaller; returns to education for immigrants fall; effects of being born in an advanced country and migrating after 2007 become smaller; other coefficients change minimally.
- Random group variance estimator (Wolters, 2007) used as alternative for GSOEP survey structure: standard errors generally similar; catching-up process significance falls from highly significant (clustered SEs) to significance at the 10 percent level with random group variance estimator. With fewer interaction terms, linear term significant at 2 percent and quadratic term at 7 percent.
- Heckman correction for selection bias applied (first-stage participation, second-stage wage): results indicate no selection bias; therefore these results are not reported in detail.
- Probit vs logit specifications for models in sections VIII and IX: probit regressions yield somewhat lower likelihoods but nearly identical substantive results; logit reported as preferred.

### Policy implications and conclusions
- Immigrants contribute substantially to the economy but face more labor market obstacles than native workers; obstacles are overcome only gradually and never fully (for wages and unemployment).
- Highly skilled immigrants experience sizable skill downgrading and relatively large and persistent wage gaps in Germany; to attract highly-skilled immigrants it is essential to understand causes of downgrading and reduce obstacles to skill transfer.
- The current large wave of refugees may face bigger obstacles than the average past immigrant:
  - Preliminary evidence points to lower education and qualification levels for this group.
  - They are more likely to come from non-advanced countries, lack German degrees, and have poor German writing skills.
  - Conditional on characteristics, these immigrants are less likely to participate and more likely to be unemployed than other immigrants.
  - When employed initially, they earn 30 percent less than otherwise similar natives; gaps decrease slowly and immigrants remain more likely to be unemployed throughout.
- The introduction of the minimum wage may lower the wage gap of new immigrants but could further increase their likelihood of being unemployed (Battisti and Felbermayr, 2015).
- Successful labor market integration of recent newcomers will likely require time, particular efforts, and decisive policy action.

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp1606.pdf*

### REFERENCES

### _wp1606 - REFERENCES

### References and literature scope
- Bibliography includes empirical and theoretical studies on immigration, labor market outcomes, assimilation, language skills, wage assimilation, over-education, fiscal effects, and measurement sources.
- Notable recurring authors and works: Adsera & Chiswick (2007); Akresh (2008); Aldashev, Gernandt & Thomsen (2008, 2009); Borjas (1999); Chiswick (1978); Chiswick & Miller (2008, 2009); Dustmann (1993, 2000), Dustmann & Frattini (2014); Ottaviano & Peri (2012); OECD (2013, 2015); SOEP Group (2001).
- Data source referenced repeatedly: GSOEP, 1984-2013.

### Appendix — Summary Statistics of Full Sample (1984–2013 and 2013)
- Full-sample composition:
  - Natives: 286,377 observations (1984–2013); 11,887 observations (2013).
  - Immigrants: 55,320 observations (1984–2013; 16 percent of sample); 4,182 observations (2013; 26 percent of sample).
- Key demographic and labor-force means (1984–2013):
  - Natives: Age mean 42.6 (Median 43); Female mean 0.49; Medium Education mean 0.57; High Education mean 0.24; Working mean 0.77; Unemployed mean 0.06; Not Part of the Labor Force mean 0.16.
  - Immigrants: Age mean 43.1 (Median 43); Female mean 0.49; Medium Education mean 0.46; High Education mean 0.17; Working mean 0.69; Unemployed mean 0.10; Not part of the Labor Force mean 0.21; Years in Germany mean 20.2 (Median 19); German Oral Ability Good or Very Good* mean 0.71; German Writing Literacy Good or Very Good* mean 0.55; German Degree mean 0.36.
  - *From those we have data. We have 8,663 observations.
- Key demographic and labor-force means (2013):
  - Natives (2013): Age mean 44.9 (Median 47); Female mean 0.50; Medium Education mean 0.59; High Education mean 0.29; Working mean 0.84; Unemployed mean 0.056; Not Part of the Labor Force mean 0.11.
  - Immigrants (2013): Age mean 43.7 (Median 43); Female mean 0.51; Medium Education mean 0.49; High Education mean 0.23; Working mean 0.77; Unemployed mean 0.10; Not part of the Labor Force mean 0.13; Years in Germany mean 22.5 (Median 21); German Oral Ability Good or Very Good** mean 0.74; German Writing Literacy Good or Very Good** mean 0.62; German Degree mean 0.36.
  - **From those we have data. We have 3,388 observations.
- Note: For dichotomous variables 0 means “no” and 1 means “yes”.

### Appendix — Summary Statistics: Working Population (1984–2013 and 2013)
- Working population counts:
  - Natives (1984–2013): 200,388 observations; Immigrants (1984–2013): 35,472 observations (15 percent of sample).
  - Natives (2013): 9,104 observations; Immigrants (2013): 2,942 observations (24 percent of working population).
- Representative means (1984–2013, Working Population):
  - Natives: Age mean 40.5; Female mean 0.43; Hourly Real Wage in EUR mean 14.9 (Min 2.81; Max 85.0); Medium Education mean 0.57; High Education mean 0.28; Working Experience mean 15.6; Unemployment Experience mean 0.40; Trained for Job mean 0.57; Not Trained for Job mean 0.33; High Edu. Job Match mean 0.16; Autonomy (Job Type) mean 2.64 (scale 0 to 5).
  - Immigrants: Age mean 41.5; Female mean 0.42; Hourly Real Wage in EUR mean 13.6 (Min 2.81; Max 85.0); Medium Education mean 0.50; High Education mean 0.19; Working Experience mean 16.3; Unemployment Experience mean 0.74; Trained for Job mean 0.37; Not Trained for Job mean 0.40; High Edu. Job Match mean 0.083; Autonomy (Job Type) mean 1.92; Years in Germany mean 20.0; German Oral Ability Good or Very Good** mean 0.77; German Writing Literacy Good or Very Good** mean 0.61; German Degree mean 0.38.
  - *From those we have data. We have 198,085 observations for natives and 33,259 for immigrants. **From those we have data. We have 5777 observations.
- Representative means (2013, Working Population):
  - Natives: Age mean 43.6; Female mean 0.48; Hourly Real Wage in EUR mean 15.0 (Min 3.87; Max 48.3); Medium Education mean 0.57; High Education mean 0.33; Working Experience mean 16.7; Unemployment Experience mean 0.64; Trained for Job mean 0.59; High Edu. Job Match mean 0.22; Autonomy (Job Type) mean 2.75.
  - Immigrants (2013): Age mean 42.4; Female mean 0.48; Hourly Real Wage in EUR mean 12.7 (Min 3.93; Max 48.3); Medium Education mean 0.51; High Education mean 0.26; Working Experience mean 17.1; Unemployment Experience mean 1.09; Trained for Job mean 0.41; High Edu. Job Match mean 0.14; Autonomy (Job Type) mean 2.20; Years in Germany mean 22.0; German Oral Ability Good or Very Good** mean 0.78; German Writing Literacy Good or Very Good** mean 0.66; German Degree mean 0.39.
  - *From those we have data. We have 8,311 observations for natives and 769 for immigrants. **From those we have data. We have 2,401 observations.
- Note: The job autonomy is measured on a scale from 0 to 5. For dichotomous variables 0 means “no” and 1 means “yes”.

### Appendix — Educational Distribution, Occupational Autonomy, and Conditional Wages (selected figures)
- Education (ISCED) distribution — Natives vs Immigrants (percent) and mean wages (EUR):
  - ISCED 1: Natives Percent 0.89, Wage in EUR 10.65 (s.e. 2.50); Immigrants Percent 4.09, Wage in EUR 9.85 (s.e. 0.40)
  - ISCED 2: Natives Percent 6.89, Wage in EUR 12.13 (s.e. 0.30); Immigrants Percent 18.34, Wage in EUR 10.37 (s.e. 0.33)
  - ISCED 3: Natives Percent 49.18, Wage in EUR 13.02 (s.e. 0.15); Immigrants Percent 36.42, Wage in EUR 12.03 (s.e. 0.35)
  - ISCED 4: Natives Percent 8.35, Wage in EUR 14.67 (s.e. 0.31); Immigrants Percent 14.66, Wage in EUR 13.17 (s.e. 0.69)
  - ISCED 5+6: Natives Percent 34.70, Wage in EUR 20.16 (s.e. 0.13); Immigrants Percent 26.49, Wage in EUR 16.79 (s.e. 0.57)
- Occupational Autonomy distribution and mean wages:
  - Autonomy 1: Natives Percent 9.67, Wage in EUR 9.76 (s.e. 0.19); Immigrants Percent 30.96, Wage in EUR 10.06 (s.e. 0.23)
  - Autonomy 2: Natives Percent 24.95, Wage in EUR 11.33 (s.e. 0.08); Immigrants Percent 28.40, Wage in EUR 11.25 (s.e. 0.32)
  - Autonomy 3: Natives Percent 37.82, Wage in EUR 15.25 (s.e. 0.15); Immigrants Percent 25.13, Wage in EUR 14.14 (s.e. 0.55)
  - Autonomy 4: Natives Percent 23.83, Wage in EUR 21.42 (s.e. 0.31); Immigrants Percent 14.77, Wage in EUR 20.27 (s.e. 0.80)
  - Autonomy 5: Natives Percent 3.74, Wage in EUR 24.53 (s.e. 0.64); Immigrants Percent 0.74, Wage in EUR 27.92 (s.e. 2.04)
- Selected autonomy/downgrading snapshot (percent distribution across Autonomy levels for education groups, presented as table excerpts):
  - High Education (All): Autonomy Low 1.12; Low-Medium 6.12; Medium 32.39; Medium-High 50.99; High 9.37.
  - Medium Education (All): Autonomy Low 7.57; Low-Medium 14.01; Medium 34.83; Medium-High 41.40; High 2.20.
  - Low Education (All): Autonomy Low 3.29; Low-Medium 8.77; Medium 39.80; Medium-High 46.01; High 2.14.

### Appendix — Cross-Correlations (selected correlations and significance)
- Immigrants’ variables (pairwise correlations; significance indicated by stars):
  - Years in Germany with Born in EU: 0.138***.
  - Speaks German Well with Writes German Well: 0.848***.
  - German Degree with Years in Germany: 0.259***.
  - High Education with Speaks German Well: 0.112***.
  - Autonomy (Job Type) with High Edu. Job Match: 0.509***.
  - Experience with Years in Germany: 0.380***.
- Cross-correlations including Hourly Real Wage (H. R. Wage) and characteristics:
  - H. R. Wage with Age: 0.210***.
  - H. R. Wage with Experience: 0.221***.
  - H. R. Wage with Education: 0.354***.
  - H. R. Wage with Female: -0.219***.
  - Immigrant indicator with H. R. Wage: -0.0731***.
  - Autonomy with H. R. Wage: 0.457***.
  - High Edu. Job Match with H. R. Wage: 0.399***.
  - Trained for Job with H. R. Wage: 0.201***.

### Appendix — Mincer-type Wage Equations: Robustness (selected coefficients)
- Regression specifications reported: Baseline (6), No Random Group Variance Est. (7), Weights (8).
- Common significant coefficients across columns (coefficient (standard error)):
  - Experience: 0.019*** (0.00081) | 0.019*** (0.00062) | 0.019*** (0.00081).
  - Experience Squared: -0.00041*** (0.000017) | -0.00041*** (0.000014) | -0.00041*** (-0.00002).
  - Medium Education: 0.045*** (0.0079) | 0.052*** (0.0063) | 0.045*** (0.0093).
  - Immigrant with ME: -0.022 (0.014) | -0.049*** (0.010) | -0.022 (0.012).
  - High Education: 0.10*** (0.011) | 0.10*** (0.0086) | 0.10*** (0.014).
  - Immigrant with HE: -0.085*** (0.024) | -0.11*** (0.019) | -0.085*** (0.023).
  - HE Job Match: 0.22*** (0.010) | 0.23*** (0.0080) | 0.22*** (0.015).
  - Immigrant with HE JM: 0.060* (0.033) | 0.061** (0.026) | 0.060** (0.023).
  - Female: -0.11*** (0.0061) | -0.12*** (0.0046) | -0.11*** (0.0052).
  - Unemployment Exp.: -0.024*** (0.0019) | -0.025*** (0.0015) | -0.024*** (0.0020).
  - Trained for Job: 0.021*** (0.0048) | 0.022*** (0.0039) | 0.021*** (0.0066).
  - Autonomy (Job Type): 0.11*** (0.0030) | 0.11*** (0.0024) | 0.11*** (0.0050).
  - Immigrant (indicator): -0.12*** (0.030) | -0.076*** (0.019) | -0.12*** (0.031).
  - Years in Germany: 0.0068*** (0.0024) | 0.0070*** (0.0015) | 0.0068* (0.0034).
  - YiG Squared: -0.00014*** (0.000053) | -0.00013*** (0.000034) | -0.00014* (0.000070).
  - Good German Writing: 0.051*** (0.018) | 0.042*** (0.012) | 0.051** (0.019).
  - German Degree: 0.0060 (0.013) | -0.00057 (0.0094) | 0.0060 (0.013).
  - Born in Adv. Country: 0.052*** (0.015) | 0.020** (0.0090) | 0.052*** (0.0071).
  - Migrated after 2007: -0.17*** (0.057) | -0.080*** (0.030) | -0.17** (0.050).
- Model diagnostics:
  - Observations: 224,272 (all specifications).
  - R-squared: 0.407 (Baseline and Weights) and 0.434 (No Random Group Variance Est.).
  - Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1.
  - Controls: region, married, working full time, the age, the number of years in the firm, as well as year and industry dummies.

*Source: GSOEP, 1984-2013; references and appendix tables as provided in _wp1606 - REFERENCES.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp1606.pdf_