Lost in translation? The relative wages of immigrants in the Portuguese labour market

Abstract

This article examines the wage gaps between immigrant and Portuguese workers using matched employer–employee data for the 2002–2008 period. We found that most of the wage gap is not due to the worst endowments of the immigrants but to differences in the returns to those characteristics and to the immigrant status effect. In particular, immigrants’ education and foreign experience are significantly less valued in the Portuguese labour market. Overall, the wages of immigrants do not fully converge to those of comparable natives as domestic experience increases. The assimilation rates tend to be stronger in the first years after migration and for immigrants with higher levels of foreign experience. Total immigrants are a heterogeneous group of different nationalities, with immigrants from the EU15 and China starring as the two extreme cases.

Keywords:

• immigration
• wage assimilation
• matched employer–employee data
• Gelbach decomposition
• Oaxaca-Blinder decomposition

JEL Classifications:

• F22
• J31
• J61

Acknowledgements

The authors thank Álvaro Novo, Manuel Coutinho Pereira and two anonymous referees for their helpful comments and suggestions. We also thank Lucena Vieira for excellent computational assistance. Any errors and omissions are the sole responsibility of the authors.

Notes

1. By law, workers formally classified as apprentices can receive a minimum wage that is, at least, 80% of the full rate.

2. In the regression analysis of the next sections, real hourly wages are the dependent variable. We also included the real monthly wage in this descriptive analysis as it results in more intuitive values and the conclusions remain unchanged.

3. In what follows, whenever we mention log points throughout the text we are referring to 100 log points.

4. Using the rate of wage convergence between immigrants and natives in the host country as the concept of economic assimilation, this sum corresponds to the linear assimilation factor or assimilation rate (see Borjas (1999) for a discussion of the interpretation of the coefficients when controlling for age and allowing for the estimates to vary between natives and immigrants). In section 5.2, the wage assimilation of immigrants is examined using simulated wage profiles of native and immigrant workers, fully taking into account the effect of the quadratic polynomials.

5. Because this decomposition is additive, one can obtain the contributions for groups of regressors, e.g. sector dummies, as the sum of group-wise components . Furthermore, robust standard errors clustered at the individual level are considered. For more details, see Gelbach (2010).

6. For the sake of simplicity, this discussion ignored the quadratic terms, which, for a reasonable time span, do not affect the main results.

7. All detailed results are available from the authors upon request.