The gender gap in earnings: an international comparison with European matched employer–employee data

Abstract

This article examines the origin of the cross-country heterogeneity of the gender wage gap in nine European countries using a unique harmonized international matched employer–employee dataset. Our novel findings suggest that cross-country differences in the intensity of female segregation into low-paying workplaces is a relevant source of international differences in the size of the gap and that international disparities in the characteristics of wage structures are not very influential. On the other hand, the evidence is not fully conclusive with respect to the influence on the variations of the gap of different macroeconomic, social and institutional country-specific features considered previously in the literature.

Keywords:

• gender wage gap
• matched employer–employee data
• Juhn–Murphy–Pierce decomposition
• segregation

JEL Classification::

• J16
• J30
• J70

Acknowledgements

This study benefits from funding support of the project European Labour Market Analysis using Firm-level Panel Data and Linked Employer–Employee Data financed by the EU under the VI Framework Programme. The author is grateful to Eurostat and the National Statistical Offices of Spain, the Netherlands, Norway, the Czech Republic, Latvia, Slovakia, Portugal, Italy and Lithuania for data access to their national samples of the European Structure of Earnings Survey 2002. In order to fulfil data protection and confidentiality restrictions, the research was conducted through online remote access (LISSY System) to the microdata. The author thanks Tanvi Desai for excellent technical support with the LISSY remote system and David Marsden, Francis Kramarz, Alex Bryson and other participants in the project for their helpful comments.

Notes

¹ These are the countries whose national statistics offices allowed researchers to access the data of their national samples of the ESES in the framework of a European research project (European Labour Market Analysis using Firm-level Panel Data and Linked Employer–Employee Data) financed by the European Union under the VI Framework Programme.

² The national samples of Norway, Latvia, Lithuania, the Netherlands and Slovakia cover workplaces with less than 10 employees. In order to work with similar types of workplaces in all the countries, they have been dropped from the samples.

³ Given that workplace specific effects also capture unobserved individual effects common to all employees in a workplace and that it is not possible to identify this effect in ESES cross-section microdata, they are relegated to the residual. Existing evidence for several countries suggests that unobserved individual effects tend in general to be weakly correlated with workplace specific effects (Abowd et al., 2001; Lane, 2009).

⁴ A recent assessment of the merits and shortcomings of the Juhn et al. (1991) decomposition can be found in Yun (2009). According to this author, the decomposition relies on a few bold assumptions that are difficult to verify.

⁵ If inter-country differences in residual inequality were interpreted as differences in the market premium for unobserved skills, this component would represent the effect of differences in the returns to unobservable skills. However, this is quite a strong interpretation of residual inequality as it may also capture, inter alia, the effect of measurement error, equation misspecification and heterogeneity in unmeasured characteristics.

⁶ These gaps should be exponentiated in order to express the estimates as percentage mark-ups.

⁷ This information may be accessed on the website of the European Commission (http://ec.europa.eu).

⁸ Although detailed information on its calculation can be found in Table 5, we should note that family conciliation policies are proxied through the recreation of the Organization for Economic Co-operation and Development (OECD) work–family reconciliation index (a summary measure of the policies prevailing in a country on work–family issues) carried out by Christofides et al. (2010) for all the EU members. Alternatively, gender role attitudes have been measured, following Fortin (2005), with the information of the 1999 to 2001 wave of the World Value Survey. In particular, perceptions of the man as main breadwinner, as well as anti-egalitarian views or discriminatory attitudes against working women are captured through the agreement to the statement ‘When jobs are scarce, men should have more right to a job than women’ and perceptions with respect to women's traditional role is measured as agreement to the statement ‘Being a housewife is just as fulfilling as working for pay’.

⁹ Interestingly, this same result holds for all the EU countries according to ESES 2002 data (European Commission, 2006).

¹⁰ A thorough analysis of international differences in wage inequality and the influence of wage-setting institutions in European economies using the ESES data can be found in Simón (2010).