Unequal effects on working time: immigrants’ vulnerability in the German labor market in the early COVID-19 pandemic

ABSTRACT

Do economic shocks increase labor market inequalities between immigrants and natives? The COVID-19 crisis reduced economic activity for almost all social groups, providing a recent case for answering this question. Research tends to focus on employment levels, overlooking potential inequalities in other job characteristics. Workers in Germany have largely kept their jobs, although their working hours were reduced. Using German high-frequency survey data, we analyze whether there was a difference in the reduction of hours for immigrants and immigrants’ descendants (IAD) compared to natives. Since IAD are overrepresented in both heavily affected and essential jobs, we argue that the effects may be heterogeneous across the distribution of the change in hours. As merely comparing averages would ignore this heterogeneity, we employ OLS and quantile treatment effect estimations to analyze working hours changes in the early COVID-19 crisis. Results show that IAD reduced hours more than natives. This effect is particularly pronounced at the lower end of the distribution of the change in working hours. Our findings suggest that IAD experienced economic hardship more often than natives and corroborate earlier findings of increased immigrant-native inequalities in times of crisis. This calls for further investigation of policies aimed at protecting vulnerable groups.

KEYWORDS:

• Economic shocks
• COVID-19
• employment
• labor market inequality
• migration
• quantile treatment effects

Acknowledgements

We wrote this article using data from the IAB High-Frequency Online Personal Panel (IAB-HOPP). Data access was provided by the Research Data Centre (FDZ) of the German Federal Employment Agency (BA) at the Institute for Employment Research (IAB). The data are now available as SUF (Volkert et al. 2021b; DOI: 10.5164/IAB.HOPP_W01-W07.de.en.v2) and SUF with administrative data included (Bellmann et al. 2021b; DOI: 10.5164/IAB.HOPP-ADIAB7519.de.en.v1). We particularly thank Bernhard Christoph for his support. Moreover, we would like to thank the editorial team and the four anonymous reviewers for their helpful suggestions and advice.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 For an overview, see Annex Table 3.B.1 in OECD (2020b, 163).

2 In the US context, studies tend to take a racial rather than a migratory perspective. Couch et al. (2020) and Montenovo et al. (2020) use the concepts of race and ethnicity. So do Mongey et al. (2021), but add citizenship and foreign country of birth in their analyses. Borjas and Cassidy (2020) also apply the concept of being foreign-born. In the UK context, Hu (2020) distinguishes between four groups: white versus ethnic minority immigrants and white versus ethnic minority natives.

3 The analyses use the first version of the IAB-HOPP. The data are now available as Scientific Use File (Volkert et al. 2021b).

4 A Scientific Use File for IAB-HOPP data combined with registry/administrative data is now available (Bellmann et al. 2021b).

5 As the questionnaire used a filter variable to retrieve this information, consolidated information was used. Only those who reported to currently have the opportunity to work from home were able to state their pre-COVID-19 home office use. Hence, it was assumed that those who did not even have the option during the crisis should also not have had it before.

6 Unfortunately, the website of the Federal Ministry of Labour and Social Affairs (BMAS) listing all essential occupations to which Giesing and Hofbauer (2020) point is no longer available. Therefore, the list provided by the authors was used.

7 As Koebe et al. (2020: 2) who also use the KldB2010 three-digit code to define essential occupations note, a definition based on said classification cannot be exact, as occupational groups might be essential in one context but non-essential in another. The authors illustrate this using the example of cleaning personnel in hospitals versus other cleaning jobs.

8 Here using data from 2018 or earlier (for data documentation on the SUF file of the IAB-HOPP data linked to registry/administrative data, see Bellmann et al. (2021a)).

9 -ivqte- command in Stata (Frölich and Melly 2008; Frölich and Melly 2010).

10 As most respondents reported integer values as their working hours, the distribution of changes in working hours is not continuous but has jumps in between values. This can lead to estimation problems as it makes observations’ ranks indistinguishable from one another (cf. Firpo 2007). This warrants the addition of random noise to slightly smooth the distribution (Machado and Santos Silva 2005). Additional checks demonstrate that the obtained estimates are robust to various amounts of added noise.

11 A Heckman model to check for selection effects (see Section 4.3) reveal that this overrepresentation has no influence on subsequent analyses.

12 Heckman models were estimated using both the default maximum likelihood estimator and the two-step estimation procedure.

13 The calibration is based on employment status, age, gender, federal state as well as last registered job, occupational position, and sector (Volkert et al. 2021a: 31ff.).

14 Considering the differentiation between immigrants and immigrants’ descendants, people who provided information that their parents were born abroad but not on their own birthplace are conservatively categorized as descendants of immigrants.

Additional information

Funding

The authors are grateful for funding received from the Federal Ministry of Labour and Social Affairs (BMAS) for ‘Promotion of research and teaching in the field of social policy’ (Grant Reference MPR.00.00024. 20).