The effect of unemployment on social participation of spouses: evidence from plant closures in Germany


This paper estimates the effect of an individual’s unemployment on the level of social participation of their spouse. Using German panel data, it is shown that unemployment has a strong negative effect on public social activities of both directly and indirectly affected spouses. Private social activities of either spouse, however, are only found to increase if the indirectly affected spouse is not working. Conflict prevention strategies or habituation may help to rationalise this finding. Our results imply that active labour market policies should account for spillover effects within couples and adopt a family perspective.

This is a preview of subscription content, log in to check access.


  1. 1.

    Note, however, that evidence on the relationship between unemployment and health outcomes turns out to be mixed. Salm (2009) and Schmitz (2011), for example, do not find any causal effect.

  2. 2.

    See also Clark (2003), Bubonya et al. (2014) and Mendolia (2014) for similar analyses using different data sources.

  3. 3.

    Classical studies about unemployment conducted in the early 1930s, long before modern welfare states have been installed, mirror both reduced social activities and increased tension within the families. Komarovsky (2004), for instance, summarises her observations from a study in a large industrial city close to New York as follows: “The unemployed man and his wife have no social life outside the family. The extent of social isolation of the family is truly striking. This refers not only to formal club affiliations, but also to informal social life. [...] Family after family gave the same story of meagre social contacts.”—(Komarovsky 2004, p. 122). Also, in the seminal Marienthal study a woman was observed reporting “I often quarrel with my husband because he does not care about a thing any longer and is never at home. Before unemployment it was not so bad because the factory provided a distraction.”—(Jahoda et al. 1974, p. 85). However, these studies are largely qualitative or cross-sectional quantitative with only a few observations and do not control for other variables.

  4. 4.

    However, it may not be completely exogenous to an individual due to anticipation effects resulting in a gradual leaving process of some workers prior to closing (Kassenböhmer and Haisken-DeNew 2009, p. 460). The presence of such a mechanism would imply an underestimation of the treatment effect [see also the discussion in Kunze and Suppa (2017)].

  5. 5.

    In Sect. 3.2.3 we provide (indirect) evidence in order to support the identifying assumption.

  6. 6.

    As a consequence, the difference in the treatment dummy equals the treatment dummy in the second period, i.e. \(\Delta D = D\).

  7. 7.

    This approach resembles the Horvitz and Thompson (1952) estimator, which uses inverse sampling probability weights to account for complex survey design or missing data problems. In fact, treatment evaluation is often conceived as a missing data problem, as the counterfactual outcome is never observed.

  8. 8.

    Other advantages are its versatility, as weights can be passed to almost any estimator and its computational efficiency.

  9. 9.

    The data used in this paper (SOEP v30, were extracted using the add-on package PanelWhiz for Stata. PanelWhiz ( was written by Dr. John P. Haisken-DeNew ( See Haisken-DeNew and Hahn (2010) for details. The PanelWhiz-generated DO file to retrieve the data used here is available from us upon request. Any data or computational errors in this paper are of our own.

  10. 10.

    According to the eigenvalue criterion, the factor analysis suggests two underlying factors, in which the items culture, cinema and volunteer do only load on the first factor, whereas socialising and helping only load on the second factor. See also Bauernschuster et al. (2014) for a similar aggregation procedure.

  11. 11.

    Note that we only use the responses to these questions when they are recorded on a 4-point scale (ranging from “weekly” and “monthly” to “less frequently” and “never”).

  12. 12.

    We have checked that none of these restrictions changes qualitatively our results. The intuition for excluding couples from the control group in case of an employer change is that social participation patterns of these couples may differ from those of the rest, for example due to differing time constraints as a result of a time-consuming new position or due to a new social environment in case the job change requires moving to another city.

  13. 13.

    In particular, in the first model potentially treated individuals have to be employed in the pre-treatment period, whereas their spouses may have any kind of labour force status, whereas in the second model the partner of the potentially treated individual must be employed in the pre-treatment period and the indirectly affected individual may or may not be employed.

  14. 14.

    Note that we do not consider the period 2001–2005 as four years are not comparable to the remaining periods. Our qualitative results, however, would be very similar if we added this period to the analysis. Similarly, dropping the observations from the period 1996–1997 would not change much.

  15. 15.

    More specifically, we experimented with a gender dummy interaction of the treatment effect. However, the results from these estimations were not entirely clear-cut, which may be due to the small number of observations, small gender-specific effects (if existent at all), or the fact that gender roles and gender-specific behaviour are in a state of flux. Tentative results suggest a slightly stronger reduction in public social activities for both spouses if the wife loses her job. Likewise, the increase in private social activities of the indirectly affected partner appears to be somewhat larger if the husband loses its job. These results are available upon request.

  16. 16.

    Note, however, that there is some evidence that employed indirectly affected partners reduce their public social participation more than non-working ones. While neither the treated coefficient, nor its interaction turns out to be significant, the sum of these coefficients is well significantly different from zero at conventional levels of significance.

  17. 17.

    Note, however, that Nikolova and Ayhan (2018) question the importance of income for the spouse’s life satisfaction, whereas Kunze and Suppa (2017) cast doubt on the importance of income for public social activities.

  18. 18.

    These calculations are based on the respective standard deviation for each outcome variable in the whole sample (not shown), which, however, approximately equals the standard deviations reported in Table 2 for individuals of the control group.

  19. 19.

    These results are available upon request.

  20. 20.

    In the early intervention studies, for instance, it is common to test dozens or even hundreds of outcomes.

  21. 21.

    Note that this conclusion is robust to alternative adjustment procedures. (The results from these procedures are available upon request.)

  22. 22.

    One explanation might be a lack of variation in the dependent variable.

  23. 23.

    See, for example, Marcus (2013) for a related discussion on health outcomes.

  24. 24.

    Reasons for unemployment entries have already received some attention in previous research, see, for example, Winkelmann and Winkelmann (1998) or Kassenböhmer and Haisken-DeNew (2009).

  25. 25.

    Results from a more detailed analysis for different reasons of unemployment are available upon request.


  1. Alesina A, La Ferrara E (2000) Participation in heterogeneous communities. Q J Econ 115(3):847–904

    Article  Google Scholar 

  2. Anderberg D, Rainer H, Wadsworth J, Wilson T (2016) Unemployment and domestic violence: theory and evidence. Econ J 126(597):1947–1979

    Article  Google Scholar 

  3. Bauernschuster S, Falck O, Woessmann L (2014) Surfing alone? The internet and social capital: evidence from an unforeseeable technological mistake. J Public Econ 117(C):73–89

    Article  Google Scholar 

  4. Bubonya M, Cobb-Clark DA, Wooden M (2014) A family affair: job loss and the mental health of spouses and adolescents, Melbourne Institute Working Paper Series 23/14, Melbourne Institute of Applied Economic and Social Research, The University of Melbourne

  5. Clark AE (2003) Unemployment as a social norm: psychological evidence from panel data. J Labor Econ 21(2):323–351

    Article  Google Scholar 

  6. Doiron D, Mendolia S (2012) The impact of job loss on family dissolution. J Popul Econ 25:367–398

    Article  Google Scholar 

  7. Hainmueller J (2012) Entropy balancing for causal effects: a multivariate reweighting method to produce balanced samples in observational studies. Political Anal 20:25–46

    Article  Google Scholar 

  8. Haisken-DeNew JP, Hahn M (2010) Panelwhiz: efficient data extraction of complex panel data sets—an example using the german SOEP. Schmollers Jahrbuch 130(4):643–654

    Article  Google Scholar 

  9. Heckman JJ, Ichimura H, Todd PE (1997) Matching as an econometric evaluation estimator: evidence from evaluating a job training programme. Rev Econ Stud 24:605–654

    Article  Google Scholar 

  10. Horvitz DG, Thompson DJ (1952) A generalization of sampling without replacement from a finite universe. J Am Stat Assoc 47(260):663–685

    Article  Google Scholar 

  11. Jahoda M (1982) Employment and unemployment: a social-psychological analysis. The psychology of social issues. Cambridge University Press, Cambridge

    Google Scholar 

  12. Jahoda M, Lazarsfeld PF, Zeisel H (1974) Marienthal: the sociography of an unemployed community. Social science paperbacks. Transaction Publishers, Piscataway

    Google Scholar 

  13. Kassenböhmer SC, Haisken-DeNew JP (2009) You’re fired! The causal negative effect of entry unemployment on life satisfaction. Econ J 119:448–462

    Article  Google Scholar 

  14. Komarovsky MA (2004) The unemployed man and his family: the effect of unemployment upon the status of the man in fifty-nine families, classics in gender studies. AltaMira Press, Lanham

    Google Scholar 

  15. Kunze L, Suppa N (2017) Bowling alone or bowling at all? The effect of unemployment on social participation. J Econ Behav Organ 133:213–235

    Article  Google Scholar 

  16. Marcus J (2013) The effect of unemployment on the mental health of spouses—evidence from plant closures in Germany. J Health Econ 32(3):546–558

    Article  Google Scholar 

  17. Mendolia S (2014) The impact of husband’s job loss on partners’ mental health. Rev Econ Househ 12(2):277–294

    Article  Google Scholar 

  18. Nikolova M, Ayhan SH (2018) Your spouse is fired! How much do you care? J Popul Econ.

    Article  Google Scholar 

  19. Putnam R (2001) Bowling alone. Simon & Schuster, New York

    Google Scholar 

  20. Salm M (2009) Does job loss cause ill health? Health Econ 18:1075–1089

    Article  Google Scholar 

  21. Schmitz H (2011) Why are the unemployed in worse health? The causal effect of unemployment on health. Labour Econ 18(1):71–78

    Article  Google Scholar 

  22. Sen AK (2000) Social exclusion: concept, application, and scrutiny, social development papers, vol 1. ADB, Manila

    Google Scholar 

  23. Stiglitz JE, Sen AK, Fitoussi JP (2009) Report by the commission on the measurement of economic performance and social progress. Technical report, commission on the Measurement of Economic Performance and Social Progress

  24. Sullivan D, von Wachter T (2009) Job displacement and mortality: an analysis using administrative data. Q J Econ 124:1265–1306

    Article  Google Scholar 

  25. Wagner GG, Frick JR, Schupp J (2007) The German socio-economic panel study (soep): scope, evolution and enhancements. Schmollers Jahrbuch 127(1):139–169

    Google Scholar 

  26. Westfall PH, Young SS (1993) Resampling-based multiple testing: examples and methods for p-value adjustment. Wiley, New York

    Google Scholar 

  27. Winkelmann L, Winkelmann R (1998) Why are the unemployed so unhappy? Evidence from panel data. Economica 65(257):1–15

    Article  Google Scholar 

  28. Wooldridge JM (2010) Econometric analysis of cross section and panel data. MIT Press, Cambridge

    Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Lars Kunze.

Additional information

We are grateful to Sebastian Garmann and Wolfram F. Richter for helpful comments and suggestions. Any remaining errors are ours.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Kunze, L., Suppa, N. The effect of unemployment on social participation of spouses: evidence from plant closures in Germany. Empir Econ 58, 815–833 (2020).

Download citation


  • Unemployment
  • Social participation
  • Plant closure
  • Entropy balancing
  • SOEP

JEL Classification

  • J64
  • I31