Skip to main content

The response of Japanese wives’ labor supply to husbands’ job loss

Abstract

This paper examines how Japanese wives react to their husbands’ involuntary job loss and tests the existence of complementarity of a wife’s labor supply to her husband’s. Utilizing panel data on Japanese households from 1993 to 2004, we found that wives’ labor supply is stimulated when husbands suffer involuntary job loss. The detailed statistics show that not only do working wives raise their labor hours but also nonworking wives begin to participate in the labor market. The added worker effect is evident during the period of job insecurity in Japan following the mid-1990s.

This is a preview of subscription content, access via your institution.

Fig. 1

Notes

  1. Exceptionally, Low (2005) and Pijoan-Mas (2006) describe an individual’s life-cycle labor supply behavior with precautionary motive, and shows that one can change labor supply flexibly in response to his/her uncertainty as well as wages. Attanasio et al. (2005) simulate the changes in consumption, savings and wives’ labor supply in relation to income uncertainty, and find that female labor supply is responsive to idiosyncratic shocks especially in those households with borrowing constraints.

  2. If we also take home production into consideration, the wife’s reservation wage for labor supply could fall according to the husband’s unemployment and his increased time for home production. The realization of the wife’s lower reservation wage raises her labor supply in the market.

  3. Unless consumption and the wife’s leisure are additively separable in the utility function, changes in a wife’s leisure are influenced by the substitutability or complementarity between her leisure and consumption. We assume an additive separability and do not treat households’ simultaneous decisions on consumption and leisure explicitly. As mentioned later, we cannot find any significant effect of consumption changes on wives’ leisure changes, even if we include consumption changes as an explanatory variable. Further consideration of simultaneous decisions between consumption and leisure remains for future research.

  4. We attempted to analyze the levels (length) of wives’ labor hours and found the same implication for the existence of the AWE as shown in the present paper.

  5. Wives’ labor hours actually have negative time dependencies in our sample.

  6. In order to control for differences in risks surrounding households, we estimated the model including income or consumption variances over the past 4 years within a household. Inclusion of them does not alter the implications of the following results at all. We also controlled for consumption changes as an endogenous explanatory variable, and the coefficient on consumption changes was not statistically significant while the exogeneity was accepted.

  7. The young sample may not be representative of the entire society. However, we cannot conclude this sample makes our empirical results on the AWE either overestimated or underestimated. It underestimates the AWE because older wives do not have a time constraint associated with child rearing, meaning that they can add labor supply more flexibly, while it overestimates the AWE because females in older generations have a tendency not to participate in the labor market.

  8. We do not divide a year into shorter time periods, since it decreases the number of households who experienced the husband’s unexpected job loss. We also do not utilize the information on unemployed periods, because the estimation results using such a small sample may be seriously affected by the existence of outliers. Our estimated sample, which does not make use of unemployed periods, includes households facing both temporary and long-lasting unemployment shocks. Thus, we cannot discuss the differential in the AWE between seriousness of husband’s job loss depending on his unemployment period.

  9. Excluding unpaid overtime working hours does not mean that unpaid working hours have no information on wives’ reaction to the husbands’ job loss. For example, working unpaid overtime leads to future income growth. However, unpaid working hours often increase for firm-specific reasons such as wife’s promotion, good sales in her firm, and so on. Because firms’ conditions are not completely observable, they could be omitted variables in explaining working hour changes, which gives us a biased estimator of the AWE.

  10. The present paper includes non-workers together with workers, since we do not want to lose the information on changes from non-workers to workers. We cannot conduct the empirical estimation for non-workers because of insufficient observations. We will discuss the difference between non-workers and workers later at the end of Section 4.

  11. Breusch–Pagan tests reject the assumption of zero variance of the stochastic individual effects, and F tests accept the assumption of no individual effects. Therefore, we attempted pooled OLS estimation with clustering robust standard errors. The results are the same as the ones shown in Table 3. The coefficient on husbands’ job loss is 3.31 with a standard error of 1.51 when excluding husbands’ past job losses, 3.38 with a standard error of 1.52 when including husbands’ 1- or 2-year lagged job loss, and 3.22 with a standard error of 1.51 when including husbands’ 3-year lagged job loss.

  12. Simply including the lagged dependent variables in the estimation of Table 3 raises a problem of autocorrelation: modified Durbin-Watson statistics in the case including wives’ past labor hour changes are 1.863, 1.866, 1.865 and 1.866, respectively, for columns (1), (2), (3) and (4) of Table 3, implying existence of AR(1) serial correlation.

  13. The present paper points out the importance of households’ precautionary savings but does not deal with it explicitly. More detailed investigation should be considered in future research.

  14. We list the results only using the specifications (1a) and (1b) in Table 4 because those two specifications give our main results in the previous section. However, the implication has not changed even if we use the other model specifications shown in Tables 3 and 4.

References

  • Arellano M, Bond SR (1991) Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. Rev Econ Stud 58(2):277–297

    Article  Google Scholar 

  • Attanasio O, Low HW, Sanchez-Marcos V (2005) Female labor supply as insurance against idiosyncratic risk. Journal of European Economic Association 3(2–3):755–764

    Google Scholar 

  • Bhargava A, Franzini L, Narendranathan W (1982) Serial correlation and the fixed effects model. Rev Econ Stud 49(4):533–549

    Article  Google Scholar 

  • Bingley P, Walker I (2001) Household unemployment and the labour supply of married woman. Economica 68(270):157–185

    Article  Google Scholar 

  • Cullen JB, Gruber J (2000) Does unemployment insurance crowd out spousal labor supply? J Labor Econ 18(3):546–572

    Article  Google Scholar 

  • Fernandes R, Felicio F (2005) The entry of the wife into the labor force in response to the husband’s unemployment: a study of the added worker effect in Brazilian Metropolitan areas. Econ Dev Cult Change 53(4):887–911

    Article  Google Scholar 

  • Heckman JJ, McCurdy TE (1980) A life cycle model of female labour supply. Rev Econ Stud 47(19):47–74

    Google Scholar 

  • Higuchi Y (2001) Koyo To Shitsugyo No Keizaigaku. Nikkei, Tokyo (in Japanese)

    Google Scholar 

  • Low HW (2005) Self-insurance in a life-cycle model of labour supply and savings. Rev Econ Dyn 8(4):945–975

    Article  Google Scholar 

  • Lundberg S (1985) The added worker effect. J Labor Econ 3(1):11–37

    Article  Google Scholar 

  • Pijoan-Mas J (2006) Precautionary savings or working longer hours? Rev Econ Dyn 9(2):326–352

    Article  Google Scholar 

  • Spletzer JR (1997) Reexamining the added worker effect. Econ Inq 35(2):417–427

    Article  Google Scholar 

  • Stephens M (2002) Worker displacement and the added worker effect. J Labor Econ 20(3):504–537

    Article  Google Scholar 

Download references

Acknowledgements

The author is grateful to two anonymous referees, James Albrecht (the responsible editor), Kenn Ariga, Jose Luis Moraga Gonzalez, Fumio Hayashi, Yoshio Higuchi, Charles Yuji Horioka, Yasushi Iwamoto, Akira Kawaguchi, Daiji Kawaguchi, Franco Mariuzzo, Colin McKenzie, Naoki Mitani, Fumio Ohtake, Soichi Ohta, Kazuaki Sakamoto, Makoto Saito, Masaru Sasaki, Ryuhei Wakasugi, Makoto Yano, seminar participants at University of Groningen, Keio University and at Hitotsubashi University, and the members of Kansai Rodo Kenkyukai for their valuable comments. The author is also grateful to the Ministry of Education, Culture, Sports, Science and Technology of the Japanese Government for Grant-in-Aid for Scientific Research number 12124207, which supported this research.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Miki Kohara.

Additional information

Responsible editor: James Albrecht

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Kohara, M. The response of Japanese wives’ labor supply to husbands’ job loss. J Popul Econ 23, 1133–1149 (2010). https://doi.org/10.1007/s00148-009-0247-6

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00148-009-0247-6

Keywords

  • Added worker effects
  • Within-family risk-sharing
  • Household panel data

JEL Classification

  • D12
  • J22
  • C23