Income after job loss: the role of the family and the welfare state

Abstract

That displaced workers suffer long-lasting earnings losses is a stylized fact, raising the question whether these losses are replaced by other means. For married men, increased labour supply by the wife may be one way. Another possibility is that the public welfare system offsets the same losses. I used a Swedish longitudinal data set containing married couples where the husband was either employed or made redundant in 1987 by an establishment closure. There was no evidence that husbands’ job loss positively affected wives’ annual earnings. Although husbands’ utilization of unemployment insurance increased significantly, government transfers including also sickness insurance, disability insurance and means-tested social benefits, did not fully replace husbands’ long-run earnings losses. Hence, displaced workers seem to suffer also from long-lasting losses in family income, which in many respects is a better measure of economic welfare than individual earnings or wages.

Notes

¹ Most of this literature have focused on wives’ labour supply response to their husband's current unemployment, and have then ignored any response before or after the occurrence of job loss. An exception is Stephens (2001) who showed that long-run increases in spousal labour supply compensated for over 25% of husbands’ lost income due to job displacement.

² See Holmlund (2003) for more details.

³ See Bergmark and Palme (2003) or Sjögren Lindquist and Wadensjö (2006) for a more thorough description of the development and details of the Swedish social security system.

⁴ Unemployment insurance is actually not a part of the social insurance system in Sweden but comes under labour market policy.

⁵ See Kasl and Jones (2000) for a review.

⁶ Income taxation and the administration of the universal Swedish welfare state provide the source for many of the variables in these registers. The employer files all wage payments to the tax authorities and, as practically all transfers in the Swedish welfare state, such as disability insurance, and sickness and unemployment benefits, are liable to tax the National Social Insurance Board also files income statements on such transfers (together with nontaxable social assistance payments).

⁷ Registers can be merged and one can link employees to their establishments since every resident and every establishment in Sweden has a unique identity number (i.e. a civic registration number or an organization number) and that the obligatory income statements filed to the taxation authorities by the employer, contain both these numbers.

⁸ This ensures that all displaced workers were employed at the same point in time as the workers in the comparison group and implies that for closures in 1987, only those who were displaced in the closing year were included in the sample and for closures in 1988, only those displaced during the calendar year before final closure given that the closing process was deemed longer than a year. The length of each individual closing process was determined ad hoc based on establishment size and worker flows during the three years preceding the closure. Most of the closing processes, however, were considered less than a calendar year.

⁹ See Falck (2007) for an empirical study of the impact of regional conditions on firm survival.

¹⁰ See Hirano and Imbens (2001) for their suggested propensity score weighted regression and Robins and Rotnitzky (1995) for their inverse-probability-of-censoring weighted estimator.

¹¹ See Hirano and Imbens (2001).

¹² Divorce is not the only reason for a couple to be censored; marriages also dissolve upon the death of a spouse, and even though migration of one or both spouses does not necessarily mean that the marriage is dissolved, they are no longer observed in the data.

¹³ Instead applying the ‘stabilized’ inverse-probability-of-censoring-weights, also proposed by Robins and Rotnitzky (1995), does not alter any results in the following analysis.

¹⁴ Various other balancing tests have been suggested in the literature (Smith and Todd, 2005), but there is no consensus on which of them to apply.

¹⁵ The SDM is the difference in covariate means between the displaced couples and the weighted nondisplaced couples, in percentage of the pooled standard deviation (before weighting) of that covariate.

¹⁶ The presented relative effects are calculated by dividing the estimate by the difference between the weighted mean for the displaced workers and the same estimate.

¹⁷ Stevens (1997) also showed that the main cause of the long-term effects of job loss on earnings was attributable to higher risks of subsequent job losses.

¹⁸ For a review, see Jones (1992) or Ström (2003).

¹⁹ For the Netherlands it has been found that approximately one quarter of the disability insurance enrolment was hidden unemployment (Koning and van Vuuren, 2007).

²⁰ Since the shame or social stigma associated to receiving means-tested social assistance implies that not all of those who are poor enough to be eligible actually apply, any measure based on receiving social assistance is likely to underestimate the true degree of economic hardship (Stenberg, 1998).

²¹ The measure of family income is not equivalent to total disposable family income but is the before tax sum of both spouses’ earnings and the transfers investigated here. As stated above the exclusion of any supplementary benefits are likely to overestimate any adverse impact although diminished by the fact that some of these benefits show up as earnings in the data. The use of family income before taxes, on the other hand, will probably overestimate the same effects to some extent as the progressive income tax is likely to offset some of the differences.