II. Prior research

View the most recent version.

Since the early 1990s, numerous studies have examined the magnitude of the earnings losses due to displacement (see the reviews by Fallick 1996 and Kletzer 1998). Using Pennsylvania administrative data, Jacobson, LaLonde and Sullivan (1993) show that the earnings losses of high-tenure prime-age workers persist well beyond the time of displacement. Earnings fall even before the displacement takes place and drop sharply at the time of the displacement. Even several years after the displacement took place, displaced workers report quarterly earnings that are about 25% lower than their pre-displacement earnings. Worse still, it seems very likely that the earnings of displaced workers do not return to their expected levels at any time. Ruhm (1991) and Stevens (1997) also analyse the earnings losses of displaced workers, using data from the Panel Study of Income Dynamics (PSID). While Ruhm (1991) finds that, four years after displacement, weekly earnings of displaced workers are 10% to 13% lower than those of their non-displaced counterparts, Stevens (1997) shows that the annual earnings of displaced workers remain about 9% below their expected levels six or more years after displacement.² Eliason and Storrie (2006) also find long-term earnings losses as a result of job displacement. In all of the aforementioned studies, no attempt is made to distinguish the long-term earnings losses of displaced husbands from those of displaced unmarried males.

Since individuals who become unemployed through mass layoffs suffer substantial and persistent earnings losses, one important question is whether various family members, especially the wives of displaced husbands, adjust their labour supply to mitigate the impact of their husbands' earnings losses. While earlier studies using cross-sectional data have failed to detect an empirically important 'added-worker' effect (e.g., Heckman and MaCurdy 1980, Cullen and Gruber 2000), Stephens (2002) uses longitudinal data from the PSID and finds that five years after husbands' displacement, wives' work hours increase significantly, compensating for 30% of husbands' earnings losses.³

Because government transfers, such as Employment Insurance (EI) benefits are generally provided for a limited period of time (e.g., 50 weeks) while others, like social assistance, are generally provided when EI benefits are exhausted, the finding that wives' earnings only partially offset husbands' earnings losses suggests that husbands' displacement induces, in the longer run, a drop in family income.⁴ However, because the tax system mitigates the income fluctuations experienced by families (Kniesner and Ziliak 2002), the extent to which families' disposable income drops following husbands' unemployment remains to be determined. Moreover, it is conceivable that other family members partially adjust their labour supply in response to husbands' unemployment (Jenkins 2000: 552), providing an additional channel through which the adverse consequences of husbands' unemployment on family income could potentially be reduced.

 

^2. While Jacobson, LaLonde and Sullivan (1993) require displaced workers to have at least six years of tenure with their employer, Ruhm (1991) and Stevens (1997) do not impose this restriction.

³ Stephens (2002) also shows that the magnitude of the increase in wives' work hours depends on the magnitude of husband's earnings losses: the larger are husbands' earnings losses, the greater is the increase in wives' work effort.

⁴ This is what Stephens (2001) finds, using data from the Panel Study of Income Dynamics.