I. Introduction

View the most recent version.

Do individuals respond to the occurrence of a layoff in the family by increasing their employment income? If so, what is the magnitude of their response? Since what happens at the family level is a key determinant of individuals' well-being, identifying the degree to which family members stabilize family income in the event of a layoff is critical for a thorough understanding of the welfare consequences of earnings shocks. Yet, even though a large body of empirical evidence has shown that worker displacement leads to long-term earnings losses at the individual level (Ruhm 1991; Jacobson, LaLonde and Sullivan 1993; Stevens 1997; Morissette, Zhang and Frenette 2007), relatively few studies have examined these questions.

One notable exception is Stephens (2002), who quantifies the labour supply response of wives to husbands' displacement. Based on U.S. data (from the Panel Study of Income Dynamics) that covers the 1968-to-1992 period, his findings suggest that, five years after displacement, increases in wives' work hours compensate for roughly 30% of husbands' earnings losses.

While these numbers shed new light on the degree to which increases in wives' labour supply mitigated the earnings losses of displaced husbands during the 1970s and the 1980s in the United States, whether they can be generalized to other industrialized countries or time periods is unknown. In particular, there are good reasons to believe that they might overestimate the degree to which wives' earnings respond to husbands' layoffs nowadays.

One reason is that women's labour supply behaviour has changed significantly over the last two decades. Recent evidence suggests that women's labour supply has become much less responsive, both to their own wages and to the wages of their husbands in the 1990s, when compared with the 1980s (Blau and Kahn 2005). Since a growing proportion of women entered the labour market and started working full time over the last two decades, fewer of them now have the ability to increase their work hours at the extensive margin (by entering the labour market) or at the intensive margin (by increasing their hours, conditional on participation) in response to declines in husbands' wages. Hence, fewer of them might also have the ability to increase their labour supply in response to husbands' displacement. The substantial growth in the risk of divorce over the last three decades, along with changes in women's attitudes toward family and work, might also have induced some women to adopt a labour-supply behaviour that is fairly independent of their husbands' outcomes. All of these factors suggest that wives' labour supply response to husbands' job loss might have been smaller in the 1990s than it was during the 1970s and 1980s. Unless the growth in female wages over the last few decades fully offsets this potentially smaller labour supply response, the implication is that wives' annual earnings are likely to have compensated for the earnings losses of displaced husbands to a smaller extent in the 1990s than they had done in the 1970s and 1980s.

Conversely, it is conceivable that families now adjust to husbands' layoffs not only through changes in wives' labour supply, but also through increases in work hours of children of working age. Given that a growing proportion of youth postponed departure from home in recent years (Card and Lemieux 1997), families may, a priori, respond to husbands' layoffs partly through an increase in youth earnings. However, many teenagers and young adults attend school full time and thus cannot devote much time to paid work; whether this happens or not is unclear and remains to be investigated.

While changes in women's labour supply and youth living arrangements have potentially altered families' response to layoffs, the substantial growth in the proportion of unmarried individuals over the last two decades has increased the relative importance of households that cannot benefit from any 'added worker effect.' In 2005, 34% of all family units in Canada consisted of unattached individuals, up from 27% in 1980.¹ These changes in family structure raise an important question: how do earnings losses of unattached workers compare with those of their married counterparts?

Earnings losses of unattached workers might differ from those of married workers for a variety of reasons. Since the former group are likely to be more geographically mobile than the latter, they might take advantage of favourable job offers in other local labour markets more frequently, after losing their job. As a result, following displacement, they might incur smaller earnings losses than the latter group. Furthermore, as long as they have lower wealth holdings than their married counterparts, unattached individuals might search more intensely for new jobs following a layoff. This may reduce the duration of the unemployment spell that they incur after job loss, thereby mitigating their earnings losses in the short run. However, this may also lead them to accept lower wage offers and thus potentially incur greater long-term earnings losses than those suffered by displaced married workers.

Even if long-term earnings losses of married workers and unattached workers were identical and no 'added worker effect' whatsoever were observed among families, the mere presence of a second earner in many families suggests that the magnitude of the relative income shocks due to layoffs might be much smaller among families than among unattached individuals. This, in turn, raises the possibility that job loss might increase income instability much more among the latter group than among the former. While recent work has investigated whether job loss can account for part of the increase in earnings instability in the United States (Stevens 2001), the question of whether earnings shocks due to layoffs generate relative income shocks—and thus, increases in income instability—of varying magnitude across household types has received little attention.

The goal of this paper is twofold. In light of the changes in women's labour supply and youth living arrangements outlined above, we first wish to provide recent evidence on whether earnings of wives and children of working age increase in response to husbands' layoffs. To do so, we use a large longitudinal Canadian dataset that covers the 1987-to-2001 period. Our dataset is based on tax records and thus contains fairly accurate information about the annual earnings of husbands, wives and teenagers. It also contains accurate information on Employment Insurance benefits and income tax paid by family members, thereby allowing us to assess the income losses (before tax and after tax) experienced by families as a result of layoffs.

Second, we wish to contrast the long-term earnings losses experienced by laid-off husbands to those of unattached individuals, thereby providing estimates of the consequences of layoffs for an increasingly important type of household that cannot rely on the presence of a second earner. When doing so, we compare the magnitude of the income losses (before tax and after tax) experienced by families of laid-off husbands, on the one hand, and unattached males, on the other. As a result, we can quantify the magnitude of the relative income shocks experienced in the long term by both groups.

Our main findings can be summarized as follows. First, we find virtually no evidence that earnings of youth increase in response to fathers' layoffs. We also find no evidence of an 'added worker effect' for Canadian wives in the aggregate. However, we do find evidence that, among families with no children of working age, earnings of wives increased during the 1990s following husbands' layoffs. Our estimates for this group of families imply that, five years after husbands' layoffs, increases in wives' earnings compensated for roughly 22% of the earnings losses experienced by husbands.

Second, we show that laid-off husbands and unattached men experience roughly the same earnings losses in the long run. We also find that families of laid-off husbands incur similar income losses before tax and after tax, compared with unattached males. However, because unattached males have much lower pre-layoff after-tax income, they end up experiencing much greater relative income shocks than families of laid-off husbands. This, in turn, suggests that layoffs among adult male earners increase income instability much more for unattached males than they do for families of laid-off husbands.

 

^1. These numbers are drawn from CANSIM table 202-0401.