Main article

View the most recent version.

Archived Content

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.

Archived

This page has been archived on the Web.

1. Introduction
2. Data
3. Which workers were laid-off during the last three recessions?
4. How did layoff rates vary across recessions?
5. How did chances of finding paid employment shortly after being laid-off vary across recessions?
6. How do wages before and after layoffs compare?
7. Conclusion

1   Introduction

Over the last three decades, Canada has experienced three recessions. As a result of these, unemployment rates rose sharply in 1981-to-1983, 1990-to-1992, and, most recently, after October 2008. To shed light on the labour market implications of these slowdowns, this study answers four questions:

1. Which workers were laid-off during these recessions?
2. How did layoff rates vary across recessions?
3. How did chances of finding employment shortly after being laid-off evolve across recessions?
4. Among workers who managed to find paid employment shortly after being laid-off, how do wages before and after layoffs compare?

Since the last three decades have witnessed important movements in the composition of employment by industry, workers' age, and workers' educational attainment, the study first assesses the degree to which: a) workers laid-off during the last recession differed from their counterparts laid-off in the early 1980s and the early 1990s; and b) temporal changes in the profile of laid-off workers can be accounted for by the aforementioned compositional effects. Layoff rates in each recession are also provided.

Next, the likelihood of laid-off workers finding employment in the short term, i.e., in a period ranging from one month to four months after layoff, is assessed. Whether the factors that facilitate or impede short-term re-employment are different now than they were in the early 1980s and in the early 1990s is also investigated.

Finally, hourly wages and weekly wages before and after layoffs are compared. Considerable attention is devoted to documenting the fact that laid-off workers experience fairly heterogeneous short-term wage changes, with substantial shares experiencing either wage gains or substantial wage losses. As a result of data limitations, the aforementioned comparison is done only for the most recent downturn.

To examine the four questions identified above, the study takes advantage of the panel nature of the Labour Force Survey (LFS), whereby households are interviewed in each of six consecutive months. However, while the LFS asks laid-off workers whether they expect to be recalled, it does not collect information on whether they actually return to the same employer (regardless of their recall expectations). Hence, the data do not allow permanent and temporary layoffs to be differentiated, whereas this is the case in Statistics Canada's Longitudinal Worker File (LWF)-another data source used to study job displacement. As a result, the layoff concept used in this article includes both temporary and permanent layoffs and thus is not equivalent to the concept of job loss (due to permanent layoffs) used in previous studies based on the LWF (e.g., Morissette 2004; Morissette, Zhang, and Frenette 2007). Despite this limitation, LFS data help paint a rich picture of the determinants and short-term consequences of layoffs during the last three recessions. The main findings can be summarized as follows.

Canadian workers were less likely to be laid-off during the most recent recession than their counterparts were in the early 1980s and the early 1990s. Assessed on a monthly basis, the risk of layoff during the early 1980s averaged 2.9%; this rate is almost 1.5 times higher than the 2.0% rate observed in 2008-to-2011. The risk of layoff averaged 2.7% in the early 1990s. ¹ 

During all three periods, chances of being temporarily or permanently laid-off were relatively high among young workers (those aged 15 to 24), individuals with no university degree, newly hired employees (those with two years or less of seniority), and those employed in the goods sector. However, such patterns are not specific to periods of economic slowdown: they are also observed during expansionary periods.

Comparing the risk of layoff associated with various characteristics across the three recessions, the study found that workers laid-off during the most recent recession were older, better educated, and less likely to come from the manufacturing sector than those laid-off during the early 1980s or the early 1990s. These temporal changes in the profile of laid-off workers resulted mainly from compositional effects, i.e., changes in the age/education profile of the Canadian workforce as well as the secular decline of the manufacturing sector.

However, compositional effects did not account for all changes in the profile of laid-off workers. For instance, employees laid-off in 2008-to-2011 had more seniority than their counterparts laid-off in the early 1980s and the early 1990s, even after controlling for changes in the seniority profile of the Canadian workforce. All else equal, high-seniority workers (those with more than five years of seniority) were less likely to be laid-off than newly hired employees (those with two years or less of seniority) during all three downturns. However, the average seniority of workers laid-off-measured in terms of the likelihood of being laid-off-was smaller during the most recent downturn than during the previous two. When controls for individual characteristics are applied, high-seniority workers were about 6-percentage-points less likely to be laid-off than newly hired employees during the early 1980s and the early 1990s. In contrast, the corresponding difference dropped to 3.0 percentage points in 2008-to-2011, as the risk of layoff of newly hired employees fell over time.

Of all workers laid-off in the 2008-to-2011 period, 50% found a paid job between one and four months after being displaced. This share is significantly larger (both statistically and quantitatively) than the corresponding proportion of 42% observed during the previous two recessions.

The workers most likely to be re-employed in the short term had the following characteristics: they initially expected to be recalled; they had a university degree; and they had more than five years of seniority.

On average, employees who were laid-off during the most recent recession and who found a job shortly after being laid-off saw their average weekly wages drop from $734 to $703. However, one-quarter saw their weekly wages drop by 23% or more, while another one-quarter saw increases in weekly pay of at least 18%.

Average declines in weekly wages amounted to at least 10% for the following workers: those who lost union coverage; those moved from a firm with at least 100 employees to a smaller firm; and those who changed both industry and occupation in the new job. Collectively, these groups represented about one-quarter of laid-off workers who were re-employed during the most recent downturn. In contrast, employees who gained union coverage or moved from firms with fewer than 100 employees to firms with 100 or more employees registered average gains in weekly wages that amounted to between 8% and 11%. Collectively, these groups represented about 17% of laid-off workers who were re-employed during the most recent downturn. Previous studies have shown that high-seniority workers who are laid-off often experience substantial and sustained earnings losses (Jacobson, Lalonde, and Sullivan 1993; Morissette, Zhang, and Frenette 2007; Couch and Placzek 2010). Given that a larger proportion (28%) of laid-off workers had high seniority during the most recent recession than during the early 1990s (17%) or the early 1980s (16%), an important question for future research is whether the long-term earnings losses of workers displaced during the most recent downturn will end up being higher or lower than those of their counterparts displaced during previous downturns.

Apart from displaying lower layoff rates and higher short-term re-employment rates than the previous two recessions, the most recent downturn was also of shorter duration. Total employment (seasonally adjusted) took 27 months to return to its pre-downturn level, compared to 53 months during the early 1990s and 40 months during the early 1980s.

The lower layoff rates observed during the last recession mirrored the relatively small peak-to-trough employment decline observed in recent years. As pointed out by Cross (2011), the peak-to-trough decline in employment observed in 2008-2009 amounted to 2.4%, compared to 5.4% in the early 1980s and 3.4% in the early 1990s. ² 

The paper is organized as follows. The data used in the analysis are described in Section 2. Section 3 sketches a profile of workers laid-off during the last three recessions. Section 4 compares layoff rates across recessions. Section 5 documents the evolution of employment rates shortly after layoff. Section 6 quantifies the wage changes experienced by employees who were laid-off during the most recent recession and who found a job shortly after being laid-off. Concluding remarks follow.

2   Data

The data used in the current study are drawn from the master file of Statistics Canada's Labour Force Survey (LFS). The LFS is a rotating panel survey in which households are interviewed for six consecutive months. ³  The total sample consists of six representative sub-samples, one of which is replaced each month after it has completed the six-month stay in the survey. This rotation results in a five-sixths month-to-month sample overlap, which in turn allows estimates of month-to-month changes in labour force status. To take advantage of the panel nature of the data, a (pseudo) individual identifier is created in order to identify a given individual within the LFS panel. ⁴ 

To identify which workers have been laid-off during the last three recessions and to assess how chances of being laid-off varied across recessions, the following sample is constructed. For each of the three recessions, the months observed between the onset of the employment downturn and the return to the pre-downturn level are chosen on the basis of seasonally-adjusted total employment. As a result, the "peak-to-peak" periods for the aforementioned downturns are the following: June 1981 to October 1984; April 1990 to September 1994; and October 2008 to January 2011. ⁵  For this reason, the study refers to the most recent recession as the 2008-to-2011 period.

To calculate layoff rates on a monthly basis, monthly transitions from paid employment to non-employment due to layoffs are captured. For each pair of months selected, inflows into non-employment (due to layoffs) between month t-1 and month t are divided by the number of workers in paid employment in month t-1. This yields month-specific layoff rates. Averaging these month-specific layoff rates across all pairs of months observed during the peak-to-peak periods yields a layoff rate for a given economic downturn. ⁶  This layoff rate is calculated for a sample of paid workers who are aged 15 to 64 in the month preceding the layoff and who are not full-time students. More details about the identification of laid-off workers are provided in Appendix II.

The layoff rates shown in this study differ from those published in Morissette (2004) and in Morissette, Zhang, and Frenette (2007) for two reasons. First, the aforementioned studies focused on permanent layoffs and, thus, excluded temporary layoffs. While the LFS asks laid-off workers whether they expect to be recalled, it cannot assess whether a worker laid-off in year t actually returns to the same employer in year t or in year t+1 (no matter what that individual's recall expectations are). Hence, it cannot differentiate between permanent and temporary layoffs as these terms are defined in the LWF. ⁷  Second, the layoff rates presented in this study are calculated on a monthly basis. In contrast, the layoff rates presented in the studies above were calculated on an annual basis. While the number of workers at risk of being laid-off (i.e., the denominator in the calculation of layoff rates) differs relatively little whether calculations are made on a monthly basis or on an annual basis, the number of workers being laid-off is, as expected, much smaller when the time interval considered is one month rather than one year. ⁸ 

To investigate how chances of finding paid employment shortly after being laid-off evolved across downturns and to quantify the wage changes of laid-off workers who found paid employment shortly after being laid-off, the study takes advantage of the fact that the LFS follows workers for six months.

The sample used for this portion of the analysis consists of individuals who: a) are observed in all six consecutive months in each panel; b) are aged 15 to 64 in all six months; c) are not full-time students in any of the six months; d) are employed as paid workers during the first month of the panel; and e) have been laid-off at any point between the second month and the fifth month. ⁹  The percentage of these workers who are re-employed during the sixth month and the magnitude of the wage changes and changes in usual weekly hours experienced by those who are re-employed are then measured.

Short-term re-employment rates?employment rates observed between one and four months after layoffs?are compared across all three downturns. ¹⁰ In contrast, wage changes are examined only for the most recent downturn. The reason is that LFS information on wages is available only for 1997 and subsequent years. To put numbers into context, wage changes during the 2008-to-2011 period are compared with those observed in the expansionary period that immediately preceded the beginning of the downturn. Twenty-three (six-month) panels from July 2006 to October 2008 are constructed for this purpose.

3   Which workers were laid-off during the last three recessions?

3.1  Background

As is well known, Canadian workers became older, better educated, and more likely to be employed in the service sector over the last three decades. As a result, the profile of workers laid-off in a given downturn may have changed over time. This question is investigated in Table 1.

Workers laid-off in 2008-to-2011 differed from their counterparts laid-off in the early 1980s and early 1990s in several ways: the former group was older and better educated and had more years of seniority in the job.

Almost 40% of workers laid-off during the most recent downturn were aged 45 or older, twice the rate of 19% observed in the early 1980s. Meanwhile, the share of laid-off workers who were aged 15 to 24 declined from 35% to 19%.

Sixteen percent of workers laid-off in the 2008-to-2011 period had a university degree, compared to 8% in the early 1990s and 5% in the early 1980s. ¹¹  Twenty-eight percent had more than five years of seniority, up from 16% in the early 1980s.

The two most populous provinces registered significant changes in the share of laid-off workers, albeit in different directions. While Quebec and Ontario accounted for about 60% of laid-off workers in all three downturns, Ontario's share rose from roughly 30% during the first two downturns to 36% in 2008-to-2011. Meanwhile, Quebec's share fell from about 30% to 25%.

The industrial and occupational profile of laid-off workers also changed over time. ^12 , 13 In the early 1980s, 46% of laid-off workers came from primary industries, construction, and manufacturing. The corresponding proportions fell to 43% in the early 1990s and to 38% during the most recent downturn. In contrast, relatively more professionals, semi-professionals, and technicians were laid-off in the most recent downturn (19%) compared to the early 1990s (13%). ¹⁴ 

Despite the growing proportion of women in the workforce (Text Table 2), the proportion of men and women laid-off remained fairly stable across all three downturns, as women accounted for between 40% and 43% of all laid-off employees.

3.2  Compositional effects

As Text Table 1 shows, changes in the profile of laid-off workers by age, educational attainment, and industry were also observed across expansionary periods. Such changes are expected, since?as was mentioned above?Canadian workers became older, better educated, and less likely to be employed in the goods sector over the last three decades (Text Table 2).

To what extent are changes in the profile of laid-off workers driven by these compositional effects? Table 2 answers this question by using shift-share analyses. The question asked is the following: what would the age (education, seniority, industrial) profile of laid-off workers have been in the 2008-to-2011 period if the age (education, seniority, industrial) profile of paid workers had remained as it was in 1981-to-1984?

For instance, the proportion of laid-off workers aged 45 to 64 increased by almost 20 percentage points between 1981-to-1984 and 2008-to-2011 (Table 2, column 3). Had the distribution of employment by age remained unchanged, the increase would have amounted to only about 3 percentage points (column 4). Thus, the aging of the workforce accounts for at least 80% (17 percentage points out of 20) of the growth in the proportion of laid-off workers aged 45 to 64 between the early 1980s and the most recent downturn.

A comparison of columns 3 and 4 also reveals that increases in workers' educational attainment explain about 80% of the increase in the proportion of laid-off workers holding a university degree. Likewise, decreases in the relative importance of manufacturing employment fully account for the drop in the proportion of laid-off workers coming from manufacturing, while increases in the relative importance of high-skill services account for almost two-thirds of the rise in the share of laid-off workers originating from this sector.

While movements in the composition of the workforce and in the industrial structure drove most of the changes in the composition of laid-off workers by age, education, and industry, ¹⁵  they explain little of the growth in the proportion of laid-off workers with high seniority. The share of laid-off workers with more than five years of seniority grew by about 12 percentage points between the early 1980s and 2008-to-2011 (Table 2, column 3). Had the distribution of employment by seniority levels remained unchanged at its 1981-to-1984 levels, the corresponding increase would have amounted to 10 percentage points (Table 2, column 4). Thus, most of the increase in the share of high-seniority laid-off workers would have occurred in the absence of compositional effects. This in turn suggests that most of this growth was driven by differential changes in layoff rates across seniority levels. Table 3 supports this contention: between 1981-to-1984 and 2008-to-2011, layoff rates fell among workers with at most five years of seniority but not among their counterparts with greater seniority.

Likewise, Table 2 indicates that the whole increase (5.0 percentage points) in Ontario's share of laid-off workers would have occurred even had the distribution of employment by region remained stable over time. In an accounting sense, the growing proportion of laid-off workers coming from Ontario is due to the fact that, relative to the Canadian average, layoff rates in Ontario grew substantially between 1981-to-1984 and 2008-to-2011 (Text Table 3).

In sum, changes in the composition of the workforce and employment shifts across industries generally accounted for most of the changes in the profile of laid-off workers between 1981-to-1984 and 2008-to-2011. However, changes in relative layoff rates altered significantly the proportion of high-seniority laid-off workers as well as the proportion of laid-off employees coming from Ontario or previously employed in retail, trade, accommodation, and food services. These conclusions hold when alternative counterfactuals, which assume a stable composition of paid employment between the early 1990s (rather than the early 1980s) and 2008-to-2011, are considered (Text Table 4).

4   How did layoff rates vary across recessions?

4.1  Descriptive evidence

Overall, Canadians' chances of being laid-off were lower during the most recent employment downturn than during the previous two. Measured on a monthly basis, the aggregate layoff rate in 2008-to-2011 averaged 2.0%, compared to 2.7% for the early 1990s and 2.9% for the early 1980s (Table 3 and Chart 1). ¹⁶  Once again, these statistics include temporary as well as permanent layoffs. ¹⁷ 

In most demographic groupings considered, layoff rates were lower in 2008-to-2011 than in 1981-to-1984 or in 1990-to-1994. A few exceptions must be noted. University graduates and workers with more than five years of seniority did not experience a lower risk of layoff in 2008-to-2011, compared to the previous two downturns. As a result, these groups also saw an increase in their relative layoff rates, i.e., their layoff rates divided by the overall layoff rate (Text Table 3).

Workers most likely to be laid-off during the most recent downturn were male, were aged 15 to 24, had two years or less of seniority with the firm, had no university degree, were living in the Atlantic Provinces, and were employed in primary industries and construction.

For instance, layoff rates among workers aged 15 to 24 amounted to 3.4%, twice the rate of 1.7% found among their counterparts aged 35 to 44. With a layoff rate of 3.6%, newly hired employees (those with two years or less of seniority) were three times more likely to be laid-off than their counterparts with 10 to 20 years of seniority. Non-university graduates had monthly layoff rates of 2.2%, while the rate for university graduates was 1.2%.

However, all of these qualitative patterns were also found during previous downturns (Table 3) as well as during previous expansionary periods. ¹⁸ 

4.2  Determinants of the probability of being laid-off

To control for the influence of potential confounders, Table 4 presents the marginal effects from a logit model of the probability of being laid-off. The dependent variable equals 1 (one) if a worker has been laid-off, 0 (zero) otherwise. The set of explanatory variables consists of a gender indicator, a quadratic term in age, a binary indicator for university graduates, seniority indicators, region indicators, and broad industry controls.

The results of this multivariate analysis confirm the findings of Section 4.1: all else equal, young workers, individuals with two years or less of seniority, workers employed in primary industries and construction, and those living in the Atlantic Provinces faced the highest layoff risk during the last downturn as well as during the previous two downturns.

All else equal, workers with a university degree were less likely to be laid-off than other workers. The difference in layoff risk amounted to 0.9 percentage points during the most recent downturn; this is down from 2.0 percentage points during the early 1980s (Table 4).

Workers with more than five years of seniority were less subject to layoffs than newly hired employees. However, the difference in the likelihood of being laid-off narrowed over time, dropping from 6.4 percentage points in the early 1980s to 3.0 percentage points during the most recent downturn. The difference narrowed mainly because newly hired employees saw their layoff rate drop by at least 2.0 percentage points between the first two downturns and the most recent one (Table 3). ¹⁹ 

On a monthly basis, inter-regional (absolute) differences in the probability of being laid-off were less pronounced during the most recent downturn than during previous ones. They reached a maximum of 1.6 percentage points in 2008-to-2011, compared to 1.9 percentage points in the early 1980s and 2.2 percentage points in the early 1990s. ²⁰  In all periods, Quebec and the Atlantic Provinces displayed a higher risk of layoff than Ontario and Alberta. However, these inter-regional differences narrowed during the most recent downturn, reflecting the differentiated impact of the last downturn across provinces. For instance, the probability of being laid-off was only 0.2-percentage-points higher in Quebec than in Ontario during the most recent downturn, compared to 1.0-percentage-point higher during the early 1980s and 1.1-percentage-points higher during the early 1990s. Likewise, differences in the risk of layoff between Quebec and Alberta fell from 1.4 percentage points during the early 1980s and the early 1990s to 0.8 percentage points during the most recent downturn. Contrary to Quebec and the Atlantic Provinces, Alberta consistently exhibited lower layoff risks than Ontario. However, the difference in layoff risk between Alberta and Ontario did not narrow during the most recent downturn. British Columbia, which was hit fairly hard during the downturn of the early 1980s, had a higher risk of layoff than Ontario in 1981-to-1984, but this was not the case in 2008-to-2011. All else equal, the risk of layoff was 0.5-percentage-points lower in the Prairie Provinces than in Ontario during the most recent downturn.

In all three downturns, chances of being laid-off were at least 2-percentage-points lower in service-producing industries than in primary industries and construction. Workers in manufacturing were also less likely to be laid-off than their counterparts in primary industries and construction: the difference varied between 1.7 percentage points and 2.9 percentage points, depending on the downturn considered.

Overall, Table 4 confirms that, during all three downturns, workers' risk of being laid-off varied consistently across the following dimensions: age, education, seniority levels, region, and industry. ²¹ 

5   How did chances of finding paid employment shortly afterbeing laid-off vary across recessions?

Of all paid workers who were laid-off in 2008-to-2011, half of them found a paid job between one month and four months after being laid-off (Table 5). This percentage is higher than the corresponding proportion of roughly 42% observed during the early 1980s and the early 1990s (Chart 2). ^22 , 23 

For many groups of workers, short-term employment rates following layoffs were higher in 2008-to-2011 than during the previous two downturns. Compared to those in the early 1980s and the early 1990s, employment rates in 2008-to-2011 rose by between 5.0 percentage points and 13.0 percentage points among male and female workers. They grew by at least 9.0 percentage points for younger workers (aged 34 or younger) and by at least 6.0 percentage points for older workers (aged 35 or older). For workers with education attainment below university, the employment rate rose by about 7 percentage points; among university graduates, the increase is between 8.0 percentage points and 12.0 percentage points. For workers with low seniority (two years or less), the increase in employment rate is about 6 percentage points. The rates grew by at least 10.0 percentage points for workers in the Atlantic Provinces, Quebec, and the Prairie Provinces. Workers employed in public services at the time of layoff also saw their employment rates increase by at least 10.0 percentage points.

For other groups, employment rates after layoffs showed relatively little improvement. This was the case for workers with high job tenure (more than five years), workers living in Ontario, Alberta, and British Columbia, as well as those employed in manufacturing and high-skill services.

Short-term employment rates also varied across worker characteristics. This is shown in Table 6, where a multivariate analysis of the probability of being re-employed is conducted. All else equal, laid-off workers who expected to be recalled were at least 14-percentage-points more likely to find a paid job in the short term than those who expected no recall. ²⁴  Employees with a university degree had greater chances of finding paid employment in the short run than others (the difference amounted to about 5 percentage points).

The likelihood of high-seniority workers experiencing relatively high short-term re-employment rates evolved differently across groups. High-seniority workers who expected to be recalled were more likely to be re-employed than their counterparts newly hired in the first two downturns: the re-employment rates of the former group exceeded those of the latter by 6.0 percentage points to 7.0 percentage points during the first two downturns and by about 3 percentage points (imprecisely measured) in 2008-to-2011. Among workers who did not expect a recall, this seniority advantage amounted to 9.0 percentage points in the early 1980s, to 6.0 percentage points in the 1990s, and to 8.0 percentage points in 2008-to-2011.

In general, employment rates varied little across industries. Public services were a notable exception. In all three downturns, workers laid-off from public services were at least 10-percentage-points more likely to be employed shortly after layoffs than workers in primary industries and construction. Further research is needed to uncover other sources underlying this difference. ²⁵ 

As expected, workers who were laid-off in the fourth month or the fifth month of the LFS interview were less likely to have a paid job during the sixth month than their counterparts laid-off during the second month of the LFS interview.

Different downturns affected regions differentially. Among comparable employees, the likelihood of being employed shortly after a layoff was consistently lower for workers laid-off in the Atlantic Provinces than for their counterparts laid-off in Ontario. However, the difference narrowed over time. In contrast, workers laid-off in the Prairie Provinces were more likely to be employed in the short term than those laid-off in Ontario in 2008-to-2011, although this was not the case during the early 1980s.

6   How do wages before and after layoffs compare?

Since LFS information on wages is available only for 1997 and subsequent years, the question of how wages before and after layoffs compare is examined only for 2006-to-2008 and 2008-to-2011.

Workers who were laid-off during the most recent downturn and found a job shortly thereafter experienced, on average, a slight drop in employment income. Average weekly wages declined from $734 to $703 (in 2008 dollars) (Table 7), and the average hourly wage fell from $20.9 to $20.4. Fairly similar declines are observed in 2006-to-2008.

Yet these averages mask considerable heterogeneity in wage changes (Figure 1). During the most recent downturn, one-quarter of re-employed laid-off workers saw their weekly wages fall by 23% or more (Table 8, Panel A). Another one-quarter registered increases in weekly pay of at least 18%. Similarly, one-quarter of these workers saw their hourly wage drop by at least 13%, while another one-quarter experienced hourly-wage increases of at least 11% (Table 8, Panel A).

Figure 1: Kernel density of log hourly-wage change and log weekly-wage change (2008 to 2011)

Measuring wage movements by using average changes in log wages, Table 9 shows that average wage changes were fairly similar in 2006-to-2008 and 2008-to-2011.

On average, managers, workers with more than 20 years of seniority, and those laid-off from high-skill services experienced weekly wage losses of at least 10% during the most recent downturn. In contrast, workers laid-off from retail trade, accommodation, and food services experienced gains in weekly pay of about 14%.

Workers who lost union coverage while moving across jobs experienced, during the most recent downturn, average hourly-wage losses of 16% and average weekly-wage losses of 17% (Table 10). ²⁶  Workers who moved from a firm with at least 100 employees to a smaller firm had hourly-wage and weekly-wage losses that averaged 11% and 15%, respectively. Workers who changed both occupation and industry saw their weekly wages fall by 10%, on average. ²⁷  In contrast, employees who gained union coverage or moved to firms with 100 or more employees registered average gains in weekly wages of between 8% and 11%. ²⁸ 

To assess the degree to which average wage losses vary with worker and job characteristics, changes in log hourly wages and changes in log weekly wages are regressed on worker attributes (age, sex, education, and seniority), a binary indicator for whether workers expect to be recalled, interaction terms between this indicator and seniority, a binary indicator capturing the end of temporary, term, or contract jobs, as well as transition-related variables. The results for 2008-to-2011 are shown in Table 11. ²⁹  Text Table 5 presents the same analysis for the 2006-to-2008 period.

Table 11 generally reveals no robust association between worker attributes (including education) and wage changes; this is consistent with Table 9. Two exceptions are the fact that, all else equal, re-employed women experienced smaller wage losses than men ³⁰  and that senior workers (with more than 20 years of job tenure) experienced larger wage losses than their counterparts with two years or less of seniority.

In contrast, transitions across job types account for part of the observed wage changes. For instance, workers who lost union coverage and workers who moved from a firm with 100 employees or more to a smaller firm experienced hourly-wage losses that were between 9-percentage-points and 13-percentage-points larger, and weekly-wage losses that were about 10-percentage-points larger, than those of workers who remained non-unionized and those of workers who remained employed in smaller firms. The substantial wage losses associated with loss of union coverage are in line with the results of Kuhn and Sweetman (1998).

The net result, conditioning on the average values of other covariates, is that expected declines in weekly wages were, at 11% to 14%, quite substantial for each of the following three groups: workers who lost union coverage; workers who moved from a firm with 100 employees or more to a smaller firm; and workers who changed industry and occupation (Table 12). ³¹  In contrast, the expected increases in weekly wages amounted to at least 4% for employees who gained union coverage or moved from firms employing fewer than 100 workers to firms with 100 or more employees.

7   Conclusion

Over the last three decades, Canada has experienced three recessions. As a result of these, unemployment rates rose sharply in 1981-to-1983, 1990-to-1992, and, most recently, after October 2008. To shed light on the labour market implications of these slowdowns, this study attempts to answer four key questions:

1. Which workers were laid-off during these downturns?
2. How did layoff rates vary across downturns?
3. How did chances of finding employment shortly after being laid-off evolve across downturns?
4. Among workers who managed to find paid employment shortly after being laid-off, how do wages before and after layoffs compare?

The findings of this study are the following:

1. Compared with their counterparts who were permanently or temporarily laid-off during the early 1980s or early 1990s, Canadian workers laid-off during the most recent recession had greater seniority, were older, were better educated, and were less likely to come from the manufacturing sector. Except for seniority, these temporal changes in the profile of laid-off workers resulted mainly from compositional effects, i.e., changes in the age/education profile of the Canadian workforce as well as the secular decline of the manufacturing sector.
2. Assessed on a monthly basis, the risk of layoff during the most recent recession was, at 2.0%, lower than the rates of 2.9% and 2.7% observed during the early 1980s and the early 1990s, respectively.
3. Of all workers laid-off in 2008-to-2011, half of them found a paid job between one and four months after being displaced. This number is significantly higher than the corresponding proportion of 42% observed during the previous two recessions.
4. On average, employees who were laid-off during the most recent recession and who found a job shortly after being laid-off experienced a slight drop in employment income. However, one-quarter saw their weekly wages drop by 23% or more while another one-quarter saw increases in weekly pay of at least 18%. Average declines in weekly wages amounted to at least 10% for re-employed workers who lost union coverage, moved from a firm with at least 100 employees to a smaller firm, or changed both industry and occupation in the new job. Collectively, these groups represented about one-quarter of laid-off workers who were re-employed during the most recent recession. In contrast, employees who gained union coverage or moved from firms with fewer than 100 employees to firms with 100 or more employees registered average gains in weekly wages that amounted to between 8% and 11%. Collectively, these groups represented about 17% of laid-off workers who were re-employed during the most recent recession.

Consistent with Farber (2005) and Riddell and Song (2009), the study showed that having a university degree was generally associated with a greater likelihood of being employed shortly after being laid-off. However, conditional on being re-employed shortly after the layoff, holding a university degree was not associated with smaller wage losses.

Workers laid-off in 2006-to-2008 and 2008-to-2011 experienced very similar changes in (hourly and weekly) wages. Thus, conditional on their being re-employed in the short term, the impact of layoffs on pay changes did not differ much across these two adjacent periods.

While the study documented short-term wage changes for the one-half of laid-off workers who found a paid job in the first few months after being laid-off during the most recent downturn, it neither distinguished temporary layoffs from permanent layoffs nor assessed the long-term wage impact of permanent layoffs. Recent research has found that high-seniority workers involved in mass layoffs experience substantial earnings losses for five years after losing their jobs (Jacobson, Lalonde, and Sullivan 1993; Couch and Placzek 2010; Morissette, Zhang, and Frenette 2007). Since a fairly high proportion (28%) of workers laid-off (temporarily or permanently) during the most recent recession had more than five years of seniority, whether this will be the case for those permanently laid-off during the employment downturn that started in October 2008 is an important question for future research.