Page content follows

Canada's employment downturn

View the most recent version.

By Sébastien LaRochelle-Côté and Jason Gilmore

Full article in PDF

• Net loss of 400,000 jobs since October 2008
• Employment changes across individual characteristics
• Employment changes across job characteristics
• Comparisons with earlier recessions
• Canada–U.S. comparisons
• Summary
• Data source and definitions

Adapted from "Canada's employment downturn: October 2008 to October 2009," Canadian Economic Observer, and accompanied by a short article in The Daily.

For an extended period of time until October 2008, employment levels were at an all-time high and unemployment rates were near historic lows in Canada. In the months that followed, a sudden downturn in the world economy caused widespread employment losses for the first time since the 1990/92 recession. Since many of these jobs were lost in the early months of the recession, many observers were concerned about the severity of the recession.

One year later, the perspective changed somewhat. Employment losses moderated in the second half of the year with declines in some months offset by gains in other months. Still, questions remained about the effects of the downturn on some specific groups.

This report examines year-over-year changes in employment levels (between October 2008 and October 2009) across demographic groups, various types of families, and associated job characteristics. It also compares how this 12-month period stacks up against the first 12 months of the Canadian recessions of the early 1980s and early 1990s. (see Data source and definitions). The employment situations in Canada and United States are also compared.

Results indicate that not all groups were equally affected by employment losses and that some groups even reported gains. Comparisons with earlier recessions indicate that although job losses were steep in the early months of the downturn, employment levels stabilized earlier than in previous recessions.

Net loss of 400,000 jobs since October 2008

In October 2009, employment in Canada was down 400,000 from the peak in October 2008, a loss of 2.3% in seasonally adjusted figures.¹ During the same period, the unemployment rate rose from 6.3% to 8.6%. Previous monthly releases have shown important variations across age groups, industries and regions.

One key feature of the downturn is that younger individuals and men from age 25 to 54 have been more affected by job losses (Table 1). Between October 2008 and October 2009, employment declined by 10.8% among young men under 25, and by 6.5% among women in the same age group. Men in their prime working years (25 to 54) were also affected as employment declined by 3.3% over the period for men in this age group. However, gains were seen among those 55 and over, especially for women, among whom employment increased by 6.0%.

Another well-known feature of this recession is that some industrial sectors—particularly manufacturing, construction, natural resources, transportation and warehousing, and retail and wholesale trade—have been more affected than others. Manufacturing industries, in particular, declined by 218,000 between October 2008 and October 2009, accounting for over one-half of the net decline in employment over the period.

Manufacturing has received more attention than some other industries for reasons other than the scale of the job losses. First, the declines in this sector began much earlier. Manufacturing employment fell by 555,900 between 2004 and 2009. Thus the current downturn merely accelerated a long-term trend in that industry (Chart A). Second, while losses in most other industries were concentrated in the first five months of the recession, employment declines in manufacturing continued into subsequent months.² This complements the findings of other studies focusing on the manufacturing sector (Bernard 2009).

The effects of the downturn varied across the country (Table 2). With a decline of 205,900 (or -3.1%) over the 12 months, the province of Ontario experienced the greatest absolute employment losses, a fact likely associated with the higher concentration of manufacturing industries in that province. Proportionately, however, Alberta experienced the largest losses (-3.3%). In contrast, employment declined much more modestly in the Atlantic provinces (-0.8%) and remained relatively stable in Manitoba and Saskatchewan over the period. Losses in Quebec (-1.6%) were slightly below the Canadian average, and British Columbia (-2.2%) had employment declines similar to Canada as a whole.

While the age, geographic and industrial dimensions of the downturn are well-known, questions remain about the impact on other population groups. In previous economic cycles, specific demographic groups and types of jobs were more affected by downturns.

Employment changes across individual characteristics³

Previous studies have shown that higher levels of education have been associated with more stable employment during previous economic cycles (Picot and Heisz 2000). The current downturn is no exception.

Between October 2008 and October 2009, core working-age men with a high school education or less experienced the greatest employment losses (-5.2%), since many were previously employed in industries like manufacturing and construction (Table 3). Women with a high school education or less also experienced relatively high job losses (-3.6%).

As previous studies indicate, the number of employees was more stable among workers with higher educational attainment. Some job gains were seen among women with a college education (+0.9%) and small losses were observed among men and women with university degrees (-0.6% and -1.2% respectively).

Recent reports have documented the relative deterioration in the economic outcomes of immigrants, especially recently landed immigrants (see Picot 2008 for a review of these studies). The situation is similar in this downturn as employment declined faster for immigrants who landed within the last five years (-12.9%) than for the Canadian-born (-2.2%). Again, the bulk of the losses for these immigrants occurred in the manufacturing industry. On the other hand, immigrants who had been in Canada for more than five years experienced much smaller losses than the Canadian-born over the 12-month period.

Among Aboriginal peoples age 25 to 54 (excluding those living on reserves), the pace of employment losses during this 12-month period was double that of the non-Aboriginal population (-4.0% vs. -1.9%). Worthy of note is the fact that Aboriginal peoples living off-reserve continue to have higher unemployment rates and lower employment rates than non-Aboriginal peoples.

The effects of the downturn also differed by family type (Table 4). Youth employment in all families was particularly affected by this downturn. Two-parent families with younger children were notably affected over this 12-month period, as employment fell by 2.5% among mothers and 2.4% among fathers in two-parent families with at least one child under age 18. In the first 12 months of the previous two downturns, the fathers of young children experienced more significant declines in employment than mothers.

Single mothers with younger children also experienced high rates of losses as their employment levels fell by 6.8%. Conversely, single fathers with younger children had an employment gain of 4.6% over the period.⁴ These recent changes in employment for both single mothers and single fathers are consistent with what occurred during the first 12 months of the previous two downturns.

Employment growth among individuals in 'other economic families' (e.g., adult siblings living together, an older parent living with an older child) was influenced by an increase in the number of individuals joining such families over this one-year period.

Employment changes across job characteristics

Other studies have shown that a period of employment downturn is typically associated with compositional changes in job type. One such example is self-employment, which tends to increase during periods of economic hardship (Picot and Heisz 2000).

Since October 2008, the number of those who were self-employed in their main job increased by 3.9%, spurred by significant growth after the first seven months of the downturn (Chart B). ⁵ Conversely, main-job employment among both private sector and public sector employees fell at roughly the same pace during the first few months of the downturn. In the seven months since then, the number of public sector employees remained stable while private sector employment continued to fall. The private sector trend reflects continuing difficulties in manufacturing, construction, transportation and warehousing.

The extent of employment losses also varied considerably by hours of work, tenure, job status, unionization and wage category (Table 5).⁶

From the beginning of the downturn, losses in full-time employment were significant (-2.2%), and larger than among part-timers (-1.6%). Declines among those with longer hours—that is, 40 or more hours (-4.6% and -4.5% respectively) were especially significant. Conversely, the number of employees with a shorter full-time schedule—between 30 and 34 hours—rose over the period (+8.2%). This decline in longer hours and growth in shorter full-time schedules is consistent with changes in hours during the first 12 months of the previous two downturns. These changes may not be exclusively the result of job losses, as they could also be the result of reduced work hours among employed workers.

Employment losses were also concentrated among permanent employees. From October 2008 to October 2009, the number of permanent employees declined by 3.8%, while the number of temporary employees increased by 0.7%.

Workers a with short employment tenure were also significantly affected by the downturn, as employment declined by 662,700 (-17.8%) among those who had a tenure of one year or less. Conversely, there was an increase (+4.2%) in the number of workers among workers who had 1 to 5 years in their current jobs, and little change in the number of workers with more than 5 years in their current jobs. The extent of the losses likely reflects both the loss of employment among short-tenured positions and the lack of hiring.

Non-unionized workers were proportionately more affected by employment declines (-4.0%) than unionized workers (-1.7%) between October 2008 and October 2009. This reflects the concentration of union jobs in the more stable public sector.

Studies have shown that periods of economic decline can alter the distribution of earnings (Heisz et al. 2002). Employees earning less than $10 per hour saw the largest decline in employment over the period (-24.8%), followed by those who earned $10.00 to $19.99 per hour (-2.2%). Among those earning less than $10, employment losses were largely concentrated in manufacturing, wholesale and retail trade, and accommodation and food services. The large loss of low-wage and short-tenured jobs is consistent with the particular difficulties noted for younger workers and very recent immigrants since they are overrepresented in these types of jobs.

Meanwhile, the number of employees who earned $30 or more per hour grew—especially those earning at least $40 per hour (+12.9%). Women accounted for two-thirds of the increase in those earning at least $40 per hour, particularly those working in industries such as health care and social assistance, educational services, and public administration, as well as finance, real estate, rental and leasing.

Comparisons with earlier recessions

In this section, recent employment trends are compared with two previous downturns (based on seasonally adjusted figures). More specifically, the number of jobs just before the downturn is indexed to 100 and then tracked for the first 12 months of the three most recent employment downturns: June 1981 to June 1982, April 1990 to April 1991, and October 2008 to October 2009.

Employment declined much faster in the early months of the current downturn compared with the first few months of the 1981 and 1990 recessions (Chart C). Five months after the October 2008 peak, employment had fallen by 2.1%, compared with 0.8% in 1981and 0.6% in 1990.

On the other hand, employment levels began to stabilize after the first 5 months of the current recession, while employment losses after the peak lasted 17 months in 1981/82 and 11 months in 1990/91. As a result, the job losses after 12 months were similar in proportion to the previous recession of the 1990s (-2.3%), and proportionately smaller than the recession of the 1980s (-3.9%). Even though such results might suggest that the labour market is getting back on track faster than in earlier recessions, history indicates that employment recovery is not always a smooth upward path. For example, in the downturn of the early 1990s, the first 11 months of employment declines were followed by 6 months of modest growth, only to be followed by another 7 months of declines.

Canada–U.S. comparisons

Comparisons with employment losses sustained by Canada's major trading partner, the United States, are also of interest due to the high volume of trade between the two countries. Employment estimates from the two countries cannot be directly compared because of differences in survey design, but some comparisons can be made using unemployment rates⁷ (Chart D). Since employment in the United States last peaked in December 2007, conceptually comparable unemployment rates for both countries are examined for the period between December 2007 and October 2009.

During the first six months of 2008, Canadian and American unemployment rates were almost at parity. Shortly thereafter—and for the first time since 1982—the U.S. unemployment rate surpassed the rate in Canada as the recession began to have a strong impact on the U.S. labour market. Since the beginning of the downturn in Canada, the unemployment rate also increased in Canada, but at a slightly slower pace than the United States. As a result, the Canadian rates have remained consistently below American figures since May 2008. During the previous two recessions, the Canadian labour market experienced the larger increase in unemployment rates.

It should be noted that the higher U.S. rate is related to greater job losses in financial, professional and business industries. According to the U.S. Current Employment Statistics (CES) survey, the financial and business sector accounted for nearly 25% of all job losses south of the border between October 2008 and October 2009.⁸ In comparison, the number of jobs in these industries rose in Canada during that period, albeit modestly.

Summary

For the first time since the 1990/92 recession, employment declined by significant margins in Canada. Since employment last peaked in October 2008, it subsequently declined by 2.3%, or 400,000 individuals. While many facts about the recession are relatively well-known—including larger employment declines among youth, men and workers in manufacturing industries—a series of questions remain about employment losses among other groups of workers and types of jobs.

Since the last employment peak in October 2008, it is now possible to examine annual variations in employment levels for a wider variety of population groups without having to deal with seasonal variation issues. In this report, year-over-year changes in employment levels were examined across a variety of personal, family and job characteristics. Comparisons with previous downturns and with the recent evolution of the U.S. labour market were also presented.

Employment losses in the current downturn were concentrated at the low end of the pay and tenure scale, thus disproportionately affecting those who tend to hold these jobs. Heavy employment losses were noted for very recent immigrants, young workers and those with lower levels of education. Other demographic groups were also proportionately more affected by losses: lone mothers, parents of younger children and non-unionized workers.

Despite the concentration of employment losses at the bottom of the pay scale, jobs typically not seen as 'vulnerable' were also disappearing. For example, employment declined faster among individuals working more than 40 hours per week and among permanent workers. And the loss of manufacturing employment that began in 2004 accelerated in the 12-month period from October 2008 to October 2009. On the other hand, the number of jobs with very high rates of pay increased over this period.

Results also indicate that this downturn differs from the previous ones in at least two ways. First, even though employment declined faster during the first few months than in previous downturns, it stabilized sooner in the current recession. As a result, employment losses after 12 months were similar in proportion to those in the early 1990s downturn and proportionately smaller than those in the early 1980s downturn. Second, the U.S. labour market was affected earlier, and continues to be in a deeper slump compared to Canada. In May 2008, the U.S. unemployment rate surpassed the Canadian rate for the first time since 1982 and that gap has yet to close.

Data source and definitions

This study uses data from the Labour Force Survey (LFS). The LFS is conducted every month to collect information about the labour market activities of the population at least 15 years of age, excluding residents of collective dwellings, persons living on reserves and other Aboriginal settlements, and full-time members of the Canadian forces. Employed individuals are defined as those who had a job during the reference week of the survey.

According to the Labour Force Survey, employment peaked in October 2008 in Canada. In the LFS, employment estimates for some demographic groups and job characteristics are not seasonally adjusted. A detailed study of employment changes since the peak therefore had to wait until the release of October 2009 data because year-over-year variations are less likely to be affected by the seasonal adjustment process.

Employment 'changes' cannot be interpreted as the total number of jobs lost during the recession. LFS employment changes should be interpreted as net changes in employment levels since they represent the differences between all losses and gains over the period.

---

Notes

1. Data not seasonally adjusted declined by 2.1%, or 360,000.
2. Losses have been particularly significant in transportation equipment manufacturing, furniture and related product manufacturing, fabricated metal product manufacturing, computer and electronic product manufacturing, and paper manufacturing.
3. In this section, employment changes are examined for prime-age workers only because overall results for personal characteristics tend to be disproportionately affected by the age composition of individuals within groups. The data have not been adjusted for seasonal variations. Although this affects absolute employment variation figures, changes in percentage terms are barely affected.
4. The sample size for lone fathers is relatively small.
5. Chart B is based on seasonally adjusted figures.
6. The numbers in Table 5 are not seasonally adjusted.
7. The Canadian unemployment rates have been adjusted to ensure that they are based on the same population covered by the Current Population Survey, the American equivalent of the Labour Force Survey.
8. The CES collects information about non-farm employment on a monthly basis. Results for October 2009 are based on preliminary data.

References

Authors

Sébastien LaRochelle-Côté is with the Labour and Household Surveys Analysis Division. He can be reached at 613-951-0803. Jason Gilmore is with the Labour Statistics Division. He can be reached at 613-951-7118. All authors can be reached at perspectives@statcan.gc.ca.