Executive summary

View the most recent version.

Canada's labour market in the 1980s and 1990s experienced major changes associated with an increasing integration with the U.S. economy and shifting trade flows, rapid advances in information technology, much increased use of non-standard work patterns in production, increased international competition and worker vulnerability, and worker-skill upgrading and high inflows of immigrants. At the macroeconomic level, the economy experienced a severe recession in the early 1990s, with a slow recovery in the labour market through to the mid-1990s. These developments set the scene for possibly major changes to the distribution of workers' earnings in the labour market.

This paper examines the variability of workers' earnings in Canada over the 1982-to-2000 period with a graphical descriptive approach using Statistics Canada's Longitudinal Administrative Data base file. Following a methodology from Gottschalk and Moffitt 1994, we decompose the total variance of workers' earnings over this period into a 'permanent' or long-run component between workers and a 'transitory' or year-to-year earnings instability component over time for given workers. The former component is related to life-cycle earnings patterns, or profile levels, and it is affected by enduring human capital and skill differentials, long-run labour force attachment, shifting returns to skill, and possible discrimination and cohort effects. It is a measure of long-run earnings inequality across workers. The latter component of year-to-year fluctuations in earnings—net of systematic life-cycle earning profiles—is related to business-cycle effects, workplace restructuring and changing industry-demand patterns. It is a measure of average-earnings instability for workers. This decomposition allows one to better interpret and evaluate alternative explanations for observed earnings-inequality changes (largely, increases) that the Canadian labour market has been experiencing.

The decomposition is applied to a five-year moving window of earnings in order that we can examine—through graphs and regression techniques—how these two earnings components have changed over the 1982-to-2000 period and to see how these changes are linked to macroeconomic indicators. We report results not just for males as a whole—as is typical in the literature—but separately for men and women and also for four separate age groups of workers.

Several major results are found. First, the general rise in total earnings variance over the period was not a steady increase, and it reflects quite different patterns of change for its separate components. Our results are more reflective of strong cyclical effects than they are of a dominating upward trend, in contrast to the situation in the United States. Long-run earnings inequality has generally increased over the period, while year-to-year earnings instability has pretty steadily decreased, except during the early 1990s recession. Changes in total earnings variability have been driven primarily by changes in long-run earnings inequality.

Second, the patterns of change in the two variance components showed substantial differences between men and women. Since the early 1990s, long-run earnings inequality continued to rise for men, but it markedly decreased for women. Since the later 1980s, earnings instability fell quite steadily for women, but it showed a more cyclical pattern for men. As a result, underlying trends in earnings instability reinforced the rising trends in long-run earnings inequality for men, but they weakened or countered the latter effect for women.

Third, the patterns across ages of the two variance components are almost opposite. Long-run earnings inequality generally rises with age, at least for younger workers and beyond, so that it is markedly highest among the older-age workers. Earnings instability, in contrast, generally declines with age, at least until workers reach prime age, so that earnings instability is markedly highest among entry-age workers.

Fourth, both unemployment rates and gross domestic product (GDP) growth rates, as macroeconomic indicators, have statistically significant net-regression effects on all the earnings variance measures. Unemployment-rate effects are positive on almost all variance measures, which are consistent with conventional expectations that tighter labour markets reduce earnings variances, while higher unemployment is associated with widened long-run earnings differentials and greater earnings instability. In proportional terms, the earnings instability impacts of unemployment are generally stronger. The GDP growth-rate effect would conventionally be expected to be negative on the different variance measures, as greater economic growth reduces earnings inequality and instability. Such estimated effects indeed hold for women and for earnings instability among men. Growth-rate effects on men's earnings inequality, however, are found to be positive—more consistent with an alternative 'new economy' set of explanations based on economic restructuring and changing demographics.