6.4 Labour Market Experience: Canada

So far we have developed the theory behind the natural rate of unemployment. We began by showing that the economy’s steady-state unemployment rate depends on the rates of job separation and job finding. Then we discussed two reasons why job finding is not instantaneous: the process of job search (which leads to frictional unemployment) and wage rigidity (which leads to structural unemployment). Wage rigidity, in turn, arises from minimum-wage laws, unionization, and efficiency wages.

With these theories as background, we now examine some additional facts about unemployment, focusing at first on the case of the Canadian and American Labour markets. These facts will help us to evaluate our theories and assess public policies aimed at reducing unemployment.

The Duration of Unemployment

When a person becomes unemployed, is the spell of unemployment likely to be short or long? The answer to this question is important because it indicates the reasons for the unemployment and what policy response is appropriate. Economists decompose the unemployment rate into two components: incidence and duration. Incidence is the likelihood that an individual suffers an unemployment spell, and duration is the average length of that spell. If the unemployment rate is 5 percent, it could be that duration is high and incidence is low (say, for example, if 5 percent of the labour force is permanently unemployed), or it could be that incidence is high and duration is low (for example, if everyone is unemployed for 5 percent of the year). In this second case, since unemployment is short-term, one might argue that it is frictional and perhaps unavoidable. Unemployed workers may need some time to search for the job that is best suited to their skills and tastes. On the other hand, long-term unemployment cannot easily be attributed to the time it takes to match jobs and workers: we would not expect this matching process to take many months. Long-term unemployment is more likely to be structural unemployment representing a mismatch between the number of jobs available and the number of people who want work. Thus, data on the duration of unemployment can affect our view about the reasons for unemployment.


Estimates on the duration and incidence of unemployment have changed over the years. In 1993, Statistics Canada estimated that about two-thirds of any increase in the unemployment rate was due to longer duration and one-third to higher incidence. But more recently it has been estimated that most of the variation in structural unemployment can be attributed to changes in the incidence of unemployment over time and that the average duration of an unemployment spell has remained remarkably constant at about 2.3 months.11

The confusing evidence on the duration of unemployment has an important implication for public policy. If the goal is to lower substantially the natural rate of unemployment, policies must aim at the long-term unemployed, because these individuals account for a large amount of unemployment. Yet policies must be carefully targeted, because the long-term unemployed still constitute a minority of those who become unemployed at some point during the year. It is still the case that most people who become unemployed find work within a fairly short time.

Variation in the Unemployment Rate Across Age Groups and Regions

The rate of unemployment varies substantially across different groups within the population. Table 6-2 presents the unemployment rates for different age groups in 2012. While the country’s overall unemployment rate was 7.3 percent, this unemployment was not equally distributed by age. The unemployment rate for younger Canadians was more than double that of people in the 55-and-over age category. While younger Canadians have always had higher unemployment rates than prime-aged adults, they now account for a smaller proportion of unemployed people than they did three decades ago. The aging of the Canadian population has resulted in a change in the age structure of unemployment. In 1976, almost half of all unemployed people were 15 to 24 years old, compared with just 26% in 2012. Not surprisingly, education affects employment prospects as well. Over the last 40 years, the unemployment rate for those with only primary-level education has been almost three times that facing those with a university education. Most analysts expect this gap to widen in years to come.

TABLE 6-2 Unemployment by Age Groups: December 2012
Age Unemployment Rate (%)
15–24 12.0
25–54 5.6
55 and over 5.8

Source: Statistics Canada, The Canadian Labour Market at a Glance Catalogue No. 71-222-X.


To explain the differences in unemployment across age groups, recall our model of the natural rate of unemployment. The model isolates two possible causes for a high rate of unemployment: a low rate of job finding, or a high rate of job separation. When economists study data on the transition of individuals between employment and unemployment, they find that those groups with high unemployment tend to have high rates of job separation. They find less variation across groups in the rate of job finding.

These findings help explain the higher unemployment rates for younger workers. Younger workers have only recently entered the labour market, and they are often uncertain about their career plans. It may be best for them to try different types of jobs before making a long-term commitment to a specific occupation. If so, we should expect a higher rate of job separation and a higher rate of frictional unemployment for this group.

Finally, unemployment varies significantly across Canadian regions. Again considering 2012, the unemployment rate in Newfoundland and Labrador was 12.5 percent (almost twice the national average), while in Alberta the unemployment rate was 4.6 percent. This regional variation helps explain why it is difficult to reform the employment-insurance system so that it operates as a true insurance policy instead of being a means for performing ongoing income redistribution. Without some other policies to support those in the depressed regions, many policymakers oppose changes and cuts in the current employment-insurance system. There is a certain irony in this outcome. It is concern for the welfare of Canadians living in the depressed regions in the short run that limits reforms to employment insurance from being implemented. Nevertheless, it can hurt the welfare of those same individuals in the long run if employment insurance is not reformed. Without reform, these regions can never escape the long-run dependency trap.

Trends in Unemployment


Over the past 50 years, the natural rate of unemployment in Canada has not been constant. If you look back at Figure 6-1, you will see that unemployment averaged well below 5 percent in the 1950s and 1960s, and above 9 percent in the 1980s and 1990s. The average unemployment rate has drifted back down somewhat in recent years. Although economists do not have a conclusive explanation for these changes, they have proposed various hypotheses.

Demographics One explanation stresses the changing composition of the Canadian labour force. After World War II, birth rates rose dramatically, producing a baby-boom generation that began entering the labour force around 1970. Because younger workers have higher unemployment rates, this influx of baby boomers into the labour force increased the average level of unemployment. At roughly the same time, the participation of women in the labour force also was increasing significantly. Since women historically have had higher unemployment rates than men (a difference that has disappeared since the early 1980s), the increasing proportion of women in the labour force may have raised the average unemployment rate, at least in the 1970s.

These two demographic changes, however, cannot fully explain the upward trend in unemployment because the trend also was apparent for fixed demographic groups. For example, for prime-aged males (men aged between 25 and 54 years), the average unemployment rate rose from below 3 percent in the 1950s and 1960s to 4.2 percent in the 1970s, 7.1 percent in the 1980s, and 9.1 percent in the 1990s.

Another point to bear in mind is that some of the demographic shifts and microeconomic rigidities that can explain rising unemployment in the 1970s have since reversed—without a corresponding decrease in unemployment. For example, as a proportion of the population, the 15–24 age group dropped from 26 percent of the population in the mid-1970s to under 18 percent in 1992. Also the ratio of the minimum wage to the average hourly wage available in Canada fell by 15 percent over this same period.12

Sectoral Shifts A second possible explanation for the upward trend in unemployment is that sectoral shifts have become more prevalent. The greater the amount of sectoral reallocation, the greater the rate of job separation and the higher the level of frictional unemployment.13 One source of sectoral shifts has been the increased pace of technological change involving robotics and major breakthroughs in the methods of information storage and transfer. Another series of sectoral shifts has resulted from the major shifts in natural resource prices. The world price of oil has been far more volatile since the mid-1970s, and the relative price of most of the natural resources that Canada exports has changed appreciably over the decades.


Productivity A third possible explanation for rising unemployment emphasizes the link between unemployment and productivity. As Chapter 8 discusses more fully, the 1970s experienced a slowdown in productivity growth, and perhaps this raised the natural rate of unemployment. Why such an effect would occur, however, is not obvious. In standard theories of the labour market, lower productivity means reduced labour demand and thus lower real wages, but unemployment is unchanged. This prediction is consistent with the long-term data, which show consistent upward trends in both productivity and real wages, but no trend in unemployment. Yet suppose that workers are slow to catch on to news about productivity. When productivity changes, workers may only gradually alter the real wages they ask from their employers, making real wages sluggish in response to labour demand. Thus, a falloff in productivity growth, such as that experienced during the 1970s, will decrease labour demand and, with workers’ real wage claims sluggish—that is, remaining high for a prolonged period—the decrease in labour demand must show up in the form of rising unemployment.14

Another consideration is skill-biased technical change. There is no question that the demand for unskilled workers has fallen relative to the demand for skilled workers. This change in demand is probably due to changes in technology: computers, for example, increase the demand for workers who can use them while reducing the demand for those who cannot. In the United States, this change in demand has been reflected in wages rather than unemployment: over the past two decades, the wages of unskilled workers have fallen substantially relative to the wages of skilled workers. In Canada and Europe, however, public programs provide unskilled workers with an alternative to working for low wages. As the wages of unskilled workers fall, more workers may view employment insurance and welfare as their best available options. The result is higher unemployment. We consider policy options to combat skill-biased technical change in the appendix to this chapter.

In the end, the upward drift in the unemployment rate over the second half of the twentieth century is probably the result of all these unrelated developments operating at the same time. Luckily, unemployment started coming down at the turn of the century to 6 percent in 2007. Of course, with the financial crisis and the world-wide recession in 2008–09, Canada’s unemployment rate rose to 8.6 percent in Canada in mid-2009. However, most economists believe that this increase was due to the cyclical downturn and was not an increase in structural unemployment. It will take the passage of more time to be sure, but if this view proves correct, then perhaps the era of the long-term rise in joblessness is behind us.


Transitions Into and Out of the Labour Force So far we have ignored an important aspect of labour market dynamics: the movement of individuals into and out of the labour force. Our model of the natural rate of unemployment assumes that the labour force is fixed. In this case, the sole reason for unemployment is job separation, and the sole reason for leaving unemployment is job finding.

In fact, movements into and out of the labour force are important. The fact that individuals enter and leave the labour force makes the unemployment statistics more difficult to interpret. On the one hand, some individuals calling themselves unemployed may not be seriously looking for a job and perhaps should best be viewed as out of the labour force. Their “unemployment” may not represent a social problem. On the other hand, some individuals, called discouraged workers, may want a job but, after an unsuccessful search, have given up looking. These discouraged workers are counted as being out of the labour-force and do not show up in unemployment statistics. Even though their joblessness is unmeasured, it may nonetheless be a social problem.

It is important to have some idea of the magnitude of these unemployed who do not usually get measured. Periodically, Statistics Canada does a study that is much more extensive than the usual monthly labour-force survey. Through these interviews, they can estimate the number of discouraged workers. Officials find that when these discouraged individuals are transferred from the “not in the labour force” category to the “unemployed” category, the unemployment rate rises by 1 percentage point. A further dimension adds to the “disguised unemployment” problem. All individuals who want to work full-time but who only have part-time jobs are counted as employed. One adjustment of the figures that can be made to accommodate this practice is to count half of these involuntary part-timers as unemployed. When this adjustment to the figures is made, it adds another full percentage point to the unemployment rate.


Comparing Unemployment in the United States and Canada

Throughout the 1960s and 1970s, the United States and Canada had similar labour markets. The rates of unemployment in the two countries were about the same on average, and they fluctuated together. By 1980, the experiences of the two countries began to diverge. Unemployment became much more prevalent in Canada than in the United States. In the 1980s and 1990s, the Canadian unemployment rate averaged about 4 percentage points above the U.S. rate. More recently, the excess of the Canadian unemployment rate over its American counterpart has shrunk. Indeed, it vanished in 2008 when the U.S. unemployment rate moved above Canada’s by about one percentage point. Why has the historical Canadian record been so disappointing, and why has our relative performance improved in recent years?


Studies15 have concluded that the earlier 4-percentage-point gap can be decomposed into three components. One percentage point can be explained by the different ways in which unemployment is defined in the two countries. Two percentage points can be rationalized by the fact that the Canadian economy was operating at a lower level of capacity utilization than the United States throughout the 1980s and much of the 1990s. Finally, one percentage point can be attributed to a set of structural factors. We discuss each of these issues in turn.

The primary measurement issue concerns how actively individuals must be looking for work to be counted among the unemployed. In Canada, simply “looking at job advertisements” is sufficient, while a more active search is required in the United States. As a result, a number of individuals who are counted as “unemployed” in Canada are counted as “not in the labour force” in the United States. There are some other differences, but they are less important. For one thing, Canada excludes Aboriginals living on reserves from the monthly labour force survey, while the Americans do not exclude them. Since the incidence of unemployment is higher among this group, the Canadian practice leads to a lower measured unemployment rate. On the other hand, both countries exclude those in prisons from their labour force surveys. Since many of these individuals would be unemployed if they were not in these institutions and since a higher proportion of the population is in prison in the United States, this exclusion leads to Canada’s measured unemployment rate being higher. It turns out that, from an overall quantitative point of view, these latter two considerations approximately cancel off, so the one-percentage-point measurement difference is mostly due to the fact that Canada deems a “passive” search strategy to be sufficient to label an individual as unemployed.

Unemployment results from both structural factors (as discussed in this chapter) and cyclical factors (discussed in later sections, starting in Chapter 9). When our economy suffers a recession, unemployment rises temporarily above the natural rate. This is cyclical, not structural, unemployment. Compared to the United States, the Canadian economy was more recessed during the 1980s and early 1990s, and the result was a higher level of cyclical unemployment in Canada. The evidence suggests that this difference averaged about two percentage points. One of the main reasons for our more recessed economic conditions was our contractionary monetary policy. Canada entered this period with higher inflation than the United States, and our central bank attacked the problem vigorously. By raising interest rates to dampen spending, the Bank of Canada lowered the demand for labour in Canada. Since workers resisted wage cuts, this decreased demand took the form of layoffs, and a temporary but prolonged increase in cyclical unemployment was the result. After eliminating this difference in cyclical unemployment from the data, only a two-percentage-point gap between the two countries’ natural unemployment rates remained to be explained.

So, after both capacity-utilization differences and measurement differences are accounted for, we are left with a one-percentage-point gap in the natural unemployment rates of the two countries. Several considerations have been studied, and the changing roles of unions in the two countries is one possible explanation for this divergence. In the 1960s, about 30 percent of the labour force was unionized in each country. But Canadian labour laws did more to foster unionization than U.S. laws did. Unionization rose in Canada while it fell in the United States.


As one might have predicted, changes in real wages accompanied the change in unionization. The real wage in Canada increased by about 30 percent relative to the real wage in the United States. This evidence suggests that unions in Canada pushed the real wage further above the equilibrium level, leading to more structural unemployment.

The divergence in the two unemployment rates may also be attributable to the increase in the availability of employment-insurance benefits in Canada. Not only does employment insurance raise search times and the amount of frictional unemployment, but it also interacts with the effects of unionization in two ways. First, employment insurance makes unemployed workers more willing to wait for a high-wage job in a unionized firm rather than take a lower-wage job in a nonunion firm. Second, because employment insurance partially protects the incomes of unemployed workers, it makes unions more willing to press for high wages at the expense of lower employment.

In the 1980s and early 1990s Canada’s employment-insurance program was more than twice as generous as that in the United States. Our program is still about 60 percent more generous. Many economists believe that this difference, along with the fact that our program contains no experience-rating feature, is one of the factors that can explain the one percentage point of Canada’s “excess” unemployment rate that can be attributed to structural factors. Canada’s unemployment rate has risen less than that of the United States in the 2007–2009 period because short-run cyclical problems have been more pronounced south of the border. The financial crisis was far more severe in the United States, and our larger primary products sector—with its level of activity buoyed up by high commodity prices—kept employment higher here.16 image