The issue of productivity growth in Canada has received considerable attention reflecting its marked slowdown since the early 1970s and concerns about its implications for Canadian competitiveness. To better understand productivity developments in Canada, it is useful to decompose total factor productivity (TFP) into investment-specific productivity change (ISP) and technologically neutral productivity change (TNP). The gap in manufacturing productivity growth between Canada and the United States originates mostly in the strong performance of specific industries, such as electrical products and commercial and industrial machinery.

Abstract

The issue of productivity growth in Canada has received considerable attention reflecting its marked slowdown since the early 1970s and concerns about its implications for Canadian competitiveness. To better understand productivity developments in Canada, it is useful to decompose total factor productivity (TFP) into investment-specific productivity change (ISP) and technologically neutral productivity change (TNP). The gap in manufacturing productivity growth between Canada and the United States originates mostly in the strong performance of specific industries, such as electrical products and commercial and industrial machinery.

II. Explaining the Difference Between Canadian and U.S. Unemployment Rates1

1. Until the early 1980s, unemployment rates in Canada and the United States had been roughly similar for most of the postwar period. Over the period 1982–89, however, the Canadian unemployment rate was on average 2½ percentage points higher than in the United States, and this differential increased to about 4 percentage points in the 1990s (Figure 1). Initially, the differential was attributed to the deeper recession in Canada. However, its persistence over the last two decades in part reflects significant deviations in the structures of the two labor markets. Empirical evidence suggests that a key structural difference in the 1980s was Canada’s relatively more generous unemployment insurance system. In contrast, the widening of the differential in the 1990s appears to be largely attributable to relatively weaker growth in Canadian aggregate demand, which appears to have more than offset a decline in structural unemployment arising from the reforms of the employment insurance system.

Figure 1.
Figure 1.

Canada, the United States, and Europe: Unemployment Rates

(Percent)

Citation: IMF Staff Country Reports 2000, 034; 10.5089/9781451806922.002.A002

Sources: U.S. Department of Labor, Bureau of Labor Statistics, “Comparative Civilian Labor Force Statistics, Ten Countries, 1959–1998”; www.bls.gov/pub/special.requests/ForeignLabor/flslforc.txt.1/ Average of Germany, France, Italy, and the United Kingdom weighted by population.

2. Various studies suggest that the Canadian labor market appears to lie somewhere between the relatively rigid labor markets of Europe, and the more flexible labor market of the United States.2 Union density, the number of collective bargaining agreements, minimum wages in relation to average earnings, and the generosity of unemployment insurance and welfare benefits in Canada are all higher than in the United States, but are considerably lower than in Europe. Since the 1960s, the level of unemployment in Canada increased by more than in the United States, but by less than in Europe. Similarly, Canadian long-term unemployment has risen by slightly more than in the United States, but significantly less than in Europe.

3. Canada’s labor market, however, has demonstrated strong job creation. Employment in Canada expanded at an average annual rate of 3½ percent in the 1970s and at a 2 percent average annual rate in the 1980s before slowing to 1 percent per year in the 1990s (Table 1). The Canadian employment-to-population ratio also increased over the same period, but it has edged down in the 1990s (Figure 2). In contrast, job creation in Europe has remained sluggish, with employment rising only by about half of a percent annually in the 1980s and contracting slightly in the 1990s. Compared to the United States, employment growth in Canada rose faster in the 1960 and 1970s, but at a roughly similar rate in the 1980s, and slower in the 1990s.

Table 1.

Canada, the United States, and Europe: Labor Market Trends

article image
Source: U.S. Department of Labor, Bureau of Labor Statistics, “Comparative Civilian Labor Force Statistics, Ten Countries, 1959-1998”; www.bls.gov/pub/special.requests/ForeignLabor/flslforc.txt.

Average annual percentage change. Data for Europe are 1991-98.

Average of Germany, France, Italy, and the United Kingdom, weighted by population.

Figure 2.
Figure 2.

Canada, the United Statess, and Europe: Employment-Population Ratios

(Percent)

Citation: IMF Staff Country Reports 2000, 034; 10.5089/9781451806922.002.A002

Sources: U.S. Department of Labor, Bureau of Labor Statistics, “Comparative Civilian Labor Force Statistics, Ten Countries, 1959-1998”; and www.bls.gov/pub/spccial.requests/ForeignLabor/flslforc.txt.1 Average of Germany, France, Italy, and the United Kingdom weighted by population.

4. Methodological differences in measuring the labor force, and hence unemployment, is one possible explanation for the differential in the Canadian and U.S. unemployment rates. Although both Canada and the United States adhere to international guidelines for measuring unemployment, there are some discrepancies that affect the comparability of unemployment rates. Most significant is the distinction between passive and active job search. Passive job seekers, (i.e., persons who respond to unemployment surveys as “looking at job ads in newspapers” as their only means of seeking employment) are classified in Canada as part of the labor force and hence unemployed; whereas in the United States, they are classified as out of the labor force, since the U.S. definition requires active job search. Sunter (1998) modifies the Canadian unemployment rate to reflect U.S. concepts and definitions, and finds that these adjustments reduce the unemployment differential by an average of 0.3 percentage point (or 15 percent of the differential) in the 1980s, and 0.7 percentage point (or 18 percent of the differential) in the 1990s (Figure 3). These results suggest that measurement issues play a role, but explain a relatively small portion of the unemployment differential.

Figure 3.
Figure 3.

Canada and the United States: Official and Modified Unemployment Differential1

(Percentage points)

Citation: IMF Staff Country Reports 2000, 034; 10.5089/9781451806922.002.A002

Source: D. Sunter, 1999, “Canada-U.S. Labour Market,” Canadian Economic Observer, December.1 Canadian unemployment data modified to be equivalent to U.S. data; adjustments include removal from Canadian data of 15 year olds, passive job searchers, and persons unavailable because of personal responsibilities, and the addition of full-time students looking for full-time work.

5. The primary explanation for the emergence of the unemployment differential appears to reflect a change in how Canadian workers behaved when not employed compared to their U.S. counterparts. Riddell (1999) and Card and Riddell (1995) decompose movements in the rate of unemployment into changes in: labor force attachment of the nonemployed (i.e., the probability that a person who is not employed will be included in the labor force), non-employment, and labor force participation.3 In comparison to the 1970s, nonemployed Canadians during the 1980s were more likely to be included in the labor force and be classified as unemployed, whereas Americans were more likely to be classified as out of the labor force. This change in behavior accounts for 80–90 percent of the 2 percentage point unemployment differential.4 The remaining 10–20 percent of the differential is attributable to changes in labor force participation and employment rates. The relative rise in labor force attachment was particularly important for women, explaining 99 percent of the unemployment differential for the group.

6. For the 1990s, this analysis reveals a very different picture. In the early 1990s, Canada experienced a longer and deeper recession than did the United States. Most of the further widening of the unemployment differential in the 1990s was related to the different cyclical positions of the two economies. The increase in the unemployment differential is attributable to the relative decline in employment and labor force participation in Canada, rather than to a widening in the differential in labor force attachment of the nonemployed as was the case in the 1980s.5

7. Although it is generally agreed that the growing generosity of Canada’s employment insurance system was a major factor in explaining the increase in the labor force attachment of the nonemployed during the 1980s, there are differing views on the mechanism and timing of how the changes in benefits affected Canadian structural unemployment. One view points to the reforms in the Canadian unemployment insurance system in 1971 which dramatically increased the generosity of the system relative to the United States.6 The sharp increase in the Disincentive Index—which quantifies the level of disincentives created by the unemployment insurance system—illustrates the magnitude of these changes in the system (Figure 4).7 As a result, structural unemployment increased, but strong employment growth in the 1970s masked this development. When growth slowed in the 1980s, the increase in structural unemployment (in particular, relative to the United States) became visible. Moreover, there can be long lags between changes in the unemployment insurance system and their ultimate impact on the labor market, especially during the expansionary phase of a business cycle. Only when activity slows and workers are laid off do they learn about the changes in unemployment benefits and adapt their work behavior accordingly.8

Figure 4.
Figure 4.

Canada: Disincentive Index and Unemployment Rate

Citation: IMF Staff Country Reports 2000, 034; 10.5089/9781451806922.002.A002

Sources: Department of Finance; and Statistics Canada.

8. An alternative view focuses attention on developments during the 1980s to explain the increase in structural unemployment. Generally, Canada has had a higher ratio of benefit recipients to unemployed than the United States, reflecting differences in the benefit structure of the two systems. In Canada, a larger share of workers who lose their jobs have been eligible for unemployment benefits; more Canadian workers have received benefits while in training, sickness, or on maternity leave; and the duration of Canadian unemployment benefits has been longer.9 During the 1980s, these differences widened as the number of workers receiving benefits increased in Canada but it decreased in the United States. The ratio of benefit recipients to unemployed in Canada increased over the 1980s from 80 percent to about 100 percent, whereas in the United States, the ratio declined from about 35 percent to about 29 percent. As a result, in the early 1980s the probability of an unemployed Canadian worker receiving benefits was about twice as great as in the United States, and by the end of the 1980s the probability in Canada was about 3.5 times as great.10

9. During the 1990s, comprehensive reforms in Canada replaced the old Unemployment Insurance program with a new Employment Insurance system.11 These reforms are likely to have contributed to a decline in benefit recipiency relative to the United States. Initially, however, their effect on relative unemployment was more than offset by the relative decline in Canadian employment (owing to the longer and deeper recession) and participation rates, leading to a widening in the differential. More recently, sustained growth in Canada has resulted in a sharp decline in the unemployment rate, and a narrowing of the differential to under 4 percentage points at the end of 1999.

10. Alternatively, there have been numerous attempts to explain the unemployment differential between Canada and the United States in terms of differences in labor market flexibility or other characteristics of the labor market (such as unionization, changes in payroll taxes, income taxation, and immigration). Overall, these studies have yielded at best mixed results. With regard to labor market flexibility, empirical evidence suggests that relative labor force adjustment costs and speed of adjustment are quite similar in the two countries.12 Evidence on the degree of wage flexibility is more mixed, with some studies (Prasad and Thomas (1998), Budd (1996), and Storer and Van Audenrode (1998)) finding no evidence that wage responsiveness has contributed to the unemployment differential, while others (Kuhn and Robb(1998)) find smaller relative wage declines for the low skilled in Canada, and hence higher unemployment rates. Other empirical studies focussing on various other labor market characteristics typically used to explain labor market performance (e.g, unionization, immigration, payroll taxes, and income taxes) have generally found that these factors have not played a major role in explaining the unemployment differential.

List of References

  • Amano, R., and R.T. Macklem, 1998, “Unemployment Persistence and Costly Adjustment of Labor,” Canadian Public Policy, Vol. 24, February, pp. S138S152.

    • Search Google Scholar
    • Export Citation
  • Budd, J., 1996, “Union Wage Determination in Canadian and U.S. Manufacturing, 1964–1990: A Comparative Analysis,” Industrial and Labor Relations Review, Vol. 49, July, pp. 67389.

    • Search Google Scholar
    • Export Citation
  • Card, D., and W.C. Riddell, 1995, “Unemployment in Canada and the United States: A Further Analysis”, Working Paper No. 352, Industrial Relations Section, Princeton University, November.

    • Search Google Scholar
    • Export Citation
  • Corak, M., “Is Unemployment Insurance Additive? Evidence from the Benefit Duration and Repeat Users,” Industrial and Labor Relations Review, Vol. 47, October, pp. 6272.

    • Search Google Scholar
    • Export Citation
  • Jones, S., and W.C. Riddell, 1998, “Gross Flows of Labour in Canada and the United States,” Canadian Public Policy, Vol. 24, February, pp. S103S120.

    • Search Google Scholar
    • Export Citation
  • Keil, M. and J. Symons, 1990, “An Analysis of Canadian Unemployment,” Canadian Public Policy, Vol. 16, pp. 116.

  • Kuhn, P. and L. Robb, 1998, “Shifting Skill Demand and the Canada-U.S. Unemployment Rate Gap: Evidence from Prime-Age Men,” Canadian Public Policy, Vol. 24, February, pp. S170S191.

    • Search Google Scholar
    • Export Citation
  • Prasad, E. and A. Thomas, 1998, “Labour Market Adjustment in Canada and the United States, Canadian Public Policy, Vol. 24, February, pp. S121S137.

    • Search Google Scholar
    • Export Citation
  • Riddell, W.C., and 1999, “Canadian Labour Market Performance in International Perspective,” University of British Columbia, and Canadian Institute for Advanced Research, unpublished paper, August.

    • Search Google Scholar
    • Export Citation
  • Riddell, W.C. A. Sharpe, 1998, “The Canada-U.S. Unemployment Rate Gap: An Introduction and Overview,” Canadian Public Policy, Vol. 24, February.

    • Search Google Scholar
    • Export Citation
  • Sargent, T. C., 1996, “An Index of Unemployment Insurance Disincentives,” mimeo, Economic Studies and Policy Analysis Division, Department of Finance, September.

    • Search Google Scholar
    • Export Citation
  • Storer, P., and M.Van Audenrode, 1998, “Exploring the Links Between Wage Inequality and Unemployment: A Comparison of Canada and the United States,” Canadian Public Policy, Vol. 24, February, pp. S233S253.

    • Search Google Scholar
    • Export Citation
  • Sunter, D., 1998, “Canada-U.S. Labor Market Comparison,” Canadian Economic Observer, December.

1

Prepared by Paula R. De Masi.

2

For a more detailed discussion comparing the Canadian, U.S., and European labor markets, see Riddell (1999).

3

This decomposition is derived by considering the unemployment rate as the probability of unemployment conditional on being in the labor force, or,

P(U|LF) = P(U[N) * P(N) / P(LF)

where P(LF) is the probability of being in the labor force (labor force participation rate), P(N) is the probability of nonemployment (one minus the employment rate), and P(U|N) is the probability of unemployment conditional on being nonemployed (the labor force attachment of the nonemployed). Taking logs:

In P(U|LF) = In P(N) + In P(U|N) − In P(LF).

Therefore, changes in the unemployment rate can be decomposed into changes in labor force participation, in nonemployment, and in the labor force attachment of the nonemployed.

4

Using data on gross labor flows between employment, unemployment, and out of the labor force, Jones and Riddell (1998) also find evidence of an increase in labor force attachment of the unemployed during the 1980s.

5

Without correcting for these cyclical differences, Riddell finds that just over half of the widening of the unemployment differential can be explained by the relative increase in nonemployment (37 percent); and the relative change in labor force participation (17 percent). The differential in labor force attachment of the nonemployed, however, remained the single most important factor, accounting for about 45 percent of the increase in the unemployment differential.

7

Disincentives introduced by the unemployment insurance system are quantified in the Disincentive Index which reflects a variety of system parameters including the replacement rate, the minimum and maximum weeks of benefit, the number of weeks required to qualify for benefits, and the waiting period. See Sargent (1996).

8

Coark (1993).

9

For example, Card and Riddell (1995) illustrate that based on data from the late 1980s, the eligibility rate for unemployment insurance was 53 percent in Canada versus 43 percent in the United States. Riddell (1999) reports that over the 1980s, the mean per capita weeks of unemployment in Canada declined from 16.8 weeks to 15.6 weeks, and in the United States from 13 weeks to 10.6 weeks.

11

For a summary of reforms, see Canada—Selected Issues Paper, SM/96/69, Chapter VI. Unemployment Insurance Reforms.

Canada: Selected Issues
Author: International Monetary Fund
  • View in gallery

    Canada, the United States, and Europe: Unemployment Rates

    (Percent)

  • View in gallery

    Canada, the United Statess, and Europe: Employment-Population Ratios

    (Percent)

  • View in gallery

    Canada and the United States: Official and Modified Unemployment Differential1

    (Percentage points)

  • View in gallery

    Canada: Disincentive Index and Unemployment Rate