Canada and the United States

PRODUCTIVITY GROWTH IN Canada and the United States BY JOHN BALDWIN, JEAN-PIERRE MAYNARD AND STEWART WELLS RÉSUMÉ Au cours des 40 dernières années, la performance du Canada au plan de la croissance de la productivité a été assez semblable à celle des États-Unis. On peut en déduire que les importantes mesures sociales prises n ont guère eu d incidence, bénéfique ou nuisible, sur la hausse de la productivité. Au cours des années 1990, cependant, la croissance du PIB par habitant était nettement moins bonne au Canada qu aux États-Unis, ce qui s explique principalement par la croissance beaucoup plus forte de l emploi aux États-Unis. (Traduction: www.isuma.net) ABSTRACT Over the last 40 years, Canada and the United States have had quite similar performance in terms of productivity growth. These more or less similar rates of productivity growth have taken place despite differences in industry structure, trade orientation, and major social programs. However, in the 1990s, while productivity growth continued to be similar, the Canadian economy did significantly worse than the United States in terms of growth of GDP per capita. Much stronger employment growth in the United States is the most important reason for the deterioration in Canadian performance with regard to our standard of living. FIGURE 1 Business Sector Labour Productivity Growth in Canada and the United States INDEX: 1961=100 CANADA UNITED STATES isuma 119

Productivity is one of several key indicators of the health of an economy. It is a measure of productive capability how much output an economy produces for the resources that contribute to production. Statistics Canada measures the growth in productivity that occurs over time. Changes in productivity capture our progress in improving our capability to produce output as we increase our inputs. Productivity growth thus captures the increase in our production efficiency. Increases in productivity stem from improvements in knowledge and in production techniques. These increases can occur because plants become larger and exploit scale economies, or use higher quality labour, or introduce better quality products. They can also arise from organizational changes in management or the way in which production is organized on the shop floor. What are productivity statistics? The most commonly used measure of productivity is labour productivity the amount of output produced per unit of labour input (usually represented by hours-worked). However, it is affected by the amount of capital that is provided to workers. Capital is simply the cumulated past investment made in machinery, equipment and buildings. Plants that have more capital tend to have a higher output per hour-worked. Since it is important to know whether changes in labour productivity occur because of changes in the amount of capital used as opposed to other factors like technical change, Statistics Canada also produces what is referred to as a multifactor productivity measure. It bundles both labour, capital and intermediate inputs together in an input measure and measures changes in output relative to changes in inputs. For example, if output increases by six percent and our inputs by five percent, there is a one percent increase in multifactor productivity. Canada-US comparisons Cross-country comparisons of productivity are invariably difficult because of differences in methodology. Output and inputs are not always measured in the same way. For example, labour can be measured as the number of jobs, the number of people employed or the number of hoursworked. Adjustments to any of these three can be made to handle changes in the quality of workers. Capital can be measured as gross or net of depreciation. Fortunately, the us Bureau of Labor Statistics provides measures of labour and multifactor productivity that are quite comparable for our purposes. Comparisons based on labour productivity are the most straightforward. Both Statistics Canada and the us Bureau of Labor Statistics report a labour productivity measure for the business sector (i.e., excluding government from the calculations). In Canada, this calculation excludes a large portion of both health and education sectors; in the United States, it is primarily the public education sector that is excluded. Both countries use hours-worked as a measure of labour input. There are differences in the methods used to reweight the base used in these calculations in order to reflect the changing structure of the economy, but these have no influence on comparisons at the aggregate level used here. 1 A comparison of the growth in labour productivity in Canada and the United States from 1961 to 1997 shows the two countries with very similar performance (Figure 1). Canada increases slightly faster than the United States in the early years and slightly less in later years, but is still marginally ahead of the United States by 1997 when growth is cumulated over the entire period. A comparison of multifactor productivity growth is more difficult because methodological differences used in constructing these measures are a bit larger here. First, the United States makes corrections for the quality of labour in FIGURE 2 Business Sector Multifactor Productivity Growth in Canada and the United States INDEX: 1961=100 CANADA UNITED STATES 120 Spring Printemps 2000

its measures of growth for the economy as a whole that are different from those made in Canada. 2 Second, while both countries use the perpetual inventory technique to calculate capital stocks from investment flows, and both use net capital stock, they use slightly different approaches to developing such measures. 3 A comparison of multifactor productivity for the business sector in both countries, using the official series of Statistics Canada and the us Bureau of Labor Statistics (Figure 2) shows Canada growing at a faster rate than the United States. Until the early 1980s, there is little difference between the two countries. But since that time, Canada has moved slightly ahead. It is also noteworthy that there are substantial cyclical effects in the measured rates of productivity growth. The rates of productivity growth for both Canada and the United States show the effects of the recession in the early 1980s. Although the recession of the early 1990s has a marginal effect on the rate of productivity growth in the United States (Figure 3), Canada experiences a more pronounced productivity slowdown during this period. This means that Canada-us comparisons of performance can be quite sensitive to the choice of endpoints. In particular, choosing years like 1988 and measuring for short periods through the recession in the early 1990s, or doing the same in the early 1980s, will give a more pessimistic view of Canada s performance relative to the United States than longer run comparisons will yield. These differences in multifactor productivity growth rates for Canada and the United States are marginal and should not be regarded as significant because of differences in the methods used. In order to demonstrate this, we have reestimated us multifactor productivity removing the correction for changes in worker quality that is normally included in us estimates. We did the same for Canada. We then compare the rates of growth in Canada and the United States by dividing the former by the latter to provide a measure of the change in relative position over time (Figure 3). The year 1961 is chosen as the base period and indexed to 100. The chart depicts the extent to which cumulative rates of growth in Canada since 1961 exceed or fall behind those of the United States. For the purpose of comparison, we do the same for the growth in labour productivity over the same period. Both labour productivity and multifactor productivity growth rates show a similar trend. Canadian productivity growth rates are higher than those of the United States from the late 1960s until the early 1970s (around the time of the first major opec oil price increase). Thereafter, there is essentially no difference in the two countries until the mid- 1980s. From 1988 until 1997, there has been a slight decline in the Canadian growth rate relative to the American. It is, however, important to recognize that the scale of these changes is small. At most, Canadian cumulative growth rates move 10 percentage points higher than American rates over the first 15 year period and then fall back by about five percentage points. These changes are within the margin of error that must be ascribed to international comparisons because of differences in the methods and data used. On balance, these data show substantial similarity between the growth of productivity in the Canadian and the us economies over the last 40 years. Several factors contribute to this: the proximity of the two economies; the large amount of foreign investment; similarities in the technologies available; and the close relationship that exists between the two countries. Is there something special about the period since 1960? Did we do more or less well in previous periods when the two economies were less integrated, when Canada s trade was more closely oriented with England and the British FIGURE 3 Relative Growth in Canada/US Labour Productivity and Canada/US MFP, with no Adjustments for Labour Quality Change (business sector) INDEX: U.S.=100 Multifactor productivity ratio Labour productivity ratio isuma 121

FIGURE 4 Historical Rates of Growth of Output per Employee COMPOUND ANNUAL UNITED STATES CANADA Commonwealth, when the economy was more heavily reliant on the extraction of natural resources? To answer this, we examine historical trends in labour productivity dating back to the years just after Confederation (Figure 4). For purposes of historical comparability, labour productivity is defined here using total output divided by employment and not output divided by hours-worked as was done in Figure 1. It is clear that similarities in labour productivity growth have been with us since Confederation. Except for a period immediately after World War 1, Canada has consistently tracked the labour productivity performance of the United States. Despite our distinctiveness in terms of trade orientation with Britain before World War 11, despite our greater reliance on natural resources, despite our adoption of a more comprehensive social safety net, our productivity growth has increased by about the same amount as that of the United States in just about every major phase of our industrial history. While a slowdown has occurred in the period after the opec oil crisis in the early 1970s, our slowdown is the same as that of the United States. Productivity and the standard of living While Canada and us productivity growth rates have generally been quite similar in the 1990s, it is important to note that differences in economic performance have arisen in a number of other areas. While the previous section suggests that our long-term productivity performance has not been grossly inferior to that of the United States, there is no reason to be complacent about our general economic performance. To understand why this is the case, it is important to stress what productivity does not measure, since some observers have confused productivity growth with growth in other important economic indicators. Productivity growth does not measure changes in profits or wages. A business manager knows that he can increase efficiency, but if the price for his product falls, he may see his profits decline and be forced to pay his workers less if he is to remain in business. The same can happen for a nation. Productivity can go up, but some other economic series can remain unaffected. We can produce resources like oil or nickel more efficiently than anyone else in the world, but if export prices for our commodities fall, profits and wages may fall, or may fail to increase over time. In particular, productivity can increase and standards of living can fall. The rate of growth in Canadian gdp per capita in the 1990s (1988-97) was less than a third of its growth rate in the 1980s (1978-88). Yet the productivity performance of the Canadian economy has been more or less the same in the two periods. This has posed a conundrum. How is it possible for Canada to do relatively well in one measure and poorly in the other? How is it possible that Canada has fallen behind the United States in recent years with regards to growth in gdp per capita, while its productivity growth has been much the same? Since productivity is one of the important determinants of the standard of living, the difference in the two measures over the last decade requires an explanation. Productivity growth is generally closely related to growth in gdp per capita; but there are periods when the two diverge. In order to explain how this can occur, we must consider the inherent differences in the way the two measures are constructed. The measure of standard of living gdp per capita differs in several respects from our labour productivity measure, though it is, by construction, linked via an identity: 122 Spring Printemps 2000

FIGURE 5 Reconciliation of Growth in Real GDP per Capita to Growth in Real GDP per job in the 1980s versus 1990s Overall Canadian Economy AVERAGE ANNUAL GROWTH Real GDP per capita Ratio: Pop 15+ to total population Ratio of jobs to population 15+ Average annual hours Labour productivity isuma gdp/capita = (gdp/hours-worked) (hours-worked/job) (jobs/potential labour-force) (potential labour-force/population). This identity means that the growth rate in gdp per capita is just equal to the sum of the growth rates in labour productivity (the first term on the right hand side) plus the rates of growth of the other three terms. Thus growth in the standard of living can increase at a different rate than labour productivity if there are any changes in the other three terms hours worked per worker, the ratio of those with a job to those who might take a job (a type of employment rate), or the ratio of the population that might take a job to the total population (a type of participation rate). The rates of growth in gdp per capita and output per worker can diverge substantially during periods when the employment rate, or the participation rate, or the hours-worked variable is either increasing or decreasing. In Figure 5, we graph the differences in the various components just discussed for Canada in the 1980s and 1990s. gdp here includes both government and private sector output. Because of the identity outlined in Equation 1, the rate of growth of gdp per capita is just equal to the sum of the growth in labour productivity and the other three components. Between the 1980s and the 1990s, the growth in gdp per capita fell while the growth in gdp per hour worked (a measure that is closely related to the official measure of private business sector labour productivity) has remained relatively constant. The difference between the two in each decade arises from several sources. First, in both decades, we had relatively similar increases in the percentage of the population who tend to work the percentage of the population over 15. In both decades, we had relatively constant decreases in the hours worked per job. Since the rates of growth of these two variables have not changed substantially over the two decades, neither explains the decline in the growth in gdp per capita relative to the growth in productivity. The cause of this decline is the decrease in the number of people holding jobs relative to the population that can take jobs. While this ratio increased in the 1980s, it fell in the 1990s. Thus, growth in gdp per capita has decreased despite the relative constancy in growth of business gdp per worker because employment growth did not keep up with population growth. This could have occurred because Canadians increasingly chose not to take jobs, for example by taking early retirement, or because not enough new employment opportunities were created to handle the increasing population. There have been substantial cyclical variations in the various components. gdp per capita and the jobs to population ratio both declined substantially in the early 1980s and the 1990s when the Canadian economy suffered a recession. But during the mid-1980s, the employment to population ratio experienced positive growth after one year of precipitous decline in 1982 thereby allowing the positive growth rates in gdp per hour to be amplified into even higher growth rates in gdp per capita during this period. However, the 1990s were quite different from the 1980s. The early 1990s experienced not just one but several years of dramatic decline in the employment/population ratio, whose cumulative effect was substantial. Moreover, the subsequent growth in this ratio has been weaker than in the 1980s and was not sufficient to overcome the damage of the early 1990s recession. The experience of job creation in the us economy has 123

been quite different. The 1990s saw us unemployment rates fall to their lowest level in decades. Employment growth has kept pace with population growth. Despite the fact that labour productivity has tended to increase at similar rates in both Canada and the United States (Figure 1), the 1990s performance of growth in real gdp per capita in the two countries was quite dissimilar because of this disparate performance in the growth of jobs relative to the growth in the population. While growth in Canadian gdpper capita tracked that of the United States before 1988, it fell considerably behind during the 1990s (Figure 6). Conclusion Productivity growth in the economy is an important contributor to improvements in our standard of living. It is affected by small, incremental changes in a host of factors that occur on the plant floor. These include new production techniques, changes in plant size, changes in organization as well as other factors that are associated with new knowledge. These changes are generally not cataclysmic. Even momentous changes involving new technologies take time to implement. And the changes are relatively steady, when measured over long cycles. Since the First World War, the annual growth rates of labour productivity have averaged very close to two percent per year. They slow down after 1973 but the slowdown in Canada has been much the same as in the United States. What is remarkable about the historical performance of productivity growth in the Canadian economy is its similarity to that of the United States. During different periods when we have experienced war and peace, a transition to a society that has a stronger safety net, and other societal changes, the rate at which new knowledge has been incorporated into the production process has been relatively steady and about the same as the United States. Over almost 40 years since the 1960s, Canada has continued to move in step with the United States over the long run. This has occurred at the same time that trade has become more liberalized between the two countries. Moreover, the similarities in performance, when measured over long periods, generally extend back 100 years. To some, our inability to outperform the Americans may be discouraging. Having comparable relative growth rates means that we have not done better than the United States in terms of developing a higher output per worker. Others may be more encouraged because these comparable rates indicate that we are part of a North American economy and the productivity gains found in the United States can be expected from our business sector. While we may not have gained on the United States, our progress has been relatively steady despite what some observers regard as large and significant differences in the two economies. While we were less integrated with the United States prior to World War ii, we became more integrated after it. Despite this shift, productivity growth plodded along at a relatively similar pace in the two countries. Admittedly, the macroeconomic environment has meant that Canada has diverged from the United States over short periods when recessions have been more severe in Canada but these divergences have been ephemeral in terms of their effect on long-term productivity growth. That we have not pulled ahead may also be frustrating to those who have argued that policy options like free trade were available to jump-start our economy, so as to narrow the gap in output per worker between the two economies and overtake the Americans. But the fact that the movement to free trade in the 1990s did not lead to a dramatic spurt in productivity growth perhaps should not be surprising. The commercial trade history during the post World War ii era was one of continuous declines in tariffs with the Kennedy round being felt in the 1970s and the Tokyo Round taking place in the 1980s. Productivity gains that resulted from these declines would be incorporated in our estimates both before and after the North American free trade pact of the early 1990s. On the other hand, the more or less constant rates of productivity growth also suggest that major social initiatives also have little effect, either beneficial or deleterious, on productivity growth in the economy at least as productivity growth is traditionally measured. Those who would argue that dramatic improvements in education or health must necessarily be reflected in improvements in productivity would not take great solace from the relatively constant growth in labour productivity. But then neither would those who argue that expanding the scope of the social safety net removes the incentives that are necessary for productivity growth. John Baldwin, Jean-Pierre Maynard and Stewart Wells are with Statistics Canada. Baldwin is Director, Micro-Economic Studies and Analysis Division, email: baldoh@statcan.ca. Maynard is Senior Economist, Productivity Measures Division, email: maynard@statcan.ca. Wells is Assistant Chief Statistician, National Accounts and Analytical Studies, email: wellste@statcan.ca. The authors are indebted to Guy Gellatly of the Micro-Economic Studies and Analysis Division for comments and suggestions. Endnotes 1. For gdp, the United States uses a chain-based index that continuously updates weights, while Canada uses a Laspeyres index whose weights are updated every five years. At the aggregate level being examined here, using a chained index rather than a Laspeyres index has almost no effect on the results. 2. The United States weights growth rates in hours worked by relative wage rates groups of workers who differ by experience and education. Canada captures some of these labour quality differences by taking account of the industry composition when we calculate growth rates in hours worked. The corrections employed by both countries increase the rate of growth in labour inputs and, therefore, decrease the rate of multifactor productivity growth compared to what would be produced if growth rates were calculated from a simple sum of hours worked. The data in Figure 3 remove both of these quality effects. 3. On the one hand, the rates of depreciation implicit in the us approach are lower than those derived in Canada. Second, the United States includes land in its estimate of capital while Canada does not. The former increases the Canadian rate of growth of capital stock relative to what would be produced using the us approach; the latter decreases it. The two effects offset one another. Research is underway to assess the magnitude of these two effects. 124 Spring Printemps 2000