Public Policy Sources #37: Productivity growth and economic performance in Canada

In recent years, Canadian economists in universities, government agencies, and private think-tanks have devoted considerable attention to productivity growth and measurement, the diffusion of technology across countries and firms, the role of the international sector in transmitting technological progress, the impact of different fiscal-policy regimes on long-term growth, and a variety of other related topics. Most of the research is empirical in nature and focuses on the Canadian economy.9

Before proceeding, however, I think it important to show why Canada has a productivity problem. To do this, it is essential to compare productivity performance in Canada with the rest of the world. The structure of this section will therefore be as follows. I will begin by discussing how Canada's productivity growth and economic performance fare relative to the rest of the developed world. I will then examine recent research that highlights some of the difficulties in measuring productivity performance. I conclude here that, while our estimates of Canadian productivity growth are imperfect, the Canadian productivity slow-down is not entirely a statistical artifact. This will be followed by an overview of some of the work being conducted by scholars in government, the academy, and elsewhere on various aspects of the slow-down in Canadian productivity.

The data in table 3 below shows business-sector productivity growth rates in the OECD countries over a variety of sample periods. Two productivity indexes are displayed. The first three columns show average annual growth in total factor productivity and the last three columns show the average annual growth rates of average labour productivity (i.e. growth rate of output per worker). While the former measures economy-wide productivity, the latter is a better measure of how much living standards have increased for the average worker.

The data in table 3 reveal a few very disturbing trends. In terms of the total factor productivity growth rate, Canada does not fare very well among the OECD countries. While most countries experienced slower total factor productivity growth during the 1970s (the so-called "productivity slow-down"), Canada was one of the few for which total factor productivity growth did not recover during the 1980s and 1990s. Indeed, from 1979 to 1996, Canada was one of only three countries to experience negative growth in total factor productivity. Among the G-7 countries (Canada, Japan, France, Germany, Italy, the United States, and the United Kingdom), Canada's total factor productivity growth was the most sluggish over the entire sample period. Hence, to the extent that one believes these estimates (more on this later), this data suggests that Canadian productivity growth has been languishing over the past two decades.

In terms of average labour productivity growth, Canada also fares relatively poorly. The annual growth rate of average labour productivity declined in Canada from 1.5 percent in the 1970s to 1.0 percent in the 1980s and 1990s. Among the G-7 countries, Canada also had one of the lowest average labour productivity growth rates. Hence, the productivity improvements of the average Canadian worker appear to be low relative to other countries.

As noted earlier, most studies suggest that slow productivity growth should result in slow increases in the material standard of living. This is borne out by the data. Table 4 below shows the percentage growth in real GDP per employed person in 13 OECD countries. Real GDP per employed person increased by only 4.3 percent in Canada from 1989 to 1996, compared to an increase of 6.5 percent in the United States, 9.8 percent in Japan, and 23 percent in Norway. Canada ties the Netherlands for the slowest growth in real GDP per employed person among these countries.

Another way to illustrate how the slow-down in productivity growth in Canada has taken its toll on material well-being is to examine per-capita GDP in Canada relative to other countries. Table 5 shows GDP per capita as a percentage of American GDP per capita in 1989 and in 1996 for the same 13 OECD countries. While almost every other country improved its per-capita GDP relative to the United States between 1989 and 1996, Canada's per-capita GDP relative to the United States actually declined, from 81.7 percent to 76.7 percent. The only other country to register a decline in this statistic is Sweden. Hence, it appears that the material well-being of the average Canadian is slipping relative to that of the United States and other countries: Canada's position has fallen from first place (behind the United States) in 1989 to sixth place in 1996.10

It should be noted, however, that not all analysts unequivocally accept this evidence of a productivity slow-down in Canada.11 Indeed, as I discussed earlier in the section on growth accounting, conventional measures of total factor productivity are only as good as the quality of the underlying data. If measurements of input and output growth are biased in some way, then estimates of productivity growth will also be biased.

The measurement of productivity growth has been studied widely. (For a very detailed discussion of the measurement issue, see Diewert and Fox 1999.) To show briefly how the measurement problem can arise, it is instructive to examine productivity growth in Canada on a sectoral basis (i.e. industry by industry). So far, my discussion of productivity has been at an aggregate level. Aggregate productivity growth is, of course, a weighted sum of industry-level productivity growth. The data in table 6 below, which was compiled by the Centre for the Study of Living Standards, shows productivity growth rates in Canada by industry from 1984 to 1996. The first column reports the average annual growth rate of total factor productivity while the second column reports average annual labour productivity growth rates.

According to these estimates, both average labour productivity growth rates and total factor productivity growth rates vary significantly by industry. In particular, we observe very rapid productivity growth in some of the traditional sectors of the economy (e.g. agriculture, fishing and trapping, forestry and logging), moderate productivity growth in transportation and storage and communications and other utilities, and sluggish to negative productivity growth rates in manufacturing, financial services, and community, business, and personal services. Why does productivity growth differ so widely from sector to sector?

One might think that investment in new capital equipment--capital equipment that embodies new productivity-enhancing technologies--has been faster in those sectors that have experienced faster productivity growth. A quick glance at the evidence refutes this explanation. Capital accumulation--particularly the accumulation of new information technology--has been fastest in those industries that have experienced the slowest rates of productivity growth, in particular, the financial services sector (Sharpe 1998). Hence, industries showing slow productivity growth have not been growing slowly because of a lack of investment in new capital and equipment; they appear to be growing slowly in spite of the rapid accumulation of new capital. This, of course, is a re-statement of the "computer paradox" of productivity measurement.12

Another explanation for sluggish productivity growth in these sectors is that our measurements of output in the service sector--financial and insurance services, community, personal, and business services--are inherently problematic. For instance, output in the banking industry is measured by the spread between lending rates and borrowing rates. Is this the basis of an adequate measure of productivity? The services provided by these industries are inherently difficult to measure. From an economic standpoint, it might be desirable to measure the output of, say, the insurance industry, as the increase in utility insurance provides the consumer by reducing his exposure to risk. However, this is not a measurable quantity. Most scholars agree that service sector output is seriously underestimated by traditional national income accounting measures. Hence, estimates of service sector productivity are also significantly underestimated. Since the service sector now accounts for a very large (and growing) fraction of total output, this underestimate of service sector productivity will result in an underestimate of economy-wide productivity.

While measurement error may explain much of the slow-down in productivity due to the service sector, however, it does not entirely account for sluggish productivity growth in the manufacturing sector, which has more readily measurable output. Total factor productivity growth in the manufacturing sector was negative from 1984 to 1996 while average labour productivity grew at a slow rate of 0.8 percent per year. According to Sharpe (1998), productivity growth in the manufacturing sector in Canada has been the weakest among the G-7 countries. Growth of output per hour from 1973 to 1996 averaged 1.8 percent a year in Canada compared with a G-7 average growth rate of 3.2 percent a year. Hence, measurement error alone cannot explain Canada's poor performance.

Indeed, the fact that all countries use roughly the same methods to estimate output growth in the service sector implies that service sector growth in all countries is biased downward. Since the growing importance of the service sector is not unique to Canada but is rather a trend reflected among nearly all industrialized countries, the measurement problem should have a more or less symmetrical impact on all countries. Hence, if the measurement problem accounts for a portion of the slow-down in productivity growth in Canada, it also accounts for a portion of the slow-down of productivity growth in other industrialized countries. What it does not account for, however, is the slow-down in Canadian productivity relative to these other countries. The fact that productivity growth, however measured, has fallen in Canada relative to other OECD countries is, therefore, evidence of a specifically Canadian productivity problem.

According to the Policy Research Initiative (1999), the major source of Canada's poor productivity performance relative to the United States is the manufacturing sector. As we noted earlier, total factor productivity growth and labour productivity growth in the Canadian manufacturing sector has been among the worst in the OECD. New growth theory suggests that investments in knowledge-related goods are a major source of productivity advance because of the spill-overs associated with such investments (see Sargent and James 1997). Since the manufacturing sector is a major source of such investments, most empirical research on Canada's productivity performance has focused on this sector.

Positive spill-overs from knowledge-related investments in the manufacturing sector have been found in most Canadian studies. Gera, Gu, and Lee (1998a) find that investments in information and communications technology have a larger impact on productivity growth in the manufacturing sector than other types of physical capital. They estimate that the rate of return on such investments is generally higher than the rate of return on investment in physical capital. Investment in human capital is also an important source of productivity advance. Baldwin and Sabourin (1996), for instance, find that those firms that train their workers enjoy higher productivity growth than those firms that do not. New innovations that raise productivity are often found in new capital investment. According to Gera, Gu, and Lee (1998b), capital accumulation is major channel through which new technology is adopted. In particular, they note that capital accumulation is an important channel through which Canadian firms catch up with their American counterparts.

The new growth theory also suggests that domestic R&D activities should also contribute to productivity growth. Gera, Gu, and Lee (1998b), and Lee and Tang (1998) have found empirical evidence of this direct linkage. Gera, Gu, and Lee (1998a) and Bernstein (1998) have also uncovered evidence of an indirect linkage between R&D expenditures abroad and Canadian productivity growth. These authors find that R&D expenditures in other countries spill over into Canada through international trade and foreign direct investment.

Most of the recent research suggests that much of the slow-down in Canadian manufacturing productivity performance can be attributed to a weakening of some of these channels through which knowledge-related investment finds its way into productivity advance. Slower rates of capital accumulation in Canada relative to the United States have contributed to a widening of the productivity gap between these two countries since new technology is usually embodied in new capital investment. R&D investment is much lower in Canada in spite of very generous R&D tax incentives. Additionally, foreign direct investment in Canada has fallen in recent years, resulting in a slow-down in international R&D spill-overs. Furthermore, there is evidence that the small size of Canadian firms precludes extensive investment in new technology and extensive human capital investment. Some of the major results from this research are highlighted below.

Lee and Tang (1998) estimate that the productivity gap (in favour of the United States) in the manufacturing sector has widened from 22 percent in the period between 1985 and 1988 to 30 percent in the period between 1993 and 1995. They attribute the growth of this productivity gap to the slow-down in capital accumulation in the Canadian manufacturing sector. The net manufacturing capital stock declined by 1.3 percent per year in Canada between 1990 and 1997, compared to a growth rate of 1.7 percent per year in the United States. Since productivity gains are often embodied in new capital investment, the result has been a slow-down in Canadian manufacturing productivity.

According to the Policy Research Initiative (1999), Canada's share of inward foreign direct investment in North America has declined significantly since the mid-1980s. Hirschhorn (1997) argues that Canada derives significant benefits from foreign direct investment since: (a) technology transfer is much faster within multinational firms rather than through licensing or other external agreements; (b) the higher average productivity of foreign affiliates has a positive impact on Canadian total factor productivity; and (c) the highly productive nature of direct foreign investment has a positive impact on economic growth and job creation. The reduction in Canada's share of inward foreign direct investment may therefore be cause for concern.

Industry Canada (1997) has found that American businesses invest almost twice as much in R&D as Canadian businesses relative to the size of the economy. In 1995, R&D investment as a percentage of GDP was 1.7 percent in the United States but 0.9 percent in Canada. A larger share of the American work force is devoted to R&D activities. Patent applications in Canada by Canadians (per person) are much fewer than the corresponding American statistic.13 Gera, Gu, and Lee (1998a) report that R&D performance in Canada is lower in nearly all industries, particularly in those industries that make up a large portion of the Canadian economy.14

Canadian firms appear to lag American firms when it comes to adopting new technologies. Sabourin and Baldwin (1997) find that in several industrial sectors, Canadian plants were less likely to use advanced technology than American plants. They attribute this to the smaller average size of Canadian firms. This, in turn, may be linked to Canada's smaller market.

Table 3: Average annual productivity growth rates in the OECD countries, in percent

Table 4: Percentage increase in real GDP per
employed person in 13 OECD Countries

Table 5: Real GDP per capita as percentage
of the American level in 1989 and 1996.

Table 6: Average annual productivity growth rates in Canada by industry, from 1984-1996, in percent

	Total Factor Productivity	Average Labour Productivity
Agriculture	5.64	3.06
Fishing and trapping	6.36	2.40
Forestry and logging	3.44	3.25
Mining	0.93	3.25
Manufacturing	-0.57	0.8
Construction	-1.12	-2.31
Transportation and storage	1.48	0.66
Communication and other utilities	1.30	0.14
Trade	1.49	2.28
Financial, insurance, and real estate sector	-3.21	0.56
Community, business, and personal services sector	-0.82	-0.49
Government services	-0.80	-0.92
Total business sector	0.65	0.48
Source: Sharpe 1998: table 5.

[Previous] [Contents] [Next]

info@fraserinstitute.ca

You can contact us at the above email address for any comments or information requests. Please report any dead links or technical problems.

Last Modified: Thursday, August 5, 1999.

Public Policy Sources #37:Productivity growth and economic performance in Canada

Table 3: Average annual productivity growth rates in the OECD countries, in percent

Table 6: Average annual productivity growth rates in Canada by industry, from 1984-1996, in percent

Public Policy Sources #37:
Productivity growth and economic performance in Canada