4 Results

View the most recent version.

In this study, we compare capital intensity, defined as the ratio of capital stock to gross domestic product (GDP), between Canada and the United States. We make no comparisons of capital or GDP across countries; rather we compare each within countries, using the respective currencies of each country and then the ratios across the two countries. If we wished to go further, we would have to derive purchasing power parities of capital stock. Since we do not do so, the reader should be aware that part of the difference in capital intensity revealed here may be due to relative price differences between investment goods and products in general in the two economies.

In the remainder of the section, we first present the relative levels of capital intensities in the business sector in Canada and the United States. We then compare the capital intensities in the total economy in the two countries.

4.1 Overall trends in investment and capital intensities in the business sector

While capital is seen to have an important impact on growth rates, there is less agreement on the relative importance of different asset types. The history of technology tends to focus on the evolution of machinery and equipment (M&E). Stories abound about the spinning jenny, the power loom and agricultural equipment that have reinforced the importance that is traditionally given to machinery. The recent high-tech boom has led to an emphasis on high-tech products within this class.

Despite the attention that is paid to M&E, it accounts for no more than 32% of total capital in Canada in 1999 (Table 3). In contrast, buildings account for about 40%. In some situations, the capital they provide is complementary to that of machines. Factories require buildings in which operations are housed. Transportation systems may use trucks, engines and airplanes, but they also require airport terminals and warehousing. In other cases, the building is an undeniable part of the product. Retail requires stores. The travel industry requires hotels and conference centres.

Large amounts of capital are also devoted to engineering construction. In fact, at 29%, its share is almost as large as that of M&E in Canada. These assets underpin the utilities sector, pipelines, railways, airports, communications, and the oil and gas sector.

Canada and the United States differ in terms of the composition of assets. The United States has a higher percentage of assets in M&E. It has about 1 percentage point more in non-information and communications technology (ICT) M&E and some 4 percentage points more in ICTM&E. It also has a greater percentage in buildings. It has considerably less in engineering assets.

Table 3
Share of assets in capital stock, business sector, 2003 (nominal dollars)

The investment-to-GDP ratio from 1987 to 2003 for Canada relative to the United States is presented in Figure 1. Canada's total investment-to-GDP ratio is higher than the U.S. ratio from 1987 to 2003.¹⁸ Over that period, total investment-to-GDP ratio in Canada is about 20% higher than the ratio in the United States. It has declined slightly over the period studied: from 1.5 to 1.2 when measured in 1997 dollars, but from only 1.3 to 1.2 when measured in current dollars (Figure B.1).

An examination of investment-to-GDP ratios by asset class reveals substantial differences both in terms of level and trends. Canada's engineering investment-to-GDP ratio is more than twice that of the United States for the period from 1987 to 2003 and has been growing relatively larger over time.¹⁹ The building investment–GDP ratio in Canada is about the same as that in the United States and has remained relatively stable over time. Canada's investment in non-ICTM&E as a proportion of GDP is also higher than that of the United States, but it has fluctuated over time. At the same time, there has been a persistent ICTM&E investment-to-GDP ratio gap between Canada and the United States that has increased over time.

Figure 1
Canadian business sector investment-to-gross domestic product ratio relative to the U.S. ratio, by asset class, 1987 to 2003 (1997 dollars)

As previously discussed, capital stocks in both countries are derived as the accumulation of past investments that are summed using the perpetual inventory method. However, if different service lives and different depreciation rates are used to compare Canada with the United States, estimates of the relative level and trend in capital intensity may be incorrect. Thus, previous comparisons of capital intensity between Canada and the United States using unadjusted depreciation rates may partly reflect the different methodologies. U.S. depreciation rates that are used by the Bureau of Economic Analysis (BEA) are sometimes lower than those used in the Canadian productivity program, particularly in engineering structures and building structures. As a result, the use of these numbers results in a lower Canadian capital-to-GDP ratio than that of the United States. In Figure 2, the line labelled "Own" depicts the course of the total capital-to- GDP ratio if we employ the capital stock estimates for Canada from the Canadian Productivity Accounts (CPA) and capital stock estimates for the United States from the BEA. Figure 2 also contains the capital–output ratios using common depreciation rates (either Canadian or U.S. rates) to produce capital stocks for both countries. Using common rates raises Canada's relative capital intensity (Figure 2). We first apply BEA depreciation rates to the Canadian stock and compare the capital intensities of the two countries. Based on common BEA depreciation rates, Canada's relative capital intensity becomes higher than that based on each country's own depreciation rates.

To undertake a sensitivity analysis of this finding, we also apply Statistics Canada's depreciation rates used in its productivity program to BEA capital stocks. Interestingly, Canada's relative capital intensity rises further with Statistics Canada's depreciation rates.²⁰ Thus, the magnitude of the difference between Canada's capital intensity and the U.S. intensity is also sensitive to the choice between BEA and Statistics Canada depreciation rates. But at least in the latter part of the 1990s, there is not much difference between the two curves—and the difference is probably not statistically significant. That is, both of these estimates show that the capital–GDP ratio in the business sector is higher in Canada than in the United States. Equally important, the trend in the two countries is virtually the same over the time period under study. The capital stock estimates based on the CPA depreciation rates show that the capital-to-GDP ratio in the Canadian business sector is about 30% higher than that in the United States over the 1987-to-2003 period (Table 4). The results based on the BEA depreciation rates show that the capital-to-GDP ratio in the Canadian business sector is about 20% higher.

Figure 2
Canadian business sector capital-to-gross domestic product ratio relative to the U.S. ratio, 1987 to 2003 (1997 dollars)

Since Canada's capital structure is different from that of the United States, it is important to note the pattern of capital–GDP ratios at the asset level. Canada's relative capital intensity trends by asset class are shown in Figure 3.²¹ Canada's engineering capital intensity has been growing faster than that of the United States and Canada's engineering capital intensity, which was three times as high as that of the United States in 1987, reached almost four times that of the United States in 2003. Canada's capital intensity in non-ICTM&E is slightly higher than that of the United States.²² Canada's building capital matched fairly closely to that of the United States. On the other hand, there has been a persistent ICTM&E capital intensity gap between Canada and the United States.

These differences suggest that there are structural differences between the two countries. The Canadian economy focuses either more on those industries where engineering infrastructure is important or it has a much different capital structure in its industries that leads it to focus more on infrastructure and less on ICT. For buildings, Canada employs a similar relative amount of capital per dollar of GDP.

Figure 3
Canadian business sector capital-to-gross domestic product ratio relative to the U.S. ratio, by asset class, 1987 to 2003 (1997 dollars)

Table 4
Average relative Canadian-U.S. levels of investment and capital intensities, by asset class, business sector, 1987 to 2003 (1997 dollars)

4.2 Decomposition of the Canadian-U.S. total capital intensity gap in the business sector

This section examines whether the differences in the capital–output ratio come from differences in intensity or from differences in the structure of the two economies. Since structure can be defined in several ways, the section approaches the question in stages. It starts by examining differences in capital structure and then differences by industry.

In the first instance, we ask how much of the difference in capital intensity comes from the difference in the intensity of specific assets as opposed to the fact that the two economies use assets in different proportions—either because of industry differences or because of different production technologies.

Table 5 presents the decomposition of the Canadian-U.S. total capital intensity gap in 2003, based on four asset types using Equation (12). In 2003, Canada's total capital intensity was 24% higher than the U.S. intensity. Canada's superior position was due to a higher capital intensity, which contributed 29 percentage points. An asset-by-asset comparison reveals that engineering capital contributed 30 percentage points to Canada's advantage. However, this was partially offset by Canada's lower capital intensity in ICT and building capital. In sum, the capital intensity of the engineering component contributed more than 100% to Canada's relatively higher capital intensity.

Differences in capital intensity may also arise because of differences in industrial structure— because Canada has a larger percentage of its assets in industries that naturally have a lower/higher capital–GDP ratio. The results of a decomposition exercise using industries are presented in Table 6.

Table 5
Decomposition of Canadian-U.S. aggregate capital intensity gap, by asset class, based on Statistics Canada depreciation rates, business sector, 2003 (1997 dollars)

Table 6
Decomposition of Canadian-U.S. aggregate capital intensity gap, by industry, based on Statistics Canada depreciation rates, business sector, 2003 (1997 dollars)

Canada's industry structure relative to that of the United States was favourable to Canada, contributing 20 percentage points to Canada's relatively higher capital intensity. That is, Canada had a larger share of output in those sectors that were generally more capital intensive, such as the utilities and mining sectors. In addition, higher capital intensity in some sectors increased Canada's structural advantage by some 4 percentage points. The largest industry contributors to Canada's relatively higher capital intensity were utilities, finance and manufacturing. The intensity effects stemming from aforementioned industries were partially offset by Canada's relatively lower capital intensity in business services, other services, wholesale and retail trade.

In summary, Canada's higher capital intensity was primarily the result of a greater focus on industries that have a high capital intensity—particularly infrastructure. However, this decomposition of the aggregate capital–output ratio hides potentially important differences by asset type between the two countries. The next section investigates this possibility.

Table 7
Decomposition of Canadian-U.S. information and communications technology capital intensity gap, by industry, based on Statistics Canada depreciation rates, business sector, 2003 (1997 dollars)

Table 8
Decomposition of Canadian-U.S. non-information and communications technology machinery and equipment capital intensity gap, by industry, based on Statistics Canada depreciation rates, business sector, 2003 (1997 dollars)

4.3 Decomposition of the Canadian-U.S. capital intensity gap by asset type in the business sector

The total capital stock-to-GDP ratios mask significant differences across asset types. Canada's relative capital intensity is lower than that of the United States in ICTM&E, it is higher in engineering structure and non-ICTM&E, and it is similar in buildings. In this section, we examine the contribution of industries to the capital intensity differences for each asset class.

The largest capital intensity gap exists in ICT. Canada's ICT capital intensity was about 34% below the U.S. intensity in 2003. Differences in industrial structure between the two countries contributed about 7 percentage points to the gap while the intensity component contributed 33 percentage points to the gap (Table 7). The largest contributions of lower ICT capital intensity were in construction; manufacturing; transportation, warehousing and utilities; and finance, insurance, real estate and rental and leasing (FIRE). However, the ICT capital intensity gap was fairly widespread across industries, with the exception of information and other services sectors.

The non-ICTM&E capital intensity in the Canadian business sector is 14% higher than in the United States (Table 8). But Canada tends to concentrate more in industries that have higher non- ICTM&E capital intensity and, if industrial structure is taken into account, the non-ICTM&E capital intensity is lower in Canada. The lower non-ICTM&E capital intensity at the industry level in Canada reduced the aggregate non-ICT capital intensity by 12 percentage points. Higher non-ICTM&E capital intensity existed in manufacturing and finance. This was offset by lower non-ICTM&E capital intensity in information and professional and business services—two of the most dynamic service sectors. However, the structural difference between the two countries increased Canada's relative capital intensity by 31 percentage points.

The capital intensity in building structures is similar in the Canadian and U.S. business sectors (Table 9). This is a result of a small intensity effect and a small structural effect. Canada's relative capital intensity in building is lower in forestry, fishing, hunting and mining; manufacturing; retail trade; and professional and business services. It is higher in agriculture; construction; transportation, warehousing and utilities; and the FIRE sector.

Table 9
Decomposition of Canadian-U.S. building capital intensity gap, by industry, based on Statistics Canada depreciation rates, business sector, 2003 (1997 dollars)

Canada's engineering capital intensity was higher than that of the United States by 274% in 2003 (Table 10). Both the industry structural and intensity components were positive, contributing 103 and 70 percentage points, respectively, to Canada's advantage. Once again, industry structure tended to increase Canadian overall capital intensity. The industries with the largest engineering intensity advantages were those in the manufacturing and utilities sectors. Although the total engineering capital intensity was significantly higher in Canada than in the United States, some Canadian industries had a lower intensity contribution compared to the United States, including construction and professional and business services.

Table 10
Decomposition of Canadian-U.S. engineering capital intensity gap, by industry, based on Statistics Canada depreciation rates, business sector, 2003 (1997 dollars)

The decomposition that has been used in this paper suffers from the problem that it is deterministic and does not facilitate hypothesis testing. This problem has long been recognized in the geography shift-share literature (Patterson 1991, Knudsen and Barff 1991). To permit statistical tests, stochastic linear regression models have been adopted that essentially test for differences in means across categories (Patterson 1991, Rubery 1998, Smith, Fagan and Rubery 1998).

In this spirit, we estimated regressions of the capital intensity of assets on binary variables representing each industry used in the above analysis and a binary variable that represented Canada (Table 11). The coefficients on the binary variable indicate the extent to which Canada suffered from a capital deficit after industry structure is considered—what is comparable to the overall intensity contribution reported in Tables 6 to 10. While the significance of the differences can be tested in this way, it should be recognized that the tests are not very powerful here because the number of industry categories being used is not large. Ultimately, to establish the significance of the differences across the two categories, one needs to take into account the statistical test provided by the data in Table 11, the economic meaningfulness of the size of the differences and the sensitivity of the size of the differences to alternate depreciation assumptions used in constructing the capital stock estimates.

The coefficients from the regression generally accord with those from the share decomposition analysis. Canadian aggregate capital intensity is 3.4% higher. ICT and non-ICTM&E is lower. Buildings are slightly lower. Engineering construction is much higher. It should be noted that there are few statistically significant differences between the two countries—which may be due to the level of industry detail available to us. Only ICT capital intensity is significantly lower in Canada than in the United States.

Table 11
Intensity differences using a decomposition analysis and a fixed effects regression, 2003 (1997 dollars)

4.4 Relative Canadian-U.S. levels of investment and capital intensities in the total economy

In this section, we examine the relative Canadian-U.S. levels of investment and capital intensity in the total economy. The results are presented in Figure 4, Figure 5 and Table 12. Canada's investment-to-GDP ratio in the total economy is about 10% higher than that of the United States in 2003. This is a result of higher investment intensity in the Canadian business sector and slightly lower investment intensity in the non-business sector.

Canada's investment intensity in the total economy is higher than that of the United States for engineering and building structures. But it is lower for M&E. The lower M&E investment intensity in the Canadian economy is a result of much lower investment intensity in the nonbusiness sector. In the business sector, the investment intensity in M&E is similar in the two countries. To some extent, the much lower capital intensity in M&E in the Canadian nonbusiness sector reflects the fact that military expenditure in the United States is treated as investment.

Table 12
Relative Canadian-U.S. levels of investment and capital intensity, total economy, 2003 (1997 dollars)

Table 13
Decomposition of Canada-U.S. aggregate capital intensity gap, by asset type, based on Statistics Canada depreciation rates, total economy, 2003 (1997 dollars)

Table 13 presents the decomposition of the Canadian-U.S. capital intensity difference in the total economy in 2003. In that year, Canada's capital intensity in the total economy is 18% higher than the U.S. intensity. Canada's higher capital intensity is due to the higher capital intensity in engineering structures, which contributed 17 percentage points. The higher capital intensity in buildings contributed 5 percentage points. The lower M&E capital intensity in the Canadian economy lowered the overall capital intensity by 3 percentage points.

Figure 4
Canadian investment-to-gross domestic product ratio relative to the U.S. ratio, total economy, 1987 to 2003 (1997 dollars)

Figure 5
Canadian capital-to-gross domestic product ratio relative to the U.S. ratio, total economy, 1987 to 2003 (1997 dollars)

 

^18. Figures in current dollars are in Appendix B. These figures are not qualitatively different from those presented in the main text, which are based on 1997 dollars.

^19. Baldwin and Gorecki (1986) report that in manufacturing, the Canadian-U.S. ratio of machinery and equipment was relatively stable from 1961 to 1979, but the Canadian-U.S. ratio for structures and engineering increased in relative terms.

^20. These results apply to all asset types in both 1997 and current dollars. The difference in the intensities arising from the use of the two depreciation rates reflects the large increase in Canadian investment on buildings and engineering relative to the United States during the 1980s. This will be weighted more heavily in the net capital stock calculation because of the relatively higher Statistics Canada's depreciation rates for these investments. For comparison, see Figure 3 and Appendix Figure D.1. The use of the lower U.S. depreciation rate for engineering and buildings reduces Canadian capital in these areas.

^21. These results are based on capital stock estimates using Statistics Canada's depreciation rates. Results based on Bureau of Economic Analysis depreciation rates are not qualitatively different; they are presented in Appendix C.

^22. The results from using the Canadian Productivity Accounts depreciation rates show that the non-information and communications technology (ICT) machinery and equipment (M&E) capital intensity is about 20% higher in Canada, while the results based on the Bureau of Economic Analysis depreciation rates show that non-ICTM&E capital intensity is similar in the two countries (Table 4).