Language selection


Archived — Working Paper Number 3: International R&D Spillovers Between Industries in Canada and the United States

Prepared by Jeffrey I. Bernstein, Carleton University and the National Bureau of Economic Research, under contract to Industry Canada, September 1994


This study has three objectives. Its primary purpose is to investigate the importance of international spillovers between Canadian and U.S. industries and to estimate the effects of both domestic and international spillovers on production cost, traditional input-output ratios (such as labour-output, and physical capital-output ratios), and research and development (R&D) intensity for eleven manufacturing industries in the United States and Canada. The second objective is to measure productivity growth rates for each of the eleven industries and to determine the productivity gains associated with international spillovers. The third objective is to estimate the private and social rates of return associated with R&D capital and to determine the extra-private returns arising from spillovers between Canada and the United States.

Investment in research and development (R&D) affects a country's standard of living. R&D activities provide both individuals and firms with new products which can be manufactured using relatively more efficient means of production. Consequently, the dynamic efficiency and competitiveness of an economy are both enhanced by R&D investment. It is a distinctive feature of R&D investment that firms undertaking such activity are unable to exclude others from obtaining the benefits of their R&D efforts. Hence, the benefits from R&D cannot be be completely appropriated and, inevitably, spillovers occur. R&D spillovers are ideas borrowed by one R&D performer from the knowledge of another performer.

In a world characterized by international trade, foreign direct investment and the international exchange of information, a country's stock of knowledge depends on its own R&D investment, and because R&D spillovers extend beyond national boundaries, on the R&D efforts of other countries as well.

The effects of domestic and international spillovers are estimated on average variable production cost, labour-output, intermediate input-output, physical capital-output, and R&D capital-output ratios for eleven Canadian and U.S. industries. These industries are: chemical products, electrical products, food and beverage, fabricated metals, nonelectrical machinery, non-metallic mineral products, paper and allied products, petroleum products, primary metals, rubber and plastics, and transportation equipment. In this context, the input-output ratios are referred to as factor intensities.

In this study international spillovers relate to R&D externalities between the corresponding industries in the two countries; spillovers that cross national boundaries but link the same industry in Canada and the United States. Domestic spillovers relate to the interindustry externalities that operate within a national boundary. Moreover, international/interindustry spillovers are indirectly captured through the domestic spillover effects.

One of the study's findings is that international spillovers generally exert greater influence on production cost and factor intensities relative to domestic spillovers. This result is not surprising since the international spillovers link the same industry in both countries. In addition, spillovers from the United States to Canada generate greater effects than spillovers from Canada to the United States. In Canada, spillovers from the United States are cost-reducing in all industries except rubber and plastics. The combination of domestic and international spillovers (that is, the joint effect of both spillovers) are cost-reducing, but occasionally one type of spillover increases variable cost. The study therefore considers the interplay between domestic and international spillovers without unduly restricting the role of each type of spillover.

The direct variable cost reductions (that is, keeping capital intensities fixed) from a 1 percent increase in U.S. R&D capital are: 0.06 percent for chemical products, 0.69 percent for electrical products, 1.13 percent for food and beverage, 0.39 percent for fabricated metals, 0.18 percent for nonelectrical machinery, 0.44 percent for non-metallic mineral products, 0.05 percent for paper and allied products, 0.36 percent for petroleum products, 0.23 percent for primary metals, and 0.39 percent for transportation equipment.

Canadian R&D capital generates cost reductions in the United States in all industries except chemical products, primary metals, and transportation equipment. Moreover, in the paper and allied products industry, the effect of international spillovers from Canada to the United States were estimated to generate cost reductions that are four-and-one-half times greater than the corresponding reductions from the United States to Canada. In the remaining six industries, where we observed cost savings for both countries, the U.S. effect is from two to 20 times greater than the Canadian effect.

International spillovers tend to increase R&D intensities in both countries, that is, international spillovers are complements to domestic R&D capital. This complementarity means that as producers in the corresponding U.S. industry increase their investment in R&D capital, Canadian producers increase the amount of R&D content in their output. The same result holds from Canada to the United States. Substitutes for international spillovers and R&D intensity are observed in the United States for only electrical products, rubber and plastics, and transportation equipment. In Canada, substitutes are found in the petroleum products, and rubber and plastics industries.

In the industries where complementary relationships exist, a 1 percent increase in U.S. R&D capital precipitates to an increase in Canadian R&D intensity from a low of approximately 0.14 percent in nonelectrical machinery to a high of 2.85 percent in fabricated metals. In the United States, a 1 percent increase in Canadian R&D capital causes R&D intensity to rise from a low of 0.01 percent in petroleum products to a high of 0.54 percent in paper and allied products. With respect to the other factor intensities, it was found that in both countries, the international spillovers generally lead to increases in physical capital intensities and decreases in labour and intermediate input intensities.

Total factor productivity growth (TFPG) rates are not much different between corresponding Canadian and U.S. industries. The differences that do occur, however, are clearly visible in the decomposition of TFPG rates. In the United States domestic spillovers generally contribute relatively more to productivity gains than international spillovers. In Canada the international spillover is the major contributor. Also in Canada, international spillovers are generally found to contribute positively to TFP growth. The percentage contributions range from a high of 100 percent in the food and beverage industry to a low of 26 percent in the chemical products industry.

In the United States R&D capital from Canada generally leads to productivity gains. In the electrical products industry, 33 percent of the 1.9 average annual growth rate is due to Canadian R&D capital expansion. In the food and beverage industry, virtually 100 percent of the 2.3 percent TFPG rate is due to international spillovers. In fabricated metals, 58 percent of the 0.3 percent growth rate arises from international spillovers. The percentage contribution is 5 percent of the 1 percent productivity growth rate in nonelectrical machinery, 8 percent of the 1 percent growth rate in non-metallic minerals, 1 percent of the 0.3 percent productivity growth rate in paper and allied products, 64 percent of the 1.1 percent productivity growth rate in petroleum products, and 47 percent of the 0.8 percent growth rate in rubber and plastics. Thus, Canadian R&D capital is found to account for a sizable portion of the productivity gains in a number of U.S. industries.

The real, after tax and net of depreciation, private rates of return are approximately 1.5 percent in Canada and 1.8 percent in the United States — rates that are not materially different from each other. Moreover, in nominal terms, before tax and gross of depreciation, the private rates are close to 13 percent in Canada and 16 percent in the United States. Due to significant domestic and international spillovers, social rates of return to R&D capital are estimated to be substantially above the private rates in both Canada and the United States. In Canada, international spillovers generally account for a greater percentage of the social returns relative to the domestic spillovers; the reverse is true in the United States. Canadian social rates of return (nominal, before tax, gross of depreciation) range from a low of 32 percent in transportation equipment to a high of 162 percent in nonelectrical machinery. Social rates of return are from two-and-one-half to twelve times greater than private returns. In the United States social rates of return are from three-and-one-half to ten times greater than the private rates, which range from a low of 44 percent for rubber and plastics to a high of 183 percent for the food and beverage industry.

These high rates of social return mean that at current R&D levels, there is in Canada substantial underinvestment in R&D. This underinvestment arises from both intranational and international spillovers. Indeed, these returns can be interpreted to mean that for a $100 increase in industrial R&D capital, Canadian industrial output increases over a range of $32 to $162 (depending on the industry), and U.S. industrial output increases over a range of $44 to $183.

The benefits being derived from international R&D spillovers imply that the Canadian government should encourage international technology transfer. This can be accomplished in a number of ways. First, with respect to R&D activities, impediments to the rate of knowledge diffusion should be eliminated, along with restrictions on the importation of R&D related equipment and the inflow of scientists and engineers. Second, with respect to the channels of technology transfer, the government should continue to forge ahead on free trade, eliminate barriers to inward foreign investment, and encourage more firms to enter into licensing agreements and joint ventures. Third, the government should investigate the possibilities for international tax harmonization policies with respect to R&D activities.

Date modified: