Correlation does not imply causation…

By Adam Holbrook

It is said that the phrase “correlation does not imply causation” is carved over the doors of most professors of statistics. Indeed, while it may be possible to correlate such diverse phenomena as stock market trends with the number of sunspots, it is difficult to imagine the linkage between them. This principle should also apply to the assumption that spending on R&D automatically results in economic growth, higher productivity and greater social benefits.

Performing R&D is not the only way an economy can innovate. Yet the government has set an ambitious target for increasing R&D performance in Canada. The Speech from the Throne states: “ We must strive for Canada to become one of the top five countries for R&D performance by 2010”.

Commentators usually interpret this to mean Canada’s research intensity — the ratio of gross expenditures on R&D (GERD) to the gross domestic product (GDP) — should be among the top five by 2010. Based on 1998 figures from the OECD, the last year for which international comparisons are available1, this would mean Canada would have to invest another $11 billion per year on R&D on a base of $13.6 billion — in other words almost double our expenditure. If this target means being among the top five R&D spenders, the gap widens to $15 billion. The Prime Minister, during subsequent debate on the Speech, stated the government wanted “to ensure our R&D effort per capita is among the top five countries in the world”. On this basis the shortfall is also $15 billion.

It is possible to show there is a strong correlation between GERD and GDP for OECD (i.e. market-driven) economies. The non-linear equation derived from this regression analysis shows larger OECD economies have higher R&D intensities (GERD/GDP). It is also possible to calculate, for a given GDP, whether or not a given country is above or below the OECD norm.

It will not surprise anyone to learn Canada falls below this norm. What’s surprising is that we are not far below the line. Based on the OECD data for 1998 (see footnote), Canada should have had a GERD/GDP ratio of 1.7% rather than the 1.6% actually recorded by the OECD. Therefore, Canada would have had to spend approximately $1.4 billion more on R&D to have met the OECD R&D spending norm for that year.

Statistics Canada recently released figures (see RE$EARCH MONEY, Nov 13, 2001) showing an estimated GERD/GDP of 1.81% in 2001, and a revised GERD/GDP of 1.77% for 1999. These revisions are possibly due to better reporting of R&D spending by the higher education sector and smaller industrial R&D performers. Based on these data, there is no shortfall at all!

Keith Smith of the United Nations University at Maastricht, in the August 2001 issue of Science and Public Policy, pointed out national R&D intensities are a function of the structure of the manufacturing sector of individual economies. Some economies have larger “research intensive” sectors, and thus higher GERD/GDP ratios. Others with either relatively smaller manufacturing sectors, or less R&D intensive industries, will fall below the norm.

All this suggests that the government’s goal (or goals) could be more difficult to achieve than previously thought. Not only will it require massive infusions of new funding from both the public and private sectors, but it would require a massive restructuring of our manufacturing sector, away from areas where we are internationally competitive to other industries with higher research intensities. Is this what the government intends? Probably not.

Linking greater economic growth, productivity and social cohesion to a single knowledge input, and force-feeding that input is an overly simplistic approach to a complex problem. Canada needs to increase its productivity and rate of economic growth, but there are other ways besides increasing R&D funding to increase the country’s level of technological innovation.

Canada has many competitive advantages. We need to understand how technology can help us innovate in those industries, including resource and services industries, where we are competitive. We need investments that draw on the innate inventiveness and innovativeness of Canadians and policies to promote these ideals.

Most successful organizations have business plans that articulate their corporate strategy. What Canada really needs is a business plan for technology development. Such a plan must include increased R&D spending, but should not tie the nation to a single and probably unattainable target. But an increase of, say, a billion inflation-adjusted dollars in government R&D funding in the upcoming Budget, to keep us ahead of the OECD norm, would not be a bad start.

Adam Holbrook is associate director of the Centre for Policy Research on S&T at Simon Fraser University.

1 “OECD in Figures, 2001”. The R&D data tables given span the period 1996 to 1999, given the different reporting practices of member nations. Canadian data are for 1999.