A Frontier Approach to Canada-U.S. Multifactor Productivity Performance

Abstract

Much of the recent Canadian productivity debate has focused on the Canada-U.S. productivity growth gap but fewer efforts have been directed to tracing and quantifying its sources. This paper fills this gap by exploiting the frontier approach to productivity measurement. Using a nonparametric programming method, we construct a North American frontier for the business sector and for the manufacturing sector based on data from the two countries. Each sector or industry is compared to that frontier. How much closer a particular sector or industry of any country gets to the North American frontier is known as technical efficiency and results from 'catching up' to best-practice production technology; how much the North American frontier shifts at each sector's or industry’s observed input mix is known as 'technical change' and results from the type of innovation that brings best-practice technology into the country. The combined effect of these two changes yields a frontier version of multifactor productivity growth that is derived from the Malmquist index of multifactor productivity growth. The paper concludes that, whether at the level of the aggregated business sector or at the manufacturing sector level, Canada’s productivity 'problem' during the 1988-2000 period, is mostly ascribed to the deterioration of its technical efficiency - or the extent to which firms are using production technology that places them behind the leaders. The results also suggest that Canada's productivity surge in the 1995-2000 period was mainly due to a recovery in its technical efficiency. We also find that, over the 1981-1997 period, the dispersion of the distribution of technical efficiency across North American manufacturing industries increased drastically.