Perspectives on North American Free Trade: Paper Number 4: The Changing Industry and Skill Mix of Canada's International Trade

This page has been archived on the Web

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please contact us to request a format other than those available.

by Peter Dungan and Steve Murphy, University of Toronto, April 1999


Executive Summary

The intention of the project is to examine in detail the industry and skill mix of Canadian exports and imports as they stood in 1997 and to ask how this mix has changed since 1961. Some important and timely questions underlie the study: How much of Canada's employment is associated with exporting, and how much has this changed over the last three decades? To what extent do Canadian exports and export employment still depend upon raw materials, and how much have our exports moved to being made up of tertiary manufacturing and services? Even if such a movement has occurred, the skill mix implications are not clear: at least some primary sector jobs require considerable skill and are very well paid, while many manufacturing and services jobs require little skill and are poorly rewarded. How has the education/skill mix of Canadian export employment changed over time, and in relation to employment in the entire economy? On the import side there is the widely-held conception that a more open economy and heavier reliance on imports has pushed aside low-skill jobs in Canada and has led to lower incomes for the less skilled. To what extent can we identify such a trend from 1961 through 1997?

The study uses Input-Output (I/O) techniques to address these questions, supplemented by detailed sectoral education data from the Canadian Occupational Projection System (COPS). We build upon earlier studies at Industry Canada and elsewhere by using annual I/O tables for 1961 through 1992 and supplementary detail to extend the analysis through 1997, and by using a somewhat greater industrial disaggregation. Results are presented in both aggregated and disaggregated form together with a variety of sensitivity tests and with decompositions of the sources of change in export and "import replacement" employment over 1961–1997.

Main Findings

  • While exports and imports have grown strongly as shares of the Canadian economy since 1961 — and massively so in the 1990s — the growth in the share of employment attributable to exports (or "displaced" by imports) has been much less rapid. The major reason is import penetration: Exports today have a much higher import content in intermediate inputs that in the past. As for imports, if they too were to be produced in Canada, today's technology would have them produced with much more imported intermediate inputs than was true in the 1960s.
  • When our detailed sectoral results are aggregated, Canadian exports have always been above the average for total business output in labour productivity — especially when the effect of agriculture in the 1960s is netted out. However, this relatively high labour productivity of exports has shown almost no tendency to increase over time. Along with higher labour productivity comes higher returns to labour in exports — but again, this relative performance in exports shows almost no trend over time. However, since the share of employment attributable to exports has been growing over time, it can be concluded that exports are helping to lift both the overall level of Canadian labour productivity and returns to labour.
  • The sources of change in the employment mix of Canadian exports are primarily the final export mix itself and the degree of import penetration for intermediate inputs. Generally less important are changes among sectors in relative labour productivity or "technical change" as embodied in changes in Input-Output coefficients. The same observation applies for the most part also to changes in the employment mix of Canadian imports.
  • Changes in export employment shares have occurred across a wider range of industries than might be imagined. Of the ten industry groups that increased their share of export employment by over one percentage point from 1961 to 1997, four are service industries (Personal and other services, Business services, Trade, and Finance), but the other six are various manufacturing categories, some of which might have been expected given the Auto Pact and changing technology — like Motor vehicles, Electrical and electronic products and Rubber and plastic products — and some that are surprising — like Leather, textile and clothing products. Of the four industries with declines in export employment shares of over one percentage point, the two largest declines are in primary production (Mining excluding oil and gas; and Agriculture, forestry and fishing), one is in raw material processing and manufacturing (Paper and allied products and printing) and the fourth is Transportation.
  • Combining the I/O results with data on education mix by industry leads to the conclusion that since 1961 the mix of Canadian export employment (both direct and indirect) has continually evolved to emphasize industries with more highly-educated workers. However, this also turns out to be true of business-sector employment as a whole. If exports are compared to the total business sector in relative uses of different education mixes, the conclusion is that employment from exports is "bi-polar" in distribution. Exports exceed the economy-wide average in employing workers with low education, and also in employing workers with relatively advanced education. The relative ratios have changed only a little over the last three decades.
  • The labour productivity of employment "displaced" by imports (if it were to be produced with Canadian technology) is invariably lower than that of exports — although it is still above the business-sector average for most years. The ratio has fallen to about the economy-wide average in recent years indicating some new inroads into replacing lower-productivity employment. Nonetheless, the evidence is that imports, on the whole, are not replacing exceptionally low-productivity employment (which is probably concentrated in non-tradable services in any event). However, imports are displacing "relatively" more jobs than exports are adding. If macro policies can keep the economy near full employment, and the pace of change is not too fast, this means that Canada is replacing low-productivity employment with high-productivity employment through expanded international trade, and is thereby made better off.
  • Comparing changes in employment shares of exports and imports, it is remarkable that the largest increases and decreases occur, in many cases, in the same sectors. Personal and business services, and Trade show among the largest increases in employment shares for both exports and imports, while Mining, Agriculture and Forestry show among the largest decreases in employment shares for both exports and imports. Both Canadian exports and Canadian imports have become more intensive in service employment and less intensive in employment for the production of raw materials.
  • The education mix calculations show that, at least by Canadian technology and education shares by sector, Canadian imports have been above the business-sector average in overall education content. In recent years, however, this discrepancy has been narrowing, indicating some increased competition from imports in sectors using lower-skill workers. Nonetheless, Canadian imports remain above the business sector average in education content. This result may seem at odds with the previous point about labour productivity and labour returns. However, it should be noted that we found productivity and returns to labour in imports also to be above the business-sector average, although they were also lower than for exports. Finding that the education mix is above the business-sector average is not inconsistent with this. The anomaly, perhaps, is that export productivity and returns are so high when at least part of the education mix is low. The anomaly can, however, be better understood if it is recalled that education, as broadly measured, is not the sole determinant of either productivity or returns to labour; high capital or natural-resource endowments also matter.