4 Differences in industrial organization across Canadian regions

View the most recent version.

As noted above, integration into world markets can result in changes in the organization of production within plants. This section provides a descriptive analysis of the degree to which regional economies have become more integrated into world markets and whether this is associated with changes in the organization of production within plants.

Over the past quarter century, Canadian manufacturing plants have become increasingly integrated into world markets. The export intensity (exports divided by sales) of the manufacturing sector as a whole more than doubled over a 25-year period—increasing from 0.18 in 1974 to 0.43 in 1999.⁵ The average export intensities of census divisions follow a similar, albeit more muted, trend. The average census division included in our study increased its export intensity from 0.23 in 1974 to 0.37 in 1999 (see Table 2).

Across regions, census divisions in Atlantic Canada and Western Canada started the period with export intensities that averaged above those in Ontario and Quebec. Over time, however, the export intensities of Ontario and Quebec's census divisions increased at a faster pace than did those in either Atlantic Canada or Western Canada. As a result, by 1999 the export intensity of Ontario had passed that of Western Canada and was nearly that of the Atlantic region. Meanwhile, Quebec had passed Western Canada, while its census divisions still lagged the average export intensity of those in Atlantic Canada.

It is noteworthy that Western Canada's export intensity declined significantly from 1996 to 1999. At least in part, this reflected declining demand resulting from the Asian financial crisis in 1997. The picture that emerges is of a manufacturing sector that has become increasingly integrated into world markets, especially in Canada's industrial heartland.

Table 2
Average export intensity of regions and Canada, selected years

The question at hand is how this growing integration is correlated with shifts in the plant size, the number of products ( NE) per plant and the length of production runs within these broad regions. We measure these characteristics, at the regional level, by taking the average of plants' characteristics for each census division and then taking the average of these across the same regions presented in Table 2.

In accordance with our models, the largest plant size is found in Ontario (Table 3). Over the past quarter of a century, all regions increased their plant sizes, but it was in Ontario, which is closest to the American heartland, where the average plant size increased the most. At the start of the period, average plant size was highest for census divisions in Ontario, followed by Atlantic Canada, Western Canada and Quebec. Although by the end of the period, census divisions in all regions had higher average plant sizes, plants in Ontario more than tripled in size, while those in Quebec and Western Canada doubled and those in Atlantic Canada increased by about 70%. This has resulted in an increasing gap between the average sizes of plants in census divisions found in Ontario compared with the other regions.

The difference in the rate of increase in plant size across regions mirrors differences in changes in export intensity over time. For instance, Ontario experienced the largest increase in export intensity and the largest increases in plant size.

Table 3
Average census division plant size by region, selected years

The largest industrial region also had the largest number of products per plant, as predicted by our models; changes in the number of products per plant increased, as predicted, with the increasing integration into North American markets (Table 4).

Trade integration can lead to scale economies being more fully exploited, by increasing plant size and/or by decreasing the number of products produced in each plant. As Table 4 indicates, the numbers of products were being reduced throughout the period at the same time that plant size was increasing. There was a decline across all regions in the number of commodities (measured using NE) produced in the average census division's plant. At the beginning of the period, the average plant within the average census division produced two and one quarter commodities in all regions, except for Ontario. In Ontario, this figure was just under two and one half commodities. By 1999, this average fell to fewer than two commodities. The rate of decline in the number of products produced was about the same across regions, with a slight advantage for Ontario.

Table 4
Average census division commodities per plant, by region, selected years

The analytical framework that informs the analysis also predicts that the average length of a production run should be higher in the larger provinces that are closer to the focal point of the North American market, and it is. Ontario has the longest production-run lengths throughout the period (Table 5).

The effect of larger plant sizes and falling commodity diversity per plant was to increase the length of production runs across all regions (Table 5). In Atlantic Canada, Western Canada and Quebec, production-run lengths, in the average census division, increased by roughly the same amount from 1974 to 1999. In Ontario, the length of production runs increased by a little over fourfold, reflecting both the stronger increases in plant sizes in Ontario and declines in number of commodities per plant.

Table 5
Average census division production run length per plant by region, selected years

Overall, the picture that emerges is of a Canadian manufacturing sector that is increasingly integrated into export markets across all regions. At the same time, there have been increases in plant size and increases in the length of production runs, both of which should have allowed scale economies to be more fully exploited. These scale-enhancing changes were most marked in Ontario. This was also the economy that experienced the most rapid integration of its manufacturing sector into export markets.

In order to further investigate these associations, we pursue a multivariate analysis in the succeeding sections that allows us to control for variables other than trade intensity that might be associated with the average characteristics of plants across geographic units.

 

⁵ . Special tabulation of the Annual Survey of Manufactures.