Sectorial relations and labour productivity

In previous posts (here and here) I've contemplated whether sectorial industry shifts have been more favourable to the median wage in Canada than in the U.S. Perhaps this idea is worth considering further. In fact, it seems like sectorial shifts are often ignored as a possible reason for the difference in the median wage between the two countries.

Case in point: a paper recently released by StatsCan (h/t Brian Ferguson in the comments section of this blog; link via the excellent Worthwhile Canadian Initiative) has isolated the manufacturing sector in an attempt to determine the contributions to labour productivity in Canada in the 1990's. The authors of Competition, Firm Turnover and Productivity Growth conclude: “the competitive process that shifts market share towards more productive firms accounted for about two thirds of aggregate labour productivity growth in Canadian manufacturing from 1989 to 1999.”

Here is my comment (this is a slightly edited version of what I wrote in the comments section of this blog, but I'll repost it here in the hopes of gaining some feedback from others):

One thing that I'm caught up on is how the StatsCan authors isolated the manufacturing sector.

Their argument seems straightforward: increased competition between firms in the manufacturing sector led to reorganization by more efficient means and labour productivity increased. At the same time, incumbant firms were forced to reorganize themselves to attract increased market shares - this is considered another source or increased labour productivity caused by between-firm competition.

However, the authors also briefly address sector-to-sector shifts:

About 45% of firms that were in operation in 1999 are new firms that entered the manufacturing sector during the period 1989 to 1999. These entering firms account for 34% of output and 39% of employment. Most firms that enter the manufacturing sector are greenfield entrants, accounting for 37% of the firms. But their shares in total output and employment are smaller than the shares of merger entrants as greenfield entrants are much smaller than merger entrants.

The authors imply that merger entrants coming from outside the manufacturing sector accounted for a source of labour productivity growth. If so, I can't help but wonder to what degree vertical integration between sectors is a contribution to labour productivity growth through increased profit shares. Yet the authors ignore this possibility in their decomposition. How do they justify this?

In our decomposition, we assume that greenfield entrants replace close-down exits and merger entrants replace divestiture exits.

I wonder if the decomposition in this paper over-emphasizes the impact of turnover by ignoring the change in profit shares earned from emerging sector-to-sector relations (such as vertical integration).