I've been following the Interpreter's debate (1, 2, 3, 4, 5) on US manufacturing with interest, not least because it took my mind back to some old debates about development and deindustrialisation. And of course we are also starting to see a fair bit of discussion here in Australia about the merits of supporting our own manufacturing base.

Economists (especially development economists) and economic historians, have spent a lot of time thinking about how the broad structure of economic output changes with the level of development. Based on this work, the 'typical' pattern of economic development is usually assumed to involve a transition from a low-income, agrarian economy to an industrial economy with a higher level of income, with big relative shifts in the share of output and employment, away from agriculture and towards industry in general and manufacturing in particular.

In other words, at the lower end of the international income distribution, development is often treated as synonymous with industrialisation, albeit with significant variations across time and space in how that industrialisation plays out. (This is one reason why there has been so much interest in India's development path, which to some observers seems to offer the possibility of an alternative path to development.)

In contrast, for countries with higher levels of income per head, the lessons of the twentieth century seemed to be that structural change typically involved deindustrialisation, or alternatively 'tertiarisation,' as industry's share of output and employment shrank, and resources were transferred into the services sector. The big drivers of this shift were relative productivity growth, the associated changes in relative prices, and shifts in the structure of demand between manufacturing and services.

For example, since labour productivity tended to grow more rapidly in manufacturing than in services, while output growth across both sectors was similar, lagging service sector productivity meant that the services sector account for a growing proportion of employment. At the same time, the demand for services has grown more rapidly than the demand for manufactured products. It's also likely that international trade has played a supporting role in the process, although the main drivers appear to have been domestic.

In this view of the world, then, (and it's basically a view I share), while successful development usually looks like industrialisation to start with, after a time, continued development looks more like deindustrialisation. It follows that worrying about a declining share of manufacturing in output or employment doesn't make too much sense: this is a shift that is largely a natural outcome of the development process, and resisting it is like trying to turn back the tide. Moreover, since many of the proposed responses to this shift involved either dubious or outright harmful economic policy suggestions, including protectionism, there is the associated risk of mistaken policies creating significant collateral economic damage.

So, a case of no worries then? Well, possibly, although I have to admit that it's probably not quite as cut and dried as all that. There are still a few issues here that I don't think are settled:

  • To the extent that the output of the manufactured sector is more likely to be highly tradable than that of other sectors, then for countries that either currently or will at some point in the future, face an external adjustment requirement, manufacturing might still matter. (That said, of course, the tradability of services is on the rise.)
  • Then there is the issue of employment, which is touched on in Hugh's post and in the comments from Geoff Miller. There seems to be evidence that the United States and other developed economies have seen a decline in middle-wage, middle-skill jobs, resulting in a polarisation of their labour markets, with growing employment opportunities at the top and bottom of the wage and skill distributions set against a contraction in the middle. The relative decline in manufacturing employment appears to be one contributory factor here, and it does appear that this trend is associated with troubling consequences for inequality and other social indicators.
  • To look at things from a slightly different perspective, it's possible that the real problem is not so much that the manufacturing sector has grown too small, but that other sectors have grown too large. (I'll take this one up in a subsequent post.)

Two last points:

Photo by Flickr user caffeina liquida.