production of capital goods

The whole philosophy of macroeconomics since Keynes more or less invented it has been to aggregate variables like "consumption" and "investment" and ignore differences between different kinds of consumption, as well as structural details about what specific goods are used to produce what other specific goods. The approach has its uses, and macroeconomists have shed some useful light on the workings of the economy this way, but even on what one would think of as a macro basis it seems likely that some details will make differences in some circumstances. One of the big changes over the past generation or two that might be important is the changing nature of the production of capital goods.

Forty years ago, "capital" meant heavy machinery — factory equipment, earth movers, that sort of thing. People who produced capital goods were to a large extent machinists and factory workers. Today a lot of "capital" is software. Software is "nonrival", meaning that producing software for 1,000,000 customers is not largely more expensive than producing the same software for 10 customers; it is also the case that a lot of software production builds on previous versions of similar software. "Fixed investment" has been supporting the recovery to a greater extent than in previous recoveries, but in 2011 that means more software and fewer tools than it would have 30 years ago.

Institutional capital, which I'm largely leaving out, may also be more important now than it was 40 years ago; more of the labor force consists of people for whom an important part of their value to their current employer is detailed knowledge of coworkers, workplace culture, and procedures than I think was the case forty years ago. This is especially true in the production of modern forms of capital as compared to production of older forms of capital; engineers and computer programmers working on projects too big for any one of them to complete alone are harder to replace with other experienced employees with the same generic training than is the case for machinists. (This is not an absolute truth, but is broadly the case; certainly any sizable company will benefit from employees who know the idiosyncrasies of that company, or even of its particular workplaces. I believe it to be more true, in general, of engineering kinds of work than of machining or factory work.) Related human capital is also likely to be more important for more modern forms of capital than for older forms.

Confident answers are not in the purview of this blog, but it seems reasonably likely to me that this contains a partial explanation for the slow recovery of employment after recessions that has been increasingly witnessed over the past 25 years. When demand shows its first signs of renewal firms may turn first toward replacing their capital, whose prices are more likely to drop than are wages; the producers of capital themselves need not hire a lot of new workers nor raise the prices of their products a great deal until demand is quite substantial, and (especially in times of uncertainty) may be slow to increase employment too quickly because of the investment this would require in institutional capital as well.