[...]

The idea that people can readily be switched from one line of work to another would appear to stem from the idea that labor time is a commodity with a coherent meaning, and this notion is an extension of nineteenth-century abstractions about labor that have lost their slight purchase in real world conditions over the course of the present century. The manual worker with general skills hired out by the day for odd jobs at a negotiable wage is a fringe case. Everybody else is linked to a social network that dictates within broad bands terms of employment specific to his or her skills and background. The small actions that lend intuitive plausibility at the micro level to the concept of a market for fish ("Atlantic Salmon $5.99! Special Today!") are never observed in the so called market for labor. Wages are not set in response to the short-term variations of supply and demand, but rather by a complex process of relevant comparisons, within and across occupations, industries, and the characteristics and qualifications of the worker.

Most economists seem to have forgotten that John Maynard Keynes quite powerfully demolished the "supply curve for labor" in the opening pages of The General Theory of Employment, Interest, and Money. Keynes showed that there was no reason to expect that, say, an excess of unemployment would drive down real wages. He showed that even with high unemployment, the remaining workers would still rationally resist reduction of their money wages; moreover, even if their resistance failed, the subsequent fall of money wages would bring down prices, leaving real wages unaffected. Thus, labor markets do not respond like fish markets to excess supply. The "second postulate" of the classical doctrine, the supply curve for labor, failed as a logical construct, and Keynes threw it away, drawing an instructive analogy to the overthrow of Euclid's axiom of parallels. But, of course, the nonexistence of a viable supply curve implies that the market for labor itself is not a market in the meaningful sense of that term. Without a supply curve, there is no market, and the "equilibrium" of wages and employment cannot be determined in it. One is forced to look outside the classical confines of the labor market to find the determination of employment and of wages. In other words, one needs once again to build a macroeconomic, and specifically a Keynesian, theory.

[...]

If you happen to think that the performance of the American economy is on the whole poor - too much poverty or unemployment or inequality, for example - and that something ought to be done about it, what place is there for you under the macroeconomists' tent? The natural rate/NAIRU answer is no place at all. Policies to attack social problems belong in the micro sphere: education and training, infrastructure, welfare and welfare reform. The frustration that this produces, as when training is provided for jobs that do not exist, goes unaddressed.

And if the concept of an aggregate labor market could be wiped at a stroke from the professional consciousness, what would happen? Plainly, there would then be no reason to associate any particular value of the unemployment rate with rising wages and hence with rising prices and inflation. The concepts of the natural rate and the NAIRU, already embattled because of the failure of empirical predictions based on them, would certainly collapse. Economists would be obliged to find ways of evaluating the evidence governing both inflation and unemployment without granting privileged status to the idea that the two are closely linked. The policy notion that unemployment is a sensible means of controlling inflation would lose most of its power.

I believe this would be an enormous intellectual improvement, for it would divert research from the ephemeral pursuit of abstract and elusive scalars (Where is the natural rate, exactly? 6.0 percent? 5.5? 5.0?) into the analysis of a much more complex realm of data, such as already characterizes the more productive veins of research in labor and financial economics. It would also expand the scope of acceptable policy discussion. It would turn many thousands of unemployed, now abandoned to fate, into reasonable candidates for reemployment on reasonable public or public-private projects - physical, intellectual, and cultural - at reasonable terms. But for this to happen, it is evidently not enough just to raise doubts about the labor market theory of aggregate wages. For if wages are not determined in the labor market but in context- and institution-specific patterns, what exactly are these patterns and how are they to be made into legitimate objects of social inquiry?

[...]