Increasing Returns and the Foundations of Unemployment Theory: Rejoinder

Abstract

At first I thought I would reply to each specific point of Mr. Hollaender's comment by addressing, more or less seriatim, the specific issues being raised as best I could understand them. Now I do not think that is the best approach. The issues are too disparate, and I fear getting entangled in a web of semantics and mathematical formulae. Instead, I think it would be more fruitful if I dealt very briefly with what I think are the main themes Hollaender is raising and then used the opportunity to restate as simply as possible my general thesis, taking into account those criticisms that seem to me most relevant. Hollaender's main initial criticism is that if one assumes that utilisation rates of all resource units are identical, rather than assuming as I do that the unemployed are uniformly distributed, then the trade-off between unemployment and the number of firms disappears altogether. This point I do not understand. The analysis should be identical in either case. If I had made Hollaender's assumption I would get the same results as before. Indeed, the expressions N and u do not appear separately in the basic formulas (22)-(25), but only appear as the combination N(i u). The remainder of his note is mostly concerned with the employment effects of flexible wages and entry. While I do not fully agree with his formulation, I think that Hollaender is here making the generally valid point that a fully rigorous argument should contain a more explicit statement about expectations than I have in the paper. Where does all this leave the macroeconomic role of increasing returns? Let me try to restate what I consider the essential issues as simply as possible, eschewing all formalism. Suppose, just for the sake of argument, we lived in a world where there was strict constant returns to scale down to the level of a person or even a grain of sand, or even, if that were necessary, an atom. The standard reasons for macroeconomic failure would continue to hold. Money might be non-neutral, there could be a genuine role for government as a Pareto-improver of social welfare, etc., etc. But, and here is the crucial distinction, there can be no involuntary unemployment with strict constant returns to scale in all aspects of technology. Note carefully the claim: some form of increasing returns is a necessary, but by no means sufficient, condition for genuine involuntary unemployment. Under constant returns, the macroeconomic inefficiencies would show themselves in the form of 'wrong' labour-leisure choices (or, more generally, wrong substitution choices among various factors and commodities). An economy of blueberry pickers, mushroom gatherers, and clam diggers cannot exhibit involuntary unemployment no matter what else is present or absent. It can show fluctuations, inefficiencies, poverty, even starvation, but it cannot show