The Costs of Recession: A Reply

Abstract

Pemberton (I 995) raises the important issue of what is the most appropriate way to measure business cycle costs from an individual perspective. As always in these matters it is not nor should it be a case of outright victory of one side over the other. Rather, it is a question of which approach is the more plausible and on this we would wish to defend our shade of grey as being the more plausible measure. This brief reply explains why. Imagine the simplest possible textbook experiment used to explain some of the principles of expected utility maximisation. A prize in our case employed consumption C can be won with a given probability and the alternative is the booby prize of unemployed consumption. Let C be the expected value of this gamble and let C be the amount for which the player would be indifferent to the gamble. The risk premium is then simply the difference between C and C. This is exactly what is given by equation (9) in Clark et al. (I994) (equation (2) in Pemberton's comment), where the risk premium is measured as a proportion of employed consumption. So there is nothing mysterious about our recession cost measure it is simply the standard textbook risk premium. It would seem to us to be a very natural experiment to consider when measuring the cost of agents' exposure to risks and we termed it the net cost measure. Our gross cost measure is given by i C/C which is described by our equation (7). This is just the risk premium plus the difference between employed consumption and expected consumption (again all measured relative to employed consumption). We argued that the gross measure was one which would appeal to Keynesians who, because they view cycles as disequilibrium events, would wish to include losses in consumption as part of the costs. Pemberton's statement that, 'A high equilibrium (natural) rate of unemployment is surely a major problem', is an implicit endorsement of the Keynesian cost measure, in our view. The net measure, because it measures the risk premium, obviously abstracts away from any consumption losses as a result of drawing the booby prize. This is exactly the same idea as Lucas in his proposed cost measure who, because of his view that business cycles are equilibrium events, does not accept consumption losses as a proper part of costs. The Lucas measure and our own compare a risky consumption path with a riskless stream. Our measure differs from Lucas, however, in one important respect. From the macro perspective that he adopts, the important comparison when measuring economic fluctuations and their impact on the welfare of representative agents is the deviation of consumption from its trend path. Thus Lucas' recession cost measure depends on the variance of consumption around that path. Unusually high aggregate consumption is as costly as unusually low