We are grateful to Whitley and Wilson both for investigating our proposed Marginal Employment Subsidy (MES) using a full-scale macro model and for correcting our arithmetic. With reference to the latter point first of all, luckily the revised numbers are not too different as a comparison of the first two columns of their Table i indicates, although they do demonstrate the superiority of computers to pencil and paper. Interestingly enough, under the revised figures the budget deficit cost per job of the MES is dramatically reduced, making it rather more desirable than we had originally supposed. This improvement is, however, more apparent than real since the budget-deficit effect is the difference between two very large and almost equal numbers (see the formula in LN). A small change in the assumption concerning, for example, the marginal rate of corporation tax makes a lot of difference. We are severely taken to task by Whitley and Wilson for asserting that the current state of the art in economics renders it impossible to analyse the policy effects in the context of a fully dynamic model. What we meant by this was not that we consider the existing UK macroeconomic models to be unsuitable but that there exists no theoretical dynamic macro model which commands wide enough support among economists to enable us to generate dynamic responses which would convince more than a small proportion of the profession. This absence of consensus does, in fact, spill over into the empirical models as well in the sense that with the key wage equation, for example, the difference in specification across the various large models is very large indeed and generates very different responses to certain types of policy, particularly in the short run. Concerning the conclusion of Whitley and Wilson that MES is only slightly superior to a non-marginal employment subsidy (NMES) (e.g. a cut in National Insurance) a reading of both papers makes it crystal clear that the important difference between us is in the impact of such a policy on price taking exporters. There are two points to be made here. First, export equations of the type which appear to have been used by Whitley and Wilson, which are specified in terms of prices but which omit costs, are obviously mis-specified with regard to the price taking sector since they cannot capture the essential supply behaviour of the price taking firm. Second, and perhaps more important, even an equation which in addition incorporates profits or better still relative labour costs (e.g. the Treasury model) cannot help us discriminate between a marginal versus average subsidy since the measures of profitability or labour costs which are used are average rather than marginal concepts. Since the average subsidy of an MES is the same as that for an NMES, use of an export equation specified [ 88I ]