Search

A Priori Economic Theory is unable to predict whether the supply of labour will increase or decrease as the wage rate changes. Political interest in the effect of income maintenance schemes on the incentive to work has stimulated a spate of empirical research into labour supply based on the simple text-book theory of labour supply. Unfortunately the empirical work to date has produced an unacceptably wide range of labour supply estimates. One reason for the poor results is that the assumptions of the simple text-book model of labour supply don't take account of important variables in the real world. However even if we assume the simple text-book theory of labour supply to be theoretically adequate, the conventional procedures used to estimate labour supply based on the simple text-book model are theoretically invalid, econometrically unsound and suffer grave measurement problems with one of the key variables required by the method. Furthermore the functional form normally used has been highly restrictive therebye reducing the tax policy relevance of this research, most of which has been undertaken with a view to estimating the effect of income maintenance schemes on labour supply. A new procedure is discussed to estimate labour supply in which hours worked is regressed on perceived net marginal wage rate, net average wage rate and non-employment income. The new model represents a marked improvement over the conventional procedure which regresses hours worked on the average wage rate and non-employment income, because unlike the conventional procedure it is theoretically correct in a world with non-linear budget lines, and because price income and substitution effects can be estimated independently of the (dubious) non-employment income coefficient therebye avoiding theoretical econometric and data measurement problems associated with non-employment income. The new average/marginal procedure is also superior because it employs a functional form which is not highly restrictive such that it allows the labour supply estimates to vary more freely over the income distribution and should therefore be more relevant for tax policy. Furthermore the average/marginal procedure overcomes the problem of spurious correlation between hours worked and the average wage rate arising from error in the measurement of hours worked. There is however one remaining econometric problem intrinsic to the conventional procedure which is not resolved by the new average/marginal procedure, namely that the average wage rate is endogenous where average and marginal wage rates are unequal. The new average/marginal procedure is used to obtain labour supply estimates for a cross section sample of British weekly paid married men. Price income and substitution effects are calculated at different points over the sample distribution of wage rates to ascertain the effect of the flexible functional form used, and this is found to have a profound effect insofar as the magnitude of the substitution effect decreases as a function of the average wage rate. The regression estimates are also used to derive the implied (skeleton) indifference map for income and leisure. Finally the implications of the estimated labour supply function with respect to income maintenance schemes are discussed, using alternative negative income tax schemes on a hypothetical population to illustrate the issues.