Introduction

The use of observed income and expenditure variables as measures of welfare has a long tradition in studies of taxation and inequality. An objection to the practice is that the variables omit the contribution of nonmarket time. With the increasing availability of household datasets attention has focused on the estimation of labour supply models for the purpose of making welfare comparisons based on a utility function defined on consumption and leisure, with leisure measured as nonmarket time. The approach has been used extensively for analysing reforms to the taxation of married couples. Examples include Arrufat and Zabalza (1986), Blundell et al. (1986, 1988), Zabalza and Arrufat (1988) and Symons and Walker (1990). An obvious criticism of the approach is that the nonmarket activities of married women do not easily fit the conventional notion of ‘leisure’. Much of their time is spent on housework, producing goods and services for which there are substitutes in the market place, and so it may be more appropriate to treat domestic activity in the same way as market activity, as suggested by Becker (1965).

A long-standing and central concern in the taxation of families is that of horizontal equity in the treatment of those with different time allocations to household production and market work. Conflicting interpretations of what the criterion implies for policy typically reflect different weightings on household production in the calculation of the welfare indicator used to assess the distributional merits of a particular reform.