Introduction

In this chapter we survey the more recent literature that attempts to take household models to the data. The first section sets the scene, by reviewing the household utility function approach, or variants of it, to modelling the labour supply decisions of couples. We show what is involved in estimating the model on datasets with missing information on the allocation of non-market time between domestic production and leisure. The following sections review the empirical studies on multi-person household models. We show what goes wrong when household production is assumed not to exist. Our basic contention here is that, as shown in chapter 1, the significance of time use for household production is an empirical fact, not a theoretical hypothesis.

We place a great deal of emphasis on the fact that available datasets, even time-use datasets, are inadequate to allow reliable estimation of structural parameters, and that significant advances in this area require more comprehensive datasets. In the meantime, we need to understand exactly what kinds of data we need that we do not have, and the ways in which assumptions implicit in commonly made empirical specifications effectively construct data. The implications of missing information on pure leisures; time allocations to domestic work, including child care; domestic outputs; and individual consumptions of market and domestic goods will be discussed in some detail.

The household utility function model

The theoretical underpinning

The empirical literature on the labour supplies and consumption demands of two adult households has tended, until quite recently, to focus on the estimation of various specifications of the household utility function (HUF) model.