ABSTRACT
This chapter aims to handle preferences while at the same time recognising the piecewise linearity of the budget constraints. It describes construction of the data sets. The chapter sets out the model and provides the estimation procedures. It also describes the results on endogeneity bias and the effects of including preference variables. The chapter presents the argument in favour of modifying the estimation procedure. It also sets out reasons for differences in endogeneity bias between samples. The chapter discusses other specification errors. It focuses on methodology and the labour supply elasticities of married women and married men. The chapter explains supply function of a quadratic form in the wage and intercept has been directly estimated for two British samples, of married women and married men. The results indicate a large bias in the linear approximation procedure for the sample of married women but very little bias for the sample of married men.
