Measuring Travel Time Values with a Discrete Choice Model: A Note

In a recent article in this JOURNAL, Truong and Hensher (I985) set out to perform a useful service by relating the theory of time allocation to the framework of discrete choice models, thereby allowing practical estimation of the value of time on a sounder theoretical basis. Unfortunately, there are a small number of crucial misunderstandings in their work, which lead to confusion and, in particular, invalidate the conclusions they draw from their presented empirical work. This note aims to correct the errors of theory in Truong and Hensher's paper, and to draw some brief conclusions from the empirical evidence that they give, in the light of these corrections. Since most of what they say can stand, only the essential points are repeated here, and considerable reference will be made to the equations in their paper (indicated by TH before an equation no.). The authors begin by outlining the seminal work in time valuation of Becker (I965) and DeSerpa (I97I). Both theories rely on utility maximisation. Since DeSerpa's approach can be viewed as an extension of Becker's, we will set out its essentials, as given in the authors' paper. For the purpose of the problem in hand, we assume that individuals have direct utility functions which are dependent on G, the volume of goods and services they can buy, L, the amount of 'leisure' time they have, and T the amount of time they have to spend travelling. Travel can be by different modes i, involving different costs and travelling time. Since total money and time budgets are fixed, these travel costs and times impact on the amount of other goods and the amount of leisure time available. The problem can be set out as follows: Max u (Gi, Li, Ti) (TH I 2') such that G1 < M-C , (TH I3)