The valuation of travel time savings plays a fundamental role on transport modelling and appraisal. The theoretical foundation for placing a value on travel time savings rests on models of time allocation. Although many alternative time allocation models exist, for many years, this area of theoretical debate was regarded as relatively uncontroversial, with several well-established theories providing a justification for conventional demand modelling practices. In recent years however, the adequacy of these frameworks has begun to be questioned. For example, a number of authors have reported both qualitative and quantitative results which they interpret as indicating that in certain circumstances there can be a positive utility associated with increasing travel time. Other authors have noted that developments in ICT have blurred the classical distinction between activities and travel, enabling some types of activities to be conducted during the course of travel.

In the light of these developments, this paper argues that there is an urgent need to revisit some of the fundamental assumptions underlying classical time allocation models. In particular, we argue that all existing models share a critical weakness in that they assume that a given quantum of time can be used for one and only one activity (or for travel). Drawing on evidence from time use surveys from a number of contexts, we argue that this is a highly unrealistic and limiting assumption, especially in an era of pervasive ICT enabled activities. We argue that failure to explicitly accommodate this type of simultaneous activity engagement underlies a number of the apparently anomalous results recently reported in the literature.

Building on this conceptual position, the paper presents a new framework for modelling time allocation which explicitly accommodates multiple simultaneous activities. This framework is based on an extension of the Goods-Leisure framework of Train and McFadden and the more recent work of Jara-Dias and colleagues (itself drawing on ideas from the earlier work of Evans and De Serpa). We demonstrate how this framework generalises earlier work and provides new insights into the impact of simultaneous activities on the valuation of travel time savings.

As an exercise in the application of this framework, we examine the consequences for the estimation of the value of travel time savings of accommodating various degrees of productive use of time spent travelling, and the consequences of mis-specifying this productivity. This is done by means of a simulation study in which the extended Goods-Leisure framework is used to generate synthetic choice data which are then modelled within a random utility framework using various specifications that omit or mis-specify the impact of the productive use of travel time. We quantify the effect of varying degrees and types of productivity mis-specification on the valuation of travel time savings. The paper discusses the implications of these results for demand modelling and project appraisal.