A utility-based framework for evaluating the distributive effects of public transport projects on perceived accessibility

Cost-benefit analysis, largely applied in transport appraisal, is criticized as utilitarian because it sums monetized gains across groups, but not because it uses utility as information. Yet the utility-based accessibility measure, the logsum of a discrete choice model, has been kept at the margin of equity analysis on the assumption that it carries the same commitment. Drawing on Sen's distinction between welfarism and sum-ranking and on the capability approach, this paper argues that a logsum of perceived accessibility can inform distinct distributive principles, provided it is expressed in units that remove its arbitrary level and model-specific scale, and proposes such a unit, job equivalents. The measure is applied to the ex-post evaluation of the metro and light rail lines of Fortaleza, Brazil, for low-income workers, with a destination-mode logsum estimated from stated and revealed preference data, segmented by gender and income and conditioned on the modes each person can use; census data show that the lines did not detectably change land use around the stations. In the case study, in job equivalents the lowest-income strata gain most, whereas in money higher-income men do, a reversal explained by modal preferences and by the weight of impedance in destination choice; the benefit concentrates along the corridors and bypasses the poorest periphery; and assuming universal mode availability understates the benefit to captive transit users two- to threefold. A structural equation model on the household travel survey shows that the benefit converts only partially into rail use, as expected, accounting for about seven percent of the association between proximity to the stations and use; the rest runs through channels outside the transport benefit as measured, such as residential self-selection, crowding on the trains and the lack of fare integration.