Effects of numerical platforms on individual choices and social welfare: The case of short-distance carpooling

This paper investigates the effects of carpooling organization on the propensity to carpool. We test whether pre-trip planning and the use of a platform influence the choice of carpooling over driving alone and public transport. In a stated choice experiment, we collected responses from 3600 residents of the Lyon urban area, France. Our econometric results suggest that platforms increase individual willingness to carpool, and that the effect of the platform is stronger for passengers than for drivers. We illustrate these results with a stylized social welfare analysis, which reveals a significant contradiction between what would be needed to make carpooling beneficial from a welfare perspective – passengers should be subsidized – and the current pricing schemes, which subsidize mainly drivers. • Carpooling platforms are effective to build trust. • Carpooling platforms are more effective for passengers than for drivers. • Carpooling platforms are more effective for women than for men. • Maximizing social welfare involves incentives that focus more on passengers than on drivers.