Autonomous Cars and Dynamic Bottleneck Congestion Revisited: How In-Vehicle Activities Determine Aggregate Travel Patterns

Abstract

We investigate the impacts of in-vehicle activities of commuters in the autonomous car on aggregate travel patterns. We allow for an autonomous car to affect the utility difference between being at home and being in the vehicle differently than the utility difference between being at work and being in the vehicle, compared to the differences experienced with a normal car. This affects the relative importance of values of travel delays, schedule delays early, and schedule delays late. Hence multiple possible changes in travel patterns may occur when autonomous cars become available. Switching to an autonomous vehicle may impose a net negative or positive externality, by raising the marginal external cost of autonomous cars themselves while lowering that of normal cars. We examine three provision regimes: marginal cost pricing, second-best pricing and profit-maximizing pricing by a private monopoly. The second-best mark-up (over marginal cost) rises with the price sensitivity, due to the increasing marginal external cost. Surprisingly, for the monopoly, mark-up may rise or fall with the price sensitivity, depending on the relative strength of the externality and of market power, where the former tends to raise it, and the latter tends to reduce it. Furthermore, the difference of the mark-up between private monopoly and second-best public provision falls as the demand becomes more price-sensitive.