Heterogeneous fare sensitivity in microtransit: evidence from a natural experiment

Small cities and outer suburbs increasingly turn to on-demand microtransit systems to provide public transportation in low-density areas. As pandemic-era subsidies expire, policymakers need evidence on how fare changes affect ridership and revenue. We study Wilson, North Carolina, which in 2020 replaced its fixed-route bus service with citywide on-demand microtransit. Leveraging detailed rider-level trip data and a natural experiment involving a fare increase, we show that aggregate stability masks large rider-level variation. A two-part structural model further reveals that fare sensitivity differs sharply by usage: low-frequency riders cut usage substantially, while high-frequency riders show only modest reductions. A simulated 40 % fare increase raises revenue by just 7 %, as losses from the large share of price-sensitive riders leave only limited net gains from the smaller group of inelastic riders. Traditional aggregate models that impose constant fare responsiveness risk overstating the fiscal benefits of uniform fare hikes, as our simulations demonstrate. Our framework can be applied to other microtransit systems and technology-enabled transit services with rider records. It also offers guidance for designing stated-preference surveys widely used by transportation agencies, enabling them to incorporate revealed rider heterogeneity into forecasts when rider-level trip data are unavailable. More broadly, this study bridges transportation policy, industrial organization, and marketing by treating microtransit as a market for mobility where pricing, segmentation, and equity interact.