Market dynamics between public transport and competitive ride-sourcing providers Renos Karamanis *1 , Panagiotis Angeloudis 1 , Aruna Sivakumar 1 , and Marc Stettler 1 1 Centre for Transport Studies, Imperial College London March 15, 2018 1 Introduction Recent advances in communications and IT technology have underpinned the rapid development of trans- portation network companies (TNCs) in recent years [8]. Although the competitive prices offered by TNCs have been well-received by users, recent studies indicate that the resulting increase in public welfare [6] is temporary, as passengers are likely to be steered away from public transport. Indeed, a recent study by [5] has presented evidence that such shifts have already occurred and can be directly linked to increased congestion in major cities. Many studies in taxi pricing [e.g. 4, 9] have adopted an economic theory perspective, using aggregate demand and supply models to represent the dynamics of urban taxi operations. Dynamic pricing has also been considered, for example by [7] who adopt a Vickrey-Clarke-Groves (VCG) bidding mechanism for shared autonomous taxi rides to maximise social welfare. The impact of dynamic pricing to TNC operations was investigated in [3], suggesting that a dynamic increase in trip price significantly increases the supply of rides in the system. However, the relationship between dynamic TNC pricing strategies and public transport provision remains unexplored to this date. To address this issue, we propose a novel, game-theoretic, dynamic pricing model that accounts for multiple TNCs operating alongside public transport services. This is applied to a city-wide service scenario, and compared to an alternative static pricing model that serves as a baseline. Finally, we perform a comparative analysis of expected utilities for travellers and operators, while monitoring mode share fluctuations across a range of competitive scenarios and market structures. 2 Model Description The key actors in our model are two symmetric TNC firms offering an identical product (rides) of equivalent quality. A centralised platform receives ride requests from the public. Both firms are expected to respond to these requests, with quotes for service and estimates of anticipated wait and travel times. Travellers can then decide (using a generalised costing mechanism) whether to accept an offer or to revert to public transport. TNCs are expected to operate with a profit maximization objective and are allowed to introduce surcharges upon the base prices for each trip. These are based upon current demand levels, and a TNC’s perception on the ability of its competitors to serve a specific trip request. Using dynamic pricing, the final bid price is the sum of a variable base price r per time of travel set by the platform, and the TNC’s choice of an extra variable tariff per time of travel f i . * Corresponding author: renos.karamanis10@imperial.ac.uk (R. Karamanis) 1