The Incentive Game under Target Effects in Ridesharing : A Structural Econometric Analysis

To wield a flexible self-scheduled supply to match the ever-changing demand and maintain market shares, ride-sharing platforms such as Uber and Didi strive to keep more registered drivers active on the road, especially during peak hours in which the demand tends to be the highest. Platforms have been providing monetary rewards to incentivize self-scheduled drivers to work longer. However, drivers' responses to bonuses can be affected by the target effect, which if present may undermine the effectiveness of such bonus schemes.In this paper, we provide a theoretical and empirical analysis of the effectiveness of the monetary bonus scheme in the context of the ridesharing industry, where the drivers are highly heterogeneous and set income targets. We first model driver's decision-making processes and the platform's optimization problem as a Stackelberg game, and show that driver's working hours do not increase monotonically with the bonus rate under the target effect and that the platform may not use all its budget on bonuses to maximize capacity or profit. Then, we develop a novel empirical strategy and perform a reduced-form analysis to provide evidence on the existence of the drivers' income-targeting behavior utilizing comprehensive datasets that were obtained from a leading ridesharing platform. Using the equilibrium outcome from the driver's problem, we also structurally probe the existence and the magnitude of the target effect for each driver. Finally, we perform a counterfactual analysis to calculate the optimal bonus rates for different scenarios utilizing the characteristics of heterogeneous drivers derived from the estimation outcomes. We find that, compared to the platform's previous bonus-setting, the optimal bonus strategy improves the capacity level during peak hours by as much as 24%, boosting the total profit by $5.1 million per month