How Do Ratings and Penalties Moderate Earnings on Crowdsourced Delivery Platforms?

Crowd-sourced delivery represents a rapidly rising segment of the global workforce. Crowd-delivery workers enjoy flexibility in choosing when and where to work. However, such flexibility brings notorious challenges to online platforms in managing the crowd-sourced workforce. Thus, understanding the behavioral and incentive issues of crowd workers in this new business model is inherently important. In this paper, we investigate how a baseline incentive effect of piece-rate earning is moderated by two unique features of online crowd-sourcing platforms, namely, ratings and penalties. Utilizing data from one leading crowd-sourced grocery delivery platform with more than 50 million active users in China, we implement a two-stage Heckman model together with instrumental variables to tackle this research question. We first show the baseline effect whereby higher piece-rate earning increases the workers' work time. Going one step further, we find this positive effect of piece-rate earning decreases when the percentage of five-star ratings increases (negative moderating effect); moreover, this positive effect increases when the monetary penalty increases (positive moderating effect). Finally, we show the magnitude of the negative (positive) moderating effect decreases when the percentage of five-star ratings (monetary penalty) increases. The ultimate goal of this paper is to provide guidelines to online platforms on the design of better incentive mechanisms with the interplay of piece-rate earning, rating, and penalty.