Prediction Algorithms in Matching Platforms

We provide a general framework to consider the effects of algorithmic demand and supply prediction in matching platforms, such as Task Rabbit for labor markets or Amazon for goods markets, where a significant proportion of matches are repeat transactions. We show that algorithmic repeat match supply prediction can significantly improve the performance of matching platforms by facilitating more flexible wage setting and pricing. The effects on labor welfare and consumer surplus are ambiguous. Demand and supply prediction can exacerbate firm competition, leading to an increase in wages and a decrease in prices. Nevertheless, during unfavorable market circumstances, firms tend to rely on their existing base of workers or consumers so resolutely that it ultimately stifles market dynamism. Consequently, the average wages decrease and the expected prices increase