Algorithm Aversion in Distribution Channels

Algorithms are taking over decision-making in business. In this paper, we study the effects associated with algorithm aversion in a distribution channel comprising one supplier and one retailer. In this context, algorithm aversion is defined as a cognitive bias that the retailer obeys inferior human decisions, even if it understands that algorithmic decisions outperform itself. Although algorithm aversion bias always leads to a lower expected profit for a centralized channel, we find that it can be a positive force for the decentralized channel, depending on the contract form between the supplier and the retailer. Specifically, under a wholesale price contract, the supplier can take advantage of the algorithm aversion at the expense of the retailer. However, under a two-part nonlinear contract, the algorithm aversion can benefit rather than hurt the supplier, the retailer, or the channel as a whole