Aligning the interests of newsvendors and forecasters through blockchain-based smart contracts and proper scoring rules

Abstract

We consider a newsvendor setting where the newsvendor elicits a demand forecast from an expert to determine the optimal inventory level for a product. Since the interests of both parties might not be aligned, we propose a proper scoring rule to elicit the expert's forecast that is tailored to the decision problem the newsvendor faces. Under the proposed scoring rule, both the newsvendor's expected profit and the expert's expected payment increase as the expert exerts a higher effort when gathering information to formulate a forecast. Trust issues arise in this setting because the ex post payment the expert receives is contingent on the accuracy of the reported forecast, which in turn depends on a future realized outcome that might be only available to the newsvendor. We show that coding the proposed proper scoring rule as a blockchain-based smart contract allows the newsvendor to signal trustworthiness while, at the same time, inducing the expert to devote more effort in information gathering. Following the design science research framework, we conclude the paper by suggesting a design and a fully functional prototype of a blockchain-based smart contract that implements the proposed tailored proper scoring rule.