Structural Estimation of the Newsvendor Model: An Application to Reserving Operating Room Time

Abstract

The newsvendor model captures the trade-off faced by a decision maker that needs to place a firm bet prior to the occurrence of a random event. Previous research in operations management has mostly focused on deriving the decision that minimizes the expected mismatch costs. In contrast, we present two methods that estimate the unobservable cost parameters characterizing the mismatch cost function. We present a structural estimation framework that accounts for heterogeneity in the uncertainty faced by the newsvendor as well as in the cost parameters. We develop statistical methods that give consistent estimates of the model primitives, and derive their asymptotic distribution, which is useful to do hypothesis testing. We apply our econometric model to a hospital that balances the costs of reserving too much versus too little operating room capacity to cardiac surgery cases. Our results reveal that the hospital places more emphasis on the tangible costs of having idle capacity than on the costs of schedule overrun and long working hours for the staff. We also extend our structural models to incorporate external information on forecasting biases and mismatch costs reported by the medical literature. Our analysis suggests that overconfidence and incentive conflicts are important drivers of the frequency of schedule overruns observed in our sample.