4. Concluding remarks
In this paper, we analytically prove that prospect theory with a decision-independent reference point can predict the decision behavior in the newsvendor problem with general distributions of demand. Our model is a typical model of prospect theory, where all probabilities are used without weights, or equivalently, the model uses a linear weighting function for probabilities. A more general model of prospect theory is to use a nonlinear weighting function for probabilities [3,11]. Obviously, the prediction power of the model with a nonlinear weighting function for probabilities is stronger than the typical model, because the latter is a special one of the former. In this sense, prospect theory (no matter typical model or general model) is powerful in predicting the pull-tocenter effect for the problem of newsvendor decision.
A number of experiments have been conducted in the literature, most of which are based on the newsvendor setting with either a uniform or a normal distribution of demand. For different motivations, other distributions can be adopted, and decisionindependent reference points when making decisions might be more viable. Our work can be a bridge to connect the experimental study with the model analysis in research on predicting the pullto-center effect.
