Abstract The article reports that on a time-state preference simulation model, it has been shown that there is little synergistic effect of combining a net present value objective function with a payback constraint and there is a possibility to increase greatly the performance of a net present value model by assigning projects to two or three risk classes and using a different discount rate for evaluating projects in each risk class. From this, several tentative conclusions may be drawn. First, the increasing use of a payback period as a secondary criterion in capital budgeting is not completely irrational since it does improve the decisions of a straight NPV model in highly uncertain environments. However, the NPV-PBK model in general leaves much room for improvement. Second, the dominant use of the NPV model seems warranted, provided different discount rates are used for different risk classes of projects. From the performance levels when only two or three risk classes are used, it appears that the NPV-RC model provides decisions nearly as good as even the most sophisticated models.
Gary L. Sundem (Mon,) studied this question.