Abstract Most mathematical programming procedures used for optimizing capital rationing decisions consider each decision in isolation. They implicitly require that the decision maker shall know, at the time of decision, all the investment opportunities that will appear between that time and the horizon time of the decision, and further, that all assets will be converted to cash at the horizon time. The research described in this article used a simulation methodology for studying the long term effectiveness of expected net present value maximization in an environment of incomplete and uncertain information. Some tentative conclusions are drawn from its use in long sequences of decisions in hypothetical firms: consistent use of a Rank:on-Net-Present-Value selection procedure will tend to maximize the expected capital growth rate of the firm if the discount rate is properly selected and expected cash flows are accurately estimated; however, failure to meet either of these conditions can have a seriously adve...