MAKING INFLEXIBLE INVESTMENT DECISIONS WITH INCOMPLETE INFORMATION

This paper presents an application of the Principle of Maximum Entropy to business investment decision making under uncertainty. This principle is derived explicitly from a set of axioms for rational inference, one of these axioms being the weak form rational expectations hypothesis (REH): that agents use all the available information efficiently. This set of assumptions are weaker than those embedded in the strong REH. It is demonstrated that the Maxent algorithm leads to rational decision solutions when the information set available to agents is very limited-so limited that there is no orthodox solution to the problem. Thus, we obtain a solution without the extreme informational assumptions necessary to obtain a strong REH solution.