ROBUSTNESS OF THE MEAN-VARIANCE MODEL WITH TRUNCATED PROBABILITY-DISTRIBUTIONS

The known sufficient conditions for the mean-variance framework to produce expected utility results are violated in the presence of truncated probability distributions. A theoretical simulation is conducted to examine the ability of the linear mean-variance model to approximate expected utility results when the income distribution is truncated by the use of commodity option contracts. The mean-variance model is shown to produce solutions that are close approximations to the expected utility model results under the assumptions of constant absolute risk aversion and normally distributed prices. However, some inconsistency was found between the comparative static results of the two models.