Comparison of Mean-Variance and Exact Utility Maximization in Stock Portfolio Selection

In the early 50's, Markowitz introduced the modern portfolio selection theory which, to this very day, constitutes the basis of many investment decisions. Given different correlated assets, how does an investor create a portfolio maximizing the expected utility? Markowitz's contribution was to show that an investor might do very well, relying only on the means and variances/covariances of the assets, which simplifies the portfolio selection tremendously. The validity of the mean-variance approximation to exact utility maximization has been verified, but only in the unrealistic case of choosing among 10-20 securities. This paper examines how well the quadratic approximation works in a larger allocation problem, where investors characterized by different utility functions can choose among nearly 120 securities. The effects of more aggressive investment strategies are also investigated, allowing for limited short selling and the inclusion of synthetic options in the security set.