The Allocation of Assets Under Higher Moments

We evaluate how deviations from normality may affect the allocation of assets. A Taylor expansion of expected utility allows us to focus on certain moments and to compute numerically the optimal portfolio allocation. A decisive advantage of our approach is that it remains operational even if a large number of assets is in-volved. We obtain that for small values of the risk-aversion parameter, non-normality does not alter significantly the optimal allocation. In contrast, when the investor is strongly risk averse and restricted to invest in risky assets, we also obtain significant changes in portfolio weights.