Extensions of expected utility theory are sensitive to the tail behavior of the portfolio return distribution and may not be approximated reliably through higher-order moment expansions. We develop a novel approach for model risk assessment based on a projection method and apply it to portfolio construction. We provide an extensive out-of-sample analysis to explore the economic gains of incorporating non-normality about financial asset returns into utility maximization with the generalized disappointment aversion (GDA) preferences. We find that the marginal utility gains of the optimal portfolio of a GDA investor are remarkably robust to misspecifications in the marginal distributions but are very sensitive to the structural assumption of stock returns implemented through a factor model.