The recent experience from the global financial crisis has raised serious questions about the accuracy of standard risk measures as a tool to quantify extreme downward risks. These standard risk measures, such as the VaR, emerge over the last decades as the industry standard for risk management and asset allocation (Basak and Shapiro [2001]; Montfort [2008]). We estimate the riskiness of risk models and we evaluate its impact on optimal portfolios at various time horizons. Based on a long sample of U.S. data, we find an inverse U-shape relation between VaR model errors and the horizon that impacts the optimal asset allocation of the representative agent.
