The accuracy of estimates of Value-at-Risk (VaR) and Conditional Value-at-Risk (CVaR) hinge on whether the assumed distributional form of returns is correctly specified. Such a correct specification, however, is a major issue, as there is no consensus on what distributional form provides the best fit with the empirical data. Indeed, many researchers argue that the best fit is obtained when mixtures of distributions are used. In this research, we study whether and how the distributional properties of stock market returns differ for bull and bear markets, and how these differences impact on the accuracy of the estimates of the tail-risk measures. We show that bull and bear markets have different distributional forms for developed and emerging stock exchanges. These differences have important implications for the accuracy of the methods used to estimate VaR and CVaR.
|Number of pages||45|
|Publication status||Published - Mar 2018|