Vine Copula Based Portfolio Level Conditional Risk Measure Forecasting

Accurate estimation of different risk measures for financial portfolios is of utmost importance equally for financial institutions as well as regulators, however, many existing models fail to incorporate any high dimensional dependence structures adequately. To overcome this problem and capture complex cross- dependence structures, we use the flexible class of vine copulas and introduce a conditional estimation approach focusing on a stress factor. Furthermore, we compute conditional portfolio level risk measure estimates by simulating portfolio level forecasts conditionally on a stress factor. We then introduce a quantile-based approach to observe the behavior of the risk measures given a particular state of the conditioning asset or assets. In particular, this can generate valuable insights in stress testing situations. In a case study on Spanish stocks, we show, using different stress factors, that the portfolio is quite robust against strong market downtrends in the American market. At the same time, we find no evidence of this behavior with respect to the European market. The novel algorithms presented are combined in the R package portvine, which is publically available on CRAN