A Structure for General and Specific Market Risk

The paper presents a consistent approach to the modeling of general and specific market risk as defined in regulatory documents. It compares the statistically based beta-factor model with a class of benckmark models that use a broadly based index as major building block for modeling. The investigation of log-return of stock prices that are expressed in units of the market index reveals that these are likely to be Student t distributed. A corresponding discrete time benchmark model is then used to calculate Value-at-Risk for equity portfolios.