ESTIMATION OF FUTURE VOLATILITY
Abstract
Estimation of market risk due to price and interest rate movements is inevitable for financial institutions with huge trading portfolios. A basic indicator of the risk exposure could be derived from the standard deviation or volatility of the financial instruments. Accounting for the effect of diversification, other parameters should be determined for the risk analysis of portfolios. Correlations could describe the connection between pairs of basic financial instruments, so they could be used for the risk analysis of portfolios. We are interested in the risk level of assets in the future, therefore we need the future estimates of these risk parameters. These parameters could be assessed in different ways. First we can assume, that their future expected values could be determined from past data. However, there is another way for the volatility, in case of there is option trading for the specific instrument. In this case implied volatility could be calculated from the option price, which could be an estimate for future volatility. In case if both methods are available, we need some hints, which is better. On the other hand, the estimation of correlations involves filling in missing values from efficiency standpoints, when incomplete time series are available. We argue that data filling raises several problems, which depreciate the supposed benefits of this approach.