Working Paper Series in Economics, Karlsruher Institut für Technologie (KIT) 52
[Statement of the problem] The present work considers the problem of investment portfolio risk estimation, including dynamic adjustment for each new transaction. Any Bank portfolio has a complex structure. It consists of stocks, bonds and a set of derivative securities. A portion of bonds and loans is riskless. For some of these assets, the methods offered cannot be applied without additional consideration of the term structure of interest rates and credit risks features. The risk estimation of this part of the portfolio containing some peculiar financial tools represents a separate issue, solving which exceeds the limits of the present research. We use as an estimation of a portfolio risk the amount of probable losses that can be sustained in case of a complete asset sale, related to current market value of these assets. The investment portfolio includes a number of shares, sale of which can significantly affect the market for a brief period of time, making the calculated estimation of risk insolvent. Thus it is necessary to estimate the quantity of shares that can be sold without having a material influence on the prices dynamics. Knowing this size, it is easy to calculate a time interval during which this portfolio can be sold. Definition of the stability of the concrete market is directly concerned with its specificity. This represents a separate practical problem, which is not considered in the submitted paper. Consequently, for a portfolio risk calculation, it is necessary to estimate dynamics of price behaviour for the time period during which controllable realization of portfolio assets is possible. Such an approach is described in many papers where estimation of risk is based on studying prices dynamics of stocks included in a portfolio (VaR - 'RiskMetrics', RiskManagement+). However, forecasting such processes represents a complicated problem itself. For example, on NASDAQ the prices of the most liquid stocks have large volatility. Deviation from average value of a stock price can run up to several percent even on ordinary days. To circumvent this problem, a new approach, which is not considered earlier, is offered in the given research.