Εκτίμηση κινδύνου σε χρηματιστηριακούς δείκτες ευρωπαϊκών χρηματιστηρίων με τη χρήση αυτοπαλίδρομων υπό συνθήκη ετεροσκεδαστικότητας υποδειγμάτων
Measuring value at risk using generalized autoregressive conditional heteroscedastic models for european indices
Οικονόμου, Νικόλαος Δ.
The objective of this thesis is to present the concept of Risk Measurement as it is achieved by using the Value-at-Risk (VaR) method. According to this technique, the user can estimate the risk by a single number which represents the worst expected loss of an asset for a given horizon at a fixed confidence level. The estimate of VaR is obtained by using a sophisticated econometric approach based on Time Series Analysis, which combines and matches AutoRegressive Integrated Moving Average (ARIMA) models based on Box-Jenkins methodology with Generalized AutoRegressive Conditional Heteroscedastic (GARCH) models. The application of these models will provide VaR as one-step-ahead forecast of the process. Thus, an efficient econometric approach based on ARIMA-GARCH models is applied to log returns of major European stock market indices with the aim to estimate volatility, as it is captured on them. Finally, the estimates of VaR will be measured for indices in total and Backtesting technique for verifying the accuracy of VaR models will be performed.