Συλλογές
Τίτλος Backtesting value at risk models
Δημιουργός Λυκούδης, Παναγιώτης, Lykoudis, Panagiotis
Συντελεστής Οικονομικό Πανεπιστήμιο Αθηνών, Τμήμα Πληροφορικής
Τύπος Text
Φυσική περιγραφή 56σ.
Γλώσσα en
Αναγνωριστικό http://www.pyxida.aueb.gr/index.php?op=view_object&object_id=5269
Περίληψη This thesis investigates the concept of risk and how to quantify this notion with measures such as the Value-at-Risk (VaR). This method provides the concerned party with a figure that expresses the maximum expected loss on an investment for a given period and a given level of confidence. The calculation of VaR as before is done through a complex econometric approach, through out various models such as the conditional Autoregressive Heteroscedasticity model, known as Arch or the generalized autoregressive conditional heteroskedastic model known as Garch models. From this point and after we are going to refer to the above models only with their initials, presented next to each one of them in parentheses. In this process VaR or any other risk measure arises as a forecast of the period under the implementation process. In other words we will quantify the notion of risk using and comparing different methods as have been proposed by the literature through out the years. As dataset, we used 4 indices with the majority of them listed in the USA’s stock exchanges and the others listed to the Europe’s stock exchanges. More specifically, in our research we used time series returns by the indices: FTSE 100, CAC 40, Nikkei 224 and Nasdaq.
Λέξη κλειδί Arch
Garch
Time series
Stochastic volatility models
Value at risk
Διαθέσιμο από 2017-02-13 22:00:42
Ημερομηνία έκδοσης 11/30/2016
Ημερομηνία κατάθεσης 2017-02-13 22:00:42
Δικαιώματα χρήσης Free access
Άδεια χρήσης https://creativecommons.org/licenses/by/4.0/