Reza Taleblou; Mohammad Mahdi Davoudi
Abstract
In this paper, in order to calculate portfolio market risk of 10 selected industries indices in Tehran Stock Exchange, two models of Value Risk (VaR) and Expected shortfall (ES) have been used. Different models of multivariate GARCH and various Coppola models have been used in order to estimate the volatility ...
Read More
In this paper, in order to calculate portfolio market risk of 10 selected industries indices in Tehran Stock Exchange, two models of Value Risk (VaR) and Expected shortfall (ES) have been used. Different models of multivariate GARCH and various Coppola models have been used in order to estimate the volatility of the portfolio and nonlinear correlation of asset portfolio. Backtesting has been done by Kupiec, Christoffersen, Engle and Manganelli and McNeill and Ferry tests. Results show that the DCC-GARCH model by t-Student distribution compared to other competing models has the best results in estimating volatility of the asset portfolio. Also among all Copula models reviewed in this paper, t-student copula model has shown better results for estimating asset dependence. Finally, the results of backtesting of different models showed that both the DCC-GARCH model with t-Student distribution and DCC-GARCH-Copula with t-Student distribution have acceptable results in estimating VaR and ES. However, the Lopez and Blanco and Ihle tests showed that the DCC-GARCH model with t-Student distribution compared to the DCC-GARCH-Copula model with t-Student distribution gives a more accurate and efficient estimate of the VaR and ES of asset portfolios.
Reza Taleblou; mohammad mahdi davoudi
Abstract
In recent years, by using extreme value theory (EVT), researchers have estimated the market risk for rare events (crises) more accurately. This paper examines the different methods of measuring market risk at different levels of reliability. According to the assumptions of the EVT methods, measuring ...
Read More
In recent years, by using extreme value theory (EVT), researchers have estimated the market risk for rare events (crises) more accurately. This paper examines the different methods of measuring market risk at different levels of reliability. According to the assumptions of the EVT methods, measuring the effects of the financial crises on the value of assets requires a lot of time-series observations. Therefore, this paper has used four indices: total index, industry index, the first market index and the second market index of the Tehran stock exchange. The backtesting results showed that among the various methods, semi-parametric approach or the EVT approach in comparison with parametric (EWMA, MA, GARCH) and nonparametric approaches (Historical simulation) is more efficient and has a higher level of reliability. Also HS method shows acceptable results at high confidence level, while in calculating the value at risk in the 0.90 and 0.95 confidence levels, parametric methods (EWMA, MA, GARCH) provide more reliable results. Also the richness of the dynamics of GARCH and EWMA models are much more than the other models. In the next step by incorporating various models, the three models EWMA-EVT, GARCH-EVT and AWHS were made. Backtesting these three patterns showed that, AWHS and EWMA-EVT have provided the best results among various patterns, and have provided acceptable adequacy in estimating the value at risk at all levels of reliability. However; GARCH-EVT model shows acceptable results only in 0.999 reliability level.
Shiva Zamani; Majid Alifar
Volume 19, Issue 59 , July 2014, , Pages 183-210
Abstract
In this paper we formulate the volatility of the Metal Index, by an ARJI-GARCH model, with reference to the effect of the volatility of dollar exchange rate. We express the jump in the dollar exchange rate by an Autoregressive Conditional Jump Intensity (ARJI) model, and then use the output to model ...
Read More
In this paper we formulate the volatility of the Metal Index, by an ARJI-GARCH model, with reference to the effect of the volatility of dollar exchange rate. We express the jump in the dollar exchange rate by an Autoregressive Conditional Jump Intensity (ARJI) model, and then use the output to model the volatility of Tehran Base Metal Index by GARCH. To test the model, we calculate VaR using the ARJI-GARCH volatility, and the GARCH volatilities without considering the dollar exchange rate volatility or the jump in it. We calculate the referred VaR also by an exponentially time weighted historical simulation method. We test the accuracy and preciseness of the resulted VaRs by the associated statistical tests, and conclude that the ARJI-GARCH model is well suited for forecasting volatility in this context
Zahra Nasrollahi; Mina Shahviri,
Volume 15, Issue 44 , October 2010, , Pages 199-230
Abstract
Management of Foreign exchange reserves is important for every country. This matter is also of particular interest for Iran as an Oil exporting developing country. This paper designs an optimal portfolio for that part of foreign exchange incomes which is used for investment. Using the data on foreign ...
Read More
Management of Foreign exchange reserves is important for every country. This matter is also of particular interest for Iran as an Oil exporting developing country. This paper designs an optimal portfolio for that part of foreign exchange incomes which is used for investment. Using the data on foreign exchange daily returns, for the period 2000-2008, and applying univariate and multivariate Garch models, we estimate a model which maximizes expected returns subject to a Value-at-Risk constraint. The results are examined using Backtesting, and then the most acceptable model is suggested. The results that the multivariate GARCH model is the most efficient method for selecting the foreign exchange optimal portfolio in Iran.
