Evaluation of multivariate GARCH models in estimating the Values at Risk (VaR) of currency, stock and gold markets

Subject Areas : Journal of Investment Knowledge

abdollah rajabi khanghah ¹ , Hashem Nikoomaram ^{2 *} , Mehdi Taghavi ³ , Mirfeiz Fallah Shams ⁴

1 - عضوهیات علمی دانشگاه آزاد اسلامی واحد اسلامشهر
2 - استاد و عضو هیات علمی دانشگاه آزاد اسلامی،واحد علوم و تحقیقات تهران، گروه مدیریت مالی،تهران،ایران
3 - Professor and Faculty Member of Islamic Azad University, Tehran University of Science and Research, Department of Economics and Management, Tehran, Iran
4 - Associate Professor and Faculty Member of Islamic Azad University, Central Tehran Branch, Department of Business Administration, Tehran, Iran

Received: 2018-11-05 Accepted : 2018-12-17 Published : 2020-08-22

Keywords: Multivariate GARCH model, Vector GARCH model (VEC-GARCH), Value at Risk (VaR),

Abstract :

The development of financial markets requires the introduction of new models, forecasting and risk management. One of the indicators that are considered in risk management and measurement is value at risk index. In this research, the multivariate GARCH model has been evaluated to predict the value of exposed portfolio risk, including currency, stocks and gold, and the combined returns of gold price data, total stock exchange index and exchange rates from 2009 to 2016 were used. The results of VECH, BEKK, DCC, and VECH diagonal models have shown that the volatility of these variables in the estimation period is effective and this confirms the hypothesis of market independence in Iran, in order to evaluate the performance of these models in predicting VAR One-day prediction of conditional variance covariance matrix of models was used. The results of the post-test of the models using the coup and kristoferson tests showed the performance of all four models was appropriate and the comparison of the mean loss function of Lopez showed that the VECH model had better performance than other models. Despite the good performance of the VECH model, however, the estimation of this model is very time-consuming. due to the large number of parameters that are estimated in the estimation of VECH and BEKK models that lead to a reduction in the degree of freedom and, as a result, a reduction in the validity of the model, the use of these two models for portfolios with more than three assets is not recommended.

References:

• ابونوری اسمعیل، عبداللهی محمدرضا، مدلسازی نوسانات بخش های بازار سهام ایران با استفاده از مدل گارچ چند متغیره، مجله علمی - پژوهشی تحقیقات مالی دانشگاه تهران، ش 1، دوره 14، تابستان 1391، ص 1-16.

• ابونوری اسمعیل، عبداللهی محمدرضا، حمزه مصطفی، ارزیابی پویایی های رابطه بین نرخ ارز و شاخص سهام ایران با استفاده از مدل گارچ دومتغیره، پژوهشنامه بازرگانی، موسسه مطالعات و پژوهش های بازرگانی، شماره 65، زمستان 1391 ص 65-86.


• دهقانی منوچهر، خیابانی ناصر (1393) اثر سرریز تلاطم در بازارهای نفت، طلا و ارزش دلار آمریکا در مقابل یورو ، فصلنامه پژوهش‌ها و سیاست‌های اقتصادی، ش ۷۲ ص۱۲۷-۱۵۴


• کشاورزحداد، غلامرضا. اسمعیل‏زاده، موسی. مدلسازی سری زمانی برای پیش­بینی تلاطم در بازدهی سهام شرکت سیمان تهران. دانشگاه تهران، مجله­ی تحقیقات اقتصادی، در شرف انتشار.


• Bollerslev, T. (1986). Generalized Autoregressive Conditional Heteroskedasticity. Journal of Econometrics, Vol.31, No.3, pp. 307-327.


• Bollerslev, Tim, Engle, Robert F. & Wooldridge, Jeffrey M (1988).A Capital Asset Pricing Model with TimeVarying The Journal of Political Economy, Vol. 96, No. 1, pp. 116-131.


• Christoffersen, P. F., 1998, "Evaluating Interval Forecasts", International Economic Review, Vol. 39, No. 4, PP. 841-862.


• Engle, Robert F. (1982). Autoregressive Conditional Heteroscedasticity with Estimation of the Variance of United Kingdom Inflation. Econometrica, Vol. 50, No. 4, PP. 987-1007.


• Engle, Robert F. & Kroner, K.F. (1995). Multivariate simultaneous generalized ARCH. Econometric Theory, Vol. 11, pp.122–150.


• Engle Robert F.(2001).GARCH 101: The use of ARCH/GARCh models in applied econometrics. Journal of economic perspectives, Vol. 15, No. 4, pp. 157-168.


• Engle Rober F.(2002). Dynamic Conditional Correlation –A Simple Class of Multivariate GARCH Models. Forthcoming Journal of Business and Economic Statistics, Vol. 20, pp. 339-350.


• He Kaijian, Lai Keung and Xiang Guocheng(2012) Portfolio Value at Risk Estimate for Crude Oil Markets:A Multivariate Wavelet Denoising Approach, Energies 5, 1018-1043.


• Kupiec, P., "Techniques for Verifying the Accuracy of Risk Measurement Mode ls", Journal of Derivatives, November 1995, PP. 73-84


• Lee, MingChih, Chiou, JerShiou & Lin, ChoMin )2006(. A study of value at risk on portfolio in stock return using DCC Multivariate GARCH. Applied Financial Economics Letters, vol. 23, pp.183-188.


• Lee, Sang Jin & Binh, Ki Beom (2008). Model Selection for Estimating Portfolio VaR in Korean Stock Market. AsiaPacific Journal of Financial Studies, vol.37, No. 5, pp. 877-913.


• Lindgren Charlie, Kayal Elias(2014) Multivariate approaches for Value-at-Risk and Expected Shortfall on electricity forwards


• Lopez, Jose A. (1999). Methods for Evaluating Value-at-Risk estimates, Federal Reserve Bank of San Francisco. Economic Review, Vol. 2, PP. 3-17.


• Ruiz E, Nogales F, Santos A (2009) Comparing univariate and multivariate models to forecast portfolio value-at-risk, Statistics and Econometrics Series, 09-72