عنوان مقاله [English]
Risk forecasting has an important role in making the right decisions of managers and financial activists to invest in companies and institutions for future periods. On the other hand, wrong decisions of managers can have undesirable consequences for their organizations. Therefore, the most important issues for investors is risk forecasting in future periods. In order to achieve their predetermined strategies and strategies, economic firms will have to undertake various activities. Investment can be considered as one of the key pillars of these activities, which involves accepting risk. Because of the importance of the issue, the concept of risk management has been developed to protect capital against harmful risk-taking effects, which is not about risk aversion but rather to turning threats into opportunities. In other words, risk management refers to a process that identifies its types in intimidating conditions to deal with the first risk, and then, in the next step, optimally controls the risk. One of the known tools for calculating risk is the Conditional Value at Risk )CvaR(. Therefore, in this paper, it has been tried to first use the Generalized Auto Regressive Conditional Heteroscedasticity methods to estimate the Conditional Value at Risk and then after determining the benchmark model Forecasting daily(one-step-ahead), weekly(five-step-ahead) and by using the classic method and the Holt Winters exponential smoothing with two and three parameters methods. The data used daily logarithmic returns of the automobile industry index from April 2010 to September 2016. Five back testing consist of unconditional Coverage test, Conditional Coverage test, Joint test, Lopez loss function test and Blanco and Ihle loss function test have been used to assess the estimating and forecasting of Conditional Value at Risk . Regarding the performance of the model of the Holt Winters exponential smoothing multiplicative method with Three parameters at 95 Percent and 99 Percent confidence levels, this model is known as a superior method in daily, weekly forecasting of Conditional Value at Risk.