Forecasting Daily Volatility and Value at Risk with High Frequency Data

Subject Areas : مدیریت

Amir Mohammad Zadeh ¹ (Assistant Professor, Department of Management, Qazvin Branch, Islamic Azad University, Qazvin, Iran)
Sahar Masoud Zadegan ² (M.Sc. Student, Department of Management, Qazvin Branch, Islamic Azad University, Qazvin, Iran)

Received: 2017-02-11 Accepted : 2017-02-11 Published : 2017-01-20

Keywords: Efficiency, Forecasting, Volatility, Value at Risk, High Frequency Data,

Abstract :

One of the key aspects in the financial markets and its development is fluctuation. Fluctuation plays a key role in option pricing, portfolio management and the market sentiment. In general, financial institutions are faced with four various kinds of risk, which are credit risk, liquidity risk, operational risk, and market risk. The most appropriate method to measure the market risk is by using the VaR (value at risk). Value at Risk is statistical technique used to measure and quantify the level of financial risk within the investment portfolio over a specific time frame. It is always expressed by the monetary amount that is at risk as well as the probability of loss. This research is to predict the VaR for a one-day period in six different industries in which three companies are monitored in each industry. The time periods of the study are 30-minute intervals between 91/11/1 to 92/4/1,  in which the GARCH model is used for predicting the variance. The research then checks to see whether the data fits the normal or t-distributions models. Thus, six models are used for six different industries. All six chosen models are deemed proper to predict the coefficients, how fit the coefficients are, and Watson statistic camera. The estimation of the variance and the Var for all models is done at a %95 confidence interval. The research concludes that the companies involved in the basic metals group are more prone to risk and have higher VaR in comparison to other industries.