A GARCH Model Method for Assessing Value at Risk (VAR)

Abstract

Financial markets have historically demonstrated their unpredictability and, when combined with poor risk management decisions and procedures, have led to various disastrous outcomes. To mitigate these issues, various tools are employed to manage risk exposure. One example of this tool is Value at Risk (VaR), which provides a statistical assessment of potential losses in an asset or portfolio over a specified period and at a defined confidence level. Despite its widespread use, VaR also has limitations due to the clustering effect of market volatility. To address this, the Generalized Autoregressive Conditional Heteroskedasticity (GARCH) model is necessary for providing more accurate data. We use the GARCH (1:1) model to estimate the Value-at-Risk (VaR) of two stocks, Nvidia and Netflix, which are selected from different industries to capture the differences in risk exposure across various sectors. The data were gathered from Yahoo Finance over two years. We find that Nvidia exhibits higher volatility compared to Netflix, which is reflected in a higher Value at Risk output, as expected. Our study highlights the significance of the Generalized Autoregressive Conditional Heteroskedasticity (GARCH) model in enhancing risk management strategies. It emphasizes its crucial role in accurately evaluating and predicting financial risk across diverse market conditions.