Mastering R for Quantitative Finance

Volatility modeling

It is a well-known and commonly accepted stylized fact in empirical finance that the volatility of financial time series varies over time. However, the non-observable nature of volatility makes the measurement and forecasting a challenging exercise. Usually, varying volatility models are motivated by three empirical observations:

• Volatility clustering: This refers to the empirical observation that calm periods are usually followed by calm periods while turbulent periods by turbulent periods in the financial markets.
• Non-normality of asset returns: Empirical analysis has shown that asset returns tend to have fat tails relative to the normal distribution.
• Leverage effect: This leads to an observation that volatility tends to react ...