The previous lessons modelled the level of a series. This one models its volatility, which in finance is often the more valuable target. Asset returns are close to unpredictable in direction, but their size is highly predictable: calm days follow calm days and violent days follow violent days. That is volatility clustering, and GARCH is the model built to capture it. This lesson explains why returns need a conditional-variance model at all, builds ARCH and then GARCH from the stylized facts of returns, works a one-step volatility forecast and a volatility term structure by hand, and covers maximum-likelihood estimation and the asymmetric extensions. Volatility is the input to position sizing, value at risk, and option pricing, so this is among the most directly useful models in the course.
Table of Contents
Already have an account? Log in!