Feb 13

Volatility forecast with GARCH(1,1) - 8 min tutorial

by David Harper, CFA, FRM, CIPM

Here is an example of using GARCH(1,1) model to forecast volatility forward. I focus on persistence (alpha + beta). The opposite of persistence is decay. Lower persistence implies greater decay: with lower persistence/more decay, the series will be pulled toward to the long run variance more rapidly. Put another way, if persistence is "unity" (alpha + beta = 1.0), then the forecast is a flat line as the forward estimate is simply today's variance. Here is the tutorial: