What is the best way to model implied volatility surface in filtered historical simulation (other than keeping it constant)? Is it appropriate to apply GARCH models on IV surface key points? Any suggestions will be appreciated.
“Real creativity won’t make things more complex. Instead, it will simplify them.”