Regime Switching

[afterworkguinness]

Hello,
In regime switching, are we modeling 2 separate normal conditional distributions ? One for the low volatility period and one for the high volatility period ? If that is the case, how wold you calculate a VaR using 2 distributions ? Would you include both the low and the high volatility ? Or would you calculate VaR low volatility and VaR high volatility and aggregate them ?

Cheers

 

[ShaktiRathore]

Hi,
in regime switching there is a paradigm shift in the volatility that is low volatile period is entirely replaced by high volatile period. One of the possible solution is that weight high volatile periods highly and low volatile returns less to take into account the high risk involved in high volatility period after weighing the returns in accordance to volatile we get a single set of compiled returns and thus calculate var.
Other way possible is that calculate two different vars for each regime and aggregate them based on their volatility that is assign higher weight to high volatile period var and lower weight to low volatile var.

thanks

 

[afterworkguinness]

Very interesting, thanks.