GARCH with additional variables

[hellen]

Hello guys,

I am a beginner in econometrics, trying to study the link between macroeconomic volatility and stock market volatility.
I estimated a GARCH to capture stock market vol and for macro vol I used first time st dev(12 M rolling window) and second time a GARCH( creating new variables with "make garch variance series" as I thought that the problem was that growth rates are not normally distributed).
I added them in "Variance regressors" window of my stock market vol model and here is where my problems begin: once I add all 3 of these macro variables, my variance coeff are either negative(GARCH/ARCH), bigger then 1(IGARCH) or statistically insignificant(EGARCH, GJR etc ).

Am I doing smth wrong? My series are stationary, seasonally ajusted etc so it shouldn't be a problem with my database.

thx!

[Marie90]

Hey Helen!

I am trying to regress macroeconomic volatility as measured by the volatility of real GDP on measures of institutional quality with yearly data of the perios 1996-2012. How exactly do you get values for the volatility variable using the "make garch variance series" option or the time standard deviation with rolling window?
Thank you very much!