I'm currently doing a research for my dissertation about R&D expenditures and stock return volatility (both firm specific volatility and systematic volatility). So, volatility is the dependent variable in my study.
Most empirical research I’ve read use the annualized standard deviation of weekly returns as a measure of total volatility and annualized standard deviation of weekly errors from CAPM model to measure firm-specific volatility.
I'm planning on using GARCH (1,1) model to measure volatility since this was suggested by my professor. I’m planning on using weekly stock returns. My question is:
- Do I have to use the conditional variance as dependent variable or the long-run volatility as the dependent variable? Also, do I need to annualize it?
- Concerning the firm-specific risk. How can I incorporate GARCH (1,1) if I would use the CAPM model?
you can use contract the conditional variance by using GARCH model and then use it as dependent variable in separate model, or you can use the other variables as independent variables while you are estimating the conditional variance. Trey to use monthly data at most if that possible since the yearly volatility does not make sense as the yearly conditional variance of any stock returns dose not reflect any fluctuation in stock returns.
Regarding to the second question you have to run weekly CAPM model first and then construct the error term after that use them as returns in GARCH model.
Try and use FIGARCH fractionally integrated generalized auto regressive conditional hetroskedasticity on eviews 12.
Hi dr. Konstantin,
You mentioned that I need to annualize the conditional volatility.
I ran in eViews the daily stock returns of a firm and calculated the volatility using GARCH (1,1) (as you can see in the attached picture). So, if I understand correctly I need to annualize the GARCH(-1) 0.738970 by multiplying it with sqrt(252)?
My ultimate research equation looks as follow:
Volatility (annualized GARCH 1,1)= alpha0+beta1*R&D_intensity+beta2*Firm_Size+beta3*Leverage
The volatility model is the appropriate model for this type of study. go ahead excellent.
- Badges
- Science topic
- Similar topics
- Methodology
- Research Methodology
- Research Papers