Hi Everyone,
I am trying to estimate rolling betas by using the past 24-60 months of data. I require at least 24 months of non-missing observations in the past 60 months. For example, if a stock has 24 observations, I want to estimate betas until that month. Or, if a stock has N observation where 24<=N<=60 with at least 24 non-missing data, I again want to estimate until that month. To do that, I run the following:
rolling _b, window(60) saving(betas, replace) reject(e(N) < 24): regress exc_ret mktrf smb hml mom
However, I can not get any result when there are less than 60 and more than 24 months of observations. Let's say a stock starts to report in month t and has non-missing values until t+23. I also want to get estimates for that month since there are at least 24 months. I guess "window(60)" option does not allow me to get results when there are less than 60 observations. As far as I understand, this code checks whether there are 60 observations and at least 24 of them are non-missing.
Is there a way to solve this problem? Thanks.
Ulas
Related Posts with Rolling Regression with 24-60 months of observations
Making new variable from first letter of string variableHi Statalist, Thanks for being a great forum which already has helped a lot! I'm new here, quite f…
Using stcrreg with the mi commandHi All, After not having much of an issue with my MI (multipley imputed) data and the stccreg comma…
How to arrange hospitalization data for sequence analysisDear Statalists, Just read a paper by Golay, P., et al., "Identifying patterns in psychiatric hospit…
add legend to one of the graphs in -graph combine-Dear all statalisters, I now face a problem in dealing with drawing a combined graph with adding a l…
Bayesian approach: what to do when it differs from the frequentist approach?Dear Forum Members, I'm somewhat puzzled by the results given from 2 down-to-earth models, one usin…
Subscribe to:
Post Comments (Atom)
0 Response to Rolling Regression with 24-60 months of observations
Post a Comment