Storing the p-values and t-stat from Paul Geertsema's regression code

[Tuyen Dinh]

Hi guys! I need to conduct rolling regressions to estimate the betas of Fama-French's 3 factors. I am using daily data so this is going to take ages! I bumped into Paul Geertsema's paper where he advocated an alternative method that was '367' times faster than Stata's regress command. I tested his code on a random sample of stocks in excel and it gets a pass from me!

His .do file and .ado file can be accessed from here: http://poseidon01.ssrn.com/delivery....106090&EXT=pdf

The problem that I have is there is no option to store the p-value and t-stat after each regression. How would I do this?

[daniel klein]

From a quick glance the mentioned program neither computes a t-statistic nor a p-value. If want these you will have to implement it.

In case you are not familiar with the formulas, see Buis (2007).

Best
Daniel

Buis, M. L. 2007. Stata tip 53: Where did my p-values go? The Stata Journal, 7(4): pp. 584-586.
http://www.stata-journal.com/sjpdf.h...iclenum=st0137

[Tuyen Dinh]

From a quick glance the mentioned program neither computes a t-statistic nor a p-value. If want these you will have to implement it.

In case you are not familiar with the formulas, see Buis (2007).

Best
Daniel

Buis, M. L. 2007. Stata tip 53: Where did my p-values go? [i]The Stata Journal[i], 7(4): pp. 584-586.
http://www.stata-journal.com/sjpdf.h...iclenum=st0137

Thanks for your prompt reply Daniel. I am not familiar with MATA. Will I need to learn MATA to implement this? Or can the change be done in his .do file?

[daniel klein]

As far as I have seen, fastreg will give you all components needed (coefficnets, covariance matrix and degrees of freedom). Aside form the DF which seem to be returned in e(dof) instead of e(r_df), you can merely copy the couple of lines of code in Maarten's article.Please do read this, as I do not think you have done so in the four minutes you needed to reply.

If you need more/other assistance, please be more precise in terms of what exactly you mean by

store the p-value and t-stat after each regression

and give code examples.

Best
Daniel

[Luce Mock]

Hello,
I have a similar problem as Tuyen. I have a panel data set with daily stock returns across a bunch of firms. I would like to estimate daily alphas (constant) and betas using the Fama French 3-Factor model with a rolling window of 250 days. As the rolling command in STATA takes ages, I also use the -fastreg- command from Geertsema (see https://papers.ssrn.com/sol3/papers....act_id=2423171).
My problem is similar to Tuyen: I do not manage to adjust the code in that it stores standard errors or t-stats after each regression. I had a look at Buis (2007). However, it still doesn't work for me.
Can anyone help me how to do that in STATA? The code that I use is as follows:

**rolling window betas**
gen mktrf_beta = .
gen smb_beta = .
gen hml_beta = .
gen _cons_beta = .
gen mktrf_beta_f = .
gen smb_beta_f = .
gen hml_beta_f = .
gen _cons_beta_f = .

local maxobs = 400000
local window = 250

timer clear
timer on 1
forvalues k = `window'/`maxobs' {
local first = `k'-`window'+1
local last = `k'
if permno[`last'] == permno[`first'] {
* if window covers same permno, then do estimation
qui fastreg eret mktrf smb hml in `first'/`last'
* save coefficients
foreach x in mktrf smb hml _cons {
qui replace `x'_beta_f = _be[`v'] in `last'
}
}
}
timer off 1
timer list 1

Thank you so much for your help.
Best,
Luce