[Shafaq Malik]

Hi,

I'm applying asreg command for Fama Mcbeth regressions
I want to add industry dummy as i.ind but it is giving me error that "factor-variable and time-series operators not allowed"
Can you please help out on this

Regards,

Shafaq

[Attaullah Shah]

You are right, version 3.1 of asreg does not support factor variables in the varlist. As a workaround, you can do the following

tab ind, gen(inddum)
asreg y x1 x2 x3 inddum*, fmb

The

tab ind, gen(inddum)

will create dummy variables for the ind categories. This approach is not memory-efficient as it creates unnecessary variables, however, it is a workaround as mentioned above. Later on, you can drop the dummies by

drop inddum*

Be careful while using the wildcard * as any variable with the prefix of inddum will be dropped.

[Shafaq Malik]

Dear Attaullah,
Thank you so much for helpful suggestion.

Regards,

Shafaq

[Attaullah Shah]

asreg version 3.2 is now available for download from SSC (thanks to Kit Baum). This version fixes a minor bug in one of the Mata functions. Mata function ASREG4s0f1() would throw an error if option fit was used on a dataset that did not have enough observations for estimating a regression. To update,

ssc install asreg, replace

[Austin Moss]

Recently used asreg to find rolling window market betas. Extremely useful program that runs quickly and smoothly. As a bonus, Dr. Shah is very responsive to questions about the code, too!

[Vince Williams]

Dear Professor Shah, I was wondering how I can export the standard errors and the degrees of freedom from the first stage Fama-Macbeth regression using asreg. You posted the code
asreg invest mvalue, fmb save(eq1) first but how can we get these additional statistics.

[Attaullah Shah]

Vince Williams I have shown one method of doing so in this blog post.

[Jasmin Hennig]

Dear Professor Shah,

I am using your asreg command combined with the Fama/MacBeth two-step procedure but it seems like the command does not work correctly although I don't get an error message.
Can you maybe help detect my error? Note that I use Fama/MacBeth for a different setting as originally used by the authers, that is I use it to analyze determinants of efficiency.

asreg EFFICIENCY L1_IO L1_MV L1_LIQU, fmb first

where EFFICIENCY is the dependent variable (efficiency measure per firm and month, estimated before) and the independent variables are estimarors for institutional ownership(IO), size(MV), and liquidity(LIQU), each lagged by one period (note that no factor variables are used, I generated the lagged variables beforehand).

Problem 1: This regression only displays missing values (.) in the first regressions table but delivers results in the second step (Note that I deleted some rows in between the table to better fit it in here).



If I adjust the regression to using contemporaneous independent variables, the first regressions table shows results, but then another problem arises:

asreg EFFICIENCY IO MV LIQU, fmb first

Problem 2: This regression displays values in the first regressions table, however, not for all time periods (161 time periods, 131 lines) and shows missing rows in between (eg between 6 and 7).

Also here I deleted rows in between so it's easier to understand the regression output.




Any hint/comment is highly appreciated!


Many thanks and kind regards,

Jasmin

[Attaullah Shah]

Please report an example dataset using dataex(from SSC). Also, please quote the asreg version

which asreg

Stata version and operating system (Windows, macOS, etc)

[Attaullah Shah]

Jasmin Hennig You can also read my blog post on how to estimate the first stage regression of Fama and MacBeth (1973) procedure. The blog entry is given at https://fintechprofessor.com/2018/12...dure-in-stata/
The blog post shows that the first stage regression can be obtained from the asreg's fmb option or using the cross-sectional regression of asreg. In the latter case, you can also get other regression components. I think this will give you a better idea of where your error is coming from.

[Jasmin Hennig]

Dear Professor Shah,

thanks for your quick reply.
Can you tell me how to display the first regressions table after running the following command (as speccified in the mentioned blog post)?
bys year: asreg invest mvalue kstock, se bys year: keep if n == _N Kind regards, Jasmin

[Attaullah Shah]

webuse grunfeld, clear
bys year: asreg invest mvalue kstock, se
bys year: keep if _n == _N
list _*

 +------------------------------------------------------------------------------------------------------+
     | _Nobs         _R2      _adjR2   _b_mvalue    _b_kstock      _b_cons   _se_mv~e   _se_ks~k   _se_cons |
     |------------------------------------------------------------------------------------------------------|
  1. |    10   .86526202   .82676546   .10249786   -.00199479    .35603339   .0157931   .2148591   23.82794 |
  2. |    10   .69639369   .60964903   .08370736   -.05364126    15.218946   .0211982   .4125528   49.72796 |
  3. |    10    .6637627   .56769491    .0765138    .21772236   -3.3864706   .0218952   .4745161   62.14382 |
  4. |    10   .70557727   .62145649   .06801777    .26911462   -17.581903   .0220019   .2076121   33.62243 |
  5. |    10   .82660153   .77705911   .06552194    .19866456   -21.154227   .0131751   .1563955   29.10151 |
     |------------------------------------------------------------------------------------------------------|
  6. |    10   .83925512   .79332801     .095399    .20229056   -27.047068   .0171077   .2206074   42.49812 |
  7. |    10   .85621485   .81513338   .11476375    .17746501   -16.519486   .0197202   .2338307   47.43406 |
  8. |    10   .85730699   .81653756   .14282513    .07102405   -17.618283   .0246973   .1966943   43.85369 |
  9. |    10   .84206394   .79693935   .11860951    .10541193   -22.763795   .0207092   .1887016    46.8604 |
 10. |    10   .87551498   .83994783   .11816422    .07220719   -15.828145   .0169881   .1537212   41.84578 |
     |------------------------------------------------------------------------------------------------------|
 11. |    10   .90679731   .88016797    .1084709    .05022083   -10.519677   .0133214   .1254533   35.10524 |
 12. |    10   .89475165    .8646807   .13794817    .00541339   -5.9906571    .018637   .1600683   45.73243 |
 13. |    10   .89123943   .86016498   .16392696   -.00370721   -3.7324894   .0280743   .1285463   37.80575 |
 14. |    10    .7888235   .72848735    .1786673   -.04255555    8.5388099   .0463983   .1661775   52.39133 |
 15. |    10   .86325678   .82418728   .16159617   -.03696511    5.1782863   .0346516   .1268614   41.07802 |
     |------------------------------------------------------------------------------------------------------|
 16. |    10   .85771384   .81706065   .17621675   -.02209565    -12.17468   .0393216   .1361792    46.6222 |
 17. |    10   .87377295   .83770808   .18314051   -.11205694    26.138157   .0358898   .1486738   53.00348 |
 18. |    10   .84612242   .80215739   .19892081   -.06749499    7.2928402    .052286   .1906835   67.84544 |
 19. |    10   .88926056   .85762072   .18267385    .09875335   -50.152546    .058579   .2164437   77.91569 |
 20. |    10   .89845005   .86943578   .13451162    .33137459   -133.39308   .0704524   .1932826   76.18067 |
     +------------------------------------------------------------------------------------------------------+

* Too see in the data
browse _*

[GEORGE PEPPAS]

Dear Professor Shah,

How do I keep the t-test value for each coeficient for regression?

Thanks in advance

[Attaullah Shah]

Here is one way one to find t statistics using asreg http://fintechprofessor.com/2018/03/...tics-se-stata/

[Kate Lussy]

Hello Mr. Shah,

I was wondering how it is possible to calculate adjusted R^2 through asreg? I only receive the average R^2 as output. Thank you kindly.