On Fri, Jul 10, 2009 at 10:13 AM, Davide
Cantoni<[email protected]> wrote:
>
> (3) -xtdata,fe- still does not work properly with the unbalanced
> dataset, and I have not found out why. Any suggestions are still
> appreciated.
>
With unbalanced data, when -xtdata- detrends the year variable you end
up with a series of unique values:
. xtdata company year invest mvalue kstock time, fe clear
. tab year
year | Freq. Percent Cum.
------------+-----------------------------------
1934.454 | 1 0.63 0.63
1934.487 | 1 0.63 1.26
1934.496 | 1 0.63 1.89
1934.775 | 1 0.63 2.52
1935.168 | 1 0.63 3.14
....
Here is one way around that (note - the random variable seed was set
as -set seed 12345-):
*** use -xtdata,fe to demean data
format year %9.0g
tsset company year
preserve
keep year
save foo,replace
restore
xtdata company year invest mvalue kstock time, fe clear
tab year
drop year
merge using foo
tsset year company
xtdata, fe clear
qui reg invest mval kst, r
est store xtdata
est table xtreg demeaned xtdata , ///
keep(mvalue kstock demean_mvalue demean_kstock) se
Which returns:
-----------------------------------------------------
Variable | xtreg demeaned xtdata
-------------+---------------------------------------
mvalue | .11547567 .1155367
| .01224137 .0182679
kstock | .39346621 .39187308
| .0513649 .050229
demean_mva~e | .11547567
| .01984418
demean_kst~k | .39346621
| .05353941
-----------------------------------------------------
legend: b/se
Scott
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
