xtivreg test for endog, instrument relevance and instrument exog

[Krissy Philips]

Hi, I want to test for endogeneuity of the variable religioisitymean where I have i.sin##c.religiostymean in my equation. What are the commands for testing for endogentuity of this variable, the instrument relevance test and instrument exog test in stata after running xtivreg?

Estimation equation is: xtivreg netreturn sin (religiositymean sinreligiositymean= statereg sinstatereg) beta lmarketcap lpb bev lgdp spread inflationrate open law year1 year2 year3 year4 year5 year6 year7 year8 year9 year10 i.country_c, re first vce(cluster company)

when doing estat endogenous
I get the error: subcommand estat endogenous is unrecognized

Thanks

[Phil Bromiley]

You will increase your chances of a useful reply if you provide Stata code (using code delimiters), Stata output, and sample data (using dateex). Please see the FAQ on asking questions.
xtivreg2 has some of these tests built in. Otherwise, look at various estat options.

[Krissy Philips]

You will increase your chances of a useful reply if you provide Stata code (using code delimiters), Stata output, and sample data (using dateex). Please see the FAQ on asking questions.
xtivreg2 has some of these tests built in. Otherwise, look at various estat options.

I did include the stata code in the question? Its the estimation equation? Sorry here is the output^^

[Eric de Souza]

Jeff Wooldridge has answered your question here:
http://www.statalist.org/forums/foru...est-panel-data
You then replied to him. If you maintain two threads on the same issue, this will create confusion.

[Krissy Philips]

Jeff Wooldridge has answered your question here:
http://www.statalist.org/forums/foru...est-panel-data
You then replied to him. If you maintain two threads on the same issue, this will create confusion.

Hi Eric,

The part about testing for instrument relevance and exogenuity hasn't been answered however.

I am running:

xtivreg netreturn sin (religiositymean sinreligiositymean= statefav sinstatefav) beta lmarketcap lpb bev lgdp spread inflationrate open law year1 year2 year3 year4 year5 year6 year7 year8 year9 year10 country1 country2 country3 country4 country5 country6 country7 country8 , re vce(cluster company)

And was told I can do xtoverid to test for the instrument relevance/exogenuity but I'm getting "o. operator not allowed
r(101);"

I haven't got any factor variables in the equation so wasn't sure why this is happening - could you assist?

[Eric de Souza]

When I run (albeit a nonsense regression):

webuse grunfeld
xtivreg invest (mvalue = kstock), re vce (cluster company)
xtoverid
xtoverid, cluster(company)

I get no error message
Try reducing the number of explanatory variables to see if that helps to get rid of the error

[Krissy Philips]

When I run (albeit a nonsense regression):

webuse grunfeld
xtivreg invest (mvalue = kstock), re vce (cluster company)
xtoverid
xtoverid, cluster(company)

I get no error message
Try reducing the number of explanatory variables to see if that helps to get rid of the error

Would that not be testing the instrumental estimation for a different equation to the one used in my analysis then? I am now getting another error message when running the same equation...:
Error - must have ivreg2/ivreg29/ivreg28 version 2.1.15 or greater installed
r(601);

I'm quite confused to what is that cause of this and what's going on. Your help would be much appreciated.

[Eric de Souza]

To get that error message you must have been running -xtivreg2- and not -xtivreg-
Could you please reproduce the commands you ran with all the variables you used as well as the error messages you got, each command and each error message on a new line.

[Krissy Philips]

Thanks Eric, please have a look at these attached, I am only running xtivreg??

[Eric de Souza]

Impossible to read.

[Krissy Philips]

Impossible to read.

Ive made the images bigger so hopefully this will be clearer. They are the succession of commands which I tried.

thanks

[Eric de Souza]

To begin with install ranktest: ssc install ranktest. xtivreg requires ranktest. Then we'll see.
Also, instead of putting your image in a Word file, save your images in png format, if you can

[Krissy Philips]

its still coming up as this Eric, will it help if I send you the data file? I've googled this problem but to no avail.

[Eric de Souza]

This time the image you posted is not visible

[Krissy Philips]

I'm not sure why not sorry, I can see it and it's a PNG file. But I basically removed the 1 out of the 8 country dummies I have put in, which was getting omitted due to collinearity. However now the constant term has been omitted?Therefore I get the error: variable _cons not found
r(111);? I think if I can prevent this from happening it should solve the problem. Why is this happening? Hopefully you can see this:

xtivreg netreturn sin (religiositymean sinreligiositymean= statefav sinstatefav) beta lmark
> etcap lpb bev lgdp spread inflationrate open law year1 year2 year3 year4 year5 year6 year7 y
> ear8 year9 year10 country1 country2 country3 country4 country5 country6 country7, re vce(clu
> ster company)

G2SLS random-effects IV regression Number of obs = 1,310,589
Group variable: company Number of groups = 11,217

R-sq: Obs per group:
within = 0.0053 min = 3
between = 0.0016 avg = 116.8
overall = 0.0039 max = 120


Wald chi2(28) = 21205.21
corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000

(Std. Err. adjusted for 11,245 clusters in company)

| Robust
netreturn | Coef. Std. Err. z P>|z| [95% Conf. Interval]

religiositymean | 8.593998 4.578544 1.88 0.061 -.3797837 17.56778
sinreligiositymean 2.629845 1.24173 2.12 0.034 .1960985 5.063591
sin | -5.628733 2.384256 -2.36 0.018 -10.30179 -.955678
beta | .4106493 1.57818 0.26 0.795 -2.682527 3.503826
lmarketcap | .0068872 .0393319 0.18 0.861 -.0702018 .0839763
lpb | 1.291663 .0651996 19.81 0.000 1.163875 1.419452
bev | -.6960829 .4117346 -1.69 0.091 -1.503068 .110902
lgdp | -1.158466 .19548 -5.93 0.000 -1.5416 -.7753326
spread | .5631453 .1378397 4.09 0.000 .2929845 .8333061
inflationrate | -.1676651 .0395968 -4.23 0.000 -.2452734 -.0900568
open | -.0139416 .002138 -6.52 0.000 -.018132 -.0097513
law | 1.566445 9.68651 0.16 0.872 -17.41877 20.55166
year1 | -3.307143 .2228099 -14.84 0.000 -3.743842 -2.870443
year2 | -2.67502 .3942239 -6.79 0.000 -3.447685 -1.902355
year3 | -.848703 .2192208 -3.87 0.000 -1.278368 -.4190382
year4 | -9.003221 .2226633 -40.43 0.000 -9.439633 -8.566809
year5 | -.934297 .181023 -5.16 0.000 -1.289096 -.5794984
year6 | -1.276434 .3654093 -3.49 0.000 -1.992623 -.5602448
year7 | -3.53848 .2776365 -12.75 0.000 -4.082638 -2.994323
year8 | -2.426716 .315858 -7.68 0.000 -3.045786 -1.807645
year9 | .5427375 .3208605 1.69 0.091 -.0861375 1.171612
year10 | -1.175674 .2956233 -3.98 0.000 -1.755085 -.5962633
country1 | .1485511 2.03223 0.07 0.942 -3.834547 4.131649
country2 | -1.130993 2.980294 -0.38 0.704 -6.972262 4.710275
country3 | 4.871609 1.030264 4.73 0.000 2.852328 6.89089
country4 | -3.829572 7.106923 -0.54 0.590 -17.75889 10.09974
country5 | 3.886175 2.093766 1.86 0.063 -.217531 7.989881
country6 | -1.465936 .6749977 -2.17 0.030 -2.788907 -.1429644
country7 | -11.46393 7.463646 -1.54 0.125 -26.0924 3.164549
_cons | 0 (omitted)

sigma_u | 5.3873446
sigma_e | 46.549328
rho | .01321734 (fraction of variance due to u_i)

Instrumented: religiositymean sinreligiositymean
Instruments: sin beta lmarketcap lpb bev lgdp spread inflationrate open law year1
year2 year3 year4 year5 year6 year7 year8 year9 year10 country1
country2 country3 country4 country5 country6 country7 statefav
sinstatefav


. xtoverid, cluster(company)
variable _cons not found
r(111);

. xtoverid
variable _cons not found
r(111);