Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: How to estimate standard error of panel variant coefficient


From   Tirthankar Chakravarty <tirthankar.lists@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: How to estimate standard error of panel variant coefficient
Date   Mon, 8 Oct 2012 10:04:05 +0530

I sent that out a little prematurely -- you need to add a -by()-
option since you want a parameter for each panel.

*-----------------------------------------------------------------
set matsize 9000
webuse nlswork, clear
xtset idcode year

xtreg ln_w  L.ln_w age ttl_exp L.ttl_exp c.ttl_exp#i.collgrad , fe
margins, expression(_b[ttl_exp]*collgrad) by(idcode)
*-----------------------------------------------------------------

T


On Mon, Oct 8, 2012 at 9:54 AM, Tirthankar Chakravarty
<tirthankar.lists@gmail.com> wrote:
> Use -margins, expression()- as in this simple example:
>
> *-----------------------------------------------------------------
> webuse nlswork, clear
> xtset idcode year
>
> generate age2 = age^2
> generate ttl_exp2 = ttl_exp^2
> generate tenure2 = tenure^2
> generate byte black = (race==2)
>
> xtreg ln_w  L.ln_w age ttl_exp L.ttl_exp c.ttl_exp#i.collgrad , fe
> margins, expression(_b[ttl_exp]*collgrad)  // expression of your choice
> *-----------------------------------------------------------------
>
> Note that "collgrad" is a panel invariant variable (Z_i) in this case.
>
> T
>
>
>
> On Mon, Oct 8, 2012 at 7:44 AM, Ram Acharya <acharya.ramc@gmail.com> wrote:
>> I am working in country and year panel data. The right-hand side
>> variables include lagged dependent and contemporaneous and lagged X
>> exogenous variable. I also have an interaction of contemporaneous X
>> and country component (Zi) which varies by country but not by year.
>> The model is somewhat like below
>>
>> Yit = a + lambda* Yi(t-1) + b1 Xit + b2 Xi(t-1) + b3 Xit* (Zi)
>>
>> My interest is in the long run coefficient (theta), which can be computed as
>>
>> Theta(i) = [b1 + b2 + b3*(Zi) ] / [1 – lambda]
>>
>> Note that theta varies by country. My question is: how to compute
>> standard error for Theta. If I did not have Zi, and had only two
>> variables, then I could use Delta method such as below to estimate
>> both coefficient and se.
>> nlcom (_b[Xit]+_b[L.xit])/(1-_b[L.Y])
>>
>> It appears to me that I cannot use the same method, as b3 varies by
>> country depending on the value of Zi. What would be the command in
>> Stata to estimate standard errors for Theta?
>>
>> Thanks in advance for the help.
>> Ram Acharya
>>
>> *
>> *   For searches and help try:
>> *   http://www.stata.com/help.cgi?search
>> *   http://www.stata.com/support/faqs/resources/statalist-faq/
>> *   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index