Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


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

Re: st: Confusion about collinearity


From   Austin Nichols <austinnichols@gmail.com>
To   "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Subject   Re: st: Confusion about collinearity
Date   Mon, 2 Dec 2013 05:11:44 -0500

Kevin Yarbrough <Kevin.Yarbrough@usma.edu> :

No doubt there are other coding errors, and you seem to want an eitc
variable in all years, coded 0 or 1 depending on whether a state has
an EITC that year. So your variable should be eitc, not eitc00, and
you need long data.

But there is a substantive problem as well: you are regressing log
wage income on a dummy for EITC, but states have different EITC values
and larger EITC benefits tend to draw more people into the labor
market, i.e. you are observing different select groups of labor market
participants in different years.  You could include people who have
zero wage income using a -glm- with log link instead; see

http://blog.stata.com/2011/08/22/use-poisson-rather-than-regress-tell-a-friend/
--referring to
http://www.stata.com/meeting/boston10/boston10_nichols.pdf

but you probably should not model only conditional mean income, as a
regression does, since the distribution of wage income will change,
not just shift upward.

Note that increased labor force participation due to the credit will
tend to depress pretax wages of otherwise comparable workers but
increase post-tax wage income of recipients.  See also:

http://ideas.repec.org/a/aea/aecrev/v95y2005i2p88-93.html
http://ideas.repec.org/a/aea/aejpol/v2y2010i1p177-208.html


On Mon, Dec 2, 2013 at 1:33 AM, Yarbrough, Kevin T CADET MIL USA USMA
<Kevin.Yarbrough@usma.edu> wrote:
> See comments below:
>
>>I followed that up with this code to replace the missing values:
>>
>>replace eitc00=0 if year==2000 & eitc00==.
>
>>If I am reading this code right, then eitc00 still = missing for all
>>years other than 2000. That means the regression will only include
>>data from 2000. Is this really what you want? How about this instead?
>>
>>replace eitc00 = 0 if eitc00 == .
>
> Thanks for that. I caught it too and made the change. However, the regressions still have an omitted coefficient on the interaction term because of collinearity.
*
*   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 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index