Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | natasha agarwal <agarwana2@googlemail.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | st: Estimating firm level data on regional level data using a within estimator. |
Date | Sun, 6 Jun 2010 11:49:41 +0100 |
Hello Everyone, I hope someone can help me. I estimate an equation where you VA of domestic firms (at the firm level)= L of dom firms (at the firm level)+ K of dom firms (at the firm level)+FDI(at the regional level)+time dummies. So it looks like VA(it)=K(it)+L(it)+FDI(rt)+D(t)+e(it) I estimate it with a within estimator, i.e. ur fixed effects. When I do so the minute I include all the time dummies, FDI goes insig but thats my main variable of interest. When Ii try to estimate it for each region the same equation, FDI drops out. The problem that I have found out is that FDI is behaving like a time dummy because all the firms within a region for each year will have the same value of FDI. So like FDI for Beijing in 2001 will be say 20000. so all the firms within the same region will have 20000 and the time dummy for 2001 will be 1 so all the firms withinhe region will also have 1 for 2001. So basically the variation for all the firms within the same region will be zero as all the firms will have the same FDI value for all the years from 2001-2005 which is in effect like a constant and like a time dummy. So at the regional level the variation is between regions and across time. But since am estimating at the firm level the variation is only across time for al the firms within the same region. Hence, when I throw in all the time dummies + fdi, FDI goes insig and when i try to estimate it for each province, FDI dropus out when i include the time dummies. I have ran out of solutions. I have tried including a time trend inst have tried including a time trend instead of a time dummy, the results for the aerage remains insig but the time trend alows me to estimate the same for each region. I have tried allot of other things as well like standarsation and all but it all doesn't seem to work for each year again. I have also tried multi-level modeling like the xtmixed in stata but frm the theoretical resources available that am reading on it, it seems that it focuses more on the random effects intercepts and slopes and I am not sure whether this is what i need to look at.... Does anybody have any suggestion?!?!?! Thanks Natasha Do you have any suggestions?!?!?!?!?!? * * 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/