Re: st: SURE in a dynamic heterogeneous non-stationary unbalanced panel with small sample size

[Christopher Baum <kit.baum@bc.edu>]

On Aug 8, 2010, at 2:33 AM, Anirudh wrote:

> 
> Yes Kit, they are. I have four regressors in each equation. What do you suggest I do?

First of all, if the regressors are numerically identical and you CAN estimate the system with SURE, it will give you exactly the OLS estimates.

If the equations' dependent variables also represent a set of shares, or adding up conditions, such that for every observation the dep.vars. add up to one, then by definition you cannot estimate the full set of equations, because the matrix of residuals must be singular in that case. This arises with a comprehensive set of demand equations or portfolio shares. Then you can drop any one of the equations and estimate the others. The coefficients for the omitted equation can be calculated, in point and interval form, from the equations you do estimate. If you use the -isure- option of -sureg-, the results are invariant to the choice of omitted equation, and equivalent to the MLE of the system.

Kit



Kit Baum   |   Boston College Economics & DIW Berlin   |   http://ideas.repec.org/e/pba1.html
                              An Introduction to Stata Programming  |   http://www.stata-press.com/books/isp.html
   An Introduction to Modern Econometrics Using Stata  |   http://www.stata-press.com/books/imeus.html


*
*   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/