Dear Statalist members,
I want to run a regression with a growth rate as dependent variable. For about 90% of the observations the growth rate is equal to zero, for 5% it is negative and for 5% it is positive.
I learned that a two-way tobit, which can treat the zeros as a discrete variable with a mass point and the negative and positive growth rates as continuous variable, would be more appropriate than OLS.
Has someone already written a programme for this problem or for a similar problem?
My data is of the pooled cross-section time-series type, therefore I would like to get standard errors that are robust to heteroscedasticity and autocorrelation.
Many thanks in advance,
Elisabeth
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
