st: RE: st: Frontier Cost & Production Models

["Clifton Chow" <clifton_chow@post.harvard.edu>]

Does anyone know what the difference is between the stochastic frontier cost and production models and why if one converges, the other does not?  In my dataset the dependent variable is measured in $ and had an extreme value of $500 (when the mean is $15 and the next highest below that value is $45).  When I ran the frontier model with the extreme value both cost and production models converged in the maximum likelihood estimation, but when I left it out, only the production model converged for all three options (Half normal, Truncated Normal & Exponential), but none of the cost models converged.  Any idea why?

Thanks in advance
PhD Candidate
*
*   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/