st: calculation of standard errors using estat mfx after asclogit

[Lisa Pfeiffer <lisapfeiffer@gmail.com>]

I'm having a problem with the calculation of standard errors through
estat mfx after asclogit. I get the "note: one or more numeric
derivatives failed; standard errors may be questionable". Depending on
the number of alternative specific constants I include (a max of 32),
many to all of the standard errors are not defined. What are the
possible causes of this failure? Is there a way I can calculate them
by hand?

Thanks,

Lisa Pfeiffer

Note: sorry if this is a double post; I received a jumbled message
from statalist and was not sure if it successfully posted.
*
*   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/