st: Tobit and correct identification of censored/uncensored observations

[Mario Veneziani <agreconucscpc@gmail.com>]

Dear all,

I would appreciate your help and hints on the following issue.

I have a strongly unbalanced panel of farms I would like to estimate
the "determinants" of the value of rent per hectare according to a
number of RHS covariates (i.e., output per hectare (yha11), subsidies
received per hectare and other exogenous variables). The time
dimension is roughly 18 years but only a few observations are repeated
for a reasonable number of years.
Since only roughly half (50%) of my sample rents land, hence reports a
value for paid rent I can use to calculate the value of rent per
hectare, my dependent variable can be censored at 0, with a minimum
value of rent per hectare being roughly 0.0029.
I have tried running the - xttobit - command in Stata 12.1 MP since a
Panel Tobit appears to be a suitable model to begin my investigation
with. Beside the significant problems in achieving convergence, when I
indicate Stata that the variable rent per hectare is left censored by
running

 - xttobit  rentpha yha11, ll(rentpha |0.0029346) -

the few times I manage to achieve convergence Stata informs me that
the (tentative and unreliable) estimates I obtain are based only on
left-censored observations while 0 uncensored observations appear to
be considered.

It this issue related to the failure to achieve convergence or am I
overlooking something? Apologies for the likely naive question but I
am new to this model's estimation.
Any help or suggestion on these matters is highly appreciated.
Best
Mario

-- 
Mario Veneziani
Research Assistant
Istituto di Economia Agroalimentare
Università Cattolica del Sacro Cuore
Piacenza, Italy

*
*   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/