Dear Statalisters,
I test for endogeneity of health insurance in health care utilization, by
jointly estimating 2 equations:
utilization equation: dependent variable = utilization
independent variables = insurance + others
Insurance equation: dependent variable = insurance
independent variable = age + others
Bivariate probit model is used. A Hausman test is then performed on the result
estimated from the bivariate probit model and the univariate probit model (of
utilization equation).
All parameters are estimated in these two models. However, in Hausman test some
values in the column of covariance matrix difference is missing, despite the
Hausman statistic is output. I want to ask whether this missing values will
affect the crediblility of the Hausman statistics, thus making the Hausman test
result untrustable?
Thanks a lot for kindly helping!
Sincerely,
Isaac Yip
*
* 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/