Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Fractional Response Model with Unbalanced Panel data

    Hello,
    I am estimating the unbalanced panel data for the model where dependent variable is a fraction. Among endogenous variables, I have one continuous and many binary variables. I would like to apply fractional probit estimator for panel data, however I am not sure whether it would work for unbalanced panel data. Could anyone give me advice and stata command how to estimate a model like this?

  • #2
    In #1 you write "Among endogenous variables, I have one continuous and many binary variables." Do you mean "exogenous" here?

    In any event, as you may already know a key reference is Papke and Wooldridge, J. Econometrics 145 (2008) 121–133, who suggest challenges using the methods they propose in the case of unbalanced panels. ("The nonlinear models we apply are difficult to extend to unbalanced panel data – a topic for future research." p.127)

    Comment


    • #3
      To follow up on John's comment, I'm attaching a paper, which I first wrote in 2009, that suggests how to extend the method to unbalanced panels. But it's only an issue if you are explicitly allowing the heterogeneity to be correlated with the explanatory variables. With an unbalanced panel, one extends the Mundlak/Chamberlain device to depend on the missing data (or selection) indicators, too.

      Attached Files

      Comment


      • #4
        John Mullahy Jeff Wooldridge Thank you so much for your comments! I appreciate your support and sharing attached manuscript is very kind of you! Well, I will try to write a command in Stata and see whether it will work with my data, given the assumption that explanatory variables can be correlated with heterogeneities.

        Comment

        Working...
        X