Dynamic binary panel models with endogeneity

Hi all,

I have a question related to dynamic panel analysis.

As we know, xtabond2 is suitable to use for the dynamic linear panel models. What happens if the dependent variable is binary (dummy variable)? Can we still use xtabond2 for a binary dependent variable by considering the model as the linear probability model?

And is it ok if we run xtabond2 with T=3? (both linear and binary models)

Model specification:
y_i,t = f(y_i,t-1, x_i,t, Z_i,t)
y_i,t-1 and x_i,t are endogenous variable
y is binary (0,1)

Thanks in advance

Share on FacebookPlus on Google+