Spatial IV estimation: First stage

[Dario Maimone Ansaldo Patti]

Dear All,

I have an issue, which I am not able to solve. I am using the command spivregress, since in my model one of the covariates is endogenous. My issue is: how does Stata estimate the first stage? Does it use a standard OLS? Or does it take into account (I do not know how) spatial dependence when the first stage is estimated? I treid to browse Stata manuals, but apparently, I was unable to fnd something.

Thanks in advance.

Dario

[Andrew Musau]

The default spivregress estimator is the generalized spatial two stage least squares (GS2SLS) estimator. See Kalejian and Prucha (1998) for a description of the 3 step procedure (Section 3). Sure, there is a regular 2SLS IV regression in the first step of the procedure, but there are a number of other steps that follow.

Reference:
Kelejian, H. H., and I. R. Prucha. 1998. A generalized spatial two-stage least squares procedure for estimating a spatial autoregressive model with autoregressive disturbances. Journal of Real Estate Finance and Economics 17: 99–121.

[Dario Maimone Ansaldo Patti]

Andrew Musau thanks for your suggestion.