Spatial Durban Model Han-Philips - spregdpd

[Bernhard Lambernd]

Hello everyone,

I performed a spatial regression with a Han-Philips estimator and used a SDM model with spregdpd:

spregdpd LN_SOX_MOV_POP_MOV GDP_POP_1000 GDP_POP_1000_SQ GDP_POP_1000_CU LN_FDI_W2 LN_POP_AREA LN_ENUSE LN_CARS LN_CAP_EMP2 LN_RaD_EXP LN_ENVEXP_GOV LN_EDU, nc(16) wmfile(weigth_50km) model(sdm) run(xtdhp) fe endog(GDP_POP_1000 LN_FDI_W2) coll vce(robust)


I expected to receive an output detailing the following variables: Y = BX + aWX* + rWy + e ; e = lWe+u

Surprisingly, the spillover effect of Y (rWy) was not reported in the STATA output.
Can anyone help me an explain me why a regression output of a SDM model with an Han-Philips estimator does not report rWy?

Thanks in advance!

Bernhard

[Sebastian Kripfganz]

spregdpd is a community-contributed command that I cannot comment on. Since Stata 15, there is the official command spregress that can estimate your kind of model.

[Bernhard Lambernd]

Thanks for your answer!
Unfortunately, I work with Stata 14 and thus can not use spregress.