When I try to add spatial lags of instrumental variable to spivregress (Spatial autoregressive models with endogenous covariates, a command added with Stata 15), I get an error saying that IVs cannot be spatially lagged.
However in spatial econometric theory there are some models (e.g. the general nesting spatial models https://www.rug.nl/staff/j.p.elhorst...trics_2017.pdf) which include lags of IVs where these are spatially autocorrelated.
Is there a way to cope with the issue?
However in spatial econometric theory there are some models (e.g. the general nesting spatial models https://www.rug.nl/staff/j.p.elhorst...trics_2017.pdf) which include lags of IVs where these are spatially autocorrelated.
Is there a way to cope with the issue?
Comment