Abstract:

This paper studies estimation and inference in the context of a dynamic spatial
model with interactive effects. An instrumental variables interactive fixed effects
(IV-IFE) estimator is proposed which provides consistent and asymptotically un-
biased estimates of model parameters, as long as the cross-sectional dimension of
the data grows sufficiently fast relative to number of time periods and the number
of instruments. This trivially includes where both the number of time periods and
the number of instruments are fixed. Nonetheless, circumstances exist where the
estimator can exhibit asymptotic bias, the extent to which depends, in part, on
the structure of the spatial dependence. A bias corrected estimator is constructed,
which, through simulation, is demonstrated to be an effective remedy to this issue.
An empirical application utilises the method to study the relationship between eco-
nomic growth, civil liberties and political rights in the 21st century