Indirect inference in spatial autoregression


  • We thank the CoEditor and two referees for helpful comments and suggestions on the original version of this paper. Phillips acknowledges research support from the NSF under Grant No. SES 12-58258.

  • This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/ectj.12084


Ordinary least squares (OLS) is well known to produce an inconsistent estimator of the spatial parameter in pure spatial autoregression (SAR). This paper explores the potential of indirect inference to correct the inconsistency of OLS. Under broad conditions, it is shown that indirect inference (II) based on OLS produces consistent and asymptotically normal estimates in pure SAR regression. The II estimator used here is robust to departures from normal disturbances and is computationally straightforward compared with quasi maximum likelihood (QML). Monte Carlo experiments based on various specifications of the weight matrix show that: (i) the indirect inference estimator displays little bias even in very small samples and gives overall performance that is comparable to the QML while raising variance in some cases; (ii) indirect inference applied to QML also enjoys good finite sample properties; and (iii) indirect inference shows robust performance in the presence of heavy tailed error distributions.

This article is protected by copyright. All rights reserved