Biases of OLS and Spatial Lag Models in the Presence of an Omitted Variable and Spatially Dependent Variables

Abstract

A number of authors have suggested that omitted variables affect spatial regression methods less than ordinary least-squares (OLS). To explore these conjectures, we derive an expression for OLS omitted variable bias in a univariate model with spatial dependence and show that positive dependence in the disturbances, regressand, and regressor magnifies the magnitude of conventional omitted variables bias. Moreover, we show that spatial dependence in the regressor exacerbates the usual bias that occurs when using OLS to estimate a spatial autoregressive data generating process (DGP).