Detecting causal relationships between spatial processes

Abstract

The paper focuses on the problem of testing for the existence and direction of causality in a group of variables, with a spatial framework. Our objective is to produce useful criteria to discuss causality on rational and objective bases. We introduce a new non-parametric test, based on symbolic entropy, which is robust to the functional form of the relation. The test has good behaviour in samples of medium to large size. The proposal is illustrated with an application to the case of migration versus unemployment, using data on 3,108 US counties for the period 2003–2008.