Post-Sample Granger Causality Analysis: A New (Relatively) Large-Scale Exemplar

Ashley and Ye (2012) exemplifies the state-of-the-art in post-sample Granger causality analysis in a small-scale (bivariate) setting, albeit with a sufficiently large sample (T = 480 months) as to make post-sample testing feasible. In the present work we extend this work in two directions. First, here we analyze four macroeconomically important endogenous variables – monthly measures of aggregate income, consumption, consumer prices, and the unemployment rate – embedded in a six-dimensional information set which also includes two interest rates, both taken to be exogenous. Second, we compare the causality results obtained using a traditional large-to-small (but partially judgmental) model identification procedure to those obtained using the objective (but mechanical) "Autometrics" identification procedure given by Doornik and Hendry (2007).