Causal Change Detection in Possibly Integrated Systems : Revisiting the Money-Income Relationship

This paper re-examines changes in the causal link between money and income in the United States for over the past half century (1959-2014). Three methods for the data-driven discovery of change points in causal relationships are proposed, all of which can be implemented without prior detrending of the data. These methods are a forward recursive algorithm, a recursive rolling algorithm and the rolling window algorithm all of which utilize subsample tests of Granger causality within a lag-augmented vector autoregressive framework. The limit distributions for these subsample Wald tests are provided. The results from a suite of simulation experiments suggest that the rolling window algorithm provides the most reliable results, followed by the recursive rolling method. The forward expanding window procedure is shown to have worst performance. All three approaches find evidence of money-income causality during the Volcker period in the 1980s. The rolling and recursive rolling algorithms detect two additional causality episodes: the turbulent period of late 1960s and the starting period of the subprime mortgage crisis in 2007