Stock returns, real activity, and money: Causality testing using a multiple hypothesis testing approach

Abstract

There is considerble debate concerning the dynamic relationship among stock returns and real activity. Two explanations that have received particular attention are those of the proxy hypothesis and the reverse causality explanation. The principle distinguishing feature between these two explanations concerns whether or not inflation merely proxies for some underlying relation among stock returns and real activity. The purpose of this article is to re-examine the evidence within the context of multiple time-series models. Our re-examination is motivated by some new developments in the area of causality testing employing Vector Autoregressive Moving Average (VARMA) models. These new developments have focused on the issue of causality testing when the hypotheses are not completely nested (as in VARMA models, by construction), and the potential bias that may arise by the inference procedure employed. Thus, this article carefully selects an inference procedure which permits one to infer the direction of potential bias and finds that the results of previous work are sensitive to the inference procedure employed. Controlling for this bias, we report that the evidence favors the proxy hypothesis.