Volatility of money growth and the U.S. velocity puzzle: Some further results

Abstract

This study reexamines the Friedman hypothesis that money growth volatility has caused the recent puzzling behavior of the U.S. money velocity. While some empirical studies reported favourable evidence, others did not. However, most previous results in this area may lack credibility due to several problems. We use a six-variable vector-autoregressive (VAR) model which allows for the endogeneity of all variables in the model, including velocity and money growth volatility. Moreover, we make adjustments for possible structural breaks in the underlying relationships and check for stationarity using unit-root tests. The empirical results appear inconsistent with the Friedman hypothesis. The results suggest that money growth volatility and velocity (however defined) are unrelated, a finding supported by some cointegration tests.