Friedman’s money supply volatility hypothesis and the quantity theory of money: the evidence from the different types of money

This paper tests the Milton Friedman hypothesis that money growth variability affects the velocity of money. It compares the Federal Reserve’s simple-sum monetary aggregates, the original (non-credit-card-augmented) Divisia aggregates, and the credit-augmented Divisia and credit-augmented Divisia inside monetary aggregates using two classes of empirical models. Standard Granger causality tests and a bivariate structural GARCH-in-Mean VAR model examine how money growth variability affects the consumption and income velocity of money. Using monthly data for the post-2006 period, when all monetary indicators are available, the findings reject the Friedman hypothesis when money supply volatility is modeled explicitly using recent advances in macroeconometrics and financial econometrics. This result is robust to a longer sample extending back to 1967 with monthly and quarterly data, across alternative model specifications. From a policy standpoint, this enhances the predictability and reliability of monetary policy actions aimed at stabilizing the economy using monetary aggregates.