Money and Inflation: Lost Cointegration

Abstract

Using a cointegration, we show that there is no long-term relationship between money in the economy M and real (and nominal) GDP and CPI (US data from 1959 to 2018). There is no empirical evidence to support the textbook claim that "inflation is always and everywhere a monetary phenomenon". Only when we shorten the time series to the period before the crisis (1959-2008), there is a cointegration between CPI and M2, but only at the 10% significance level and only according to one of two co-integration tests. The relationship that existed before the crisis either had to fall apart or change. There are three possible explanations: (1) The growth of M in low-inflation economies (CPI below 10% annually) is distributed more equally between CPI and real GDP than in the event of significant changes in M. (2) The falling velocity of money after the crisis of 2008/2009. (3) The last possibility is an increase in the adequacy problem of inflation - the CPI does not adequately reflect the economic definition of inflation.