In the early 1940s, Friedman’s own analysis of monetary policy adhered closely to the dismissive tone prevalent in much other Keynesian literature of that vintage. His solo-authored contribution to 1943’s Taxing to Prevent Inflation, written while he was at the Treasury, plotted growth rates of the nominal money stock and nominal income for the United States for the period 1899-1929. To the modern reader, the scatter plot in Friedman’s paper indicates that the monetary growth/income relationship is clearly positive, and reasonably tight by the standards of rate-of-change data. That was not, however, the judgment Friedman reached in his 1943 paper, in which he concluded instead that the relationship was “extremely unstable.”
That is from p.95 of the recent Edward Nelson two-volume set on Milton Friedman — one of the best books written on any economist!