From section III, "How Should Monetary Policy Be Conducted," of the 1968 AEA presidential address.
My own prescription is still that the monetary authority go all the way in avoiding such swings by adopting publicly the policy of achieving a steady rate of growth in a specified monetary total. The precise rate of growth, like the precise monetary total, is less important than the adoption of some stated and known rate.
Having argued that monetary policy cannot reliably steer real magnitudes, Friedman draws the constructive conclusion: stop trying. A pre-announced constant growth rate for a chosen aggregate removes the central bank as an independent source of disturbance and lets the private economy plan against a known nominal anchor. The striking move is the second sentence — Friedman is nearly indifferent to which rate and which aggregate, because the value of the rule lies in its being a rule at all rather than in its calibration. The prescription was never adopted in the literal monetarist form; the monetary-aggregate relationships it relied on grew unstable through the 1980s. But its deeper logic — that credible, rule-like, pre-committed policy beats period-by-period discretion — survived into the inflation-targeting and central-bank-independence consensus that followed the stagflation decade.