The Return of the Liquidity Effect: A Study of the Short-Run Relation Between Money Growth and Interest Rates

This paper asks whether short-run co-movements between money growth and interest rates reflect the traditional “liquidity effect” – money growth temporarily lowers real and nominal rates before eventually raising inflation – or the competing “anticipated inflation effect,” in which money growth signals future inflation and raises nominal rates with no offsetting decline in real rates. The two views make opposite predictions for the sign of the short-run correlation between money growth and interest rates (negative versus positive), which Cochrane tests using weekly U.S. M1, three-month Treasury-bill, and 20-year government bond data. He restricts the sample to the Federal Reserve’s October 1979-November 1982 nonborrowed-reserve-targeting episode specifically because, outside that period, the Fed’s active smoothing of interest rates creates reverse-causality (simultaneous-equations) bias in any regression of rates on money growth. To isolate the short-run relationship without needing to estimate the full long-run-plus-short-run-plus-noise relation between the series, he applies two-sided band-pass (spectral-window) filters that pass only cycles in a chosen frequency band – for example 12-26 or 26-52 weeks – generalizing the low-pass-filter approach Lucas (1980) and Summers (1983) had used to study long-run monetary neutrality and the long-run Fisher effect. Regressing filtered interest rates on filtered money growth across several such windows, Cochrane finds a consistently negative correlation for the three-month bill rate and, more surprisingly, for the 20-year bond rate as well, which he interprets as the liquidity effect dominating the anticipated-inflation effect at these short-run frequencies; a Geweke-style feedback test does not reject the absence of reverse causation from interest rates to money growth within the studied windows. The same negative correlation is largely absent when the same method is applied to the surrounding 1976-79 and 1982-86 subsamples, consistent with the finding depending on the Fed’s reduced interest-rate smoothing during the targeting experiment specifically. Cochrane reads the result as consistent with money-growth changes being largely unanticipated, money growth being a poor predictor of its own future path, and inflation following money growth only with a long lag, while explicitly cautioning that the reduced-form correlation documents the existence of a liquidity effect without quantifying its structural magnitude or duration.

Journal of Business & Economic Statistics | Macro Paper Warehouse

The Return of the Liquidity Effect: A Study of the Short-Run Relation Between Money Growth and Interest Rates