A New Approach to Estimating the Natural Rate of Interest

Building on the finding that the velocity of M1 is, to a close approximation, the permanent component of short-term nominal interest rates, this paper proposes estimating the natural rate of interest directly from velocity — projecting the monetary policy rate onto M1 velocity to obtain the nominal natural rate, then subtracting inflation’s sample average or target to obtain the real one. Applied to eight economies over samples ending in 2019Q4, the estimated real natural rate trends down throughout, and by 2019 the point estimate is negative in all of them except New Zealand and Norway. The paper defines the natural rate as a pure unit root process — the permanent component of the ex post real short rate — and claims two advantages over the two existing families of estimates, Laubach–Williams-style unobserved-components filters and DSGE-based estimates: under regimes that make inflation stationary the real natural rate is, up to a linear transformation, observed rather than filtered, and because M1 and interest rates are available at least weekly, the natural rate can in principle be computed at monthly or even weekly frequency given a high-frequency estimate of nominal GDP. In the United States, the euro area and Canada the monthly estimates fall sharply in the months around the collapse of Lehman Brothers, and the 1929 stock market crash is followed in the United States by a comparably dramatic decline. Two qualifications travel with the results: the whole approach rests on the assumption — argued from Benati (2020), not re-derived here — that M1 velocity is the short rate’s stochastic trend under a stable Selden–Latané demand for M1 balances; and the author flags explicitly that the higher the data frequency, the less credible it becomes that agents can actually perform the permanent–transitory decomposition the method attributes to them.

Luca Benati | Macro Paper Warehouse

A New Approach to Estimating the Natural Rate of Interest