On a Standard Method for Measuring the Natural Rate of Interest

Daniel Buncic

April 20, 2021, 11:30–12:30

BDF Paris

Room Online

Séminaire Banque de France


This paper corrects the implementation of Median Unbiased Estimation (MUE) in Stage 2 of Holston, Laubach and Williams’ (2017) framework to estimate the natural rate of interest and provides corresponding corrected estimates. The correction is quantitatively important. It yields substantially smaller point estimates of the signal-to-noise ratio parameter λ(z) which determines the size of the downward trend of ‘other factor’ z(t) in the natural rate. For US data, the point estimate of λ(z) shrinks from 0.040 to 0.013 and is statistically highly insignificant. For data on the Euro Area, the UK and Canada, the λ(z) point estimates are exactly zero. Natural rate estimates from this model are up to 100 basis points larger than originally computed by Holston et al. (2017).


Natural rate of interest; Median Unbiased Estimation; Kalman Filter; misspecified econometric models; correction to Stage 2 model.;

JEL codes

  • C32: Time-Series Models • Dynamic Quantile Regressions • Dynamic Treatment Effect Models • Diffusion Processes
  • E43: Interest Rates: Determination, Term Structure, and Effects
  • E52: Monetary Policy
  • O40: General