Working Paper

The Zero Lower Bound and Estimation Accuracy


Abstract: During the Great Recession, many central banks lowered their policy rate to its zero lower bound (ZLB), creating a kink in the policy rule and calling into question linear estimation methods. There are two promising alternatives: estimate a fully nonlinear model that accounts for precautionary savings effects of the ZLB or a piecewise linear model that is much faster but ignores the precautionary savings effects. Repeated estimation with artificial datasets reveals some advantages of the nonlinear model, but they are not large enough to justify the longer estimation time, regardless of the ZLB duration in the data. Misspecification of the estimated models has a much larger impact on accuracy. It biases the parameter estimates and creates significant differences between the predictions of the models and the data generating process.

Keywords: Bayesian Estimation; Projection Methods; Particle Filter; OccBin; Inversion Filter;

JEL Classification: C11; C32; C51; E43;

https://doi.org/10.24149/wp1804r1

Access Documents

File(s): File format is application/pdf https://www.dallasfed.org/-/media/documents/research/papers/2018/wp1804r1.pdf
Description: Revised version

File(s): File format is application/pdf https://www.dallasfed.org/-/media/documents/research/papers/2018/wp1804.pdf
Description: Original version

Authors

Bibliographic Information

Provider: Federal Reserve Bank of Dallas

Part of Series: Working Papers

Publication Date: 2019-02-01

Number: 1804

Pages: 36 pages

Note: A previous version of this paper circulated with the title, "The Accuracy of Linear and Nonlinear Estimation in the Presence of the Zero Lower Bound."