Locally Stationary Quantile Regression for Inflation and Interest Rates

Abstract

Motivated by the potential time-varying and quantile-specific relation between inflation and interest rates, we propose a locally stationary quantile regression approach to model the inflation and interest rates relation. Large sample theory for estimation and inference of quantile-varying and time-varying coefficients are established. In empirical analysis of inflation and interest rates relation, it is found that the estimated functional coefficients vary with time in a complicated manner. Furthermore, the relation is quantile-specific: not only do the selected orders differ for different quantiles, but also the coefficients corresponding to different quantiles can display completely different patterns.

Keywords:

• Inflation
• Interest rates
• Locally stationary model
• Nonparametric estimation
• Quantile regression
• Time-varying model

Supplementary Materials

All technical proofs are presented in the Appendix (separate supplement file).

Acknowledgments

We are grateful to an associate editor and two referees for their constructive comments. Part of this work is based on Zhuying Xu’s PhD dissertation at Penn State University.