Do Monetary Aggregates Improve Inflation Forecasting in Switzerland?

Authors

  • Logan J. Kelly University of Wisconsin
  • Jane M. Binner University of Birmingham
  • Jonathan A. Tepper Aston University

DOI:

https://doi.org/10.33423/jmpp.v25i1.6922

Keywords:

management policy, Divisia monetary aggregates, inflation, recurrent neural networks, Swiss monetary policy

Abstract

This study examines whether or not Swiss monetary aggregates enhance inflation forecasting in Switzerland during the out-of-sample period, December 2008 to November 2019. We use a state-of-the-art multi-recurrent neural network endowed with a sluggish state-based memory to approximate a non-linear auto-regressive moving average model. Conventional monetary aggregates have been shown to lose dynamic information, potentially explaining why many deem traditional measures of the money supply to have minimal economic relevance. Our findings suggest that when conventional monetary aggregates, Divisia money measures, and a short-term interest rate are combined, forecasts of Swiss inflation over the 12, 24 and 36-month forecasting horizons are significantly improved compared to a model that excludes a measure of the money supply.

Downloads

Published

2024-04-26

How to Cite

Kelly, L. J., Binner, J. M., & Tepper, J. A. (2024). Do Monetary Aggregates Improve Inflation Forecasting in Switzerland?. Journal of Management Policy and Practice, 25(1). https://doi.org/10.33423/jmpp.v25i1.6922

Issue

Section

Articles