https://doi.org/10.52903/wp2026360
TREND INFLATION AND INFLATION EXPECTATIONS IN HIGH DIMENSIONAL VECTOR AUTOREGRESSIONS
Dimitrios P. Louzis
Bank of Greece
ABSTRACT
This paper introduces a new class of large vector autoregression (VAR) models with a hybrid trend structure that explicitly incorporates both trend inflation and inflation expectations, proxied by long-term survey forecasts and statistical filters. We develop efficient Bayesian estimation methods leveraging recent advances in matrix precision algorithms, substantially reducing computational costs and enabling large-scale forecasting exercises. Using a dataset of 20 U.S. macroeconomic variables, we show that incorporating trend inflation and survey-based expectations within a high-dimensional VAR framework markedly improves inflation forecast accuracy relative to widely used large-VAR benchmarks.
Keywords: Inflation Forecasting; Survey-Based Inflation Expectations; Large-Cross Section; Efficient MCMC algorithms
JEL-classifications: E31; E37; C51; C53; C55
Disclaimer: The author would like to thank Michele Modugno, Marta Bańbura, and participants at the CRETE 2023 Conference and the WGF 2024 Workshop (Athens) for helpful comments and suggestions. The views expressed in this paper are those of the author and do not necessarily reflect the views of the Bank of Greece. Any remaining errors or omissions are the sole responsibility of the author.
Correspondence:
Dimitrios Louzis
Economic Analysis and Research Department
Bank of Greece
El. Venizelos 21, 10250 Athens, Greece
Tel.: +30-2103202648
email: dlouzis@bankofgreece.gr