Latent Variable Phillips Curve
Daniil Bargman, Francesca Medda, Akash Sedai Sharma
TL;DR
The paper rethinks the Phillips curve by positing an unobserved price-pressure process and evaluates a latent-variable formulation (LVPC) via Latent Shock Regression (LSR). It conducts an out-of-sample, cross-factor comparison across 3,968 specifications using US core PCE inflation data from 1983 to 2025. LVPC forecasts outperform traditional PC models in the medium term (six to eight quarters ahead) and show competitive gains versus univariate benchmarks, with MA(1) residuals providing additional improvements. The findings offer a conceptual advance and practical guidance for improving Phillips-curve-based forecasts, including models that account for moving-average residuals and seasonal patterns.
Abstract
This paper re-examines the empirical Phillips curve (PC) model and its usefulness in the context of medium-term inflation forecasting. A latent variable Phillips curve hypothesis is formulated and tested using 3,968 randomly generated factor combinations. Evidence from US core PCE inflation between Q1 1983 and Q1 2025 suggests that latent variable PC models reliably outperform traditional PC models six to eight quarters ahead and stand a greater chance of outperforming a univariate benchmark. Incorporating an MA(1) residual process improves the accuracy of empirical PC models across the board, although the gains relative to univariate models remain small. The findings presented in this paper have two important implications: First, they corroborate a new conceptual view on the Phillips curve theory; second, they offer a novel path towards improving the competitiveness of Phillips curve forecasts in future empirical work.