Learning-Based Inflation Expectations in an Unobserved Components Model

• Activity Number: 330 - Bayesian Analysis of Latent Variable Models in Economics
• Type: Topic Contributed
• Date/Time: Tuesday, July 31, 2018 : 10:30 AM to 12:20 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Srikanth Ramamurthy*
• Companies: Loyola University Maryland
• Keywords: Output Gap; Unobserved Components Model; Adaptive Learning; Inflation Expectations; New Keynesian Phillips Curve

Abstract

We examine, within the context of an Unobserved Components (UC) model, the role of adaptive learning-based inflation expectations in determining the output gap. Keeping in line with the literature, the forward-looking New Keynesian Phillips curve (NKPC) serves as the backbone for modeling inflation dynamics. The principal goal of working with endogenously derived forecast rules, as opposed to, say, using survey data as proxy for inflation expectations (Basistha and Nelson, 2007), is to quantity directly the extent to which such forecasts drive cycles and trends, with particular focus around turning points. For completeness, we benchmark learning-based forecast rules against both rational expectations and the aforementioned survey data for inflation expectations. We tackle this problem empirically with the aid of modern Bayesian tools. Preliminary results suggest that the implied output gap from our modeling approach is more persistent compared to the benchmarks, and that the predictive importance of forecast errors can change substantially over time.