The Phillips Curve and Inflation Expectations: A Machine Learning Perspective

Key findings from this study

This research indicates that:

• Monetary policy transmission to inflation strengthened during the inflation-targeting era, indicating more effective policy channels.
• Phillips curve flattening characterized the pre-pandemic period, reflecting reduced inflation responsiveness to economic slack.
• Long-horizon inflation expectations became less sensitive to short-term expectation measures, suggesting anchored credibility.

Overview

This research develops a machine learning framework to examine time-varying transmission mechanisms of U.S. monetary policy under inflation targeting. The authors employ ridge regression estimation of local projections with time-varying parameters to detect gradual shifts in macroeconomic relationships. The framework accommodates parameter heteroskedasticity, enabling detection of regime-dependent dynamics in monetary policy effectiveness and Phillips curve behavior.

Methods and approach

Time-varying parameter local projections estimated via ridge regression constitute the primary methodological contribution. Ridge regression addresses potential multicollinearity and overfitting inherent in high-dimensional macroeconomic datasets. The approach permits heteroskedasticity across parameters, capturing differential volatility in estimated coefficients. This design enables identification of structural breaks and gradual transitions in policy transmission channels without imposing predetermined breakpoint dates.

Results

Monetary policy's effect on inflation has intensified throughout the analysis period, indicating strengthened transmission mechanisms. The Phillips curve exhibits flattening across most of the pre-pandemic era, reflecting reduced sensitivity of inflation to slack measures. However, pass-through from short-horizon to long-horizon inflation expectations diminished substantially, consistent with stronger expectation anchoring and enhanced monetary policy credibility.

Post-pandemic dynamics reveal regime dependency in Phillips curve behavior. During the inflation surge following the pandemic, the Phillips curve temporarily steepened, suggesting that nonlinearities and threshold effects characterize inflation-slack relationships. This empirical pattern indicates that Phillips curve stability varies with economic conditions rather than remaining constant across regimes.

Implications

The findings suggest monetary policy transmission operates through expectation channels more prominently than through conventional slack-based mechanisms. Anchored long-term expectations appear to have insulated inflation dynamics from short-term policy surprises, indicating successful institutional credibility accumulation. This stability enabled policy authorities to pursue aggressive accommodation during the pandemic without proportional inflation consequences until supply-side disruptions overwhelmed demand-management frameworks.

Regime-dependent Phillips curve dynamics complicate conventional monetary policy guidance predicated on stable relationship estimates. Policymakers encounter distinct inflation-output trade-offs under different economic conditions, requiring adaptive policy frameworks rather than static reaction functions. Machine learning methodologies that accommodate time variation provide empirical tools for real-time detection of structural transitions.

Scope and limitations

This summary is based on the study abstract and available metadata. It does not include a full analysis of the complete paper, supplementary materials, or underlying datasets unless explicitly stated. Findings should be interpreted in the context of the original publication.