The yield curve strikes back: New evidence of its predictive power for economic activity and inflation

Overview

This study examines the predictive capacity of yield curve factors—level, slope, and curvature—for economic growth and inflation across 40 countries spanning developed, Central and Eastern European, and emerging market economies from 2010 to 2021. The research differentiates countries based on monetary policy credibility and economic stability to assess whether these institutional characteristics modulate the informational content of yield curve measures.

Methods and approach

Yield curve factors were extracted from sovereign yield curves for each country in the panel using factor extraction techniques. Slope and curvature were subsequently incorporated into predictive regression frameworks using panel methodology. Out-of-sample forecasting accuracy was evaluated through novel panel forecasting techniques that integrate both time-series and cross-sectional information. Countries were stratified according to monetary policy credibility levels and economic stability measures to enable heterogeneous analysis of predictive relationships.

Key Findings

Yield curve slope and curvature demonstrate predictive power for economic growth in Central and Eastern European and developed economies, with this relationship materially strengthened in jurisdictions characterized by lower monetary policy credibility. In emerging markets, yield curve factors shape expectations about future growth and inflation but exhibit limited out-of-sample forecasting performance. Monetary policy credibility emerges as a critical determinant of yield curve informativeness, whereas economic stability does not substantively condition forecasting reliability. Yield curve factors provide restricted and temporally unstable signals for inflation prediction across the sample.

Implications

The heterogeneous predictive performance of yield curve factors across country groupings and institutional contexts indicates that the informational content of yield curve signals varies substantially according to domestic monetary policy framework credibility. The finding that predictive relationships are stronger in low-credibility jurisdictions reveals an important dimension of cross-country variation in yield curve informativeness for growth outcomes. These findings imply that yield curve analysis requires context-specific application rather than uniform deployment across heterogeneous economic environments.