Revisiting a Post Keynesian Explanation of US Inflation

Key findings from this study

• The study found that the Post Keynesian wage-cost markup model exhibits limited robustness when applied to 2002–2024 U.S. data, failing to maintain explanatory power across alternative specifications of wages, unemployment, and inflation measures.
• The researchers demonstrate that unit labor costs display a non-robust negative relationship with inflation in initial specifications, becoming statistically insignificant after controlling for energy costs and imports and correcting for serial correlation.
• The authors report that coefficients in the aggregate demand–augmented wage-cost markup equation declined substantially in magnitude and significance during the 2002–2024 period compared to earlier time frames, indicating potential structural shifts in inflation transmission mechanisms.

Overview

This paper evaluates the explanatory power of a Post Keynesian wage-cost markup model for U.S. inflation across 2002–2024, incorporating recent pandemic-era inflation dynamics. The analysis employs a reduced-form equation linking aggregate demand, unit labor costs, and wage growth to inflation outcomes. Testing proceeds with alternative specifications for labor market slack, wage measures, and inflation metrics to assess model robustness.

Methods and approach

The study utilizes quarterly U.S. data spanning 2002 to 2024 in an econometric framework centered on a wage-cost markup equation augmented for aggregate demand effects. Model specification includes a wage growth equation to capture labor market dynamics. Robustness testing incorporates multiple operational definitions of labor market slack, wage variables, and inflation indicators. The Prais–Winsten estimation technique addresses persistent serial correlation in residuals. Control variables for energy costs and imports are successively introduced to assess their influence on the relationship between unit labor costs and inflation.

Results

The comprehensive Post Keynesian model demonstrated limited robustness across the full 2002–2024 period, irrespective of alternative variable specifications. Initial model estimation revealed a negative relationship between unit labor costs and inflation, inconsistent with theoretical expectations. Introduction of energy cost and import controls eliminated this negative relationship, rendering the unit labor cost coefficient statistically insignificant after Prais–Winsten correction. The revised specification restored alignment with the Atesoglu wage-cost markup equation's sign conventions, though the unit labor cost coefficient remained statistically non-significant. Coefficients in the aggregate demand–augmented wage-cost markup equation declined substantially in both magnitude and statistical significance during the 2002–2024 period relative to earlier periods.

Implications

The deterioration in model fit and coefficient significance suggests fundamental shifts in inflation transmission mechanisms within the U.S. economy during this period. The inability of the wage-cost markup framework to maintain explanatory power despite theoretical refinement indicates that either structural changes in price-setting behavior have occurred or additional economic variables not captured in the original Post Keynesian specification now exert dominant influence. The distinction between wage growth and broader inflation measures requires investigation, as the pass-through mechanism appears weaker than historical precedent would predict. Energy costs and import price pressures emerged as material factors that interact with domestic wage dynamics to determine inflation outcomes.
The study contributes to Post Keynesian empirical work by documenting the conditional applicability of foundational inflation models across different economic periods. Rather than validating the comprehensive model, these findings highlight necessary areas for theoretical extension or augmentation. The 2002–2024 sample encompasses diverse macroeconomic regimes: the pre-crisis expansion, the financial crisis period, subdued recovery, and supply-shock-driven inflation following COVID-19. This heterogeneity may require period-specific or regime-dependent analytical frameworks.
The findings inform ongoing debates about the causal mechanisms underlying inflation. If unit labor costs no longer predict inflation movements with statistical significance, economists must reconsider wage dynamics' role in price determination or examine whether measurement issues obscure true relationships. The reduced explanatory power over recent decades warrants investigation into whether central bank credibility, international supply chains, or sectoral recomposition have fundamentally altered how domestic labor market conditions transmit to consumer price changes.

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.