EcoFin Research Papers

This paper revisits the theoretical and empirical evolution of the Phillips Curve through the lens of modern New Keynesian macroeconomics. While the traditional unemployment–inflation relationship has long been viewed as unstable, recent advances attribute its variability to the interaction of expectations, nominal rigidities, and structural features such as openness and imported marginal costs. The review synthesizes key developments in the New Keynesian Phillips Curve, including forward-looking price setting, hybrid indexation, sticky-information dynamics, and small open-economy extensions. Empirical evidence across advanced, emerging, and transition economies reveals substantial heterogeneity in slope, persistence, and the relative weight of backward- and forward-looking components, particularly across tranquil and recessionary periods. The findings highlight that the Phillips Curve is conditional rather than structural, with inflation dynamics fundamentally shaped by the credibility of monetary policy, the structure of expectations, and the sensitivity of marginal costs to domestic and external shocks.

This paper provides a comprehensive meta-analysis of advances in the New Keynesian Phillips Curve (NKPC) literature, synthesizing theoretical developments, empirical findings, and methodological innovations over the past three decades. Rather than aggregating coefficients mechanically, the study adopts a structured qualitative–quantitative meta-analytic approach to evaluate how inflation dynamics vary across economic regimes, institutional settings, and model specifications. The analysis reveals that apparent instability and flattening of the Phillips Curve largely reflect regime dependence, expectation anchoring, and openness to global cost pressures rather than a breakdown of the underlying NKPC mechanism. Evidence across advanced, emerging, and transition economies indicates that forward-looking inflation behavior strengthens in tranquil macroeconomic environments with credible monetary frameworks, while backward-looking inertia dominates during recessions and in economies with histories of volatile inflation. Hybrid and sticky-information NKPC formulations consistently outperform purely forward-looking specifications in capturing inflation persistence, particularly during periods of heightened uncertainty. Recent methodological contributions, including time-varying, Bayesian, frequency-domain, and machine learning approaches, further demonstrate that the Phillips relationship is nonlinear, state-dependent, and horizon-specific. Overall, the findings suggest that the NKPC remains a valid but conditional framework for understanding inflation dynamics, with its empirical performance critically shaped by expectations regimes, institutional credibility, and global integration.

The COVID-19 pandemic has reignited fundamental debates about the stability, slope, and empirical relevance of the Phillips Curve. While the relationship between inflation and economic slack deteriorated during the pandemic, recent scholarship offers new insights that complicate simplified narratives about the curve’s failure. This meta-analysis synthesizes evidence from pre-COVID panel studies and the rapidly growing body of post-2020 research. Earlier cross-country analyses illustrate that the inflation–unemployment relationship is highly conditional on macroeconomic tranquillity, institutional structure, and inflation expectations, particularly during recessionary periods. Post-COVID research advances this perspective by uncovering substantial nonlinearities, structural breaks, frequency-dependent dynamics, anchored expectation effects, trend inflation interactions, and regional heterogeneity. The pandemic shock provides an empirical environment in which many conventional mechanisms behave atypically, yet recent studies show that the Phillips relationship re-emerges once expectations stabilize and supply disruptions ease. Integrating evidence from hybrid NKPC models, time-varying instrumental variable techniques, Bayesian structural break approaches, spatial panels, sectoral analyses, and explainable machine learning, this study concludes that the Phillips Curve has neither collapsed nor returned unchanged. Instead, its apparent instability reflects the interaction between forward-looking expectations, inflation persistence, structural labor-market transformations, capacity constraints, and global supply linkages. These findings reaffirm the importance of hybrid modeling frameworks and caution against interpreting reduced-form flattening as evidence of structural irrelevance.

This paper develops an industry-based perspective on the Phillips Curve, arguing that the inflation–unemployment nexus varies systematically across sectors due to differences in labor intensity, market power, pricing flexibility, and moral or regulatory constraints. While the traditional Phillips Curve assumes a homogeneous transmission from labor market slack to inflation, modern economies are increasingly characterized by sectoral asymmetries. In labor-intensive industries such as manufacturing and personal services, wage pressures transmit strongly into prices, generating a steep Phillips relationship. In contrast, capital-intensive, digital, or morally constrained industries— including information technology and sin-related sectors—exhibit weak or unstable inflation–unemployment linkages due to pricing power, non-labor cost dominance, reputational considerations, and demand inelasticities. Integrating sectoral real marginal cost composition into a hybrid New Keynesian Phillips Curve framework, the paper provides a conceptual foundation for understanding why aggregate Phillips Curve estimates appear flattened. The findings imply that monetary policy transmission is inherently sector-specific and that uniform stabilization policies may produce uneven inflationary outcomes across industries.

This paper revisits the Phillips Curve by incorporating financial volatility, investor sentiment, and moral valuation shocks as structural cost-push components within the New Keynesian Phillips Curve framework. While conventional explanations for the flattening of the inflation–slack relationship emphasize expectation anchoring, globalization, or labor market changes, we argue that financial market instability and sentiment-driven moral re-pricing systematically alter firms’ marginal costs and pricing behavior. Financial volatility raises external financing premia, tightens balance sheet constraints, and amplifies precautionary pricing, while sentiment and moral shocks reshape demand elasticities and equilibrium markups. Embedding these channels into a hybrid NKPC highlights a previously underexplored transmission mechanism through which financial and ethical perceptions influence inflation dynamics. The framework reconciles empirical instability in Phillips Curve estimates with observed episodes of inflation persistence during periods of weak real activity, suggesting that modern inflation is increasingly driven by non-traditional cost pressures rather than labor market slack alone.

This study examines whether investor sentiment constitutes a priced systematic risk factor that spans traditional equity markets and cryptocurrency markets. Using monthly data for 42 equity markets and 30 major cryptocurrencies over the period 2012–2024, we construct a global sentiment factor based on Google Trends search intensity and news-based tone indicators. We integrate this sentiment factor into standard asset pricing frameworks, including CAPM and Fama–French-type multi-factor models, and estimate factor loadings using time-series regressions and Fama–MacBeth cross-sectional tests. The results indicate that sentiment exposure is positively priced across both asset classes, with higher expected returns required for assets that are more sensitive to adverse sentiment shocks. The pricing effect is economically stronger for cryptocurrencies and for equity portfolios characterized by high growth orientation and low institutional ownership. These findings suggest that investor sentiment operates as a common non-fundamental risk factor across markets with heterogeneous fundamentals, challenging the view that sentiment effects are confined to speculative niches.

This paper investigates whether the pricing of investor sentiment varies systematically across market regimes. Using monthly equity market data for 35 countries from 1995 to 2024, we identify bull, bear, and crisis states through a Markov-switching volatility framework. Investor sentiment is measured using a composite index combining news-based sentiment and search-engine attention indicators. We estimate regime-specific asset pricing models in which sentiment enters both as a direct pricing factor and through interaction terms with regime indicators. The empirical results reveal strong state dependence: sentiment risk is not priced during tranquil bull markets but becomes significantly priced during bear and crisis regimes, when limits to arbitrage intensify. Assets with high sentiment exposure earn higher expected returns exclusively in downturns, consistent with compensation for crash-state risk. These findings reconcile mixed evidence in the sentiment literature and demonstrate that sentiment premia emerge primarily as conditional rather than unconditional phenomena

This paper examines whether investor sentiment originating in cryptocurrency markets spills over into equity markets with overlapping investor bases. Using daily data for major cryptocurrencies and technology sector stock indices across the United States, Europe, and Asia from 2017 to 2024, we construct a crypto-specific sentiment index based on search behavior and social media tone. Spillover effects are analyzed using vector autoregression models and industry-level panel regressions that control for market-wide sentiment and macro-financial conditions. The results indicate that positive and negative sentiment shocks in crypto markets significantly affect subsequent returns and volatility of technology stocks, but not defensive or value-oriented sectors. The spillovers are asymmetric and stronger during high-volatility periods, suggesting that sentiment transmission operates through shared speculative investor demand rather than information fundamentals. These findings highlight cryptocurrencies as an independent source of behavioral shocks with measurable effects on equity pricing.