Working Paper

• Quasi-Bayesian Inference for Grouped Panels

  abstract pdf highlight slides

  This paper considers the problem of conducting inference in panel data models with latent group structures. I introduce a quasi-Bayesian framework that combines general classes of loss functions and priors for joint inference on the latent group structures, including group-level parameters and group assignments. Theoretically, I establish consistency of the proposed framework and derive posterior contraction rates for the quasi-Bayesian posterior distribution. Simulation results demonstrate significant improvements in bias and coverage for group-level parameters compared to existing methods, particularly when group assignments cannot be precisely estimated. Using the quasi-Bayesian clustering approach, I revisit the heterogeneous income risks of households and identify two previously undetected groups. The first experiences income increases in response to higher unemployment rates, and the second suffers substantial income losses despite being wealthy. These findings cast doubts on conventional shock amplification mechanisms in heterogeneous agent models where poor households are most vulnerable during recessions.

• Group Local Projections

  abstract pdf supp code highlight cite

  This paper considers the estimation of heterogeneous impulse responses in large panels. I introduce an efficient data-driven clustering methodology for grouping heterogeneous responses within the local projection-IV framework. The proposed group local projection (GLP) estimator consistently recovers the latent group structure and the group-specific impulse responses when the panel dimensions increase. Simulation evidence illustrates the reliable finite sample performance of the estimator even under misspecification of the group structure. With the GLP estimator I revisit the debate on the effects of monetary policy shocks on house prices and document significant price appreciation after a contractionary shock in an economically large cluster of MSAs in the US. Importantly, this cluster is ignored by conventional grouping criteria.

• Functional VAR

  Marcelo Reyes Award, 14th Workshop in Time Series Econometrics (Zaragoza, Spain, 2024)

  abstract pdf highlight

  This paper models the joint dynamics of macro aggregates and functional variables within the Structural VAR framework. I reduce the dimension of the system using functional PCA and show that the proposed functional VAR (FVAR) consistently recovers the responses of the functions. The FVAR is easy to implement and fully compatible with conventional SVAR tools. Simulation evidence shows that it performs satisfactorily in finite samples. Applying FVAR to study the impact of tax shocks on income distributions in the UK, I find that tax cuts persistently reduce the density of lower-middle-class households, which is offset by a substantial increase in the richer range and a moderate increase in the poorer range. However, this pattern is not captured by VARs with conventional inequality measures.

• Firm Hierarchy and Wage Rigidity, joint with Yameng Fan (UPF)

  abstract pdf highlight slides

  This paper presents new evidence on how the internal organization of firms shapes asymmetric wage risks over the business cycle, using matched employer-employee data in Germany from 1979 to 2010. We document three results. First, wage cyclicality is significantly more left-skewed for workers at lower hierarchical levels within firms. Second, there is substantial heterogeneity among low-ranked workers, depending on firm organizations. Third, the wage cyclicality for production workers becomes more left-skewed as the span of control for executives widens. Overall, the findings highlight the importance of firm organizations in driving the polarization of wage risks.