Working papers

Abstract: This paper introduces a novel approach for estimating heterogeneous-agent macroeconomic models adding information from micro data. The methodology applies to both panels and repeated cross sections, with applications to a wide class of dynamic structural models used in macroeconomics. The routine involves the estimation of dynamic moments over subgroups of the cross-sectional dimension of agents. Micro moments differ from each other in the informative content that they carry for point estimation of the structural parameters. For instance, variability of moments over the cross-sectional distribution of households' wealth contain relevant information for the correct estimation of the subjective discount rate. However, data from the cross section are not relevant for the identification of a technology shock.
  • Estimation of continuous-time linear DSGE models from sampled measurements (joint with B.J. Christensen and J.C. Parra-Alvarez) [latest version - preliminary]

Abstract: We provide a framework to estimate continuous-time linear dynamic stochastic general equilibrium (DSGE) models when data is only available at discrete points in time. Any system of linear stochastic differential equations (SDE) has an equivalent discrete-time VAR(1) representation. We can therefore cast the continuous-time DSGE model as a discrete-time state space model (SSM). This accommodates for the discrete nature of the data, while keeping the fundamental characteristics of the continuous-time model unchanged. The discrete-time representation allows for standard maximum likelihood estimation of frequency-independent structural parameters of the economic model. One implication of the continuous-to-discrete-time mapping is that innovations in the state-space representation do not represent structural shocks in the economy, but the combination of all the shocks occurring between measurements. Following the structural VAR literature, we provide a framework to identify structural shocks from the reduced-form innovations of the estimated model.
  • Firm uncertainty and labor composition dynamics (joint with with E. H. Partsch) [latest version]

Abstract: We document the effects of uncertainty shocks on firm-level employment of high- and low-skilled labor. To investigate the potential effects of uncertainty on employment growth, we use that different industries are differentially exposed to a number of aggregate shocks. We use this fact to identify industry-specific uncertainty shocks. We show that while low-skilled labor growth is negatively affected by uncertainty shocks on impact, high-skill labor growth is not. Our dynamic approach shows that high-skill labor falls with a lag. Low-skilled labor shows similar dynamics, with the effect of uncertainty being strongest one year after impact. Our results highlight that the labor misallocation effects ascribed to uncertainty shocks seem to affect low-skilled labor most and that there is persistence in the effects. We contextualize our empirical findings within a heterogeneous firm model with high- and low-skill labor inputs and heterogenous labor adjustment costs.

Work in progress

  • Identification of structural shocks in a VAR augmented by factors (joint with F. Carlini and P. Gagliardini)

  • Estimation of weakly identified parameters with macroeconomic and financial data (with B. J. Christensen and M. van der Wel)


  • Essays on Estimation of Dynamic Macroeconomic Models (link)