Eugenio González Flores

WORKING PAPERS

Borrowing to Do, Doing to Learn (JMP)

Draft available upon request

This paper examines how learning by doing interacts with financial frictions and explores the quantitative consequences of this interplay for firm growth and allocative efficiency. I build a heterogeneous-firm model in which productivity accumulates through input use and where firms face borrowing limits. I calibrate the model to match new estimates of the elasticity of future revenue-based productivity (TFPR) to current input use from Spanish and Portuguese firm-level data. From the calibrated model, I find that the elasticity of future physical productivity (TFP) to current input use is 0.09. I then show that output losses from financial frictions are 50% larger with learning by doing. This amplification arises because allocative efficiency losses are twice as large, and because financial frictions directly lower productivity in constrained firms by 4% on average. Furthermore, I show that a planner that ignores learning by doing can recover only 10% of productivity losses from misallocation in the long run and worsens allocative efficiency in the short run.

WORK IN PROGRESS

Becoming Informal: Selection or Displacement?

With Alejandro Castillo

This paper quantifies the share of selection and displacement into informality using a revealed preference approach and a structural search model. We use Mexican and Peruvian panel data to show that around 45% of transitions into informality are associated with utility gains, even though transitions into informality are, on average, associated with lower consumption and income. Then, we develop an equilibrium search model where workers can quit formal jobs, but formal firms can also fire formal workers. We estimate the model using empirical evidence as disciplining moments to estimate structural parameters. Finally, we use the estimated model to decompose transitions into informality: quits represent selection, while firings represent displacement.

Wedge-Accounting in Multi-Sector Models

With Paulo Lins and Gian Paulo Soave

Aggregate business-cycle accounting identifies distorted margins but not the sectors in which they arise. We develop a nine-sector model with input--output linkages and estimate it using annual U.S. data for 1963--2024. In leave-one-wedge-out counterfactuals, model GDP falls by 7.28 log points in 2007--09 and 5.33 in 2019--20. Removing the labor wedge attenuates those declines by 4.61 and 3.15 log points. Removing the supplied-good intermediate wedge attenuates the private-gross-output declines by 7.31 and 2.65 log points, the largest class comparisons for that outcome. At the sector level, trade generates the largest model-GDP comparison in 2007--09 (3.89 log points), while education, health, and other private services dominate 2019--20 (3.58 for GDP and 3.87 for labor). These signed, nonadditive comparisons show that model GDP and gross output rank distorted margins differently.

Page updated

Google Sites

Report abuse