A0527
Title: Revisiting localized technical change: Insights from a mixed-type endogenous partially linear framework
Authors: Davide Golinelli - Università degli Studi di Ferrara (Italy) [presenting]
Antonio Musolesi - University of Ferrara (Italy)
Alexandra Soberon - Universidad de Cantabria (Spain)
Abstract: The aim is to revisit the concept of localized technical change (LTC) by analyzing the heterogeneous effects of innovation on firms' production technology within a mixed-type endogenous partially linear framework. Traditional models often assume Hicks-neutral technical change, implying uniform productivity gains across all factor proportions. Challenging this view, it explores how technical change can be localized to specific techniques, as suggested by a past study. A three-stage estimation procedure is proposed to address endogeneity in a partially linear model with mixed-type (continuous and discrete) endogenous variables. The methodology combines nonparametric techniques for flexibility with parametric components for interpretability, using instrumental variables to reduce bias. Monte Carlo simulations show the estimator's robustness across different levels of endogeneity and sample sizes. An empirical application with firm-level data from the Italian manufacturing sector incorporates innovation as a binary variable to test the model's practical value. By relaxing strict functional form assumptions, the model more accurately reflects production processes. The contribution is to the econometrics of productivity and innovation, offering a flexible tool for analyzing causal effects in complex, mixed-type settings.
