The Matthew effect: Evidence on firms' digitalization distributional effects.

This paper explores the relationship between Information and Communication Technologies (ICT) diffusion and labor productivity at the firm level. We propose a model that explains labor productivity based on firm-specific characteristics (sector, size, location and year fixed effects), on the level of diffusion of digital technologies, and on human capital. We use quantile regression techniques to gain a more precise understanding of the association between ICT's diffusion and labor productivity for a heterogeneous firm landscape. Our results show that more productive firms are able to extract higher gains from digitalization than low productive ones, which exacerbates productivity differences between them. This requires from a tailor-made policy agenda to promote firm convergence, focused on helping firms to extract the most of digitalization. • We explore the relationship between Information and Communication Technologies (ICT) adoption and labor productivity through firm-specific characteristics, the level of diffusion of digital technologies, and human capital. • The article focuses on how technological change can trigger economic transformations, making more productive firms to achieve higher economic gains from digital advances, in contrast to lagging firms, which are expected to be relegated far behind. • Our results also show how technological diffusion impacts are different across the labor productivity distribution. Differential impacts of the different technologies are widening the gap between less productive and more productive firms. • We can expect inequalities to increase in the future unless some policy actions are taken in advance to help less productive firms to make the most of digitalization. • Methodologically, we conduct estimates based on Unconditional Quantile Regressions to understand the heterogeneous effects that advanced digital technologies have on firm performance depending on their position in the productivity distribution. [ABSTRACT FROM AUTHOR]