Automation and Functional Production: What are the effects of automation on functional upgrading in 40 advanced and emerging economies between 1996 and 2011?

Abstract

Se investiga los efectos de la adopción de robots en el ascenso en la cadena de suministros rastreando cambios en las fracciones de empleo intensivas en conocimiento en 40 economías avanzadas y emergentes entre 1996 y 2011. Se construye un modelo económico de automatización de la fabricación basado en tareas para determinar los mecanismos y la dirección del reemplazo del trabajo por capital en las funciones de fabricación (un estadio de la producción). Se contrasta el modelo con la evidencia empírica sobre adopción de robots industriales y se derivan hipótesis testeables. Se intenta identificar efectos causales empleando un diseño de variable instrumental shift-share para disminuir problemas de endogeneidad. La investigación tiene contribuciones teóricas, metodológicas y tres hallazgos empíricos sore los efectos de la adopción de robots. Primero, en economías emergentes, está asociada con reorganización de la producción: incrementa la fracción del empleo dedicada a la fabricación y reduce la fracción dedicada al conocimiento intensivo. Segundo, incrementa la profundización de la división internacional del trabajo en lugar de ayudar a la convergencia en actividades intensivas en conocimiento. Tercero, se asocia con una dislocación entre la fuerza doméstica de trabajo en ingeniería (intensiva en conocimiento) y las ganancias en productividad.

This dissertation attempts to shed light into the mechanisms and effects of automation in functional upgrading. More precisely, it investigates the effects of robot adoption in functional upgrading and changes in the knowledge shares of employment in 40 advanced and emerging economies during 1996 and 2011. In doing so, it bridges the literatures of business functions and automation. It starts by establishing that business functions are groups of tasks accomplished by firms, and that they are performed by specific occupations. Countries with more knowledge-intensive occupations can perform more knowledge-intensive functions. Knowledge intensive functions (design, R&D, management) have several advantages over fabrication (the stage of production when goods are physically transformed). It proceeds by building a task-based model of fabrication automation. By uncovering the mechanisms at play when automation technologies replace labour with capital in the fabrication process, it shows 3 counteractive effects and derives 2 implications: automation is technology-specific and ‘country-specific’. It continues by linking the model implications to the rich findings of the robot adoption studies, to derive testable hypotheses. It takes several steps to make progress in the identification of causal effects, including the design of a shift-share instrumental variable design to alleviate endogeneity issues. Beside its theoretical and methodological contributions, this dissertation has 3 empirical findings: reorganisation, deepening and dislocation. Robot adoption is associated with reorganisation of production in a counterintuitive way: increasing the fabrication share and decreasing the knowledge shares in emerging economies. Because its differential effect detrimental to emerging economies, it contributes to deepening the division of labour working against convergence to knowledge-intensive activities. Finally, it shows that productivity gains from automation are dislocated from the domestic engineering workforce, which is consistent with a pattern of ‘global-connectedness and local-disconnectedness’ and lack of incentives to build domestic knowledge-intensive complementarities.