Fragile Robots, Economic Growth and Convergence

Technological progress leads to the development of robots that are more error-prone and fragile than their predecessors. As a consequence, the utilization of the existing automation capital stock is associated with higher wear and tear, CPU overload or communication downtime and, as a consequence, an increase of depreciation costs. This in turn affect new investments in the future. Considering a growth model with physical and automation capital utilization, we argue that in a fully automated society, the utilized automation capital is a perfect substitute for labor, not the automation capital stock per se. We show that it is not necessarily the introduction of capital utilization by itself, but the relationship between the elasticities of utilization of automation and physical capital that plays a crucial role in slowing down the convergence speed in a model that reflects an automated society.