Automation, economic growth, and the labor share - a comment on Prettner (2019)

Prettner (2019) studies the implications of automation for economic growth and the labor share in a variant of the Solow-Swan model. The aggregate production function allows for two types of capital, traditional and automation capital. Traditional capital and labor are imperfect substitutes whereas automation capital and labor are perfect substitutes. In this paper, we point to a flaw in Prettner's analysis that invalidates his main analytical and computational findings. In contrast to Prettner, we argue that both kinds of capital are perfect substitutes as stores of value, and, therefore, must earn the same rate of return in equilibrium. Our computational analysis shows that the model dramatically overestimates the actual decline in the US labor share over the last 50 years.