Reconsidering Non-Neutral Technical Change

Jaap Bos' website at Maastricht University

Abstract: These days, understanding the drivers of changes in inequality is key to effective policy making. One of those drivers can be non-neutral technical change, especially when the latter favors high-skilled labor over low(er)-skilled labor. Technical change can be said to be especially biased if it is wrongly measured, in which case a presumed equality between a marginal rate of technical substitution and the relative price of inputs actually reflects an overpayment of one input compared to at least one other input, thus contributing to a rise in inequality that is not reflected in changes in productivity.
In this paper, we tackle this issue. We start from the premise that old units of input – through experience – are more efficient than new units. We then show, first in a simple two-firm setting, and then in a more generalized model, how this can result in a bias in the measurement of non-neutral technical change. Subsequent simulations demonstrate that this bias is significant, both statistically and economically. When we take our approach to industry data covering 40 countries over the period 1995-2009, we indeed find evidence that non-neutral technical change is biased in line with our theoretical model. These findings allow us to assess to what extent this bias has contributed to rising inequality over the sample period, as well as the conditions that made this bias particularly strong in certain industries.