John Komlos has a new paper on this topic, here is the abstract:
Schumpeter’s concept of creative destruction as the engine of capitalist development is well-known. However, that the destructive part of creative destruction is a social and economic cost and therefore biases our estimate of the impact of the innovation on GDP is hardly acknowledged, with the notable exception of Witt (1996.“Innovations, Externalities and the Problem of Economic Progress.” Public Choice 89:113 –30). Admittedly, during the First and Second Industrial Revolutions the magnitude of the destructive component of innovation was no doubt small compared to the net value added to GDP. However, we conjecture that recently the destructive component of innovations has increased relative to the size of the creative component as the new technologies are often creating products which are close substitutes for the ones they replace whose value depreciates substantially in the process of destruction. Consequently, the contribution of recent innovations to GDP is likely upwardly biased. This note calls for further research in innovation economics in order to measure and decompose the effects of innovations into their creative and destructive components in order to provide improved estimates of their contribution to GDP and to employment.
Think of Uber being a relatively close substitute for taxicabs, for instance. Speculative, as they say, and the paper does not in fact actually demonstrate these conclusions, but at least we should be asking such questions more often.