The purpose of (Aghion, Bergeaud, Boppart, Klenow, & Li, 2017) (hereafter ABBKL) is to identify a wedge between the true economic growth and the one measured in practice. The approach of the authors is to identify the true growth in a theoretical model with creative destruction and compare it to a model for the actual measure of growth (which misses the creative destruction effect on growth). From this comparison, ABBKL find that a sizeable growth component is missing: in practice, growth could be underestimated by 0.5 to 1 %.
The original paper is here
My comments are threefold
- The authors define as growth a utility based quantity and this is not what statistical offices intend to measure. I propose a correction to align the definition of growth with the national account’s concept. This correction diminishes the missing growth identified by ABBKL (possibly to 0) but the contribution of creative destruction to growth mismeasurement is unchanged.
- The authors base their comparison on CES aggregates while more realistic Laspeyres indexes can be computed. I revise ABBKL’s formula accordingly.
- The fact that missing growth is explained by creative destruction is related to the assumption on the imputation made by the statistical office. By assumption ABBKL leave little room for other channels to explain missing growth. I discuss this assumption and argue that ABBKL estimate an upper bound of the missing growth.
To properly account for the effect of correction 1 and suggestion 2 on missing growth, a reestimation of ABBKL’s indirect inference seems necessary. Nevertheless, the model developed is valid under general assumptions regarding the statistical office’s imputations. Therefore, it remains a great framework to understand why growth can be mismeasured.