Econometrics for Grumblers is the title of a recent paper in the Journal of Economic Surveys by Markus Eberhardt and Francis Teal (Working paper version). It is a survey of the cross-country economic growth literature with an emphasis on the econometric problems encountered. The biggest problem is that the state of technology or productivity is unobserved - the famous "residual" in the Solow model - and that:
1. The levels of the inputs - human capital and labor etc - which are usually explanatory variables in a growth model are affected by this unobserved variable. This results in a correlation between the error term in the model and the explanatory variables. The explanatory variables are then said to be endogenous.
2. The state of technology is non-stationary and probably has a stochastic trend.
3. The level of technology and rate of change are different in every country.
4. But the level of technology and rate of change are likely correlated across countries.
These aren't the only problems of course :) These issues have lead many to reject the usefulness of doing cross-country economic growth econometrics. Eberhardt and Teal review the literature thoroughly highlighting the underlying assumptions of each study and suggesting a new approach that they have used in their work.
I have been struggling with the issues in 1 to 4 above in a lot of my recent research, particularly on the environmental Kuznets curve and energy efficiency. I've used a variety of econometric techniques to try to deal with the issues but this gives me some new ideas to check out in more depth.
No comments:
Post a Comment