NOTE: The Growth Economics Blog has moved sites. Click here to find this post at the new site.
The next installment of reading lists for my grad growth and development class this fall has to do with one explanation for the fundamental source of cross-country differences: institutions. “institutions” is a very nebulous concept, and so most of the recent, empirically sound, research on institutions focuses not on “institutions” in general but on some specific type of institution. Almost without fail, this leads to studies of former colonies. This is because they tend to have a very useful quality: a specific institution may have been thrust upon them, or changed arbitrarily.
Empirically, we want to find the causal effect of some specific institution on development (output per worker, wage levels, education levels, what have you). In general, what we call institutions are hopelessly endogenous, so we’re looking for natural experiments where an institution was dropped from the sky, so to speak, on top of some country. If so, we can possibly isolate the independent effect of that institution on development. Colonies make for a good test bed, as one can make a (perhaps) plausible argument that the colonizer exogenously altered, added, or subtracted some specific institution to the colony.
I try to avoid any papers that look at simple cross-country regressions of income per capita on some index of institutions, although I do talk through the Acemoglu, Johnson, and Robinson papers because of their prevalence across the literature. This is because of a point I raised before, which is that one can arbitrarily make an “institutions index” significant in a regression if you simply wiggle around the index values the right way. [I could code the U.S. a 1, Turkey a 0.5, and Zimbabwe a 0, or code them as 10,9, and 0, respectively. I preserve the ordering, but completely change the possible implications in a regression].
The reading list, which is posted under the “Papers” page on this site, thus focuses on recent within-country work related to specific institutions. One exception is the African work by Stelios Michalopoulos and Elias Papaioannou, which covers most of the continent, and tends to look at institutions as a generic concept. The interesting comparison here is the differing results: one paper finds that pre-colonial institutions do have persistent effects on development of ethnic groups, while the other finds that national institutions are not relevant. This just highlights that “institutions” taken as a whole are not necessarily a good predictor of development. Rather, one can find examples of specific institutions that matter.