The “persistence literature” is a strand of papers in economic history, which seek to find how big events in the past persist to this day. The seminal Acemoglu, Johnson and Robinson paper “The Colonial Origins of Comparative Development” in 2001 was the first and most influential, but the literature is also typified by Melissa Dell’s paper “The Persistent Effects of Peru’s Mining Mita”, Nathan Nunn’s “The Long Term Effects of Africa’s Slave Trades”, and Nunn, Alesina, and Giuliano’s paper “On the Origins of Gender Roles: Women and the Plough”. (There are numerous other papers; this one is a good guide to the literature). Their method is to find plausible exogenous variation in exposure to some event or system, and then compare the affected regions to today. Some of the papers are using the event as an independent variable to isolate the effects of institutions, but many are simply concerned with the effect of the event in and of itself.
These papers are heavily contested in terms of accuracy. The data is old and often speculative, the outcomes fuzzy and difficult to quantify. Demonstrating that it was not due to other factors is often impossible. However, fighting about this would be missing the point. Most of the literature is not even trying to find the effect of institutions through an IV of some kind. This is deeply misguided. Let us grant that we have very solidly shown that this or that historical event is responsible for the divergence in the outcomes we see today. What does this tell us about the future? Do we have any way to predict what events occurring now will persist into the future? The point of studying economic history should always be the guidance of future action. We study the great divergence so that we may understand why economic growth occurs, and how the underdeveloped countries in the world can do it today; finding out why the deep roots of how the plow led to gender norms today, or how pogroms in the 1500s affected Nazi vote share, tell us nothing. It is but a source of amusement. If you are going to write a paper investigating this, you should be able to identify how similar things in the present or recent past will lead to divergent outcomes in the future, or else the paper is trivial. It is very easy to go backwards, from what is now to what is past, and tell a hundred perfect little stories about how the past influenced the present day – it is hard, if not impossible, to go in reverse.
Those papers which are using the events as an IV to find the true impact of institutions are on sounder footing. AJR’s paper uses differences in the disease environment between countries to show the effects of institutions on development; places which had high settler mortality, they hypothesize, formed extractive institutions which wrung out as much as they could, and were ill equipped to grow in the future. Those with low settler mortality formed inclusive institutions, which were more concerned with the long term, and so grew more than their (often initially wealthier!) but more disease ridden neighbors. Think of the difference between the northeast of America, and of the malarial swamps of Virginia – the former an inclusive, highly democratic region, and the latter a repressive slave aristocracy.
There are problems with the data of AJR. Indeed it is almost certainly wrong! David Albouy’s replication in 2012 lays out the problems with the data; assuming countries had the same disease environment, unreliable data sources (such as mortality of soldiers while out on campaign), and that changes to the data to correct this will reverse the effect. Vollrath covers the reasons why the paper is in error; the measure of institutional quality is not linear (what does it mean for a country to be rated 5/10 or 10/10? Is it twice as good?), there’s no way to control for everything (disease environment makes a place worse to live in, not just through institutions), and their independent variable results are higher than their OLS results (which indicates that settler mortality has a negative effect on wealth on its own). Beyond simple challenges of factual accuracy, Glaeser et al. argues that they have it backwards; they find that growth is due to specific policies (which can be adopted in any environment), and that institutional policy improves as a consequence of growth. The measures of institutional quality do not point to any fundamental condition, but instead wildly alternate with elections. They point to, as an illustrative example, the case of South and North Korea, which were both dictatorships at the end of the Korean War. The south picked good policies, the North bad, and so the South later liberalized. They also find that human capital (the skills that people have) leads to growth and better institutions. Their telling of the difference between the north and south of America would simply be that they attracted different people, or that the different mortality risks lead to different rewards for investing in skills.
Notwithstanding such disputes, papers like “The Colonial Origins of Comparative Development” are at least worth discussing at all. The finding of persistence is simply a means to an end, not an end in itself. Those that do not try to explain anything beyond the influence of one event upon the present, are a triviality – a waste of time.
hi. you’re cool.
You might not have the necessary skills, but you perfectly illustrate the economist *attitude*: you took a look at the one subfield that gets any interest whatsoever from the general public, that goes beyond inside baseball whose importance is utterly impossible to communicate to anybody, and decided to write a post saying "this is trivial, this is dispensable, this is a waste of time."