Bryan Caplan  

Dr. Pritchett's Six Bitter Pills

What empirical evidence should... Henderson at Indiana Universit...
When Chris Blattman publicly asked, "As a development economist, what is one politically-incorrect research finding (a finding reflecting an unpopular belief) that you wish more people cared about?," Lant Pritchett swiftly rattled off six responses.  I stitch them together here, with Lant's permission.  Lant speaking:

1. One of course is that the income gains to movers from migration are an order of magnitude bigger than any in situ development project/program gain.

Another is that the variance of economic growth rates is much lower among democracies than non-democracies and hence (descriptively) nearly all episodes of rapid, sustained, (and hence poverty reducing) economic growth were initiated by non-democratic regimes (e.g. Indonesia, China, Korea, Taiwan, Vietnam).

Another is that on many individualized indicators of well-being (education, health, malnutrition, self-reported subjective well-being) the gaps between the sexes within poor countries are at least an order of magnitude smaller than the gaps between males in poor countries and females in rich (OECD) countries.

Another is the point Dani Rodrik has made (and Branko and me using Engle curves and food shares not income) that the rich in poor countries (e.g. 95th percentile) are much poorer than the poor in rich countries (say 20th percentile).

Another is that the high scoring students in poor countries on examinations like PISA (or equivalent) are much lower than even the average of the OECD--so while there are rich-poor gaps in the quality of education in poor countries this is not because the rich get a good education and the poor get a bad one but because the rich get a bad one and the poor get none at all (e.g. on a recent OECD study tertiary graduates in the capital city of Jakarta had lower measured literacy than high school drop-outs in Denmark, or when Indian states participated in PISA there were literally no students in the sample in the top two categories (5 and 6) of a six level scale of performance).

The reason I wish more people paid attention to the last three is that in my view in development today there is too much attention to inequality within poor countries and not enough to the very low levels such that things are pretty bad even for the, say, 80th percentile.

COMMENTS (13 to date)
robc writes:

Is #2 politically incorrect? Isnt it kind of obvious for at least 2 reasons?

1. Non-democracies have a lower baseline, hence they can have growth that isn't possible in a democracy.

2. Democracies had their rapid growth prior to being democracies, at least as we think of them today. England, for example, exploded well before the power of Kings was extinguished. It may have been trending democratic since Magna Carta, but it still wasn't really there, at least as of the lifetime of George III.

David Boaz writes:

About number four: really? Surely some of the rich in poor countries live in gated estates and drive Mercedes.

E. Harding writes:

"Another is that the high scoring students in poor countries on examinations like PISA (or equivalent) are much lower than even the average of the OECD--so while there are rich-poor gaps in the quality of education in poor countries this is not because the rich get a good education and the poor get a bad one but because the rich get a bad one and the poor get none at all"
-This is far from a hard and fast rule; Vietnam's test scores are comparable to those of the United States, even when adjusted for nonparticipation. South Korea's Baby Boomers score slightly higher than France's and substantially higher than those of most of Southern Europe on literacy tests, despite South Korea in the 1970s being poorer than Brazil, and having a poverty rate comparable to that of North Korea at the time.
Similarly, Qatar, the UAE, and Saudi Arabia are far from poor countries, yet their natives' test scores are substantially lower than those of any part of Europe.
The truth is more politically incorrect than what you've written.

@robc Well-established democracies mostly have never had rapid growth. Rather, they have had consistent growth for a long time.
Also, if one compares poor democracies to poor non-democracies, one finds much greater variation in the growth of poor non-democracies.

Shahid writes:

#4. I agree with David on this one. I live in a 'developing' country. The rich here enjoy a kind of lifestyle that even the upper middle income folks can only dream of. For example, they have a whole Army of servants, and the one's working on their lands may not even be paid in cash but in kind. Can the rich do that in developed countries? Not sure. That's why not sure about #4. Would appreciate if somebody could explain this one a bit more, and how this conclusion came about?

gwern writes:

David: '5th percentile' obviously does not cover the Mercedes-driving class; those plutocrats and corrupt officials would be more like 99th percentile. Consider the implications if it did: 5% of the population of, say, Ethiopia are driving around 5.2 million Mercedes? (That would be like 3 years of their total production, incidentally.) 2.8 million Mercedes are driving around South Africa? And so on.

RPLong writes:

Here's a link to one of Dr. Pritchett's paper re: #4:

If anecdotes are relevant here, I've spent a number of months in a rich household in a developing nation (although not the one they studied) and I think Pritchett & Spivack's findings are thoroughly consistent with my anecdotal observations.

Although, I admit that house servants tend to cost much more in the developed world.

PStu writes:

Under the condition that non-democracies have higher variance, isn't the flip side of sustained, rapid growth a pronounced economic decline? In Pritchett's last CGD paper he shows that the largest growth decelerations were in Iran (1976), Brazil (1980), Honduras (1979), and Cote d'Ivoire (1978).

David R Henderson writes:

Well done.

V. L writes:

The role that violence, especially organized collective violence that is sustained, plays in poorer countries is also worth considering. The impact of relatively large scale violence on a country is significant but the spillover effect can create a regional security problem AND become an obstacle to the national economic development of other regional countries. Since serious rates of economic growth are a basic requirement for a country to move beyond the range of conflict recidivism that is very likely, the failure to achieve needed GDP growth rates opens the door to repetition of violence and continued obstacles to growth. The loss of allocational efficiency due to actual conflict or preparation to contain conflict spillovers -- especially in these countries in which there are too few resources already poorly used -- is important. Not recognizing or failing to acknowledge the problems of violence, especially organized sustained violence, is a problem with long term impact. Not recognizing the complex, dynamic relationship between growth and violence us a serious analytical issue.

mm4 writes:

Where can one find more reading on #2, and what is the appropriate prescriptive interpretation of this data?

Phil Vernon writes:

#1-5 are all fairly uncontroversial. Whether or not one agrees with #6 depends on one's perspective: if global, then LP's view is understandable; from the political viewpoint of a poor person in a poor country, then inequality within that country is perhaps a more useful frame?

Blaise writes:

As a development economist, I on't find these findings so controversial. 6 is maybe the most controversial and goes against the narrative in the new SDG of Leave Noone Behind. I have a lot of reservations with this agenda but, in some contexts, it is a very useful lens of analysis. Resource allocation is often really biased in developing countries and targeting marginalized groups is both fair and efficient. I don't think there is a trade-off with focusing on global progress.

bill greene writes:

I submit Alexander Tytler's politically incorrect theory on why democracies rise and fall: Over a few hundred years the citizens go from being self-reliant, confident, and patriotic to being apathetic, dependent, and pessimistic. Thus, a nation's fortune is all about its people. New democracies grow from their energy and strength and eventually decline when the population regresses to a dependent and critical mass. This inevitable and fatal sequence may render all educational systems and economic management irrelevant.


Return to top