Daron Acemoglu, Simon Johnson, and James Robinson were recently awarded a Nobel Prize in Economics, partly for work emphasizing the role of good institutions in economic development. They used a variety of creative statistical techniques to get around the thorny problem of how to establish causality.
One problem in establishing causality is that a group of successful countries might share a number of distinctive characteristics. If most successful countries share characteristics X, Y and Z, then it is difficult to know whether any one of those characteristics played a decisive role in development.
Normally, we tend to assume that a large number of observations provides more “statistical significance” than does a single anecdote. But is that always true?
Consider the following claim:
Differences in institutions, differences in culture, and differences in natural resource endowments all play a major role in explaining the relative wealth of nations.
I believe that these three claims are true. I also believe that the strongest evidence in favor of these three claims does not come from reams of statistical analysis, rather it comes from three very powerful anecdotes.
Good institutions: Korea provides a near perfect example of the importance of having good institutions, which is a term economists use to represent an economic and legal system conducive to wealth creation. When Korea was divided after WWII, there was relatively little difference between either the culture or the resource endowments of the two halves of the country. If anything, the North was somewhat richer.
Today, South Korea is vastly richer than the North. It’s almost impossible to think of a plausible explanation that does not in one way or another reflect differences in institutions. Yes, Korea is just a single data point. But it is overwhelmingly persuasive, in some ways more so than a regression of 200 nations that tries to establish the importance of good institutions.
If you are the sort of person who likes a simple explanation for complex phenomena, you might wish to stop here. Korea seems to clearly demonstrate the overwhelming importance of institutions. Unfortunately, further south there’s another powerful anecdote, with a very different implication.
Good culture: People who emphasize the role of culture in economic development, often point to the impact of Chinese diaspora on the economies of Southeast Asia. Singapore (76% Chinese) is much richer than Malaysia (23% Chinese), which is significantly richer than Thailand (14% Chinese), which is significantly richer than Indonesia (1% Chinese.)
But is that information actually decisive? Perhaps Singapore is rich due to the policies of Lee Kuan Yew. Malaysia has more natural resources than Thailand. Thailand was not exploited by colonial powers in the way that Indonesia was. One can always dream up all sorts of ad hoc explanations for any given correlation.
I’m actually much more persuaded by a single anecdote: Malaysia. The 23% of Malaysia’s population that are Chinese are much richer than the remaining 77% of the population (mostly ethnic Malays, but also some Indians.) Does that prove that culture matters? Perhaps the Malays are being discriminated against. In fact, it is the Chinese that are being discriminated against, as Malaysia has a strong affirmative action policy favoring the Malays. Thus it’s very hard to think of any explanation for the relative success of Chinese Malaysians that does not involve some aspect of culture (which is admittedly a broad category.)
As an aside, I’m perplexed when progressives make the argument that economic disparity is prima facie evidence of discrimination. Jewish Americans have considerably higher average incomes than Christian Americans (even white Christian Americans). Do progressives actually wish to argue that this disparity reflects the fact that America discriminates against Christians? How would that differ from offensive arguments made by anti-Semites?
Natural Resources: When I was young, the north coast of South America had three little colonies called British Guiana, Dutch Guiana and French Guiana. Today, British Guiana is just Guyana, Dutch Guiana is Suriname, and French Guiana is a part of France. Even five years ago, both Guyana and Suriname were fairly poor.
Today (in 2024), Suriname has a per capita GDP of $6700, while Guyana has shot up to over $26,000 (and will be two or three times higher by the end of the decade.) There is simply no plausible explanation for Guyana’s success that does not reference Exxon’s recent discovery of large oil deposits off the country’s north coast.
To summarize: From Korea, we know that good institutions are really, really important. From Malaysia we know that culture is really, really important. And from Guyana we know that natural resource endowments can be very, very important. Lots of things are very, very important!
I worry that people often have trouble holding multiple explanations in their mind at the same time. Consider that right next to Guyana you have poverty-stricken Venezuela. But Venezuela is even more oil rich than Guyana (although that’s partly offset by its larger population.). A person obsessed with monocausal explanations might look at Venezuela and conclude that natural resources were a curse.
A better way to approach the problem is to note that more than one factor is very, very important. From Guyana, we know that natural resources can be very conducive to growth. But from North Korea we know that bad institutions can be extremely negative for growth. Venezuela has ended up with really bad institutions in a country with large oil reserves.
Here’s how I approach the tricky problem of establishing causality in economic development. Start with the single best anecdote, the single example that seems to be the cleanest test. Then look for confirmation from slightly less clean examples.
After the two Koreas, you might look to East and West Germany prior to 1989. Or to Taiwan and Mainland China during the Mao era.
After Malaysia, you can look at Chinese minorities in other Asian countries, or Indian immigrants to Africa, or Jewish immigrants to America. Thomas Sowell has documented many more such examples showing the importance of culture.
After Guyana you can look at other oil-rich low population places such as the UAE and Kuwait.
If the single most powerful anecdote is mostly confirmed by the next best examples, then you can have even more confidence that it wasn’t just some sort of fluke. You know that you are on the right track.
But now the hard part starts. We have general sense that extreme communism in North Korea hurt their growth, but which specific aspects of the system were most damaging? That’s a much harder question. You might look to other cases that share some characteristics of North Korean communism, but not all.
We have a general sense that Chinese culture is favorable for growth, but which specific aspects? We might look at other Confucian cultures in East Asia. Or we might look at other “middleman minorities” in various parts of the work.
We have a general sense that oil and gas can be very important, but what other natural resources seem to matter?
One piece at a time, we gradually fill in the jigsaw puzzle, and a coherent picture starts to emerge. There are no short cuts, no easy answers, no monocausal explanations.
READER COMMENTS
TGGP
Oct 16 2024 at 9:03pm
I normally downplay anecdotes, but I have to give credit that this is a good post. Multiple factors are relevant, but people tend to just focus on one when making arguments. I have recently been irritated by Richard Hanania mocking culture as an explanation, and the perfect version of this post might contain arguments against him, but for most audiences this gets the point across.
On the subject of Sowell and “middlemen minorities”, since 2007 I have hosted his “Are Jews Generic?” essay on my blog.
Scott Sumner
Oct 17 2024 at 4:13am
Yes, I was greatly influenced by Sowell’s work in this area.
Ahmed Fares
Oct 16 2024 at 11:24pm
Lots of reasons for divergence in this article by Noah Smith:
Haiti vs. the Dominican Republic
Scott Sumner
Oct 17 2024 at 4:09am
Good example. Zimbabwe/Botswana is another case.
Mactoul
Oct 17 2024 at 12:32am
Are good institutions defined in the same way by Austrians on one extreme and Keynesians and other left-wing economists on the other extreme?
Some economists might regard a strong tax-collecting institute as good but an economist of libertarian bend might regard it is bad institute.
Scott Sumner
Oct 17 2024 at 4:07am
There might be some disagreement among economists, but they’d certainly agree far more than they’d disagree as to what are good institutions. Rule of law, competitive markets, simple tax system, etc.
steve
Oct 17 2024 at 12:40am
I thought AJR actually claimed that culture is important too. Am I misremembering? Also, I think Venezuela shows that natural resources alone do not lead to prosperity. Good institutions and culture seem necessary to make that work. Other examples are Norway and Saudi Arabia. The latter has had incredible amounts of oil wealth but it’s still not actually prosperous. Norway has both oil and good institutions and it is wealthy.
Scott Sumner
Oct 17 2024 at 4:05am
I think Norway has better institutions than Saudi Arabia, but even the Saudis are extremely prosperous relative to where they’d be without oil.
And yes, I believe that AJR also discussed culture, not just good institutions.
Rajat
Oct 17 2024 at 6:40pm
One factor you don’t mention is IQ, and the tendency for average IQ to vary by race (or ethnicity, if you prefer). There may be cultural and socio-economic reasons for those differences in IQ, but whatever it is, there does seem to be some relationship between measured average IQ, race, and economic success, other things being equal. For example, the economic success of Jews, Asians and high-caste Indians in America.
john hare
Oct 17 2024 at 7:14pm
One problem is that even mentioning your thought gets one branded racist immediately. IMO effective IQ does vary by race within a given area. Selection bias and culture of various races or ethnicities.
Jim Glass
Oct 18 2024 at 2:15am
A very excellent article. Thanks.
Here are some more anecdotes that come to mind…
[] Institutions – Why Nations Fail points to the town of Nagales, Arizona/Sonora, which has the US-Mexican border run right through it – and you really want to live on the US side.
[] Culture, here in the USA — Asians are famously successful immigrants, but Asia is a big place with many cultures. Just checking the numbers, Indians here earn near 3x and Filipinos 2x as much as Nepalese and Burmese … Back at the turn of the 20th Century, when poor immigrants were making Ellis Island famous, poor Jewish children graduated from NYC schools at a higher rate than even children of the “ruling” white Protestants, while Irish and Italians had a near 0% graduation rate. The Jewish immigrants greatly valued learning, while the Irish and Italians had learned to distrust government and its schools. This greatly affected how the groups assimilated … Black immigrants to the US from Africa and the Caribbean have a long history of significantly out-earning native American blacks. (The worst racism I ever in my life heard with my own ears was back in the 1990s, when welfare reform was a big political issue, Caribbean blacks going after American “welfare” blacks.)
[] Resources — Lack of resources hasn’t kept countries with good institutions from prospering greatly: Britain, Japan, the tiny Dutch state and its world economic empire in the 17th Century. But the “resources curse” is a thing too, allowing extractive authoritarian regimes to live off them for as long as they last. (Blood diamonds and Russia.) Perhaps resources and the quality of institutions have a multiplier effect on each other?
[] A fourth factor economists never mention but geo-politicians often do is the “land state” vs “sea state” difference. Land states exist throughout their history with violent mortal threats on their borders (Mearsheimer explains). So they tend to have extractive, authoritarian military regimes with zero-sum policies, taking land from others for security, ‘just win the inevitable next war’.
Sea states are physically protected by water (more or less) and profit by trade, which requires ‘fair rules’, so they tend to rules-based orders and institutions. Trade-by-water is positive sum and can scale up fast, so they can get very rich. It’s not coincidence that Athens invented democracy and Britain and the Dutch led in advancing the liberal orders. (Russia is the par excellence land state, with no natural borders and so fighting endless wars of over-expansion to gain security against enemies dating back to the Mongols. The West vs Russia & China can be seen as “land powers v sea powers” conflict.)
Jim Glass
Oct 18 2024 at 2:52am
BTW, the USA hit the grand slam of these four factors:
Superior institutions … Superior culture, near entirely populated by immigrants who traveled across an ocean to a land where most of them didn’t speak the language, all with the goal of improving themselves … Superior endowment of natural resources … Protected by and with full access to an entire ocean on each side, in the 20th Century it became the best-positioned sea state ever. (In the 19th it acted mostly like a land state conquering its way to the western sea.)
nobody.really
Oct 18 2024 at 10:46pm
For another example, Palm Springs, CA, is divided in a checkboard pattern between land subject to convetional US property rights (albeit the California version of these rights) and land subject to Native American law. For a long time, the disparities between these lands were quite stark: Lands subject to conventional law sprouted lots of development–say, fancy resort complexes–while across the street the land subject to native law would be a gravel parking lot. As CA real estate has grown more expensive, as tribes have grown wealthier, and perhaps as native property rights have become more developed, the native lands have started showing more growth.
nobody.really
Oct 18 2024 at 1:41pm
Two points of definition.
I use the term prima facie to mean “at first glance”—basically, evidence supporting a working hypothesis. When I observe correlation between two variables, this causes me to look for some causal relationship between them. (And yes, I enjoy Spurious Correlations, too.)
I use the term discriminate to mean “choose (based on criteria or taste).”
Thus, I find the fact that Jewish Americans have unusually high incomes as prima facie evidence of discrimination. And behold, Jewish Americans appear to have an unusually high amount of education. Employers who regard education as a criterion upon which to hire may well discriminate in favor of Jewish Americans. QED.
Of course, this prompts the question, why do Jewish Americans obtain an unusually high amount of education? This can lead to other causal factors (a religious admonition that young men learn to read the Torah; history of oppression leading people to invest in capital that can’t be easily appropriated—e,g., human capital; etc.) The answer to these questions will prompt still more.
Bottom line: I find no problem with developing working hypotheses—even when those hypotheses conflict with prevailing dogmas—so long as the hypotheses get subjected to testing. (We kinda had a discussion along those lines in the comments here.) Admittedly, people will disagree about the appropriate degree of testing, with some advocating the maxim “Extraordinary claims require extraordinary proof” while others reject this and argue that all claims should be subject to the same degree of rigor. We’ll save that discussion for another post.