Does varying rainfall make people collectivists?

Lewis Davis has a newly published paper on that topic with the more elegant title “Individual Responsibility and Economic Development: Evidence from Rainfall Data.”  Here is the abstract:

This paper estimates the effect of individual responsibility on economic development using an instrument derived from rainfall data. I argue that a taste for collective responsibility was adaptive in preindustrial societies that were exposed to high levels of agricultural risk, and that these attitudes continue to influence contemporary social norms and economic outcomes. The link between agricultural risk and collective responsibility is formalized in a model of optimal parental socialization effort. Empirically, I find a robust negative correlation between rainfall variation, a measure of exogenous agricultural risk, and a measure of individual responsibility. Using rainfall variation as an instrument, I find that individual responsibility has a large positive effect on economic development. The relationships between rainfall variation, individual responsibility and economic development are robust to the inclusion of variables related to climate and agricultural and institutional development.

This kind of investigation is always going to be fraught with uncertainty and also controversy, given imperfections of data and methods.  Nonetheless I find this one of the more plausible macro-historical hypotheses, perhaps because of my own experience in central Mexico, where varying rainfall still is the most important economic event of the year, though it is rapidly being supplanted by the variability of tourist demand for arts and crafts.  And yes, they are largely collectivist, at least at the clan level, with extensive systems of informal social insurance and very high implicit social marginal tax rates on accumulated wealth.

Have you noticed it rains a lot in England?

Here are earlier and ungated/less gated versions of the paper.


Guadalajara, Mexico has unusual weather: it rains like crazy in June, July but is dry most of the time. Or so it appears (I've visited but not stayed there) from Google.

Philippines is #1 for rain, the east side is mostly wet except for two months of the year, typically around Jan, Feb. And during the rainy season it will rain up to 18 inches a month or more, which is a lot.

'Empirically, I find a robust negative correlation between rainfall variation, a measure of exogenous agricultural risk, and a measure of individual responsibility.'

I wonder how thiw would apply to the southwest of the U.S.

'Using rainfall variation as an instrument, I find that individual responsibility has a large positive effect on economic development.'

So extinct cultures play no role in rainfall variation measurements, it appears.

'The relationships between rainfall variation, individual responsibility and economic development are robust to the inclusion of variables related to climate and agricultural and institutional development.'

So, the developers and beneficiaries of corn cultivation - how do they stack up? After all, the U.S. has been extensively reliant on corn based agriculture since essentially its founding. A bit of background for those who may not have watched those maize commercials, or wondered why its called Mazola -

I think you are reading the paper wrong (or I am). It's saying that in SW USA which has very variable rainfall (think of Arizona flash floods), the people are socialists, while in England, west Europe, or, most amazingly, the tri-state mid-Atlantic region of the USA (VA, MD, West VA) which has 'steady rain' year round (especially the DC area, which makes me laugh when I sometimes read the DC area has a 'water shortfall'--that can only be caused by mismanagement), the people are more 'libertarian'. To test this thesis, we can compare native Americans in the tri-state region of the mid-Atlantic (the Potomac tribe, etc) to the SW USA native Americans (the Pueblo, etc) and see if the latter are more 'socialist' than the former (they are). But, how to explain the Aztecs, Olmecs, Mayans however? Were they not socialist? Compare them to the high-and-wet Incas, were they not also socialist? Puzzles for a future data mining historian.

No idea who is reading this correctly or not, but the point about rainfall variability (see comment below) is extemely dependent on time scale and how it is measured. To the point that pretty much any argument can be made to support one's own position when applied to investigating such nebulous concepts as whether a 500 or 5000 year old culture that is only known through archeological remnants was 'collectivist.'

Not to mention that rainfall patterns could really change of 5K years,

Time scale seems like a red herring, here. The point is surely correct that rainfall variability can be highly dependent on the time scale, but the time scales we're interested here would be relatively short ones on the order of annual to 5-year variability. Longer than that and the concept of social insurance in a subsistence setting would be moot, anyway; longer droughts would result in everybody being dead or moving, Anasazi style.

The objection that it wouldn't be straightforward to classify the cultures in question, especially at a great remove, seems more germane.

The relevant rainfall conditions for American culture would be England, not SW USA. Because that's where a lot of the culture came from. And, anyways, by the time they were doing a lot of farming there, various insurance mechanisms were available to reduce the need for the forms of insurance through collectivist thinking or practices/habits/traditions.

'Have you noticed it rains a lot in England?'

Well, sort of. One of the things taught at the start of my GMU meteorology class was that there is more than one way to measure precipitation, and one needs to be aware of that when talking about either weather or climate or both. For example, SF and Richmond, Va. (at least back in the early 80s) had the same average precipitation totals. However, 'sunny' Richmond's contained a significant amount of heavy summer thunderstorm rain, while 'rainy' SF had more extensive periods of light rain. Here is a bit of information from wikipedia, by the way - 'London has a temperate oceanic climate (Köppen: Cfb ), similar to all of southern Britain. Despite its reputation as being a rainy city, London receives less precipitation (601 mm, 24 in, in a year) than Rome, Bordeaux, Toulouse, Naples, Sydney and New York.'

And really, does Egypt have more or less rainfall variability than Virginia or NJ or Massachusetts? Or could it be that other geographical factors than rainfall variability played a major role in the wold's first agricultural societies?

"Have you noticed it rains a lot in England?" Where I live we get "semi-arid" precipitation in some years e.g. in 2011 with 13.67 inches. How was 2011 for you?

This matches up with people wanting to share "lucky" high variance hunting kills, but not high effort gathering produce.

(Cato event on why socialism is appealing.)

I had the same thought. See

From my experience growing up in rural central Mexico.

--Tourist demand for arts and crafts is almost nonexisting for 95% percent of the towns and cities in Mexico, which have zero tourism.

--In a handful of places arts and crafts may be important, but still a relativelly small source of income.

--People are not particularly "collectivist". Where did you get that idea?

--I also do not see the "extensive systems of informal social insurance and very high implicit social marginal tax rates on accumulated wealth"

I am sorry, but I think you are making stuff up, or repeating some noble savage stories about collective live in some indigenous towns.

Please, you are just someone who knows what they are talking about. Besides, this web site has never been about facts anyways.

First, reread the definition of collectivism.

I was tempted to rant against Tyler, then I remembered the perennial 2nd place in presidential elections in Mexico always talk about groups, individuals do not exist.

@ramon: Mexicans are also tourists. If I remember well, domestic tourism is larger than foreign.

That´s why 5% have some form of tourism (mostly national). Foreign tourists visit only a handful of places.

"very high implicit social marginal tax rates on accumulated wealth." translated to popular culture means that if you're the "rich" in your family you're padrino(godfather) for every party in town: weddings, baptisms, local(s) saint(s) festivities, etc. In small towns, having some money, or some cattle, and refusing to make a large contribution to religious festivity is the equivalent of Ebenezer Scrooge cursing Christmas. In urban areas, individualism is just fine.

Thanks for translating this into English.

So maybe this explains why African cities are hubs of consumption by the rich, per a recent Economist article.

That´s the noble savage stories I was talking about.

Even in rural áreas, weddings, baptisms, etc. are paid mostly by the families involved, with some contribution by a padrino, a close relative not necessarily the richest relative available (money is not the only variable here).

Public festivities are paid mostly by the government. Municipal governments pay musicians, fireworks, and all sorts of preparations.

Washington DC has higher rainfall than London, maybe that`s why it is full of Communists:

look this is all a pretty pointless discussion, but at least to be clear the thing we're talking about is variability of rainfall, not the total amount of rain. i.e. a useful metric might be average number of days between rainy days.

And it's not exactly variability of rainfall. It's variability of sufficient rain to effect local crop growth and output.

Curse those hothouse growing individualists, selfishly trying to avoid the weather.

Jeffersonian America, out to about the 100th Meridian on the Great Plains, was a place where rainfall was both abundant and fairly regular. So, a family could pioneer their own farm with no need for large scale water infrastructure or intrusive regulations on water use. Clear the forest, plow the soil, plants some seeds, and the rain that falls from the sky would be enough.

Here in California, however, the laissez-faire common law water law has proven troublesome, however.

I'd say it's also the milking operation scale. A farmer could milk 30-50 cows, 100 with some extra help outside the family. Neighboring farmers with a similar low cows/farm area ratio. Then, industrial milking comes and there are 10K cows in an area for 1000 in the previous system. Result: local water more intensively used, need to bring water from neighboring valley, etc.

With "rainfall variation", we are talking about variation within the year (with strong differences between "dry" and "wet" seasons), or variation between years? If we are talking about high risk, I imagine that what is more relevant is the variation between years.

Economists can be so amusing. At what point does the increasing stack of papers, each claiming rainfall is a great instrument for its own purported causal relationship, suggest that maybe it's not as good at meeting the exclusion restriction as each author claims (hopes)?

Solve for the collective action problem, as Tyler might say.

Exactly my thoughts when I saw the post title.

Rainfall and water management are hugely relevant in understanding various things about managing collective issues.

For example, the first female Economics Nobel winner did lots of work on systems to deal with water-related issues.

The first futures market was created in response to varying rainfall and supplies of agricultural goods (rice) - the futures market smoothed both supply and price. That's collectivist in a sense; indeed, financial markets generally, and all markets, are collectivist in a sense; economic development requires cooperation of individual actors. It's unfortunate that "collectivist" has taken on such negative connotations, when it has taken humanity from a primitive existence of individual actors to highly developed and economically advanced societies. Many are so enamored of individual actors they cannot see that progress is dependent on the collectivist.

I've moved from California to the Netherlands. In California, annual variation has been reduced with massive government projects. People therefore fight more than cooperate over the water.

The Dutch face constant rainfall as well as river flows TO THE POINT OF WORRYING ABOUT FLOODS, therefore this "model" would need to include more cooperation from variation in runoff (floods).

Yes, it's a stretch to say rainfall variability = agriculture yield variability due to weather, I'd also look at frosts & hail. But, a model has to be as simple as possible to understand what the model does. If with 1 variable there's a good correlation (rainfall variability), a more complete model (floods, frosts, hail) would give a better result.

The more complete model is not in the article, thus my inner skeptic ponders about what happened: a) the more complete model it's a follow-up article, 2 publications are better than 1, b) it's already a lot of work to have this result, reader patience is needed, c) the more complete model contradicts the simple model results, why spoil a good publication?. Until the more complete model results are published, the question remains open.

Ps. I liked reading Living with Water Scarcity. It's a good and simple introduction for people that know nothing about the topic.

It's easier to model historical rainfall than historical hail intensity. With irrigation and the ability to mine water tables, hail may be even more relevant than water in some places, but that did not apply in the historical situation.

Rainfall / water and its effect on the human mindset is a key theme in the bible, particularly the Exodus. Egypt, in Hebrew, is called "Mitzrayim" which translates as "narrow" and refers to the physical geography of the country clustering around the narrow stretch of the Nile (for agricultural reasons), and the narrow-mindedness that comes from dependence on the Nile for irrigation, rather than rain fall. The regularity of it was one of the key factors shaping Egyptian culture.

In contrast, agriculture in Israel was historically dependent on variable rainfall. The mindset that creates is a consciousness of omnipresent risk and dependence on God.

Connect this with the concept of "Oriental despotism," also called hydraulic despotism. Lands that need massive irrigation, such as the Middle East, turn autocratic, but a land such as Greece that depends on consistent rainfall turns individualistic.

There is huge variation in systems across all types of conditions. We can learn from many of them, but generalizations, not so much so. Places with high need to manage rainfall and irrigation water issues may also be highly relevant in the development of individual property rights systems in contexts which require degrees of collective resource management.

Optimal foraging theory: By the way, many papers have been published on food availability in the animal behavior and the field anthropology literatures -- some useful search terms include optimal foraging theory, cooperative hunting, and patchiness. As an example of a recent paper (and bibliography), .. (Marco A. Janssen and Kim Hill 2014 in Human Ecology) (Though as far as I'm aware these studies only look at very small samples of populations.)

If I say "Texas," are you going to say "outlier"? Or do I get to hear the sound of a hypothesis crumbling?

" they are largely collectivist, at least at the clan level, with extensive systems of informal social insurance"

In other words, people look out for their friends and relatives, a situation regarded as normal until the nation/state arrived with progressive programs to take the place of interpersonal relationships.

The number of papers using rainfall as an instrument only goes to show why it shouldn't be used as an instrument. If it manages to correlate with all sorts of phenomenon that have political and economic effects then the exclusion restriction is probably violated in nearly every one of these studies! (For the laymen, the researchers are trying to show a causal relationship and not just a correlation. It all depends on whether rainfall affects the independent variable they're interested in, which in this case is agricultural conditions, BUT it must do so without affecting the outcome variable(s) through any other channels besides its effect on agricultural conditions. This last part is the exclusion restriction. If rainfall affects the outcome variables through any other pathways, then the whole analysis is junk. It can't tell you that agricultural conditions have an effect on the outcome variables. It may just be correlation and not causation after all. Sad.)

Or ... maybe it's actually relevant, and that's why so many people use it?

Rain and sun affect just about everything, some more so than others.

My pet theory is that disease vector transmission via currency and along trade routes engenders disgust feelings about money and markets, which is what is responsible for persistent anti-capitalist sentiment in most countries. Also noteworthy that the Black Death in Europe occurred relatively recently, and it spread along the trade routes from Asia.

Before the Black Death there was the 6th century Plague of Justinian, which may have been even deadlier, and which did play a large role in the collapse of late Roman civilization. For sure pandemics make people suspicious of outsiders since disease is generally brought to a locality by someone from outside. I'm not sure it makes people suspicious of financial activities specifically. The Black Death after all was followed by the Renaissance with the birth of modern capitalism.

When was paper currency introduced into Europe?

Per Wikipedia, the first European bank notes were issued in Sweden in 1661.

Karl Wittfogel described this phenomenon as the basis of the "Hydraulic State" in his 1957 book, "Oriental Despotism." Wikipedia's synopsis of his argument follows:

"Wittfogel argues that climate caused some parts of the world to develop higher levels of civilization than others. He is known for claiming that climate in the Orient led to despotic rule. This environmental determinism comes to bear when considering that in those societies where the most control was exhibited, this was commonly the case due to the central role of the resource in economic processes and its environmentally limited, or constrained nature. This made controlling supply and demand easier and allowed a more complete monopoly to be established, as well as preventing the use of alternative resources to compensate. However, Diamond points out that complex irrigation projects predated states in Madagascar, Mexico, China and Mesopotamia.

"The typical hydraulic empire government, in Wittfogel's thesis, is extremely centralized, with no trace of an independent aristocracy – in contrast to the decentralized feudalism of medieval Europe. Though tribal societies had structures that were usually personal in nature, exercised by a patriarch over a tribal group related by various degrees of kinship, hydraulic hierarchies gave rise to the established permanent institution of impersonal government. Popular revolution in such a state was impossible: a dynasty might die out or be overthrown by force, but the new regime would differ very little from the old one. Hydraulic empires were only ever destroyed by foreign conquerors."

Sometimes hydraulic societies were destroyed by major natural disasters. That is what seems to have brought down the Bronze Age Indus Valley civilization.

Food surplus feeds soldiers, and this is probably the most relevant interaction between irrigation in ancient empires and the fact of there being an empire. I don't think the monopoly aspect is relevant. Dissidents would have been addressed by means other than restricting their water (which might have been technically difficult in many early systems).

it rains a lot in the American South too...and I think there's a preference for gun collecting over collectivization.

And I'm thinking Dutch dikes..
“God created the earth, but the Dutch created the Netherlands.”

It doesn't rain that much in England. It rains far more in Ireland.

I have noticed that there seems to be some correlation between colder climates and peace/stability, but I doubt that there's a causal element there.

Comments for this post are closed