Learning to Compete and Cooperate

What drives individualism and competitiveness as opposed to collectivism and cooperation? Leibbrandt, Gneezy and List have a great paper studying this question with an ingenious List2experiment. LGL study two types of fishermen in Northeastern Brazil. The two types live within ~50km of one another but one type are lake fishermen and the other sea fishermen. Lake fishing favors individual fisherman in small boats while sea fishing favors team production on larger boats.

LGL ask the fisherman to participate in a simple experiment, throw 10 tennis balls into a bucket. The participants choose how they are paid, 1 monetary unit per successful attempt or 3 units per successful attempt if they have more successes than an unknown competitor (chosen randomly and without their knowledge to avoid social effects; in case of a tie they are paid 1 unit per success). Fishermen could earn 1-2 days of income for less than an hour of work depending on how successful they were and the payment scheme chosen.

Perhaps you won’t be too surprised to learn that 45.6% of the lake fishermen chose to compete compared with just 27.6% of the sea fishermen. What makes the paper great is all the secondary tests the authors do to understand this result at a deep level. The result, for example, is not due to differences in throwing ability or risk preferences.

You might suspect that the different choices about whether to compete or not are driven by cultural differences. But that too is incorrect. The authors, for example, show that women–who do not fish in either the sea or lake villages–do not show differences in the choices to compete (both chose to compete less than the men but at the same rates in lake or sea villages).

List 3Instead, what the authors demonstrate is that differences in the choice to compete or not appear to be learned differences. First, the lake villagers who chose to compete are among the most successful lake fishermen–that is, they have learned that competition increases income. In the sea villages there is no correlation between choosing to compete and fishing income.

Finally, and most tellingly, there is a dose-response relationship between competition and learning. In particular, the choice to compete or not increases with fishing experience with the experienced lake fisherman choosing to compete more and the experienced sea fishermen choosing to compete less (as shown at left).

The paper appears on the surface to be affirming the importance of cultural differences and to be agreeing with the kind of literature that stresses the idea of self-interest and individualism as western and contingent. Yet, in fact, the paper is suggesting that at a deeper level so-called cultural differences may not be transmitted down through the generations but instead are learned responses to very particular production techniques. Note that such learned responses may change rapidly as production techniques change and that the sea and lake villages are both unusual in the modern world in relying on just one dominant production technique with few other options for learning.

More generally, learning needs to be added to incentives, genetics, and culture as an independent yet entangled determinant of choice.


you're getting all of this from throwing tennis balls in a bucket?

...yup, sounds fishy. Full study is walled off, but seems very soft -- non-random sample & small sample size, no control group, etc.

In today's world where results of most published behavioral studies can not be replicated -- this study is just noise.

Every popular study is 1 part observation, 99 parts conjecture.

Sometimes I wish journals decided not to publish such studies until the authors studied 5 such lake-sea pairs and demonstrated their correlation holds.

It's too easy to be swayed by spurious correlations that lack external validity.

Actually the study studies several lake-sea pairs....8 different villages at a lake and 3 different villages at the sea.

As a Church of the Supergenius junior research fellow, I cannot agree more. Without historical benchmarks for doing large dataset regression analysis, this is deep dive in a shallow 5 gallon tank of ethnographic aquascaping.

How is "learning" any different than "culture", which is just a catch-all for all sorts of environmental causes? Who restricts "culture" to "multi-generational traditions," or whatever? You don't need to add anything.

Tyler Cowan for one? There is a long debate in economics about the effect of culture on prosperity, and whether these cultural effects outweigh policy, and if so under whuch circumstances. The definition of "culture" used thus implicitly is long-term. Words are formed by yse, and their definitions often differ between specialised sectors and the mainstream. Notoriously mathematics uses the word "obvious" in a way that's completely opposite ttoeveryone else's meaning.

This isn't a Cowan [sic] post.

Discussion of culture in economics is not limited to discussion of *national* cultures. It's not even restricted to generational transmission, but also includes socialization and learning.

Insofar as culture is restricted to particular kinds of learning, it might be restricted to learning that has to do with identity formation. So it's not "learning" tout court that needs to be added - if something needs to be added - but learning other than learning that's part of identity formation.

I doubt economists have been missing the fact that learning which affects preferences isn't restricted to learning related to identity formation, but it's possible.

Oops sorry on the name.
Of course discussion of culture in economics is not limited in the ways you describe, and I said nothing that said it was. What I am saying is that the word has acquired overtones of long-termness. So when the distinction needs to be drawn between different forms of learning, it's not surprising that Alex contrasts this form of learning with "culture". Words' definitions change as more distinctions need to be drawn. Though mathematicians really can't be excused by this process.

And personally I think some economists are too quick to attribute differences to culture in Alex's sense of the word, and don't test alternative theories enough. It's very difficult to keep in mind all plausible alternatives when hypothesis testing. I know I'm lousy at it myself.

"The Linguistic Wars" ISBN-10: 019509834X.

Don't bite the hook of economics as being the answer. The comment section is going to send me to the hospital from laughing so hard.

I would love to have been born into the lake fisherman society and be sitting around the campfire talking about this. And is upsetting if ya'll don't see how good this "ethnographic" study is. Can we please take off our warped lens that this is an economics thing? LOL. I can't wait to do some studies on this around our college of higher learning where we do things like this because, well, we are not tier 1 and driven by this insane need for grant money and big data regression analysis. LOL.

Very Alchian and Demsetz, property right paradigm style. They predict that in a sphere where things are organized privately, no strong norms will emerge to govern the resource (few spillover effects), while common pool resources will cause strong norms to emerge as an alternate means to governance of that resource. People living where things tend to be ordered privately will have a more individualistic society, while people living where things are ordered communally will have a collectivist society.

Are Alchian and Demsetz right? Communist countries were notorious for corruption and lack of trust between citizens (partly because of secret informers), which doesn't sound very collectivist culturally to me.

Unless of course Alchian and Demsetz are merely stating definitions.

Families and ascetic religious orders seem to do fine with collective ownership, as do thousands of successful business entities. But the socialist countries relied on an economic theory that doesn't describe reality; they didn't work because they couldn't work.

Sure, but do families or successful business entities have collectivist societies?
(Ascetic religious orders are no good for testing A&D's theory as it seems reasonably safe to assume that anyone who wants a dedicated religious life that's also individualistic becomes a hermit, or joins a religious order whose acolytes wander or some such.)

20th century national Communism isn't the only form of collectivism (or "small-c" communism) around.

I never said it was the only form. But it's a famous form, and if A&D's theories are meant to go beyond tautology then historical Communism then it's a form I'm interested in how they explain. Perhaps the question is in their definition of communal order.

Two things: It isn't necessarily trust, but a willingness to punish people who use communal resources in unapproved ways. In the US, moralistic sanctions tend to be strongest where there are perceived spillover effects, especially regarding the environment (though there are paternalistic moral sanctions as well). These moral sanctions, if they are appropriate in size and direction, increase efficiency. Moral outrage is a scarce resource, too, so a country that is largely ordered through private property will better manage its collective resources through moral outrage than a country that orders all its property as collective resources. Business arrangements within a firm seem to be subject to more social sanctions/approbation than arrangements across property boundaries, and there's quite a lot of effort put into developing a positive corporate culture.

If you buy the Hofstede measures (which are certainly problematic), Russia measures as highly collectivist, as does China. South Korea does, too, and Japan is a sort of middle ground. Obviously causation is a huge problem, but East Germany does rate more collectivist than West Germany, even still.

What does Hofstede mean by collectivist society? I found his website but couldn't find any definitions.

20th century Communism wasn't collectivist in this ressource-sharing way of seeing things. Even the kolkhozy - the collective farms - were under strict management by state officials.

A much better example would be the pre-industrial Russian obschina or mir (or indeed any community in pre-modern times)

Yes, that matches with my understanding of Soviet Russia: communal ownership of resources but no collectivist society.

Though I don't know why you say that other examples would be better. Do A&D limit the applicability of their theory somehow to only certain types of communal ownership?

Hoping for success against an unknown, randomly-selected opponent sounds more like gambling than competing.

That was my first reaction too. This is an indicator of risk tolerance, not a marker of individuality.

A collectivist society can also have a high risk tolerance.

Actually, I think that's addressed. They note that the ones most likely to choose to compete were the ones most likely to have been successful on the lake for a while in the first place. Presumably they're estimating their chances based on what they normally experience and realize that the competition is being drawn from the pool of guys they compete with every day anyway.

Actually, the study controls for risk tolerance. The participants' risk tolerance is measured and the findings are robust taking risk preferences into account

How did they measure for risk tolerance?

Did they have the participents play other wagering games for real money? Perhaps the sea fishers knew the wisdom of reducing the volatility of their payout?

Tyler is absolutely salivating over the thought that, possibly (Nay! Experimentally proven!), his Great Big Ideas for the transformation of human society will gain acceptance in a single generation.

Seems fishy, like a poster upstream said. Also note the chica that is recording results. She's hot enough to make you take the eye off the ball so to speak--did both groups have the same recorder?

Further, how do we know that sea fishermen, who take bigger risks (the sea is harder to survive in than a lake, all things being equal), are not more risk adverse from years of hard experience than the more reckless lake fishermen? The years spent working support this thesis: the sea fisherman, the longer they work, become more risk adverse since they see dangers that jejune sea fisherman do not see; conversely the lake fisherman become more reckless, less risk adverse, the longer they work since fishing on the lake is a dream job free of risk.

the authors control for risk attitudes. They don't find risk attitudes to change over time.

How do they control for such a hard-to-measure parameter? A survey where you self-describe yourself? Another bogus social science report along the lines of that Dutch researcher, or, like finding a sinusoidal pattern in a string of random numbers, a mere coincidence that falls within the 95% confidence interval.

I guess we're learning to 'farm' the Government.

"what the authors demonstrate is that differences in the choice to compete or not appear to be learned differences."

Or perhaps that fishermen who do not want to be so competitive choose to fish in the sea, while those without aversion to competition fish in the lake? Just from what you've described, I don't think that it's necessarily clear whether their jobs explain their competitiveness or their competitiveness explains their jobs.

Fishing in the sea requires a lot more cooperation than the lake. If a ship is in trouble, neighboring ships will always respond. This may suppress some competitiveness.


The research relates in interesting ways to economic anthropology research published about fishing communities in Northeast Brazil. The book *Sons of the Sea Goddess* examines some of the community effects of the shift from self-paddled canoe sea fishing to motorized sea fishing. It has been nearly 20 years since I've read the book, but what I recall is that canoe fishermen tended to be solitary and home-focused: they'd fish in the morning, bring the catch in to sell, then spend the afternoon in front of their home mending their nets. Those fishermen joining motorized commercial fishing ventures would find work at a community bar, then be aboard ship for two to four days, get a share of the take when they return to port, then hang out around the bar looking for their next job.

Whether the shift had implications for community/cooperative outlooks versus individual/competitiove outlooks I don't recall, but if the Leibbrandt, Gneezy and List result is any good then one would suspect such an effect.

"First, the lake villagers who chose to compete are among the most successful lake fishermen–that is, they have learned that competition increases income."

The statement is false because its composition requires it to read

First, the lake villagers who chose to compete are among the most successful lake fishermen–that is, they have learned that WINNING the competition increases income.

The learning seems to be "once a winner, always a winner".

Which is why pitches start off, "sure fire way to get rich", then include the warning "past performance does not predict future results", followed by, you will never regret putting every dime you have plus all you can borrow into this investment because you will get ten to one hundred times that back."

The lake fishermen could be competing against professional ball player in a rational world - agreeing to invest in something you don't understand is agreeing to be robbed.

Cool study. You should read their new book The Why Axis!

This study sounds like garbage, for many reasons.

Comments for this post are closed