Arnold Kling  


Mass Transit and Happiness... Health and Taxes...

Newsweek has a survey of what I think it should have called neuroeconomics. They use the term "behavioral economics," which I think of as looking at cognitive biases in decision making. Neuroeconomics links cognitive biases to brain science.

Observing that some societies are consistently richer than others, social scientists have invoked such ingenious explanations as "the Protestant ethic" (of working and saving for the future) or "the resource curse" (when an elite controls a valuable natural resource, such as oil, and has no incentive to encourage political and economic modernization). One of the newest explanations is "trust," which varies widely between societies and is strongly correlated with economic growth, says Paul Zak, an economist at Claremont Graduate University. Trust encourages savings and investment, and reduces the "transaction cost" of investigating the people you do business with. But, compared with well-studied behaviors such as aggression, relatively little is known about the biological basis for trust. (Zak's own research is not on brain function directly, but on oxytocin, a hormone that seems to promote trust. It is usually studied in relation not to the stock market but to lovemaking and breast-feeding.)

The question of what makes some societies richer than others is much more complex than the preceding paragraph implies. And I would argue that the very fact that it there are differences across socieities implies that individual hormone levels are not the answer. In my view, institutional and cultural differences are likely to provide better explanations.

I mean, if you want to try to do research showing that South Koreans have more trust than North Koreans, go ahead. But I think you are more likely to find that the difference between the two societies is that one uses a market system and one uses a Communist system.

For Discussion. If economic success and high levels of trust are correlated, which way is the causality arrow to be drawn?

Comments and Sharing

COMMENTS (4 to date)
Lawrance George Lux writes:

I infer their use of trust implies an intergration of legal systems, so that same argument defense is accepted throughout. Even this does not show high correlation with economic success. The prime consideration for economic success must be social stratification--or rather, the lack of it.

Social stratification always brings reduction of Profit levels for Production operations, reduction in the level of Business inventiveness, increase in overall established Business Profits without increase in Product or Service provision. Societies begin a multiplication of population, without rising standards of living, or effective Job opportunities. lgl

Peter Gallagher writes:

Hi Arnold,

The 'communism' aspect is probably irrelevant, too. Vietnam remains a one-party communist state that has done extremely well since deciding in 1987 that markets work if you allow them to.

The difference between the Koreas is probably markets alone. They seem to generate the sort of 'social capital' (I'm uncomfortable with this idea, but I can't think of another) that is reflected in a more robust network of relationships between individuals that -- seen in the right light -- could be called 'trust', or 'mutual obligation' or 'enlightened self-interest'.

So to answer your question: correlates such as "economic success" and "trust" don't need to be causally linked. They're probably both the products of something else.

Paul J. Zak writes:

Dr. Kling is correct that institutional and social factors are much more important to understanding general trust levels across countries and a country's resulting wealth or poverty than are hormones (see Zak, PJ, and Knack, S., "Trust and Growth", Econ. J., 111:295-321, 2001). In this paper, we demonstrate that trust affects growth (changes in income) and not vice-versa (though income levels affect trust).

The work of mine mentioned in the Newsweek article examines the neurophysiology of interpersonal trust holding the institutional environment (the rules of the interaction) constant. The two are of course related, with the hormone data telling us how to design environments to foster trust. This basic research can then be applied to understanding poverty. Again, the big leverage on country-level trust is in institutional design. Generalized trust is raised by increasing education, increasing freedoms, and decreasing income inequality (Knack, S., Zak, PJ, "Building Trust: Public Policy, Interpersonal Trust, and Economic Development", Supreme Court Economic Review, 10:91-107, 2002). A nontechnical overview of this work can be found in my paper "Trust" downloadable at

Colin Camerer writes:

Hello all-- Love your thoughtful blog, which is asking *all* the right questions about behavioral economics.

Some comments on Arnold's post. Yes, the Newsweek article is about neuroeconomics. This is an area which is a part of behavioral economics (using evidence from neighboring social sciences, especially psychology, on rationality limits to renovate economics) because it is new facts & tools; but it's also part of experimental economics.

On trust-growth, of course there is a leap. Cross-country work by Paul Zak and others show that assessed trust correlates with growth *across countries*. The causality question is the key one. Keep in mind, too, that the hormone surges can be quite situationally dependent (in Paul's study, for example, when a random device determines how much the second person is "trusted" there is no oxytocin). This means, for example, if somebody sends me money for an Ebay item I'm about to mail, I might feel a surge of oxytocin. But if a regulation compels them to send money, no oxy. So institutions can *cause* trust. It is not the *hormone levels* that are probably biologically universal, but the capacity of those levels to respond to expressions of trust and empathy cues.

One more thing: The argument that markets automatically create wealth can be tested experimentally by installing and uninstalling free markets. (Keep in mind that policy institutions like the World Bank are sometimes trying to do this too, except with small samples and no control groups!). We are starting to work on this in our lab. Probably it is not just markets, per se, but also that supporting infrastructure is needed-- accounting, some capitalist spirit, a certain base level of trust which generates moral obligation and limits lemons problems, rule of law, and so forth. Stay tuned.

Comments for this entry have been closed
Return to top