Garett Jones  

Group IQ: One Source of Trustworthiness

How Stagnant Are We? A Time D... Robert Murphy: A New Result on...
In Arnold's new essay entitled "Libertarians and Group Norms," he writes:

[W]e live in a world that demands enormous levels of trust among strangers...I doubt that anyone fully comprehends what holds this fabric of trust together. 

I agree.  But we're building comprehension, one journal article at a time.  A new line of experimental  research shows that groups with higher cognitive skills are likely to have higher levels of trust.  More importantly, they're likely to have higher levels of trustworthiness.   A big component of the wealth of nations.  

Let's use the term "IQ" as a shorthand for these cognitive skills.  Some of the studies in the literature use conventional IQ tests, while others use tests that have a strong positive correlation with IQ (like the SAT).  Either way, we'll be checking to see if there's a positive relationship between these mental skills and what most people would consider personality traits.  

One of the standard games of trust and trustworthiness is the prisoner's dilemma. Two people deciding whether to work hard on a team project when the credit will be shared 50-50 no matter what; two tribes each deciding whether to make its own output or take the other's output; less naively, two firms deciding whether to carve up the auto market or instead engage in a price war. 

Each is a prisoner's dilemma: My best option imposes negative side effects on you and vice versa.  So, shall you and I grow the pie or shall we fight over it?  An eternal dilemma for humans.  

The usual way to escape the prisoner's dilemma is to have the same people play each other quite a number of times.  If you know that your tribe and mine will be seeing each other for a while, you just might find it wise to refrain from night raids on my tribe.  After all, if you raid my tribe tonight, I might reciprocate tomorrow night.  But if we both refrain tonight, and the next night, we might be able to sustain a truce for quite some time.   

I often think of civilization as one grand truce.  

Back in 2005, I wanted to know whether high IQ groups were more cooperative--whether they were better at keeping the truce, whether they were better at seeing life as a repeated game where we play "tit for tat". 

I didn't have access to an experimental economics lab back then, but I did have access to a library and a research assistant.  So I collected every prisoner's dilemma study I could find, along with data on the average SAT score of the school and other characteristics (whether the school was private, whether the students were paid in cash, etc.).  The finding, published in 2008

A meta-study of repeated prisoner's dilemma experiments run at numerous universities suggests that students cooperate 5% to 8% more often for every 100 point increase in the school's average SAT score. 

This finding was the first of its kind: In prisoner's dilemmas, smarter groups really were more cooperative.  Since then other researchers have found similar results, some of which I discuss in Section III of this article for the Asian Development Review.  It looks like intelligence is a form of social intelligence.  

However you measure it--with IQ tests or math and science tests--it appears that nations currently differ in their levels of cognitive skill.  It would be excellent to raise cognitive skills around the world, and in section IID of my ADR paper I discuss some crucial public health interventions that can do just that.  

But as a macroeconomist, not a public health researcher, my comparative advantage is at investigating whether average cognitive skills matter for national prosperity.  If this link between group cognitive skills and trustworthiness holds across countries like it does within countries, I'd expect nations who do better on these tests to have economic institutions that are better at creating win-win outcomes.  Nations with high average cognitive skills probably have less rent-seeking for instance.  

Does that happen in the real world?  If it does, does it mean that there are negative political externalities to low-skill immigration?  That's a topic for a later time.  

Another worthy question: Why would high IQ groups be more cooperative anyway?  Isn't cynicism intelligent?  Sure, sometimes, but the political entrepreneur who can find a way to sustain a truce can probably skim quite a lot of the resulting prosperity off for herself.  And people who are better  at solving the puzzles in an IQ test are probably better at solving the puzzles of human interaction.  

Coda: Toward the end of his essay on libertarianism and group norms, Arnold makes this claim, one well worth reflecting upon: 

Indeed, I believe that libertarians are on much stronger ground if they support a Tocquevillian ideal of voluntary associations rather than a Randian ideal of individuals living under an objective moral code. 

COMMENTS (12 to date)
Ted Levy writes:

Well, Galt's Gulch WAS "a Tocquevillian ideal of voluntary associations." Granted, the American State Rand supported, not so much...

MikeP writes:

And people who are better at solving the puzzles in an IQ test are probably better at solving the puzzles of human interaction.

Up to a point...

Steve Sailer writes:

Here's my long American Conservative article on trust and trustworthiness:

Steve Sailer writes:

"Idiocracy" does a memorable job of pointing out that societies of low IQ individuals tend not to be sweet-natured Forrest Gumps, but surly and suspicious.

MingoV writes:

One reason why no nation has a libertarian government is that there are too many people who lack the intelligence (or the will) to be responsible for themselves AND to cooperate wisely with others. Governments gain power because such people cannot achieve optimal outcomes for economic "prisoners' dilemma" situations: those people want the 'wise' nanny-state to ensure that they won't lose too badly in their economic interactions.

Floccina writes:

You may be right but Sweden has about the same average IQ as Italy. So it seems trust can vary independent of IQ and it seems we might get more traction attempting to increase creating than attempting to increase IQ.

Martin Hewson writes:

There has been a long term trend to more cooperation (see Robert Wright, Nonzero; Matt Ridley, The Rational Optimist). Might a long-term Flynn Effect of rising IQ be behind the trend of growing cooperation?

Steve Sailer writes:

Minnesota remains more honest, trustworthy, and trusting than New Jersey. That suggests we barely have any clue how to "get more traction," so we ought to be careful not only about the IQ of immigrants but about other characteristics as well. Thus, low IQ immigrants from narco-violent Sinaloa would be about last on the preferred list, but they seem to be close to first on the actual source of immigrants.

John T. Kennedy writes:

It seems to me that Rand's code obviously requires voluntary associations. What other kind of association would be consistent with the code?

OneEyedMan writes:

It doesn't jive exactly with the prisoner's dilemma game form, but perhaps the trust issue is simply a form of k-level thinking. That is, perhaps that since high IQ people are less likely to miss-understand social cues and the game, they can be relied upon to play the game more consistently which enables more cooperation. I remember reading that Cal Tech students choose a higher k-level implied choices in the Keynesian Beauty Contest than other schools. While this may be because they are smarter it may just because they expect others to be smarter. If so, the same thing could be happening with trust.

I wonder if better pairs from better schools do better in Split or Steal.

Garett Jones writes:


Arnold is pointing toward the power of multiple types of voluntary associations, even associations based on moralities with which we disagree.


This paper coauthored by Cesarini, "Billiards and Brains," finds just what you predicted: High IQ individuals do better at k-level thinking, informally known as the Keynesian Beauty Contest.

More importantly, they apparently do a better job thinking about what the k-level is of the people around them.

Here, the intelligent have a better model of the minds of others.

Ian Luria writes:

"even associations based on moralities with which we disagree"

What would be an example of a morality that is incapable with an Objectivist society?

Comments for this entry have been closed
Return to top