Bryan Caplan  

Intelligence and Economic Beliefs

PRINT
Socialist Hate Speech of Antiq... My Advice to Tyler on the Econ...

One of the clearest facts about economic beliefs is that more educated people think more like economists. A lot of economists say their experience in academia completely contradicts this, but (a) the highly-educated folks who hang around universities are heavily self-selected for leftism, and (b) however bad the economic beliefs of the typical college graduate are, the economic beliefs of the typical high school drop-out are far worse.

Once people accept the reality of this pattern, many leap to the assumption that education is a proxy for income or wealth. But this is incorrect, too. The effect of education on economic beliefs remains about as strong after you statistically control for income (as well as job security, income growth, gender, age, and race).

No one, however, has explored the possibility that seems most plausible to me: More educated people think more like economists because they are unusually intelligent. Until now, that is. Today I submitted a new paper, "Economic Beliefs, Intelligence, and Ability Bias: Evidence from the General Social Survey," co-authored with Stephen C. Miller, to the Economic Journal. Long story short: Not only does controlling for intelligence reduce the estimated effect of education on economic beliefs; intelligence turns out to be the single strongest overall predictor of economic beliefs.

Education still has a large independent effect on beliefs - especially beliefs about the international economy. But controlling for intelligence, the average magnitude of the effect of education falls by about one-third. This change is particularly impressive because the General Social Survey's IQ test only has ten questions, which biases estimates of the effect of IQ downward. The true effect of intelligence is probably even larger, and the true effect of education is probably even smaller.

What does all this mean? The simplest explanation is also the best: Basic economics is true, and smart people are more likely to believe what is true. Psychometricians already know that almost all forms of knowledge are positively correlated. Now we can officially add economic knowledge to the list.


Comments and Sharing





TRACKBACKS (6 to date)
TrackBack URL: http://econlog.econlib.org/mt/mt-tb.cgi/534
The author at Half Sigma in a related article titled What the GSS really says about intelligence and economic attitudes writes:
    Bryan Caplan of the blog EconLog has a post which he titles Intelligence and Economic Beliefs. Bryan has coauthored a paper with Stephen Miller in which he alleges the above statement to be true based on data in the General [Tracked on July 14, 2006 9:36 PM]
The author at De Gustibus Non Est Disputandum in a related article titled Economistas podem ter um futuro brilhante sim writes:
    Basic economics is true, and smart people are more likely to believe what is true. Psychometricians already know that almost all forms of knowledge are positively correlated. Now we can officially add economic knowledge to the list. Por isto, digo... [Tracked on July 18, 2006 5:19 AM]
COMMENTS (24 to date)
Fritz writes:

Being a Leftist is easy, requires no intelligence. Intelligent people can understand intangible effects on behavior. I noticed in college that the distribution of economic exam scores had many C's & many A's but very few B's.

FMB writes:

So, Bryan, what about the converse? How much can you say about someone's intelligence based on their economic beliefs? I'm sure people here (including the first commenter) have their opinions, but I'm wondering what your data says.

dearieme writes:

In my experience of academia, engineers tend to be more economically rational than scientists and mathematicians, but the latter two groups tend to congratulate themselves on being more intelligent than engineers. Any views?

On the one hand, the conclusion sounds plausible.

On the other hand, can a mere ten questions reliably test IQ?

Martin Kelly writes:

Without wishing to seem facetious, by what bias do you presume all economists are smart?

The Engineer writes:

>In my experience of academia, engineers tend to be more economically rational than scientists and mathematicians, but the latter two groups tend to congratulate themselves on being more intelligent than engineers. Any views?

Are you talking about academic engineers (i.e. professors)? If so, they ARE scientists and mathemiticians! They're certainly not engineers.

Engineers, academic or otherwise, are a pretty conservative bunch. That probably translates into a classical economics outlook, belief in the invisible hand and all that.

Steve Miller writes:

"On the one hand, the conclusion sounds plausible.

On the other hand, can a mere ten questions reliably test IQ?"

pp. 3-4 and pp. 7-8 of the paper linked above address that question. For more details, see the cited works by Wolfle, Huang and Hauser, Wechsler, and Miner. As brief tests go, the vocabulary subtest from the WAIS (and even shorter versions like WORDSUM) are remarkably good substitutes for overall WAIS IQ scores. Not perfect, but the main effect appears to be a loss of data at the tails.

Half Sigma writes:

I already blogged about this, and it's not clear whether more intelligent people understand the Truth, or if they are just voting their economic interests because income and social class are correlated with IQ.

Steve Miller writes:

HS: it's quite a bit clearer if you control for income, job security, etc. If you control for those things and still find the IQ effect, as we did, then you can't very easily dismiss the "Truth" hypothesis in favor of the "class interest" one.

Steve Sailer writes:

Good work.

And if you had an IQ test that measured non-verbal intelligence as well as just vocabulary, I bet the effect would be even stronger.

aaron writes:

As we're all aware, also, highly-educated does not necessarily mean highly-intelligent.

Steve Sailer writes:

People who score higher on the SAT Math subsection than on the SAT Verbal tend to be more "tough-minded," while verbalists tend to be more "tender-minded." So, I bet if the GSS had asked math questions and visual/performance questions as well as vocabulary questions, you would have found an even higher correlation between high cognitive scores and "thinking like an economist."

Here's a possible research opportunity for you. There's an annual survey of the political views of college students that breaks out results by college. You could try correlating the answers on economics-related questions to each college's average SAT scores. I bet the ratio of SAT-Math to SAT-Verbal would be significant.

Half Sigma writes:

In the GSS, Wordsum has a tiny correlation with income after education is taken into account. But Wordsum has a much stronger correlation with various other questions, including ones that aren't necessarily smart economic thinking.

Wordsum is probably a better determination of social class than any other measure on the GSS, including income or education. So smarter people are more likely to be pro-environment because that's an issue that the upper classes like.

Generally it's poor people who favor redistributiong government income, because they want some of it. And of course, poor people tend to be less intelligent.

Smart people are more libertarian, and the correlation is a lot stronger on non-economic issues. Smart people are far more likely to support abortion rights, to be in favor of free speech (an abstract concept that stupid people don't get), and to be against prayer in school.

Steve Miller writes:

HS: Don't be ridiculous. The whole point of multiple regression analysis is to be able to find one effect independent of another. IQ independent of income, job security, income growth, race, etc. is just that -- and it still has a stronger effect than any other control. You can't pretend WORDSUM is a measure of class, when nearly any other class variables you could name are used as controls. Our results flatly contradict your speculation.

Steve: I think you are probably right about non-vocab subtests, though simply using an error-in-variable correction for both education and WORDSUM strengthens our results dramatically.

avm writes:

Regarding:"More educated people think more like economists because they are unusually intelligent"", I have to laugh out loud for the next couple of hours...Please, can you at least pretend to have a shread of modesty and reformulate?

What about people which, maybe you don't consider intelligent, but which were quite respected in their times? Where do you suppose Sartre would fall in this world of yours? what about different types of intelligence? IQ is not a complete measure, and you have to be an economist to think that a diplomat and a mathematician are not inteligent in different ways.

Steve Miller writes:

Scoff all you want, but it's not any kind of a stretch to suggest that professional Ph.D. economists (the basis for Bryan's original SAEE papers) are unusually intelligent compared to the general population.

Half Sigma writes:

"You can't pretend WORDSUM is a measure of class"

Wordsum measures vocabulary skills, but that's a great correlate of class. Much more so than income.

Steve Miller writes:

WORDSUM is a much better proxy of IQ than it is of social class. Then there's the question of why you'd think social class is relevant when we talk about "class interest," when that phrase generally refers to economic class interest. There's also the problem that IQ is far less determined by social or economic class, than, say, education -- which is also controlled for.

Steve Miller writes:

Half Sigma:

Let me follow up, in hopes of making my point as clear as possible: for the relationship we observe between IQ and economic beliefs to reflect class interest, you'd need a definition of class that is independent of income, job security, income growth, education, race, sex, age, political party identification, and political ideology. I cannot conceive of a definition of class that doesn't include many or most of those components, and simply using the label of "social class" doesn't get around that problem.

The Engineer writes:

Good paper. Maybe what you were doing was less complicated than what Levitt did with abortion and crime, but I was more easily able to follow your paper than any of his. I think that I know what you did, I still don't understand what he did.

When you guys say that you were "controlling" for something (IQ, for example) does that just mean that you have it in your multivariate regression?

avm writes:

As it happens, being a PhD requires to be more intelligent than most of the people - in some sense of the word intelligent. Not necessarily in economics, but in physics, maths, literature. There ar ealso people who are extremely intelligent and are not educated.

From here to saying that intelligent people think like economists because they are unusual intelligent is quite a leap of faith. Do you mean that if somebody doe snot reason like an economist, they are not intelligent - or intelligent but not unusually intelligent. Or do you mean that economist are unusually intelligent (compared to, say, neurobiologists)?

and by this, i am not attacking economists in general, or PhD-s in economics (which you seem to believe), but particular line of argument.

cpurick writes:

"for the relationship we observe between IQ and economic beliefs to reflect class interest, you'd need a definition of class that is independent of income, job security, income growth, education, race, sex, age, political party identification, and political ideology."

Half Sigma doesn't need definitions; he simply makes them up when necessary.

Anon writes:
There's also the problem that IQ is far less determined by social or economic class, than, say, education -- which is also controlled for.
Sharon Begley wrote a small article on how intelligence, relative to others, is constant throughout one's life. At adulthood, the heritability of intelligence is 70-80%, I believe. You may not want to venture farther in this marsh, lest a libertarian sacred cow be slaughtered. Here it is at Sigma's blog, because a login is necessary for WSJ http://www.halfsigma.com/2006/06/wall_street_jou.html#comments
C-mac writes:

Perhaps the tie of of intelligence with economics is a bell curve, with people on the lower end of intelligence being leftist, people with median to high intelligence thinking like capitalist economists, and at the other end people with extreme intelligence go back into leftism. I ponder this because Einstein was a socialist was he not? Perhaps he understood something people with average intelligence could not.

Comments for this entry have been closed
Return to top