Arnold Kling  

Is This a Swindle?

Two Rothbardian Reductios... Common Genius...

Brad DeLong writes,

If inherited genetically-based IQ were the source of the extra edge that the children of the rich get in our society, than we would expect a parent with 4 times average lifetime full-time earnings--say $200,000 a year--to have a kid with a lifetime average income of $51,500 instead of the average of $50,000. But it is not $51,500. It is $150,000.

What he is doing (read the whole post) is arguing that number that are less than one, when multiplied together, approach zero. So the correlation between my percentile in IQ and my kids' percentile in IQ will be less than one. The correlation between someone's percentile in IQ and their percentile in the income distribution is also less than one. Multiply the two together, and you get a small number. But the actual correlation between parent income and kids' income is high. Ergo, there must be some social transmission of inequality, which, for someone of Brad's inclination, means that there is an opportunity for technocrats to step in and "fix" the distribution of income.

Is this argument a swindle? What bothers me about the DeLong story is that if you multiply enough small numbers, then the predicted heritability of anything approaches zero. It's the definition of regression to the mean. After enough generations, all heritable characteristics should even out. But they don't. What actually happens?

Comments and Sharing

CATEGORIES: IQ in Economics

COMMENTS (22 to date)
Jody writes:

Assuming independence of the small numbers is the swindle.

Lots of factors influence success (e.g., looks, height, IQ, social skills, drive) and are heritable. I.e., if your parents are hot uber-nerds (see Cindy Crawford) or really driven nerds (see Bill Gates), odds are you'll bee too. Or in otherwords, you inherit a package of characteristics.

razib writes:

After enough generations, all heritable characteristics should even out. But they don't. What actually happens?

huh? i know you know genetics dude, so you must be asking something more sophisticated than i can figure out (i can't think like an economist). populations regress back to the breeding population value, right? but the new breeding population value doesn't go all the way back to the old mean, and unless you apply more selection it won't. in terms of the biophysical substrate, genetics is discrete and not blending, so the part of the variance that is additive genetic alters the population's architecture. if you assume heritability remains constant you can just iterate the breeder's equation to model the population (mean value after selection = heritability X selection differential). in terms of assortative mating you start to increase heritability and the population starts to develop substructure so that there may be multiple breeding values to which individuals may regress. that is, the population turns multi-modal.

hope i didn't make myself look stupid ;-)

Peter Twieg writes:

What bothers me about the DeLong story is that if you multiply enough small numbers, then the predicted heritability of anything approaches zero. It's the definition of regression to the mean. After enough generations, all heritable characteristics should even out. But they don't. What actually happens?

All regression to the mean implies in this case is that the correlation between one's self and one's offspring X generations down the line will decrease to 0 as X becomes large. And as the correlation approaches 0, the expected value of the offspring's intelligence will be the mean of the intelligence distribution.

Naturally, this doesn't say anything about what the distribution itself will look like X generations down the line. It certainly doesn't mean that "all heritable characteristics will even out" (their distribution will tighten.) It simply means that the states of past generations will not wholly determine the states of future generations. There's still plenty of other factors that allow for variability, including the intelligence of one's more recent ancestors.

Brad Hutchings writes:

So he's saying that IQ is not the source of the income edge. This could be true. While higher IQ generally correlates with higher income, the correlation might be directional, i.e. higher IQ ==> higher income. But higher income might not imply higher IQ. It may rely on different skills, such as athleticism or artistic creativity or salesmanship or leadership ability or good looks or just kissing plenty of other people's back sides. I know plenty of people who are quite wealthy but not geniuses. Most really IQ smart people I know have figured out how to keep their asking price high, thus easily earning salaries that average people would have to work hard for. It's entirely consistent with what DeLong is noticing and doesn't call for any intervention. Nobody in the private sector is earning more than what they're actually worth to the person writing the check.

Steve Sailer writes:

Yeah, it's a swindle.

Correlation isn't the same as effect size. In a complex system with many factors, a single factor can have a low correlation but still have a sizable effect (or vice-versa). As an economist, DeLong should be thinking about the effect of IQ all else being equal, but instead he's trying to pull the wool over our eyes.

Charles Murray quantified this in an ingenious study of pairs of American siblings raised together in non-poor homes. Murray described his findings in the Sunday Times of London in 1997:

"Each pair consists of one sibling with an IQ in the normal range of 90-110, a range that includes 50% of the population. I will call this group the normals. The second sibling in each pair had an IQ either higher than 110, putting him in the top quartile of intelligence (the brights) or lower than 90, putting him in the bottom quartile (the dulls). These constraints produced a sample of 710 pairs. How much difference did IQ make? Earned income is a good place to begin. In 1993, when we took our most recent look at them, members of the sample were aged 28-36. That year, the bright siblings earned almost double the average of the dull: £22,400 compared to £11,800. The normals were in the middle, averaging £16,800." [IQ Will Put You In Your Place, Charles Murray, Sunday Times, UK, Day 25, 1997]

Fly Fisher writes:

DeLong's narrative is similar to many: the children of rich parents tend to be rich, and this tendency is stronger than can be explained by IQ, so we must use 'our collective hands' to level the playing field. So, high IQ people who have successful parents for role-models don't tend to drop back to 3% above the national average for income. Is this a problem that we should use our collective hands to solve?

What about the children of poor parents? Surely someone must have looked at the correlation of IQ for children who become much richer than their parents. When someone moves up in wealth relative to their parents, how much of the mobility is explained by IQ? This is the social issue that concerns me more. Having been born to very, very poor parents, I am less concerned about bringing down the rich.

By DeLong's standards, my children have been born into a very wealthy family. If my wife and I help our children attain education and access to opportunities that bring them wealth and happiness, why should this now be used as evidence that our system needs to be 'fixed'? Isn't it a good thing that successful parents have successful children? Why is this a problem?

Again, I am much more interested in the factors that allow children of poor parents to attain education, wealth, and happiness. Is inherited genetically-based IQ an extra edge here?

Troy Camplin writes:

First, he's assuming that the children will start off in the same place as the average person. This, of course, is a fallacy. A wealthy person can provide many things for their children. The money they have can, of course, provide better education and more educational materials, not to mention more complex social situations. The parents also likely have more social connections, allowing the children to start off life with many social connections. Most high-paying jobs are gotten through social connections rather than objective merit. I would say this, more than the money per se, has a larger effect on what the children of rich people earn than anything else.

Dr. T writes:

DeLong's argument shows ignorance of biology and statistics. The reason we do not see regression to the mean with IQ is that human procreation is not random. Few high IQ people have children with partners of below average IQ. If a person with a 140 IQ marries a person of 130 IQ, the likelihood that their child will have a 100 IQ is less than 5% (unless there is premature birth or brain damage).

Since wealth correlates with high IQ, and since high IQ couples have high IQ children, it is unsurprising that most of those children become wealthy. I believe this would happen even if the parents were killed, the estate was squandered, and the baby was raised by average IQ guardians. The genetic component of IQ dominates, and it takes an awful environment to significantly reduce a child's IQ.

John Thacker writes:

The twin studies mentioned by Steve Sailer are a problem first.

Another problem is that, even granted all of this, the desire to pass along rewards to one's children is a powerful motive, and thus it's not always efficient to try to eliminate it (and often not even possible-- much of the support for "public" schools and opposition to vouchers is from suburban parents who want to keep the schools their kids go to good.)

Onto more technical complaints, using heritability alone strictly speaking ignores the covariance of genetics and environment, or assumes it to be zero. In other words, it assumes no correlation between genetics and environment. Real life is rarely such a nice experiment; they are correlated, so determining the true impact of genes is very tricky.

Another mathematical complaint is the mathematical assumption of linearity. All the papers are assuming a linear response between IQ and earnings. Otherwise E(f(X)) != f(E(X)) in general, as they assume. Correlation coefficients only measure the linear relationship, not any nonlinear relationship. It's quite possible that there is a nonlinear relationship. Ordinarily linear relationships dominate with two normally distributed r.v., so it's okay. However, if there's covariance between IQ and environment then the mathematical assumptions and simplifications fail, the chain of reasoning falls apart, and the conclusion does not hold.

DeLong's summary is also first off confusing because he's not speaking in terms of standard deviations. But basically, yes, they're talking about multiplying correlations to get a small number. It all breaks down if there's covariance and nonlinear responses, though.

At the same time, note he's arguing that since IQ doesn't play that large of a role, that equality of opportunity does exist; it seems he's leaving a large room for effort, etc.

Bruce G Charlton writes:

Temperament is inherited, as well as IQ (see Judith Rich Harris's 'The Nurture Assumption') - and a Big 5 personality factor 'conscientious' temperament is positively correlated with job success. So is honesty, which is also probably inherited.

But, back to IQ. It's the broad sweep of data that counts, not mathemetical modelling.

Daniel Nettle used long term follow up data from the UK. Subjects were IQ tested at 11 and their social class measured at 42. There was considerable IQ based mobility in all classes; and there was no evidence that it was harder for people born into lower social classes to rise. In other words, the data were conistent with a near-perfect IQ meritocracy.

Of course the IQ average and distribution was notably stratified between social classes.

Read the whole thing:

jaim klein writes:

People tends to marry other people like them, and intelligent people tend to marry and procreate with intelligent partners. It is most probable that Prof. De Long himself is married to an educated woman and his children's IQ revert to their parents's (high) mean. Prof. Kling, you are wrong to call it a swindle, it is prudence aka political correctness, i.e. purposeful blindness to social & racial reality. In the view of what happened to Prof. Summers and Nobel Prize Prof. Watkins, prudence is most indicated. No one is entitled to question actions taken with a gun aimed at one's head. But I may be wrong.

Steve Sailer writes:

"No one is entitled to question actions taken with a gun aimed at one's head."

DeLong is perfectly free to avoid the topic of IQ on his blog. Instead, he goes out of his way to spread disinformation and to encourage the harassment of heretics.

Why did so many so enthusiastically sign up as auxiliaries of the Thought Police?

Because it's fun.

The psychology of those who rushed to attack Watson was memorably outlined in Orwell's 1984, when the interrogator O'Brien explains to his prisoner Winston Smith the exciting future envisaged by the Party:

"Always, at every moment, there will be the thrill of victory, the sensation of trampling on an enemy who is helpless. If you want a picture of the future, imagine a boot stamping on a human face—for ever. …The heretic, the enemy of society, will always be there, so that he can be defeated and humiliated over again. … The more the Party is powerful, the less it will be tolerant: the weaker the opposition, the tighter the despotism. Goldstein and his heresies will live for ever. … Always we shall have the heretic here at our mercy, screaming with pain, broken up, contemptible—and in the end utterly penitent, saved from himself, crawling to our feet of his own accord."

TGGP writes:

The reason we do not see regression to the mean with IQ is that human procreation is not random.
Even with non-random mating, we should still expect regression to the mean. If Andre Agassi marries Venus Williams, their kids are probably not going to be as good at tennis as they were, even though they will likely be better than average. Both people are flukes, statistically unlikely outliers in the field (by that I mean tennis ability in the general population). In the random pairing of genes that will produce their children, it is unlikely lightning will strike yet again (though I admit that did basically happen with both Venus and Serena giving their folks a perfect batting average).

tc writes:

IQ is not the only way that genetic inheritance can affect income - there could be personality, looks, health etc. Later on in the paper, Gintis & Bowles do a heritability calculation of earnings using twins (which is subject to all the usual caveats) and find that genetics accounts for
about a third of the intergenerational correlation.

Now, when it comes to psychological variables such as IQ, the usual approach is to partition the environmental effect into shared and nonshared (based on the similarity of people raised in the same family). This requires twins raised apart or adoptees - hard to find, but there have been a few (e.g. this one on Korean adoptees). All find much lower transmission (sometimes zero) for adoptees compared to biological children.

stefano writes:

The general tone of the comments here seems to be that given that IQ correlates with wealth, that means that people are poor because they are stupid. And since IQ is inherited, their children will be stupid and poor too. It's a fact of life, and "interventions" to change their state is only a waste of money.

What DeLong is hinting to, in my opinion, is that is not the full story. Perhaps (just perhaps) poor people are less stupid than their wealth would imply, and moreover maybe IQ is not 100% inherited, so that their children may be even less stupid. Therefore, some "intervention", like better access to education could do something to improve their lot.

jaim klein writes:

Thanks to Steve Sailer for discretely correcting me: I meant Nobel Prize Prof. Watson, not Watkins. But I could not digest his suggestion of sadism as De Long's motive.

Bruce G Charlton writes:

Stefano - take a look at the data on social class and IQ - you could start here:

The opposite of high IQ is not 'stupid' - and you are misusing this word for rhetorical purposes.

IQ measures a fairly specific cognitive attribute called the g-factor or general intelligence, which is mainly concerned with reasoning ability and speed of learning.

Many other mental factors such as most aspects of personaility (or empathizing versus systemizing ability) are distinct from IQ.

IQ is very important to surviving and thriving in modern societies, and IQ is substantially hereditary; but IQ is not *everything* and nobody is saying it is everything.

jb writes:

We know that the children of smart parents tend to be somewhat smarter than average, but I don't know how much the bell curve is shifted for the kids.

So IQ is one of the inputs into the child's 150k, but not all of it.

Certainly, smarter parents are better informed - they retain info about education and strategies to improve their children's lot in life. This means better preparation for college, better coaching on doing well in school.

So some of the benefit to the child is the parent's ability to work the system, because the parents are smart.

Smart people are also better at investing and managing their money, sticking to budgets, etc. They end up with more cash to pay for college, which helps raise the possibility that the child can afford to go to a quality school.

So, some of the benefit to the child is the parent's ability to manage money, because the parents are smart.

Of course, when you're making a lot of money, you probably have relatively high-powered friends and colleagues. You have a social network that you can use to help your children do well once they're out of college.

This also, can be a benefit of being smart - more connections in more places.

In the end, what would be interesting is to see if there is a strong correlation between the parent's IQ and the children's income. We know that there's a correlation between parent's income and child's income - is that correlation controlled for IQ?

In other words - maybe the fact that the parents are smart is what makes the children more successful than average, even if the child isn't particularly smart.

Chuck writes:

Maybe all DeLong is saying is keep what you earned, not what your parents earned.

Certainly one can see that humans are a social animal and that a parent will often give their children all the advantage they can. What, from the perspective of a kid who will basically recieve nothing from their parents, is fair about that?

Tom West writes:

Why did so many so enthusiastically sign up as auxiliaries of the Thought Police?

Because it's fun.

Garbage. People attacked him because they do not like the idea of living in a society that embodies the principles he espoused.

I would say most such people despise the idea of a racial caste system, but see it as an almost inevitable outcome of a society that fully embraces the idea of non-environmental racial intelligence differences. The point of their attacks is preservation of the current social equilibrium that has provided a fairly fruitful culture (with many flaws) for the last half-century or so.

(To forestall the obvious - an almost inevitable outcome does not mean a logical outcome, it means an almost inevitable one)

MarkT writes:

I find that any argument that requires justifying why one person has more than the average is very difficult to sustain. (I do think it seems to become easier to justify differences when you have sets of specific individuals to compare.) But more importantly, it is also very hard to sustain the argument DeLong implies but cagily avoids presenting, that, because the difference versus the average cannot be justified, eliminating it can be justified. I think this is because questions of status and outcome within a society really lack a moral component when you examine them. They are what they are. You can try to change them for various reasons but the only reasons that can be defended logically are equally positivist ("the society has decided to do it this way instead") or purely immoral or amoral utilitarian arguments, like "it will stave off civil unrest." So I recommend calling on the opponent to justify from start to fhinsh with coherent reasoning, his case for change of outcomes. I have yet to see anyone who could.

Brad DeLong writes:

May I ask that people please go and read Bowles and Gintis (2002), "The Inheritance of Inequality"? It is not hard.

Comments for this entry have been closed
Return to top