Why does the son rise?

John Cochrane recently gave a speech where one of the main threads involved talking down the importance of income and wealth inequality. Poverty, and generally not having as much as we would like are bad, he says, but is there anything bad about inequality per se? That is: is there at least one respect in which things would be better if some people who are very well off were made worse off? He argues that there isn’t, or if there is, it is of only trivial importance, and outweighed by all of the costs of actually ‘doing something’ about inequality.

In a response on Bloomberg View, Noah Smith argued that economists should respect people’s actual preferences, and since people show strong preferences against wealth and income inequality, we should respect them. He uses the example of how people prefer to take nothing over free money when they are made offers they perceive as ‘unfair’ in the Ultimatum Game. On top of that, he says, inequality leads to socio-political unrest, which we can all agree is very bad and costly, citing a 1993 paper.

Finally, Tony Yates adds some extra arguments on his blog. He says luck has a big role to play in success, but success can also buy some of the non-luck factors in success (e.g. education), meaning that it can ‘set off path dependence’—according to Yates this can lead to inefficient outcomes by distorting the allocation of talent. He says inequality reduces public good provision (e.g. education—although I’m not sure that is really a public good). And he says that inequality might make ‘crony capitalism’ more likely.

I’ve written twice about equality before: once saying that Rawlsian-style justice demands inequality of wealth/income in certain very relevant circumstances; another time arguing that Hayekian-style information economics militates towards equality of wealth. There are lots more things to say in this debate, but here I intend to take issue only with one of Yates’ claims: the idea that luck + path-dependence means inequality is passed down through the generations (I can’t see why exactly he thinks this distorts the allocation of talent, but here I’m only questioning the mechanism).

Luck is certainly a huge factor in success. And people do pay big money for better education to try and make sure their kids are more likely to succeed. But does this work? Let’s look at some studies. Random selection into a better school in Beijing has no effect, random selection into a better school in Chicago has close to no effect, random selection into a better Kenyan school has no effect, nor does it in Missouri, nor in New York City. Once you control for student characteristics, Australian private schools didn’t outperform state schools on the 2009 PISA. Conscription into extra education didn’t much affect life outcomes in late 1970s France. In 1950s England, going to an elite school made no difference to a youth’s job market outcomes. The literature is huge and there are many many more examples.

And other literatures point to the same conclusion. For example, we now know that the heritability of intelligence increases through life (to hit around 50-90% in adulthood), while ‘shared environment’—upbringinging, parental inputs and schooling—falls to around zero. This is supported by traditional twin studies, twins reared apart studies, adoption studies, and now whole-genome analysis.

So it should not be surprising that it’s actually really really really really hard to make sure your descendants stay rich with the proceeds of luck. In fact, we know that that’s not why the descendants of the rich often are rich because we have a couple of pretty good experiments showing it! For example:

We track descendants of those eligible to win in Georgia’s Cherokee Land Lottery of 1832, which had nearly universal participation among adult white males. Winners received close to the median level of wealth – a large financial windfall orthogonal to parents’ underlying characteristics that might have also affected their children’s human capital. Although winners had slightly more children than non-winners, they did not send them to school more. Sons of winners have no better adult outcomes (wealth, income, literacy) than the sons of non-winners, and winners’ grandchildren do not have higher literacy or school attendance than non-winners’ grandchildren. This suggests only a limited role for family financial resources in the formation of human capital in the next generations in this environment and a potentially more important role for other factors that persist through family lines.

The same is true for modern lottery winners—the truest example of pure luck in success. And it took only two generations for the descendants of slaves to catch up with the much more advantaged & wealthy free blacks. Whether rich parents split an inheritance between eighteen kids or one, their grandkids are equally rich. Basically luck mixed with path dependance explains almost nothing.

Five intriguing papers I discovered this week

In what might become a recurring feature, I am going to summarise the findings of a few research papers, potentially of interest to ASI blog readers, that were either first released this week, first published this week, or first come upon by me this week.

1. Fernandes, D., Lynch, J. G., and Netemeyer, R. G., “Financial Literacy, Financial Education and Downstream Financial Behaviors” (Jan 2014)

This paper is a large meta-analysis of 168 other papers, which in turn refer to 201 different studies and experiments. They find that at least 99.9% of financial behaviour in life cannot be explained by differences in financial education, or conversely at most 0.1% of the difference in people’s financial decision-making and choices is down to education interventions designed to improve their financial literacy. In the words of their abstract: “even large interventions with many hours of instruction have negligible effects on behavior 20 months or more from the time of intervention”.

While other correlational studies appear to show some relationship between financial behaviour and educational schemes (i.e. one explaining more than 0.1% of the variance between individuals) they explain that this is only because those typically getting financial education already have various psychological traits associated with careful management of finances. They therefore suggest that big schemes designed to improve lifetime financial decisionmaking are futile and a waste of money; the best we can hope for is “just in time” interventions, perhaps at the point of financial transactions, that are more likely to be taken in and not forgotten.

2. Wang, M-T., Eccles, J. S., and Kenny, S., “Not Lack of Ability but More Choice: Individual and Gender Differences in Choice of Careers in Science, Technology, Engineering, and Mathematics” (May 2013)

In this paper the authors find that a substantial fraction of the male-female “gap” in STEM (Science, Technology, Engineering and Medicine) fields can be explained by the fact that women who are talented at maths tend to also have high verbal skills, skills that mathematically talented men are much more likely to lack. This means they have a wider range of choices available to them, and also possibly identify less closely with maths as part of their personality, and it is this choice not to pursue STEM further that drives the gap, rather than, for example, discrimination in the area or a perceived unfriendly atmosphere.

3. Liu, S., Huang, J. L., and Wang, M., “Effectiveness of Job Search Interventions: A Meta-Analytic Review” (Mar 2014)

Liu, Huang and Wang found, reviewing 47 different experiments testing if schemes “teaching job search skills, improving self-presentation, boosting self-efficacy, encouraging proactivity, promoting goal setting, and enlisting social support” could boost the unemployed’s chances of getting a job. In fact, on average those in the treatment groups—i.e. those actually subject to the intervention, and not in the control group—were 2.67 times more likely to get a job. Since the studies all used randomness of quasi-randomness to assign treatment, this suggests, they say, that schemes that develop skills and self-motivation can be effective. However, the schemes were more likely to help the young than the old, the short-term unemployed than the long-term unemployed, and job-seekers with special needs, as compared to the population at large.

4. Karwowski, M., and Lebuda, I., “Digit ratio predicts eminence of Polish actor” (Jul 2014)

In a slightly surprising study, the two authors looked at 98 Polish actors, both male and female, and compared the ratio between their second and fourth digits on their hand (a measure of prenatal testosterone exposure) and their productivity and fame. For both men and women, even controlling for age, a higher ratio predicted more pre-eminence.

5. Aisen, A., and Veiga, F. J., “How Does Political Instability Affect Economic Growth?” (Jan 2011)

In a classic example where economists do extensive research to tell us what we already know, this IMF paper from 2011 shows us how bad political instability is for economic growth. Actually, the paper is a great one because it allows us to estimate the size of the impact of different political elements on instability and then the size of instability’s own impact on economic aggregates.

Their findings are highly interesting: whereas primary school enrolment has a pitifully small impact on economic growth, and the impact of investment, economic freedom and the security of property rights comes out quite small, violence, political instability and cabinet changes have substantial negative effects, as does, surprisingly, population growth. And while the most productive regions in Europe are the most ethnically diverse, in this study ethnic homogeneity is very strongly associated with growth. Of course, the conclusions of the paper—that countries should address the root causes of political instability—are much easier said than done!

Equality: as cheap as 50p?

Peter Oborne argues that Ed Balls’ pledge to raise the 45p top tax rate back up to 50p is a good idea. While the extremely high marginal rates (top main rate 83%, plus a 15% surcharge for “unearned income”) of the 1960s and ’70s might have been driven by “socialist envy”, George Osborne’s dropping the rate from 50p to 45p in was “profoundly shaming and offensive”, Oborne contends. This is because, echoing Stanley Baldwin and his brand of Toryism, the conservatives should represent the whole country, not the rich or any other factional interest.

Apparently the Coalition has “devoted a great deal of effort to lowering the living standards of the poor”, and this move to “make the rich richer” is inappropriate when the poor are getting poorer. I contend this by arguing that inequality is down to 90s levels under chancellor Osborne, while the worst-off in society are the only group to actually see their living standards improve the since the recession hit. And the (ugly, unpleasant, and regrettable) attitudes that have emerged towards benefits claimants are probably driving government rhetoric in that area, rather than vice versa.

In general, it annoys me when a columnist writes something apparently trading on what everyone just knows. Sometimes the common view is incorrect. Funnily enough, politics is the area where people err most profoundly and with the most regularity. And I would argue that Oborne is trading on falsehoods in his piece; would it still be a coherent argument if it started with the factual premise that inequality in the UK fell back below its 1997-8 low in 2011-12, 0.34 measured by the GINI coefficient? That the top 10% of earners endured the biggest blow to their incomes since the onset of the recession? And that the bottom 10% by income were the only one to see a rise in living standards taking inflation into account? I don’t think so.

The IFS reports I link above predict that by 2015-16 inequality will rise back to roughly its pre-recession level, so perhaps Oborne could refocus his attack on the future inequality Osborne possibly has a hand in. But in all likelihood there is probably little the government can do about inequality over the long-term, caused as it is by very fundamental trends and robust as it is to institutions even such as the USSR’s. Most of the extra inequality since the 60s and 70s has come from couples engaging in much more assortative mating. And very long-term trends are mainly dominated by heritability of social class—those with Norman surnames are 28% more likely than a random sample of similar others to get an Oxford place.

This is not an IQ test

Boris Johnson’s recent CPS speech, which among other things discussed intelligence and inequality, has caused such a storm that David Cameron and George Osborne were forced to distance themselves from his comments. There is an interesting point to be made about how distressingly and obnoxiously ugly this “everything is offensive” aspect of modernity is, where people are punished for neo-blasphemy either by the mob, or even by laws against heresy. These heresies needn’t be correct, and indeed can often be repellent, but the manner of their repression is so depressing and generally grim. But I don’t want to make this point. I don’t even want to talk much about how misguided the IQ ignorers are, as Tom Chivers has already made this point, in his excellent Telegraph blog post.

Instead I want to just point out how different the “IQ test” (three trick questions) Boris was given live on LBC radio was to an actual intelligence measure.

A man builds a house with four sides of rectangular construction each side with southern exposure. A big bear comes along. What is the colour of the bear?

Supposed answer: white, because you can only have four south-facing walls if you’re in the Arctic.

 Take two apples from three apples and what do you have?

Supposed answer: two, because you’ve taken two apples.

 I went to bed at eight in the evening last night and I wound up my clock and set my alarm to sound for nine o’clock in the morning. How many hours sleep did I get?

Supposed answer: one, since wind-up alarm clocks do not discriminate between am and pm.

These questions are ridiculous, meaningless ambiguities and the third one might even be incoherent (how can you wind it until 9am if it doesn’t recognise the distinction between am and pm?). I’d guess they are more likely to test knowledge of the questioner’s intention than anything like g (what scientists in the field call general intelligence).

Have these people seen any actual IQ tests that are used in real research? For example Raven’s Progressive Matrices, where one must discern a pattern from matrices of simple diagrams, or any of the many different tests used in Wechsler’s Adult Intelligence Scale, or for developing minds, Wechsler’s Intelligence Scale for Children.

Funnily enough, I highly suspect that the correlation between answering those “lateral thinking” questions and life outcomes is much weaker than that between IQ measured on the WISC or WAIS. And this linkage, which articles in the Guardian and Observer have not so much disputed (though letters have) as attempted to associate with badness (see Andrew Rawnsley here), is pretty important. In fact, it usually trumps upbringing when it comes to children born out of wedlock, income, number of weeks worked and many other indicators of well-being. Noticing these, combined with not condemning greed as Worse Than Hitler (TM) should not make Boris himself Worse Than Hitler (TM).

And I didn’t even mention what brave and good policies London’s mayor has on housebuilding, migration, and London transport!