Links this week:
And no posts this week, although with Robert Frank joining Twitter, here are three book reviews from the vault:
Links this week:
And if you missed them, my posts this week:
Over four years since release of the working paper (and two and half years since I posted about it), Henrik Cronqvist and Stephan Siegel’s paper The Origin of Savings Behavior has been published in the Journal of Political Economy (follow the working paper link for an ungated copy). The abstract is as follows:
Analyzing the savings behavior of a large sample of identical and fraternal twins, we find that genetic differences explain about 33 percent of the variation in savings propensities across individuals. Individuals are born with a persistent genetic predisposition to a specific savings behavior. Parenting contributes to the variation in savings rates among younger individuals, but its effect decays over time. The environment when growing up (e.g., parents’ wealth) moderates genetic effects. Finally, savings behavior is genetically correlated with income growth, smoking, and obesity, suggesting that the genetic component of savings behavior reflects genetic variation in time preferences or self-control.
As I posted last time, the finding is unsurprising and matches findings from behavioural genetics about other traits. Genes matter, heritability increases with age, family environment has little influence and there is a large non-shared environmental effect.
That it takes so long for economics papers to be published makes me thankful for the practice of releasing working papers. We’ve known of and have been able to talk about this result for several years now. It’s always interesting to see the multiple waves of press and attention as different audiences first become aware of a paper at different stages of its life.
But this does have some perverse effects, particularly across disciplinary boundaries. I recently received a referee report suggesting we had neglected some recent literature. Yet the main omission, a paper published in the Proceedings of the Royal Society B, cites the working paper on which our submission was based. Our working paper has been out long enough for people in fields with a faster track to publication to cite it and be published themselves.
So, if we reference that paper, we create a circular set of citations. I’ve spent plenty of time over the last few years following citation chains that do not ultimately establish the point claimed, but I haven’t ended up back where I started too many times yet.
The Economist discusses a new OECD report on a growing gender gap in schools:
It is a problem that would have been unimaginable a few decades ago. Until the 1960s boys spent longer and went further in school than girls, and were more likely to graduate from university. Now, across the rich world and in a growing number of poor countries, the balance has tilted the other way. …
The reversal is laid out in a report published on March 5th by the OECD, a Paris-based rich-country think-tank. Boys’ dominance just about endures in maths: at age 15 they are, on average, the equivalent of three months’ schooling ahead of girls. In science the results are fairly even. But in reading, where girls have been ahead for some time, a gulf has appeared. In all 64 countries and economies in the study, girls outperform boys. The average gap is equivalent to an extra year of schooling.
The gap is particularly stark at the bottom, with teenage boys “50% more likely than girls to fail to achieve basic proficiency in any of maths, reading and science.” While Larry Summers got crucified for referring to male-female differences in variation in different human traits, the higher rates of low proficiency among boys reflects the often ignored consequence of Summers’s statement – more males at the bottom.
The article notes other explanations for the gap such as time doing homework, attitude to school and reading at home, but one of the more interesting explanations might be discrimination:
Perhaps because they can be so insufferable, teenage boys are often marked down. The OECD found that boys did much better in its anonymised tests than in teacher assessments. The gap with girls in reading was a third smaller, and the gap in maths—where boys were already ahead—opened up further. In another finding that suggests a lack of even-handedness among teachers, boys are more likely than girls to be forced to repeat a year, even when they are of equal ability.
There are likely ways to improve test outcomes for some of the lowest performing students. Anonymised tests could be one.
But will the net result of interventions to increase male or female school performance open or close the gap? Making the environment the same for everyone will exacerbate innate differences. Different teaching environments for boys and girls may maximise performance, but again, it’s not clear what size the gap would be under that arrangement (Or teaching could be adjusted based on traits such as IQ and big five personality traits – females tend to score higher on agreeableness and conscientiousness – removing direct discrimination, but effectively treating boys and girls differently on average)
The advantage of females over males continues through to tertiary education:
In the OECD women now make up 56% of students enrolled, up from 46% in 1985. By 2025 that may rise to 58%.
Even in the handful of OECD countries where women are in the minority on campus, their numbers are creeping up. Meanwhile several, including America, Britain and parts of Scandinavia, have 50% more women than men on campus. …
According to the OECD, the return on investment in a degree is higher for women than for men in many countries, though not all.
So why does the income gap persist in the workforce? The Economist article closes with research by Claudia Goldin:
In a recent paper in the American Economic Review Ms Goldin found that the difference between the hourly earnings of highly qualified men and their female peers grows hugely in the first 10-15 years of working life, largely because of a big premium in some highly paid jobs on putting in long days and being constantly on call. On the whole men find it easier than women to work in this way.
It is not easy to develop a policy response to this, nor for a business to try to close the gap without trade-offs. One possibility is gender quotas, although this may largely operate to the advantage of women who do not take time off for family reasons.
Another option is to require men to take paternity leave when they have children. But this penalises both men and women who have children. I am all for people bearing the costs of their children, but a policy to exacerbate those costs in the workplace seems misplaced.
Another question around this option is why time out of the workplace relating to children is specifically targeted. What of those who want to travel, study or volunteer for charity, each of which surely come with costs?
A less prescriptive response is reflected in the closing part of Goldin’s paper, where she writes:
The last chapter must be concerned with how worker time is allocated, used and remunerated and it must involve a reduction in the dependence of remuneration on particular segments of time. It must involve greater independence and autonomy for certain types of workers and the ability of workers to substitute seamlessly for each other. Flexibility at work has become a prized benefit but flexibility is of less value if it comes at a high price in terms of earnings. The various types of temporal flexibility require changes in the structure of work so that their cost is reduced.
Goldin mentions that some jobs necessarily involve long hours and availability at all times. These professions will be resistant to a closing of the gap. However, for jobs that do not require long hours, how easy it is to replace time-based measures? Outputs in many of those jobs are credence goods (think of a lot of output from consulting or even law). Hours worked may be the only tangible measure the employers and clients have.
Links this week:
And if you missed them, my posts this week:
I have been working through The Behavioral Foundations of Public Policy, edited by Eldar Shafir, and have mixed views so far. As I go through, I will note some interesting points.
The opening substantive chapter by Curtis Hardin and Mahzarin Banaji is on bias – and particularly implicit bias. Implicit biases are unconscious negative (or positive) attitudes towards a person or group. Most people who claim (and believe) they are not biased because they don’t show explicit bias will nevertheless have implicit bias that affects their actions.
There is no shortage of tests out there on implicit bias (here’s one set, although you have to fill out a set of surveys before you get to play) and they consistently show that implicit bias exists. Even when you know it is occurring, it’s hard to overcome. Playing with the tests when writing this post, I came up with a strong automatic preference for thin over fat people.
As the chapter is in a book on public policy, it turns to how policy makers should deal with implicit bias. It has a generally optimistic tone about the potential to reduce implicit bias – one that I don’t necessarily share from a public policy perspective – so the paragraphs that stood out for me indicated how complicated any plans to intervene would be.
Research also suggests that the interpersonal regulation of implicit prejudice is due in part to a motivation to affiliate with others who are presumed to hold specific values related to prejudice, as implied by shared reality theory (e.g., Hardin and Conley, 2001). For example, participants exhibited less implicit racial prejudice in the presence of an experimenter wearing a T-shirt with an antiracism message than a blank T-shirt, but only when the experimenter was likeable (Sinclair et al., 2005). When the experimenter was not likeable, implicit prejudice was actually greater in the presence of the ostensibly egalitarian experimenter. In addition, social tuning in these experiments was mediated by the degree to which participants liked the experimenter, providing converging evidence that interpersonal dynamics play a role in the modulation of implicit prejudice, as they do in other dimensions of social cognition (Hardin and Conley, 2001; Hardin and Higgins, 1996).
As regards public and personal policy, these findings suggest that a public stance for egalitarian values is a double-edged sword, and a sharp one at that. Although it may reduce implicit prejudice among others when espoused by someone who is likeable and high in status, it may backfire when espoused by someone who is not likeable or otherwise of marginal status. This finding suggests one mechanism by which common forms of “sensitivity training” in service of the reduction of workplace sexism and racism may be subverted by interpersonal dynamics, however laudable the goals.
I’m guessing that in many scenarios government and its agents would fall into the “not likeable or otherwise of marginal status” category.
What is Evonomics?
Evonomics is an online magazine and intellectual movement built on the pillars of complexity science and evolutionary principles, and includes key insights from the synthesis that has slowly been growing across disciplines in areas like behavioral, experimental, institutional, and ecological economics. The magazine showcases the new scientific foundations for human nature and society and demonstrates its relevance to contemporary economic and political issues.
Sign up on the Evonomics site for updates as it gears up for launch. You can also follow Evonomics on twitter.
Links this week:
And if you missed them, my posts this week:
At Stumbling and Mumbling, Chris Dillow writes:
[M]aybe some lefties do reject the heritability of IQ on ideological grounds. I want to make another point – that there’s no need for them to do so. You can accept that IQ (or ability generally) is heritable and still be a strong egalitarian.
I say this because of a simple principle: luck egalitarianism. This says that inequalities are unjust if they are due to circumstances beyond one’s control. If we grant that ability is inherited, then differences in ability are obviously a matter of luck. Insofar as these give rise to inequalities of income, a luck egalitarian can thus claim they are unjust.
…
That said, there is a sort of leftie who would be discombobulated by the heritability of ability. I’m thinking of that sort, like Tessa Jowell, who – in their optimism about the malleability of humankind – think that education can significantly reduce inequality.
But that leftism isn’t mine. I agree with Ed Smith that social mobility – even if it could be achieved – is an unattractive ideal. It’s no substitute for a just society.
Peter Singer made a related argument in A Darwinian Left: Politics, Evolution, and Cooperation, suggesting that the left needs to incorporate an updated understanding of the malleability of human nature into its framework – although Singer’s arguments focused on our tendency to cooperate.
Arnold Kling suggests the discombobulation of some on the left comes from the need to maintain a narrative:
In the three-axes model, progressives want to squeeze every issue into an oppressor-oppressed narrative. To suggest that ethnic groups differ in average income for reasons other than oppression would be to weaken that narrative. So even if from a policy perspective a belief in heritability is tolerable, from a narrative perspective a book like The Bell Curve represents a huge threat.
My sense is that this produces a great deal of cognitive dissonance on the left. I have many friends on the left, and I do not know a single one who would instinctively deny the heritability of intelligence. On the other hand, they have been instructed to regard Murray and Herrnstein as vile racists.
My own experience is that plenty of people are willing to argue whether behavioural traits are heritable. I sense Kling’s narrative story is part of the reason, but I also suggest that it comes from a general unwillingness of people to concede any points in a debate. (Does this “bias” have a name – or is this just a manifestation of confirmation bias or a desire to reduce cognitive dissonance?)
Take arguments about climate change. Many libertarians or conservatives fight at every step of the way – the earth is not warming, the warming is not caused by human activity, the warming will be mild, the warming will be beneficial – all this before they get to arguments about the costs and benefits of different policy responses. Yet, whether warming is occurring or harmful would not seem to be a core part of the libertarian philosophy. Debates about heritability have a similar character.
I have just finished Leo Tolstoy’s War and Peace, and along the way marked a couple of passages.
The first two fit with the story that much behavioural science is formalisation of common sense. First, hindsight bias:
In historical works on the year 1812 French writers are very fond of saying that Napoleon felt the danger of extending his line, that he sought a battle and that his marshals advised him to stop at Smolensk, and of making similar statements to show that the danger of the campaign was even then understood. Russian authors are still fonder of telling us that from the commencement of the campaign a Scythian war plan was adopted to lure Napoleon into the depths of Russia, and this plan some of them attribute to Pfuel, others to a certain Frenchman, others to Toll, and others again to Alexander himself—pointing to notes, projects, and letters which contain hints of such a line of action. But all these hints at what happened, both from the French side and the Russian, are advanced only because they fit in with the event. Had that event not occurred these hints would have been forgotten, as we have forgotten the thousands and millions of hints and expectations to the contrary which were current then but have now been forgotten because the event falsified them. There are always so many conjectures as to the issue of any event that however it may end there will always be people to say: “I said then that it would be so,” quite forgetting that amid their innumerable conjectures many were to quite the contrary effect.
Next, confirmation bias:
But these were only suppositions, which seemed important to the younger men but not to Kutuzov. With his sixty years’ experience he knew what value to attach to rumors, knew how apt people who desire anything are to group all news so that it appears to confirm what they desire, and he knew how readily in such cases they omit all that makes for the contrary.
Tolstoy also knew something of statistics and selection bias:
One would have thought that under the almost incredibly wretched conditions the Russian soldiers were in at that time—lacking warm boots and sheepskin coats, without a roof over their heads, in the snow with eighteen degrees of frost, and without even full rations (the commissariat did not always keep up with the troops)—they would have presented a very sad and depressing spectacle.
On the contrary, the army had never under the best material conditions presented a more cheerful and animated aspect. This was because all who began to grow depressed or who lost strength were sifted out of the army day by day. All the physically or morally weak had long since been left behind and only the flower of the army—physically and mentally—remained.
And the value of medicine in those days:
He had what the doctors termed “bilious fever.” But despite the fact that the doctors treated him, bled him, and gave him medicines to drink, he recovered.
