I have been listening to a podcast of an excellent talk by David Spiegelhalter on “Thinking and Feeling About Risk”. The video of the lecture is below.

The lecture covers a lot of interesting material – from the misrepresentation of cancer screening statistics to bicycle helmets – and I recommend listening to or watching the whole thing.

One interesting point was about the presentation of estimates of GDP growth. The Bank of England produces quarterly forecasts of GDP growth, but when they present them graphically, they don’t include their central estimate. The May 2015 graphic is to the left.

The story Spiegelhalter tells is that providing a central estimate leads to everyone focusing on that, rather than the considerable range of uncertainty. He shows a similar example where removing the central line for prediction of hurricane movements results in people who sit within the “cone of uncertainty” taking the risk to them more seriously.

I see another benefit of this GDP growth forecast chart. It effectively communicates that The Bank of England has little idea what the level of growth will be. In fact, there is a large range for what they believe the growth rate was. If people are going to insist on publishing forecasts such as this (whatever their merits), the more people who come to understand that we have no idea, the better.

Michelle Meyer and Christopher Chabris write:

Companies — and other powerful actors, including lawmakers, educators and doctors — “experiment” on us without our consent every time they implement a new policy, practice or product without knowing its consequences. When Facebook started, it created a radical new way for people to share emotionally laden information, with unknown effects on their moods. And when OkCupid started, it advised users to go on dates based on an algorithm without knowing whether it worked.

Why does one “experiment” (i.e., introducing a new product) fail to raise ethical concerns, whereas a true scientific experiment (i.e., introducing a variation of the product to determine the comparative safety or efficacy of the original) sets off ethical alarms?

In a forthcoming article in the Colorado Technology Law Journal, one of us (Professor Meyer) calls this the “A/B illusion” — the human tendency to focus on the risk, uncertainty and power asymmetries of running a test that compares A to B, while ignoring those factors when A is simply imposed by itself.

….

[A]s long as we permit those in power to make unilateral choices that affect us, we shouldn’t thwart low-risk efforts, like those of Facebook and OkCupid, to rigorously determine the effects of those choices. Instead, we should cast off the A/B illusion and applaud them.

Amen.

I posted this paper on SSRN a few months ago, but neglected to blog about it – I’ve written (with my supervisors) a review of the literature incorporating evolutionary theory into economics. The abstract:

The Evolutionary Foundations of Economics

As human traits and preferences were shaped by natural selection, there is substantial potential for the use of evolutionary biology in economic analysis. In this paper, we review the extent to which evolutionary theory has been incorporated into economic research. We examine work in four areas: the evolution of preferences, the molecular genetic basis of economic traits, the interaction of evolutionary and economic dynamics, and the genetic foundations of economic development. These fields comprise a thriving body of research, but have significant scope of further investigation. In particular, the growing accessibility of low cost molecular data will create more opportunities for research on the relationship between molecular genetic information and economic traits.

The paper is fairly flat in tone as I wrote it as the introductory review chapter for my thesis. If you’re familiar with the blog, you will have read some more critical pieces on the papers covered in my article before. Links to some of those critiques can be found down the bottom of my evolutionary biology and economics reading list page.

And the disclaimer – this paper isn’t about “evolutionary economics” in the way that term is typically used. I’m interested in the biological angle:

The subject matter of this paper needs to be distinguished from what is commonly called “evolutionary economics”. Evolutionary economics uses biological concepts, such as natural selection, and applies them to the dynamics of firms, business processes and institutions. The economy is seen as a complex adaptive system in which innovation and change are central considerations. The origin of evolutionary economics is often traced to Veblen (1898), and was revived by Alchian (1950) and later Nelson and Winter (1982), whose seminal work inspired a vast literature. The subject matter of this paper differs from evolutionary economics in that we focus on human biology rather than seeking to apply a biological analogy to higher levels such as firms. This paper is about the application of evolutionary biology to economic processes at the level of humans and their genes and their interactions at the population level.

Links this week:

1. Storytelling about famous experiments tends to go a bit askew.
2. Noah Smith takes on Deirdre McCloskey.
3. Chimps on the drink.
4. A review of Richard Thaler’s ‘Misbehaving: The Making of Behavioural Economics’.
5. The gender gap in tech.

And if you missed them, my posts from the last week:

1. MSiX 2015 is on July 30 in Sydney, and features yours truly.
2. Humans cause accidents.

The below passage is from a neat article on how mistakes can save lives.

CRM [crew resource management] as born of a realisation that in the late 20th century the most frequent cause of crashes wasn’t technical failure, but human error. Its roots go back to the Second World War, when the US army assigned a psychologist called Alphonse Chapanis to investigate a curious phenomenon. B-17 bombers kept crashing on to the runway on landing, even though there were no apparent mechanical problem with the planes. Rather than blaming the pilots, Chapanis pointed to the instrument panel. The lever to control the landing gear and the lever that operated the flaps were next to each other. Pilots, weary after long flights, were confusing the two, retracting the wheels and causing the crash. Chapanis suggested attaching a wheel to the handle of the landing lever and a triangle to the flaps lever, making each easily distinguishable by touch alone. Problem solved.

Chapanis had recognised that human beings’ propensity to make mistakes when they are tired is much harder to fix than the design of levers. His deeper insight was that people have limits, and many of their mistakes are predictable effects of those limits. That is why the architects of CRM defined its aim as the reduction of human error, rather than pilot error. Rather than trying to hire or train perfect pilots, it is better to design systems that minimise or mitigate inevitable human mistakes.

In the 1990s, a cognitive psychologist called James Reason turned this principle into a theory of how accidents happen in large organisations. When a space shuttle crashes or an oil tanker leaks, our instinct is to look for a single, “root” cause. This often leads us to the operator: the person who triggered the disaster by pulling the wrong lever or entering the wrong line of code. But the operator is at the end of a long chain of decisions, some of them taken that day, some taken long in the past, all contributing to the accident; like achievements, accidents are a team effort. Reason proposed a “Swiss cheese” model: accidents happen when a concatenation of factors occurs in unpredictable ways, like the holes in a block of cheese lining up.

James Reason’s underlying message was that because human beings are fallible and will always make operational mistakes, it is the responsibility of managers to ensure that those mistakes are anticipated, planned for and learned from. Without seeking to do away altogether with the notion of culpability, he shifted the emphasis from the flaws of individuals to flaws in organisation, from the person to the environment, and from blame to learning.

The science of “human factors” now permeates the aviation industry. It includes a sophisticated understanding of the kinds of mistakes that even experts make under stress.

I recommend reading the full article. Among other things, it has a lot of interesting material about mistakes in medical settings.

The 2015 Marketing Science Ideas Xchange – MSiX – has been announced for 30 July in Sydney. As it says in the blurb, MSiX “is dedicated to exploring how brands can benefit from the interface between behavioural science and marketing.”

The headline speaker is Michael Norton, Harvard professor, author of Happy Money: The Science of Happier Spending and developer of the first set of experiments on the IKEA effect (that last point is the reason I knew his name when I heard he would be speaking).

The rest of the speakers and further detail on the conference are here – with the speaking line-up including me. I’ll be talking about how behavioural economics (science) could benefit from a good dose of evolutionary biology, and how that evolutionary lens can be valuable in understanding consumer behaviour.

Last year’s event – headlined by Rory Sutherland (who linked me into this MSiX world) – was a pretty good day, and this year looks promising too.

And here’s Norton doing the TED thing.

Links this week:

1. Nobody is doing more to save the NHS than the “drinkers, smokers or fatties”.
2. Some bashing of the benefits of education: Did schooling drive the industrial revolution? Against tulip (education) subsidies.
3. Is war on the wane?
4. The Dead Sea lives.

And if you missed them, my posts from the last week:

1. Why family friendly policies backfire.
2. The winner effect in humans.

I am using some material from John Coates’s excellent The Hour Between Dog and Wolf for a presentation I am giving next week, and decided it was worth sharing here:

During moments of risk-taking, competition and triumph, of exuberance, there is one steroid in particular that makes its presence felt and guides our actions – testosterone. At Rockefeller University I came across a model of testosterone-fuelled behaviour that offered a tantalising explanation of trader behaviour during market bubbles, a model taken from animal behaviour called ‘the winner effect’.

In this model, two males enter a fight for turf or a contest for a mate and, in anticipation of the competition, experience a surge in testosterone, a chemical bracer that increases their blood’s capacity to carry oxygen and, in time, their lean-muscle mass. Testosterone also affects the brain, where it increases the animal’s confidence and appetite for risk. After the battle has been decided the winner emerges with even higher levels of testosterone, the loser with lower levels. The winner, if he proceeds to a next round of competition, does so with already elevated testosterone, and this androgenic priming gives him an edge, helping him win yet again. Scientists have replicated these experiments with athletes, and believe the testosterone feedback loop may explain winning and losing streaks in sports. However, at some point in this winning streak the elevated steroids begin to have the opposite effect on success and survival. Animals experiencing this upward spiral of testosterone and victory have been found after a while to start more fights and to spend more time out in the open, and as a result they suffer an increased mortality. As testosterone levels rise, confidence and risk-taking segue into overconfidence and reckless behaviour.

Could this upward surge of testosterone, cockiness and risky behaviour also occur in the financial markets? This model seemed to describe perfectly how traders behaved as the bull market of the nineties morphed into the tech bubble. When traders, most of whom are young males, make money, their testosterone levels rise, increasing their confidence and appetite for risk, until the extended winning streak of a bull market causes them to become every bit as delusional, overconfident and risk-seeking as those animals venturing into the open, oblivious to all danger. The winner effect seemed to me a plausible explanation for the chemical hit traders receive, one that exaggerates a bull market and turns it into a bubble. The role of testosterone could also explain why women seemed relatively unaffected by the bubble, for they have about 10 to 20 per cent of the testosterone levels of men.

Coates tested this on a London trading floor:

I set up an experiment on the trading floor of a mid-sized firm in the City of London. The floor employed 250 traders, all but three of whom were men. They were all engaged in high-frequency trading, … meaning they bought and sold securities, sometimes in sizes ranging up to $1 or $2 billion, but held their bets only for a matter of hours or minutes, sometimes mere seconds. They therefore occupied the same market niche as the black boxes.

These traders were therefore up against some of the world’s most sophisticated and well-capitalised competitors. They lacked the large capital base and informational advantages of the flow traders at the big banks, and the deep pools of capital and inhuman processing speeds of the black boxes. Yet they were astonishingly successful: David against Goliath, John Connor against the Terminator. In fact they were some of the best traders I have ever seen: highly disciplined, consistent, and profitable.

I sampled testosterone from these traders and recorded P&L over a two-week period. What we found was that their testosterone levels were significantly higher on days when they made an above-average profit. More intriguing, though, was what we found when we looked at testosterone levels in the morning, because these predicted how much money the traders would make in the afternoon. When the traders’ morning testosterone levels were high, they went on to make a lot more money in the afternoon than they did on days when their morning testosterone levels were low. Moreover, the difference in P&L between high- and low-testosterone days was large, amounting in statistical terms to one full standard deviation, a difference that if annualised could amount for some of the traders to over £500,000 in pay.

Last month a NYT article by Claire Cain Miller documented some of the backfires associated with family friendly policies. For instance:

Unlike many countries, the United States has few federal policies for working parents. One is the Family and Medical Leave Act of 1993, which provides workers at companies of a certain size with 12 weeks of unpaid leave.

Women are 5 percent more likely to remain employed but 8 percent less likely to get promotions than they were before it became law, according to an unpublished new study by Mallika Thomas, who will be an assistant professor of economics at Cornell University. …

The child-care law in Chile, the most recent version of which went into effect in 2009, was intended to increase the percentage of women who work, which is below 50 percent, among the lowest rates in Latin America. It requires that companies with 20 or more female workers provide and pay for child care for women with children under 2, in a location nearby where the women can go to feed them.

It eases the transition back to work and helps children’s development, said María F. Prada, an economist at the Inter-American Development Bank and lead author of a new study on the effects of the law. But it has also led to a decline in women’s starting salaries of between 9 percent and 20 percent.

I am not sure there exists a family friendly policy that doesn’t “backfire” in some dimension. These policies tend to have multiple objectives and there are trade-offs between these objectives. That holds even if these policies are publicly funded and place no burden on employers.

First, those objectives. Policy makers want women to be able to have children while having a career. They are pronatalist. They want women (and sometimes men) to be able to take time out of the workforce to care for their children. They want high-quality care for children. And they want men and women to have the same level of pay. You can’t have them all.

Consider this trade-off in the light of the AER article by Claudia Goldin suggesting gender pay disparities are because women find it difficult  to work the long hours many jobs now require. How can you facilitate those long hours?

One option is to discourage having a family, which tends not to be the preferred route (unless you are a tech company offering to subsidise egg freezing services).

Alternatively, you can provide subsidised or free childcare, which will reduce the cost of having a family and enable the mother to return to work faster. But this also increases the attractiveness of having a family, which increases the number of children and constrains potential hours worked. And even with free childcare, the mother will likely take some time off.

Alternatively, take Paul Seabright’s argument in War of the Sexes that much of the wage gap is due to gender differences in networks. When you take time out of the workforce, your networks suffer. Women are more likely to take time out of the workforce.

One of Seabright’s ideas to counter this is compulsory paternity leave. Gender neutral parental leave arrangements are also floated at the end of the NYT article. However, compulsory paternity leave effectively converts the penalty on women into a penalty on families. The gap will be between those with and without children. And now that the birth may be timed to fit in with the man’s career, it is possible that the birth may be timed even less suitably for the woman.

Links this week:

1. Two perspectives on the chocolate study – Scott Alexander and Andrew Gelman. I would say that the chocolate study didn’t tell us anything that we didn’t already know.
2. Self-deluded leaders.
3. The education myth.
4. If only chimpanzees had ovens. HT: Tyler Cowen

And if you missed them, my posts from the last week:

1. Ration information. Avoid news.
2. Measurement error on 23andme.
3. Merton on retirement incomes.