behavior

The stochastic overload model

pnrj cognitive science, mathematical models, personal, science and academia , , , , , , , ,

The stochastic overload model

Mar 12 JDN 2460016

The next few posts are going to be a bit different, a bit more advanced and technical than usual. This is because, for the first time in several months at least, I am actually working on what could be reasonably considered something like theoretical research.

I am writing it up in the form of blog posts, because actually writing a paper is still too stressful for me right now. This also forces me to articulate my ideas in a clearer and more readable way, rather than dive directly into a morass of equations. It also means that even if I do never actually get around to finishing a paper, the idea is out there, and maybe someone else could make use of it (and hopefully give me some of the credit).

I’ve written previously about the Yerkes-Dodson effect: On cognitively-demanding tasks, increased stress increases performance, but only to a point, after which it begins decreasing it again. The effect is well-documented, but the mechanism is poorly understood.

I am currently on the wrong side of the Yerkes-Dodson curve, which is why I’m too stressed to write this as a formal paper right now. But that also gave me some ideas about how it may work.

I have come up with a simple but powerful mathematical model that may provide a mechanism for the Yerkes-Dodson effect.

This model is clearly well within the realm of a behavioral economic model, but it is also closely tied to neuroscience and cognitive science.

I call it the stochastic overload model.

First, a metaphor: Consider an engine, which can run faster or slower. If you increase its RPMs, it will output more power, and provide more torque—but only up to a certain point. Eventually it hits a threshold where it will break down, or even break apart. In real engines, we often include safety systems that force the engine to shut down as it approaches such a threshold.

I believe that human brains function on a similar principle. Stress increases arousal, which activates a variety of processes via the sympathetic nervous system. This activation improves performance on both physical and cognitive tasks. But it has a downside; especially on cognitively demanding tasks which required sustained effort, I hypothesize that too much sympathetic activation can result in a kind of system overload, where your brain can no longer handle the stress and processes are forced to shut down.

This shutdown could be brief—a few seconds, or even a fraction of a second—or it could be prolonged—hours or days. That might depend on just how severe the stress is, or how much of your brain it requires, or how prolonged it is. For purposes of the model, this isn’t vital. It’s probably easiest to imagine it being a relatively brief, localized shutdown of a particular neural pathway. Then, your performance in a task is summed up over many such pathways over a longer period of time, and by the law of large numbers your overall performance is essentially the average performance of all your brain systems.

That’s the “overload” part of the model. Now for the “stochastic” part.

Let’s say that, in the absence of stress, your brain has a certain innate level of sympathetic activation, which varies over time in an essentially chaotic, unpredictable—stochastic—sort of way. It is never really completely deactivated, and may even have some chance of randomly overloading itself even without outside input. (Actually, a potential role in the model for the personality trait neuroticism is an innate tendency toward higher levels of sympathetic activation in the absence of outside stress.)

Let’s say that this innate activation is x, which follows some kind of known random distribution F(x).

For simplicity, let’s also say that added stress s adds linearly to your level of sympathetic activation, so your overall level of activation is x + s.

For simplicity, let’s say that activation ranges between 0 and 1, where 0 is no activation at all and 1 is the maximum possible activation and triggers overload.

I’m assuming that if a pathway shuts down from overload, it doesn’t contribute at all to performance on the task. (You can assume it’s only reduced performance, but this adds complexity without any qualitative change.)

Since sympathetic activation improves performance, but can result in overload, your overall expected performance in a given task can be computed as the product of two terms:

[expected value of x + s, provided overload does not occur] * [probability overload does not occur]

E[x + s | x + s < 1] P[x + s < 1]

The first term can be thought of as the incentive effect: Higher stress promotes more activation and thus better performance.

The second term can be thought of as the overload effect: Higher stress also increases the risk that activation will exceed the threshold and force shutdown.

This equation actually turns out to have a remarkably elegant form as an integral (and here’s where I get especially technical and mathematical):

\int_{0}^{1-s} (x+s) dF(x)

The integral subsumes both the incentive effect and the overload effect into one term; you can also think of the +s in the integrand as the incentive effect and the 1-s in the limit of integration as the overload effect.

For the uninitated, this is probably just Greek. So let me show you some pictures to help with your intuition. These are all freehand sketches, so let me apologize in advance for my limited drawing skills. Think of this as like Arthur Laffer’s famous cocktail napkin.

Suppose that, in the absence of outside stress, your innate activation follows a distribution like this (this could be a normal or logit PDF; as I’ll talk about next week, logit is far more tractable):

As I start adding stress, this shifts the distribution upward, toward increased activation:

Initially, this will improve average performance.

But at some point, increased stress actually becomes harmful, as it increases the probability of overload.

And eventually, the probability of overload becomes so high that performance becomes worse than it was with no stress at all:

The result is that overall performance, as a function of stress, looks like an inverted U-shaped curve—the Yerkes-Dodson curve:

The precise shape of this curve depends on the distribution that we use for the innate activation, which I will save for next week’s post.

What would a new macroeconomics look like?

pnrj critique of neoclassical economics, heuristics and biases, inequality, macroeconomics , , , , , , , , , , , , , ,

Dec 9 JDN 2458462

In previous posts I have extensively criticized the current paradigm of macroeconomics. But it’s always easier to tear the old edifice down than to build a better one in its place. So in this post I thought I’d try to be more constructive: What sort of new directions could macroeconomics take?

The most important change we need to make is to abandon the assumption of dynamic optimization. This will be a very hard sell, as most macroeconomists have become convinced that the Lucas Critique means we need to always base everything on the dynamic optimization of a single representative agent. I don’t think this was actually what Lucas meant (though maybe we should ask him; he’s still at Chicago), and I certainly don’t think it is what he should have meant. He had a legitimate point about the way macroeconomics was operating at that time: It was ignoring the feedback loops that occur when we start trying to change policies.

Goodhart’s Law is probably a better formulation: Once you make an indicator into a target, you make it less effective as an indicator. So while inflation does seem to be negatively correlated with unemployment, that doesn’t mean we should try to increase inflation to extreme levels in order to get rid of unemployment; sooner or later the economy is going to adapt and we’ll just have both inflation and unemployment at the same time. (Campbell’s Law provides a specific example that I wish more people in the US understood: Test scores would be a good measure of education if we didn’t use them to target educational resources.)

The reason we must get rid of dynamic optimization is quite simple: No one behaves that way.

It’s often computationally intractable even in our wildly oversimplified models that experts spend years working on—now you’re imagining that everyone does this constantly?

The most fundamental part of almost every DSGE model is the Euler equation; this equation comes directly from the dynamic optimization. It’s supposed to predict how people will choose to spend and save based upon their plans for an infinite sequence of future income and spending—and if this sounds utterly impossible, that’s because it is. Euler equations don’t fit the data at all, and even extreme attempts to save them by adding a proliferation of additional terms have failed. (It reminds me very much of the epicycles that astronomers used to add to the geocentric model of the universe to try to squeeze in weird results like Mars, before they had the heliocentric model.)

We should instead start over: How do people actually choose their spending? Well, first of all, it’s not completely rational. But it’s also not totally random. People spend on necessities before luxuries; they try to live within their means; they shop for bargains. There is a great deal of data from behavioral economics that could be brought to bear on understanding the actual heuristics people use in deciding how to spend and save. There have already been successful policy interventions using this knowledge, like Save More Tomorrow.

The best thing about this is that it should make our models simpler. We’re no longer asking each agent in the model to solve an impossible problem. However people actually make these decisions, we know it can be done, because it is being done. Most people don’t really think that hard, even when they probably should; so the heuristics really can’t be that complicated. My guess is that you can get a good fit—certainly better than an Euler equation—just by assuming that people set a target for how much they’re going to save (which is also probably pretty small for most people), and then spend the rest.

The second most important thing we need to add is inequality. Some people are much richer than others; this is a very important fact about economics that we need to understand. Yet it has taken the economics profession decades to figure this out, and even now I’m only aware of one class of macroeconomic models that seriously involves inequality, the Heterogeneous Agent New Keynesian (HANK) models which didn’t emerge until the last few years (the earliest publication I can find is 2016!). And these models are monsters; they are almost always computationally intractable and have a huge number of parameters to estimate.

Understanding inequality will require more parameters, that much is true. But if we abandon dynamic optimization, we won’t need as many as the HANK models have, and most of the new parameters are actually things we can observe, like the distribution of wages and years of schooling.

Observability of parameters is a big deal. Another problem with the way the Lucas Critique has been used is that we’ve been told we need to be using “deep structural parameters” like the temporal elasticity of substitution and the coefficient of relative risk aversion—but we have no idea what those actually are. We can’t observe them, and all of our attempts to measure them indirectly have yielded inconclusive or even inconsistent results. This is probably because these parameters are based on assumptions about human rationality that are simply not realistic. Most people probably don’t have a well-defined temporal elasticity of substitution, because their day-to-day decisions simply aren’t consistent enough over time for that to make sense. Sometimes they eat salad and exercise; sometimes they loaf on the couch and drink milkshakes. Likewise with risk aversion: many moons ago I wrote about how people will buy both insurance and lottery tickets, which no one with a consistent coefficient of relative risk aversion would ever do.

So if we are interested in deep structural parameters, we need to base those parameters on behavioral experiments so that we can understand actual human behavior. And frankly I don’t think we need deep structural parameters; I think this is a form of greedy reductionism, where we assume that the way to understand something is always to look at smaller pieces. Sometimes the whole is more than the sum of its parts. Economists obviously feel a lot of envy for physics; but they don’t seem to understand that aerodynamics would never have (ahem) gotten off the ground if we had first waited for an exact quantum mechanical solution of the oxygen atom (which we still don’t have, by the way). Macroeconomics may not actually need “microfoundations” in the strong sense that most economists intend; it needs to be consistent with small-scale behavior, but it doesn’t need to be derived from small-scale behavior.

This means that the new paradigm in macroeconomics does not need to be computationally intractable. Using heuristics instead of dynamic optimization and worrying less about microfoundations will make the models simpler; adding inequality need not make them so much more complicated.

“But wait, there’s more!”: The clever tricks of commercials

pnrj cognitive science, heuristics and biases, market failure , , , , , , , , , , , , , , , ,

JDN 2457565

I’m sure you’ve all seen commercials like this dozens of times:

A person is shown (usually in black-and-white) trying to use an ordinary consumer product, and failing miserably. Often their failure can only be attributed to the most abject incompetence, but the narrator will explain otherwise: “Old product is so hard to use. Who can handle [basic household activity] and [simple instructions]?”

“Struggle no more!” he says (it’s almost always a masculine narrator), and the video turns to full color as the same person is shown using the new consumer product effortlessly. “With innovative high-tech new product, you can do [basic household activity] with ease in no time!”

“Best of all, new product, a $400 value, can be yours for just five easy payments of $19.95. That’s five easy payments of $19.95!”

And then, here it comes: “But wait. There’s more! Order within the next 15 minutes and you will get two new products, for the same low price. That’s $800 in value for just five easy payments of $19.95! And best of all, your satisfaction is guaranteed! If you don’t like new product, return it within 30 days for your money back!” (A much quieter, faster voice says: “Just pay shipping and handling.”)

Call 555-1234. That’s 555-1234.

“CALL NOW!”

Did you ever stop and think about why so many commercials follow this same precise format?

In short, because it works. Indeed, it works a good deal better than simply presenting the product’s actual upsides and downsides and reporting a sensible market price—even if that sensible market price is lower than the “five easy payments of $19.95”.

We owe this style of marketing to one Ron Popeil; Ron Popeil was a prolific inventor, but none of his inventions have had so much impact as the market methods he used to sell them.

Let’s go through step by step. Why is the person using the old product so incompetent? Surely they could sell their product without implying that we don’t know how to do basic household activities like boiling pasta and cutting vegetables?

Well, first of all, many of these products do nothing but automate such simple household activities (like the famous Veg-O-Matic which cuts vegetables and “It slices! It dices!”), so if they couldn’t at least suggest that this is a lot of work they’re saving us, we’d have no reason to want their product.

But there’s another reason as well: Watching someone else fumble with basic household appliances is funny, as any fan of the 1950s classic I Love Lucy would attest (in fact, it may not be a coincidence that the one fumbling with the vegetables is often a woman who looks a lot like Lucy), and meta-analysis of humor in advertising has shown that it draws attention and triggers positive feelings.

Why use black-and-white for the first part? The switch to color enhances the feeling of contrast, and the color video is more appealing. You wouldn’t consciously say “Wow, that slicer changed the tomatoes from an ugly grey to a vibrant red!” but your subconscious mind is still registering that association.

Then they will hit you with appealing but meaningless buzzwords. For technology it will be things like “innovative”, “ground-breaking”, “high-tech” and “state-of-the-art”, while for foods and nutritional supplements it will be things like “all-natural”, “organic”, “no chemicals”, and “just like homemade”. It will generally be either so vague as to be unverifiable (what constitutes “innovative”?), utterly tautological (all carbon-based substances are “organic” and this term is not regulated), or transparently false but nonetheless not specific enough to get them in trouble (“just like homemade” literally can’t be true if you’re buying it from a TV ad). These give you positive associations without forcing the company to commit to making a claim they could actually be sued for breaking. It’s the same principle as the Applause Lights that politicians bring to every speech: “Three cheers for moms!” “A delicious slice of homemade apple pie!” “God Bless America!”

Occasionally you’ll also hear buzzwords that do have some meaning, but often not nearly as strong as people imagine: “Patent pending” means that they applied for the patent and it wasn’t summarily rejected—but not that they’ll end up getting it approved. “Certified organic” means that the USDA signed off on the farming standards, which is better than nothing but leaves a lot of wiggle room for animal abuse and irresponsible environmental practices.

And then we get to the price. They’ll quote some ludicrous figure for its “value”, which may be a price that no one has ever actually paid for a product of this kind, then draw a line through it and replace it with the actual price, which will be far lower.

Indeed, not just lower: The actual price is almost always $19.99 or $19.95. If the product is too expensive to make for them to sell it at $19.95, they will sell it at several payments of $19.95, and emphasize that these are “easy” payments, as though the difficulty of writing the check were a major factor in people’s purchasing decisions. (That actually is a legitimate concern for micropayments, but not for buying kitchen appliances!) They’ll repeat the price because repetition improves memory and also makes statements more persuasive.

This is what we call psychological pricing, and it’s one of those enormous market distortions that once you realize it’s there, you see it everywhere and start to wonder how our whole market system hasn’t collapsed on itself from the sheer weight of our overwhelming irrationality. The price of a product sold on TV will almost always be just slightly less than $20.

In general, most prices will take the form of $X.95 or $X.99; Costco even has a code system they use in the least significant digit. Continuous substances like gasoline can even be sold at fractional pennies, and so they’ll usually be at $X.X99, being not even one penny less. It really does seem to work; despite being an eminently trivial difference from the round number, and typically rounded up from what it actually should have been, it just feels like less to see $19.95 rather than $20.00.

Moreover, I have less data to support this particular hypothesis, but I think that $20 in particular is a very specific number, because $19.95 pops up so very, very often. I think most Americans have what we might call a “Jackson heuristic”, which is as follows: If something costs less than a Jackson (a $20 bill, though hopefully they’ll put Harriet Tubman on soon, so “Tubman heuristic”), you’re allowed to buy it on impulse without thinking too hard about whether it’s worth it. But if it costs more than a Jackson, you need to stop and think about it, weigh the alternatives before you come to a decision. Since these TV ads are almost always aiming for the thoughtless impulse buy, they try to scrape in just under the Jackson heuristic.

Of course, inflation will change the precise figure over time; in the 1980s it was probably a Hamilton heuristic, in the 1970s a Lincoln heuristic, in the 1940s a Washington heuristic. Soon enough it will be a Grant heuristic and then a Benjamin heuristic. In fact it’s probably something like “The closest commonly-used cash denomination to half a milliQALY”, but nobody does that calculation consciously; the estimate is made automatically without thinking. This in turn is probably figured because you could literally do that once a day every single day for only about 20% of your total income, and if you hold it to once a week you’re under 3% of your income. So if you follow the Jackson heuristic on impulse buys every week or so, your impulse spending is a “statistically insignificant” proportion of your income. (Why do we use that anyway? And suddenly we realize: The 95% confidence level is itself nothing more than a heuristic.)

Then they take advantage of our difficulty in discounting time rationally, by spreading it into payments; “five easy payments of $19.95” sounds a lot more affordable than “$100”, but they are in fact basically the same. (You save $0.25 by the payment plan, maybe as much as a few dollars if your cashflow is very bad and thus you have a high temporal discount rate.)

And then, finally, “But wait. There’s more!” They offer you another of the exact same product, knowing full well you’ll probably have no use for the second one. They’ll multiply their previous arbitrary “value” by 2 to get an even more ludicrous number. Now it sounds like they’re doing you a favor, so you’ll feel obliged to do one back by buying the product. Gifts often have this effect in experiments: People are significantly more motivated to answer a survey if you give them a small gift beforehand, even if they get to keep it without taking the survey.

They’ll tell you to call in the next 15 minutes so that you feel like part of an exclusive club (when in reality you could probably call at any time and get the same deal). This also ensures that you’re staying in impulse-buy mode, since if you wait longer to think, you’ll miss the window!

They will offer a “money-back guarantee” to give you a sense of trust in the product, and this would be a rational response, except for that little disclaimer: “Just pay shipping and handling.” For many products, especially nutritional supplements (which cost basically nothing to make), the “handling” fee is high enough that they don’t lose much money, if any, even if you immediately send it back for a refund. Besides, they know that hardly anyone actually bothers to return products. Retailers are currently in a panic about “skyrocketing” rates of product returns that are still under 10%.

Then, they’ll repeat their phone number, followed by a remarkably brazen direct command: “Call now!” Personally I tend to bristle at direct commands, even from legitimate authorities; but apparently I’m unusual in that respect, and most people will in fact obey direct commands from random strangers as long as they aren’t too demanding. A famous demonstration of this you could try yourself if you’re feeling like a prankster is to walk into a room, point at someone, and say “You! Stand up!” They probably will. There’s a whole literature in social psychology about what makes people comply with commands of this sort.

And all, to make you buy a useless gadget you’ll try to use once and then leave in a cupboard somewhere. What untold billions of dollars in wealth are wasted this way?

Nature via Nurture

pnrj cognitive science, core principles , , , , , , , , , , , , , , , , , , , , , , , ,

JDN 2457222 EDT 16:33.

One of the most common “deep questions” human beings have asked ourselves over the centuries is also one of the most misguided, the question of “nature versus nurture”: Is it genetics or environment that makes us what we are?

Humans are probably the single entity in the universe for which this question makes least sense. Artificial constructs have no prior existence, so they are “all nurture”, made what we choose to make them. Most other organisms on Earth behave accordingly to fixed instinctual programming, acting out a specific series of responses that have been honed over millions of years, doing only one thing, but doing it exceedingly well. They are in this sense “all nature”. As the saying goes, the fox knows many things, but the hedgehog knows one very big thing. Most organisms on Earth are in this sense hedgehogs, but we Homo sapiens are the ultimate foxes. (Ironically, hedgehogs are not actually “hedgehogs” in this sense: Being mammals, they have an advanced brain capable of flexibly responding to environmental circumstances. Foxes are a good deal more intelligent still, however.)

But human beings are by far the most flexible, adaptable organism on Earth. We live on literally every continent; despite being savannah apes we even live deep underwater and in outer space. Unlike most other species, we do not fit into a well-defined ecological niche; instead, we carve our own. This certainly has downsides; human beings are ourselves a mass extinction event.

Does this mean, therefore, that we are tabula rasa, blank slates upon which anything can be written?

Hardly. We’re more like word processors. Staring (as I of course presently am) at the blinking cursor of a word processor on a computer screen, seeing that wide, open space where a virtual infinity of possible texts could be written, depending entirely upon a sequence of miniscule key vibrations, you could be forgiven for thinking that you are looking at a blank slate. But in fact you are looking at the pinnacle of thousands of years of technological advancement, a machine so advanced, so precisely engineered, that its individual components are one ten-thousandth the width of a human hair (Intel just announced that we can now do even better than that). At peak performance, it is capable of over 100 billion calculations per second. Its random-access memory stores as much information as all the books on a stacks floor of the Hatcher Graduate Library, and its hard drive stores as much as all the books in the US Library of Congress. (Of course, both libraries contain digital media as well, exceeding anything my humble hard drive could hold by a factor of a thousand.)

All of this, simply to process text? Of course not; word processing is an afterthought for a processor that is specifically designed for dealing with high-resolution 3D images. (Of course, nowadays even a low-end netbook that is designed only for word processing and web browsing can typically handle a billion calculations per second.) But there the analogy with humans is quite accurate as well: Written language is about 10,000 years old, while the human visual mind is at least 100,000. We were 3D image analyzers long before we were word processors. This may be why we say “a picture is worth a thousand words”; we process each with about as much effort, even though the image necessarily contains thousands of times as many bits.

Why is the computer capable of so many different things? Why is the human mind capable of so many more? Not because they are simple and impinged upon by their environments, but because they are complex and precision-engineered to nonlinearly amplify tiny inputs into vast outputs—but only certain tiny inputs.

That is, it is because of our nature that we are capable of being nurtured. It is precisely the millions of years of genetic programming that have optimized the human brain that allow us to learn and adapt so flexibly to new environments and form a vast multitude of languages and cultures. It is precisely the genetically-programmed humanity we all share that makes our environmentally-acquired diversity possible.

In fact, causality also runs the other direction. Indeed, when I said other organisms were “all nature” that wasn’t right either; for even tightly-programmed instincts are evolved through millions of years of environmental pressure. Human beings have even been involved in cultural interactions long enough that it has begun to affect our genetic evolution; the reason I can digest lactose is that my ancestors about 10,000 years ago raised goats. We have our nature because of our ancestors’ nurture.

And then of course there’s the fact that we need a certain minimum level of environmental enrichment even to develop normally; a genetically-normal human raised into a deficient environment will suffer a kind of mental atrophy, as when children raised feral lose their ability to speak.

Thus, the question “nature or nurture?” seems a bit beside the point: We are extremely flexible and responsive to our environment, because of innate genetic hardware and software, which requires a certain environment to express itself, and which arose because of thousands of years of culture and millions of years of the struggle for survival—we are nurture because nature because nurture.

But perhaps we didn’t actually mean to ask about human traits in general; perhaps we meant to ask about some specific trait, like spatial intelligence, or eye color, or gender identity. This at least can be structured as a coherent question: How heritable is the trait? What proportion of the variance in this population is caused by genetic variation? Heritability analysis is a well-established methodology in behavioral genetics.
Yet, that isn’t the same question at all. For while height is extremely heritable within a given population (usually about 80%), human height worldwide has been increasing dramatically over time due to environmental influences and can actually be used as a measure of a nation’s economic development. (Look at what happened to the height of men in Japan.) How heritable is height? You have to be very careful what you mean.

Meanwhile, the heritability of neurofibromatosis is actually quite low—as many people acquire the disease by new mutations as inherit it from their parents—but we know for a fact it is a genetic disorder, because we can point to the specific genes that mutate to cause the disease.

Heritability also depends on the population under consideration; speaking English is more heritable within the United States than it is across the world as a whole, because there are a larger proportion of non-native English speakers in other countries. In general, a more diverse environment will lead to lower heritability, because there are simply more environmental influences that could affect the trait.

As children get older, their behavior gets more heritable—a result which probably seems completely baffling, until you understand what heritability really means. Your genes become a more important factor in your behavior as you grow up, because you become separated from the environment of your birth and immersed into the general environment of your whole society. Lower environmental diversity means higher heritability, by definition. There’s also an effect of choosing your own environment; people who are intelligent and conscientious are likely to choose to go to college, where they will be further trained in knowledge and self-control. This latter effect is called niche-picking.

This is why saying something like “intelligence is 80% genetic” is basically meaningless, and “intelligence is 80% heritable” isn’t much better until you specify the reference population. The heritability of intelligence depends very much on what you mean by “intelligence” and what population you’re looking at for heritability. But even if you do find a high heritability (as we do for, say, Spearman’s g within the United States), this doesn’t mean that intelligence is fixed at birth; it simply means that parents with high intelligence are likely to have children with high intelligence. In evolutionary terms that’s all that matters—natural selection doesn’t care where you got your traits, only that you have them and pass them to your offspring—but many people do care, and IQ being heritable because rich, educated parents raise rich, educated children is very different from IQ being heritable because innately intelligent parents give birth to innately intelligent children. If genetic variation is systematically related to environmental variation, you can measure a high heritability even though the genes are not directly causing the outcome.

We do use twin studies to try to sort this out, but because identical twins raised apart are exceedingly rare, two very serious problems emerge: One, there usually isn’t a large enough sample size to say anything useful; and more importantly, this is actually an inaccurate measure in terms of natural selection. The evolutionary pressure is based on the correlation with the genes—it actually doesn’t matter whether the genes are directly causal. All that matters is that organisms with allele X survive and organisms with allele Y do not. Usually that’s because allele X does something useful, but even if it’s simply because people with allele X happen to mostly come from a culture that makes better guns, that will work just as well.

We can see this quite directly: White skin spread across the world not because it was useful (it’s actually terrible in any latitude other than subarctic), but because the cultures that conquered the world happened to be comprised mostly of people with White skin. In the 15th century you’d find a very high heritability of “using gunpowder weapons”, and there was definitely a selection pressure in favor of that trait—but it obviously doesn’t take special genes to use a gun.

The kind of heritability you get from twin studies is answering a totally different, nonsensical question, something like: “If we reassigned all offspring to parents randomly, how much of the variation in this trait in the new population would be correlated with genetic variation?” And honestly, I think the only reason people think that this is the question to ask is precisely because even biologists don’t fully grasp the way that nature and nurture are fundamentally entwined. They are trying to answer the intuitive question, “How much of this trait is genetic?” rather than the biologically meaningful “How strongly could a selection pressure for this trait evolve this gene?”

And if right now you’re thinking, “I don’t care how strongly a selection pressure for the trait could evolve some particular gene”, that’s fine; there are plenty of meaningful scientific questions that I don’t find particularly interesting and are probably not particularly important. (I hesitate to provide a rigid ranking, but I think it’s safe to say that “How does consciousness arise?” is a more important question than “Why are male platypuses venomous?” and “How can poverty be eradicated?” is a more important question than “How did the aircraft manufacturing duopoly emerge?”) But that’s really the most meaningful question we can construct from the ill-formed question “How much of this trait is genetic?” The next step is to think about why you thought that you were asking something important.

What did you really mean to ask?

For a bald question like, “Is being gay genetic?” there is no meaningful answer. We could try to reformulate it as a meaningful biological question, like “What is the heritability of homosexual behavior among males in the United States?” or “Can we find genetic markers strongly linked to self-identification as ‘gay’?” but I don’t think those are the questions we really meant to ask. I think actually the question we meant to ask was more fundamental than that: Is it legitimate to discriminate against gay people? And here the answer is unequivocal: No, it isn’t. It is a grave mistake to think that this moral question has anything to do with genetics; discrimination is wrong even against traits that are totally environmental (like religion, for example), and there are morally legitimate actions to take based entirely on a person’s genes (the obvious examples all coming from medicine—you don’t treat someone for cystic fibrosis if they don’t actually have it).

Similarly, when we ask the question “Is intelligence genetic?” I don’t think most people are actually interested in the heritability of spatial working memory among young American males. I think the real question they want to ask is about equality of opportunity, and what it would look like if we had it. If success were entirely determined by intelligence and intelligence were entirely determined by genetics, then even a society with equality of opportunity would show significant inequality inherited across generations. Thus, inherited inequality is not necessarily evidence against equality of opportunity. But this is in fact a deeply disingenuous argument, used by people like Charles Murray to excuse systemic racism, sexism, and concentration of wealth.

We didn’t have to say that inherited inequality is necessarily or undeniably evidence against equality of opportunity—merely that it is, in fact, evidence of inequality of opportunity. Moreover, it is far from the only evidence against equality of opportunity; we also can observe the fact that college-educated Black people are no more likely to be employed than White people who didn’t even finish high school, for example, or the fact that otherwise identical resumes with predominantly Black names (like “Jamal”) are less likely to receive callbacks compared to predominantly White names (like “Greg”). We can observe that the same is true for resumes with obviously female names (like “Sarah”) versus obviously male names (like “David”), even when the hiring is done by social scientists. We can directly observe that one-third of the 400 richest Americans inherited their wealth (and if you look closer into the other two-thirds, all of them had some very unusual opportunities, usually due to their family connections—“self-made” is invariably a great exaggeration). The evidence for inequality of opportunity in our society is legion, regardless of how genetics and intelligence are related. In fact, I think that the high observed heritability of intelligence is largely due to the fact that educational opportunities are distributed in a genetically-biased fashion, but I could be wrong about that; maybe there really is a large genetic influence on human intelligence. Even so, that does not justify widespread and directly-measured discrimination. It does not justify a handful of billionaires luxuriating in almost unimaginable wealth as millions of people languish in poverty. Intelligence can be as heritable as you like and it is still wrong for Donald Trump to have billions of dollars while millions of children starve.

This is what I think we need to do when people try to bring up a “nature versus nurture” question. We can certainly talk about the real complexity of the relationship between genetics and environment, which I think are best summarized as “nature via nurture”; but in fact usually we should think about why we are asking that question, and try to find the real question we actually meant to ask.

Happy Capybara Day! Or the power of culture

pnrj cognitive science, core principles, critique of neoclassical economics , , , , , , , , , , , ,

JDN 2457131 EDT 14:33.

Did you celebrate Capybara Day yesterday? You didn’t? Why not? We weren’t able to find any actual capybaras this year, but maybe next year we’ll be able to plan better and find a capybara at a zoo; unfortunately the nearest zoo with a capybara appears to be in Maryland. But where would we be without a capybara to consult annually on the stock market?

Right now you are probably rather confused, perhaps wondering if I’ve gone completely insane. This is because Capybara Day is a holiday of my own invention, one which only a handful of people have even heard about.

But if you think we’d never have a holiday so bizarre, think again: For all I did was make some slight modifications to Groundhog Day. Instead of consulting a groundhog about the weather every February 2, I proposed that we consult a capybara about the stock market every April 17. And if you think you have some reason why groundhogs are better at predicting the weather (perhaps because they at least have some vague notion of what weather is) than capybaras are at predicting the stock market (since they have no concept of money or numbers), think about this: Capybara Day could produce extremely accurate predictions, provided only that people actually believed it. The prophecy of rising or falling stock prices could very easily become self-fulfilling. If it were a cultural habit of ours to consult capybaras about the stock market, capybaras would become good predictors of the stock market.

That might seem a bit far-fetched, but think about this: Why is there a January Effect? (To be fair, some researchers argue that there isn’t, and the apparent correlation between higher stock prices and the month of January is simply an illusion, perhaps the result of data overfitting.)

But I think it probably is real, and moreover has some very obvious reasons behind it. In this I’m in agreement with Richard Thaler, a founder of cognitive economics who wrote about such anomalies in the 1980s. December is a time when two very culturally-important events occur: The end of the year, during which many contracts end, profits are assessed, and tax liabilities are determined; and Christmas, the greatest surge of consumer spending and consumer debt.

The first effect means that corporations are very likely to liquidate assets—particularly assets that are running at a loss—in order to minimize their tax liabilities for the year, which will drive down prices. The second effect means that consumers are in search of financing for extravagant gift purchases, and those who don’t run up credit cards may instead sell off stocks. This is if anything a more rational way of dealing with the credit constraint, since interest rates on credit cards are typically far in excess of stock returns. But this surge of selling due to credit constraints further depresses prices.

In January, things return to normal; assets are repurchased, debt is repaid. This brings prices back up to where they were, which results in a higher than normal return for January.

Neoclassical economists are loath to admit that such a seasonal effect could exist, because it violates their concept of how markets work—and to be fair, the January Effect is actually weak enough to be somewhat ambiguous. But actually it doesn’t take much deviation from neoclassical models to explain the effect: Tax policies and credit constraints are basically enough to do it, so you don’t even need to go that far into understanding human behavior. It’s perfectly rational to behave this way given the distortions that are created by taxes and credit limits, and the arbitrage opportunity is one that you can only take advantage of if you have large amounts of credit and aren’t worried about minimizing your tax liabilities. It’s important to remember just how strong the assumptions of models like CAPM truly are; in addition to the usual infinite identical psychopaths, CAPM assumes there are no taxes, no transaction costs, and unlimited access to credit. I’d say it’s amazing that it works at all, but actually, it doesn’t—check out this graph of risk versus return and tell me if you think CAPM is actually giving us any information at all about how stock markets behave. It frankly looks like you could have drawn a random line through a scatter plot and gotten just as good a fit. Knowing how strong its assumptions are, we would not expect CAPM to work—and sure enough, it doesn’t.

Of course, that leaves the question of why our tax policy would be structured in this way—why make the year end on December 31 instead of some other date? And for that, you need to go back through hundreds of years of history, the Gregorian calendar, which in turn was influenced by Christianity, and before that the Julian calendar—in other words, culture.

Culture is one of the most powerful forces that influences human behavior—and also one of the strangest and least-understood. Economic theory is basically silent on the matter of culture. Typically it is ignored entirely, assumed to be irrelevant against the economic incentives that are the true drivers of human action. (There’s a peculiar emotion many neoclassical economists express that I can best describe as self-righteous cynicism, the attitude that we alone—i.e., economists—understand that human beings are not the noble and altruistic creatures many imagine us to be, nor beings of art and culture, but simply cold, calculating machines whose true motives are reducible to profit incentives—and all who think otherwise are being foolish and naïve; true enlightenment is understanding that human beings are infinite identical psychopaths. This is the attitude epitomized by the economist who once sent me an email with “altruism” written in scare quotes.)

Occasionally culture will be invoked as an external (in jargon, exogenous) force, to explain some aspect of human behavior that is otherwise so totally irrational that even invoking nonsensical preferences won’t make it go away. When a suicide bomber blows himself up in a crowd of people, it’s really pretty hard to explain that in terms of rational profit incentives—though I have seen it tried. (It could be self-interest at a larger scale, like families or nations—but then, isn’t that just the tribal paradigm I’ve been arguing for all along?)

But culture doesn’t just motivate us to do extreme or wildly irrational things. It motivates us all the time, often in quite beneficial ways; we wait in line, hold doors for people walking behind us, tip waiters who serve us, and vote in elections, not because anyone pressures us directly to do so (unlike say Australia we do not have compulsory voting) but because it’s what we feel we ought to do. There is a sense of altruism—and altruism provides the ultimate justification for why it is right to do these things—but the primary motivator in most cases is culture—that’s what people do, and are expected to do, around here.

Indeed, even when there is a direct incentive against behaving a certain way—like criminal penalties against theft—the probability of actually suffering a direct penalty is generally so low that it really can’t be our primary motivation. Instead, the reason we don’t cheat and steal is that we think we shouldn’t, and a major part of why we think we shouldn’t is that we have cultural norms against it.

We can actually observe differences in cultural norms across countries in the laboratory. In this 2008 study by Massimo Castro (PDF) comparing British and Italian people playing an economic game called the public goods game in which you can pay a cost yourself to benefit the group as a whole, it was found not only that people were less willing to benefit groups of foreigners than groups of compatriots, British people were overall more generous than Italian people. This 2010 study by Gachter et. al. (actually Joshua Greene talked about it last week) compared how people play the game in various cities, they found three basic patterns: In Western European and American cities such as Zurich, Copenhagen and Boston, cooperation started out high and remained high throughout; people were just cooperative in general. In Asian cities such as Chengdu and Seoul, cooperation started out low, but if people were punished for not cooperating, cooperation would improve over time, eventually reaching about the same place as in the highly cooperative cities. And in Mediterranean cities such as Istanbul, Athens, and Riyadh, cooperation started low and stayed low—even when people could be punished for not cooperating, nobody actually punished them. (These patterns are broadly consistent with the World Bank corruption ratings of these regions, by the way; Western Europe shows very low corruption, while Asia and the Mediterranean show high corruption. Of course this isn’t all that’s going on—and Asia isn’t much less corrupt than the Middle East, while this experiment might make you think so.)

Interestingly, these cultural patterns showed Melbourne as behaving more like an Asian city than a Western European one—perhaps being in the Pacific has worn off on Australia more than they realize.

This is very preliminary, cutting-edge research I’m talking about, so be careful about drawing too many conclusions. But in general we’ve begun to find some fairly clear cultural differences in economic behavior across different societies. While this would not be at all surprising to a sociologist or anthropologist, it’s the sort of thing that economists have insisted for years is impossible.

This is the frontier of cognitive economics, in my opinion. We know that culture is a very powerful motivator of our behavior, and it is time for us to understand how it works—and then, how it can be changed. We know that culture can be changed—cultural norms do change over time, sometimes remarkably rapidly; but we have only a faint notion of how or why they change. Changing culture has the power to do things that simply changing policy cannot, however; policy requires enforcement, and when the enforcement is removed the behavior will often disappear. But if a cultural norm can be imparted, it could sustain itself for a thousand years without any government action at all.

How following the crowd can doom us all

pnrj heuristics and biases , , , , , , , , , , , , , , , , , , , , , , , ,

JDN 2457110 EDT 21:30

Humans are nothing if not social animals. We like to follow the crowd, do what everyone else is doing—and many of us will continue to do so even if our own behavior doesn’t make sense to us. There is a very famous experiment in cognitive science that demonstrates this vividly.

People are given a very simple task to perform several times: We show you line X and lines A, B, and C. Now tell us which of A, B or C is the same length as X. Couldn’t be easier, right? But there’s a trick: seven other people are in the same room performing the same experiment, and they all say that B is the same length as X, even though you can clearly see that A is the correct answer. Do you stick with what you know, or say what everyone else is saying? Typically, you say what everyone else is saying. Over 18 trials, 75% of people followed the crowd at least once, and some people followed the crowd every single time. Some people even began to doubt their own perception, wondering if B really was the right answer—there are four lights, anyone?

Given that our behavior can be distorted by others in such simple and obvious tasks, it should be no surprise that it can be distorted even more in complex and ambiguous tasks—like those involved in finance. If everyone is buying up Beanie Babies or Tweeter stock, maybe you should too, right? Can all those people be wrong?

In fact, matters are even worse with the stock market, because it is in a sense rational to buy into a bubble if you know that other people will as well. As long as you aren’t the last to buy in, you can make a lot of money that way. In speculation, you try to predict the way that other people will cause prices to move and base your decisions around that—but then everyone else is doing the same thing. By Keynes called it a “beauty contest”; apparently in his day it was common to have contests for picking the most beautiful photo—but how is beauty assessed? By how many people pick it! So you actually don’t want to choose the one you think is most beautiful, you want to choose the one you think most people will think is the most beautiful—or the one you think most people will think most people will think….

Our herd behavior probably made a lot more sense when we evolved it millennia ago; when most of your threats are external and human beings don’t have that much influence over our environment, the majority opinion is quite likely to be right, and can often given you an answer much faster than you could figure it out on your own. (If everyone else thinks a lion is hiding in the bushes, there’s probably a lion hiding in the bushes—and if there is, the last thing you want is to be the only one who didn’t run.) The problem arises when this tendency to follow the ground feeds back on itself, and our behavior becomes driven not by the external reality but by an attempt to predict each other’s predictions of each other’s predictions. Yet this is exactly how financial markets are structured.

With this in mind, the surprise is not why markets are unstable—the surprise is why markets are ever stable. I think the main reason markets ever manage price stability is actually something most economists think of as a failure of markets: Price rigidity and so-called “menu costs“. If it’s costly to change your price, you won’t be constantly trying to adjust it to the mood of the hour—or the minute, or the microsecond—but instead trying to tie it to the fundamental value of what you’re selling so that the price will continue to be close for a long time ahead. You may get shortages in times of high demand and gluts in times of low demand, but as long as those two things roughly balance out you’ll leave the price where it is. But if you can instantly and costlessly change the price however you want, you can raise it when people seem particularly interested in buying and lower it when they don’t, and then people can start trying to buy when your price is low and sell when it is high. If people were completely rational and had perfect information, this arbitrage would stabilize prices—but since they’re not, arbitrage attempts can over- or under-compensate, and thus result in cyclical or even chaotic changes in prices.

Our herd behavior then makes this worse, as more people buying leads to, well, more people buying, and more people selling leads to more people selling. If there were no other causes of behavior, the result would be prices that explode outward exponentially; but even with other forces trying to counteract them, prices can move suddenly and unpredictably.

If most traders are irrational or under-informed while a handful are rational and well-informed, the latter can exploit the former for enormous amounts of money; this fact is often used to argue that irrational or under-informed traders will simply drop out, but it should only take you a few moments of thought to see why that isn’t necessarily true. The incentives isn’t just to be well-informed but also to keep others from being well-informed. If everyone were rational and had perfect information, stock trading would be the most boring job in the world, because the prices would never change except perhaps to grow with the growth rate of the overall economy. Wall Street therefore has every incentive in the world not to let that happen. And now perhaps you can see why they are so opposed to regulations that would require them to improve transparency or slow down market changes. Without the ability to deceive people about the real value of assets or trigger irrational bouts of mass buying or selling, Wall Street would make little or no money at all. Not only are markets inherently unstable by themselves, in addition we have extremely powerful individuals and institutions who are driven to ensure that this instability is never corrected.

This is why as our markets have become ever more streamlined and interconnected, instead of becoming more efficient as expected, they have actually become more unstable. They were never stable—and the gold standard made that instability worse—but despite monetary policy that has provided us with very stable inflation in the prices of real goods, the prices of assets such as stocks and real estate have continued to fluctuate wildly. Real estate isn’t as bad as stocks, again because of price rigidity—houses rarely have their values re-assessed multiple times per year, let alone multiple times per second. But real estate markets are still unstable, because of so many people trying to speculate on them. We think of real estate as a good way to make money fast—and if you’re lucky, it can be. But in a rational and efficient market, real estate would be almost as boring as stock trading; your profits would be driven entirely by population growth (increasing the demand for land without changing the supply) and the value added in construction of buildings. In fact, the population growth effect should be sapped by a land tax, and then you should only make a profit if you actually build things. Simply owning land shouldn’t be a way of making money—and the reason for this should be obvious: You’re not actually doing anything. I don’t like patent rents very much, but at least inventing new technologies is actually beneficial for society. Owning land contributes absolutely nothing, and yet it has been one of the primary means of amassing wealth for centuries and continues to be today.

But (so-called) investors and the banks and hedge funds they control have little reason to change their ways, as long as the system is set up so that they can keep profiting from the instability that they foster. Particularly when we let them keep the profits when things go well, but immediately rush to bail them out when things go badly, they have basically no incentive at all not to take maximum risk and seek maximum instability. We need a fundamentally different outlook on the proper role and structure of finance in our economy.

Fortunately one is emerging, summarized in a slogan among economically-savvy liberals: Banking should be boring. (Elizabeth Warren has said this, as have Joseph Stiglitz and Paul Krugman.) And indeed it should, for all banks are supposed to be doing is lending money from people who have it and don’t need it to people who need it but don’t have it. They aren’t supposed to be making large profits of their own, because they aren’t the ones actually adding value to the economy. Indeed it was never quite clear to me why banks should be privatized in the first place, though I guess it makes more sense than, oh, say, prisons.

Unfortunately, the majority opinion right now, at least among those who make policy, seems to be that banks don’t need to be restructured or even placed on a tighter leash; no, they need to be set free so they can work their magic again. Even otherwise reasonable, intelligent people quickly become unshakeable ideologues when it comes to the idea of raising taxes or tightening regulations. And as much as I’d like to think that it’s just a small but powerful minority of people who thinks this way, I know full well that a large proportion of Americans believe in these views and intentionally elect politicians who will act upon them.

All the more reason to break from the crowd, don’t you think?