Kluge: The Haphazard Construction of the Human Mind (Houghton Mifflin; 2008)

I'll start with good news. On occasion, human choices can be entirely rational. Two professors at NYU, for example, studied what one might think of as the world's simplest touch-screen video game — and found that, within the parameters of that simple task, people were almost as rational (in the sense of maximizing reward relative to risk) as you could possibly imagine. Two targets appear (standing still) on a screen, one green, one red. In this task, you get points if you touch the green circle; you lose a larger number of points if you touch the red one. The challenge comes when the circles overlap, as they often do, and if you touch the intersection between the circles, you get both the reward and the (larger) penalty, thus accruing a net loss. Because people are encouraged to touch the screen quickly, and because nobody's hand-eye coordination is perfect, the optimal thing to do is to point somewhere other than the center of the green circle. For example, if the green circle overlaps but is to the right of the red circle, pointing to the center of the green circle is risky business: an effort to point at the exact center of the green circle will sometimes wind up off target, left of center, smack in the point-losing region where the green and red circles overlap. Instead, it makes more sense to point somewhere to the right of the center of the green circle, keeping the probability of hitting the green circle high, while minimizing the probability of hitting the red circle. Somehow people figure all this out, though not necessarily in an explicit or conscious fashion. Even more remarkably, they do so in a manner that is almost perfectly calibrated to the specific accuracy of their own individual system of hand-eye coordination. Adam Smith couldn't have asked for more.

The bad news is that such exquisite rationality may well be the exception rather than the rule. People are as good as they are at the pointing-at-circles task because it draws on a mental capacity — the ability to reach for things — that is truly ancient. Reaching is close to a reflex, not just for humans, but for every animal that grabs a meal to bring it closer to its mouth; by the time we are adults, our reaching system is so well tuned, we never even think about it. For instance, in a strict technical sense, every time I reach for my cup of tea, I make a set of choices. I decide that I want the tea, that the potential pleasure and the hydration offered by the beverage outweigh the risk of spillage. More than that, and even less consciously, I decide at what angle to send my hand. Should I use my left hand (which is closer) or my right hand (which is better coordinated)? Should I grab the cylindrical central portion of the mug (which holds the contents that I really want) or go instead for the handle, a less direct but easier-tograsp means to the tea that is inside? Our hands and muscles align themselves automatically, my fingers forming a pincer grip, my elbow rotating so that my hand is in perfect position. Reaching, central to life, involves many decisions, and evolution has had a long time to get them just right.

But economics is not supposed to be a theory of how people reach for coffee mugs; it's supposed be a theory of how they spend their money, allocate their time, plan for their retirement, and so forth — it's supposed to be, at least in part, a theory about how people make conscious decisions.

And often, the closer we get to conscious decision making, a more recent product of evolution, the worse our decisions become. When the NYU professors reworked their grasping task to make it a more explicit word problem, most subjects' performance fell to pieces. Our more recently evolved deliberative system is, in this particular respect, no match for our ancient system for muscle control. Outside that rarefied domain, there are loads of circumstances in which human performance predictably defies any reasonable notion of rationality.

Suppose, for example, that I give you a choice between participating in two lotteries. In one lottery, you have an 89 percent chance of winning $1 million, a 10 percent chance of winning $5 million, and a 1 percent chance of winning nothing; in the other, you have a 100 percent chance of winning $1 million. Which do you go for? Almost everyone takes the sure thing.

Now suppose instead your choice is slightly more complicated. You can take either an 11 percent chance at $1 million or a 10 percent chance of winning $5 million. Which do you choose? Here, almost everyone goes for the second choice, a 10 percent shot at $5 million.