Can Economic Utility Be Measured?

A friend of mine (who happens to be Ph.D student in economics) sent me a skeptical email regarding a recent article that sought to measure marginal utility:

I'm really not convinced that marginal utility can be so easily correlated with activity in the midbrain. I think one of the virtues of the economic definition of marginal utility is that it's ultimately vague in definition. Depending on the context it can be happiness or money or satisfaction or whatever that person wants. I'm not sure it benefits from a strict neuroscientific definition.

I understand the skepticism. But I think there is solid scientific evidence that the brain has a generalized reward system. In fact, the reward system inside our brain actually supports the vagueness of economic utility theory since just about anything can become a reward, at least according to these "reward neurons". We use the same dopaminergic cells no matter what reward we actually desire, be it crack cocaine or an HDTV or a loved one. The vagaries of human desire have a common source.

And it's also important to note that evidence for this system doesn't just come from fMRI experiments. In fact, I think some of the most promising neuroeconomic research goes beyond brain imaging, and actually maps approximations of utility directly onto the firing rates of neurons. (For obvious reasons, this research usually can't be done on humans.)

Take this elegant paper, which was published in Nature in May 2006. Dr. Camillo Padoa-Schioppa, a neuroscientist in John Assad's Harvard lab, spent several months watching monkeys choose what to drink. He offered the primates a dozen different beverages and let them select their favorite. Schioppa discovered that the monkeys had a complicated hierarchy of preferences. Kool aid was better than grape juice, which was better than fruit punch. Apple juice was better than cranberry juice, and peppermint tea was preferred over water. Nobody liked diet Kool Aid.

Of course, observing the beverage habits of monkeys wasn't what got Schioppa's research paper into Nature. Schioppa's real experimental insight was that he could get the monkeys to change their mind, and choose the less preferable drink. If he was offering the monkey equal amounts of peppermint tea and apple juice, then the monkey would always choose the apple juice, because juice tastes better. (It had a higher utility.) But if he increased the amount of peppermint tea on offer, then, at a certain point, the monkey would change its mind, and choose to drink tea instead. It took hundreds and hundreds of experimental trials, but Schioppa eventually found the moment at which each monkey began to change its mind, and chose milk over cranberry juice, or water over fruit punch. This is called "the indifference point," since the monkey seemed "indifferent" between the two options. For example, if the monkey was equally likely to choose four drops of tea or one drop of apple juice, then Schioppa knew that the monkey valued the taste of apple juice four times more than tea. Apple juice has greater utility on a per drop basis.

What was the point of all this liquid bartering? Schioppa wanted to obtain a precise measurement of the monkey's subjective preferences, so that he could figure out what was going on in the monkey's brain when it was making these economic decisions. His question was simple yet profound: How do the cells in our brain choose between beverages? How does the mind represent our desires?

When Schioppa measured the activity of individual cells in the orbitofrontal cortex, he found something truly cool. These neurons were responding to the subjective value of the drinks - their utility - rather than the drinks' physical properties. For example, the brain cells representing Kool Aid showed a higher firing rate than the neurons representing cranberry juice when the monkey was offered equal amounts of both drinks, since Kool Aid has a higher value on a per drop basis. However, the same neurons had an equivalent firing rate when the monkey was choosing between one drop of Kool Aid and four drops of cranberry juice ("the indifference point"). Obviously, this pattern of neuronal activity can't be explained in terms of the drinks' volume, because equal volumes of the drinks produce different neuronal firing rates. Nor can it be explained in terms of the drinks' taste, since different volumes of the drinks produce equal levels of neuronal firing. According to Schioppa, this left only one answer: the neurons were encoding the monkey's subjective preferences. By monitoring these small changes in electrical activity, he could accurately predict which drink the monkey would actually choose. He knew what they wanted before they did.

And I probably don't have to remind people that the orbitofrontal cortex has previously been implicated in all sorts of decision-making tasks. Experiments like this will, I believe, eventually help us define economic utility in terms of specific neurological events. I can only believe that this sort of reductionist research is crucial for the future of economics. Herbert Simon said it best:

It has sometimes been implied that the assumptions of rational behavior underlying the classical theory of economics are not merely irrelevant, but are not even empirically testable in any direct way, the only valid test being whether these assumptions lead to tolerably correct predictions at the macroscopic level. That would be true, of course, if we had no microscopes, so that the micro-level behavior was not directly observable. But we do have microscopes.