Scarcity: Why Having Too Little Means So Much (8 page)

The second component of bandwidth is executive control. As discussed above, executive control is multifaceted, so we begin by considering one of the many important functions to which it contributes, namely, self-control. In the late 1960s
Walter Mischel and his colleagues
performed one of the most interesting (at the very least, the cutest) psychology experiments on impulsivity. Mischel’s research staff would seat a four- or five-year-old in a room and put a marshmallow in front of him. Some children would stare entranced
at it, some would fidget with excitement; all of them wanted it. And the child could have it. But, before he could eat it, he was told there was a catch. More of an opportunity, really. The researcher was going to leave the room. If the child hadn’t eaten the marshmallow before the researcher returned, he would get a second marshmallow. The children were faced with one of the oldest problems known to man, what the social scientist Thomas Schelling calls “
the intimate contest for self-command
,” the problem of self-control.

Self-control remains one of the more difficult parts of the study of psychology. We know many ingredients go into the manufacturing of self-control. It depends on how we weigh the future. And we appear to do it inconsistently. Immediate rewards (a marshmallow now) are salient and receive a heavy weight. Rewards in the distant future (two marshmallows later) are less salient and thus receive lower weight. So when we think about one versus two marshmallows in the abstract future, two is better than one. But when one marshmallow is right in front of us now, it suddenly beats two. Selfcontrol also depends on willpower, a resource whose functioning we do not fully understand, but which is affected, among other things, by
personality, fatigue, and attention
.

Self-control relies heavily on executive control. We use executive control to direct attention, initiate an action, inhibit an intuitive response, or resist an impulse. In fact, a less publicized but often replicated part of Mischel’s study is highly instructive here.
The children who were most successful in resisting
the marshmallow temptation did so by focusing their attention elsewhere. Instead of looking at and thinking about the marshmallow, they thought about other things. Instead of having to resist the desire, they simply arranged not to notice it. As Mischel put it, “
Once you realize that willpower
is just a matter of learning how to control your attention and thoughts, you can really begin to increase it.”

This provides a telling link between executive control and selfcontrol. Since executive control helps direct attention and control impulses, reduced executive function will hamper self-control. A
number of experiments have vividly illustrated this connection. One experiment gave subjects
a memory task
. Some were asked to remember a two-digit number; some were given a seven-digit number. The subjects were then led to a lobby where they would await further testing. In front of them in the waiting area were slices of cake and fruit. The real test was what they would choose while they waited, while rehearsing those numbers in their heads. Those whose minds were not terribly occupied by the two-digit number chose the fruit most of the time. Those whose minds were busy rehearsing the seven-digit number chose the cake 50 percent more often. The cake is the impulsive choice. It requires conscious action to prevent the automatic choice. When our mental bandwidth is used on something else, like rehearsing digits, we have less capacity to prevent ourselves from eating cake.

In another study, white Australian students were served food, but in this case it was something they found revolting:
a chicken foot cooked in a Chinese style
that preserved the entire foot intact, claws included. The challenge for the subjects was that this was served by a Chinese experimenter, creating some pressure to act civilized. As in the cake study, some subjects’ minds were loaded: they were asked to remember an eight-digit number. Those whose minds were not loaded managed to maintain composure, keeping their thoughts to themselves. Not so with the cognitively loaded subjects. They would blurt out rude comments, such as “This is bloody revolting,” despite their best intentions.

Whether it is eating cake we would rather resist or saying things we do not mean to say, a tax on bandwidth makes it harder for us to control our impulses. And because scarcity taxes bandwidth, this suggests that scarcity not only can lower fluid intelligence but can also reduce self-control. Hence, the Australian student snaps at the Chinese experimenter, the executive consumed by the impending presentation snaps at her daughter, and the employee thinking about his unpaid bills snaps at a rude customer.

To explore whether scarcity reduces executive control, we gave subjects at the New Jersey mall a test that is frequently used to measure
executive control, one that directly tested their ability to inhibit automatic responses. First, the subjects were presented with the hypothetical financial scenarios, either easy or hard, as before. They would then see pictures such as these:

or

in rapid succession on a computer screen. They placed the fingers of both hands on the keyboard, and their task was to press the same side as the heart and the
opposite
side of the flower. So if the heart appears on the right, you press right. And if the flower appears on the right, you press left.

The flower creates an automatic impulse that needs to be resisted: hitting the same side as the heart comes easy; hitting the opposite side of the flower is hard. Doing well requires overriding your impulse to quickly hit the same side. The more executive control you have, the better you will do. This test measures how capable you are at inhibiting your first impulse in favor of a different response, be it resisting a cake, biting your tongue, or, in this case, resisting the flower.

Though this task tests executive control, quite different from fluid intelligence, the results were the same. After the financially easy questions, the poor and the well off looked similar. They were able to control their impulses to the same degree, and they made about the same number of errors. But the financially hard questions
changed things dramatically for the poor. The well-off subjects continued to do just as well as if they had seen the easy scenario. They exhibited the same level of executive control. The poorer subjects, on the other hand, now did significantly worse. They were more impulsive, mistakenly hitting the same side as the flower more often. While they had hit the correct key 83 percent of the time in the context of the financially easy scenarios, correct key presses went down to 63 percent in the context of scenarios that were financially more challenging. A small tickle of scarcity and they were suddenly more impulsive. Beyond fluid intelligence, scarcity appears to reduce executive control.

These experiments at the mall test our hypothesis. But in a way, they are artificial. They show how people respond when we trigger in them thoughts about scarcity, which we induce through hypothetical questions about financial hardship. Our interest, though, is in people’s everyday lives outside the confines of an experiment. Does scarcity tax people’s cognitive resources even when there are no experimenters lurking at the mall to get them to think about it?

Showing this is essential to our argument. But it is hard. We cannot simply look at how poor people compare to rich people in cognitive capacity or self-control. Too many other things—health, friends, education—differ between the rich and the poor for us to be able to attribute any observed differences to scarcity. Such comparisons have been attempted endlessly with no obvious solution to the statistical problems that are inherent to such comparisons. How could we see the effect of scarcity free from all these intricacies?

It was around this time that we were doing fieldwork on farming in India with the economist Anandi Mani, when we noticed something interesting. Farmers get their income in a big lump, all at once at harvest time. This means the farmer has a very different financial
life from most workers, who get paid regularly (daily, weekly, or monthly). Instead, a farmer might get paid twice a year or sometimes even once a year. Now picture a farmer who gets paid in June. The next few months are quite good: he’s got cash. But even if he’s prudent and tries hard to smooth his spending over this period, by the time next April or May rolls around, he will be tight on cash. So the
same farmer
is rich in the months after harvest and poor in the months before harvest.

This was quite close to what we needed: we could examine the same farmer’s bandwidth in the months before harvest and in the months after harvest. Instead of comparing rich and poor people, we’d be seeing how the same person behaves differently when tight for cash and when flush with cash. But there was one wrinkle. Might not harvest months impose different obligations from ordinary months? For example, festivals and weddings are common during harvest months—exactly because people are cash rich. So instead of seeing the effects of scarcity, we might just see the effects of celebrations.

To get around this, we used sugar cane farming, which has a peculiar feature. Sugar cane requires an enormous factory to crush the cane and extract the juice (which, once evaporated, forms sugar). The factories can only process so much and the crop can’t sit after harvesting for long. So sugar cane is harvested during a four-to-five-month window. In some areas it is harvested throughout the year. Neighboring plots are often on very different harvest cycles. One farmer may be harvesting while his neighbor to one side harvested several months ago and his other neighbor has months to go before harvesting. This rather obscure fact gave us the break we needed. We could now study the same farmers when they’re poor and rich
and
know that there’s nothing specific about the preharvest and postharvest calendar months. After all, the same month was preharvest for one farmer and postharvest for his neighbor.

As we expected
, the data showed that the farmers were more strapped for cash preharvest. Seventy-eight percent of them had pawned something in the month before harvest (and 99 percent took
some sort of loan), but only 4 percent pawned something in the month after harvest (and only 13 percent took any kind of loan). Before harvest, they were also more likely to report having trouble coping with ordinary bills.

As at the mall, we again measured executive control and fluid intelligence. We gave the farmers a Raven’s Matrices task, but we could not do the heart–flower task because it was difficult to administer it in the field. So for an executive control task, we chose a close cousin, something called the Stroop task. In this task, subjects see strings of items, such as
F F F F
, and have to quickly say how many items are in the string. (In this case, the answer is four.) When you see
2 2 2 2
, quickly saying “four” is quite hard.
It is hard for the same reason
that it is hard to quickly hit the opposite side each time you see the flower.

Using these tasks, we found that farmers performed much worse before harvest than after harvest. The same farmer fared worse on fluid intelligence and executive control when he was poor (preharvest) than when he was rich (postharvest). Much like the subjects at the mall, the same person looked less intelligent and more impulsive when he was poor. Yet in this case it was not us who triggered scarcity-related thoughts or even tried to bring them to the surface. These thoughts were there naturally when the farmers were poor (the harvest money dissipated to a small amount) but not when they were rich (still flush with cash from the harvest).

And again the magnitudes were large. The postharvest farmers got about 25 percent more items correct on Raven’s. Put in IQ terms, as in the earlier mall study, this would correspond to about 9 or 10 IQ points. Not as big a gap as at the mall, but that is to be expected. After all, here we hadn’t induced them to think about money. We simply measured their mental state at an arbitrarily selected point in time, their
latent
tendency to have their bandwidth taxed by scarcity.
On the executive control task
, they were 11 percent slower in responding and made 15 percent more errors while poor, quite comparable to the mall study. Had we met a farmer when he was poor, we would have been tempted to attribute his limited capacity to a personal trait. But we know from our study that his limitation has
little to do with his genuine capacity as a person. The very state of having less money in the months before harvest had made him perform less intelligently and show less cognitive control.