*Scarcity: Why Having Too Little Means so Much*

That is the new book by Sendhil Mullainathan and Eldar Shafir, and as you might expect it is one of the most significant economics books of the year.  Here is their bottom line:

The poor are not just short on cash.  They are also short on bandwidth.

For an example, imagine giving both rich and poor an intelligence test with this question:

Imagine that your car has some trouble, which requires a $300 service.  Your auto insurance will cover half the cost.  You need to decide whether to go ahead and get the car fixed, or take a chance and hope that it lasts for a while longer.  How would you go about making such a decision?  Financially, would it be an easy or a difficult decision for you to make?

In their answers to that question, we are told, rich and poor look equally smart.  Now run the same question with different groups, but change the first sentence to this:

Imagine that your car has some trouble, which requires an expensive $3000 service.

All of a sudden the poorer individuals did much worse in response to this question and the authors claim this result has been replicated repeatedly.  Control studies suggest it is not about the number being larger per se, but rather that the poor individuals see this as a more stressful decision, which lowers their measured fluid intelligence.

Overall I find this all very intriguing, but would like to have a better sense of how this fits in with other results about the relative rigidity of IQ.  I also worry about tests where there is an exogenous increase in stressfulness, to which test participants must submit.  There are various ways that an examiner could stress me out, but part of one’s smarts, whether at high income levels or low, is exercising some control over matching your talents to the environment.

Here is a good review of the book by Oliver Burkeman.  By the way, Alex Tabarrok and MR make a cameo appearance on p.103.

Addendum: Here is Alex on their work.


Imagine that your car has some trouble, which requires a $300 service. Your auto insurance will cover half the cost.

Snarky reply: I want to know where you get car insurance where it covers even part of "trouble" in the generic sense, rather than liability and possibly collision repair.

Maybe due to the language barrier they meant warranty?

From a health insurance company.


'and as you might expect it is one of the most significant economics books of the year'

Affiliate link? Check.

'and it is likely to be the best and most important economics book of the forthcoming year.' - Just enough different from a post the previous day to look like something other than pandering? Check.

Self-referencing link? - 'By the way, Alex Tabarrok and MR make a cameo appearance on p.103' Check.

A perfect post, if one believes in checklists.

>A perfect post, if one believes in checklists.

Only when you replied.

I felt compelled to look up "bandwidth." Still can't see how it fits here.

Seems, President Obama should have handed out millions of PC's instead of smart phones.

Of course, a "poor" person would look upon a one-time $3,000 "bill" and murmur, "We won't eat this month." Whereas a "rich " person would write a check: firm grasp of the obvious (which is not a compliment).

No one handed out millions of smart phones. "Dumb" phones that can do calls and texting, yes, but not smart ones.

I'm not sure the clarifying detail mitigates the point.

I think Alex's post, which Tyler links to, gives a little clearer explanation than this post by the otherwise usually excellent Tyler. (I myself was confused as to how a hypothetical question induced stress based on the dollar amount involved.) In Alex's post, he explains that the poor encounter many more "urgent" tasks and thus can devote fewer cognitive resources, i.e., bandwidth, to solving longer-term problems.

As Alex's post also explains, the authors of this book simulate poverty in various experiments/games by giving some players fewer game resources. They find that those poor players devote more attention to current rounds and less attention to future rounds. Alex also notes that these experiments allow one to study the *effect* of poverty rather than the poor themselves, i.e., no selection issues since the "poor" in the games are not poor in reality.

I have noticed a tendency I've labeled "telling the poor to save money by buying in bulk" -- folks trying to be helpful by advising the poor to do things they are in a position to do themselves but the poor are not.

You think the people shopping at Costco are rich?


"the average salary of a Costco member is $95,333."


From a 2004 Slate article:

"the average salary of a Costco member is $95,333."

So, yes?

The customers at my Costco in Van Nuys look a lot higher up the social scale on average than the people at the DMV in Van Nuys.

Cliff: pwned. Sorry dude.

It's been my experience that Costco caters to yuppies to a large extent. I think Sam's Club tends to cater more to the lower end of the income scale. (But that's purely a single POV anecdotal observation.)

"Bandwidth" in the sense of "capacity for processing". See also "ego depletion". Stress creates ego depletion, so it reduces the available capacity to think things through carefully. And that can result in increased stress. Turns out, one of the most effective things I've seen for getting people out of that loop is to magically fix a problem once at a time when they were looking like they'd have made it out but something came up. A single $60 bus pass or the like can result in someone being in pretty good shape a year later.

It's like asking, "The tiger is hungry - how will you resolve this issue?" and then using the collected responses to declare that zoo-keepers have more "fluid intelligence" than Bangladeshi villagers in the sundarbans.

"Stress" impacting "fluid intelligence" seems like an odd way to put it. $3000 is not an existential threat to a middle class family, but it certainly is to many members of the working poor. I wouldn't say the stress of the decision impedes their ability to solve the problem; instead, I'd say that whether a problem presents an existential threat materially changes the way we approach that problem. It's not a question of intelligence. Solving an existential threat just looks different than solving a monetary expenditure problem.

To finish off your analogy, the poor man is given the choice of giving all the meat for his family to the tiger so his children starve, or picking a child and throwing her to the tiger, while the better off man buys twice as much chicken instead of steak and feeds half the chicken to the tiger and feeds the kids with the other half.

The poor man has an unsustainable path when the tiger returns, while the better off man will miss his steak dinners until he can arrange for his fellow better off men to deal with the tiger.

I think you're not familiar with the budgets and expenditures of the "poor".

There is a reason why the number one health problem of our poor is obesity.

They eat bad food?

Not as bad as starving people. During Holodomor people were fighting over the last pair of shoes in the village, and not to wear them.

A poor person is also much more likely to own a car that is only worth $1500 (or less).

Even if it were a $6,000 car, that's a quarter of the cost.


$300 in a minor repair; $3000 indicates that your car is in a lot of trouble.

Really the problem is the lack of a 300 dollar repair (not fixing tranmission leak) leads to new transmission (3000).

And the only time I've been told I had a 3000 bill was when the shop didn't want the work.

Maybe the book addresses this, or event his excerpt does and I don't understand it well, but the "smart" decision is not static. It is situational, with poverty being a whole situation, not just a factor. Poor people live (by necessity) differently overall than rich people do, so the same questions mean different things in those very different contexts.

First, it makes me skeptical of the study writer's sense when he talks of insurance covering car troubles.

Second, why is it strange that people get stressed outside their comfort zones? A Saudi prince would probably fumble while shopping at the dollar store & I would fumble with a million dollar withdrawal at a Swiss bank.

I echo the concerns raised above. Plus, how to determine what is a "good" answer to this question? Seems that "fix it immediately" and "ride it out for a while" could both be decent enough answers to this question, absent further details.

Especially, because, as noted above, a lot of poor people have cars that may not even be worth $1,500. In that case, "ride it out for a while" may make more sense as you save up for a replacement rather than dump more money that what the car is worth to keep it running a bit better.

I once had this really old clunker of a car in this exact situation. I was practically willing to give it away. The dealer took one look at it and started begging me to do a trade in. "Why the heck does he want this 12 year old car worth no more than $2,000 that needs $1,500 in repairs so badly?" I asked him. It turns out a lot of their clientele were poor Mexican immigrants and there was huge demand for cars at that price point, even if they were clunkers with "trouble."

This is one case where I don't see I.Q. making much of a difference, except very indirectly, since the person with the higher I.Q. is probably more likely to have saved more money and/or have higher take home pay. If the conditions are the same, the behavior is likely to be the same, unless the low I.Q. person is willing to ride around chancing disaster while having the money needed to fix it in the bank. I suppose that's a possibility, but it doesn't seem like the one the quotes you pulled are pointing to.

It's a threshold thing. Ask a lawyer who deals in parking and speeding tickets. He'll tell you the greater percentage of clients are poorer. Some of it is poor decision making (speeding, or doing so in the wrong place), but some of it is exposure. Poor driver cannot afford indoor parking (in a city) and is more exposed to parking tickets, street cleaning tickets, etc. Anecdotally, when I was younger and had less money, I seemed to be much more susceptible to municipal revenue schemes. And, because I had less money, the amounts hurt all the more.

May also be that you've wisened up to the schemes. Once bitten, twice shy. And if we assume that the younger are more likely to get caught; and we know the younger tend to be poorer, it would make sense for those fines to hit the poorer more.

Also could be that cops spend more time in poorer neighborhoods, for reasons unrelated to giving parking tickets, but since we're here anyway what the heck. Fun to speculate about these things at least.

No because in basically every human society ever:

poorer=less politically powerful

Things like alternate day parking would not exist if they harmed the rich, or even the better parts of the middle.

Or I could put it this way, once I was drunk, high, naked and in a car that made a u-turn across the median on I-35 and was pulled over by the police, and only received a warning. The fact that it was a brand new Cadillac Fleetwood Brougham registered to the father of the underage girl who made that u-turn, a gentleman who also happened to be a person that every ambitious peace officer in Central Texas knew the name of because of his distinct position in the Texas Democratic Party, in the days when that sort of thing really really mattered, of course all of those details had nothing to do with anything.

Surely you understand that is something that almost never happens to anyone, and so is practically irrelevant

I always wondered whether kids in poor areas might have difficulty performing well academically given more uncertain, and more debt-laden, pathways. Even wealthier kids have to negotiate between local high school incentives, and long-term incentives. But if you attend a high school where most kids see a clear relationship between their scholastic efforts and outcomes, then you might expect a cultural emphasis on intelligence and academic performance. Perhaps rising inequality even enhances cultural differences as the incentives for good performance improve over time. If cost is a negligible factor, then society applies enormous leverage to fairly minor accomplishments (straight As, good SAT scores, extra-currics). It would be interesting to see whether wealthier kids, or those who can reasonably expect financial support, focus less upon the costs of those potential accomplishments relative to poorer kids.

I don't think "intelligence" is the concept most relevant here -- perhaps "defeatedness" would be a more useful term. For example, I've seen rural working class families demoralized by the immense sticker prices for college education, and thus fail to jump through all the hoops necessary to receive financial aid.

Of course, defeatedness can impact the upper end of the social stratum as well -- Dave Barry, B.A. in English Lit from Haverford, has written illuminatingly about his feelings of inadequacy and incompetence when considering home improvement projects that most rural working class families would consider right in their wheelhouses.

A good point. Having watched my father (high school dropout, with somewhat limited reading/writing skills) read construction blueprints (and figure out what to do to handle the inevitable errors in those blueprints) I think an immense part of what we call "intelligence" is very difficult to isolate well from context and experience -- and the feeling that one can defeat the task, rather than be defeated by it.

There's a National Geographic article about a safari in the 1920's (written then, when PC wasn't even a term) that describes the difficulty the porters had in setting up a simple folding table. A good source of amusement for the westerners. A few short paragraphs later, these porters were in the river, catching fish with their bare hands.

I wonder if your idea of "defeatedness" is akin to the idea of "bandwidth" in the OP - a rural, working class family might not have the resources (time, patience, paperwork, etc.) to complete the FAFSA. That poverty is demoralizing is probably pretty uncontroversial.

What a stupid question! And increasing the cost of the repair to $3000 makes it stressful?! Duh! That means the car is gonzo and I have to buy a new one, pretty soon if my $12 / hour job depends on my being mobile. And I don't have the bandwidth right now to go shopping...

LOOOOL at "bandwidth"

Whatever the merits of the book, the example cited sounds like extremely sketchy research. The kind you need to take with a shaker of salt, especially if the proponent is not up-front with the methodology.

"The poor are not just short on cash. They are also short on bandwidth."

Is this anything more than an failed attempt to use a buzzword?

Usually bandwidth to me in such a context translates to time available to spend on a solution. For example, I've can make the problem go away quickly by spending a significant amount of money or I can dedicate some bandwidth (time to carefully plan things out, comparison shop, see if there is an alternate I've missed) and thus avoid spending as much money.

Generally, in my experience, the poor have lots of bandwidth and little money. So they are more apt to come up with a solution that requires more effort but less money. So for example,

Imagine that your car has some trouble, which requires an expensive $3000 service.

Rich person: I drive my car to a full service dealer, obtain a car from them and they call me when the problem is fixed.

Middle class person: I drive my car to an auto shop, that offers good prices. My wife follows me in her car and takes us both home. We carpool, till the car is fixed. We pick it up Saturday morning, because we both can't take off work earlier to get it before then.

Poor person: I keep driving the car till it breaks down. Then I call a friend with a pickup truck to give me a tow to my house. I then use the bus or a relative to get to work. I start working on the car using a Chilton's manual, light cord, junkyard parts and 6-pack after work that night.

That really syncs up accurately with the rich, middle class, and poor people I know.

Another point is that "trouble" means different things. My husband's truck has a broken door, so he has to open the driver's side door from the inside. It's a real pain every morning, leaning in, opening the door, going around. I can't imagine any wealthy person dealing with that for a second; most middle class folks would take it in when they had the time; but the poor look at fixing that as a luxury, maybe one that can be purchased after IRS money or a bonus comes in.

awesome explanation, but perhaps it's better to meet some real poor people before assuming they have the knowledge to repair cars.

Well, this matters too, because "poor" isn't a monolith. Rural poor forty-five year old who grew up fixing Fords or Chevys and has an old model he owns outright and keeps up -- that I see a lot of among folks near the poverty line. But the young man I know working minimum wage in an urban area with a bitty cute modern car on payments, new car that is very hard to make any repairs to without shop equipment, he's not fixing his vehicle, no. He's more likely to just put off until tomorrow, hoping it will hold together until he can fix it later, and when it breaks he goes back to the dealer who puts him in another loan, a worse position, etc.

I'm sure there are seventy other kinds of poor in the U.S., too.

Chilton manual. Bigger obstacle may be the tools

awesome explanation, but perhaps it’s better to meet some real poor people before assuming they have the knowledge to repair cars.

That's where the Chilton Repair manual comes in. And I have more than a few "real poor people" in the extended family. The kind that I go pick up when their pickup truck breaks down on the highway.

Chilton manual. Bigger obstacle may be the tools

You are right on both accounts. Though you can still get an awful lot done with a cheap set of imperial/metric sockets. I think the biggest negative to the poor working on their cars anymore is a) modern cars are much harder to repair without a lift and b) due to increased regulation and insurance costs, junk yard parts are far more expensive than they used to be.

Don't forget the cash for clunkers program -- where people were federally encouraged to trade in usable, repairable older vehicles in exchange for credit on loans for new, harder to repair cars. My understanding is that this severely affected the parts market also.

What does Tyler think is the correct answer?

The correct answer is the one Tyler picks.

Marginal tautology.

I am very skeptical of claims of general human maladaptiveness, and that seems to be what we're talking about here. The idea that people are dumbest and generally perform worse precisely when stakes are highest and effective action is most important -- that makes no evolutionary sense whatsoever. Might it be the case that people become more risk averse and that their time horizons shrink when the stakes are personally higher? Sure -- I wouldn't be surprised by that at all. But the idea people get stupider when it matters most? No way.

The study seems a little detached from reality.

Three thousand dollars is not just a stressfully large amount for the poor; it's not just an amount that requires cutting into essentials. It's an amount they can't (often) acquire at all. Period. So the question posed to a rich person is "what would you do if you had a $3000 bill and that was hard?" but the same question for a poor person is "what would you do if you had a $3000 bill and you couldn't pay it?" It's an entirely different question. The idea that someone might be considered less (even simply momentarily) intelligent because he answers a different question differently seems pretty strange.

As others have noted, it is a strange claim to say that the "poor" have a bandwidth problem rather than an intelligence problem. The very poorest in America are likely to be not working and receiving benefits from the Government. Thus, all things being equal (including their intelligence) they should have more time to plan the best strategy than a working person. Similarly, even if working, poorer people tend to have less cognitively demanding jobs, again giving them the ability if they chose to spend more time thinking about their plan. Isn't it more likely that the troubles that poor people have with money comes from lower intelligence? After all we know that the very poor tend to have the lowest intelligence, and low intelligence people struggle with even simple maths (like simple division and addition). Occam's razor surely suggests that this is the most dominant factor.

Libraries are free.

My local libraries are extremely lacking in material above the 6th grade level.

It doesn't seem likely "intelligence" as a genetic limitation is much of a factor. More likely, the poor choose not to invest time in learning skills, especially marketable skills (there are some very smart poor people btw!).

But learning is a cumulative process, so intelligence is sticky.

You say that people should have more time to plan, but I think you miss key points. The most important is, "time" is not the relevant resource; it's something closer to "energy" or "focus" that we don't have good language for talking about. The second is, I think you may underestimate just how much of a time and energy sink it can be trying to keep social services happy.

Really? Spending on repair the amount comparable to the total car's price is "smart" for the poor driving that car? This from "one of the most significant economics books of the year". Are you trying to compete with The Onion?

It's a bullshit question. Marie explains exactly why.

Axa, at least in my hometown, you better believe every poor person does some of their own repairs or knows a local mechanic and then barter/negotiates the price down. Could be trading services (plumbing work for van fixed, babysitting for family friends), could be cooking for them and their family, could be you buy the parts and then owe them for installing it. My brother became a MechE because he loved working on cars and maintained all the family vehicles for anything short of work needing an engine hoist or tranny work.

I think IQ isn't rigid, but rather sticky. Learning is a cumulative process, past the age of 15 or so there's just an incredibly daunting amount of ground to make up, and the people who are better habituated to being able to do so aren't the people who have to do so.

As I noted over at The Atlantic article on this, this is complete rubbish.

Wishful thinking, I'm afraid.

These guys are trying to pull a fast one and overturn all the evidence linking IQ and wealth, both on an individual and group level.
Why then is IQ predictive of earnings within families? Why of IQ and childhood SES, IQ is by far the stronger predictor of future earnings?

The differences between higher IQ and lower IQ people become more evident on harder tasks, because there's only so good one can be on easy tasks.
I'm a big proponent of making the lives of the poor easier; even a poor citizen of should have a place in society. But let's not fool ourselves into thinking that that's going to perform miracles and boost their brainpower. Mountains of evidence simply say otherwise. See:

HBD Fundamentals: On the reality of IQ | JayMan's Blog

This is a clearer explanation:

It is clearer. But it does still hold on to some assumptions that don't seem verifiable, particularly the assumption that poor people are making bad decisions.

The article uses as its examples thing like not taking medications, meeting appointments, or being attentive parents. These are all very heftily subjective in both their definitions and their weights. Take attentive parents -- a poor parent may be assumed inattentive if she gives her child a good deal of independence, or if her older children have responsibility for younger children, etc. But a middle class or wealthy parent who sees her kid a third of the time the poor parent does will be considered attentive if she's on the PTA board. The kind of people who create the studies are biased towards assuming their own values and expressions are universal truths.

Taking meds is another good example -- I'd believe a theory that says people with medical problems are more likely to struggle and therefore to be poor. Therefore, poor people will be much more likely to have had much more experience with medical professionals than wealthier people. Therefore, poor people will be much more likely to view with a skeptical eye promises of medications or therapies, give much less weight to the importance of taking X medication religiously. The thing is, there's nothing objectively that says that perspective is inaccurate, it's just that study authors tend to be wealthier and professionals who tend to trust doctors and pharmaceutical companies more than poor people might. But their trust is not any more solidly founded than the poor man's distrust. So to say the poor man's decision to underplay the importance of medicine is a bad one, without noting the possibility that the rich man's potential decision to overplay its importance (and overuse it -- we're a hypermedicated society by anyone's standards) might also be a bad one, makes the foundation of the study shaky.

However, having just filled out Medicaid paperwork I can attest to the fact that my brain now hurts and I probably shouldn't be called upon to do algebra right now. Or, say, write a coherent comment.

Comments for this post are closed