Who gets what wrong?

My colleague Daniel Klein reports from the front:

…under the right circumstances, conservatives and libertarians were as likely as anyone on the left to give wrong answers to economic questions. The proper inference from our work is not that one group is more enlightened, or less. It’s that “myside bias”—the tendency to judge a statement according to how conveniently it fits with one’s settled position—is pervasive among all of America’s political groups. The bias is seen in the data, and in my actions.

And what do the “right-wing” thinkers get wrong?

More than 30 percent of my libertarian compatriots (and more than 40 percent of conservatives), for instance, disagreed with the statement “A dollar means more to a poor person than it does to a rich person”—c’mon, people!—versus just 4 percent among progressives. Seventy-eight percent of libertarians believed gun-control laws fail to reduce people’s access to guns. Overall, on the nine new items, the respondents on the left did much better than the conservatives and libertarians. Some of the new questions challenge (or falsely reassure) conservative and not libertarian positions, and vice versa. Consistently, the more a statement challenged a group’s position, the worse the group did.

A college education, by the way, doesn’t help much.  Here is another statement of the conclusion:

A full tabulation of all 17 questions showed that no group clearly out-stupids the others. They appear about equally stupid when faced with proper challenges to their position.

That’s a lesson David Hume would have appreciated.


“A dollar means more to a poor person than it does to a rich person”

A dollar allows both a rich person and a poor person to buy a soda. It won't change either of their lives much. $10,000 on the other hand...

A dollar might be able to keep a poor person from watching their kid starve to death. That likely can't be said about the rich person.

$10 would probably be a better amount to use.

In a Third World country maybe. Not in a country where 50M people are getting food stamps.

Umm.... decreasing marginal utility.... maybe.

So, you're stupid if you don't interpret the ancient word "mean" consistent with a particular and still juvenile academic sector's model? That's stupid.

Not necessarily stupid, just ideological to the point of being pedantic.

You're right, but i've been told parallel construction is an effective rhetorical tool.

They should have said a marginal dollar, or one additional dollar, or something like that. It is literally true that a dollar buys the same goods and/or services whether you're rich or poor. It's also true that money had diminishing marginal utility. The question is just not clear. In this case I think the pollsters clearly out-stupid everyone.

I think the key is the "means more" phrasing. They are trying to imply marginal utility in a way that is understandable to the average respondant to the survey. I guess there is an issue with that being sort of a personal emotion of the recipient of the dollar though. I am not sure the best way to phrase that question to mean the same thing to everyone who reads it. If it said "A dollar does more for a poor person..." you could still argue that it does the same thing for everyone, increases net worth by $1. This seems like some strange psych project.

Correct. It's a ridiculous query. There are poor people who'll buy Hawaiian Punch rather than milk or eggs. There are rich people who buy stupid stuff they can't use just because it's on sale.

Someone can legitimately disagree with that statement on the simple grounds that you can't make interpersonal utility comparisons.

Now if only we could believe that 40% of conservatives understand that.

@joshua: You might want to re-read that bit about "no group clearly out-stupids the others. They appear about equally stupid when faced with proper challenges to their position."

You would have to word it like "an additional dollar has relatively more value for a poor than it does for a rich person". Even though it is quite apparent what concept they are "referring to" - like it stands the question does not have a right answer (although I'm not sure it would have changed anything).

About halfway through a Law and Economics course in law school the prof. through out, as a given, that a dollar means more to a rich man than a poor man. Needless to say I stopped attending the class.

Many silly things happen under the L&E banner. Its kind of like the law's version of deconstructionism.

"through out"?

You probably should have kept going...

Perhaps the dollar does mean more to the rich person. That's why the rich person held onto all his dollars and became rich, while the poor person spent all his dollars and became poor. A rich person might beat you to death for trying to steal a single dollar, while a poor person might see that you need the dollar more than him and give it freely.

If I've swayed (or reinforced) your opinion, you need to understand that what I've really done is given the question a context that changes its meaning. The question certainly does not imply that the rich person was OCD with his money and lived like a pauper so he could amass a fortune, while the poor person had a complete disregard for his money and didn't care to keep any of it. Of course this can happen, but it is an exception. It's a manufactured scenario specifically designed to defeat the general concept of rich and poor.

If a rich person does not have that one dollar to spend on something, he'll use a different dollar. A presence or absence of a single dollar does not change the rich person's purchasing power enough to sway a purchasing decision. The one dollar means nothing. If the poor person does not have that one dollar to spend on something, the poor person may not be able to buy something he otherwise would have. The dollar means something. Maybe it's life or death. Maybe it's a soda. Either way, it's something. Something is more than nothing, so in the general case where no other information is available, the dollar means more to the poor than the wealthy.

Not entirely fanciful. For a century, the Scots were notorious skinflints in this country. Guess what their per capita millionaire rate was in 1990?

If a rich person does not have that one dollar to spend on something, he’ll use a different dollar.

Or he'll go without.

If the poor person does not have that one dollar to spend on something, the poor person may not be able to buy something he otherwise would have.

That presupposes the nonexistence of a welfare state. In fact, it is even possible that a marginal dollar will have negative value because of income cutoffs.

We'd better yank some Nobel prizes from some rather famous economists like Kenneth Arrow, Paul Samuelson and others who recognize what others seem to dispute, diminishing marginal utility.

The only debate is whether or not the questions are too narrow to be useful. If you use a historic definition of poor (that is, often hungry), then two things happen. 1) Almost NO ONE in the US today qualifies and 2) the correct result is not highly disputed.

If you use the outrageous "poverty line" spew from our government, then it is HIGHLY debatable, for reasons we see in these comments.

And there is at least one Nobel prize winner whose articles are regularly bandied about this blog who just as regularly demonstrates himself to be a fool, so your appeal to authority gets a big raspberry from me.

OK, let's play a game.

If you say rich man with a million dollars values his marginal dollar as much as a poor man with $20k, let's see where that takes us.
Let's equate both sides of the equation and subtract $10k from both persons, because, as you said, they each value a dollar equally.
The person who had $1 million, and must be feeling bad because he lost $10k, and the other person now has $10k.

Then let's subtract 10k again from both sides.

OOOP--sorry, lost the other guy.

He must be hurting as much as the millionaire because they both suffered the same loss, but the other guy is dead from starvation.

NOT HERE, NOT NOW, HE IS NOT! Because the government is going to swoop in and give him benefits the value of which has be computed to be as high at $40k/year. (I have no idea if that number is accurate today, but it ain't chicken feed.)

Can't do that. Got to stick to your guns even if it leads to absurdity.

Maybe the other guy could sell himself into slavery.

Huh? I'm sticking to my guns just fine. My point is that the perversities of government-run "welfare" include the fact that a marginal dollar of income might well be taxed in excess of 100% for a "poor" person in this country. A person in such a situation might well face negative marginal value from the marginal dollar.

Right Wing, you may be sticking to your guns, but logic and the example just shot you down, as you had to bring in welfare support to solve the obvious problem.

The only obvious problem I see is that we have a welfare state. If you wish to restrict your arguments to universes other than the one we inhabit, please specify in advance.

"They appear about equally stupid when faced with proper challenges to their position."

I wish I'd written that.

Link to the quoted Klein article: http://www.theatlantic.com/magazine/archive/2011/12/i-was-wrong-and-so-are-you/8713/

"no group clearly out-stupids the others"
That phrasing cracks me up. Anyway, kudos to Dan Klein. I tend to think of him as a whiner for how libertarians are a minority in academia, but here he's making admissions against (political) interest that simultaneously advance his research program.

A dollar itself is only valuable to the extent that it can purchase goods. So the value that the poor person places on the goods, vis a vis the rich person, is what matters. I don't make blanket statements comparing the subjective utility of different people, so I would have gotten that question wrong.

Also, there's this question "when two people complete a voluntary transaction, they both necessarily come away better off." They have agree as the incorrect response. Again, I think this is a rather murky proposition. Ex-ante people would not have made the choice if they did not think they would be better off, but ex-post it is possible for them not to be better off. If the question specified ex ante vs. ex post better, I think more people would have answered the question as the authors expected.

These two questions are also a bit tricky: drug prohibition fails to reduce people’s access to drugs and gun-control laws fail to reduce people’s access to guns. They have agree as the incorrect answer to both. Very clearly, drug prohibition has not reduced people's access to drugs, it has only increased the cost to obtain them. The same should be the case for gun control. However, if people simply followed the law, then yes drug prohibition would reduce access to drugs. The whole point of the economic analysis of black markets is that people get around them.

John: Don't you think a poor person would do more for a dollar than a rich person would? Like, say, bend over to pick up a dollar on the sidewalk. Or think of it another way. If a rich guy loses a dollar through inadvertence and a poor person loses a dollar through inadvertence which is likely to feel worse about it? And if a dollar doesn't work for you, make it a hundred.

Rich people will often tip a waiter $50. I suspect poor people rarely do.

"Rich people will often tip a waiter $50. I suspect poor people rarely do."

Maybe rich people value a waiter's service more than poor people do. And they value the dollar the same.

"Don’t you think a poor person would do more for a dollar than a rich person would? Like, say, bend over to pick up a dollar on the sidewalk."

Only because at $1000 per hour once he gets to work that dollar isn't worth the lost revenue.

Come on, people. Replies like this make you look like assholes when you all know damn well that a poor person values an individual dollar (as opposed to a collection of, say, 100 million of them) far, far more than a rich person does. You can get a sandwich for a dollar - it can literally prevent starving.

I don't know, I mean revealed preference or something...

It's not like there are many people in America starving for want of money. If you include the poor in war-ravaged Africa, sure, I'll grant you a dollar means a lot. But if a dollar meant a lot to American poor, don't you think they'd have fewer TVs and more healthy food? Don't you think they'd be working more?

The question--declining marginal utility--should have been phrased more generically. Many people are poor precisely because they have no sense of value due to their high present-orientation. They'll use the first $5 in their pocket to buy cigarettes, then go yell in Zucotti Park because they can't afford bread. There are rich people who spend hours clipping coupons. The questions are pretty loaded with the pollsters preferences.

You didn't get the point yet; the implication of ordinalism is that while we have a clear meaning for 'I value A more than B' (it defines some ordering), the proposition 'X values $1 more than Y' lacks any meaning.

Yes, $1 may prevent the poor starving. Still, 1+1=2 and it is perfectly reasonable to argue that it is incorrect to say that "Rich value $1 less than the poor". (it is also incorrect to say that "rich value $1 more or equal than the poor; the question is undecidable)

To explain it in clear terms: I value $1 more than an apple means: I have some ordering (at least somewhat transitive...), according to which I decide, and on this ordering $1 is higher than an apple.

I value $1 more than you value a $1 is.... well... akin to I love you more than John does. It seems to me perfectly reasonable to argue, that a question "rich John usually/always loves me more than poor Andy" is undecidable based on the fact, that I have no idea what it means "love more" (or "value more") in this context.


Yes, we know what progressives think they mean by one of their slogans.

You can get a couple loaves of bread for a dollar. At one point in my life, I spent my last dollar on two loaves of bread and that's all I had to eat for about a week and a half. (Actually, it was something like 96 cents, mostly in pennies). I agonized over spending that dollar - I had one dollar and spent over two hours in a grocery store trying to figure out what I should buy. What would feed me the longest for the least amount of money? I'm sure I spent more than half an hour just looking through the bread to make sure I got the cheapest bread. I lost 20 lbs in 10 days, and I was not overweight when I started. (Low cash diets are surprisingly effective).

You can say that I could have or would have found another way to eat, and you're probably right. But you sure as hell can't tell me that the one dollar didn't matter. It meant I didn't have to beg for money from strangers. It meant I could look for a job instead of looking for food. Not in Africa. In Chicago.

This is what it means to be poor. Dollars matter. The "American poor" you are talking about aren't actually poor. You're talking about the welfare recipients who choose to let the state support them rather than work. I hate that we pay people to sit around watching big screen TV's and playing XBOX who are fully capable of earning their own living. Don't confuse these people with the poor.

> Don’t confuse these people with the poor.

I didn't invent the term.

Actual "poor," in any real sense, are vanishingly rare in the USA.

I guess I'm saying: "Is the question about poor people like how we normally use that term in this country?" If that's the case, then, revealed preference, the answer is no the poor don't value a dollar more than rich people, they value it less. That's most of the reason why they're poor, if they aren't children.

If it's about mythical poor people, or poor people in really awful situations, then revealed preference is probably not a good tool for disambiguation and you get into the what's-my-utility-relative-to-yours questions everyone is going on about.

A la Mark M, it would be very interesting to have a description from each person as to what question they thought they were answering and what they think they meant.

Mark M, I would urge you to read the reply from andy. He's making the economic argument. We're not talking about your individual case. I can't see inside your mind to know how much you really prefer anything. All I'm saying is that I cannot compare the preferences of one person to another using economics. If you can using some other means, besides economics, then fine and there might be value in that. But that's not economics.

Andy, I got the point. My own point is that when people start arguing ordinalism, they also start looking like assholes to people who have actually been in the situation of having to count every last dollar, or every last penny for that matter. And for people like myself who have been on both sides - not having any money at all, and then having a good amount of money - it's abundantly clear that dollars, like just about everything else, mean more to you when you have fewer of them. What is there to argue here?

Many of the comments here, well-meaning as they may be, just come across as obfuscations from a bunch of people who would have given the wrong answer to a very simple question.

Jason, do you see why a person who stopped being poor because he valued dollars enough to do something about being poor might not be the best person to tell people what the poor think about the value of dollars?

I agree with Andrew' about having a description of the question you thought you were answering.

Let's not forget the context of the questions in this survey we're discussing. All of these questions are designed to trick those with a certain philosophy to interpret it incorrectly - or rather to showcase bias by allowing you to frame the question in the context of your belief system. To get the "right" answer, you need to stick to the information provided in the question. When you try to apply examples from your experience, you get interpretations that do things like change "poor" to "American poor," which is not the same question.

Don't over-think it. It's a general question. In common usage of the words rich and poor, rich people have more money than they need, and poor people have less money than they need. Rich people would neither buy more nor less if they gain or lose a single dollar. They would simply be $1 more or less wealthy. Poor people would spend it on something. Maybe bread or a soda or a lottery ticket. Whatever it is, it's more than nothing. That's it. The end. There's nothing more to think about.

If you start to think about it, you start thinking about whether the poor person needs it to stay alive or whether the poor person deserves it. You might start thinking this is the beginning of an argument in favor of redistributing wealth from the rich (who have more than they need) to the poor (who don't have enough). It isn't an argument for that. You might think that you'll be rewarding the shiftless poor who squandered their money, and penalizing the wealthy who scrimped and saved and went without to become wealthy. Why should those who've worked for their money give it to those who squandered theirs? They shouldn't, but that doesn't have anything to do with the question that was asked.

I don't think this is exactly what the "wrong" people are saying, but you could ask if the poor have a higher propensity to consume. The rich will save or invest. The "wrong" answer is to think that saving and investing is just as meaningful as consumption.

John Hall -

I've read Andy's posts, and they are pretty non-responsive to the original question, which asked about the relative meaning of the dollar between two general groups of people who have too many and too few dollars. Andy believes that it is impossible to compare what the dollar means to them - that it's not quantifiable.

But we aren't talking about nuance here. We aren't talking about which of two chocolate lovers likes chocolate more. We're also not talking about two specific people. We're talking about general groups of people. We're talking about whether people, in general, who have a shortage of a vital resource would value a small amount of that resource more than people, in general, who have a surplus. A has more than enough X. B has too little X. Does a small fraction of X mean more to B than to A? Absolutely.

While you can't apply generalizations to individuals within a group you can use generalizations to come up with general answers.

In general, if you have sufficient surplus, it doesn't matter if your surplus is reduced or increased by a small fraction. It has no meaning. In general, if you have a significant shortage of a vital resource, you will notice when you receive or lose a small fraction. It means something.

"Does a small fraction of X mean more to B than to A? Absolutely."

Me: The word "more" in "X means more to B than to A" - the comparison operator - is not defined. Therfore this sentence lacks meaning, and therefore it cannot be true.

You: I got the point. Still, X means more to B than to A, it's obvious isn't it?

You need to get a class in logic.

>You can get a sandwich for a dollar

Where? Myanmar?

"drug prohibition has not reduced people’s access to drugs, it has only increased the cost to obtain them."

Isn't this the exact same thing?

Not necessarily. During the whole debate about the "access" to health care I was pulling my hair out. If I knew the reformers were thinking marginally I wouldn't have been as irritated.

Is "reducing access to guns/drugs" the same as "makes it more difficult to obtain guns/drugs"

I think you'd get a different answer profile with those two different wordings. Access to me has more of a threshold meaning. I had more access to drugs in high school in the sense that today I have no idea where I would go to get them.

The rich person will not purchase either more or less due to the gain or loss of a single dollar. A poor person will. It's as simple as that. The meaning does not need to be profound in order for it to have meaning. It could just be a candy bar.

The voluntary transaction question says "come away better off." They are implying a future state of being better off, so you can't simply dismiss the consequences of the transaction. If you want to disregard future outcomes, then I am certainly no better off after buying insurance, even though I buy it voluntarily. I may eventually be glad I bought it, but I may also eventually be sorry that I bought it. You can't argue that one of these future states is valid and the other is not.

Raising prices, all by itself, reduces access to goods. I've had friends that definitely would have bought more drugs if they had more money. Controlled substances are routinely discovered and destroyed. Reducing the supply reduces access. Drug dealers are arrested and put in jail. Not all of them are arrested and those arrested do not stay in jail, but these arrests disrupt distribution, which reduces access.

Control measures certainly do not eliminate access. But they do reduce access. The same with gun control.

(I'd have gotten all the questions right, by the way)

I guarantee you that the vast majority of people would find that they had greater access to drugs if they could find them next to cigarettes at the local convenience store.

What if they were by the cigarettes, but cost twice as much as from your dealer?

Again, the term "access" is problematic. I suspect many people answer such a question with the sentiment that anyone who really wants something can get their hands on it.

You would still have greater access then if the price from your dealer was the same and you couldn't buy them at all from the convenience store.

Is there a meaningful difference between "I can get them at a price I'm willing to pay, without any particular trouble" and "I can get them even cheaper and with less trouble"?

After all, in both cases, the target person can acquire the item in question, again, without any particular effort.

There is a difference, but to call the answer right vs. wrong depending on how you weight the definition of "access" and its ease or magnitude is... bad design, at best.

(If the price of a good is raised 1/1000 of a percent, such that one person, once, decides to forgo purchasing it one time, "access was reduced" that one time, right?

Should any of us ever take seriously a claim that that's "reduced access" for purposes of the correct answer to a question purporting to ferret out bias?

I can't speak for anyone but me, but I sure as hell don't take it seriously.)

Agreed - "reduce access" is uselessly vague, in use and interpretation.

Gun control laws as we have them do "reduce people's access to guns" ... law abiding people, that is, and the non-professional criminal who isn't going to put much effort into getting one*. And they might make it marginally more expensive for a criminal to manage it.

The problem there is not, I think, confirmation bias, but confirmational interpretation (to coin a term for lack of knowing a better one); someone on the Left is more likely to think any access reduction to any actor counts, I think.

Likewise, someone on the Right, or a Libertarian, is more likely to interpret the question itself as asking a different thing - the question they themselves view the issue as relating to, which is reducing effective access to guns to those gun control laws are notionally intended to disarm; the criminal element.

It's right that the "correct" answer on the strictest possible reading is that gun control does do what the Left tends to think... but it's rather less meaningful than presented.

I read the entire question about gun control as a complicated (and ironically probably not intentional, in this sense) "trick question" about careful parsing. This would be useful if one wanted to ferret out confirmational parsing... but that's not the stated goal.

(* Eg someone with a disqualifying conviction, but who also isn't a career criminal who'd go after a black market gun. Of course, as a matter of judging the utility of the policy, these are the least important people to disarm... thus it rather undercuts the question's utility for measuring confirmation bias.)

I wish we could know the behavioural basis for this: Rationally all of us should prefer to be correct than be ideological, to solve each problem as it is, than as we feel it should be, and yet reason happens so seldom.
The marginal utility of lazy positions is higher perhaps than working things out from first principles; or it may just be that we identify with the ideology and it becomes a part of our identity and losing our identity is expensive, or it may be that we just cannot solve problems as they are; if we are to understand a problem it is via our ideological selves and hence a lot of understanding is faulty. The safety of fixed positions is immense, it is as if you belong to a multitude and are not alone, and possibly a lot of us would prefer if we are a part of the Many in Error than the lone ranger who is right.
After all, what matters more? Is it Safety or The Right Answer to a problem you may not care about?

We do. The conscious mind is primarily a confabulator that rationalizes the choices we have already made emotionally. Being a partisan is making an emotional commitment to a "team," be it a sports team, brand, or political party. This just piles on the already ample empirical evidence that being a partisan makes you dumber.

And here I thought that interpersonal comparisions of marginal utility were not meaningful. If I could just shed that foolish belief it would be clear to me that the dollar does mean more to a poor person than a rich one.

I do know that a dollar meant more to me when I had fewer of them, but that's just due to diminishing marginal utility.

Wait, SEVENTY-EIGHT percent think it doesn't limit access to guns? What if, instead of answering incorrectly, they just know something the rest of us don't. Those might be some really well armed libertarians.

It might limit my access to guns, but maybe not so much for malefactors. In fact, if the law-abiding have a harder time arming themselves, the value of a gun to a criminal increases.

You use the word "limit" while the quote above uses the word reduce. If you read the question as 'limit' and assume that means limits entirely you will marginally answer that gun control laws fail to completely eliminate gun access.

Your local hood rat would probably find his access to guns less limited if he could buy one at the local convenience store without a background check instead of having to seek out a black market seller and pay him a premium.

You may be right, but Rich Berger's point is non-trivial. It is like the zero-tolerance policies in school. The person with nothing to lose doesn't care.

It's kind of a poorly-worded question, since the target group isn't specified. Gun control certainly does limit access for law-abiding folks, but has no demonstrated effects one way or another for criminals. So I guess when taken together that means the answer is a yes, but clearly many of the people who read the question will be thinking of its effects on criminals—the target group of the laws themselves (ostensibly).

When I worked in a quality department, several times per year I had to explain to someone our definition of "limit." It would go something like "We are only slightly over the limit!" "But you are way over the target." "But we are only slightly over the limit!"

"Seventy-eight percent of libertarians believed gun-control laws fail to reduce people’s access to guns"
= gun-control laws increases access to guns ?

on the other hand i think gun-control laws as implemented dont work.
is that the point of the question ?
i agree with that and am not a libertarian.

I believe that gun control laws increase access to guns, because it creates a black market for guns.

Since a huge portion of what you pay for guns is liability, taxes, licencing, black market guns tend to be much cheaper and unregulated in functionality (i.e., you can purchase full-auto).

“A dollar means more to a poor person than it does to a rich person”—c’mon, people!—versus just 4 percent among progressives..

96% of progressives get it wrong, "only" 60/70% libertarians are wrong. C'mon, libertarians, back to school, this is still pretty bad number.

C'mon, if you accept ordinalism, the question is undecidable....

Ambiguity seems the glue binding together the questions on this study.

I think it's way simpler than most of the explanations and not exactly about being wrong. I'd bet the more closely worded the question is to the other side's mantras (or their sides characterization of those mantras) that someone knows is "wrong" then they will answer the question as the statement being wrong.

For example, some people will almost define the question of gun laws away because someone absolutely determined to get one can still get one. And yet, it is harder for some marginal person to obtain a gun and that is the problem because we shouldn't be concerned with the average law-abiding citizen getting a gun.

I would have gotten them all correct and can, therefore, declare myself free of bias and stupidity. And I would have gotten them correct because I would have recognized the traps being set. Drug prohibition is mostly (but not totally) ineffective (with many horrible side effects), voluntary exchanges nearly always make both parties better off (but there are rare exceptions -- such as when one party is misinformed), and anything two parties do (or one party does alone), whether an economic activity or just having a conversation or scratching an itch, can have some effect on others (the butterfly effect and all that) -- but that does not give others the right to interfere.

"drug prohibition fails to reduce people’s access to drugs (agree) "

I guess this one confuses me... We've had drug prohibition in this country for most of the 20th century, drugs are cheaper and just as available (if not more so). How can "Drug prohibition reduces people's access to drugs" be the correct statement?

Is it just relative to no prohibition at all? A slight inconvenience versus complete availability?

In answer to your last two questions, yes. Don't make it more complicated than necessary. You can buy cigarettes in more places than you can buy pot. Just because drugs are widely available and you might know where to find them doesn't mean that prohibition hasn't reduced access at the margin.

Who said the economists are right?

Economic denialist!

"A cookie means more to a skinny fellow than a fat one"

Surely not. Does obesity violate diminishing marginal utility?

It certainly violates diminishing!

I would guess that many rich people are willing to work more hours for more dollars, while many poor Americans are not.
"means more" is too unclear.

I would guess more "rich" folk eat at home more often than eat at fast food restaurants, where many poor folk eat more often out at fast food (including pizza). Looking at actual behavior that people really did in order to save a dollar would give more insight -- but reality doesn't fit well within low cost experiments.

People mentioned some of the problems with the questions already. The funny thing is to try to imagine how to phrase these questions in a non-biased way. If you ask whether a poor person and a rich person would choose to buy the same things with a dollar I bet 100% of people would say no. If you ask whether the poor person would choose to buy something more important... What is the answer here anyway? Remember those bums who have the 'Give me some money to buy beer' signs?

I think it is impossible to measure economics knowledge this way.

What do they mean by "more?" To a poor person, a dollar means lunch off the dollar menu. To a rich person, a dollar means "with this, I can get some poor person to do some crazy ass shit."

Okay, so a neutral way to ask it would be, "a poor person given one more dollar is more likely to spend it on food than a rich person given one more dollar".

Then again, you would have to specify "how poor" because if we're talking hobos, that money's going stright to cheap liquor, since they can get their nourishment from homeless shelters.

How about

"If you had a dollar to give away would you rather to a poor person or a rich person"

And introduce an orthogonal value element to the question?

Hehehe, now the follow up thought I had to this little discussion was: how come the people who did the study didn't think of all this? maybe their agenda was to make one of the sides look more stupid than the other.

But then again, which one? :-)

maybe that one dollar to the rich person goes to funding "rich person's school of fishing". In that case I would give it to the rich person.

While I think the dollar means more to the poor person, that dollar used by the rich person probably means more to the poor person than the poor person thinks.

Shouldn't they just phrase it to talk about the same person. "A person would tend to value a dollar less after they win the lottery than before."

Maybe questions seem more poorly worded the more they deviate from my established beliefs.


Basically, the proper wording of a question that 96% of liberals agree with is "do you agree with liberals?"

Well, I would argue that questions asking about general tendencies are better than trick questions about edge cases. So, the question, "Do voluntary exchanges between two people usually make both better off?" is a much better question than "Do voluntary exchanges between two people always make both better off?". That is a trick question intended to induce people who (correctly) believe the first statement is true to get the second one wrong. And it's also intended to give credit for a correct answer on the second question to people who incorrectly believe the first is false (which is a much more important question to be right about). So, it's clearly designed to trip up libertarians AND give a free pass to progressives who are wrong about voluntary exchanges generally.

Better said than about 5 pages I produced in the last iteration.

Additionally, they should ask questions NOT about the standard political footballs to get rid of rhetorical baggage.

“Do voluntary exchanges between two people usually make both better off?” is a much better question ... Yes! That is exactly what is wrong with these questions. Trick questions about edge cases.

How you frame a question matters. Behavioral economics 101.

Sure, but then the same standard should be applied to the original set of questions. "Free trade leads to unemployment." is false in the aggregate, but "free trade has caused some people to lose their jobs" is true. "Minimum wage laws raise unemployment." could just as easily be written as "Minimum wage laws eliminate jobs that would otherwise be offered at less than $7.00 an hour." The point is when you say 'raise unemployment' people are assuming you are reducing the availability of well-paying jobs*, not the availability of any job at all.

*I realize some of you may think that minimum wage laws also reduce middle class jobs as well, by hampering small businesses and the like, but I think that is up for considerable more debate and is not on the same level of economic truthfulness as those potential $4 an hour jobs which would be eliminated by a minimum wage.

The whole point is that no affiliation answers all correctly.

But it would be interesting to see if the 96% who answer the question correctly really understand the question or are just answering consistent with their rhetorical slogans.

Btw, there was a lot of criticism to the original set of questions, which is why I suspect they did it again. Now they get to rinse and repeat!

Yes. But notice that those two questions aren't about edge cases. In general, free trade does not lead to greater unemployment while, again, in general, minimum wage laws do lead to greater unemployment. It's worse to be wrong on the general principles than on exceptions to the rule.

The point is when you say ‘raise unemployment’ people are assuming you are reducing the availability of well-paying jobs*, not the availability of any job at all.

Huh? Why on earth would anybody think 'raise unemployment' meant anything other than the obvious -- namely 'increase the unemployment rate'?

I agree with you on free trade, and concede that a lot of people (including a lot of tea partiers) are misinformed on its benefits. I disagree on the minimum wage. How are the unemployed people priced out by a $7 minimum wage not marginal, edge cases? How much unemployment today do you think is attributable to minimum wage laws? Do you honestly believe that if we abolished all minimum wage laws we would see unemployment fall in any meaningful sense?

I actually don't think people think of the unemployment rate when you say "raise unemployment." The unemployment rate, whether it's 9.3% or 8.7%, is meaningless to most people. They think of the the number of people in their community who are working for a decent wage versus those who want to but can't find the jobs. I think if you said: "hey, in order to reduce unemployment, we're going to offer more jobs at less than 7$ an hour," they'd laugh in your face.

looking at some of the questions for normative bias is completely circular since the question have normative terms e.g., "Third World workers working for American companies overseas are being exploited."

Some of the questions totaly depend on the circumstances-- e.g., free trade leads to unemployment? The theory of comparative advantage suggests that those employed in the domestic production of a good that is now imported will be rendered (at least temporaly) enumployed. otherwise, what would be the point?

Do minimum-wage laws raise unemployment? Does the author think that a reduction in the minimum wage law would reduce unemployment currently? I hardly take that as a given accepted wisdom in the economic profgession.

Don't poo-poo a dollar... A dollar will buy you a steak and kidney pie, a cup of coffee, a slice of cheesecake and a newsreel. With enough change left over to ride the trolley from Battery Park to the polo grounds.

Cut and pasted. I've been trying to track that one down for a while.

No doubt each side would insist their errors would do less (and probably no) damage if translated into policy. So they wouldn't acknowledge much of a need to correct themselves even if they accepted they were wrong.

Your supposed "right answer" to the "means-more" question is a typical inter-personal value comparison nonsense, which is the moral foundation of most of the "social engineering" policies of destruction.

It's also the moral foundation of the economic argument for capitalism.

A dollar may mean more to a rich person than a poor person. Rich people spend lots of time and effort in order to acquire dollars. Poor people may say they value dollars more but do little to demonstrate that with their actions. Rich people may spend long hours at work to acquire dollars, while a poor person may spend his whole day on the porch doing nothing to get additional dollars. Rich people spent countless hours studying and acquiring specialized skills to acquire dollars. Would they do this if they didn’t value the dollar as much as poorer people? Poor people do little to get the proper training and skills to earn dollars. Certainly one can argue that a dollar means more to Rich people since they spent so much time and effort attempting to acquire them. To say C’mon people – like there is a right answer to that question shows bias of the questioner. Each individual is different – rich and poor. There is no real way to compare who values a dollar more.

Revealed preference suggests the rich value a dollar more. Endowment of ability is an important confound.

Certainly diminishing marginal utility is, if not quite bunk, then limited in its application.

A rich person and poor person are both on their way to catch a train (to meet their spouse/go to a ballgame/go home for dinner). On the way, they discover that they dropped $1 ($10/$100) out of their pocket. Do they go back to look for the missing money if it means missing the train?

Of course not, the poor person wouldn't just drop something as precious as a dollar.

I have to admit, being stupid is almost worth annoying progressives.

Where's his evidence that gun-control laws significantly reduce access to guns? I really hate one-line poll questions of this sort as there's a whole bunch of assumptions baked into the poll question that are subject to vast interpretation, particularly when he presents the poll as a sort of "examination" with right and wrong answers.

In this case, it's obvious that criminals have no problem getting guns, even in cities with "tough" gun control laws, so I would have said "no" to that question.

Would you find it easier or harder to buy a gun if you didn't need a background check and you could find them on a shelf at your local grocery store next to the bread and milk? Would a criminal?

Gosh, I wish he'd taught me the "right" answers before I took the survey. Silly me, I relied on incorrect sources, like what I have experienced or read. And I thought, in many cases "Hmm, depends on what you mean, or what the context is". Thus, I came up with the "wrong" answers.

"conservatives and libertarians were as likely as anyone on the left to give wrong answers to economic questions"

Not possible! Paul Krugman is NEVER wrong.

Upon further thought about "means more",

5. Not sure, and
6. Other

are the two correct answers.
Most of the respondents got it "wrong"; and I would have, too, probably, by agreeing -- since a dollar did mean more to me when I was poorer.
Terrible question -- which is at the heart of the problem of gov't forced redistribution.

Note that if the poor do indeed value dollars less than the rich, it follows that (even if perfectly informed) they will spend relatively little effort agitating for redistribution. Since the rich have a considerable advantage in material resources (by definition), it seems to me that the disinterested poor would be unlikely to overcome the preferences of the rich in order to enact redistributive policies. So it follows, naturally, that forced government redistribution is not a problem.

you ignore the vindictiveness of the middle class and upper middle class redistributionists.

Who foist the income of the wealthy upon the indifferent lower classes, who, having already acquired everything they find useful, fritter it into Welfare Cadillacs? An interesting if implausible scenario. I believe this should result in steeply progressive taxation, since presumably every mid-to-upper class person would attempt to lay the burden of taxation upon those wealthier than themselves, and thus each successive bracket would have a larger and more vehement constituency. If we presume that vindictiveness is proportional to the ratio of wealth between the tax-proponent and the taxee, then we should expect punitive levels of taxation on the highest brackets, modulo the efficacy of money at thwarting political change.

Notwithstanding the crack about Cadillacs, I think a scenario in which the poor value dollars less than the rich, but significant amounts of money are redistributed from rich to poor, should result in the poor having a significantly higher savings rate than the rich. For a given pre-tax income, the poor would have less desire to spend their marginal dollar, and the rich a greater desire to spend theirs. The confounding factors would be their relative discount rates and the risk-free interest rates available to both groups.

my point was that middle class redistributionists use the poor as an excuse to take wealth from the rich and then proceed to spend it on themselves. lets not pretend that government intervention in mortgages and student loands had to do with helping the poor instead of the middle class. just look at OWS. they are mostly middle class and upper middle class. meanwhile, the actual low income groups who work minimum wage, dont support OWS. many are even angry at the protesters, because they drove out customers and cost them their jobs.

Well, as I said, if it's mostly the middle/mid-upper class ganging up on the rich, then you'd expect steeper rates at the top, and (unless I'm mistaken about the distribution of capital gains) a higher capital gains rate than we have now. That's not to say there isn't misdirected redistribution, but most of the policies that benefit the middle class in that regard also benefit the rich to an even greater degree (student loans the exception, probably).

I think the most recent polls show fairly broad support for OWS, though I could be mistaken, and it's not really relevant to whether redistribution is a good idea or not.

There seem to be a lot of willfully ignorant folk in this thread.

No, these are essay answers to the questions--still wrong, but essay answers.

It's more like when one of my professors asked what kind of animals should we use for animal testing. I answered rats, or pigs if we are feeling frisky. Her answer was that I was wrong and we should use chimpanzees. Well...yeah...but NO!!

I'm curious, does everyone on here complaining about interpersonal utility comparisons also reject modern micro-founded macroeconomics on grounds that studying the utility problem of a representative agent is useless because people all have vastly different utility functions for money?

David K raises a good point, but it may be better made in a rephrased manner. It isn't just utility for money (which may or may not be directly incorporated into an individual's utility function). It's also a question of heterogeneous preferences for consumption or labor. You are correct that the representative agent approach tends to be a bit narrow in the way it approaches heterogeneity. For instance, it might consider people with different initial endowments or ages, but (as far as I know) they mostly have the same underlying utility functions.

An alternate approach would be that the parameters to the utility functions would be like random variables themselves. They could be conditional on income or time-varying or what have you. While it would not be possible to use DSGE techniques to solve these types of problems. It should be in agent-based approaches. I'm not sure what value there would be in doing so, but it should be possible.

you could probably aggregate various groups of people with similar utility functions and then make intelligent analyses based on their group behavior. It's probably a lot more complex than simply rich vs poor.

Aren't you asking if models are reality?

This is another problem with the survey. Are they asking for your understanding of economics 101, or are they asking if the world conforms to economics 101 models?

"drug prohibition fails to reduce people’s access to drugs" --- I think many people would effectively read this statement as saying "drug prohibition fails to prevent people from accessing drugs." Which I would say is correct. But I guess the point of the test is that people will misunderstand, reword or otherwise fight the statements that challenge their core believes.

Indeed, all the bellyaching above about poor people and a $1 proves this. It is obviously a very general question about marginal utility and to "disagree" is the worst answer you can give. For everyone nit-picking about context, recall that "not sure," "other," and "refuse to answer" were not considered incorrect responses.

Well, one of the points of marginal utility is that it is 'subjective' (pertaining to the one subject - individual), therefore you can quite surely say that interpersonal comparisons in general are impossible. Considering that all questions rely on a fine wording (i.e. forbdding drugs fails to reduce access to drugs vs. forbidding drugs fails to significantly reduce access to drugs), I am very surprised that for this particular question actually a very non-general, specific and I would say quite a controversial meaning is required.

A full tabulation of all 17 questions showed that no group clearly out-stupids the others. They appear about equally stupid when faced with proper challenges to their position.

If true wouldn't it be an argument for government to do less?

The most amusing part of this is the mass of commenters on the thread arguing that they're not really wrong and that liberals really are dumber than anyone else in the world.

"Seventy-eight percent of libertarians believed gun-control laws fail to reduce people’s access to guns."

I am picking on a libertarian question here, but I think it may apply to others as well. Basically, the english language is imprecise which makes interpretation bent to ones own conclusions quite easy. Specifically, gun control laws do "reduce" absolute access. They are not sold in the corner gas stations for instance. But if I wanted to get an illegal gun could I? Certainly I could. Just like drug control laws it wouldn't eliminate access to anyone. And I mean anyone, who goes looking. Ownership would certainly be reduced if guns were illegal, but that is primarily because most people are law abidding. Those who wanted a gun anyway could get one. Hence, while technically incorect, many gun freedom advocates interpret "reduced" access as essentially meaningless since access is never really eliminated.

The questions were selected to make the results look as if no group out-stupids the others. Everyone knows that [...] are objectively stupid.

OMG politics really does bring out the crazies from the woodwork. It's a simply question of marginal utility decreasing in wealth and while it wasn't explicitly stated so in the question, most people should've inferred that the question meant marginal dollar. This is not some economic theory artifact. Just ask everyone how much more time they're willing to trade for another dollar of income and plot that as a function of the responder's wealth.

The problem with claiming that the respondents should infer something unsaid from the question is that the questioner can get any result they want by changing the inference from question to question.


1)when two people complete a voluntary transaction, they both necessarily come away better off (agree)

2)rent control leads to housing shortages (disagree)

These are both declarative statements that if interpreted consistently would have the same answer, but Mr. Klein makes his bias clear, and shows how he can get any result he wants.

"Like many surveys, this one throws a single sentence at the respondent, who has to conjure a likely context. Restrictions on housing developments may have certain redeeming features. But by and large, they do make housing less affordable. After all, restrictions restrict. The odd case might turn out otherwise, but the range of responses allows the respondent to register reservations without contradicting the general claim (by answering “somewhat agree,” for instance). "

If you are trying to put together a study based on absolute questions, and you concede that there are instances when the absolute isn't correct, then the only reason "somewhat agree" is more accurate than "somewhat disagree" is because the questioner wants it to be so.

Which of these two statements is accurate:

I somewhat agree with statement that rent controls lead to housing shortages, because it's true in general, but there are certainly exceptions.

I somewhat disagree with the statement that rent controls lead to housing shortages, because it's true in general, but there are certainly exceptions.

He might as well show people a picture of a half full glass of water, and call them stupid for claiming it's half empty.

So, if I understand the dominant arguments in this comment thread, conservatives weren't wrong on the dollar/poor question because there aren't really any poor people and means doesn't mean what everyone, everywhere (including here) knows that it means?

Good to know.

I think you guys are missing the arguments. People aren't defending conservatives or libertarians or attacking liberals. They are complaining about the design of the survey.

For example, is there really decreasing marginal utility of a dollar? Yes, a rich guy won't spend 5 more minutes to earn a dollar, but that is mostly because he can spend 5 more minutes earning 10 dollars.

I'm not missing the argument, but if you have to torture the English language to support your argument, maybe you should rethink your position.

I'm not an economist, so I can't go into great detail debating marginal this and utility that, but I have made my living through various types of writing and editing over the last 10 years, so I know a little bit about language. Here we have a question asked to regular people using regular people language being dissected through the lens of technical jargon. It's just ridiculous. It's the left shoe that matches the evolution "debate's" right shoe--there a lay term is being conflated with a technical term with a different meaning that happens to be the same word.

Do your own poll. Go out on the street and ask regular people what they think that question means. How many do you think will go all Clinton-testimony on the meaning of the word "means"? Or how many of them will be confused about whether you mean "actual" poor people in Africa or America's "non-poor" (because, you know, you can't be poor if you have a refrigerator)?

I'm sure there are some problems with this survey or with the analysis. Just as there were with the previous paper. But the discussion here is plain silly.

I just read back through all the comments (what can I say, the work day is over), and not a single proposed alternative to that question is more clear and less leading, and they certainly aren't any more demonstrably "correct."

I just took a look at the actual paper. It asks everyday people, but it makes the claim that certain answers are the "unenlightened" response. Our argument is that the opposite "enlightened" response depends on economic theory, not on the plain regular everyday usage of a few thousand Americans. We're disputing the conclusions reached by the paper because we don't think some of the answers are related to ideology and are instead related to the economic way of thinking.

You ask for alternate statements/questions. Consider: "Redistributing a modest amount of income from a high-income person to a low-income person is good because the possibility of slower economic growth is offset by having a fairer society." The problem is that I'm not sure there is an "unenlightened" answer to this. It is almost entirely dependent on ideology how a person will respond (in other words how the respondent values fairness vs. economic growth). Suffice it to say, it is a sh!tty question to evaluate whether someone is economically enlightened or not.

The questions about drug prohibition and gun control are still just mystifying to me, but I might rather phrase them as "Drug prohibition reduces the amount of drugs consumed" and "Gun control reduces deaths from gun violence." I think you could safely say that the first is untrue, but I think the second one might be uncertain depending on the evidence (I recall some disagreements over this). I would say that the question about people being better off with transactions is rather obvious if stated ex-post, but just clearing it up to be that people expect to be better off with voluntary transactions would be simple enough.

"Our argument is that the opposite “enlightened” response depends on economic theory, not on the plain regular everyday usage of a few thousand Americans."

You are entirely missing the point of the paper, then.

the statement “A dollar means more to a poor person than it does to a rich person”

The Millionaire Next Door is basically written around the premise this seemingly obviously true statement is actually false..


Also, we need to add a "[User_name] likes this" feature to avoid all the clutter of people having to register that view with a comment.

Bonus points for a "... dislikes this" feature!

Reading these comments has been much more instructive than the original post. The magnitue of the intellectual contortions and rationalizations that people people will use to avoiding having their beliefs challanged is breathtaking.

the survey was riddled with weasel words. pointing this out is not an "intellectual contortion."

> Reading these comments has been much more instructive than the original post. The magnitue of the intellectual contortions and rationalizations that people people will use to avoiding having their beliefs challanged is breathtaking.

Precisely. These comments prove the point better than the survey.

This project assumes there are definite "right" answers, but questions are worded so vaguely than the opposite interpretation is at least arguable for most of the questions.

For example "Restrictions on housing development make housing less affordable".

What restrictions? 99% of such restrictions make no and pretty much no difference. Like not being able to build a house in national part, not being able to use asbestos, minimum distance between houses for legitimate fire safety reasons, etc. are all real restrictions that do absolutely nothing to make housing less affordable.

And then there's that 1% of restrictions that will make a big negative impact.

And then there are even some restrictions that have positive impact - like prohibition from racial discrimination in restrictive covenant, which is a government imposed restriction on housing development, but has definitely major positive impact.

So both "somewhat disagree" and "somewhat agree" are good answers here. "Strongly agree" and "strongly disagree" both rely on very particular reading of a vague question, and I'd argue they're both wrong.

You can do the same to almost every question they asked.

Klein is almost certainly right that you can create a survey that shows most people of a political persuasion have false beliefs convenient to that persuasion, but this is still a really bad survey. if you're going to do a survey like this, you could at least bother to ask factual questions. When you throw in imprecisions like "means more to" that everyone will disagree over, you just make the whole exercise futile.

It would have been really easy to just ask purely factual questions about things like consumer banking policy, income and wealth distributions, tax statistics, etc, and the point could have made with a lot less subjectivity.

Your concerns will all be clarified if you go to "Econ Journal Watch" and look under Support. You will see: "EJW is a project legally existing within the Atlas Economic Research Foundation." Checking out the board members of the foundation show them to be part of the GMU, Mercatus Center, et al, cabal of the conservative Koch foundation beneficiaries.

Mr. Klein should understand how to conduct an unbiased survey. However, as Upton Sinclair said: "It is difficult to get a man to understand something when his salary depends on his not understanding it.”


I don't think you actually read the article.

The declining marginal utility of wealth and income insure that a dollar means more to a poor person than a rich one. It is this utility function which is the very basis of socialism, and as a static matter without regards to the means of the re-allocation of resources, it implies that society will be the happiest (from a financial point of view), when all are about equal. Oddly, egalitarians hardly ever seem to argue using this model.

I believe the declining utility of wealth and income explains why dictatorships decay into democracies. Thus, it is a very powerful explanatory tool and the chief engine of social progress.

But make no mistake, the marginal utility or wealth and income does decline as it increases.

The 1%: explain?

That a graph of different individuals might show a decline doesn't necessarily prove a decline for an individual.

The folks who do not believe in diminishing marginal utility of a dollar, and do believe in the equivalent loss of a dollar, should think twice about how this could affect tax proposals.

So, are you in favor of a dollar cap on, say, housing interest deductions in the form of a housing allowance. Everyone gets an equal credit of a $20k deduction, true, some poor people won't use it their income is low or their rents are low, but the millionaires shouldn't care--each person gets to deduct a $20k housing allowance from their taxes and I'd even let you have a low flat rate.

Capped dollar deductions should make the rich happy because they apply to the poor as to the rich themselves, and they each have the same marginal utility for the last dollar as we have been told.

See, aren't you happy.

So, are you in favor of a dollar cap on, say, housing interest deductions in the form of a housing allowance.

Yes, specifically a cap of $0.

and I’d even let you have a low flat rate.


You didn't ask what it was for you income level and how it would step, or how all types of income should be treated.

Leaving steps aside, I an glad we can agree on caps to use that to reduce rates, making deductions even less valuable.

It seems that this question reveals more about the person answering than any "objective" answer. After all I have met plenty of rich people that value a dollar highly. Thrifty like Warren Buffet. I also know rich people that are careless with their money. A dollar here or there is a trifle. I know poor people that scrimp and save every last nickel. And I know poor people that don't worry about a few bucks here and there. So the answer of who values a dollar more is really a case of which rich person and which poor person we are discussing. Or what you happen to believe about rich people and poor people.

Benny, you seem to be the only person responding to this article that actually gets it. The more one parses the question, and how they parse the question, is more revealing about the person answering the question than the "correct" answer. We, as a society, seem to be lacking the ability to question ourselves and our belief systems, even in the face of potentially overwhelming data. To reject realities of being able to pay for social programs while wringing hands in angst is as unenlightened as dismissing any government solution as "socialist". This is the reason we have gridlock in Congress.

I might defend the conservatives/libertarians thus. I begin by admitting that, if two people are exactly alike except that one is rich and the other poor, a dollar will mean more to the poor one than it does to the rich one. But I would insist that a randomly chosen rich person and a randomly chosen poor person should be expected to differ in many ways other than their wealth. In particular, perhaps the poor person is less “materialistic”: perhaps he places a lower value on things that can be bought as opposed to things that cannot be bought. Then even though he had fewer dollars at his command, an extra dollar might be less important to him than it was to the rich person.

Perhaps the discrepant answers of conservatives/libertarians and liberals/leftists is due to the latter’s tacit assumption that rich people and poor people *are* just alike except for their wealth. Does Klein accept this assumption? Is it in fact true?

The truth is that the poor are more likely to spend on what most people would refer to as basic necessities, right?

"Psychologists would count this tendency as a manifestation of “myside bias,” or “confirmation bias...Buturovic and I openly acknowledged that the set of eight statements was biased. But these were the statements we had available to us."

Nice to see that a Libertarian can create such an effective piece of right-wing propaganda while learning well known basic psychological tendencies like "confirmation bias." "Econ Journal Watch" must either be unrefereed or the editor (Mr. Klein - co-author of the study) had undue influence. No real surprise that the WSJ editors didn't catch it - rather justifying Krugman's assessment of that publication.


Repeat after me: interpersonal utility comparisons are impossible.

Also there's no justification whatsoever for claiming that drug prohibition and gun control work. They don't.

Comments for this post are closed