Interpersonal Utility Comparisons

One of my least favorite arguments against redistribution is the claim that you can't make interpersonal utility comparisons. For all practical purposes, this is dead wrong. Granted, it's true that you can't make interpersonal utility comparisons with 100% certainty, and it's also true that in many cases you can't even come up with a reasonable guess.

But in some cases, particularly those in which proponents of redistribution are most keenly interested, you can guess with near certainty which of two people will value a particular resource more. In virtually all cases, taking $1,000 from a billionaire and giving it to a starving beggar will help the beggar more than it hurts the billionaire. No, it's not true 100% of the time, but you don't need complete certainty; if you're right 95% of the time, that's good enough for government work.

What I particularly dislike about this argument is that the flaw is obvious to anyone who isn't already a libertarian. When we use this argument and defend it to the death against obvious counterexamples, it confirms their view of libertarianism as a utopian philosophy that can't deal with the vagaries of the real world. That's not true, and we shouldn't give them cause to think it is.

I suspect that the reason many libertarians cling so tightly to this claim is that they believe that to concede the possibility of interpersonal utility comparisons is to concede the case for redistribution. Nothing could be further from the truth. Granted, if we're dealing with a brief, isolated slice of time, ignoring both past and future, redistribution makes sense. If there are a bunch of people who are much better off than others for no apparent reason, and if we don't have to worry about the future, why not divide things up a bit more equitably?

But life isn't a one-off game. When we consider the past, questions of desert arise. Redistribution may seem just when we pretend the world was created yesterday but manifestly unjust when we acknowledge that differences in wealth are generally earned.

The case against redistribution becomes even stronger when we consider its effects on the future. Redistributing wealth from the young and the wealthy to the old and the poor tends to reduce savings, which ultimately depresses wage growth. Ditto taxes on wealth and inheritance. Taxing wealth-creating activities discourages them, and subsidizing behavior that leads to poverty encourages it. Confiscatory rates of taxation on very high incomes discourages high-expected-value, high-variance activities like entrepreneurship. And the list goes on.

The impossibility of interpersonal utility comparisons is not a hill to die on. So don't.


Related posts:

Why IUCs?
Cardinal Schmardinal, Ordinal Schmordinal
Encoding Happiness
IUCs and the Law of Large Numbers
No Soul Suggests IUCs
Futilitarianism
I-CDDFP
I-CUP
Love and Intrapersonal Utility Comparison
What color does a submarine weigh? (True or False?)
Exploding IUCs on the roadside
Pareto Efficiency and Justice
Can the Paradox of the Non-Comparability of Interpersonal Utility be Resolved?

Share this

Brandon, Your post strikes

Brandon,

Your post strikes me as wrong on multiple levels. First of all, libertarians frequently argue for capitalism on the basis of interpersonal utility comparisons. That is what Kaldor-Hicks efficiency is - it is a brave (but I think ultimately flawed) attempt to compare utility across people. Using Kaldor-Hicks efficiency, it is possible to make one person better off and another person worse off and still have a more efficient economy, which is the hallmark of a theory that treats utility across people as comparable. (The particular mechanism of the comparison is interesting: one imagines a hypothetical transfer of some quantity of good from one person to another. The good transferred plays the role of the "common unit" of the two utilities, making them commensurable.)

When I argue against interpersonal utility comparisons I do so knowing that I am depriving the defenders of capitalism of their tools. How in the world are you going to compare two whole economies and say one is better than the other unless you have some way of aggregating all the benefits and costs over everybody in the economy? I don't really see how. How, then, are you going to argue for the superiority of capitalism over socialism without it? Again, I don't know how. Well, I can begin to come up with an answer, but the argument is nevertheless greatly complicated.

So, no, interpersonal utility comparisons are a major tool in the pro-capitalist arsenal since they are what allows us so easily to pass judgement on whole economies and therefore to declare capitalism far superior to socialism, and I don't get rid of them in order to bolster the case for capitalism.

As you yourself have touched on, there is the matter of history, which makes things a lot more difficult for the redistributionists. If we redistribute today then in the long run we will *in the aggregate* be much poorer than we would have been if we did not redistribute. Because when we redistribute in effect we are eating the seed corn of our collective future.

My reason for rejecting interpersonal utility comparisons is entirely theoretical. Once you understand what utility is - that it is a matter of preference, and therefore must be given the mathematical structure of preference, which involves statements like A-is-preferred-by-P-to-B which do not have any place for interpersonal comparisons - you don't want to go back to the naive, and false, theory of utility as a scalar quantity. That's what it's all about. Similarly, once you understand that prices are created by the market, you no longer want to go back to the naive, and false, theory of value as somehow inherent in goods.

Well luckily for the IU

Well luckily for the IU crowd the question of a million dollars for a beggar vs. the millionaire is not, in reality, the most popular form of distribution. It's mostly from the middle to the middle.

You don't need

You don't need commensurability of utility to argue that there should be redistribution. It's a conceptual error to think so, because you are mistakenly confusing the everyday concept of a benefit with the economic concept of utility. People who argue for redistribution needn't invoke utility at all.

People who use the word "utility" either know economics or they do not. If they do not then they are almost certainly misusing the word to mean something ordinary. If they know economics then they understand the point about incommensurability and either accept it or do not accept it but in any case they understand where the point is coming from. It's a theoretical point of economics, similarly to the point about price not being intrinsic to the good.

If someone is suicidal people don't try to help them kill themselves. Someone who was trying to maximize the suicide's utility would help the suicide. Most people do not, because they perceive suicide as a harm to the person regardless of the person's preferences. Their goal is, first, to prevent the suicide from being successful, and second, to alter the suicide's preferences.

So benefit is not the same as utility, and people who treat them the same just are confused. And people who argue for giving money to the poor on the basis of a utility comparison are just making fools out of themselves, because the argument is silly. The advocacy is not silly but the argument is silly because it confuses concepts, it misapplies the idea of utility.

Why do people do that? Hasty simplification. Benefit and utility are hastily and wrongly identified. A similar error is common with "happiness research" - happiness and preference are hastily, and wrongly, confused.

Anybody who wants to give

Anybody who wants to give their own money to the poor wants to redistribute money to the poor. They're for redistribution from rich to poor. There's nothing wrong with that. There's only something wrong with coercion. It's not the redistribution but the coercion that that is wrong with coerced redistribution.

Leftists dishonestly combine these two issues - the poor, and coercion - and try to impose the impression that their opponents, the advocates of freedom, are hostile to the poor. They are not. They are hostile to coercion.

Here's how stupid the argument is. Suppose someone says, "I want you to leave your child in the care of a Jewish child molester," and suppose you say, "no thanks, I'll find other arrangements." Then suppose that this person then goes out and says to all your acquaintances that you are anti-semitic because you declined to have your child cared for by someone that you were told was a Jew.

Would that be dishonest to the point of being depraved? Yes. And yet anti-capitalists do the equivalent on the issue of coerced redistribution. They use this to paint the advocates of freedom as hostile to or uncaring of the poor. It is dishonest to the point of being depraved.

Suppose I want to

Suppose I want to redistribute my money to the poor. Have I, in doing so, compared my utility function with the utility function of the poor? No, of course I have not.

How so? Very simple: when I give my dollar to the poor, I am revealing that my preference is for that dollar to be used for whatever the poor person prefers. In other words, in giving that dollar to the poor man I am revealing that my preference for the use of that money is identical to his preference.

Now let's talk marginal utility, let's see how this enters into it. The marginal value of the dollar to me is of course the value of my preferred use of the dollar, which is identical to the value, to me, of the recipient's preferred use, whatever that is - an identity revealed by my giving it to him to spend as he likes.

The marginal value of the dollar to him is the value of his preferred use of the dollar, which is the very same thing.

So it's the very same thing: say the man spends the dollar on P (say, P is "the man has lunch"). Then the marginal value of the dollar to me is the value to me of P, and the marginal value of the dollar to him is the value to him of P.

The supposed argument for redistribution is that the marginal value of the dollar is higher for the poor man than for me. Let's see how that works out. It works out thus: it comes to the statement that:

"I gave the man this dollar because the value to him of his having lunch is higher than the value to me of his having lunch."

Huh?????

That makes no sense.

That's right, it makes no sense. That's obviously not the reason. Even if we admit commensurability of utility, clearly the reason I gave him the dollar is not that he values his having lunch more than I value his having lunch. The reason I gave him the dollar is that *I* value his having lunch more than I value other uses of the dollar. The comparison is entirely within me, it is entirely about my set of preferences.

Constant: Why do you value

Constant:
Why do you value his having lunch more than whatever else you might have used that dollar for? Doesn't the fact that he's really hungry and you don't really need the dollar anyway enter into the calculus?

And a dollar for lunch? Are you posting from 1965?

Also, Kaldor-Hicks

Also, Kaldor-Hicks efficiency uses willingness to pay as a proxy for utility, so it is subject to some of the vagaries of interpersonal utility comparison, but it ignores the part that's most interesting to redistributionists: diminishing marginal utility of wealth. If transaction costs are zero, robbing the poor to give to the rich is Kaldor-Hicks efficient. Or vice-versa. Distribution isn't taken into account in any way.

Brandon, You can get a

Brandon,

You can get a delicious chicken sandwich from Wendy's for 99 cents. (Obligatory product placement - I have to make a living too.) :)

Okay, I'll try to eliminate lingering doubts about my point by explaining why I might value his having lunch more than my having a second lunch (say). Let's make it real simple - I am altering the example to simplify it. I am no longer giving him the dollar but buying him something quickly perishable. Suppose I have two dollars and I'm at Wendy's and someone else is there and there's no sales tax. I can eat two delicious chicken sandwiches or I can eat one delicious Wendy's chicken sandwich and buy a second one for him to eat. These are (let's say) my two choices and suppose I choose the latter. Why might I choose the latter? Is it because I have compared my utility function with his utility function?

I reply that this is not it, that this idea of comparing utility functions is a mistake, is a collapsing of two different things into one thing. We are collapsing economics talk about preference, with everyday talk about benefit. Recall my point about suicide: something that a person prefers may not, in our view, be a benefit to him, though we may readily concede that he prefers it.

Here it is: I care about the man. Maybe I care about his being well fed about as much as I care about my being well fed. How is this different from comparing our utility functions? Well, I don't want you to take this the wrong way, but imagine a gardener tending two plants. He cares from them equally. Does that mean he is comparing their utility functions? No - as far as he is likely to know they do not have utility functions. He cares for them because he wants them both to be well, he cares about them, and this is not on account of a comparison between their utility functions, which they do not have.

Of course we recognize that other people have utility functions but, seeing as we also take good care of plants, perhaps one may wonder whether their utility functions per se are really key to our caring for them. Other people have utility functions but they also have eyes, a nose, fingers, etc. Why pick out their utility function as the key reason we sometimes take care of other people? Again - plants do not have utility functions and we take care of them (some of them).

Consider parents and how they care for their kids. Do parents take care of their kids because they are trying to maximize their kids' utility functions? No, because the way they take care of their kids is far from maximizing the kids' utility functions. In fact quite often parents are at odds with their kids over things. Parents do not send their kids to school because their kids have voiced a preference for school. They send their kids to school because they consider an education to be beneficial.

Let's return again to the two people at Wendy's. Why do I feed this other person about as well as myself? (Or rather why, hypothetically, might I?) Is it because I have compared our utility functions? I say that is not the reason, and I add that utility functions are raised as a possible answer only in haste, possibly out of insufficient patience for the right answer.

Remember the movie The Titanic? I never saw it myself, but I know how it ends. Somehow I managed to either see or imagine the last scene. The boy sacrifices his life to save the girl. Has this boy compared his utility function with hers and found that she derived greater utility from survival than he derived from survival? I do not believe so. I believe that he loved her and that he preferred that she live than that he live. It was not that he considered her utility of survival to be greater than his, but that he himself valued her survival more than he valued his own. Surely that is the right explanation. Surely it is wrong and convoluted to argue that he compared their utility functions and decided that she preferred surviving to dying more than he preferred surviving to dying.

So I think it is simply wrong to try to reduce love or concern to a comparison of utility functions.

If transaction costs are

If transaction costs are zero, robbing the poor to give to the rich is Kaldor-Hicks efficient.

I call that a feature, not a bug. :)

If interpersonal utility

If interpersonal utility comparisons were truly impossible, I think we wouldn't engage in many of our everyday behaviors.

If anything, I'd say the suicide example is an attempt to make an IUC. The person trying to convince the other not to committ suicide is carrying out the following thought process:

1) That guy is suicidal.
2) I've been sad before; it passes. It would suck to be dead right now.
3) That guy is probably a lot like me. In the future, he'll be glad he didn't kill himself. Right now, he's temporarily in an irrational frame of mind.
4) I will convince him to not kill himself.

It's an exercise in empathy. Empathy is the product of an assumption that other people are probably a lot like us, and that there are some things constant about human nature, an assumption that most people intuitively possess.

If IUCs were truly impossible, we wouldn't give gifts. We wouldn't try to figure out what one gift would make a person happier than other gifts.

If IUCs were truly impossible, we would treat everyone in exactly the same way. But we don't. We treat people differently, because different things make different people happy.

Well, regarding the initial

Well, regarding the initial article... you're right, in politics or moral philosophy. You're not right in positive economics since you can't/didn't provide "scientific" criteria for such comparisons.

Again, what welfare economics means by utility is rather different than the value judgements involved in the case of any millionaire or pauper.

On the other hand if you don't mean your post to be scientific, and instead you meant it as a moral-political philosophy text, then you don't have to deal with the impossibility of inter-subjective comparisons anyway.

"In virtually all cases,

"In virtually all cases, taking $1,000 from a billionaire and giving it to a starving beggar will help the beggar more than it hurts the billionaire."

No, in "virtually all cases," people would vote for it, especially "virtually all the people" who are not the billionaire.

Which is NOT the same thing and matters not one damn bit.

"Constant" commented: If

"Constant" commented:

If someone is suicidal people don't try to help them kill themselves.

Sure they do, under the proper circumstances. See generally, "terminal illness."

In fact, Constant has actually provided the perfect counterexample to Brandon's very thesis. Who are "virutally all the people" to usurp the private decisions of others whose utility they cannot possibly perceive or compare to their own?

I love it... a friend of

I love it... a friend of mine refered me to this blog and the first post I see is about interpersonal utility comparisons. It sounds sarcastic, but I'm serious: awesome.

If IUCs were truly

If IUCs were truly impossible, we wouldn’t give gifts. We wouldn’t try to figure out what one gift would make a person happier than other gifts.

No, I disagree. When we try to figure out what gift would make a person happier than other gifts, we are thinking only about his utility function, not comparing his utility function to ours. We are trying to figure out what he prefers. We are not trying to figure out whether he prefers the gift more than we prefer something else.

Think about it. If I think, "I will buy her chocolate rather than candy because she likes chocolate better than candy," I am not at any point in this brief thought comparing any of her preferences to any of my preferences. I am not thinking, "I will buy her chocolate rather than buy myself chocolate because she likes chocolate better than I like chocolate."

Good God, if we did an IUC at every gift giving occasion to decide whether to give a gift at all, we would pretty soon discover our mistake.

Constant, I think you're

Constant,

I think you're maybe missing Jonathan's point. It's not that giving the gift itself is an act of IUC. Rather, it's that the act of giving gifts demonstrates that I can and do make determinations about other people's utilities. Once I can do that, though, I can make interpersonal comparisons all over the place. That is, if I can estimate your utility and my utility, then I can pretty much compare the two, no?

Constant, No, you're right

Constant,

No, you're right that I can't do what you're talking about in this particular case. You also make Brandon's point rather nicely about libertarians being utopians. It just doesn't follow from the fact that I can't make a reasonable determination in the case that you outline that I cannot make a reasonable determination in any case whatsoever. When the comparison is something like whether you prefer 6 cars to 5 cars more than I prefer eating today over starving to death, then yes, I think that for the most part we can determine which preference is greater. Put it in any language you like. To deny that we can do this, at least roughly, is just to fail to live in the same world as the rest of us.

That is, if I can estimate

That is, if I can estimate your utility and my utility, then I can pretty much compare the two, no?

No, because they're not comparable. I think part of the problem is the unfortunate words chosen. Let's use the word "preference" and see whether the illusion persists. Suppose I know that you prefer candy to chocolate, and suppose I know that I prefer chocolate to candy. We may be very simple creatures with very simple preferences so this may be the sum total of our preferences. In which case I have complete knowledge of both your preferences and mine: I know what you prefer to what, and I know what I prefer to what.

From this knowledge, can I now derive whether you prefer candy to chocolate more than I prefer chocolate to candy?

No. I think it should be obvious that I cannot derive that.

Brandon, But in some cases,

Brandon,

But in some cases, particularly those in which proponents of redistribution are most keenly interested, you can guess with near certainty which of two people will value a particular resource more. In virtually all cases, taking $1,000 from a billionaire and giving it to a starving beggar will help the beggar more than it hurts the billionaire. No, it’s not true 100% of the time, but you don’t need complete certainty; if you’re right 95% of the time, that’s good enough for government work.

The net effect of a broader transfer on an economy-wide basis is not likely to be distinguishable from monetary supply inflation (not credit expansion). Only people that are sensitive to money possessed and prices are capable of causing price inflation. How much of the benefit of the transfer will be eaten up in rising prices?

Regards, Don

Joe, Brandon is wrong about

Joe,

Brandon is wrong about utopianism. That has nothing to do with it. Utopianism is a belief that a perfectly libertarian society is possible. What I'm talking about is clarity about the concepts that we're using.

Please realize that as I have repeatedly pointed out, I am not saying we can't do what we obviously can. I am saying that you guys are thinking the wrong way about it. You're conflating different ideas which are really different from each other. In no way have I even started doubting anything about what we normally do when it comes to doing things for other people, I have only doubted the way you and Brandon have chosen to think about it. I think you are mis-applying concepts from economic theory to it.

You are taking people's obvious concern for other people and imputing a utilitarian theory to them. That is a common mistake, the mistake of taking empirical reality as proof that our concepts are valid. Similarly, the ancient norse might take a thunderclap as proof that Thor has done whatever the Norse believe a thunderclap means he did. On one side there's empirical reality, and on the other side there's the ideas we use to explain it, and I'm saying that your explanation of what people are doing when they do nice things for other people is wrong.

Constant, Since when does

Constant,

Since when does theoretical economics equal reality? Economics is a science, much like physics or biology is a science. As such, it is exactly the sort of thing that you criticize--an explanation for why it is that people do the sorts of things that they do. To take that science as somehow constitutive of reality is to make the very mistake you accuse me of making.

My point here is that your explanation (namely, the science of economics) fails to map on accurately to the sorts of things that people actually do. It doesn't fail in a radical way; by and large, economics does correctly explain what people really do. It fails, however, in that economics cannot actually quantify utility functions and so economics the science cannot therefore measure interpersonal utility. But the fact that our economic theory cannot do something that humans can do doesn't mean that humans are therefore not actually doing the thing in question.

I think of this as being somewhat analogous to humans' ability to recognize faces. We don't quite know how it is that humans are able to do so. There isn't (yet anyway) any codified set of rules that we use to recognize faces. As a result, AI programs that are designed to recognize faces are pretty unreliable. My four year old is better at it than are most AIs. But from the fact that computer science can't determine how it is that humans recognize faces (that is, from the fact that we can't quantify or measure or develop rules for the process), it in no way follows that humans can't really recognize faces. The limitation is in the theory, not in reality itself.

The same is true of economics. We don't have a way to measure IUCs, so we can't use economics to make them. But that still in no way implies that we are unable to do this very sort of thing.

In short, I'm not guilty of cherry-picking evidence to fit my theory. The better explanation is that you are denying reality where it doesn't fit your theoretical model. It's what makes ideologues so frustrating to deal with...and, as Brandon says, what gives libertarians a bad name outside of narrow libertarian-oriented blogs.

All you have to do is assume

All you have to do is assume any decreasing marginal utility of wealth to see that the lazy drug addict, at the margin, will get more out of $1000 than the billionaire.

You can't quantify and compare the utilities, which is where this whole IUC thing gets messy. I can't point at the billionaire and and say "Uncle Moneybags is at point (X,Y) on his utility of wealth function, while Begger McPoorerson is at point (Z,Q) on his," but I can say that given any old diminishing marginal utility of wealth "total utility" can be increased by such a transfer.

That doesn't really change the argument against coercion, and it doesn't make the transfer right, but it helps to at least admit this fact and then argue from a different position about the nature and duty of government, or explaining the tricky nature of thinking about utility in terms of an aggregate, or how the government has no reason to value the utility of the beggar over the billionaire.

Since when does theoretical

Since when does theoretical economics equal reality?

I didn't say it did. I said that the economic concept of utility is one thing and not another. Similarly, the mathematical concept of multiplication is one thing and not another. Your argument is like saying that the multiplication operation in arithmetic fails to map on accurately to the multiplication performed by rabbits when they have sex (as in be fruitful and multiply). Well, yeah, it sure does fail to do that, but that's not a problem with arithmetic, that's a problem with someone confusing two different ideas, taking them to be the same idea (arithmetical multiplication and rabbit multiplication).

You insist, wrongly, on using the word "utility" to describe what we compare between people, and on claiming it to be the same concept as the concept of utility used in economics.

We don’t quite know how it is that humans are able to do so.

And yet you want to jump the gun and use the word "utility" to describe something that you admit you don't understand. I say you're being hasty.

The better explanation is that you are denying reality where it doesn’t fit your theoretical model. It’s what makes ideologues so frustrating to deal with…and, as Brandon says, what gives libertarians a bad name outside of narrow libertarian-oriented blogs.

It has nothing to do with libertarianism. I don't even know why Brandon brought that in. That's yet another layer of error in his post, which is mountain of error piled deep one on top of the other. He's giving libertarians a bad name by supplying people like you with erroneous explanations which you readily lap up as to what lies behind arguments about terminology.

All you have to do is assume

All you have to do is assume any decreasing marginal utility of wealth to see that the lazy drug addict, at the margin, will get more out of $1000 than the billionaire.

Even assuming you want to construct a theory of utility that makes utility a scalar with a common unit across everyone (nothing is preventing you from coming up with such a theory, other than, possibly, internal inconsistency, though I'm not aware of any) - even in that case, you can have the rich man's utility function be a lot steeper at a billion dollars than the poor man's is at one dollar. And that's even if you have decreasing marginal utility.

The marginal utility of each additional dollar for the rich man might be 1/x + 1 (for x dollars). So the marginal utility keeps going down but never reaches zero. And the marginal utility of each additional dollar for the poor man might be 1/(x+1). That is also decreasing but however small x gets it never quite reaches 1. You can plot the two functions here if you don't believe me:

http://www.coolmath.com/graphit/

Now, why am I making this point? Is it because I am a libertarian ideologue? Or is it because I am careful and I try to encourage others to be careful? Let me tell you that if you made the assertion that you just made in a mathematics class that I was teaching, I would mark your assertion incorrect. Would that be because I was a utopian libertarian ideologue, or would it be because I was holding you to a certain standard of rigor? How many utopian libertarian ideologues do you think teach math classes?

If Brandon had made the criticism that libertarians can sometimes be pedantic because they are somewhat literate in economics and, enthusiasts that they are in economics they can sometimes get carried away on points of rigor, just as some people can get carried away on points of grammar, then he might have a point. But instead, what he said was that the problem was that libertarians are "utopian", which has nothing whatsoever to do with it.

Constant, First, I'm not

Constant,

First, I'm not quite sure what's up with your use of phrases like "people like you." I'm a rather sympathetic libertarianism, not some raving lunatic leftist. You might also want to cut out the assumptions about what I do and do not understand. I'll readily admit that I'm no expert on economics, but I have done a bit on questions related to utility. Perhaps you're not aware, but there are actually competing notions of what the term "utility" really means.

Anyway, this is all really silly. Even if I were to accept your point entirely, you are still completely missing the point of Brandon's post. Whatever you want to call it, Brandon is still right to point out that we can make comparisons between the amount of satisfaction/happiness/whatever that you get from some bit of wealth and the amount of satisfaction/happiness/whatever that I get from it. We do this roughly and can't always do so perfectly (especially where the values are close), but we can nonetheless do just that. Your claim that whatever it is we do when we do this is not really an IUC is just simply not relevant to Brandon's main point.

Look, whatever name you want to give it, the phenomenon that Brandon describes, that Jonathan points to, and that I defend is a real one. And it shows that redistribution is not in principle wrong. There may well be other sorts of reasons for thinking that redistribution is a bad thing, but the claim that we can't possibly ever determine whether a particular act of redistribution would create increase happiness is just plain false. To insist on such a point does a disservice to libertarians. Or perhaps you'd rather stick with a pure ideology and ignore the messy real world. False consciousness anyone?

Let's go back to Brandon's

Let's go back to Brandon's central point, which is found in this statement: "In virtually all cases, taking $1,000 from a billionaire and giving it to a starving beggar will help the beggar more than it hurts the billionaire." Not true. There simply is no way to measure how much it helps the beggar and how much it hurts the billionaire. Humans are capable of empathy, but they are incapable of actually thinking what other persons think -- and that's what it takes to make a valid interpersonal utility comparison (IUC).

Having empathy is not the same thing as making a valid IUC. Empathy is an attribute of imagination. Through empathy I imagine how *I* would feel if *I* were in another person's shoes. But *I* cannot actually be in another person's shoes and actually feel what that other person is feeling.

And so, I begin my opposition to redistribution by stating my belief in the impossibility of IUCs. To that I add this: Taking money from the "rich" to help the "poor" probably hurts the "poor" because the "rich" finance economic growth. (See this, for example.) In saying that I am not undercutting my case against IUCs. Rather, I am simply pointing out the wrongness of the belief that the "poor" (as a group defined by some arbitrary measure of income) are made better off (in the long run) through redistribution, where "better off" is measured in terms of income. So, those who believe (wrongly) in IUCs might nevertheless be convinced that redistribution fails to make the "poor" happier (as a class) than they would be in the absence of redistribution.

(If most redistribution really is from "middle" to "middle" -- as one commenter claims -- that merely indicates the futility of trying to redistribute through the political process. It does not touch on the argument about the validity of IUCs.)

First, I’m not quite sure

First, I’m not quite sure what’s up with your use of phrases like “people like you.”

People who do what you're doing. E.g. say things like The better explanation is that you are denying reality where it doesn’t fit your theoretical model. It’s what makes ideologues so frustrating to deal with…and, as Brandon says, what gives libertarians a bad name outside of narrow libertarian-oriented blogs. Statements like that. You haven't engaged my arguments, you've merely brushed them aside and spat on them as they went by, and you were aided and abetted in this by Brandon, who provided you with a ready excuse. He made you think it was okay to call people names (ideologues, utopians) in order to avoid dealing with their arguments.

You might also want to cut out the assumptions about what I do and do not understand.

I took you as implying by your own statements that you did not understand, because nobody understands. I was not implying that I understand more than you.

Whatever you want to call it, Brandon is still right to point out that we can make comparisons between the amount of satisfaction/happiness/whatever that you get from some bit of wealth and the amount of satisfaction/happiness/whatever that I get from it.

What we can do is we can sometimes do nice things for other people. But this stays within our set of preferences (which sometimes encompasses the good of other people) and our set of preferences is subjective. Different people will have a different set of preferences, and so some people will go out of their way to help some people and other people will go out of their way to help other people. Who gets helped and in what way will vary depending on who is doing the helping.

Now, when I go out of my way to help some people and you go out of your way to help other people, I don't say to you, "hey you, you're helping the wrong people." I accept that you may care about or be interested in some people more than I am. I respect that your preferences are different from mine. And this contradicts the notion that (a) I am trying to maximize social happiness/whatever and (b) I believe you are trying to maximize social happiness/whatever. Because if (a) and (b) were true then I would get together with and argue with you about which of us was doing more to maximize social happiness. I don't do that because I don't see you as trying to do the same thing I am trying to do. I am trying to help certain people in certain ways, and you are trying to help certain other people in certain other ways, and I recognize this and so I recognize that we are not trying to do the same thing. And that's why I don't try to convince you that you're failing to do the things that would maximize the social utility function. Because that's not what either of us is trying to do and I recognize that intuitively. You care about Bob more than Bill and I care about Bill more than Bob, and that's fine.

So, no, I don't agree that what people do is rightly described as trying to maximize some objective universal goodness function over all society. That description does not match what they are trying to do. I am increasing things that I care about increasing, and you are increasing things you care about increasing.

There may well be other sorts of reasons for thinking that redistribution is a bad thing, but the claim that we can’t possibly ever determine whether a particular act of redistribution would create increase happiness is just plain false.

I think that that's not a good description of why we perform acts of redistribution (acts of charity).

Ok, I'll grant you that you

Ok, I'll grant you that you can make interpersonal utility comparisons probabilistically by looking at your experience.

"Looking at experience" is

"Looking at experience" is not scientific.
The appeal to fuzzy/probabilistic "stuff" isn't helping because what we have here is a categorical issue, not a quantitative one. It's not like: 1 inch is sort of like 2 centimeters. It's more: 1 inch is sort of like half a litter.

Scientifically, I don't see an acceptable criterion for interpersonal comparison of preferences. It's meaningless, in scientific terms to say that X prefers something more than Y does. Until von Neumann-Morgenstern in was impossible to do quantitative claims about preferences even for the same person.

Outside of science, you can easily apply something like Wittgenstein's argument regarding private language and go from there to the (correct) claim that everything in language in interpersonal and use a Wittgensteinian analysis of preferences language games to explore the issue. But that won't provide you with clear quantitative criteria nor will these criteria be scientific in the regular sense.

Also, and this can't be stressed enough, the economic treatment of revealed preferences (utility functions) has little to do with "the amount of satisfaction/happiness/whatever". Keep that utilitarian mambo-jumbo away from our pretty little axiomatized theories of consumer behaviour, please! ;-)

Seriously now, if this discussion is placed in the context of the positive science of economics, then I haven't seen "the goods", the supporter of inter-personal comparisons simply haven't delivered. If the discussion is placed in the context of everyday life, then we can make comparisons, obviously, but we can't quantify and theorize them.

We treat people equally on

We treat people equally on various occasions for a variety of reasons and according to a variety of measures. For example we might give our two nephews same-sized slices of cake so that neither one feels as if the other got preferential treatment. That has nothing to do with maximizing collective happiness and everything to do with an ad hoc criterion of equality employed and hopefully shared by all parties so as to prevent strife. People seek out balance points in their dealings with each other so as to minimize friction and maximize the likelihood of agreement and harmony. The complex of all these practices in turn underly and along with other things inform our moral behavior which in turn inform our moral ideas.

If one of our nephews is hurt and the other one is not, we'll rush to the aid of the hurt nephew and, necessarily, reduce our attention to the nephew who is unhurt. But we can only do this so far without creating jealousy. So we need to find a balance point here too, a balance point between helping the hurt nephew and not entirely abandoning the well nephew. The greater the hurt, the easier it is to convince the unhurt nephew that, on balance, the hurt and the attention cancel out so he has no cause for complaint about unfairness and favoritism.

These sorts of considerations extend to the general society. In our charitable actions we find a balance between assisting those who are especially deprived - of clothing, of food, of housing, of education, etc. - and not neglecting those who are not as deprived. The greater the observable deprivation, the more likely we are to devote extra care to them, the easier it is for us to explain such attentiveness to ourselves and to others.

So the question is, does this mean we are comparing utility or happiness or whatever across people and trying to maximize the quantity? The visible behavior superficially resembles an attempt to maximize overall utility in the context of declining marginal utility.

But I argue that the comparisons we make are ad hoc, they are on the surface, and they should not be taken too seriously. If we give each of two nephews the same size slice of cake, we do so in order to avoid strife, not because we are trying to maximize collective utility between the two boys in a context of declining marginal utility. The latter explanation also explains the equal-sized slices, but surely that is not the real reason for it.

From the top: Constant: If

From the top:

Constant:
If the point you're trying to make is that the idea of interpersonal utility comparison is incoherent because of the way utility is defined, that may be true (I don't know), but it doesn't strike me as particularly relevant. Redistributionists argue that the marginal benefit of a given quantity of money to a pauper is much greater than the marginal benefit of that same quantity of money to a billionaire, and this is almost certainly true in almost all cases. This is a real phenomenon; whether or not it can properly be called an interpersonal utility comparison is beside the point.

Regarding charity, you are not in the habit of giving alms to the wealthy, are you? Why not? My guess is that it's because you believe that your money will benefit a poor person much more than it would benefit a wealthy person.

I do agree that there's a serious flaw in this sentence:

I gave the man this dollar because the value to him of his having lunch is higher than the value to me of his having lunch.

It's more an error of omission, though, in that there's an additional condition which must be met. More in a follow-up post.

Regarding the Titanic example, personal affection is different from charity, although the difference may be one of degree. More in the aforementioned follow-up post.

Gabriel:
That may be true. I intended this as a response to libertarians who answer millionaire/pauper comparisons with, "You can't make interpersonal utility comparisons." If I'm misusing the term, then so are they. But we're talking about the same thing.

Kip:
No one's proposed usurping anything. The whole point of my post is that the case against such usurpations isn't contingent on denying the obvious fact that the marginal benefit of a dollar to a pauper is usually greater than the marginal benefit of a dollar to a billionaire.

Don:
Good point, and that's another argument against redistribution. I suppose it would depend on what the wealthy are doing with their money (hoarding cash vs. investing in capital goods) and how efficiently the economy can shift production towards the sort of consumer goods preferred by the poor.

Constant again:
Now that you mention it, I guess "utopian" was the wrong word. Perhaps "dogmatic?"

Yes, you could hypothetically create a utility function for a rich man which is steeper at around a billion dollars than it is for some poor man at zero. Maybe the poor man is an ascetic and the rich guy really likes watching his bank balance go up. But such people are exceptional---most of us don't have utility functions that look like theirs. As I said, such comparisons are at best probabilistic, and you can never be 100% certain.

Joe:
Look, whatever name you want to give it, the phenomenon that Brandon describes, that Jonathan points to, and that I defend is a real one. And it shows that redistribution is not in principle wrong.

I wouldn't go that far. All I said was that one particular argument against redistribution is flawed. Redistribution may be wrong in principle for other reasons.

Constant: After the 2004

Constant:
After the 2004 presidential campaign, I could quite happily have gone the rest of my life without ever again hearing the phrase "band of brothers." Thanks a lot.

Anyway, I think you're begging the question. Why is it better to help the injured than the healthy?

Also, you're the one who brought up your motives for almsgiving. It could be that they really are as you describe (though you haven't explained them in a way that makes sense to me), but my point is not contingent on your motives.

Sometimes I can tell why a comment gets stuck in the moderation queue---there's a misspelled word or a proper noun that the spam filter hasn't seen before---but I'm not sure what the problem with yours is. You misspelled "judgment," but the filter should be used to that by now. "Sublimated," maybe?

Redistributionists argue

Redistributionists argue that the marginal benefit of a given quantity of money to a pauper is much greater than the marginal benefit of that same quantity of money to a billionaire, and this is almost certainly true in almost all cases. This is a real phenomenon; whether or not it can properly be called an interpersonal utility comparison is beside the point.

I don't agree at all that this is obviously real. In my opinion, I merely do, or rather may (sometimes) place a higher value (in my own scheme of preferences) on spending that dollar on a poor man than on spending that dollar on a rich man. Why must you go further than this fact about my own preferences? Why do you feel the need to justify my decision about how I will spend my charity dollars by arguing that in some sense other than merely my own preference, the poor man benefits more than the rich man?

Regarding charity, you are not in the habit of giving alms to the wealthy, are you? Why not? My guess is that it’s because you believe that your money will benefit a poor person much more than it would benefit a wealthy person.

There are many reasons. Analogy is a powerful psychological mechanism and we take our behaviors in one context and often extend them by analogy to other contexts. I will now describe one context from which the behavior you describe can emerge. Here is a situation which in one or another variant we sometimes encounter, not necessarily in the following precise form: Suppose I am part of a band of brothers in a war zone. Suppose one of us is hurt. Will I then devote more of my energy to giving a back rub to a healthy member of my band, or will I devote more of my energy to tending to the hurt member of my band? There are powerful reasons of self-interest for everyone in the band to adopt and deeply internalize the policy of going out of our way to take care of those members who are hurt. If we try to trace back to root causes why groups of warriors behave that way, we will probably find naked self-interest. That is to say, my own selfish preferences about my own future are transformed, sublimated if you like, by my participation in the group into a deep attachment to the other members of my unit. There's massive reciprocity going on in the group, so pure narrow self-interest gives way to, in fact actually serves as motivation for, ultimately, selfless service to our fellows. And of course people don't just want us to give them back rubs. They want us to have their backs: they want us to be there for them if something goes horribly wrong for them, or something might go horribly wrong for them. It's a mutual insurance scheme if you want. (And I don't mean that they are cynically manipulating each other by the way, I mean that they have literally traded away their souls to each other, in a sense, they've fully committed themselves, and I doubt there's any conscious calculation. So I'm not trying to make some cynical comment about it.)

My point is that, to explain the behavior that you see displayed in a band of brothers, you don't need to imagine that a member of the band is some sort of utilitarian impartially weighing the good of different members of the troop. Each of them, by doing what he does (e.g., protecting the fallen) is doing what is best for himself, what is best in an entirely practical and self-interested way. Each one is serving his own interests. Each one is responding to his own preferences. They are not doing this because they are trying to maximize a collective happiness function, they each are doing this because it is best for himself.

The collective happiness function that they are supposedly responding to is a mirage created in the minds of utilitarians who have hastily concocted a theory to explain the behavior of people who behave like bands of brothers, caring disproportionately for the fallen. The behavior that people display is part of a bargain that they have unconsciously struck. It may be part of a bargain that, in a sense, is ingrained into all of us through the millions of years of evolution that produced what we are.

If I take care of you because of a bargain we have made, because of a trade, then I am not taking care of you because I have made a judgement about which of us gets greater utility from the last dollar spent. If I make a trade with you then I am not comparing your utility to mine, or to anyone else's.

Some of my posts (like the

Some of my posts (like the one I just now sent) are caught in the maw of moderation and other ones pass right through appearing instantly on the site, and I have no clue what's the difference. Are there key words that I should be avoiding?

This reminds me of an

This reminds me of an episode of Star Trek (really!). The crew landed on a planet where environmental radiation had made all the inhabitants sterile- they had no children and could have none. They then proceeded to kidnap most of the kids off the Enterprise. They justified it by saying "You have plenty more kids, your race has plenty more kids, and you have the capability to have plenty more kids. We need these kids more than you do. Goodbye." If you can make interpersonal comparisons of utility, then weren't their actions correct? Does that mean we can't rule out redistribution of anything, including kids, if I can claim my utility (say, if I were infertile) is greater than yours (say, if you're expecting your sixth kid), and you can't prove otherwise?

COnstant- The ways of our

COnstant-

The ways of our moderating system are strange and inscrutable. We do what we can to pop them out when we find them, apologies for the meantime. :(

Egads Brandon, don't tell me

Egads Brandon, don't tell me you've bought into the soul-destroying heresy that suggests you can make a soft "g" sound without the aid of an e?

JUDGEMENT - Aye!
JUDGMENT - Nay!!

Unless you're also prepared to ask for Fudg Brownies and hail the Edg as a brilliant guitarist...

Ignoring the spelling issue

Ignoring the spelling issue (though you can guess what my position is, since I spelled it that "wrong" way and I do spell it that way over and over, despite the many efforts over the years of Microsoft Word Wavy Red Underline to teach me otherwise).

Anyway, I think you’re begging the question. Why is it better to help the injured than the healthy?

I have already given you my answer. When I buy insurance, I pay the insurance company to - what? To help me when I am "injured" (literally when the insurance is medical). The insurance company's behavior towards me is that it will help me when I am injured and not otherwise. Its behavior towards all of its customers is that it will help the ones who are injured but not the ones who are uninjured.

Does that involve a cross-personal comparison of utility? No, it does not. I am making a trade with the insurance company. I am participating in this trade on the basis of my own preferences, and the insurance company is participating in this trade on the basis of the preferences of its owners and employees.

The band of brothers is no different.

Also, you’re the one who brought up your motives for almsgiving.

Maybe I did do that at some point a few iterations ago, but in your last post you did ask me a direct question about my motives for almsgiving here:

Regarding charity, you are not in the habit of giving alms to the wealthy, are you? Why not? My guess is that it’s because you believe that your money will benefit a poor person much more than it would benefit a wealthy person.

I answered it, and now in response to my answer you claim that I'm the one who brought the subject up. Does that mean are you retracting the question? Or what?

my point is not contingent on your motives.

As far as I can see your point is mere assertion, and re-assertion, and re-assertion, that obviously we are performing an interpersonal utility comparison, and that obviously utility really can be compared between people. Because it's obvious. End of explanation. For my part, I've tried to give you an explanation for the behavior which does not involve any mention of interpersonal utility comparisons and does not rely on the possibility of comparing utility across people.

Let me add one final point to this submission. Suppose that there is a utility monster who is otherwise normal. You know what those are, I won't elaborate. Now put him in our society. What will happen? Well, the utility of everything for him will be vastly magnified relative to the rest of us, but he will still have the same set of preferences as a normal person (part of the assumption). So typically, such a utility monster will do what all the rest of us do - e.g., buy insurance. And then he will get help from the insurance company when he is injured and not otherwise. So the fact that he is a utility monster is transparent. No utility comparisons are - apparently - being made across people.

Now take this same utility monster and put him down in the midst of a band of ... if you don't like the term would you prefer BOB? So we put this utility monster into a BOB, and again, the fact that he is a utility monster has no effect on how he is treated and how others treat him. He is helped when he is injured, and even though a tiny cut to him is supposedly as much a private tragedy as losing a leg is to someone else, he does not get everyone's attention when he gets a tiny cut. (Of course someone might argue that if a tiny cut were truly so tragic to a person then he would be writhing on the floor in pain, but I think that violates the point of the idea of the utility monster, because once you reduce utility to outward behavior then we can explain things in terms of that outward behavior, *again* ignoring utility.)

Oh no, not again.

Oh no, not again.

Isnt that what the pot of

Isnt that what the pot of petunias said before hitting the ground at Magrathea?

Brandon, "Redistribution may

Brandon,

"Redistribution may seem just when we pretend the world was created yesterday but manifestly unjust when we acknowledge that differences in wealth are generally earned."

This is an odd thing to say when you've offered nothing that suggests redistribution is just in the first place. Why may it seem just?

By invoking desert you're invoking rights and morality. Do you reaaly mean to say that rights trump the net utility?

Brandon, What rate of

Brandon,

What rate of taxation is not confiscatory?