Scott Sumner  

Interpersonal utility comparisons are unavoidable

PRINT
Government Works Badly... Hooves and Priors...

When I advocate policies aimed at redistribution on utilitarian grounds, some commenters object that interpersonal utility comparisons are impossible. Actually they are unavoidable.

Consider a policy that almost all well-informed people would support; ending sugar subsidies. Ending sugar subsidies would increase the efficiency of the American (and global) economy, boosting real GDP. It would redistribute money from relatively wealthy sugar producers to average Americans. But real GDP is just a number, and redistribution is not necessarily good, as we (supposedly) can't make interpersonal utility comparisons. Maybe the wealthy sugar producers get amazing happiness out of the extra sugar subsidies, far more than the loss in happiness to average taxpayers.

If one is going to argue to for or against any public policy on pragmatic utilitarian grounds, then one has to make interpersonal utility comparisons. Of course you can also make non-utilitarian arguments, as when people oppose government intervention on "natural rights" grounds. But it is not easy to convince the broader public without some sort of utilitarian argument.

I tend to rely on common sense when making interpersonal comparisons. People seem to be strongly influenced by relative conditions, and thus I believe that absolute changes in utility are roughly proportional to percentage changes in wealth. Thus the disappointment in falling from $10 billion to $5 billion in wealth, is about the same as the disappointment in falling from $10 million to $5 million, or from $10 thousand to $5 thousand. Furthermore, I'd argue that if this rule of thumb is incorrect, it's probably because it underestimates the extent to which marginal utility diminishes at high wealth levels. That is, it's quite possible that the decline from $10 million to $5 million is actually a bigger psychological blow than falling from $10 billion to $5 billion. I can't prove that, it's just an educated guess based on 59 years of observing people. But it underpins my advocacy of high marginal tax rates on very high levels of consumption. It also explains my advocacy of wage subsidies at the low end of the spectrum.

Here's some information on Larry Ellison:

Indian Wells is a vacation paradise full of resorts with luscious green golf courses, vibrant flower gardens, and abundant pools. Some resorts even have sandy beach wading areas for the grandchildren. The weather is perfect eight months of the year.

In 2011, Larry Ellison, co-founder of Oracle, paid $43 million for Porcupine Creek, a 249-acre estate here. Visitors sign a nondisclosure agreement to get past the armed guard and front gate. Once inside, they find an exquisitely maintained 18-hole golf course decorated with sculptures of naked women lounging among pink and purple tropical flowers, a "slidetacular" pool, and a Bellagio-worthy fountain that shoots water into the air when cars approach. These diversions surround a 27-room residence reminiscent of an Italian villa. Each year, Ellison's architect adds another signature feature to the estate to keep him from getting bored.

Ellison comes to Porcupine Creek for at least two weeks in mid-March to take in the BNP Paribas Open tennis tournament at the Indian Wells Tennis Garden, 20 minutes away. Ellison is worth an estimated $47 billion, making him the world's seventh-wealthiest person.

. . .

Ellison, 70, has rebuilt a sport before. His love of sailing led him to his highly controversial domination of the America's Cup, the capstone to decades of exploits that have included brushes with death while sailing in the Pacific, the construction of a real estate empire in Malibu and Hawaii (where he owns 98 percent of the island of Lanai), and liaisons with far younger women, including current girlfriend Nikita Kahn, a young actress and model from Ukraine.


Unlike many on the left, I don't envy the rich. I'm really happy that Larry is really happy. If Larry was even a bit happier, and if that boosted total global happiness, that would be fine with me. But I can't get away from the implication of (what I perceive as) diminishing marginal utility. Some redistribution is justified.

Larry Ellison is extremely lucky to live in a universe where:

1. Government is very inefficient.
2. Incentives have a surprisingly strong effect on behavior.

If not for those practical limitations to redistribution, it would be hard to argue against people who favor taking away almost all of Ellison's wealth, and reducing his consumption levels back to roughly average. But that's not the world we live in; Larry's very lucky that incentives do matter a lot and that governments are very inefficient. Even Sweden tolerates multi-billionaires for pragmatic reasons. Thus even a pure utilitarian like me favors MTRs on top consumption peaking at perhaps 80%, not 99.999%. As a result, I get hammered by people on the left for opposing hair-brained schemes like minimum wages, maximum wages, and taxes on capital income. Gene Marsh left the following remarks after my previous post:

Please correct me if I'm wrong but doesn't each and every post here align perfectly with the desired messaging of the wealthy and the business elite (not counting appeals for the deregulation of hair braiding and legalization of soft drugs which (for the time being) run orthogonal to those interests.)

What bothers me is that this fealty to a single interest group is elided, effaced, concealed, whatever.


At the opposite extreme, more deontologically-inclined libertarians accuse me of being a Maoist.

I plead innocent to both charges. I'm a pragmatic libertarian who just happens to think that global happiness would be maximized by small government, free market economic systems, with a few interventions to deal with externalities like pollution, and also to redistribute consumption.

It just so happens that most of my policy recommendations help the super rich (but not redistribution). But it also just so happens that all of my policy recommendations help the poor. There may be some potential policies out there that boost Larry Ellison's happiness by more than it hurts the happiness of the poor, and where a true libertarian like me would have to favor the rich over the poor. But I haven't yet seen such a policy, and I very much doubt that I ever will. Larry's already a pretty happy guy, and there is only so much happiness that can be squeezed into one human brain, especially at age 70.

Gene also expressed disbelief that policies intended to help the poor might not actually do so, for "counterintuitive" reasons:

You only object on the absurdly counterintuitive grounds that, in all cases all policies designed to help labor are paradoxically doomed to harm labor.
I'm tempted to give a two-word response: Chavez, Venezuela.

More seriously, I do favor those policies that I believe actually favor the working poor, such as tax cuts and wage subsidies for the less well off.

PS. I hope it's obvious that some of this is written in jest, I don't really know whether Larry Ellison is happy or not. I think that most people on both the left and the right overestimate the role of money in happiness. My hunch is that the observed correlation in the survey data reflects (mostly) other factors---optimism, wisdom and energy leads to both wealth and happiness. (At least that's the view of a guy that is somewhat lacking in those areas.)


Comments and Sharing






COMMENTS (34 to date)
D. F. Linton writes:

You say that without interpersonal utility comparisons one can't make greatest-good-of-the-greatest-number utilitarian arguments. This is true. Either you can validly make such arguments because such comparisons (and sums) are possible or such comparisons are not possible and the comparative utilitarian makes a faulty argument.

You seem to argue that because you want to offer utilitarian grounds for your favored policies interpersonal utility comparisons must be possible.

Can you really confidently judge the truth of the statement: "Better 50 men have an extra beer a week for life than a young girl be trained as an Olympic gymnast"?

andy writes:

@D.F.Linton - what if someone claimed that 'some interpersonal utility comparisons are possible'? Your question may not have answer, but different question may have one.

That said, the case against non-existence of interpersonal utility comparisons seems to me quite solid. Wouldn't it be better so simply accept it and argue for redistribution on different grounds? You can appeal to sympathy, being a human and helping fellow men without resorting to pseudo-argument of interpersonal utility comparison, can't you?

Brent writes:

Because convincing "the broader public" about what is (your?) 'Common Sense' is likely to be effective...

And why do you think an 80% tax is even necessary to pay for such a 'small government'? It seems designed to be spiteful - i.e., taking pleasure from the act of (advocating) taking income (err, "consumption power") away from more monetarily successful individuals.

D. F. Linton writes:

My question (or any such question) can be answered by saying: I would prefer to see the girl trained than to see the men enjoy their beers or vice versa. But those are the speaker's preferences about the condition of his fellow men not preferences of those affected. Utility is an internal, subjective, non-stationary thing.

Monk always said "You'll thank me later." Maybe yes, maybe no, but that was just his opinion of what the other person should want.

This way of thinking is very popular with those who enjoy treating people as chessmen in some game rather than as equals.

MG writes:

I would agree that Interpersonal Utility Comps should not be seen as a "Red Light", but neither should they be seen a "Green Light". When adducing them, I see: Proceed with caution -- consider context and application. Re. Application: e.g., as an intellectual exercise, Green Light; as a justification for redistribution, Yellow. Re Context or starting position: e.g., denying a subsidy is different from taking what was earned. Is there a way to reconcile this with your thinking? (More truculently: Is there a limit to the ability of a third party to psychoanalyze the relation between another, specific person and the object being redistributed?)

ThomasH writes:

I'd call someone who wants government to stick to dealing with externalities like ACC, redistributing consumption and purchasing some kinds of public goods (NASA is my favorite) a "Liberal."

My guess is that such a government would not be small, though it might be a bit smaller in the US than the one we have. It would not include the TSA, for example. In it one could have a wage subsidy instead of a "hare-brained" minimum wage and exempt everyone's saving, not just those who have some of their income from "capital."

Scott Sumner writes:

DF, You asked:

"Can you really confidently judge the truth of the statement: "Better 50 men have an extra beer a week for life than a young girl be trained as an Olympic gymnast"?"

No.

Andy, There are lots of potential arguments, but I find utility to be the most persuasive. Others might prefer other arguments. "Sympathy" seems even vaguer, harder to link up to specific public policies (and even utility is pretty hard to pin down.)

Brent, You said:

"It seems designed to be spiteful"

Not at all. I believe the effect on Larry Ellison would be trivial. If I wanted to be spiteful I wouldn't leave him with $10 billion in consumption, after taxes. I want him to be very happy.

DF, You said:

"This way of thinking is very popular with those who enjoy treating people as chessmen in some game rather than as equals."

We all treat people as chessman. Great wealth comes from things such as specific public policies like copyright laws, which makes it illegal to counterfeit software. There is no "neutral" public policy that does not treat other humans as chessman.

MG, I don't put much weight on the starting position. All of society is artificial. Starting positions often reflect luck, or past injustices.

Thomas, I'm considered a libertarian by others, although words don't really matter much to me, as long as people understand my views. I sometimes call myself a right-wing liberal, or a classical liberal, or a pragmatic libertarian.

Andrew_FL writes:
But it is not easy to convince the broader public without some sort of utilitarian argument.

This is not the same thing as such comparisons being "unavoidable." It merely means, if true, that the broader public are pseudo-utilitarian in their outlook.

Which, frankly, is unfortunate.

But whether one is able to make convincing arguments is something very different from whether one is making correct or incorrect arguments. Anyone constructing an argument using interpersonal utility comparisons is making an incorrect argument. Period.

Don Geddis writes:

Gene Marsh appears to have very little economics education. Economics has many, many results that are counterintuitive to common sense. It should not be "absurd" to consider yet more cases. Also, Gene seems to think that if you are "for the poor", you must necessarily be "against the rich". He betrays a non-economist's model that there must be some kind of fixed pie, and "fairness" is all about how we choose to divide it up. The idea that there are many "win-win" scenarios in economics, where everybody winds up better off, doesn't seem to have occurred to him.

Tiago writes:

Great post. The only thing that makes me uneasy about the debates I've been seeing between utilitarians and libertarians is that while there is a real philosophical difference between them, at the current state of affairs their policy menus would be virtually indistinguishable. I sometimes wish we could focus on what everyone agrees on and try to implement it.

D. F. Linton writes:

Perhaps there are no "public policies" that do not treat our fellow men as chessmen. Perhaps some such policies do bring great wealth to their proponents or their supporters. Nonetheless not everyone treats others as mere pawns. Great wealth has been gained by simply exchanging value for value as well as by plunder.

Malcolm Kirkpatrick writes:

You make a case --against-- progressive taxation. If the marginal utility of wealth (or income) decreases as wealth (or income) increases, then less-wealthy recipients of the services of the highly-compensated would benefit from a tax regime that presented to high-income professionals a tax rate that declines as income rises.
Consider an orthopedic surgeon who specializes in knee and elbow replacement. Suppose s/he performs four operations a week and takes a four-week break every year. That's 192 operations per year. Suppose this person earns $10,000 per operation. The marginal utility of an additional operation to the patient declines far more slowly than does the marginal utility of the surgeon's income from one additional operation, since most patients will get one operation, or none, or, in extreme cases, perhaps, two. Every operation is life-changing to the patient, while it's just another day at the office to the surgeon. If the marginal utility of income declines as income rises and the surgeon's current allocation of leisure and work reflects the utility of income, to call this surgeon out for one additional operation will require an increase in the income per operation to the surgeon. On the assumption that the price remains the same, the tax rate must fall.
Perhaps rock stars and concert musicians are a better illustration. How to persuade them to add one more city to the tour?

Nick writes:

Great Post!
I think prof summer actually touched on the real issue that bothers a lot of folks like Gene by accident. Sorry if I'm stuffing words into your mouth here, Gene, but here I go. First, a quote from prof sumner:

'Larry Ellison is very lucky to live in a world where:
1. Government is very inefficient
2. Incentives have a surprisingly strong effect on behavior'

But here's the thing. Some of us see huge variety in the inefficiency of governments across the wrold. And since the best reason these really rich guys have for keeping their money is government inefficiency, some of us fear that deep down all these guys really want to do when they engage with politics is make our government even less efficient.
After all, incentives have a surprisingly strong effect on behavior.
To me, our comparative advantage in the world has come as much from our relatively efficient government (compared to 90% of the world) as from being more 'free market'. Our market policies may explain the gap between us and France, but not why we are so much richer than Brazil. Our history of stronger poltiical institutions explain far more of that gap.
It would be nice to hear prof summer acknowledge the possibility that even some of the fear of extreme concentration of wealth on the left comes from anything other than jealousy.

Gene Marsh writes:

Don,
Since the close of the First World War and the demise of the pince-nez, haughty and supercilious speech has ceased to be an effective means of shutting down inquiry. The tone of contempt for those who cannot or will not recite the gospel of your religion makes teaching, learning, and original thinking triple impossibilities. Having risen to such a high position in the field of economics, you'd do well to relax the tenor of your own pedantry so as not to drive away every other doubting mind that crosses your path.

Scott Sumner writes:

Andrew. I don't agree. It's not a matter of correct or incorrect, it's plausible or implausible.

Malcolm, That policy implication is possible, but seems very unlikely to me.

DF, True, but my point still stands. Great wealth is difficult to achieve without support from governmental institutions like police and copyright laws.

Nick, I doubt that any plausible tax scheme would have much impact on things like lobbying. But for what it's worth, a progressive consumption tax reduces their incentive to lobby, by reducing after-tax wealth.

You said:

"And since the best reason these really rich guys have for keeping their money is government inefficiency, some of us fear that deep down all these guys really want to do when they engage with politics is make our government even less efficient."

I'd add that people like Gates and Buffett are lobbying for higher taxes on the rich, and are giving away most of their wealth. So I'm not convinced the rich are as corrupt as you imply, although obviously some individuals are. I see groups like the teachers unions having as much political clout as bankers, and using that clout for evil purposes.

Roger writes:

"Ending sugar subsidies would increase the efficiency of the American (and global) economy, boosting real GDP. It would redistribute money from relatively wealthy sugar producers to average Americans."

Honest question here... Haven't you just used a second version of the term "redistribution"?

In the normal version we are referring to an action which requires something to be transferred from A to go to B. Taxing A to give to B, or making a regulation which forces A to pay more than otherwise to B are both valid versions of this definition.

But you just extended this definition above to something quite different. In your above case, the supposed redistribution is the elimination of forced redistribution (dropping the tax transfer or the regulation).

Now, you are free to use whatever definition you chose, but using the same term for forced redistribution and for no longer forcing redistribution (since it leads to a different ultimate distribution than would occur with force) seems guaranteed to confuse the issue.

Seems to me it is important to call the first redistribution and the second is "no longer redistributing" and thus leading to a different than currently expected future distribution.

Roger writes:

Here is my second friendly pushback...

"If one is going to argue to for or against any public policy on pragmatic utilitarian grounds, then one has to make interpersonal utility comparisons."

Disagree. The other way to argue on pragmatic utilitarian grounds is from behind a veil. Assume you are entering this society and unaware of your position details. What types of rules, institutions and norms would you establish to optimize the desirability of outcomes for everyone whom you could potentially become?

In other words, if you were designing the game, what rules would you suggest knowing your utility preferences and assuming that you could play any position?

In the case of the higher marginal tax rates, this is something many, probably most, people would agree to behind a veil assuming the increase wasn't so large as to harm growth rates in the size of the future pie.

I would suggest conflicting utility preferences and conflicting values is extremely likely to deteriorate into a zero sum situation which actively destroys total utility. I would approach it instead from a constitutional level with choices and exit options.

By the way, just because I mention a veil, don't assume I am accepting Rawl's arguments -- I don't. I would never choose maximin nor would I sacrifice future growth for current welfare.

Cornflour writes:

"I get hammered by people on the left for opposing hair-brained schemes"

Instead of slyly mocking aging hippies, you need to attack the rabbits. Harebrained are their schemes.

More at The Grammarphobia Blog.

Of little importance, but I couldn't get that picture out of my mind.

Levi Russell writes:

Robbins said it well:

"[I]t is one thing to assume that scales can be drawn up showing the order in which an individual will prefer a series of alternatives, and to compare the arrangement of one such individual scale with another. It is quite a different thing to assume that behind such arrangements lie magnitudes which themselves can be compared as between individual scales."


khodge writes:

I don't think Brent put it nearly strong enough: small government and a redistributive (= punitive) tax are mutually exclusive.

(Malcolm): "You make a case --against-- progressive taxation. ... to call this surgeon out for one additional operation will require an increase in the income per operation to the surgeon. On the assumption that the price remains the same, the tax rate must fall."
(Scott): "Malcolm, That policy implication is possible, but seems very unlikely to me."
...
One of the contributors to Steuerle, et. al., __Vouchers and the Provision of Public Services__ asserted this conclusion (the welfare-economic benefits of regressive income tax rates) as an aside in a general discussion of income taxes.

Examples are easy to construct: the surgeon, the rock star, the housing developer, the home-improvements contractor. Anyone who sells big-ticket items to ordinary people fits in this category.

Ali Bertarian writes:

Mr. Sumner is an oxymoronic "pragmatic libertarian."

The reason that there even exists a "science" of economics is because we have Pikettys on one side, Sumners on the other side, and endless, unresolved debates on the maximization of the production of goodies, and on the optimum distribution ratios of those goodies among all the people. This endless debate on how to produce and distribute the stuff that we make allows them to get paid by their universities, and allows them access to the formation of knowledge bases upon which politicians draw. We get to wait for their decisions on what to do with our stuff.

There is no algorithm for measuring utility. There is no utilo-meter. No amount of paper publishing will substitute for that unicorn that is a utilo-meter.

"But it is not easy to convince the broader public without some sort of utilitarian argument."

This busy-ness in debate immersion allows Sumner to ignore what a libertarianism is, which is not a materialistic philosophy, but a moral one. Leftists complain that some people make too much money. I find it quite easy to tell them that they are envious and covetous, and that such emotions are not good foundations on which to form government policy. I have yet to get a response from them to this statement of fact.

If the left has been winning the last hundred years, then it is because insufficient numbers of people have bothered to put up a fight having the same assertiveness that the left has. Of course you will lose if you do not speak what you know to be true. Of course you will lose if you bring a knife to a gunfight.

Gene Marsh writes:

"I find it quite easy to tell them that they are envious and covetous, and that such emotions are not good foundations on which to form government policy. I have yet to get a response from them to this statement of fact."

Ali, envy is a fact but so is psychological projection.

I dont presume an anti-intellectual is envious of those whose interest area is reading abstruse literature.

It seems to me that people who prefer sports, detective novels, or management theory to Henry James really do! Am I supposed to believe that hostility to intellectuals is motivated by a secret longing for surrealist poetry? A longing for Daddy? Plausible. Dada? No way.

What's more likely is that you're of two minds. You're the average ambitious and avaricious bloke who was raised to believe that an excessive love of wealth and power sows suffering in the world.

But how does one honor tradition AND satisfy swollen ambitions? Jung knows. The shadow knows.

To eject what the conscious self knows is wrong and destructive without having to actually change, simply project it onto those people whose very existence threatens your fantasy that you have changed and that your monomania isnt pathological. Cast it onto those people whose expressed preferences seem to say, "im more interested in helping, or learning, or creating than building wealth".

Because if they're not lying, you're a singular threat to the future. Because if you're insatiable, all of our wisdom literature, everything passed down to us from our ancestors cautions us against you. It's just intuitive.


Gene Marsh writes:

"I get hammered by people on the left for opposing hair-brained schemes"

Seem like they might be saying "hair-braiding" schemes, since that's one third of the libertarian three-part plan to help the poor.

andy writes:

Scott, what Andrew wrote seems to me correct - how can an argument by 'plausible' when it is 'incorrect'? It can be be 'persuasive' for the general public, though.
My gripe with this comes in the moment when one asks: what horrible ideas can be justified by utilitarian arguments while using interpersonal utility comparison. Quite many, seems to me.

I think one can appeal to some basic humanity. The argument can be 'we are humans, some are unfortunate, we should take care of those, who are. The very basics are some food and clean shelter - rich people are far away from the basics, it makes sense to take from them and give to the poor to achieve these basics. I don't care if the 'total social utility' increases or decreases; the rich could be very unhappy about the situation, the poor might not be made happier - yet this policy is justified'.

It seems to me one can make an argument to alleviate 'absolute poverty' without interpersonal comparisons, the arguments against 'relative poverty' are much harder. Which is consistent with the libertarian view that most 'relative poverty' arguments are incorrect.

Daniel Kendrick writes:

Scott, you may not be a Maoist, but you are a collectivist. Utilitarianism is inherently collectivistic (and, for all intents and purposes, altruistic, since your own happiness is only one unit out of seven billion). You agree with the Maoists as to the desired end; your only point of disagreement is the means.

But the essential point is this: why should anyone be concerned with advancing the greatest happiness of the greatest number of people? There is not a single earthly reason you can give why each individual agent ought not to be concerned with his own happiness—but rather with that of the greatest number. That is the fraud of every kind of altruistic-collectivistic duty-based morality (and make no mistake: utilitarianism is just as duty-based as anything Kant came up with): there is no reason to follow it.

Charlie writes:

David Henderson says that one of the things we are "most sure about in economics" is that we can't do interpersonal utility comparisons.

http://econlog.econlib.org/archives/2015/05/tyler_cowen_on_14.html

I can't believe Scott Sumner completely sat out of that debate.

Is it wrong of me to demand that you two duke it out for our amusement? What else is a blog good for? Surely, it will increase blog readers utilities more than it will decrease blog writers (or maybe it's impossible to know).

Ann on a mouse writes:

Like Dr. Sumner, I'm a libertarian. This means that I favor violent redistribution of wealth (I mean "consumption"!), massive government big enough to violently control all polluting big business (hopefully world-wide!), and a complete abolition of the meaning of words.

Daublin writes:

Here here, Scott!

I'm pretty sure that just about any individual will gain more from having their own place to live, than any other individual would lose by going from two yachts down to one yacht. That's an inter-personal utility comparison.

Ali Bertarian writes:

Daublin wrote:
"I'm pretty sure that just about any individual will gain more from having their own place to live, than any other individual would lose by going from two yachts down to one yacht. That's an inter-personal utility comparison."

I'm pretty sure that you are considering only goodies in your definition of "gain." Too bad your utility formula includes only material goods as something to be valued by humans.

I'm also pretty sure that the person who had two yachts, but now has one, in the future will not work nearly as much as he did before losing that yacht, if it was forcibly taken from him. How does that decrease in work affect your utility formula?

Finally, I am certain that your utility formula does not have a copyright on it, because that formula does not exist for two people, let alone 6 billion.

foosion writes:

Unlike many on the left, I don't envy the rich.

Is there any good evidence those on the left envy the rich at a different rate than those on the right?

foosion writes:

I'm also pretty sure that the person who had two yachts, but now has one, in the future will not work nearly as much as he did before losing that yacht, if it was forcibly taken from him.

That's far from clear. He might work harder to get it back.

David S writes:

foosion:

While he may work harder to get it back, he is unlikely to do it in a way that you would like. Just working harder obviously won't work under a confiscatory regime.

If the government is confiscating your wealth, you have 2 choices: join the government, so that you can confiscate wealth as well; or invest in a private military. Neither of those options will help society, though...

J Mann writes:

I am 100% serious when I say that this is one of my all time favorite comment threads ever.

I have gained several utils from reading the back and forth, and thank you all.

Comments for this entry have been closed
Return to top