Bryan Caplan  

Utility Isn't Everything

Krugman on Federal Debt... Securitization and Credit Defa...
Scott Sumner is one of the great monetary thinkers of our age.  But the news isn't all good - he's also a utilitarian

I am frankly mystified by the enduring popularity of a moral theory subject to so many simple but devastating counter-examples.  My best guess is that people stick with utilitarianism because competing moral theories seem even more ridiculous.  But the main reason that competing moral theories seem even more ridiculous is that utilitarians only consider ridiculous alternatives!  In particular, if you question utilitarians' view that only utility matters, they assume that you believe that utility doesn't matter at all.  And if you grant that utility matters at all, they assume that you believe that utility is all that matters.

Scott provides an excellent example of what I'm talking about in Appendix B of his "Great Danes" paper:
My view of liberal values is probably a minority view, as I think most would want to augment utilitarian concerns with some sort of concept of "fairness" or "human rights". Let's consider four possible principles that might be viewed as "rights"; liberty, private behavior by consenting adults, free speech, and non-traditional marriages. We have already seen where liberals are willing to discard the right to liberty--the military draft in a "just war." "Consensual private behavior" won't work either--as liberals often support vice laws on utilitarian grounds--as with highly addictive drugs like heroin. And liberals often oppose free speech in areas such as commerce and hate speech. Finally, liberals do tend to support gay marriage, but not based on any abstract principle that one should be able to marry whomever one chooses--as most liberals oppose legalizing incest and polygamy. What do all four of these cases have in common? I would argue that utilitarianism is at work in all four cases. Liberals will easily discard any abstract "human right" if they think it that we can improve aggregate utility by restricting freedom. [emphasis mine]
If I were a liberal, my one-word response would be "Easily?!?!"  My longer response would be: "We do not discard human rights merely because we think we can 'improve aggregate utility.'  Human rights aren't absolutes; we are willing to discard them if the consequences would be terrible.  But we are willing to give up a lot of utility in order to protect these rights.  Heck, we're even willing to lose elections over gay marriage, even though we know that compromising on this one issue would allow us to mitigate a whole array of Republican evils."

Of course, as a libertarian I reject many of liberals' favorite "rights," and think that their utility calculators are woefully inaccurate.  But I'd still accept the basic structure of the reply I just sketched on their behalf: Utility matters, but it's far from the only thing that matters.

Comments and Sharing

COMMENTS (21 to date)
bbb writes:

Shouldn't two issues be separated when one criticizes utilitarianism?, namely:

its collectivism (its focus on aggregate "social" utility) and its consequentialism (the position that political measures are to be evaluated by their evaluated consequences = "utility")?

I only view the first position as wrong, the second only poses a problem if it is naively interpreted.

The first position was criticized by J.M. Buchanan as an illegitimate extension of the economic maximizing paradigm from the individual sphere to the "social" sphere - a departure from normative individualism of economics, i.e. the premise that the basis of evaluation can only lie in individual preferences.

Regarding the second position there arises a problem from the fact that statist policies recommended on utilitarian grounds which infringe liberty often focus only on the seen benefits of the measures and neglect the long-term unintended and unseen consequences that they will cause. However, this is not a problem of the consequentialist position, but only reflects the neglect of unintended consequences.

If the political advice of the defenders of liberty wants to be accepted, one should give reasons why the defense of liberty should be in the listener's own interest. The most convincing argument for this proposition is that ultimately a liberal (as in "free") social order will produce outcomes, that are, as evaluated by the preferences of the individuals themselves, superior than those of a "less free" order. That is, outcomes that yield a higher "utility" for the individuals. The personal freedom to choose could well be included in this utility calculus as an "option value" of freedom.

Tyler Cowen writes:

Welcome to messy pluralism!

IWantCookieNow writes:

Bryan, you are aware that your colleage Tyler is (self-described) "mostly utilitarian"? Talk to him...

Gary Chartier writes:

The counter-examples are, indeed, devastating for anyone with an ordinary moral conscience. Utilitarianism in its modern form tends to see one of its missions as the eradication of this conscience.

But even someone who isn't sure what to make of the counter-examples can be confident that utilitarianism is implausible. For the utilitarian has to suppose that welfare is some discriminable thing. And this is clearly false. The most obvious utilitarian proposals, like “happiness” and “pleasure,” don't work because, for instance, there are lots of things people want for their own sake that aren't either happiness or pleasure (life, friendship, knowledge, play, aesthetic experience, for instance). And wanting these things for their own sakes means (as should be obvious to anyone who wants them) not wanting them as means to some independently desirable psychological state. We do want some psychological states for their own sake (the pleasure of eating chocolate, say); but that's not true of many of the things we want. So utilitarianism is phenomenologically way off base here.

But if there's no single thing at which all of our actions aim (or should aim), then there's no quantity to maximize, so that the whole notion of a greatest good for some greatest number (however the relevant terms are interpreted) is incoherent. It's not that it's wrong or foolish to seek this greatest good; it's that there is no greatest good—there are lots of different goods, which are incapable of rational aggregation.

I hesitate to venture into social scientific territory, but doesn't the denial that interpersonal utility comparisons are possible have much the same practical consequence?

Robert Wiblin writes:

I am a down the line utilitarian because I don't give weight to any of the moral intuitions that drive your so-called 'counterexamples'. I am a moral non-realist so it's all arbitary anyway, but I would rather be truly compassionate in my morality, rather than try to pass off a selfishly strong concern for my own moral 'preferences' as a noble concern for some imagined objective moral law.

Utilitarianism = applied compassion.

Our ideas of fairness and justice evolved as heuristics to benefit small groups of humans in a vastly different situation to that which we exist in now. Why care what your intuitions in regard to those things say? Why not think about what alternative moral intuitions would produce the most of what people want, and make yourself use those ones.

Fundamentally: Why protect 'human rights' or anything else if that's not what people *want the most or benefit the most from*? If you're an atheist, there's *nothing else in the universe to please except the preferences of sentient beings*.

Paul Zrimsek writes:

I am a down the line utilitarian because I don't give weight to any of the moral intuitions that drive your so-called 'counterexamples'.

Thanks for the warning.

SRF writes:


I hereby assert that my personal utility will be 1,000,000x the combined possible maximum utility of all other humans in existence if you become my personal slave for the rest of your life. You now have in your power the single most important factor for maximizing the utility of humanity.

My first command: Reply to this post with your email address and phone number so that I can contact you and give you your next command.

Robert Wiblin writes:

SRF: I don't believe you.

SRF writes:

So can I (or anyone) use that excuse whenever someone says I should do something because it will increase utility? Or do you have a way to measure it (in which case I invite you to measure me, you will then find that you should indeed become my slave)?

Robert Wiblin writes:

SRF: Utility is hard to measure at this point, I'll grant you, but I can infer it to a practically useful level from my knowledge of the human experience/how the brain works/how people and myself behave, etc.

In any case, pointing out "what you want is hard to measure" is hardly likely to make me change what I want. Measuring a good education is harder than measuring how much money we spend in schools, but I doubt that will persuade you to choose the 'easy to measure' thing as your terminal goal.

Philo writes:

*Devastating* counter-examples? You must be kidding. These are just grist for the mill of classroom discussion.

SRF writes:

I'm not claiming "hard to measure", I'm claiming "impossible to measure". Utilitarianism is fundamentally incoherent because there is no conceivable way to measure or compare utility.

Beyond that, do you grant that if it were possible to establish that my enslaving you results in a positive utility gain that you would submit to it?

It seems to me that the moral non-realism doesn't imply that you should try to maximize human happiness. Just the reverse: you should try to maximize your own happiness and give no thought at all to anyone else.

Might I suggest that your strategy should be to publicly argue for Natural, Humand and God-given rights so that as many people as possible will respect your (fictional in you view) rights. Then at every moment do what you believe will maximize your utility with no thought about anyone else (expect to the extent that keeping others happy will be good for you).

For example, if you happen to meet (or intentionally follow) a young girl with $1000 in her pocket into a dark alley you should rape her (assuming your into that sort of thing), steal the $1000 and murder her. Clearly you would need to do a risk-benefits analysis regarding the possibility of being caught, but assuming this chance is low, go ahead. The question of her (and her family's) loss of utility as balanced against your own gain is really irrelevant from your point of view, isn't it?

The bottom line is that utilitarianism is just another arbitrary moral code. If you're going to support some morality you should choose one that has less chance of resulting in a net utility loss for you personally.

Robert Wiblin writes:

It's impossible to measure directly at this point, but lots of our knowledge comes from indirect inference and this is no different. Are you a Bayesian?

If someone is screaming in rage at another person who has done something bad to them, would you really say "I have no evidence whatever as to whether they are angry or not, as I can't measure their state of mind directly. Working out whether people are angry is just impossible." That is just silly.

"It seems to me that the moral non-realism doesn't imply that you should try to maximize human happiness. Just the reverse: you should try to maximize your own happiness and give no thought at all to anyone else."

Moral non-realism implies nothing about what you should do. It doesn't say selfishness is good.

Yes utilitarianism is an arbitrary moral code - it's the one that appeals most to me given my compassion, love of consistency and study of evolutionary psychology.

Robert Wiblin writes:

Sorry to double post, but this is a pretty great blog post you might like:

Robert Wiblin writes:

"Thanks for the warning."

If what really gives you pleasure is fairness, justice, etc then utilitarianism would require me to care about providing those things - I just wouldn't prioritise them over other things people want just because they're 'moral' rather than 'hedonic' preferences.

SRF writes:

I consider myself an Austrian/Libertarian.

I don't deny the reality of utility/happiness (or anger), just the possibility of an objective external measurement. An individual can know what make him more or less happy, but a third party can't verify this, quantify it (say that Jack likes apples 3x more than he likes oranges) or compare it to another person. The exercise of maximizing utility is hopeless. In fact it makes me very unhappy when someone tries to compute utility :).

"Moral non-realism implies nothing about what you should do. It doesn't say selfishness is good."

If I understand correctly the implication is stronger than that. It says that there is no such thing as what "should" be done and it is meaning less to talk of anything being "good" or "bad" in any objective sense.

I'll grant that if you get your jollies by attempting to maximize other people's utility, it is consistent for you to do so. But you also are committed to saying that someone else's preference for rape, murder and genocide is just as legitimate.

What I was pointing out is that you are better off not trying to convince others of this view. If you succeed there is a chance they will prefer to do something nasty to you.

As for the blog post, I can't say I agree. The author ignores the fact that throughout history politics has been the enemy of technological and economic progress. It is no coincidence that the drastic reduction of political oppression in western countries resulted in a massive increase in wealth and technology (as explained here: He also assumes the moral judgment that wealth and progress are "good".

Phil writes:

As I learned it, utility means nothing more than value. What we commonly refer to as utilitarianism makes assumptions about what people value, and our ability to know and measure such. That makes it a problematic (at best) approach. But if utility simply means matters to us, how can it not be all that matters, as Bryan says? How can there be a moral philosophy that is not a variant on utilitarianism? If you object to the arbitrary set of values commonly labeled utilitarianism because it allows for violations of liberty, and you value liberty, it seems to me that you haven't found a contradiction in the notion of maximizing utility, you've simply pointed out that this system is not in fact maximizing utility because reducing liberty reduces utility. I don't think it's helpful to argue about whether we should maximize an abstract concept of "the good". It's just that this concept is so vacuous that the real question becomes what is "the good" or "utility" and what policies will maximize it.

Kevin writes:

Tyler: ARG. Consequence-sensitive deontology is not the same thing as messy pluralism! Bryan, avoid the Cowenian swamp of moral theory-less-ness!!

Bryan: Your post is inaptly title. Utility isn't something that could be everything. It's the measure of preference and preference, not a substantive thing in itself. In fact, it might be appropriate to say that utility measures everything in terms of valuations of states of affairs.

Paul Zrimsek writes:

Kevin's last point is illustrated well by Wiblin's reply to me, which tries, at one and the same time, to disparage moral-realist values and to bring them within the utilitarian system. The message seems to be, "We'll take into account the disvalue you place upon murder and slavery, but only after you acknowledge that you're not supposed to disvalue them as much as you do."

Robert Wiblin writes:

"to disparage moral-realist values and to bring them within the utilitarian system."

If I gave that impression I take it back. It's all arbitrary - I just like this moral system.

"only after you acknowledge that you're not supposed to disvalue them as much as you do."

While obviously I would prefer that you shared my precise moral 'utility' function, my caring about your interests is not dependent on that at all.

Tim K writes:

Utilitarians evaluate the utility of third parties based on their own preferences (perhaps subconsciously). If a utilitarian rejects slavery on the grounds of reduced aggregate utility, then they must believe the lost utility of the slave is greater than that gained by owner.

In this respect it's not necessary to exactly measure changes in utility, as its really a reflection of the utilitarians' own preferences. No one can claim the absolute right or wrong answer in such a situation, it can be equally plausible for a utilitarian to favour slavery if he belives aggregate utility would improve.

No moral theory provides the course of action agreeable to all, but utilitarians do allow you to easily understand what they value. If nothing else, its a great indicator of personal preferences.

Comments for this entry have been closed
Return to top