April 29, 2003

I've Never Understood Sen's Paradox

Robert Waldmann is thinking about Sen's Paradox and the contradictions between utilitarianism and liberalism. I'm confused, because I never understood Sen's Paradox at all. The way Robert states it, Sen's Paradox is as follows:

Sen's paradox is well known, even if the example is painfully dated. Once upon a time Lady Chatterley's Lover was considered obscene. (Amazingly, recently, the courts came to their senses.)

Sen considered what [Vilfredo] Pareto might say of a world in which lived two agents, Lewd and Prude. Lewd wants to read porn, but finds the thought of Prude being forced to read Lady Chatterley's Lover even more exciting. Prude finds Lady Chatterley's Lover appalling, but is even more disturbed by the thought of Lewd reading it and enjoying it.

Freedom [Sen says] is not Pareto optimal in this case. [Freedom leads to Lewd reading Lady Chatterley's Lover and Prude not reading it. But it] is Pareto better[, that is, everyone is happier in a utilitarian sense,] to block Lewd from reading LCL and to force Prude to read it.

This is crazy. And, in any case, it shows that Liberalism and the [utilitarian] Pareto principal might hypothetically be in conflict.

It's at this point that my brain goes, "Booiiinnnggg!" The Liberal--libertarian--outcome is not for Lewd to eagerly read Lady Chatterley's Lover and for Prude to stand angrily by. The Liberal outcome is for Lewd to say, "I'll pay you $100 if you read this book," for Prude to accept (there is a bargain here that both would agree to at some price), and for Prude to then say "I'll pay you $150 if you don't read this book" (for here, also, there is a bargain that both would agree to at some price). Through these mutual acts of commercial intercourse by consenting adults, you attain a point on the first-best contract curve.

So where is the incompatibility between Paretian-style utilitarianism and classical liberalism? What's the paradox? I don't see it at all. I never have.

Posted by DeLong at April 29, 2003 01:31 PM | TrackBack

Comments

Brad, I'm really surprised you ahven't read "Choice, Welfare and Measurement."

Anyway, in the Chatterly example there are four states of the world, so the side payments option doesn't exist. It's introduced just as an example of a broader assertion that under reasonable assumptions, you cannot have a decision process that will both yield a Pareto-optimal outcome and treat certain sets of outcomes as "private" -- i.e. all but one agent's preferences over them are disregarded.

At least that's how I recall it -- been a couple of years...

Posted by: jw mason on April 29, 2003 02:12 PM

I would add that in your example, allowing payments, the outcome is pareto-optimal but there are no strictly private devisions. That's the point of the paradox.

This is clearer if we take the identical example you gave, but describe it slightly differently: Prude will fine Lewd $150 if he does read the book.

Posted by: jw mason on April 29, 2003 02:25 PM

Brad, I think the paradox comes only when there is a COMMAND-AND-CONTROL LAW THAT MAKES PRUDE READ PORN AND PROHIBITS LEWD FROM READING PORN. However, Brad is discussing a voluntary trade. THE VOLUNTARY TRADE DOES NOT VIOLATE LIBERTY. They have the same effect of course.

Occasionally voluntary trade may usually not be possible, and the command-and-control law is the only way to achieve the Pareto efficient allocation. In this case the paradox is unavoidable. Is this right?

Posted by: Bobby on April 29, 2003 02:25 PM

So the paradox applies to command and control Pareto improving laws, which violates liberty.

The paradox does not apply to voluntary Pareto improving trades, which do not violate liberty.

They have the same effect, but the paradox applies to the first and not the second.

Posted by: Bobby on April 29, 2003 02:30 PM

Bobby,

Sure, that's sometimes the case, but it's not Sen's paradox. Sen's paradox is that liberalism requires a set of genuinely private choices and that this is in general not compatible with the Pareto criterion.

(Not that your question was addressed to me.)

Posted by: jw mason on April 29, 2003 02:33 PM

So you are saying that Sen's paradox is that if there are genuinely private choices (and we equate these with liberalism) "this is in general not compatible with the Pareto criterion." This sounds more like a simple denial of the First Fundamental Theorem of Welfare Economics. By renaming this Sen's paradox, the only new contribution we are making is that we're tacking on the word liberalism to describe private choices. Is this what Sen's paradox is? I'm sure that he meant more than this.

Anyway, I really don't know what Sen's paradox is beyond DeLong's post and something from Slate by Steve Landsburg. So I'm just speculating.

Posted by: Bobby on April 29, 2003 02:48 PM

... But wouldn't the transaction costs of monitoring either Lewd or Prude would eat up all of the potential gains from trade?

Posted by: Lou D on April 29, 2003 04:33 PM

Lou,

I think for Lewd and Prude would each find that monitoring the other's compliance would be an additional benefit.

;-)

Posted by: Tom on April 29, 2003 06:05 PM

Just in case you thought this was easy:

http://www.theihs.org/libertyguide/hsr/hsr.php/47.html

Posted by: Tom Maguire on April 29, 2003 07:11 PM

It was always pretty improbable that Sen might have ignored this argument and as pointed out above, he devotes a long discussion to it in his book. As noted, the Chatterly case is just a heuristic modelling assumption. If you don't like it, substitute something else ...

Brad, do you not realise that this line of thinking commits you to supporting the following gangs of people:

* People who want to pay crack addicts to get sterilised.
* People who want to pay all black immigrants a sum of money to go back to Africa
* People who want to offer Jews money to convert to Christianity
* People who provide televisions to schools in exchange for the pupils watching advertising

There are some contracts that a liberal state doesn't get involved in enforcing. Extend your example with another common libertarian trope; if some day after the contract is made, Lewd happens to pick up a copy of Lady Chatterly, would you support the use of "men with guns" to make him put it down?

Posted by: dsquared on April 29, 2003 11:12 PM

Actually, I've just thought of an attractively Goedelian "short way" to show that there is a genuine paradox here.

For Lewd and Prude substitute Naomi Klein and Brad DeLong. For Lady Chatterly's Lover substitute "the practice of making horse-trading financial exchanges about everday behaviour".

Brad wants to be able to make financial contracts with people to carry out or refrain from carrying out everyday actions which others happen to disapprove of, but gets even more excited about the thought of Naomi Klein being forced to do things his way. Naomi finds the practice repugnant, but is even more disturbed by Brad doing it all the time.

If a dictator came in and forced Brad to stop this practice and Naomi to start, let us assume that would be a Pareto improvement. But if Brad offers this contract to Naomi, she will tell him to stuff it; it is true that this is a state of the world in which she is /worse off/, but she has higher level procedural views which prevent her from making herself better off by making repugnant contracts. In other words, a Pareto improvement can be made by *forcing* the global optimum on Brad and Naomi but (for sufficiently high levels of disutility of making the contract for Naomi), it can't be achieved by *contracting* to the global optimum, since making the contract is a really bad thing for Naomi.

And a *liberal* state can't rule out this kind of higher-order preference about the contracting process; it's as valid as any other. Note that although this argument has a bit of a Goedelian flavour, it's not a rinky-dink corner case; what it shows is what Sen says; that utilitarianism can't cope with private choices and that liberalism has to accomodate them.

Posted by: dsquared on April 29, 2003 11:30 PM

>>Extend your example with another common libertarian trope; if some day after the contract is made, Lewd happens to pick up a copy of Lady Chatterly, would you support the use of "men with guns" to make him put it down?<<

Ah. Clever.

Posted by: Brad DeLong on April 30, 2003 06:02 AM

>>It was always pretty improbable that Sen might have ignored this argument and as pointed out above, he devotes a long discussion to it in his book.<<

Which I do not understand...


Brad DeLong

Posted by: Brad DeLong on April 30, 2003 06:04 AM

I am a fine one to talk about this, by the way, as my entire knowledge of Sen comes from other people's summaries of him. I own both his big books (Choice, Welfare and Measurement and Welfare as Freedom) and have made several good faith attempts to read both, but I just always seem to wake up dribbling an hour later. Some prose styles just sit heavy on my stomach. The version of Sen above comes via Bernard Williams, unless I misremember, which I probably do.

Posted by: dsquared on April 30, 2003 08:22 AM

>>* People who want to offer Jews money to convert to Christianity

Abe and his friend Sol are out for a walk together in a part of town they haven't been in before. Passing a Christian church, they notice a curious sign in front that says "$1,000 to anyone who will convert." "I wonder what that's about," says Abe. "I think I'll go in and have a look. I'll be back in a minute; just wait for me."

Sol sits on the sidewalk bench and waits patiently for nearly half an hour. Finally, Abe reappears.

"Well," asks Sol, "what are they up to? Who are they trying to convert? Why do they care? Did you get the $1,000?"

Indignantly Abe replies, "Money. That's all you people care about."

-- snip --

Off topic (you can delete it, if you want, Brad), I know, but I dunno... I can see how the four instances you provided -- sterilizing crack users, paying blacks immigrants to go to Africa, paying Jews to become Christians, or providing TVs for kids to watch advertizing: I know those go against my values, dsquared values, and many others' values, but is that really a welfare issue? It seems to me that the "liberal" condition of allowing people to do the things that dsquared mentions only violates Pareto optimality if third-party observers like us would be willing to compensate the anti-immigrant people and the black immigrants to NOT make the contract to migrate to Africa, or willing to pay schools and TV suppliers NOT to place TVs in exchange for advertizing eyeballs (or whatever), and if that's the case, we third-party types can do just that: make our own contracts for the other contracts not to be made!

And so on and so forth...

Of course, I'm sure Sen knows a lot more about this than I do. That's why he's a Nobel-prize winning economist and I'm a loser college freshman who trolls various obscure blogs, but I don't see how potential contracts that go against the values of some third parties NECESSARILY violate liberalism or Pareto optimality.

Posted by: Julian Elson on April 30, 2003 09:12 AM

Economists are so good at completely missing the point. Hasnt it occurred to anyone to ask WHO would be the one to decide which contracts to force on people? There isnt any such person and even if there were, their decisions on Day 1 might not hold for Day 2 since society continues to evolve. This whole train of thought strikes me as something that a TA I once had called "things only middle aged economists with tenure can take seriously" (and not even all of them)
As for being liberal, that means (a) you can read porn if you like (b) if you dont like it then dont look at it and if thinking about other people's sex lives or lack of them troubles you then seek therapy because it isnt their problem its yours (c) for those who are not yet able to make fully informed decisions (e.g. school kids) we make decisions for them through our school boards which either do or dont allow advertising. For the rest of us, there may indeed be a price at which any one of us might, e.g. leave the country. I certainly have mine, so please start bidding.

Posted by: steve on April 30, 2003 09:57 AM

And what's with this equating of liberalism and libertarianism? By any definition I'm familiar with they are not even close to the same thing. I consider myself to be a liberal, more or less... and the opinions of those people on the net who self-identify as libertarian almost uniformly make my hair stand on end.

Posted by: Canadian Reader on April 30, 2003 11:59 AM

From my side of the Atlantic the word liberal has not exactly the same meaning it has in the USA, don't know about Canada, I am under the impression that most economists, at least those not from the USA apply this other meaning, it may be what happens with Sen which seems to have grown in India and UK ( http://www.nobel.se/economics/laureates/1998/sen-autobio.html ).

DSW

Posted by: Antoni Jaume on April 30, 2003 12:23 PM

Surely there are some issues on which even libertarians would feel uncomfortable denying the force of this paradox. For instance: we are surely not expected to PAY criminals to refrain from engaging in immoral acts, are we? Sen's point seems to be that defining immorality as an economic parameter assumes away the problems of judging right from wrong in a liberal society.

What if the issue at stake isn't whether someone has the freedom to read a book, but rather the the freedom to kill?

Posted by: david on April 30, 2003 01:38 PM

A non-economist's point of view.

First, liberal here means approximately weak-sense libertarian or Friedmanite, not welfare liberal as it does in the states.

To liberals, society is a group of atomic individuals, each with their own individual desires. The liberal assumption is that the best system is one which allows each individual the freedom to act to satisfy his own desires, with no social or collective interference. A market society.

The utilitarian says that the best society is the one which, in the end, satisfies the most needs of the most people. (Yeah, lots of questions about what that means).

The Sen paradox says that the ideal liberal society will not be the ideal utilitarian society. The personal-freedom market society is not the best at satisfying the most needs of the most people.

What I think that the Sen paradox does is ask what happens if there are second-orer desires, i.e. if the satisfaction of party A's desire is significantly dependent on the satisfaction of party B's desire. He chose the negative case of two people who hate each other because of the economics profession's taboo against moralizing arguments. He could have chosen a case in which A and B were related by affection rather than hatred. In either case you get complex feedback relationships making the simple individual-choice free-market solution actually non-utilitarian in terms of outcomes, since no one's desire is autonomous but also dependent on the desire of others. So communitarian or statist interference in the market becomes justifiable.

Posted by: zizka on April 30, 2003 02:19 PM

"Surely there are some issues on which even libertarians would feel uncomfortable denying the force of this paradox. For instance: we are surely not expected to PAY criminals to refrain from engaging in immoral acts, are we?"

Libertarianism doesn't care about the "morality" of actions. It's a completely Amoral (not IMmoral!) philosophy.

Libertarians support government actions that involve one person committing (unwanted) violence or fraud against another person. But that has nothing to do with whether those actions are "moral." (Though just about everyone would agree that such unwanted actions are immoral.)

But libertarians do NOT support government interference in matters between consenting adults...REGARDLESS of whether the particular libertarian considers the act to be moral or immoral.

For example, there are libertarians that think homosexual acts are morally OK. And there are libertarians who think that homosexual acts are immoral. But no TRUE libertarian supports government actions to either discourage or promote homosexuality.

An example: There was a Libertarian in the state legislature in...New Hampshire? Vermont?...I forget. But a bill came through the legislature to allow homosexual marriages. He opposed the bill, for moral reasons. He was *roundly* criticized within the Libertarian Party, and eventually joined the Republican Party.

The point is that Libertarians would not have criticized him for speaking out against homosexual marriage; but when his religious views caused him to involve the State in the matter, he was no longer acting in a libertarian manner.

Another example would be prostitution. Some libertarians would have moral problems with prostitution, while others would not. But no REAL libertarian would support laws against prostitution (or government funds to promote or regulate prostitution).

As for "paying criminals to refrain from engaging in immoral acts": libertarians would oppose *governments* paying criminals to *not do* anything...regardless of whether libertarians thought those acts were moral or immoral.

For example, libertarians absolutely oppose the federal government paying farmers NOT to grow crops. (Or to grow crops!) That would be true, regardless of whether the crops were "moral" (e.g., wheat) or "immoral" (e.g.,....well, some people would say marijuana, or opium, or coca).

"What if the issue at stake isn't whether someone has the freedom to read a book, but rather the the freedom to kill?"

No libertarian supports a "freedom to kill"...at least not to kill "persons," unless it is in self-defense.

Every libertarian supports the right of self-defense...e.g., a person comes into your home, you think that person is going to hurt you, so you shoot him. And every libertarian (that I know of) supports the right of people to kill animals, without government interference, either for eating, or for fun. (Again, libertarians may find this practice MORALLY reprehensible--I think hunting for fun is morally wrong--but they wouldn't involve the government.)

A more tricky situation occurs regarding abortion. Some Libertarians argue that the government should get involved in abortion (i.e., pass laws against abortion). But a majority of Libertarians, as represented by the Libertarian Party platform, have held that the government shouldn't be involved in passing laws against abortion. (And absolutely no true Libertarian would support *federal* government involvement, either to promote or oppose abortion.)

Posted by: Mark Bahner on April 30, 2003 02:59 PM

Oops. This doesn't seem to have posted:

"And what's with this equating of liberalism and libertarianism? By any definition I'm familiar with they are not even close to the same thing."

I don't know what your definition of liberalism is. I do know the definition of libertarianism. In fact, it is on the back of every U.S. Libertarian Party membership card:

"We hold that all individuals have the right to exercise sole dominion over their own lives, and have the right to live in whatever manner they choose, so long as they do not forcibly interfere with the equal right of others to live in whatever manner they choose."

In other words, libertarians think that the only purpose of government is to address when people are committing violence or fraud against other people.

Libertarians would not support laws to either: 1) force Mr. Prude to read the book, or 2) forcibly prevent Mr. Lewd from reading the book.

Libertarians would only support government intervention if: 1) Mr. Lewd spray-painted excerpts from the book on Mr. Prude's property, or 2) Mr. Prude stole Mr. Lewd's book, in order to burn it.

A complicating fact is that libertarians of the time of the founding of the U.S. (e.g., Thomas Jefferson) were called "liberals," at least beginning in the early 1800's. In essence, their philosophy was defined by the motto, "laissez faire" ("let do," or "allow to do.")

http://www.britannica.com/eb/article?eu=47953

Another significant problem is that Thomas Jefferson and others of the time failed to denounce and completely oppose slavery...which is about the worst example of "forcibly interfering with the right of others to live as they choose" that I can think of.

Also, libertarians of today often avoid calling themselves liberals, because of the bad things that "liberals" like Franklin Roosevelt and Lyndon Johnson did with the term. Sometimes libertarians call themselves "classical liberals."

http://www.libertarianism.org/ex-3.html

Posted by: Mark Bahner on April 30, 2003 03:07 PM

In few words, libertarianism is the way of things, not of peoples. Adquired Psycopathy.

DSW

Posted by: Antoni Jaume on April 30, 2003 03:18 PM

"The Liberal--libertarian--outcome is not for Lewd to eagerly read Lady Chatterley's Lover and for Prude to stand angrily by. The Liberal outcome is for Lewd to say, "I'll pay you $100 if you read this book," for Prude to accept (there is a bargain here that both would agree to at some price), and for Prude to then say "I'll pay you $150 if you don't read this book" (for here, also, there is a bargain that both would agree to at some price)."

See, this is why I come here. :-)

I would have said that the libertarian outcome WAS to have the two fools stand angrily by, while each did actions that was offensive to the other.

But I can think of an even better solution than Lewd paying Prude $100 to read the book, and Prude paying Lewd $150 not to read the book. My solution would be for Lewd to pay Prude $100 to read the book, and Prude to pay Lewd the exact same amount not to read the book.

And then the State, borrowing from the Monty Python writers, could drop a 100-ton weight on the both of them. ;-)

Posted by: Mark Bahner on April 30, 2003 03:18 PM

The contemporary American usage of "liberal" is anomolous. Basically Roosevelt tried to smuggle a bit of social democracy into liberalism under the "Four Freedoms", which included "freedom from want" which was not a freedom granted by any classical liberal. It was sort of a con job, though it made sense given that classical liberalism leaves some people free, but destitute and totally at the mercy of others.

Actually, John Stewart Mill already moved in this direction, but he was functioning more as a utilitarian than as a liberal then. IMHO.

Classical liberalism was free-marketism or near-libertarianism.

Posted by: zizka on April 30, 2003 03:22 PM

"Basically Roosevelt tried to smuggle a bit of social democracy into liberalism under the "Four Freedoms", which included "freedom from want" which was not a freedom granted by any classical liberal."

And don't forget "freedom from fear."

In fact, G.W. Bush could easily have used FDR's "right" to a "freedom from fear," as a justification for attacking Iraq.

http://www.libertynet.org/~edcivic/fdr.html

"It was sort of a con job, though it made sense given that classical liberalism leaves some people free, but destitute and totally at the mercy of others."

Yeah, right. Go to Cuba or North Korea (you may stay there, if you'd like ;-)), and then try to tell me that the absence of classical liberalism leaves people LESS "destitute and totally at the mercy of others."

In a classical liberal (libertarian) society, no one who is mentally and physically capable of performing ANY service is "totally at the mercy of others."

And whether the society is libertarian or totalitarian, or anything in between, anyone who is completely physically and mentally incapable of doing ANY service is "totally at the mercy of others." It's simply a matter of whether the "other" is a private individual or the government.


Posted by: Mark Bahner on April 30, 2003 03:43 PM

"In few words, libertarianism is the way of things, not of peoples. Adquired Psycopathy."

Complete BS, and 180 degrees wrong. Libertarianism stands above *any* other political ideology, in terms of consideration for individual people.

Communists (e.g. Stalin) build canals, and though the building of the canal causes the deaths of 100,000 (mostly political prisoners/slaves), still considers the canal to be a triumph.

Democrats and Republicans build the Big Dig in Boston, and though the creation of that tunnel/road/bridge involves the legalized extortion of $10+ BILLION dollars, mostly from people who will NEVER use the structure, call the project a success.

Communists, Fascists, Democrats, Republicans...they don't even come CLOSE to Libertarians, when compared Libertarians' caring about individual people, over things.

Posted by: Mark Bahner on April 30, 2003 03:55 PM

Mark, if had to choose between classical liberalism and Kim Il Sungism I would choose classical liberalism. You have a very, very coarse mesh in your sieve. There are more than two choices here.

The deficiencies of classical liberalism were documented by a flood of fiction and non-fiction books written between 1850 and 1920. Perhaps you should read a few of them.

Mark, you should go to Antartica and live on penguin eggs and krill. ;-))

Posted by: zizka on April 30, 2003 04:04 PM

One place where I part company with libertarians is over the issue of the power of aggregated wealth, which, it seems to me, is every bit as capable of interfering with the freedoms of large numbers of people as any government has ever been. I want a society where there are also checks on corporate power.

And then there's the question of safety nets. Libertarians seem to imagine a world composed of spherical human beings of uniform density, who never encounter bad luck or injustice, and are all equally capable of fending for themselves from birth -- or at least puberty -- until death. Poverty, accident, disease, and old age, do nevertheless exist. I want a society where a decent life is possible even for those who can't -- possibly could never, or perhaps, can no longer -- provide for themselves. Not a do-all "nanny" state, no, but one that does something to help smooth things out. In case, you know, through no fault of my own, one of those things happens to me!

But mention that to a libertarian, and right away they're ranting about Cuba.

In fact, that's my main beef with libertarians -- they're so all-or-nothing. A real society has to be a balancing act. What's wrong with a bit of moderation? A modern liberal (forget "classical," which is only a polite way of saying obsolete) is perfectly okay with making compromises. The goal is to improve the quality of life for everyone, and if that requires balancing governments, corporations, unions, tribal councils, languages... well, we'll do whatever it takes.

Posted by: Canadian Reader on April 30, 2003 07:29 PM

OK, got the book in front of me. (The relevant essay, "The Impossibility of a Paretian Liberal," is just 5 pages long. Maybe Brad and Daniel Davies should give it another try.)

Here's what Sen claims to prove.

Given condition P:

If every individual prefers x to y, then society must prefer x to y.

Condition L:

For every individual there is at least one pair of alternatives (x,y) such that if the individual prefers x to y, society must prefer x to y.

Then there is no social decision function which can simultaneously satisfy P and L for all preference orderings.

P is a weak form of the Pareto criterion, so this result holds a foreteori for the stronger form.

L is the liberal criterion: "ain't nobody' business if you do."* There are some outcomes over which my preferences are dispositive, regardless of what others might prefer.

Now Sen might have made a mistake in his proof, tho it's not complicated -- if you find one then, fine, no paradox.

It's also the case that many decision functions can satisfy P and L for some preference orderings. This is one of the points he makes in the essay -- that "the ultimate guarantee for individual liberty may not rest on rules for social choice but on developing individual values that respect each other's personal choices."

And of course if you reject condition L then there is no paradox either -- that is, if you regard my preferences about your reading habits as having exactly the same status as my preferences about my reading habits. But this is obviously not how most of us think, and in particular it seems incompatible with traditional ideas of liberalism.

It's significant that the essay is in the section of the book on "Non-utility Information." I would mildly suggest that an inability to process such an information is a hazard of "thinking like an economist."

* A Libertarian catchphrase, even if the Billie Holiday song is arguably suggesting just the opposite.

Posted by: jw mason on April 30, 2003 08:57 PM

umm, scratch the bit about thinking like an economist. It's sophomoric & annoying.

Sorry.

Posted by: jw mason on April 30, 2003 09:31 PM

But the whole substance of the argument is a mixture of a) whether L is a suitable representation of classical liberalism and b) whether there are in fact incompatible preferences of the kind which would make Sen's impossibility theorem practically relevant.

Posted by: dsquared on April 30, 2003 11:33 PM

I'm confused by syntax here.

I likes x but likes forcing x on J even better. J dislikes x but likes seeing I enjoy x even less.

Is that the general idea? I don't understand this usage of "even better" and "even less" -- the enjoyment or revulsion experienced by I and J in the two cases -- having an experience vs. watching the other have it -- seem like apples and oranges to me. What am I missing here?

Posted by: Jeremy Osner on May 1, 2003 06:29 AM

"I don't understand this usage of "even better" and "even less" --..."

I think those are superfluous to the issue at hand. Prude doesn't want to read the book, but ALSO wants to force Lewd not to read the book. Lewd wants to read the book, but ALSO wants to force Prude to read the book.

I don't see that their degrees of wanting are important. That the wanting is non-zero is the only thing that's important, as I see it. (Little Stalinists...makes me want to puke! ;-))

Posted by: Mark Bahner on May 1, 2003 09:41 AM

"Now Sen might have made a mistake in his proof,..."

Yes, the mistake he made was in thinking that there was any such thing as "society." "Society" is a completely fictional social construct. Individuals are real. "Society" is an invention of the human mind.

Rewriting the conditions, to remove the fictional social constructs:

Condition P: "If every individual prefers x to y, then ***every individual*** must prefer x to y."

Condition L: "For every individual there is at least one pair of alternatives (x,y) such that if the individual prefers x to y, ***the individual*** must prefer x to y."

Now there's no paradox. These are simply two statements of the reflexive property of equality ("a=a").

The problem ees solfed.---->Economic Inspector Clouseau. :-)

Posted by: Mark Bahner on May 1, 2003 09:52 AM

Mark: that's not a mistake in the _proof_. You are free to reject the axioms, but that's not relevent to the proof. A proof is correct if the conclusion follows logically from the axioms. It makes no difference whether the axioms apply to the real world.

Posted by: Walt Pohl on May 1, 2003 10:19 AM

Delong's point, I take it, is that the externalities creating by meddling preferences are no more of a problem than any other externalities. Of course, nor are they any less of a problem. Imagine that there are two Prudes to whom Lewd offers to trade his right to read the offending matter for the promise that it will be read in turn by Prude. Now we have a chicken game: If one knows the other will deal, he will get the gain with no pain and will not deal, while if the other will not deal, then he will deal, since the gain, by hypothesis, exceeds the pain.
In a mixed strategy equilibrium, each will deal with the probability p. Then there is the square of (1-p) probability that the welfare-improving deal doesn't get done. Bring in the State: Liberty and Utility part ways.

On an entirely different tack, isn't a hallmark of most Liberalisms, Libertarian or not, the certain trades are proscribed, certain rights inalienable. Just as, for most liberals, an individual cannot sell himself into slavery, mightn't it be anti-lberal to allow individuals to sell their right to free expression, or to read what they choose to read, etc? Understood this way, the Liberal/ Paretian Paradox reappears.

Posted by: kevin quinn on May 1, 2003 12:31 PM

dsquared:

On the preferences question, the problem arises if people have preferences regarding the private choices of others. I feel confident that describes the world we live in.

On the question of whether L describes liberalism, I guess it depends if you see any important difference between "You are free to read what you like" and "You are free to read what you like, you'll just have to pay a fine corresponding to its offensiveness to others in the community."

Posted by: jw mason on May 1, 2003 12:55 PM

Sorry. I thought society existed. My bad.

Posted by: zizka on May 1, 2003 03:43 PM

Ok, after re-tightening your brain realize that I can buy the book off e-bay for under $10. So go ahead, pay me $150 to not read it. I'll just not honor the deal, buy the book and walk away with a little over $140 dollars. Unless, of course, society comes to and end because I pissed off a paradox. And trust me, one should not piss off a paradox.

Posted by: Guy H. on May 1, 2003 07:33 PM

Ziska, remember that some people think that "no law then no crime"

DSW

Posted by: Antoni Jaume on May 2, 2003 03:13 AM

Brad's proposed solution to the L/P paradox has gotten under my skin! Let me ask again: would you call a society which allows enforceable contracts to be made alienating one's right to read or not read what one chooses a Liberal society? If not, then there is a conflict between value maximization and Liberalism. This seems to me a special case of the more general question of whether a liberal society allows people to freely choose to trade away personal freedom, or whether one's property in one's self is to be treated any differently than one's freely alienable property in everything else.

Posted by: kevin quinn on May 2, 2003 08:16 AM

"Sorry. I thought society existed."

Nope. Figment of your imagination. You've made the same mistake as Rick Santorum.

Individual people are real, though.

"Ziska, remember that some people think that 'no law then no crime'."

If they are speaking in strictly legal terms, they are 100% correct.

Posted by: Mark Bahner on May 2, 2003 09:30 AM

"Brad's proposed solution to the L/P paradox has gotten under my skin! Let me ask again: would you call a society which allows enforceable contracts to be made alienating one's right to read or not read what one chooses a Liberal society?"

Yes, of course. I would rather be reading Dr. DeLong's thoughts about potential paradoxes than reading/analyzing emission monitoring protocols. But here at work, I signed a contract to do work. So emission monitoring protocols it is, for 8 hours a day.

On the other hand, I would gladly accept...ummmm...$500, to be prohibitted from ever visiting this site again. (Better make it high enough that no one will take me up on the offer. ;-))

So yes, a classical liberal society could easily have enforceable contracts restricting one's right to read what one chooses.

Posted by: Mark Bahner on May 2, 2003 09:45 AM

Last post for lunch:

"Mark: that's not a mistake in the _proof_. You are free to reject the axioms, but that's not relevent to the proof."

OK. Point taken. :-)

But don't you think it's Sen's fault that he starts with axioms that aren't true in the real world?

If I start with the axiom that pigs can fly, and use it to prove that I can therefore have my pork BBQ airmailed to me...is my proof worth much? ;-)

Do you agree with the scientist who, when confronted with evidence that his conclusions were clearly wrong in the real world, replied: "The real world need not be considered! It is merely a special case!"? ;-)

Posted by: Mark Bahner on May 2, 2003 09:53 AM

Mark: In this case, precisely the opposite. Sen has proven that a "social welfare function" can't satisfy certain natural criteria. Since you seem to not believe in a "social welfare function", Sen's paradox is a point in your favor.

But of course you believe in a social welfare function. As a libertarian, you believe that some possible worlds are better than others. One where property rights are protected is better than one where all property is confiscated by the government, etc. Assign the first one a value of 10, the second one a value of 5, and voila, you have a social welfare function.

Posted by: Walt Pohl on May 2, 2003 11:43 AM

Walt -

Right exactly. Society exists, and has preferences, in the limited but important sense that individual choices are not independent. Given a set of possible outcomes over which individuals have conflicting preferences, somehow or other one outcome will be selected. That's all that social preference means.

As for Mark Bahmer's example -- didn't somebody or other say that liberal values ceased to apply once you passed through that door reading, "No entry except on business"?

Posted by: jw mason on May 2, 2003 12:47 PM

"Mark: In this case, precisely the opposite. Sen has proven that a "social welfare function" can't satisfy certain natural criteria. Since you seem to not believe in a "social welfare function", Sen's paradox is a point in your favor."

So Sen invented a paradox that I already knew never really existed. Give that man a Nobel prize! ;-)

"But of course you believe in a social welfare function. As a libertarian, you believe that some possible worlds are better than others. One where property rights are protected is better than one where all property is confiscated by the government, etc."

Yes, but my "belief" (actually, I have no "beliefs"...so it's really my "thought") that some possible worlds are better than others is based on looking at individuals, rather than inventing a "society."

I guess one could say that "society" could then be represented by the arithmetic sum of the happiness of each of the individuals. (Maybe that was the point of the numbers you used.) And perhaps that would lead back to some sort of paradox.

In any case, I've changed my mind on this whole matter.

Professor DeLong proposed to have a system whereby Lewd paid Prude, and then Prude paid Lewd, each for the privilege of needlessly interfering in the other's life. I thought that was better than my thought that they both could simply suck eggs, for wanting to be little Stalins.

But then Kevin Quinn pointed out that, if there were two Prudes and a Lewd, the Prudes might try to wait for the other to make a deal, to get a free ride (gain without pain).

This seems logical, and causes my head to hurt. So I'm back to where I was before I read Dr. DeLong's solution: while all the Prudes and Lewds are debating whether to engage in free trade, or use of government, to interfere in each others' lives, I will drop a JDAM (or MOAB, or MPHTW*) on them all. And the world will be a better place. Just me and the darn fruit flies that seem to have invaded my house.

*MPHTW=Monty Python Hundred Ton Weight.

Posted by: Mark Bahner on May 2, 2003 02:51 PM

One more contribution to this very interesting thread.

The key question here is whether you believe, as Brad claims to, that the right to trade is prior to and subsumes all other rights – whether the right to read Lady Chatterly’s lover (to stick with the original example) has the same status as ownership of a given copy of Lady Chatterly’s lover. If you believe this, no paradox. But liberalism, if it means anything, means that there are other rights which are equally fundamental – rights of conscience, of free speech and assembly, to vote, and so on. None of these rights are subsumed other any of the others. To a liberal, the right to make consensual trades doesn’t mean we can trade away these other rights, just as the right to vote doesn’t mean we can vote away the right to free speech, or to make consensual trades. That’s why we have a Constitution, and why we talk about illiberal democracies – which in fact are the starting point of Sen’s essay.

In a way, this is obvious. Sen’s contribution was just to formalize it.

Clearly, the right to healthcare or housing, say, means something more than that they’re available in the market. Now I don’t know if Brad believes in those rights. I’m sure that most of his readers don’t. But what about genuinely liberal rights, like voting?

Should votes should be freely tradable in the market? Such trades would be Pareto-improving by definition, but I suspect (and hope!) that Brad’s against them anyway. But in that case he’s acknowledging the force of the paradox, even if he doesn’t like Sen’s presentation.

Voting is a clearer example, since it doesn’t involve any problems of enforcement – just strike the seller’s name from the rolls, and add one vote to the buyer’s. And it doesn’t require any special assumptions about preferences. It’s helpful, too, that we can imagine what the outcome of such a system would be: all the votes soon gathered in a few hands, or at least distributed proportionately to private wealth. This doesn’t matter for the logic of the paradox. But it does matter for its practical relevance – it means there is an important difference between the liberal and Paretian outcomes.

It doesn’t diminish the paradox that the contradiction is sometimes resolved in favor of Pareto. As Mark Bahmer notes, the workplace is not a liberal regime. In my mind, that only highlights the distinction between genuine liberalism and Brad’s “solution” to the paradox in the original post.


Posted by: jw mason on May 2, 2003 06:07 PM

Mark-

I know you're joking, but you've got a valid point. As I noted earlier, Sen himself suggests that one solution is for people not to have such "intrusive" preferences. That's one of the reasons I brought up voting -- to show that that doesn't make the problem go away.

"Give that man a Nobel prize! ;-) "

You know he's got one, right?

Posted by: jw mason on May 2, 2003 06:38 PM

At the risk of seeming obtuse (and perhaps flogging a dead horse), is there actually a principled basis for the claim that society is a fictional construct, but individuals are real?

I mean, for instance, if we're conceiving of individuals as persons capable of making meaningful remarks, rather than, say, multicellular colonies, or conglomerations of space-time points.

Posted by: john beer on May 2, 2003 08:54 PM

"As Mark Bahner notes, the workplace is not a liberal regime."

Whoa! Wait a second there, pardner! ;-)

*I* was saying that my signing a contract to not read what might be first on my list of interest, for 8 hours a day, in exchange for a day's pay, IS a liberal trade!

There is no coercion between my employer and me. We both, completely of our free will, sign an agreement, whereby I only get paid if I work on the things they want me to work on. (Which, happily, are generally interesting things...even if they aren't quite as intersting as the things I do in my free time.)

And as for my right to vote, I'd very happily sell my right to vote in federal elections for even a 5% reduction in my federal tax bill.

Posted by: Mark Bahner on May 2, 2003 09:57 PM

Another Paretian/Liberal paradox. Tom suffers a loss of 10 dollars when humiliated. The rest of us gain a total of $30 from the pleasure of humiliating him. A la Coase and Brad, value maximization requires that we contract with Tom, offering him more than 10 and less than 30 if he will let us humiliate him, a trade that will be mutually beneficial. If transactions costs are too high, Coase teaches, then where we put the rights matters for efficiency. In this case, the public nature of Tom's humiliation (in the economist's sense) creates perhaps insuperable transactions costs when it is a question of our buying Tom's right not to be humiliated, as each of us hopes to free ride, seeing Tom humiliated without paying for it. Efficiency would then be better served by taking away Tom's right not to be humiliated (suppose the species of humiliation involves what we would otherwise call criminal acts: de-criminalize them!). Alternatively, as is frequently done with public goods, the State might tax the humiliators something between $10 and $30 and transfer the money to poor Tom.

The point of course is that whether this problem is solved privately in Coasian style or by the State, the solution to the Paretian's problem is a disaster for the Liberal. Liberalism, Kant style, would be appalled at the solution. At the basis of liberalism is autonomy. Kant's ethics prohibit us from treating human beings, including *ourselves*, as mere vehicles for welfare; we must treat ourselves and others with dignity. As in this example, this can lead to conflicts between liberalism and value-maximization. A liberal state in the best case would see one of its roles as preventing the conditions that would make trades such as the one Tom is contemplating desirable, by raising his income and lowering the his marginal utility of money; or failing that, it would ban such trades.

Posted by: kevin quinn on May 4, 2003 12:59 PM
Post a comment