I apologize Gil. I obviously have not been clear. I may have been using technobabble.

With your permission, perhaps I can try again.

I will get back to the pony in a minute.

Imagine that each of 100 people has 100 dollars that he can anonymously place in an envelope and place in a container. Everyone pools their money together and no one knows who gives what to the total, but each can give 100 dollars or any portion of 100 dollars.

The rules of the game are as follows. 10% of the TOTAL contribution of all 100 people is added to the total by an anonymous source. So if all 100 people give 100 dollars, that would make $10000.00 and the anonymous source then adds 10% of the total or 1000 dollars. So if everyone gives 100 dollars, there is now $11,000 in the container, once the 10% is added.

Now, the total is divided by all 100 people, so $11,000 divided by 100 people is $110 dollars per person. Everyone is happy and makes 10 dollars, since each has $110 dollars, and started with $100.

But let's allow some people to be selfish. So if no one knows who contributes what, one person might think, "Hmm, if I don't donate my 100 dollars then the total of everyone else will be $9,900 plus 10% equals $10,890. If that is divided by 100 people, everyone will receive ($10,890 divided by 100 which is) $108.90, not quite as much as 110 dollars, but close. But since I never donated the 100 dollars, I will have 100 dollars plus $108.90 which is $208.90, and that total is greater than 110 dollars, the amount I would have received had I contributed. I think I won't contribute and more than double my money."

Clearly if someone is selfish, he gets the benefit of everyone pooling their money (the extra 10% plus a portion of everyone elses contribution), but none of the cost (the contribution of 100 dollars). But of course the situation is symmetrical for everyone, so if everyone is selfish, no one contributes $100 dollars and the benefit (the extra 10% added to the total) is lost to EVERYONE.

So in this "game", the voluntary cooperation of people (if all are self-interested), will not lead to the optimal outcome. No one will contribute much of anything (or less will be contributed), and so the benefit of the extra 10% is lost to everyone. Voluntary cooperation will not lead to the optimal outcome. By the way, in economics, the people who don't contribute but expect the benefits anyway, are called "free riders".

But let's say there is a smart person in the group who says, "You know, we could get an extra 10% if we had some way of guaranteeing that everyone will pay, or excluding the people who don't contribute from enjoying the benefit."

But let's say there is no way of "excluding" the people who don't contribute from the benefit.

A smart person might continue, "if we hire a 'police-man', say for 10 dollars and he makes sure (able to use coercion) that everyone pays their 100 dollars, then at the end of the night we will have $10,000 plus 10% is $11,000, minus 10 dollars for the policeman, leaves $10,990 dollars. This, divided by 100 leaves $109.90, not quite the $110 dollars and 10 dollar profit, but still a profit of $9.90.

So the smart person asks everyone whether he would like to take a vote to decide what to do. Should everyone voluntarily cooperate without the policeman, and pool his contributions, or should the group hire a policeman and pool the money that way?

One can see that if this were a real world scenario, it is very likely that 100% of people in the group would agree *TO VOTE TO BE COERCED*. That is, 100% would vote to hire a policeman, provided the policeman did his job.

Of course the problem is always the "who polices the policeman?" situation. Could not the policeman abscond with everyones money? Obviously we need "checks and balances", to try to hold policemen accountable, too, and these checks and balances will never be perfect.

By the way, the outcome of hiring a policeman could be considered an entirely libertarian outcome. (Why?) Because 100% of everyone votes for the outcome. Everyone agrees to be coerced, provided that his neighbor is equally coerced, to create an outcome that 100% of everyone favors. It is using coercion to allow free citizens the option of *unanimous rule*, the essence of libertarianism. So paradoxically, in this hypothetical example, libertarians NEED A COERCIVE force (like the government) to realize their highest ideals.

Back to the Pony. Let's imagine that it is a "pony of defense" that enables our 100 villagers to protect their farmlands from theft, when ridden by a police officer. After much debate, all 100% of the villagers agree that it would be worth 100 dollars for each to pool their money and buy the needed one pony of defense, riden by a policeman, per year. They come to this rational conclusion, because after arguing back and forth for a while, they all believe that their best theory (they all read "The Fabric...!") argues that more than 110 dollars is stolen from each farm every year, so it is worth pooling their money to buy the Pony.

Gil, do you see the analogy to the situation previously described when people were pooling their money? If everyone decides "voluntarily" to pool their money, the fact that there is a policeman on a Pony, protects everyones farm from attack by marauding evil-doers. So should a selfish individual contribute? If one person does not contribute, and if everyone else does, everyone still has a pony and a policeman protecting everyones farm, but a given individual who does not contribute, has the pony and the policeman and gets to keep his 100 dollars. (Well, because one 100 dollars contribution is missing, perhaps the policeman and pony take a day or two off per year). Since the situation is symmetrical for everyone, self-interested individuals do not contribute and "free ride" on everyone elses contributions. So everyone loses. But if the group hires the policeman not only to protect their property, but also to collect taxes (err...collect the contributions), if everyone shares the theory that more than 110 dollars will be stolen per farm unless each contributes 100 dollars, there will be a vote in which 100% of everyone (unanimous rule) prevails. A perfect libertarian solution (unanimous rule) requires the coercion of a government (err...voluntarily hired policeman).

Now, enter the real world. a. Policemen steal, too, and they have guns! Obviously, in deciding to vote for the policeman, the townspeople will need to take into account their best theories about the amount that a policeman might steal.

But another problem is: What if there is 1 person in the group who does not agree with the "best theory" of the other 99. In fact his theory is that "defense ponies" attract "evil-doers" and that if a pony and policeman are hired, the town will surely lose $1,000,000, the entirety of the town.

So when the vote is taken, 99 say hire the "defense pony and policeman" and 1 says don't hire the defense pony and policeman.

So what's a good unanimous rule libertarian to do? There are several options. The 99 could agree to hire the defense-pony by themselves, and ignore the 1 person who disagrees, and at least don't ask him to pay 100 dollars, even if he thinks he's going to lose his entire $10,000 farm if the others get the pony.

But if the 99 know before-hand that someone who votes against the defense-pony, won't have to pay for it, (even if this hypothetical selfish person secretly wants a defense pony), an actual selfish person amongst the 99 could say, "I ought to pretend to not want the defense-pony."

Then when the vote is taken, 98 will vote for the defense pony, and the person who does not want the defense pony will vote against, as will the selfish man who wants the defense pony but would like others to pay for it for him.

But the situation is symmetrical for everyone, so if you can opt out of paying by voting against, all selfish people will vote against, and pretend to have the same viewpoint as the person who actually believes the "defense-pony" will be bad for the village. So if the majority in a vote does not have the ability to coerce the minority in a vote to also pay, all selfish people will lie, to get their neighbor who votes for something to pay instead. (Their neighbor being the one who "foolishly" is honest, votes for what he actually thinks and ends up being the only one contributing to the global good of the community by contributing to the "defense pony.") So, people wisely agree that in the case of "defense-ponys", they will not allow the minority to prevent them from increasing their efficieny.

Note that in my first example, where 100 people pool their $100 to get an extra 10% on the total collected, if a person can "opt out" of contributing by voting against, yet he still gets 1/100th of the total, a selfish person will "opt out" and vote against pooling money. But this is true for all selfish people in the group. So virtually everyone will vote against pooling their money, and the group forfeits its extra 10%! If a minority can prevent the majority from pooling its money, the minority is coercing the majority into forfeiting its extra 10%, and this is wrong.

So one of the reasons that there is no "opt-out" clauses in votes on defense, for example, is that it would prevent *the majority* from coercing itself into a contribution that each and every person in the majority wants.

Yes the majority coerces the minority in a democratic vote. But if there is a "good" that is produced (like defense or the reconstruction of New Orleans) in which, regardless of someones contribution, everyone gets to enjoy the benefits: If the minority can "opt out" of paying taxes, then a group of individuals pursuing their "rational self-interest," who actually want Defense or New Orleans reconstructed, will also "opt-out" and vote against it, even if they want it. Why? Because if everyone else votes for it, they won't have to pay taxes because they "opted-out" and they still get New Orleans reconstructed or proper defense, with no contribution. Since this situation is symmetrical for everyone, *allowing the minority to "opt-out" of paying taxes, coerces the majority into not being able to coerce itself, into an outcome that each one in the majority wants.*

This is why we do not allow a minority to "opt-out" of paying taxes when they disagree with the majority. Obviously, part of the reason is because (as Gil suggests), the majority wants to take money from the minority, but part of the reason is to allow efficient collective action of the majority, in the production of a good from which others cannot be easily excluded (like national defense or the reconstruction of New Orleans.)