Voluntary Exchange: A Prisoner's Dilemma

[Borodog]

Consider a voluntary exchange. Praxeological analysis tells us that the two parties to the exchange find themselves in a double coincidence of wants, i.e. each has a good or service that the other wants. Furthermore, because the value scales of each party are subjective, their ordinal ranking of the value of the two goods are subjective. If each values the other's good more highly than their own, an exchange will occur, and the subjective state of each party's "satisfaction" ("happiness" whatever you want to call it) increases (or his state of "unease" decreases). The exchange takes place because each party believes before the exchange that he will benefit after the exchange. Either party (or both) may discover subsequent to the exchange that they have made a mistake, and that their subjective level of satisfaction has actually decreased, or their unease has actually increased. A particularly obvious case of this is where one party cheats the other somehow; they pay for their good with counterfeit money, for example, or the car they sold turns out to be a lemon, or they take payment but do not render a promised service. We could generally classify this as "fraud."

It becomes clear that any voluntary exchange is in fact, a Prisoner's Dilemma. In the Prisoner's Dilemma, a two-player game, each "player" (party to the exchange) has two strategies, to play fair (to ooperate, in the parlance of the PD), or to cheat (to defect). Since each player has two possible moves, there are 4 possible outcomes of the game. Consider the Payouts for Player 1 and 2 for each strategy.

The Reward For Fair Dealing is that each player sees his state of satisfaction increase (we shall neglect for now the case where an exchanger made a mistake, and is disatisfied subsequent to the exchange through no fault of the other party).

The Punishment for Double Cheating is that neither party has their state of satisfaction increase, plus they've wasted their time, energy, and possible a small amount of resources.

The Temptation to Cheat is that if you cheat, you gain the satisfaction granted by the thing you wanted, but don't have to give up the satisfaction to be gained from the thing you would have had to exchange. Clearly, this is higher than the satisfaction of gaining the thing you want, but giving up the thing you have.

The Sucker's Payoff is that if you give up your good but get cheated, you not only don't gain the satisfaction of the thing you wanted, but you also lose the satisfaction of the thing you had.

Quite clearly, this meets all the criteria of the Prisoner's Dilemma.

Any game theorist knows that the proper strategy for the PD is to Always Defect (Cheat, here). This is because no matter what your opponent plays, your best move is to cheat. Under this analysis, we might wonder why civilization, cooperation, the division of labor and exchange, exists at all, since man really should be mankind's wolf.

However, consider a small community of men where each man can expect to interact with another man not once, but an indefinite number of times. We see then that each exchange is not a single Prisoner's Dilemma, but rather a single instance within an Iterated Prisoner's Dilemma. Without getting into a detailed analysis, in the IPD, the correct strategy is NOT Always Cheat; this is infact the worst possible strategy. The correct strategy is a Nice, Retaliatory, but Forgiving strategy. A "Nice" strategy is a strategy that is never the first to cheat, in other words a trusting strategy. A "Retaliatory" strategy is one that punishes being cheated by retaliatory cheating in future games. A "Forgiving" strategy is one that ceases punishing a former cheater once they demonstrate "remorse", i.e. stop cheating. A strategy like this is "Tit For Tat", where a player starts off by Playing Fair, and thereafter simply repeats the other player's last move. You should convince yourself that this is a Nice, Retaliatory, Forgiving strategy.

Hence, we can see that economic life, production and exchange, for the members of the small group is a set of IPDs, and the optimal strategy is for everyone to play NRF strategies. This is a "stable" strategy in the sense that NRF strategies do well in populations dominated by other NRF strategists. A populations dominated by cheatscan be stable for the same reason; a Sucker will do poorly in a population of Cheats; the best option is to be a Cheat yourself.

However, NRF communities have a "metastability" that Cheater communities do not. Because NRF communities enjoy an absolute advantage over Cheater communities, small NRF groups that arise can spread at the expense of Cheater communities, while the reverse is not true. Dawkins goes into this in more detail in The Selfish Gene.

So players who play NRF strategies in their economic lives will prosper relative to those who play Nasty (first to cheat) strategies, Wimpy (non-retaliatory), or Grudge-Holding (unforgiving) strategies.

This neatly explains the human condition. It is perfectly inline with the Austrian analysis, which stems from the fact that people prefer more goods to less, and hence will engage in specialization, division of labor and exchange, even if they hate the people they are exchanging with, is also right in line with Adam Smith's "propensity to truck and barter" (which I think is both genetic and cultural).

In short, in general man is not usually other men's wolf because it doesn't pay to be one. This is of course not to say that cheaters don't exist; they do. A more detailed analysis would show that an NRF community can be "invaded" by a mixture of other types of strategies, particularly some mix of Wimps and Cheaters, but their proportions will stay relatively small compared to the NRFs.

Also, consider the Con Man. What does the Con Man do? He arranges it so that his exchanges are not IPDs but rather single PDs. He does this by disappering and moving to another community after Cheating some Sucker.

Lastly, consider the fact that because people can actually talk to each other, the IPD effect is even stronger. We don't have to wait for a Cheat to attempt an exchange with the person he last Cheated for Retaliation to occur; the Sucker can tell the rest of the community he was Cheated, and others can Retaliate for him! In otherwords, the community wields the power of economic ostracism, a power which can be so terrible in small groups that the temptation to Cheat must be greatly restricted.

This kind of analysis decisively destroys the Hobbesian justification of the state, homo homini lupus est is simply false.

To my knowledge, this application of the PD and IPD to economic exchange is original. However, it seems so incredibly obvious that someone must have thought of it. If anyone knows of a source where this sort of application is discussed, I would greatly appreciate it.

[moorobot]

Maybe this destroys the vulgar "Hobbesian" justification for the state, but it doesn't really have anything to do with Hobbes actual viewpoint, which in no way is based on "[censored] homini lupus est".

Next, your analysis doesn't apply to large anonymous groupings of people with high degrees of geographically mobility, which you appear to realize. Under modern conditions, cheating is a far more effective of a strategy than it was in small groups in which everyone knew one another.

Also, your analysis doesn't apply to intangible things e.g. religous or political differences/conflicts which are usually by nature zero/negative sum struggles where one's loss is another's gain (e.g. conflicts over abortion, drug use, school policy) etc; it applies only to cooperating for matieral goods and services.

[Skidoo]

It also tends to ignore the sort of transaction where both parties voluntarily sacrifice for an emergent gain that neither would be able to realize individually.

[WillMagic]

Um...this is pretty [censored] awesome. Nice post, Boro.

[thylacine]

I am purely speculating, and I have no references, but my guess would be that the number of published papers with these kinds of ideas would number in the thousands.

[Borodog]

[ QUOTE ]
I am purely speculating, and I have no references, but my guess would be that the number of published papers with these kinds of ideas would number in the thousands.

[/ QUOTE ]

I would think so to. It's just too obvious not to be. But I asked a Ph.D. economist friend of mine, and didn't know off the top of his head anyone that had treated voluntary exchange specifically as an IPD. Also, PD and IPD type analyses are common in "sociobiology", but I don't know of any analysis that talks specifically about human specialization, the division of labor, and exchange (i.e. "real" economics, rather than "evolutionary" economics) in terms of the IPD.

[Borodog]

[ QUOTE ]
It also tends to ignore the sort of transaction where both parties voluntarily sacrifice for an emergent gain that neither would be able to realize individually.

[/ QUOTE ]

No, it doesn't. Do you see why?

[hmkpoker]

There is one huge difference between the PD and voluntary exchange: in the PD, communication is deliberately restricted. If both prisoners were allowed to communicate, as is the case in any significant voluntary exchange, they would both serve a year and get out.

[Borodog]

[ QUOTE ]
Maybe this destroys the vulgar "Hobbesian" justification for the state, but it doesn't really have anything to do with Hobbes actual viewpoint, which in no way is based on "[censored] homini lupus est".

[/ QUOTE ]

Yes, it is. The Hobbesian justification is essentially that two parties A and B in a "state of nature" (anarchy) vis a vis each other will be at each other's throats (defect-defect), and hence a third party S (the state) is required to prevent and settle disputes. This of course is farcical, since the last time I checked S has to be comprised of human beings! Even if we concede the Hobbesian position, a state is then worse than no state; S might indeed appear to act to prevent and settle disputes between A and B, but only so that they are the fatter for S to plunder and exploit.

[ QUOTE ]
Next, your analysis doesn't apply to large anonymous groupings of people with high degrees of geographically mobility, which you appear to realize. Under modern conditions, cheating is a far more effective of a strategy than it was in small groups in which everyone knew one another.

[/ QUOTE ]

I've already anticipated this in earlier posts. It is precisely where repeat interactions are the least likely that people stray from Nice strategies. Examples are the rudeness of New Yorkers and drivers on the highway.

However, technology has a remarkable way of "shrinking the world". For example credit reports, consumer advocacy publications, product and service reviews and guides, the media, etc. act to shrink the effective community by increasing the coverage of information about "cheating" individuals and businesses.

[ QUOTE ]
Also, your analysis doesn't apply to intangible things e.g. religous or political differences/conflicts which are usually by nature zero/negative sum struggles where one's loss is another's gain (e.g. conflicts over abortion, drug use, school policy) etc; it applies only to cooperating for matieral goods and services.

[/ QUOTE ]

I didn't claim that it did.

[Borodog]

[ QUOTE ]
There is one huge difference between the PD and voluntary exchange: in the PD, communication is deliberately restricted. If both prisoners were allowed to communicate, as is the case in any significant voluntary exchange, they would both serve a year and get out.

[/ QUOTE ]

Not really. Do you see why?

[Dan.]

[ QUOTE ]
[ QUOTE ]
There is one huge difference between the PD and voluntary exchange: in the PD, communication is deliberately restricted. If both prisoners were allowed to communicate, as is the case in any significant voluntary exchange, they would both serve a year and get out.

[/ QUOTE ]

Not really. Do you see why?

[/ QUOTE ]

Hmk offers a very legitimate criticism of this work and you offer no explaination to why his claim is unfounded. That makes you an [censored], dysw? BTW Hmk, I think point is excellent.

[Zygote]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
There is one huge difference between the PD and voluntary exchange: in the PD, communication is deliberately restricted. If both prisoners were allowed to communicate, as is the case in any significant voluntary exchange, they would both serve a year and get out.

[/ QUOTE ]

Not really. Do you see why?

[/ QUOTE ]

Hmk offers a very legitimate criticism of this work and you offer no explaination to why his claim is unfounded. That makes you an [censored], dysw? BTW Hmk, I think point is excellent.

[/ QUOTE ]


borodog has made a point. He saying that even though they can communicate, neither party can ever be sure that the communications are genuine.

This leaves you in the same case as the PD, since you can never be sure the other party isn't lying. Regardless to what a person communicates, you can never be sure as to what they will do, so you must make a foundational choice of strategy beyond this communication.

[Borodog]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
There is one huge difference between the PD and voluntary exchange: in the PD, communication is deliberately restricted. If both prisoners were allowed to communicate, as is the case in any significant voluntary exchange, they would both serve a year and get out.

[/ QUOTE ]

Not really. Do you see why?

[/ QUOTE ]

Hmk offers a very legitimate criticism of this work and you offer no explaination to why his claim is unfounded. That makes you an [censored], dysw? BTW Hmk, I think point is excellent.

[/ QUOTE ]

hmk is a good friend of mine; I was [censored] with him.

The reason that his point doesn't change anything is the same reason that communicating in the Prisoner's Deliemma doesn't actually change anything. Prisoners A and B get together and agree to cooperate. What then is the proper strategy for either player once this agreement has been made?

Edit: Or, what Zygote said.

[Speedlimits]

[ QUOTE ]
[ QUOTE ]
It also tends to ignore the sort of transaction where both parties voluntarily sacrifice for an emergent gain that neither would be able to realize individually.

[/ QUOTE ]

No, it doesn't. Do you see why?

[/ QUOTE ]

LoL@ the Sklanskian responses.

The NFR strategy is incredibly effective. I actually use this strategy daily but never defined it as such. Good post.

[DougShrapnel]

Boro very good post. Pretty much the way I view society. Could you explain who you get to these.

[ QUOTE ]
We don't have to wait for a Cheat to attempt an exchange with the person he last Cheated for Retaliation to occur; the Sucker can tell the rest of the community he was Cheated, and others can Retaliate for him! In otherwords, the community wields the power of economic ostracism, a power which can be so terrible in small groups that the temptation to Cheat must be greatly restricted.

[/ QUOTE ] The rest of the community's influence is named government. It's a good idea to call it that becuase it gives the power of ostracsim(e.g justice) visiablity. Your post may point to specific changes in the use of government and the rights govenrment has. However, it is an excellent justification for a state. Albiet not the state we have currently but a state non the less.

[Borodog]

[ QUOTE ]
Boro very good post. Pretty much the way I view society. Could you explain who you get to these.

[ QUOTE ]
We don't have to wait for a Cheat to attempt an exchange with the person he last Cheated for Retaliation to occur; the Sucker can tell the rest of the community he was Cheated, and others can Retaliate for him! In otherwords, the community wields the power of economic ostracism, a power which can be so terrible in small groups that the temptation to Cheat must be greatly restricted.

[/ QUOTE ] The rest of the community's influence is named government.

[/ QUOTE ]

No, it isn't. It's the individual choices made by the individuals in the community within the context of their own interactions with the Cheater.

[DougShrapnel]

[ QUOTE ]
[ QUOTE ]
Boro very good post. Pretty much the way I view society. Could you explain who you get to these.

[ QUOTE ]
We don't have to wait for a Cheat to attempt an exchange with the person he last Cheated for Retaliation to occur; the Sucker can tell the rest of the community he was Cheated, and others can Retaliate for him! In otherwords, the community wields the power of economic ostracism, a power which can be so terrible in small groups that the temptation to Cheat must be greatly restricted.

[/ QUOTE ] The rest of the community's influence is named government.

[/ QUOTE ]

No, it isn't. It's the individual choices made by the individuals in the community within the context of their own interactions with the Cheater.

[/ QUOTE ]I think you should at least reconsider grouping those individual choices together. And giving it a name. A company is not just individuals exchanging capital, labor, profits, and risks. You'd consider me ridiculous if I forced you to call a company a group of individulas exchanging capital, labor, profits, and risks. Although it's a partial correct view, it's much simplier an also correct to give it a name. Company, Government.

[Borodog]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
Boro very good post. Pretty much the way I view society. Could you explain who you get to these.

[ QUOTE ]
We don't have to wait for a Cheat to attempt an exchange with the person he last Cheated for Retaliation to occur; the Sucker can tell the rest of the community he was Cheated, and others can Retaliate for him! In otherwords, the community wields the power of economic ostracism, a power which can be so terrible in small groups that the temptation to Cheat must be greatly restricted.

[/ QUOTE ] The rest of the community's influence is named government.

[/ QUOTE ]

No, it isn't. It's the individual choices made by the individuals in the community within the context of their own interactions with the Cheater.

[/ QUOTE ]I think you should at least reconsider grouping those individual choices together. And giving it a name. A company is not just individuals exchanging capital, labor, profits, and risks. You'd consider me ridiculous if I forced you to call a company a group of individulas exchanging capital, labor, profits, and risks. Although it's a partial correct view, it's much simplier an also correct to give it a name. Company, Government.

[/ QUOTE ]

Giving things names is only useful when it's useful, and not obfuscating. Me not taking my car to Bob's Auto Body Shop because my neighbor told me Bob ripped him off is not "government."

[DougShrapnel]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
Boro very good post. Pretty much the way I view society. Could you explain who you get to these.

[ QUOTE ]
We don't have to wait for a Cheat to attempt an exchange with the person he last Cheated for Retaliation to occur; the Sucker can tell the rest of the community he was Cheated, and others can Retaliate for him! In otherwords, the community wields the power of economic ostracism, a power which can be so terrible in small groups that the temptation to Cheat must be greatly restricted.

[/ QUOTE ] The rest of the community's influence is named government.

[/ QUOTE ]

No, it isn't. It's the individual choices made by the individuals in the community within the context of their own interactions with the Cheater.

[/ QUOTE ]I think you should at least reconsider grouping those individual choices together. And giving it a name. A company is not just individuals exchanging capital, labor, profits, and risks. You'd consider me ridiculous if I forced you to call a company a group of individulas exchanging capital, labor, profits, and risks. Although it's a partial correct view, it's much simplier an also correct to give it a name. Company, Government.

[/ QUOTE ]

Giving things names is only useful when it's useful, and not obfuscating. Me not taking my car to Bob's Auto Body Shop because my neighbor told me Bob ripped him off is not "government."

[/ QUOTE ]Of course not Boro. It's the organization that makes it a government or a company. I needed to rebuild my 3 season room the other month. The wood was rotten. I employed the use of my neighbor. In turn I helped him install his network, and some light automobile mateniance. We aren't a company. However If these exchanges becomes highly orginazed and contractual, we name it a company. You and Bob aren't government, it's only when these agreements become orginized that we need the term government to fully account for this orginaztion of NRF. It's important to keep it's use limited as these types of orginaztions will always attemp to use the power they have for unjust purposes.

[Borodog]

It doesn't help matters to give voluntary organizations that already have names like "businesses", "partnerships", "companies", or "firms" the label of a coercive organization like "government", which should be rightly lumped with similar coercive organizations like "protection racket", "organized crime ring", or "gang."

[hmkpoker]

[ QUOTE ]
[ QUOTE ]
There is one huge difference between the PD and voluntary exchange: in the PD, communication is deliberately restricted. If both prisoners were allowed to communicate, as is the case in any significant voluntary exchange, they would both serve a year and get out.

[/ QUOTE ]

Not really. Do you see why?

[/ QUOTE ]

Borodog and I just had a very interesting coversation about this. Consider the classic PD, except now the prisoners are allowed to communicate. Both are still the same rational maximizers that we always assume in this situation. The prisoners are given one hour to talk their strategy out, then must make their decisions. What will the result be?

Both will defect, and serve seven years.

The reason is that neither prisoner has the capacity to retaliate. Think about it; no contracts are being signed, no lawyers involved, no one's going to go to jail for defrauding the other for defecting (that's actually the whole point); even if one player leaves the room having shaken hands to an agreement of mutual cooperation, that doesn't change the fact that cooperation is still the wrong play; he is choosing one year in prison over walking out that day.

Now if communication is strategically irrelevent in such a situation, one might ask why the PD is always deliberately constructed so as to exclude it. The reason is that, while the players in the PD are perfectly rational, the real people to whom the PD is being delivered are not. Human beings have evolved to instinctively apply behavior useful to the IPD, not the PD. Very different. The result is seemingly irrational behavior in situations unreflective of the real world like the ultimatum game. However, allowing this PD-irrational behavior to take place in the PD in the form of communication will more than likely result in increased cooperation, and two people both achieving the best overall solution, even though it is completely irrational!

Of course, time preference and stakes are very important variables to consider; one with high time preference (like a criminal) is less likely to cooperate than one with low time preference, and a situation with low stakes is more likely to see cooperation than one with less, but nevertheless, introducing communication must increase the incidence of cooperation, even if it is completely irrational to do so.

[Borodog]

I dig it.

[Borodog]

Although I do think part of the prohibition against "communication" in the PD and IPD is specifically so that one player cannot see his opponent actually making his game play, say by writing on a slip of paper ala the Dawkins video "Nice Guys Finish First", and adjusting his play accordingly (i.e. if you can simply see your opponent starting to draw a D, you would obviously choose D, even if you would otherwise play C).

[DougShrapnel]

[ QUOTE ]
It doesn't help matters to give voluntary organizations that already have names like "businesses", "partnerships", "companies", or "firms" the label of a coercive organization like "government", which should be rightly lumped with similar coercive organizations like "protection racket", "organized crime ring", or "gang."

[/ QUOTE ] Businesses, partnerships, companies and firms have a different goal then government. The goal is to profit. Protection rackets, Organized crime rings, and gangs have the same goal, these should be lumped together with businesses, sure often they are fraudulent. But's it's a fradualent business. A just governments goal has zero to do with profit.

[Skidoo]

The power of economic ostracism held by the community at large as exercised at the level of the individual consumer can be overwhelmed by the undue influence, economic or otherwise, wielded by unregulated concentrations of wealth.

[Borodog]

[ QUOTE ]
A just governments goal has zero to do with profit.

[/ QUOTE ]

And you think I want to live in Fantasyland!

[Borodog]

[ QUOTE ]
The power of economic ostracism held by the community at large as exercised at the level of the individual consumer can be overwhelmed by the undue influence, economic or otherwise, wielded by unregulated concentrations of wealth.

[/ QUOTE ]

Wealth accrues most to those who over time. in the parlance of the OP, continuously "treat fairly" with the most people.

[Skidoo]

Doesn't that assume perfect information?

[Borodog]

[ QUOTE ]
Doesn't that assume perfect information?

[/ QUOTE ]

No.

[DougShrapnel]

[ QUOTE ]
[ QUOTE ]
A just governments goal has zero to do with profit.

[/ QUOTE ]

And you think I want to live in Fantasyland!

[/ QUOTE ]Boro, so in your last you make the distiction between business on the up and up, vs fraudelent business. Claim that fraudulent business is not business but is catagorilized with government in EVILland. And then you answer back at me with a similiarly ridiculous catagorical separation in failing to recognize the distinction betwix government and corrupt officials. It's your right as an american citizen to do so, and more improtantly as a person to catagorize the world any way you please. But don't you see the bias in your demarcation? All bad business belongs with the likes of government. Some corrupt officials spoil the whole bunch. I could as easily say that the corrupt officials are really businessmen taking advangtage of bussiness oportunities with the help of human greed, and it's business that ruins all government. Therefore socialism is the only way. But I won't becuase it a ridiculous proposition. As is AC. The closest justifible position to AC is pure darwinism. Might makes right. Economic might, monetary might, technological might. What ever you can take you take. No laws, no government none of this "It's my proprety stuff..

Edit: Will Magic told me to chill.

[WillMagic]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
A just governments goal has zero to do with profit.

[/ QUOTE ]

And you think I want to live in Fantasyland!

[/ QUOTE ]Boro, so in your last you make the distiction between business on the up and up, vs fraudelent business. Claim that fraudulent business is not business but is catagorilized with government in EVILland. And then you answer back at me with a similiarly ridiculous catagorical separation in failing to recognize the distinction betwix government and corrupt officials. It's your right as an american citizen to do so, and more improtantly as a person to catagorize the world any way you please. But don't you see the bias in your demarcation? All bad business belongs with the likes of government. Some corrupt officials spoil the whole bunch. I could as easily say that the corrupt officials are really businessmen taking advangtage of bussiness oportunities with the help of human greed, and it's business that ruins all government. Therefore socialism is the only way. But I won't becuase it a ridiculous proposition. As is AC. The closest justifible position to AC is pure darwinism. Might makes right. Economic might, monetary might, technological might. What ever you can take you take. No laws, no government none of this pansy, It's my proprety whinny BS.

[/ QUOTE ]

Chill.

[DougShrapnel]

Boro, I've decided to stay out of your ACist threads. I will give you the last word, if you want it. Te only reason why I get so worked up about this kinda stuff, is becuase if there was a violent revolution, I would take up arms against this government. However, I hope that they get thier [censored] together and I won't have to. I'm off to have a drink, cheers bud.