Slashdot Mirror
Contracts in Cyberspace
phutureboy writes "In his online Journal of Interesting Economics, economist David D. Friedman (son of Nobel laureate Milton Friedman) presents some interesting ideas about the enforcement of contracts in cyberspace. The gist is that he sees a gradual shift away from public enforcement of contracts (i.e. government courts) to private enforcement (e.g. third party arbitration, reputationial enforcement). The rest of his site is interesting as well - he even has an archive of his open-source economics software, which includes a neat trading simulation game called Hansa."
There are also questions of jurisdiction. If you buy something from a company in Mexico, which is a subsidiary of a US company, but they outsource their credit collection to a group in Canada, and you are a resident of the U.K., which country's courts does the company use to seek redress against you if you fail to live up to your contract?
---
When you come to a fork in the road, take it! --Yogi Berra--
The trend toward third-party arbitration has been happening for a while. The arbitration process, as I have read, is cheaper for both parties and quicker to resolve. So, I'm not certain why this is seen as a novel trend.
On the other hand, the development of contract assurance techniques for dealings with anonymous parties for open purposes is rather novel. Ordering a pallet of widgets from digital signature 7YkkeL33Tphubar6 is a little more shaky than ordering them from ACME Co.'s.
--- "It annoyed me, so I fixed it." -- Tom's First Principle of Engineering
Whoa, Friedman made Slashdot. Not bad.
:-)
Dr. Friedman is quite a character -- I was lucky enough to chat with him a couple times; he teaches at Santa Clara University and by some peculiar twist of fate I now possess a degree from there. So we ended up going to the same talks every once in a while -- quirky guy, occasionally reminds you of the late 90's when people really acted like the Internet Was Going To Change Humanity or something.
Friedman's paper is overall pretty reasonable, but his calculations seem to ignore the *tremendous* nonlinearity in our responsiveness to bad news. Ten people can tell you a restaurant is good, but if one says it's bad, you're probably not going to eat there. Paypal can pass a million good transactions a year, and it only takes a couple thousand questionable ones to really make a visible impact on their quality of service. Simply having one's reputation questioned tarnishes it -- indeed, one reason so many cases settle, or go to binding arbitration, is to keep major conflicts quiet.
It's in this context that arbitration servers have a problem. If they downgrade reputation as humans do, those who are downgraded may complain -- with apparent statistical cause -- that their otherwise good service is being mucked up by the inevitable screwups. But if they *don't*, their data is quite useless relative to the weighing the human mind does, and nobody would ever trust them.
Now, as an economist, Friedman would probably use this as an example of how humans are irrational...I doubt that. Consider the nonlinearity a form of damage multiplication...one transaction may be tweaked to make more money, but this will impact a hundred other transactions, that will thus cause a net loss of money. This means nobody can be screwed over -- everyone must get fair service. Without nonlinearity, it's always worthwhile to screw 1-5% percent of your clientele.
The news have it right -- bad news is much more interesting, and it should be.
For what it's worth, I do suspect that a cross-jurisdictional system will spawn to handle global commerce, and I think it'll be a combination of Friedman and Visa. Anonymity in financial transactions is pretty doomed -- we've just gotten accustomed to handing over our wallet, and hoping we get it all back -- but my expectation is that person-to-person cash transfer over distance will be formalized, arbitrated, and most of all -- insured. There will be reputational work, but it'll be in the same context we already have institutionalized reputations -- credit checks. It's *much* more likely that we'll see ebay feedback on your Visa web page than we'll see some funky distributed arbitration notices system.
Visa, you see, has one major advantage: The arbitration systems that Friedman describes are great for stopping the *next* fraud, but they don't do anything about *this* one. This is the big deal about government, folks -- they may not prevent crime directly, but they sure as hell respond to it. It's not like there's a murder, and the cops are like, wow, we better prevent any gun stores from letting this guy in, but that's all we can do
Yours Truly,
Dan Kaminsky
DoxPara Research
http://www.doxpara.com
I have to say that I like the idea of this a lot. Laws seem either too big, and stop legitimite services, or are too little, leaving some hole through which the con jobs may continue. It seems that reputation protects consumers from unethical behavior more than simply illegal behavior, which is probably better for the consumer. I am still unsure how you could keep track of the validity of positive/negative comments on a person's reputation. If reputation really grows -that- important, then a kiddie could ruin a legitimate business. At first, I would think you could just give each person a second reputation, to determine the legitimacy of their comments, but that's really more complicated than it seems. I mean, how would you really cope with say, a worm that spreads and tries to post complaints about your company?
On the one hand, business gets separate, cooperative environments for dirt-cheap labor, pollution, safety issues, and so forth. On the other hand, the increasingly confusing legal issues between jurisdictions make it difficult to estimate liability. A balkanized world would provide better business opportunities if other people's lawyers weren't so clever. Oh, what to do, what to do?
We can debate this in terms of political philosophy, if we want to sound like those peasants by the ditch in that Monty Python movie. But the ultimate fate of the world's separate governments is a function of their utility to people who have enough money to force their decisions on politicians. My guess is that eventually we will have a single world government, but it will happen when the bigger corps decide that it's better for them.
"Information Wants to be Free!" is the mantra of Ken MacLeod's SF novel Stone Canal, which may sound like a cliche to Slashdot readers. However this and it sequels are actually thoughtful, engaging looks at how a future society based on true anarchic principals would work.
Prominently featured are the notions of private courts and private enforcements of contracts. A respected judge arbitrates a violent feud between the leader of the colony and a would-be-rebel. The extreme case is the privitization of nuclear deterrence on Earth, when Kazakhistan sells "rights" to left over USSR nuclear arms, permitting countries (even welathy private individuals) to enjoy the power of threatened mutually assured destruction.
One problem I see with all this reputation-based stuff - how does it prevent attacks like this: http://zdnet.com.com/2100-11-530012.html?legacy=zd nn
where someone builds up a good rating through legit deals, then "turns to the dark side" and suddenly commits a big fraud?
Do you limit the size of trade that can be made to be less than the accumulated 'value' (by some measure) of the existing reputation?
I am a big free market fan, but free market only works well with enforcement of agreements made in the market. Concerns I would have in this case:
1) Would the enforcement be impartial? Private enforcement tends to be prone to vigilante behavior.
2) Would the enforcement be legal? In the various countries, there are laws of libel, monopoly power, etc. that could raise some concerns.
3) Would the enforcement be respected? What power could it really yield without violating the first two issues?
4)Would the enforcement be accountable for its actions? How can the community concerned be sure to have the right people making the important decisions?
All in all, these are the reasons ours and other legal systems do a great job of fostering a relatively free market. Private enforcement tends to fall prey to this and other problems. Just look at all of the recent issues with ICANN for an example of some of these problems.
Boom Shanka
Mr. Friedman makes a number of good points in his case, but he overlooks one major requirement for a reputation: a fixed public identity. In the case of the diamond dealers cited in the article, each dealer was known throughout the community, if not to every individual, at least to every dealer's uncle's best friend, etc. Getting caught doing something dishonest brought down the censure of the entire community: extended family, friends, congregation members & leaders, neighbors. Other dealers, especially would be likely to boycott and direct customers away from an unethical persont to protect the entire industry.
In the case of the internet, there is no single "community." The anonymous nature of the media makes pinning a particular crime to a particular individual much more difficult. Use of aliases and private identities allows people to engage in virtual activities they would not consider in real life. For example, many pedophiles have managed to amuse themselves without families or spouses being aware of nature of their online activities until the police showed up. Your neighbor may regularly rip people off on eBay, but if you don't shop there or simply don't know his eBay identity, then you'll never know. Were he caught shoplifting down the street, he'd be likely to move out of embarassment, but on the internet he can safely defraud people hundreds of kilometers away.
No, if anything *more* legal survelliance and action is needed to improve the cybereconomy, not less. Until there is as great a certainty that online crimes will be caught and punished to the same degree as their Real Life counterparts online business will not have the credability (much less the overall profitability) of brick-and-mortar stores.