ReputationCurrencies

Questions

Most questions culled from BayleShanks original post.

References

See Also

CategoryEconomics?

Discussion

There’s a lot of talk on the possible future usage of “reputation” as form of currency (for instance, in CharlieStross?’s book Accelerando) or even as a replacement for the whole economic system (for instance, in CoryDoctorow?’s book Down and Out in the Magic Kingdom).

It’s an interesting topic for discussion.

Exactly how would such a “reputation currency” or “reputation economy” work?

Is it really a good idea?

If it’s so great then why hasn’t it happened already? Is there some tech advance we are waiting for? Or it is that we haven’t figured out how to make a workable currency out of it yet? Or is everything ready, and it’s just that no particular reputation system has reached critical mass yet?

Who is working on this? What books should we read to learn about it? Which parts of traditional monetary economic theory, high finance, or the legal aspects of banking would members of this economy need to know?

OK here are my opinions.

Exactly how would such a “reputation currency” or “reputation economy” work?

Not sure, I’d like to discuss that on this page.

Is it really a good idea?

Not sure. But it’s worth trying. I think the reason people like it is that they see it as fulfilling the same economic function as MicroPayments?; a way to economically gain from each small contribution of intellectual property. With the added advantage that if it’s reputation, not money, being gained, then all intellectual content could be OpenContent, and yet authors could still pay their bills.

I’d sure like it if “doing a good job editing CommunityWiki” could somehow help to pay the bills 😊

If it’s so great then why hasn’t it happened already?

I think the main problem is that no one sees how it could work “for real”. But I’m a WorseIsBetter kind of guy; I think if we start trying stuff out, for real, that’s more efficient that waiting for a group of economic theorists to prove that it’s possible.

So I think what’s really need is for people to start trying stuff out; even though we don’t have the economic theory yet, I think that’ll come through tinkering, and the tinkering is waiting on critical mass. I’m not saying we need people to say, “I’m giving up money and quitting my job and hoping that this reputation network that I just joined pans out”. But just setting up reputation networks and assigning reputations in them.

Who is working on this? I don’t know of anyone who is. People are working on trust metrics, though. There is a TrustMetricsWiki.

What books should we read to learn about it?

I don’t know.

Which parts of traditional monetary economic theory, high finance, or the legal aspects of banking would members of this economy need to know?

I think we need to know a lot.

More details on how it should actually work:

Why haven’t social networking sites jumpstarted a reputation economy?

You go on those sites, and you see, “Bob is a friend of Jane; I think Jane is very cool”. But you don’t actually conclude, “OK, I can trust Bob very much; in fact, if Bob asks, I’d paint his house for free, trusting that he’d repay me later”. Why? Because most people on these networks are promiscuous in who they connect to. Jane has 100 friends. Even if she only has 4 friends, out of those, 2 of them may be casual acquaintences that she friended because they friended her.

In other words, the value of “friendedness” is low because “friendedness” is not scarce. Maybe if we could make it scarce, then it would become valuable enough to pay the bills.

How to do this? There are a bunch of possibilities:

Make friendedness more like real friendships Most people don’t really trust 100 other people to a very high degree. But their social network in the system doesn’t match their true social network. Maybe we could find a social means to make it match better.

Allow more gradations of friendedness Of course if you can only say “this person is my friend” or “i don’t know this person”, you will say everyone is your friend. But if friendedness was a number, you would be able to specify a middle ground and to reserve the highest numbers for close friends.

Unfortunately, I think that people still have every incentive to inflate their true estimation of others in public. Also, a system for ranking your friends is a good way to hurt everyone’s feelings.

Make friendedness less like real friendships

If you want a reputation economy to pay the bills for authors of OpenContent, you’d better have reputation points coming in from people who aren’t their friends, who wouldn’t even trust them with most personal matters, but who just respect something they’ve done. So you don’t really want to make the system mirror deep friendships.

In addition, if it’s clear that “reputation points” are NOT tied to personal friendship, people will feel more free to not give a high reputation to every acquaintence (“it’s nothing personal”), to trade reputation points for services, and to give reputation to people whom they don’t personally know.

Create different namespaces for different types of reputation If there is just one kind of “reputation”, I’m going to make sure to rate my friends and family higher than everyone else. This will make it hard for me to adequately reward some guy who I don’t personally know, but whose work I like, with a comparatively high score. But if there are many kinds of reputation, such as “personal reputation” (people you personally know and trust), “professional reliability”, “programming skill”, etc, then I will have no problem giving my sister a high “personal reputation” but a 0 in “professional reliability” (I think she is reliable, but I’ve never done business with her; this sort of rule could be a general normative guideline that comes with the “professional reliability” standardized namespace), and a 0 in “programming skill”.

Limit the total amount of friendedness, thereby creating scarcity This gives people an incentive not to inflate the points they assign to other people, because they only have so many points to distribute. Therefore, other people can trust that if they see that you have assigned a high reputation to me, that that really means something.

I think the limit should be per-person, rather than a limit in the total economy.

This is the core of my ReputationBudget proposal.

I think alternative currencies clearly work, (because there are several already, that have been running for a while,) but I’m not sure what a “reputation” currency would be.

I mean: “Your reputation is our story about you.” It’s not something you can give away. If you could just give it away, maybe trade it for a bike, then a guy with a bad reputation could just suddenly buy himself a good reputation, by paying a guy who had a good reputation, and will subsequently have only an okay reputation (or perhaps a bad one.)

“Mafia don purchases positive reputation from Mahatma Gandhi. Everyone’s mad at Gandhi, but whoah! Look at the Don! What a great guy he is!” It’s just never going to happen.

I think we’re going to see metrics, bestowed by the OrganizedCulture, (by the membership of the culture,) and those metrics are going to grant priviledges in treatment and circumstance. People will work to meet the metrics, like high school kids, but focused on much more than just clothing. But I don’t know that I would call it a currency.

But perhaps we mean something different by “currency” here? If so, what?

The phrase “reputation currency” brings to mind: “In God we trust.” Do we need to be theistic, in order for money to work? No. Calling the currency “reputation” seems more like: “In community, we trust.” The concept of reputation and community, connected with the exchange. But the bills could say anything, and they’d still work.

I think you are right: reputation can’t be traded, and so isn’t a currency.

I guess what I was talking about was actually two related ideas:

  • A reputation-based economic system without currency (ReputationEconomicSystem?)
  • A new type of currency that is sort of like “reputation” but is not quite the same as it (PsuedoReputationCurrency?)

These ideas are closely related, however; I think that what is described in Doctrorw’s book could be either one of them, depending on the details that he didn’t specify in the book.

6 you could have an economic system in which having a high metric essentially takes the place of money in “paying the bills”, even without trading; if someone has a high metric, you do things for them. It’s not a “reputation currency” but more of a “reputation economic system” without currency.

PsuedoReputationCurrency?: there could be something that could be traded, that isn’t “reputation” exactly but for which we don’t yet have a name (we could say whuffie but i think that name is so dumb sounding!). For instance, you eat at a restaurant for free; but there is an implicit understanding that you bump up your “reputation” score for that restaurant a little bit in proportion to how often you eat there and how good the food and service are. If you stop going to that restaurant and go to another one, you lower the “reputation” score that you assign to the first one, and raise the score of the second one. This is more of a currency, but it’s not quite reputation (after all, you think just as highly of the first restaurant after you stop going to it). Even here, though, it isn’t quite “currency” if your own reputation can’t be lowered when you raise someone else’s. But perhaps we could find a way to ensure that happens, indirectly? Hmmm…

I feel like the following sort of “transaction” would be key to either system: a “quid pro quo” of providing a service to someone in hopes that they will raise your reputation; the Mafia don might do all sorts of favors for the Pope, build homeless shelters etc, hoping that the Pope will raise his reputation opinion of the Mafia guy; the Pope is not giving away his own repute, however, so it’s not quite an exchange of a service for currency.

If we have a “ReputationEconomicSystem?” you might have people getting mad at the Pope for letting the Mafia guy “buy” goodwill by doing nice things – “he’s obviously faking it; the Pope is being corrupt”. If we have a “PsuedoReputationCurrency?”, on the other hand, there would be nothing wrong with the Pope and the Don even signing a contract beforehand, since the Pope’s “reputation” opinion of the Don isn’s supposed to be what he really thinks, but is rather just a “psuedo-reputation” economic construct.

Note to self: Write SoftEconomy when I get home. 😊

If all transactions must (by social convention) be done “in public” (i.e., registered, whatever,) then you get these neat things, where: A person registers a need for help. Someone who’s trying to get into the system fills that need. They get a positive reputation in return. (The person who needed help gets a reputation as well! Overly cranky “customers,” or people who abuse the system, will be penalized..!) Now they have a positive reputation, and people in general who are paying money will trust them enough to hire them, for pay.

It’s like Craig’s list, but with persistant reputation, and a Wiki-like RecentChanges page, showing all the transactions going on. You could comment on transactions.

It’d be incredibly humane. If someone was in a dire situation, and needed something, and can’t pay anything, and you help, then that transaction is visible, and goes onto your reputation, your permanent record.

You can think of it as a sort of “soft debit,” that plays on all the various debit/credit systems that go into manners, as well. (Perhaps that’s where the “currency” part comes in?)

I suppose there is a currency, and there is an exchange, in reputation. But the currency isn’t what’s traded; Rather, reputation is the ledger.

So you do a favor for me, and I’m in the direction of “mooch” territory, and you go in the direction of “helpfulness.”

What’s interesting is that new currency can be “issued.” If some terrible accident happens, and I need help, and you help me, then you gain “helpful” territory, but I don’t become a “mooch,” just because a comet fell on my house.

What’s being traded are personal favors, time, attention, and care. Your tally is your reputation. If you do good things for people, you get a greater tally. Note that you can’t hoard! If you went and did a bunch of favors for everybody, you start getting diminished returns. People will start to looking at you, like: “Uh, that’s a little weird. This kind of makes me uncomfortable.” If you do it too much, people will even resist you! So, it’s not like capitalism, where you can hoard capital, and then turn it against people.

Damn. I’m writing the SoftEconomy page, right here. I’ll move it later.

I think this is a special case of what I was calling a ReputationEconomicSystem?.

The difference is that you are saying that instead of just keeping track of each individual, the system keeps track of each transaction.

Very software-intensive, hence SoftEconomy? Or do you mean soft like touchy-feely?

I think the human labor of monitoring all those transactions and discussing them would be too much. But maybe not.

BTW, why can’t you hoard? I would be just as grateful to someone who washed my car, whether he just did it once or whether he also painted my house and babysat my kids and did the same things for everyone on the block.

If someone did that without receiving anything in return, people would think he was batty, and force him to receive favors from them, thus deflating the haording. At this point, you’re basically working out a trade.

(right?)

Sort of like: You can’t give a girl a teddy bear, and that mean that she has to like you.

That kind of thing.

So what if you want to work really hard for 20 years and then take a 10 year vacation? Is that option not available here?

One problem with any ReputationEconomicSystem? is that it creates more risk because you don’t know before you do something how many points you will ultimately be rewarded for it. People would be very reluctant to start long-term projects, such as today’s “startups”, because not only do they not know if they will succeed, but they don’t know how much credit it will buy them if it does. This is similar to not knowing the demand of your future product in capitalism, but it seems like it may be worse.

Another problem with almost any reputation-ish system is that grumpy, unpopular, unfriendly, but hard-working people are unfairly penalized.

And another problem is security. People (especially people with a bad temper) will be stressed out knowing that not only their future earnings, but their past earnings “in the bank” could be blown away if they one really mean thing.

And this ties in with a free-speech problem; in a reputation-based system, it will be even more economically hazardous than today to say something politically unpopular, particularly if what you have to say insults people.

(not that I think we shouldn’t try out these systems anyway; after all, capitalism has its own problems, maybe if both systems existed at once, there would be “something for everyone”)

Gotta run, but:

In some respects, you can take more risks in a SoftEconomy! It’s because people are, on the whole, forgiving, where-as the capital economy is most certainly not.

If you really blow it in a SoftEConomy?, then all that’s required is a ton of apology, and dedication to work for a long time, and avowal to never do it again, whatever it was.

In the capital economy, if you really blow it, you go to jail, or you end up out on the street.

A normal person can’t very easily start their own business. They can dream about it, but it’s psychologically enormously difficult to do it, to take those risks.

But in a commune, like TwinOaks, you can say: “Hey! Here’s this idea. I think it will work. Can we try it?” People talk about it, and if they think it will work, they say, “Sure!”

If it tanks, people are sad, but no lasting damage. Try that in a capital economy!

So I think the SoftEconomy is very good at encouraging business and the like.

I do think we will have mixed economy, and I strongly believe there will be multiple groups that you can choose to work and live in.

Another potential problem, but probably a solvable one: you need to make sure that complicated, technical services that few people understand are still correctly values. For instance, some sort of specialized job that in today’s economy would earn a lot of money, but that few people know what it is or understand why it is worth so much.

Another potential SoftEconomy problem is making sure that supply/demand is factored in. People might be tempted to say, “why should we reward this Wall Street corporate accountant more per hour than this other guy who fixes people’s sinks?”. In this economy, it’s because high finance corporate accountants are more scarce than people who can fix sinks; this encourages more people to become corporate accountants.

In some respects, you can take more risks in a SoftEconomy!

OK but if people are willing to forgive failed startups, that has its own risks.

Whoah Whoah Whoah! I didn’t say anything about “we’re getting rid of conventional currency.”

Rather, I think there will be a complex network of currencies.

Ex: SoftSecurity does not spell the end of HardSecurity! (By any means!)

Nor will a SoftEconomy spell the end of the conventional currency based economy.

I think there will be a mix of soft, currency, and alternative economy.

I like to tell people: “Soft Security is security that exists in people’s heads.” “Soft technology is technology that exists in people’s heads.” So a “Soft Economy” is an economy that exists in people’s heads.

It already exists, it just doesn’t go very far today. I’m saying I think this has the potential to be a major currency: favors, traits, kindnesses, and showing in: reputation.

Yes, but similar problems may crop up whether or not the reputation economy is the only game in town, right? I think it is still useful to explore how the various reputation systems would fail to solve certain problems, even if we don’t actually intend to rely on them in those situations.

Had a conversations with some friends (including RuadhanOflanagan?) in which more issues were pointed out:

  • if reputation ratings are not anonymous, people get their feelings hurt
  • if they are, then what’s to prevent someone rendering a service to you only to find that you don’t bother to repute them up?

My friends also raised one of the issues I raised above more forcefully. Essentially, part of the conversation went like:


OK, what if I meet someone who has earned a high reputation, but she is from a different city so I’ve never heard of her.

Then hopefully, you’ll have a friend who has a friend who has a friend who knows her, so you can hear of/trust/compute her high reputation rating via this chain. For example Ghandi would enjoy the benefits of his high reputation rating even with people who had never heard of him, because so many people would rate him so highly that even if those people were a few hops away from the person interacting with Ghandi, the vast numbers of people giving Ghandi a high rating would have a large effect

What about a hermit, then, who is very nice and who is a productive member of society, but who knows almost no one? There will be few people to vouch for the hermit so even if they give him a high rating, he doesn’t do so well. Also, when he is far from home, the chain connecting the people he interacts with to people who can vouch for him is very long. Whereas someone who is well-known doesn’t have these problems, even if they receive similar ratings from the people who know them. So this system is really designed to reward popularity, right?


I’m beginning to get a better idea of the design space of reputation economic systems. In the following list, the top-level entries are choices (usually binary choices); subordinate nodes are either explanations, or sub-choices (choices which are only available if the parent choice goes a particular way).

  • the “reputation” in the system can be either expected by everyone to actually mean what we mean by “reputation”, or not
    • for example, if the Pope can in some sense sell some reputation to a Mafia don, and no one thinks that anyone is “cheating”, then “reputation” in the system is not really reputation as we usually mean it
  • reputation judgements can be anonymous, or not
  • reputation judgements can be annotated and discussed, or not
  • individual transactions can be recorded within the system, or not
    • if not, then the only thing the system records is which ratings people assign to other people
      • (the relationship between ratings and transactions is similar, but not necessarily identical to the relationship between the integral of a function f() and f() itself; that is, each transaction adds to or subtracts from reputation, but all you really need to know later is what the reputation is at the end of the day)
    • if so (transactions are recorded), then each transaction could be annotated or discussed (SoftEconomy), or not
  • either you can effectively trade away your own “reputation”, or not
  • either you can enter into a contract with someone else to raise your appraisal of them in exchange for a service, or not
  • either reputation is a locally scarce quantity, or not
    • by which i mean, can each individual rate many other individuals as high as they want without limit? or is there some mechanism like ReputationBudget that means that you can only give a higher reputation to one person by effectively lowering your score for someone else?
    • either reputation is a globally scarce quantity, or not
      • by which i mean, is there a limit of the sum of all of the “reputation points” in the entire world?
    • Is it considered “cheating” if (1) someone does something for you (2) you bump up your rating of them in response (3) a week later, someone else does something for you; in order to free up points to give to them, you lower your score of last week’s person
  • there could be many types of reputation, or just one
  • reputation could be hoarded, or not
    • can you build up a vast store of reputation to be “spent” later?
    • reputation could be securely hoarded, or not
      • can you lose your “life savings” of reputation all at once, or is there some way that you can have the security of not losing everything?
  • Is reputation a (good) medium of exchange? (WikiPedia:Concept_of_money)
  • Is reputation a (good) store of value? (WikiPedia:Concept_of_money)
  • Is reputation a (good) unit of account? (WikiPedia:Concept_of_money)

There are also constraints between these questions. So far, we have:

  • If scores are assigned anonymously, then you can’t make a contract with someone to raise their score in exchange for something
  • if scores are not anonymous, then system “reputation” probably can’t be the same as reputation in the usual sense, because people would get insulted
  • if you can trade away your reputation for something else, then “system reputation” != reputation

Some possible interpretations of “system reputation”:

  • reputation (whatever that means)
  • ethical “goodness” (the Pope vs. the mafia don)
  • “is this a trustworthy/competent person?” (i.e. “i don’t care what they have done for society, but if i hire this person for a job, will they follow through?”)
  • “how much this person has done for society/how much society owes her” (i.e. “this highly-rated person has done a lot of work for others, so i’ll do some for them”)
  • “how much this person has done for society/how much society owes her” times popularity (i.e. “everyone says this famous author has done a lot of good work, so i’ll do some favors for them”)
  • a currency (the amount they have on hand has no deep meaning, but is simply the result/sum of their transactions to date)

Anyone interested in going through (& expanding upon) the “design space” as I’ve given up above, thinking of each possible permutation of answers to those questions, describing each resulting economic system along with obvious strengths and weaknesses, writing this up into a paper, and submitting it to FirstMonday? I’d like to do this myself even if others aren’t interested, but co-authors would be welcome (man, writing a paper would necessitate doing a literature search at some point; darn!).

I think this is an interesting conversation; I’m sorry I dropped the ball on it somewhere in the past; Trying to figure out what’s going on here.

The last post, Bayle, I have a hard time following it. You’re using the word “Reputation” in ways that I’m not getting it: It’s very abstract here, and I’m not sure what it means.

I guess I’m asking you for more language. More words. Something to make it a little clearer.

I don’t know if writing for FirstMonday is the best thing to do with this right now; How about writing out the ideas on this wiki?

Take what you wrote above, and put it on AlternativeCurrencyDesignSpace?, or something like that..? Maybe “CurrencyAttributes?,” to go alongside the MediaAttributes? page?

I do think that a lot of what you are talking about is currency, not reputation. Once you start talking about transferring bits of number from person to person, it feels much more like currency, and a lot less like reputation, to me. I think you need my authorization, before your reputation in my eyes changes.

I think we’ve talked about reputation currency type stuff a few times here now; it just seems so pregnant with LinkLanguage possibilities.

So, … let’s start! Would we agree with ReputationIsNotaCurrency?, as a founding point? I mean, it’s a clear, distinct idea, in my eyes, and I think we agree. But let’s suppose you disagree. (I don’t know: I’m just sitting here typing.) Then lets figure out a discussion, and name it, or a piece of idea that we both agree on, and name it.

OK! I’m excited too – but I’ve gotta be in lab a lot this week so I will be moving slow for awhile. Don’t interpret that as lack of enthusiasm.


Possibly relevant links, unexplored as of yet:

The six grants have paid about about $400k apiece. So, in total, about $2mil has been spent researching this sort of thing so far (“this sort of thing” = electronic reputation systems != reputation currencies; this just seemed like a good page to put this on).


At the RecentChangesCamp2007, I was encouraged to think of currency from it’s root term: Current. This suggested the word: Flow.

I had looked at reputation not as a currency, because reputation is non-transferable. (I can give you $5, but I can’t give you my reputation.)

But it does seem to be part of flow. Reputation is lost, or at least depreciates, if you don’t continue to engage in activity. So reputation requires flow, and thus can be properly given the label “currency.”

Aren’t there different forms of reputation? One can have the reputation to be a great visionary but hard to communicate with. One can have the reputation to be reliable or a bad reputation as as a troll.

Does this really fit to the model of a currency? A currency is convertable into another currency. How is reputation converted?

Lion: That is interesting, the root of the word, I mean. In TrustMetric applications, it does seem to be an exchange, a reciprocation between transactors. Like the mutual-rating on ebay.

Helmut: Hmmm…I think that “ReputationCurrency?” only really applies to TrustMetric-type reputation. “ReputationCurrency?” is converted into actual trust 😊

If do an hour’s worth of work, I am given let’s say $5.44 US, which is convertable to both other currencies, and converatble to physical goods and services.

If I complete a transaction on ebay, I am given a rating, which is convertable on ebay to trust from future users (if it is a good rating). Also, if it were allowed, it could be convertable to another type of reputation currency from another system, like Amazon, or whatever. Although, I think ebay tries to avoid allowing people to export their reputation out of ebay.

external transclusion (from a text file http://janus.datafix.com/hwo/blocks/070215.dcn) …