Wikipedia, Prices, and Hayek
Hayek's big claim about the price system was that it aggregates widely dispersed information and tastes. For this reason, he said that it was a "marvel." We've been discussing other ways of aggregating information, and it might be useful to start with Wikipedia, if only because Jimbo Wales refers to Hayek in his comments.
Wikipedia does aggregate dispersed information -- amazingly so. In a general way, it's definitely a Hayekian process. But there are at least two differences between Wikipedia and the price system. First, Wikipedia doesn't rest on economic incentives. People aren't participating because they're getting a commodity or money. There are no trades. Second, Wikipedia generally works by a "last in time" rule. The last editor. and hence a single person, can do a lot. But in the price system, the last purchaser usually can't have a huge effect. (Even if you buy 10,000 copies of each of Larry's books, you won't affect the price.) The upshot is that Wikipedia is different from the price system; it aggregates dispersed information in a distinctive and less reliable way.
Two qualifications. 1) Wikipedia nonetheless works, at least for the most part. 2) The price system doesn't always work, in the sense that bad information, sometimes spreading like wildfire, can produce inflated and deflated prices. (So Hayek was too optimistic, as behavioral economists have shown.)
What is this discussion missing?
| Permalink | technorati
Comments (34)
"What is this discussion missing?"
I would suggest: Some sense that "aggregation" is a very broad phenomena.
As much as I can make out, the overall concept seems to be:
1) Hayek said prices were *agreggation* of *information* for *markets*
2) Wikipedia does *agreggation* of *information* via *the Internet*
3) Therefore, drumroll ... Wikipedia/Internet is like Hayek/prices!
But just because two different phenomena are both aggregation at some sort of macroscopic level, hardly means they have the exact same mechanism at a microscopic level.
I think this analogy may break down as it relates to software. Software is unlike a "price" and more like "art". The one is measuring some extant quantity and the other is the creation of something new.
Wikipedia is interesting: it tends toward perfection, in the sense of a price, but it is certainly developing new value. It is enabled by and expressed in software, but it is not software.
"Peer pressure", when you are the last one adding a finishing touch to some great work, you do not want to screw up ...
>> aren't participating because they're getting a commodity or money. There are no trades.
People are receiving the encyclopedia they are authors of. That is, each person receives the benefit of the labors of all others who are participating.
It's not necessary to participate to receive those benefits so there is the possibility of doing nothing and not influencing the process.
Those who participate may also receive intangible benefits such as community membership and a sense of achievement.
There can sometimes be more tangible benefits. As one of the technical team, I learned about a particular product because that was of assistance to the projects. I was subsequently hired by the vendor of that product, having demonstrated my skills in the traditional open source way: by doing things and helping others on their support areas using the knowledge I'd acquired.
For me, the fact that people enjoy creating encyclopedias without any monetary benefit proves that a more general gift economy can possibly work, at least for intangible information products and services.
Aside from Seth's well-put response, I think you're giving Wikipedia too much credit. Sure, by and large it is an accurate source of info, but, by and large, prices are accurate indicators. Aberrations exist in both.
Making babies and making prototype cars are both about creating something new, about birthing. What can a GM engineer learn from an obstetrician? Nothing, unless you keep the answer as vague as the comparison "the beauty of creation" but not anything specific like "cars need stirrups"
If I buy ten thousand copies of a book I will impact the price. Directly. I will impact it for myself, as the publisher will most certainly agree at that level to grant me a discounted price, and my purchase will drive the publisher that much closer to recouping the initial investment which will then lead to the book being produced that much quicker in a discounted paperback (or even electronic) form.
And if my purchase leads to an increased audience for the book (which is very likely given that most folks would not buy ten thousand copies of a book unless they planned to resell them) then it is much more likely someone in that audience will scan that book and post it to an e-book newsgroup - which will likewise increase its circulation that much more and, equally likely, lead to even more sales and even lower prices.
Wikipedia does the same thing. I'm writing this nonsense here because I want to be heard. If I have an agenda (and everyone has an agenda about something) and I am driven to express my opinion on a matter to as wide an audience as possible, wikipedia is one potentially good soapbox for that. The more popular the entry the more likely my voice wil have to compete with others, but this doesn't stop me from competing - going back again and again to encure my agenda is well reflected in the entry.
This is not entirely intangible. If I am selling a product and my entry helps shape societal views on the product I am selling, the added profits I would reap from a favorable public perception wouldbe entirely tangible. Hollywood does not send Millions of dollars to DC lobbyists every year out of some desire for touch feely "intangible benefits."
“(Even if you buy 10,000 copies of each of Larry's books, you won't affect the price.)”, is an interesting parenthetical proposition. The question is, is that necessarily the case?
For me the fundamental similarity between price systems and open source approaches is individual freedom. Both systems aggregate information (an emergent property) in an unplanned (to a large extent) way due to individual free decisions of participants in markets, wikipedia or software projects.
The upshot is that Wikipedia is different from the price system; it aggregates dispersed information in a distinctive and less reliable way.
exactly right.
if we can provide an additional clarification to seth finkelstein's erudite observation. prices are not themselves an aggregation of information, rather they are reflection of the perception created by that information. when, for example, the information suggests that google will continue to show explosive growth - and that information is perceived as reliable - the price of google shares rises. the same information about google's prospects - when perceived as unreliable - might result in a different price even if the information is accurate. similarly, as professor sundstein observes, bad information when perceived as reliable can have a similar effect on prices. the price system is certainly not infallible, but it is generally quite good as an indicator.
the wikipedia is, by definition, an aggregation of information, but since there is no price associated with its use there is no "reliable" measure of the validity of this information. this is why, on better blogs, linking to a wikipedia entry is considered as indicative but by no means authoritative.
this is not to say that the wikipedia does not have value, but it is no replacement for a commericial encyclopedia because: it aggregates dispersed information in a distinctive and less reliable way.
"What is this discussion missing?"
Maybe a quantitative analysis that proves the Hayekian price system is less reliable than the Wikipedia knowledge system. Both differ by output but I suggest that both can be disorganised by the same element (bad info).
There's an obvious methodological problem here (e.g. one could compare price distortion on the dishwashers market and knowledge distortion on cognitive psychology W. pages, but is that feasible and pertinent at the same time, I doubt it).
If I buy ten thousand copies of a book I will impact the price. Directly. I will impact it for myself, as the publisher will most certainly agree at that level to grant me a discounted price
I know from experience that I can impact the price by buying seventy copies or so -- I've done that.
As for Wiki, what is interesting is that it managed to work, while the various directory systems have pretty much failed. What is also missing from the analysis is that Wiki is pretty much static, while the open directories (consider the one Google used to support, then abandoned, or Yahoo's old core).
They were dealing with dynamic information and failed.
So, you are missing that Wiki deals with static information, with payment (usually) being those pushing an agenda, the market deals with dynamic information.
There's a distinction (which appears, for example, prominently in Habermas's work) between strategic action and communicative action. The former is aligned toward one's own interests, while the latter is aligned toward shared projects.
It would seem that this distinction would be relevant here. I hate to rain on the parade of our economist friends, but when people interact with each other in a strategic mode, they have incentives to do things that screw up information-aggregating functions, like manipulate the market, propagate disinformation, exercise coercive power, etc. The same can not be said for a collaborative mode.
Why is wikipedia seemingly as reliable as the price system?
Simply because the information represented by an encyclopedia is generally a more static set of data. Henry VIII is not going to resurrect and marry a 7th wife, Columbus is not going to make a fifth visit to the New World. This is why wikipedia works, because the information is based on verifiable historical facts that don't change over time. And it takes a lot of time to build these articles up. If you look at wikipedia you will find that now that it has matured, people are not usually erasing articles and rewriting them from scratch (or someone else will probably revert it back), it's an evolutionary rather than revolutionary process. The consistency of data means that people can slowly add, expand, and debate what belongs.
On the other hand, the price system is representing something that is extremely dynamic. There is nothing more fickle than the subjective taste of public consumption. Something as ridiculous as Paris Hilton wearing a top on television could start a whole fashion trend. Rumors spreading like wildfire can change the value of the stock market by billions of dollars. This doesn't mean the price vs the wiki system would be any more/less susceptible to bad information if they were analyzing the same thing, just that the price system is lifting a heavier dynamic load.
Past comments have touched on this, but I feel they haven't made it explicit enough. The problem with Sunstein's post is that the information Wikipedia is aggregating can just be looked up elsewhere; it's not really that different from any other encyclopedia, is it, except that there are more editors/contributors because of the internet's efficiency? Aggregation is ALL it's doing. The price system is doing *much much* more. It's drawing out tacit knowledge and producing relative valuations that can't be "looked up" anywhere. It's serving an epistemological function that can't be compared with mere aggregation. And it's doing so despite the lack of any intent (and, with respect to tacit knowledge, *ability*) of people to reveal their preferences or their relative valuations or their tacit knowledge or ....
You claim that Hayek was "too optimistic" because he didn't understand that the "price system doesn't always work". This is a misreading of Hayek. Can you provide a citation in which he claims that the price system "always works"? I suspect not. Nothing "always works," and Hayek was smart enough to realize that.
Hayek's claim is that the price system works better on average than any other system.
Cass:
1. Your second comparison of the market and Wikipedia isn't apples-to-apples. The market also works on a "last-in-time" rule—if I go onto the trading floor, and offer to buy gold at $600/ounce, the price will be $600/ounce until someone outbids me or I run out of money. The market is just so much larger than Wikipedia that one person's "last-in-time" decision rarely has an impact. But if Wikipedia consisted of millions of people editing the same article, you'd see the same lack of impact, as the LA Times found out. You see the same sort of discrete bumpiness you describe in Wikipedia in smaller artificial markets such as the market for free agent quarterbacks.
2. Can one really say that Wikipedia works better than the market? That seems to be a testable claim. I've found lots of misinformation and repetition of bad information on Wikipedia when I explore subjects I know a lot about. Sometimes I take the time to correct it; sometimes I don't. Pick ten things you know a lot about, and see what Wikipedia has to say. I personally don't see any reason to believe that Wikipedia is any more "efficient" than the markets described in the efficient markets hypothesis.
If the price system aggregates widely dispersed information and tastes, it does so in a manner remarkably different from Wikipedia.
The on-line encyclopedia aggregates information and tastes in the sense that many different pieces of (even conflicting) information and many different (even opposite) tastes may be entered into the system without contradiction. Horses, for example, may be described in Wikipedia as pretty animals. But they may also be described, in Wikipedia, as ugly brutes.
When the price system aggregates widely dispersed information and tastes, it aggregates that data into an average that may be a less reliable indicator of value. In the above example, the horse's value in a price system, based on the input data, is neutral -- one positive recommendation, one negative -- and is therefore an ineffective indicator of how people actually value horses.
That means that in a Wikipedia pricing system there's a place for the idiosyncratic party. Mr. Peevyhouse can go into Wikipedia and type "I think horses are really, really, really pretty" and add a picture of the most beautiful horse in the world, but there's simply not enough room to reflect that valuation in the price system.
The Wikipedia fails to distribute this information properly.
A more proper model for the Wikipedia would be a distributed system where the information is sent out by torrents and stored locally on the user's hard drives. This distributed network would eliminate the centralized server idea that is holding wikipedia back.
Tell me the truth. If I don't care about page views, and you don't care about page views, why are we still structuring our information distribution on a page view (per click advertising) model?
read more rants: http://thunt.net
But wikipedia does have an economic incentive. People want to show off how much they know, or maybe they just want to see their words on a big, successful website. Or maybe they just want to contribute to that particular community.
Free public service has economic incentives also.
Based on the Misean premise that the reason a free market economy works is that "people want."
I think that the element of "the power of large numbers" is missing from the analysis.
;-) http://en.wikipedia.org/wiki/Law_of_large_numbers
At any given moment, the price of a particular something might be momentarily skewed, and the wikipedia article about something might be incorrect. Some buffoon might have posted wildly inaccurate information about widgets on wikipedia, or the lunatics on Wall Street might have bid a stock to an idiotic price/earnings ratio.
But when taken over a long period of time or a large sample of somethings, both systems settle into what can be called correctness.
How about not defining economic incentives to be strictly about exchanging commodities? Time is a limited resource (therefore the a topic of economics). Maybe wikipedia is barter system exchanging services (collecting and disseminating information)? Maybe they're trading in social status? As a primer for understanding non-traditional internet economies, you might try reading Homesteading the Noosphere by Eric Raymond.
See Yochai Benkler 2005 Business Week article ("The Sharing Economy") and his 2002 Yale Law Journal in depth article on peer production vs. the price system as a better modality to produce valuable information goods ..."Wikipedia provides a rich example of a medium-sized collection of individuals who successfully collaborate to create an information product of mid- to highbrow quality. In particular, it suggests that even in a group of this size, social norms coupled with a simple facility to allow any participant to edit out blatant opinion written by another in contravention of the social norms keep the group on track."
The price system is what we call a theoretical entity. Wikipedia, in contrast, is straightforwardly real (and easy to find). One could doubt the reality of the price system and thereby start a big old argument where various dogmas of free market economics might be refined or discarded. Not so with Wikipedia.
To take a step in that direction, compare "informations": Like I said we know where to look with Wikipedia. And when we do look there we find aggregated information par excellance. With the price system the "information" is spoken of metaphorically (cf. the "information" in DNA). As far as the aggregation goes, we can only look for prices and quantities (and I don't know where to find the latter, do you?) which fail to deliver all sorts of information that the economically trained blithely talk about. Consider for example, how much I would be willing to pay for a copy of James Burnham's "The Managerial Revolution": That info is not to be had by looking at any transactions, although I did buy the book. Most we can say is that I was willing to pay what I did (at least).
Main points: There may be lots of aggregated information in the price system. But unlike Wikipedia there does not seem to any place where it is aggregated. If there are people (call them "market researchers") who have this information in an aggregated form, what they have access to is not what Hayek was talking about. But what forms aggregation does take when it comes to prices and quantities of goods is itself an empirical question and has (I am certain) changed very much since Hayek developed his theories.
Sorry for being so scatter-brained.
Bill,
The price is the place where the information is aggregated. Individuals have their private information about their preferences and costs, and each time we make a choice to buy/sell something, or not to, we reveal information that, when aggregated with the choices of all others, will show up in the price.
Thus the "place" where the information aggregation shows up is in the enormous price vector (sorry!), or big set of prices for everything that is transacted through market processes.
“(Even if you buy 10,000 copies of each of Larry's books, you won't affect the price.)
Two qualifications. 1) Wikipedia nonetheless works, at least for the most part. 2) The price system doesn't always work, in the sense that bad information, sometimes spreading like wildfire, can produce inflated and deflated prices. (So Hayek was too optimistic, as behavioral economists have shown.)
What is this discussion missing?"
What's missing is the notion that pricing isn't the only form that information aggregation about markets can take (in some ways pricing is a singularly useless signal). And other forms of information aggregation *can* be affected by a 10k item purchase.
The example that comes most immediately to mind is Amazon's Sales Rank, where a 10k copy book purchase will likely push a book to #1 at least for a few hours, but there are older examples as well, such as the NYTimes Bestseller list, and Bowkers Bookscan data.
These forms of aggregation have their own flaws and are subject to manipulation as well (eg. Church of Scientology telling adherents they must go out and buy X copies of a particular L. Ron Hubbard book), but that does not negate the utility of these forms of aggregation as signals. There is after all no point in choosing what books to read based solely (or even mostly) on their price.
One may, of course, choose what subject to study (and therefor what subjects to buy books on) based on indirect pricing informaties on (for example, studying a specific computer programming language because salaries for Software Engineers with that skill are rising), but that doesn't help you choose a *particular* book at all.
So with regard to Wikipedia etc, it is clear that there *is* a market in place, but pricing just isn't the form that aggregated signals take, other means of information agregation are being used, just as they are used in other markets.
One place where Wiki seems to fail is in political commentary. That is because it seems to end up with the views of the group that most actively participates in it, as opposed to an accurate view, or even the view of a majority.
This is somewhat akin to the problem of on-line polling. The participants in it are self-selected, and, indeed, results can be, and often are, skewed by organized campaigns. I would suggest that Wiki has the exact same problem.
The Wikipedia entry on "Pricing" may be of some interest.
To continue the theme of my earlier post:
What would Hayek think of the phenomenon of "transfer pricing" (about which you can read on Wikipedia)? Probably it was much less significant in the 1940s than it is today. I know from experience that it is very important in international taxation, and thought to be a hard- to-police way in which firms escape taxes.
I alluded in an earlier post to the empirical question of just how much "firm" and how much "market" (as Coase might say) we have. The importance of transfer pricing may be one variable to consider in an inquiry along these lines.
Bill,
I worked in transfer pricing in my short sojourn out of academia, and I repeatedly asked "where's Coase here?" My fellow PhD colleagues laughed, and agreed that in theory transfer pricing is all about Coase, but in reality transfer pricing is all about jointly maximizing profits while minimizing the probability that the tax authorities of the involved polities would challenge your pricing. In other words, the pragmatic politics of compliance meant that the fallacy of the transfer pricing comparison of internal transactions with "comparable" external transactions got lost.
What Wikipedia/open media/copyleft has shown is that IP cant continue to be commoditized in the same way as say durable goods (DG), or more generally tangible goods (TG). "The economy" refers to spectrum of different "goods," each of which have different value curves for different economic environments. The web has changed the environment particular to many forms of IP, while leaving most TG untouched.
"What's missing" is a thorough understanding of the harm done by economic-political leverage being directed toward keeping IP economics inline with TG economics. Unrealistic centralized controls simply means a weak centralized system.
It would be interesting to add "democracy" to the economics/wikipedia conceptual discussion.
For example. In presidential elections, there's incredibly high granularity -- in effect, it's binary: each citizen has exactly one vote, which they can (ignoring electoral college etc.) choose to cast for exactly one candidate, or for none of them. (Hm, make that trinary.) The people who care most about the election's outcome don't get more votes, and neither do the people with the most money.
In a market, there's lower granularity: you have as many votes as you're willing to spend -- $1 or $1.01 or $38936. But money isn't divied up fairly (nor can it be; I'm not advocating maoism), which is why there are more and better positions available in the yacht-design-and-maintainence industry than the teaching-impovershed-children industry.
Wikipedia, on the other hand, is a give-a-shit-ocracy: the people who care the most have the most power. You can't leave to your offspring your passion for editing WP entries on waterfowl (at least not in one fell mortuary swoop, not without putting at least as much passion into raising them).
Yes, that's a little oversimplified -- there are elements of give-a-shit-ocracy in persuading people to vote for your party, and in working long hours in a market economy (and, too, there are problems with systems in which people who care the most have the most power -- in wikipedia they have proportionally more power, but not, I think, so much more that one of them could seize control in coup de etat and refocus the encyclopedia on rebuilding the galactic civilization), but it's a matter of scale: wikipedia is a much purer form of adhocracy than the others.
"What is this discussion missing?"
One point of difference between "Wikipedia as aggregator" vs "Market price as aggregator" is some notion of stability or convergence over time (vs. oscillation or divergence).
Wikipedia amuses and frustrates when its articles on controversial subjects oscillate between extremes. As I recall, its article on the new pope went through a rapid flurry of edits from fevered and diametrically opposed contributors in a time series that went {"he's a Nazi" , "he's holy" , "Nazi!" , "holy" , "antichrist" , "NAZI" , ...} and oscillating on and on.
Even controversial stocks converge on a single price. You rarely see a series of trades that consistently runs {$1, 900, 1, 900, 1, 900, ...}. Far more often it converges around some stable value {$50, 70, 60, 65, 62, 63, ...}.
Or in the case of my investment in Enron, {$100, 50, 20, 10, 5, 3, 2, 1, ...} Oh, well. ;-)
I suppose the difference is in the notion of closing a deal; if I hate Ratzinger, then no amount of your advocacy will change my opinion of him. But if I "hate" a stock (think it's worth $1) and you "love" it (think it's worth $100), then you and I can still reach some sort of deal, e.g. I sell to you at $80 and go laughing to the bank. Until you turn around and flip it for $100. ;-)