Smart Crowds and Dumb
Blogger-dot-com uses a flagging system, whereby blog readers report offensive material for censorship. I have no idea what is supposed to qualify; I assume it's primarily intended for illegal postings like copyright infringements and child pornography. But while I was reading about it they (the website authorities) turned me on to a concept: the Wisdom of Crowds. There's a recent book about it by James Surowiecki called, of all things, The Wisdom of Crowds, of which I have only read an excerpt. But Wikipedia sums it up, and if you faithful 'dlog readers know anything it's that you can trust information that is posted on the World Wide Web. On the Wikipedia page about the Wisdom of Crowds it is pointed out that the term is a counterpoint to the Madness of Crowds, described at length in Memoirs of Extraordinary Popular Delusions and the Madness of Crowds by Charles Mackay in 1852.
Are crowds wise or mad? I think feedback is a key factor. According to Wikipedia, one of Surowiecki's necessary conditions for crowds to display wisdom is that one crowd member's opinions don't affect anyone else's, nor are they affected by others'. If this is the case, the common wisdom of the people is likely to outweigh the ignorance of the individual. I read an example where townspeople tried to guess the weight of a cow. I imagine this is pretty much the ideal scenario for crowd wisdom. Nobody talks to each other about it (I assume there is a prize at stake); everyone just takes a stab. It would be like everyone in town throwing a dart at a dartboard; the average throw would quite probably be exactly in the bull's-eye (given enough contestants). Or like that game where you try to guess how many jelly beans there are in the jar. Ask 5,000 adults and average their guesses - that will tell you.
But what about when people's opinions do affect one another's? Mackay describes the craze over tulips in Europe in the 1600s. It is regarded as an early description of an economic bubble, and it is just what happens when feedback runs rampant in crowd behavior. I am a tulip importer (really!) and I understand that tulips are all the rage in Holland's upper class. The Turks I buy them from don't speak the language, so they don't know to raise the price, so it's my opportunity to make a killing. I can still buy them from the Turks at 20 florins a bushel, but my selling price goes from 30 florins to 100. People will pay this - these are the Air Jordans of 1642 (this is of course a feedback loop in itself - Lord and Lady Alabaster Buntcake have a beautiful tulip garden - we'll follow suit or be a laughingstock at the autumn colly-wad-hollow regalia! Lord and Lady Filigree DuPont follow suit, paying a bit more, who of course are not to outdo the Viscount of Hamslice and his ample-bosomed mistress, who will gladly pay double what those charlatans can afford. People's opinions do, of course, affect one another's).
So anyway, almost perversely, from a Smithian supply and demand sort of point of view, people start buying the crap out of tulips. The higher the price, the more buyers. This is of course partially because tulips are all the rage, and the regalia is in three weeks. But most of the buyers are now speculators: they have seen the dramatic rise in tulip price over the last few weeks, and they want a piece of the action. They're looking to unload the tulips at an inflated price and make a fast buck. Danger! Add to this the fact that the Turks have figured out something's up by my ten-fold increase in requests, plus the cropping up of Tulip Time and Tulips On Your Face, the new chains that are driving me out of business by buying ten times my volume at half the price, plus the King of Holland suddenly whipping the citizenry into a fervor over how Turks treat their prisoners and swearing that the imminent war is not over the tulips, and the next thing you know the Turks raise the price. Tulipmania is out of control.
This bubble bursts when people realize you can't eat tulips (for much nutritional value, anyway), when they are no longer the rage, or for whatever reason people start to speculate that it's time to jump ship. The last people to sell before the bubble bursts will be the most fortunate. But it will invariably be the case that people realize the stupid flowers are way overvalued, and a large bunch of folks will be left holding the bag. The Tulip Bag, which cost them 50,000 florins, and for which they now can't get 500. They're just flowers, after all. They pretty much grow on trees. Stems, anyway.
What would have happened without feedback? Let's remove the causes. Each seller now has no idea how much people would pay for flowers, so he charges what it cost him to pick the things, plus enough for him to pay the mortgage and buy the cow. Making less would make necessary a new line of work. Making more would, I think, require some idea that people will pay more. (Of course this is the case. I'm not saying it's possible for feedback not to exist in an economy; I'm just trying to describe this absurd scenario.) Buyers no longer have human envy and competitiveness, which cause fashion rages, so they will only buy tulips if they really want a tulip. There is no record-keeping of prices, so no speculation about future prices, so no investors buy tulips who really just want money. It's almost as though we're playing a game at the county fair and the prettiest tulip in the world is at stake. Everybody has to guess the price, and whoever comes closest wins the tulip. Nobody says anything to anybody else; we all just quietly pray that we know the value of a florin, and the value of a flower.
Conversely, what if the jelly bean game were to be infected by feedback? What if, instead of looking at the jar and trying to do some math, we could only look at our immediate neighbors' guesses at how many jelly beans were in the jar? We might well start off with reasonable guesses based on regular-sized jelly beans and jars. Guesses would range in the low hundreds. But this is a potentially volatile system; a small perturbation could send it spinning off to God-knows-where. What if one guy discovers the jelly bean keeper has stock in the industrial-sized jar market? Estimations could go to tens of thousands of jelly beans. What if one guy takes the The Price Is Right route and guesses one bean, hoping everybody else overbids? His neighbors will call him out and wager 15 beans or so, and now just the opposite is going on, and the bean number is driven down. Madness ensues, all because of feedback. How Nature Works by Per Bak is a good related read.
So two questions are a) which scenario, if either, dominates in actual human crowds; and b) what are the consequences of madness? There might be advantages. Our economy is broad and complicated and not everybody is a farmer or some other worker who provides some necessity. And as long as everybody is working and living, isn't that all that matters? (Of course, it isn't the case.) But everybody has different ideas about what the current housing bubble is going to do, and how bad the pop will be, and was it possibly created intentionally by the Federal Reserve to keep us out of recession, &c. I'm personally fascinated by this kind of stuff, and if you're not, you're probably not reading this post to this point, so I imagine you are, too. I plan on addressing these questions, with my unending cavalcade of expertise, in future posts.
Are crowds wise or mad? I think feedback is a key factor. According to Wikipedia, one of Surowiecki's necessary conditions for crowds to display wisdom is that one crowd member's opinions don't affect anyone else's, nor are they affected by others'. If this is the case, the common wisdom of the people is likely to outweigh the ignorance of the individual. I read an example where townspeople tried to guess the weight of a cow. I imagine this is pretty much the ideal scenario for crowd wisdom. Nobody talks to each other about it (I assume there is a prize at stake); everyone just takes a stab. It would be like everyone in town throwing a dart at a dartboard; the average throw would quite probably be exactly in the bull's-eye (given enough contestants). Or like that game where you try to guess how many jelly beans there are in the jar. Ask 5,000 adults and average their guesses - that will tell you.
But what about when people's opinions do affect one another's? Mackay describes the craze over tulips in Europe in the 1600s. It is regarded as an early description of an economic bubble, and it is just what happens when feedback runs rampant in crowd behavior. I am a tulip importer (really!) and I understand that tulips are all the rage in Holland's upper class. The Turks I buy them from don't speak the language, so they don't know to raise the price, so it's my opportunity to make a killing. I can still buy them from the Turks at 20 florins a bushel, but my selling price goes from 30 florins to 100. People will pay this - these are the Air Jordans of 1642 (this is of course a feedback loop in itself - Lord and Lady Alabaster Buntcake have a beautiful tulip garden - we'll follow suit or be a laughingstock at the autumn colly-wad-hollow regalia! Lord and Lady Filigree DuPont follow suit, paying a bit more, who of course are not to outdo the Viscount of Hamslice and his ample-bosomed mistress, who will gladly pay double what those charlatans can afford. People's opinions do, of course, affect one another's).
So anyway, almost perversely, from a Smithian supply and demand sort of point of view, people start buying the crap out of tulips. The higher the price, the more buyers. This is of course partially because tulips are all the rage, and the regalia is in three weeks. But most of the buyers are now speculators: they have seen the dramatic rise in tulip price over the last few weeks, and they want a piece of the action. They're looking to unload the tulips at an inflated price and make a fast buck. Danger! Add to this the fact that the Turks have figured out something's up by my ten-fold increase in requests, plus the cropping up of Tulip Time and Tulips On Your Face, the new chains that are driving me out of business by buying ten times my volume at half the price, plus the King of Holland suddenly whipping the citizenry into a fervor over how Turks treat their prisoners and swearing that the imminent war is not over the tulips, and the next thing you know the Turks raise the price. Tulipmania is out of control.
This bubble bursts when people realize you can't eat tulips (for much nutritional value, anyway), when they are no longer the rage, or for whatever reason people start to speculate that it's time to jump ship. The last people to sell before the bubble bursts will be the most fortunate. But it will invariably be the case that people realize the stupid flowers are way overvalued, and a large bunch of folks will be left holding the bag. The Tulip Bag, which cost them 50,000 florins, and for which they now can't get 500. They're just flowers, after all. They pretty much grow on trees. Stems, anyway.
What would have happened without feedback? Let's remove the causes. Each seller now has no idea how much people would pay for flowers, so he charges what it cost him to pick the things, plus enough for him to pay the mortgage and buy the cow. Making less would make necessary a new line of work. Making more would, I think, require some idea that people will pay more. (Of course this is the case. I'm not saying it's possible for feedback not to exist in an economy; I'm just trying to describe this absurd scenario.) Buyers no longer have human envy and competitiveness, which cause fashion rages, so they will only buy tulips if they really want a tulip. There is no record-keeping of prices, so no speculation about future prices, so no investors buy tulips who really just want money. It's almost as though we're playing a game at the county fair and the prettiest tulip in the world is at stake. Everybody has to guess the price, and whoever comes closest wins the tulip. Nobody says anything to anybody else; we all just quietly pray that we know the value of a florin, and the value of a flower.
Conversely, what if the jelly bean game were to be infected by feedback? What if, instead of looking at the jar and trying to do some math, we could only look at our immediate neighbors' guesses at how many jelly beans were in the jar? We might well start off with reasonable guesses based on regular-sized jelly beans and jars. Guesses would range in the low hundreds. But this is a potentially volatile system; a small perturbation could send it spinning off to God-knows-where. What if one guy discovers the jelly bean keeper has stock in the industrial-sized jar market? Estimations could go to tens of thousands of jelly beans. What if one guy takes the The Price Is Right route and guesses one bean, hoping everybody else overbids? His neighbors will call him out and wager 15 beans or so, and now just the opposite is going on, and the bean number is driven down. Madness ensues, all because of feedback. How Nature Works by Per Bak is a good related read.
So two questions are a) which scenario, if either, dominates in actual human crowds; and b) what are the consequences of madness? There might be advantages. Our economy is broad and complicated and not everybody is a farmer or some other worker who provides some necessity. And as long as everybody is working and living, isn't that all that matters? (Of course, it isn't the case.) But everybody has different ideas about what the current housing bubble is going to do, and how bad the pop will be, and was it possibly created intentionally by the Federal Reserve to keep us out of recession, &c. I'm personally fascinated by this kind of stuff, and if you're not, you're probably not reading this post to this point, so I imagine you are, too. I plan on addressing these questions, with my unending cavalcade of expertise, in future posts.
posted by john at
10:42 AM
1 Comments:
Damn dude, and I thought I was long-winded.
Post a Comment
Subscribe to Post Comments [Atom]
<< Home