Review of "Mind over money"
Just watched a PBS Nova documentary called "Mind over money". It's a film about two competing models of economics, referred to in the show as 'rationalist economics' and 'behavioralist economics'. 'Rationalists' model each human being as being self interested and perfectly rational in pursuing the maximization of their own wealth; this results in certain mathematical assumptions that allow for an elegant theory. 'Behavioralists' claim that human beings do not act that way; even if maximization of personal wealth were pursued (which is not always the case), mistakes are often made.
I for one believe that the behavioralists are right. It seems common sense to me to assume that nobody can reliably do the kind of mental calculation required by perfect economic rationality, and there are various cognitive biases to skew things the wrong way. But, I stress, this is merely a belief and I will defer to experts that study these things. In any case, the rationalist economic model seems to work, most of the time, because most people, most of the time, behave close to what that model assumes. That's why, often, it can make useful predictions (though, when its predictions fail, they fail spectacularly).
The PBS Nova piece also tries to convince the viewer that the behavioralists are right. So, how well do they achieve that goal? Keep in mind, I'm part of the choir here.
First piece of experimental evidence is an auction, with a twist. The item being auctioned is a 20$ bill, which is exactly that. Nothing special about it. Not worth more than 20$. Now, the bill will be given to the highest bidder, however- the twist- the second highest bidder also has to pay, and will get nothing for the trouble.
The auction happens, and surprise surprise, two people eventually enter a bidding war above 20$. Rationality failed, claims Nova, because if they were rational they wouldn't have bid more than 20$. It was the 'fear' of losing that brings the bids above 20$.
If rationality failed, I'd say it failed when entering the auction, if at all. The auction is a trick, and one needs to look several steps ahead to realize the trap and avoid it. So, yeah, I can see why you'd expect a rational agent not to commit that blunder. But the 'fear' of losing is a rational fear, once the trap's been sprung. Once a bidder enters the above 20$ zone, they must, rationally, try and minimize their loss. If said loss involves bidding higher, at least part of the high bid is offset by the 20$ bill. If they lose, none of the bid is gained back.
Second piece is a little poll meant to illustrate 'present bias'. Suppose you are given an option: either receive 100$ in a year, or 102$ in a year and a day. Most people choose the 102$ in a year and a day, which is deemed consistent with rationalist economics.
Now, let's say we change the option: either 100$ now, or 102$ tomorrow. What's your choice? Most people choose 100$ today. This illustrates the 'present bias'- we want our rewards now, now, now!
The problem I see here is that, while I'm already convinced present bias exists, and also that the little poll shows it in action, I'm still not convinced it's irrational. What if I -do- need 100$ (or less) right now? What if having those dollars now opens a good prospect of earning more later, but if I were to wait one day to get those extra 2$, I lose my chance?
This also reveals that 102$ after a year and a day is only 'better' than 100$ in a year because I have no idea about what my situation is after that hypothetical year, so might as well pick the 102$. But I know my situation today. It may well be such that I need the money, fast. {EDIT, 11th of June 2013: oh wow. Now that was a hell of a brainfart on my part. Even I'm having trouble understanding what I might have thought when writing that about the present bias illustration. Maybe that's because it makes no sense. Oh well. At least some of the time, I'm stupid. Hopefully just some of the time.}
The anchor effect though was neatly presented, even for my standards. Here's the setup- again an auction for some product, say, a wine bottle. But, before bidding starts, the participants are asked to write down the last two digits of their social security number on a sheet of paper. It turns out that those who wrote the larger numbers, also tended to make the higher bids.
Incidentally, the documentary claims (and lacking a reason not to, I'll believe it) that the social security number is random. That means, it is not associated with any trait of its owner. Having a social security number with a high pair of last digits reveals nothing about the person holding it. Even so, just writing a high number down tends to push people to bid higher.
For my money, the explanation of economic bubbles is spot on, and the example of the Dutch tulip bulb bubble is a neat piece of history. It went like this- in 17th century Holland, people started buying tulip bulbs at a high price because, since prices were increasing, one could expect to bump into a bigger fool that would pay even more. Same thing is supposed to drive other speculative bubbles, like what happened in the housing markets. And in all cases, the bubble eventually bursts because one runs out of fools to trick.
A brief look at the wikipedia article on tulip mania paints a more complex picture however, which may indicate that the Nova doc is a bit hasty in its conclusions. For one, it's difficult to ascertain just how high the prices were, apparently, and also there may have been some kind of safeguards in place that were meant to reduce loss- or so pro-speculative historians claim. What can I say, it's a very controversial issue, where people have actual stakes in. The worst kind of controversies, therefore. Only uber nerds were ever personally invested in whether light is a wave or a stream of particles. Attack someone's means of making money though, or their beliefs, and -then- you have a fight on your hands.
Back to experimental stuff. The little experiment is, a group of students is asked, how much would they pay in order to buy a travel mug. Each student writes down an answer, and the average is 6$. Some of them are given (free of charge) a travel mug identical to the one put on display, and after an hour they are asked how much are they willing to sell it for. The average turns out to be 9$.
This is supposed to show that the 'pleasure of owning something' for one hour increases its price, whereas, if the subjects were rational, they'd be willing to sell it for the same price.
I'm thinking that some deeper care must be taken here. I don't doubt that the students would want to sell the mug for as much as they can. And I presume the question they were asked went something like, 'what is the absolute lowest price that you'd be willing to sell this mug for (lower than this, you will reject the deal)?' I still think though that, if they were to actually sell it, they'd be happy to get any amount for it. After all, they got it free of charge, and if they met someone driving a hard enough bargain, they might still settle for less than they claimed.
More convincing (but I have a serious caveat here) is the experiment where subjects are asked to say how much they would pay for a water bottle. Some are in a 'neutral' state, some have been primed to be at 'low level sadness' through some method. It turns out, or so I assume, that the sad group would be willing to pay much more.
In the doc, they only have time to compare and contrast two cases. Neutral person says 1$, Sad person says 10$. Boom, effect. Obviously that's not how the study was conducted. Obviously two large groups must have been compared, and some statistical analysis deployed to show how dissimilar the responses were. It's a pity no time was found in the doc to present that (to their credit, they indicate that one can find more info on their website).
Oh, an observation. I'm willing to believe that if someone claims they'll pay 10$, then they'll pay 10$ or more. You've noticed that if someone claims that they'll sell for 9$ or more, I'll believe they might still settle for less. In other words, I believe that people will settle for worse deals than they claim they will.
Vernon Smith's experiment is an interesting piece that I'll need to investigate more. It's, again, only sketched in the doc, but as far as I can tell it pits a class of college students against each other, each student being a stock trader. There's real money in it for the winner, so there's an incentive to play well. The thing they are trading is something that declines in 'real' value (maybe some perishable commodity?), and the initial prices are too cheap. The product is undervalued. Apparently, the collective behavior of the profit seeking traders drives the prices up, until they plateau way above the declining 'real' value of the thing being traded. Eventually, as the thing reaches 0 real value, prices drop- the bubble bursts.
Why I like this in principle and think it deserves more attention is that it's a clear, simple, simulation of a bubble. This is not a historic case like the Dutch tulip mania where one can wave hands and say we don't have enough evidence to judge. It's a controlled experiment, where some factors are emphasized to illustrate their effects.
I do wish more time was allocated to presenting the method though. What sample size of subjects was used, how were the bubbles- if they appeared- analyzed later?
The documentary closes with a brief recap of the 2008 crash, which is presented as a case of the rationalist theory failing. I think they leap a bit far, but only because as far as my limited knowledge allows me to see, there is no reason why even perfectly rational agents would create a stable, efficient, prosperous market. If the game is such that the rational choice is to stab everyone in the back, then the rationalist model predicts a lot of perforated people.
In particular, the documentary seems to imply that the 2008 crash was "nobody"'s fault. We just had the wrong model of the market, is all, we blindly assumed that it will take care of our prosperity on its own, we acted on emotional impulse and got fooled by ourselves.
I'm not above assuming that at least some parts of a bubble are driven by rational agents who are indeed selfishly- and efficiently- pursuing their own gain, longer term consequences to others be damned. So I'm not ill-disposed to explanations of the crash that go beyond systemic failure. It's pretty clear that the system failed because the magical safeguards it was supposed to generate failed to work. It's pretty clear that it was not resilient against manipulation in ill-faith.
So yeah, what to say on the whole documentary, as a conclusion? It won't convince you, if you're not part of the choir. It's very lean on information, for both sides of the debate. It's not just that the rationalist economists interviewed were edited into sounding like old codgers (there's at least some truth in what they say, but you wouldn't easily guess it by the edits), it's also that the arguments marshaled for the behavioralist explanation are barely sketched.
"At a mere 52 minute running time, what did you expect?" I may be asked. Well, for one, less tea-bagging with mathematical formulas. This trope needs to die, now. If you're not going to explain what that formula is, don't show it.
I for one believe that the behavioralists are right. It seems common sense to me to assume that nobody can reliably do the kind of mental calculation required by perfect economic rationality, and there are various cognitive biases to skew things the wrong way. But, I stress, this is merely a belief and I will defer to experts that study these things. In any case, the rationalist economic model seems to work, most of the time, because most people, most of the time, behave close to what that model assumes. That's why, often, it can make useful predictions (though, when its predictions fail, they fail spectacularly).
The PBS Nova piece also tries to convince the viewer that the behavioralists are right. So, how well do they achieve that goal? Keep in mind, I'm part of the choir here.
First piece of experimental evidence is an auction, with a twist. The item being auctioned is a 20$ bill, which is exactly that. Nothing special about it. Not worth more than 20$. Now, the bill will be given to the highest bidder, however- the twist- the second highest bidder also has to pay, and will get nothing for the trouble.
The auction happens, and surprise surprise, two people eventually enter a bidding war above 20$. Rationality failed, claims Nova, because if they were rational they wouldn't have bid more than 20$. It was the 'fear' of losing that brings the bids above 20$.
If rationality failed, I'd say it failed when entering the auction, if at all. The auction is a trick, and one needs to look several steps ahead to realize the trap and avoid it. So, yeah, I can see why you'd expect a rational agent not to commit that blunder. But the 'fear' of losing is a rational fear, once the trap's been sprung. Once a bidder enters the above 20$ zone, they must, rationally, try and minimize their loss. If said loss involves bidding higher, at least part of the high bid is offset by the 20$ bill. If they lose, none of the bid is gained back.
Second piece is a little poll meant to illustrate 'present bias'. Suppose you are given an option: either receive 100$ in a year, or 102$ in a year and a day. Most people choose the 102$ in a year and a day, which is deemed consistent with rationalist economics.
Now, let's say we change the option: either 100$ now, or 102$ tomorrow. What's your choice? Most people choose 100$ today. This illustrates the 'present bias'- we want our rewards now, now, now!
The problem I see here is that, while I'm already convinced present bias exists, and also that the little poll shows it in action, I'm still not convinced it's irrational. What if I -do- need 100$ (or less) right now? What if having those dollars now opens a good prospect of earning more later, but if I were to wait one day to get those extra 2$, I lose my chance?
This also reveals that 102$ after a year and a day is only 'better' than 100$ in a year because I have no idea about what my situation is after that hypothetical year, so might as well pick the 102$. But I know my situation today. It may well be such that I need the money, fast. {EDIT, 11th of June 2013: oh wow. Now that was a hell of a brainfart on my part. Even I'm having trouble understanding what I might have thought when writing that about the present bias illustration. Maybe that's because it makes no sense. Oh well. At least some of the time, I'm stupid. Hopefully just some of the time.}
The anchor effect though was neatly presented, even for my standards. Here's the setup- again an auction for some product, say, a wine bottle. But, before bidding starts, the participants are asked to write down the last two digits of their social security number on a sheet of paper. It turns out that those who wrote the larger numbers, also tended to make the higher bids.
Incidentally, the documentary claims (and lacking a reason not to, I'll believe it) that the social security number is random. That means, it is not associated with any trait of its owner. Having a social security number with a high pair of last digits reveals nothing about the person holding it. Even so, just writing a high number down tends to push people to bid higher.
For my money, the explanation of economic bubbles is spot on, and the example of the Dutch tulip bulb bubble is a neat piece of history. It went like this- in 17th century Holland, people started buying tulip bulbs at a high price because, since prices were increasing, one could expect to bump into a bigger fool that would pay even more. Same thing is supposed to drive other speculative bubbles, like what happened in the housing markets. And in all cases, the bubble eventually bursts because one runs out of fools to trick.
A brief look at the wikipedia article on tulip mania paints a more complex picture however, which may indicate that the Nova doc is a bit hasty in its conclusions. For one, it's difficult to ascertain just how high the prices were, apparently, and also there may have been some kind of safeguards in place that were meant to reduce loss- or so pro-speculative historians claim. What can I say, it's a very controversial issue, where people have actual stakes in. The worst kind of controversies, therefore. Only uber nerds were ever personally invested in whether light is a wave or a stream of particles. Attack someone's means of making money though, or their beliefs, and -then- you have a fight on your hands.
Back to experimental stuff. The little experiment is, a group of students is asked, how much would they pay in order to buy a travel mug. Each student writes down an answer, and the average is 6$. Some of them are given (free of charge) a travel mug identical to the one put on display, and after an hour they are asked how much are they willing to sell it for. The average turns out to be 9$.
This is supposed to show that the 'pleasure of owning something' for one hour increases its price, whereas, if the subjects were rational, they'd be willing to sell it for the same price.
I'm thinking that some deeper care must be taken here. I don't doubt that the students would want to sell the mug for as much as they can. And I presume the question they were asked went something like, 'what is the absolute lowest price that you'd be willing to sell this mug for (lower than this, you will reject the deal)?' I still think though that, if they were to actually sell it, they'd be happy to get any amount for it. After all, they got it free of charge, and if they met someone driving a hard enough bargain, they might still settle for less than they claimed.
More convincing (but I have a serious caveat here) is the experiment where subjects are asked to say how much they would pay for a water bottle. Some are in a 'neutral' state, some have been primed to be at 'low level sadness' through some method. It turns out, or so I assume, that the sad group would be willing to pay much more.
In the doc, they only have time to compare and contrast two cases. Neutral person says 1$, Sad person says 10$. Boom, effect. Obviously that's not how the study was conducted. Obviously two large groups must have been compared, and some statistical analysis deployed to show how dissimilar the responses were. It's a pity no time was found in the doc to present that (to their credit, they indicate that one can find more info on their website).
Oh, an observation. I'm willing to believe that if someone claims they'll pay 10$, then they'll pay 10$ or more. You've noticed that if someone claims that they'll sell for 9$ or more, I'll believe they might still settle for less. In other words, I believe that people will settle for worse deals than they claim they will.
Vernon Smith's experiment is an interesting piece that I'll need to investigate more. It's, again, only sketched in the doc, but as far as I can tell it pits a class of college students against each other, each student being a stock trader. There's real money in it for the winner, so there's an incentive to play well. The thing they are trading is something that declines in 'real' value (maybe some perishable commodity?), and the initial prices are too cheap. The product is undervalued. Apparently, the collective behavior of the profit seeking traders drives the prices up, until they plateau way above the declining 'real' value of the thing being traded. Eventually, as the thing reaches 0 real value, prices drop- the bubble bursts.
Why I like this in principle and think it deserves more attention is that it's a clear, simple, simulation of a bubble. This is not a historic case like the Dutch tulip mania where one can wave hands and say we don't have enough evidence to judge. It's a controlled experiment, where some factors are emphasized to illustrate their effects.
I do wish more time was allocated to presenting the method though. What sample size of subjects was used, how were the bubbles- if they appeared- analyzed later?
The documentary closes with a brief recap of the 2008 crash, which is presented as a case of the rationalist theory failing. I think they leap a bit far, but only because as far as my limited knowledge allows me to see, there is no reason why even perfectly rational agents would create a stable, efficient, prosperous market. If the game is such that the rational choice is to stab everyone in the back, then the rationalist model predicts a lot of perforated people.
In particular, the documentary seems to imply that the 2008 crash was "nobody"'s fault. We just had the wrong model of the market, is all, we blindly assumed that it will take care of our prosperity on its own, we acted on emotional impulse and got fooled by ourselves.
I'm not above assuming that at least some parts of a bubble are driven by rational agents who are indeed selfishly- and efficiently- pursuing their own gain, longer term consequences to others be damned. So I'm not ill-disposed to explanations of the crash that go beyond systemic failure. It's pretty clear that the system failed because the magical safeguards it was supposed to generate failed to work. It's pretty clear that it was not resilient against manipulation in ill-faith.
So yeah, what to say on the whole documentary, as a conclusion? It won't convince you, if you're not part of the choir. It's very lean on information, for both sides of the debate. It's not just that the rationalist economists interviewed were edited into sounding like old codgers (there's at least some truth in what they say, but you wouldn't easily guess it by the edits), it's also that the arguments marshaled for the behavioralist explanation are barely sketched.
"At a mere 52 minute running time, what did you expect?" I may be asked. Well, for one, less tea-bagging with mathematical formulas. This trope needs to die, now. If you're not going to explain what that formula is, don't show it.
Comments
Post a Comment