6 ms·

Why asset bubbles are a part of the human condition that regulation can’t cure

44 points by robg 18y ago

helveticaman 18y ago
I would like to point out that Chile has less regulation than US and fewer bubbles.
- mdasen 18y ago
  So, what is a bubble? I would define a bubble as when we get irrationally exuberant just to come crashing down when reality strikes. Such a bubble could be measured by looking at the up and down of the GDP growth rate. Does that sound fair? The WorldBank actually keeps track of a lot of economic data. So, what does data say? Well, the United States (in the period of 1961-2007) has had fewer bubbles and those bubbles have been smaller. When counting bubbles, I used this methodology: a bubble is a decline in the growth rate of at least 3.5%. When a decline is continuous without a growth of at least 1% in between, it is considered part of the same bubble (so that I wasn't counting several years that Chile had large declines in a row as multiple bubbles). Under that assumption, the US has seen 6 bubbles. Chile, on the other hand, has seen 8. Move that assumption to 4% and the US has seen 5 and Chile has still seen 8. Moving the assumption up doesn't help Chile. Moving that assumption to 2% and the US has seen 7 and Chile has seen 10. Moving the assumption down doesn't help Chile. The facts show that Chile has more bubbles and the bubbles are larger in size. The US bubbles average at around 5% while the Chilean bubbles average around 10%. So, Chile has more bubbles and their bubbles are much greater than the US'. You can see it in a cool chart: http://mda-cxxtx.posterous.com/ http://mda-cxxtx.posterous.com/ A basic look at the chart and you can see that Chile fluctuates a lot more than the US.
  - nradov 18y ago
    GDP isn't a useful way to analyze bubbles. Bubbles are asset price inflation significantly above the trend line.
jerf 18y ago
Traditional differential equations is not one of the more useful college math classes(star), but a few useful things can be extracted from it. One of the most interesting is that under certain circumstances, oscillation is inevitable, by the very nature of how the system reacts to its own changes. A classic example is the well-known "simple harmonic oscillation", where the only solution that doesn't oscillate is to start right at 0. The market is more complicated. If simple harmonic motion is caused by negative feedback (the further you get from the origin, the stronger the force pulling you back), market instability is caused by positive feedback and the fact that there is a cap in how far bubbles can inflate. Excitement engenders excitement, so bubbles inflate. Depression engenders depression, so contractions also tend to overshoot vs. the "true" state of the economy. Stripping this effect out of the economy would take more than just making it "smarter", you'd have to rewrite the whole foundation of it. Mitigating recessions (and depressions) may be possible. Trying to stop them just makes them hit harder. ((star): DiffieQs are of course themselves useful; it is the class itself that is considered a bit of a waste by professors. Closed form solutions to diffieqs are the exception, not the rule, and mathematicians in general are not very interested in further pursuing the problem.)
pragmatic 18y ago
The author makes the mistake of equating efficient markets with rational markets. The market is not rational (whatever that means) but rather efficient. All known information (whether it is correct or not) is reflected in the price of a security. Rational markets lead us to the communist/central planning trap. If the market can be shown to not be "rational" the government can assume control (nationalize) of the market. Does the weather/climate appear "rational" on a large scale? No of course not. It consists of the local weather at each measurable point on the planet. Assuming we can make systems "rational" is plain foolish. See "The Systems Bible" by John Gall.
- create_account 18y ago
  The market is not rational (whatever that means) but rather efficient. Oh boy. Careful to whom you say that.
- steveplace 18y ago
  All known information (whether it is correct or not) is reflected in the price of a security. Eventually, yes. But not intermediate to short term. EMT removes subjectivity of information. People perceive (or fail to) certain data differently than others. The market is a collection of that perception combined with uncertainty. To ignore that fact is not practical over the short term. Evidence: see the premarket futures today.
zandorg 18y ago
I haven't read the article, but Isaac Newton once fell prey to the South Sea bubble, showing that even geniuses get hit.
- mynameishere 18y ago
  I haven't read your comment, but there's no way our shirt cured AIDS.
jderick 18y ago
Bubbles may be inevitable, but that doesn't mean we can't do anything to mitigate them. To me, it seems clear that the "Trader's option" -- big bonuses in good years, no downside in bad -- is a recipe for disaster in any market. I doubt this bubble would have been nearly as bad if the banks were still owned primarily by their management.
- Tichy 18y ago
  What I don't get is why should we expect the government to do a better job at managing banks than the banks themselves? If "trader's options" are bad, why can't the banks figure it out for themselves?
  - mattobrien 18y ago
    Because trader's options are good for bankers, but not good for the public. These options create incentives for bankers to take large risks to maximize short term profits, and hence year end bonuses. But if risky trades blow up, a banker merely loses his job - not exactly nontrivial, but certainly mitigated by getting to keep bonuses from past years. And if enough risky trades blow up within a firm, sending it into insolvency, it can always get government bailouts because of the systemic risk its failure would produce. Essentially, gains are privatized and losses are socialized. This is why the government needs to restructure bank pay incentives, with a key probably being making banks privately held again, as well as reducing leverage and regulating the system better (or at all). Granted, some firms, like Morgan Stanley and UBS, are beginning to experiment with clawback provisions, that have longer 3 year horizons for bonuses. But until we reform the system, this won't be in banks', or bankers', interests, except from a PR angle, which clearly matters. And government pressure can help us get there faster.
    - Tichy 18y ago
      Good for bankers, yes, but not good for banks, or are they? And banks employ bankers. Remembering Joel's writings on performance measurements, I am simply skeptic that there can be any remedy, short of people watching out for themselves. I suspect even with longer windows for bonuses, there will be ways to game the system.
mattmaroon 18y ago
"These lab results should give pause not only to people who believe in efficient markets..." The author is mistaking why markets trend toward efficiency. The 10% who don't overvalue the equities will have enough money (if not at first, but later as their realism makes them much more profit) to keep the price in line. Just because 90% of people overvalue something doesn't mean the markets can't be efficient.
- muerdeme 18y ago
  "Bubbles start to pop when the momentum traders run out of money and can no longer push prices up." She states credible evidence that the market is efficient, then claims that the evidence undermines those who believe in efficient markets... Weak.
  - stcredzero 18y ago
    Why asset bubbles are a part of the human condition that regulation can’t cure "Human Condition" is sometimes a code word for something people don't want you to fix. There are those who have the means to profit from the large scale business cycle certainly want you to think that it has to be this way. They want the forced selling so that they can profit "on the buy." There are those out there who want to blow up bubbles and ride valuations up, then get out before the fall. Death is inevitable. The inevitability of taxes is propaganda.
    - DLWormwood 18y ago
      > "Human Condition" is sometimes a code word for something people don't want you to fix. snip Death is inevitable. The inevitability of taxes is propaganda. Dang, that's kind of profound. Are you quoting somebody?
      - stcredzero 18y ago
        Nope. I wrote that.
- time_management 18y ago
time_management 18y ago
Great article. I wonder how the experiments would have turned out if people were allowed to short-sell. My guess is that the bubbles and crashes would be a lot less common. When people can only go long, there's a desire to buy in order to be "participating", and there's also the effect of the Winner's Curse. One interesting result I read about in Aaron Brown's The Poker Face of Wall Street involved a mock security that would mature either to $0 or $20, with a 50% chance of each. The participants would be given index cards; half of the cards had the security's correct value, and half were blank. The price would start around $10 (since those who have information don't want to show it by bidding $19/$1) but rapidly converge to the fair value. When the experiment was done with everyone receiving blank cards (but no one knowing that all cards were blank) trading would start, as usual, right around $10. In many cases, it would converge to $0 or $20 as the market "figured out" the fair value.
- eru 18y ago
  Does anybody care to set up the experiment with short-selling enabled? Would be interesting to watch.
Tichy 18y ago
Is it possible that the expected gain is not the proper measurement for the value of the shares? For example, the expected value of the national lottery is negative (-50% in Germany), but I can understand people who play nevertheless: if you lose, you only lose 1€, but if you win, you can win a million €. Similar for the experiment, maybe the chance on the 90cents yield was enough to warrant higher prices? It is simply gambling? Phrased differently (using example from other comment): if you had to choose between 10$ and a 50% chance to get 0$ or 20$, which would you chose? Depending on your circumstances, either option might be more valuable to you? If you are about to starve in the next minute, you'd probably take the 10$ (50% chance of dieing is too much), but if you already have enough money, why not take a 50% shot at earning 10$? What if the choice is between getting 1 million $ and a 50% chance between 0 and 2 million $? I definitely don't think the two options have the same value, even though the expected outcome is the same. Spontaneously I would take the 1 million, otoh I have read somewhere that in TV game shows, participants routinely gamble away such sums (the kind of game show where you can either quit and take the money you already won, or put it all at stake for the next question, with a chance to double it). I wonder, in the experiment of the article, if prices rose even above the maximum possible value of the shares?
- schtog 18y ago
  Well they are not the same because your first million has a lot more utility than the second for most people, unless say you need 2million to start he company you always dreamed of starting or to do a surgery in a foreign country on your child that is superexpensive and if not the child will die.
  - Tichy 18y ago
    Yes that is what I mean - it seems "expected value" is not equal to real value.
DLWormwood 18y ago
Wow. Given the common complaint here and at other news sites about economics lacking scientific vigor, I assumed that it was because it was impossible to experimentally isolate human economic behavior. Not only does this article point out one way to do just that, but that it even has a "stereotypical" base experiment used for decades that future scenarios can be derived from. I need to bookmark this article as rebuttal or "ward against trolls" in the future...
- eru 18y ago
  Our university even has its own lab for 'experimental economics', where they do nothing else but run such games.

44 points by robg 18y ago

helveticaman 18y ago

I would like to point out that Chile has less regulation than US and fewer bubbles.

mdasen 18y ago
So, what is a bubble? I would define a bubble as when we get irrationally exuberant just to come crashing down when reality strikes. Such a bubble could be measured by looking at the up and down of the GDP growth rate. Does that sound fair? The WorldBank actually keeps track of a lot of economic data. So, what does data say? Well, the United States (in the period of 1961-2007) has had fewer bubbles and those bubbles have been smaller. When counting bubbles, I used this methodology: a bubble is a decline in the growth rate of at least 3.5%. When a decline is continuous without a growth of at least 1% in between, it is considered part of the same bubble (so that I wasn't counting several years that Chile had large declines in a row as multiple bubbles). Under that assumption, the US has seen 6 bubbles. Chile, on the other hand, has seen 8. Move that assumption to 4% and the US has seen 5 and Chile has still seen 8. Moving the assumption up doesn't help Chile. Moving that assumption to 2% and the US has seen 7 and Chile has seen 10. Moving the assumption down doesn't help Chile. The facts show that Chile has more bubbles and the bubbles are larger in size. The US bubbles average at around 5% while the Chilean bubbles average around 10%. So, Chile has more bubbles and their bubbles are much greater than the US'. You can see it in a cool chart: http://mda-cxxtx.posterous.com/ http://mda-cxxtx.posterous.com/ A basic look at the chart and you can see that Chile fluctuates a lot more than the US.
- nradov 18y ago
  GDP isn't a useful way to analyze bubbles. Bubbles are asset price inflation significantly above the trend line.

jerf 18y ago

Traditional differential equations is not one of the more useful college math classes(star), but a few useful things can be extracted from it. One of the most interesting is that under certain circumstances, oscillation is inevitable, by the very nature of how the system reacts to its own changes. A classic example is the well-known "simple harmonic oscillation", where the only solution that doesn't oscillate is to start right at 0. The market is more complicated. If simple harmonic motion is caused by negative feedback (the further you get from the origin, the stronger the force pulling you back), market instability is caused by positive feedback and the fact that there is a cap in how far bubbles can inflate. Excitement engenders excitement, so bubbles inflate. Depression engenders depression, so contractions also tend to overshoot vs. the "true" state of the economy. Stripping this effect out of the economy would take more than just making it "smarter", you'd have to rewrite the whole foundation of it. Mitigating recessions (and depressions) may be possible. Trying to stop them just makes them hit harder. ((star): DiffieQs are of course themselves useful; it is the class itself that is considered a bit of a waste by professors. Closed form solutions to diffieqs are the exception, not the rule, and mathematicians in general are not very interested in further pursuing the problem.)

pragmatic 18y ago

The author makes the mistake of equating efficient markets with rational markets. The market is not rational (whatever that means) but rather efficient. All known information (whether it is correct or not) is reflected in the price of a security. Rational markets lead us to the communist/central planning trap. If the market can be shown to not be "rational" the government can assume control (nationalize) of the market. Does the weather/climate appear "rational" on a large scale? No of course not. It consists of the local weather at each measurable point on the planet. Assuming we can make systems "rational" is plain foolish. See "The Systems Bible" by John Gall.

create_account 18y ago
The market is not rational (whatever that means) but rather efficient. Oh boy. Careful to whom you say that.
steveplace 18y ago
All known information (whether it is correct or not) is reflected in the price of a security. Eventually, yes. But not intermediate to short term. EMT removes subjectivity of information. People perceive (or fail to) certain data differently than others. The market is a collection of that perception combined with uncertainty. To ignore that fact is not practical over the short term. Evidence: see the premarket futures today.

zandorg 18y ago

I haven't read the article, but Isaac Newton once fell prey to the South Sea bubble, showing that even geniuses get hit.

mynameishere 18y ago
I haven't read your comment, but there's no way our shirt cured AIDS.

jderick 18y ago

Bubbles may be inevitable, but that doesn't mean we can't do anything to mitigate them. To me, it seems clear that the "Trader's option" -- big bonuses in good years, no downside in bad -- is a recipe for disaster in any market. I doubt this bubble would have been nearly as bad if the banks were still owned primarily by their management.

Tichy 18y ago
What I don't get is why should we expect the government to do a better job at managing banks than the banks themselves? If "trader's options" are bad, why can't the banks figure it out for themselves?
- mattobrien 18y ago
  Because trader's options are good for bankers, but not good for the public. These options create incentives for bankers to take large risks to maximize short term profits, and hence year end bonuses. But if risky trades blow up, a banker merely loses his job - not exactly nontrivial, but certainly mitigated by getting to keep bonuses from past years. And if enough risky trades blow up within a firm, sending it into insolvency, it can always get government bailouts because of the systemic risk its failure would produce. Essentially, gains are privatized and losses are socialized. This is why the government needs to restructure bank pay incentives, with a key probably being making banks privately held again, as well as reducing leverage and regulating the system better (or at all). Granted, some firms, like Morgan Stanley and UBS, are beginning to experiment with clawback provisions, that have longer 3 year horizons for bonuses. But until we reform the system, this won't be in banks', or bankers', interests, except from a PR angle, which clearly matters. And government pressure can help us get there faster.
  - Tichy 18y ago
    Good for bankers, yes, but not good for banks, or are they? And banks employ bankers. Remembering Joel's writings on performance measurements, I am simply skeptic that there can be any remedy, short of people watching out for themselves. I suspect even with longer windows for bonuses, there will be ways to game the system.

mattmaroon 18y ago

"These lab results should give pause not only to people who believe in efficient markets..." The author is mistaking why markets trend toward efficiency. The 10% who don't overvalue the equities will have enough money (if not at first, but later as their realism makes them much more profit) to keep the price in line. Just because 90% of people overvalue something doesn't mean the markets can't be efficient.

muerdeme 18y ago
"Bubbles start to pop when the momentum traders run out of money and can no longer push prices up." She states credible evidence that the market is efficient, then claims that the evidence undermines those who believe in efficient markets... Weak.
- stcredzero 18y ago
  Why asset bubbles are a part of the human condition that regulation can’t cure "Human Condition" is sometimes a code word for something people don't want you to fix. There are those who have the means to profit from the large scale business cycle certainly want you to think that it has to be this way. They want the forced selling so that they can profit "on the buy." There are those out there who want to blow up bubbles and ride valuations up, then get out before the fall. Death is inevitable. The inevitability of taxes is propaganda.
  - DLWormwood 18y ago
    > "Human Condition" is sometimes a code word for something people don't want you to fix. snip Death is inevitable. The inevitability of taxes is propaganda. Dang, that's kind of profound. Are you quoting somebody?
    - stcredzero 18y ago
      Nope. I wrote that.
time_management 18y ago

time_management 18y ago

Great article. I wonder how the experiments would have turned out if people were allowed to short-sell. My guess is that the bubbles and crashes would be a lot less common. When people can only go long, there's a desire to buy in order to be "participating", and there's also the effect of the Winner's Curse. One interesting result I read about in Aaron Brown's The Poker Face of Wall Street involved a mock security that would mature either to $0 or $20, with a 50% chance of each. The participants would be given index cards; half of the cards had the security's correct value, and half were blank. The price would start around $10 (since those who have information don't want to show it by bidding $19/$1) but rapidly converge to the fair value. When the experiment was done with everyone receiving blank cards (but no one knowing that all cards were blank) trading would start, as usual, right around $10. In many cases, it would converge to $0 or $20 as the market "figured out" the fair value.

eru 18y ago
Does anybody care to set up the experiment with short-selling enabled? Would be interesting to watch.

Tichy 18y ago

Is it possible that the expected gain is not the proper measurement for the value of the shares? For example, the expected value of the national lottery is negative (-50% in Germany), but I can understand people who play nevertheless: if you lose, you only lose 1€, but if you win, you can win a million €. Similar for the experiment, maybe the chance on the 90cents yield was enough to warrant higher prices? It is simply gambling? Phrased differently (using example from other comment): if you had to choose between 10$ and a 50% chance to get 0$ or 20$, which would you chose? Depending on your circumstances, either option might be more valuable to you? If you are about to starve in the next minute, you'd probably take the 10$ (50% chance of dieing is too much), but if you already have enough money, why not take a 50% shot at earning 10$? What if the choice is between getting 1 million $ and a 50% chance between 0 and 2 million $? I definitely don't think the two options have the same value, even though the expected outcome is the same. Spontaneously I would take the 1 million, otoh I have read somewhere that in TV game shows, participants routinely gamble away such sums (the kind of game show where you can either quit and take the money you already won, or put it all at stake for the next question, with a chance to double it). I wonder, in the experiment of the article, if prices rose even above the maximum possible value of the shares?

schtog 18y ago
Well they are not the same because your first million has a lot more utility than the second for most people, unless say you need 2million to start he company you always dreamed of starting or to do a surgery in a foreign country on your child that is superexpensive and if not the child will die.
- Tichy 18y ago
  Yes that is what I mean - it seems "expected value" is not equal to real value.

DLWormwood 18y ago

Wow. Given the common complaint here and at other news sites about economics lacking scientific vigor, I assumed that it was because it was impossible to experimentally isolate human economic behavior. Not only does this article point out one way to do just that, but that it even has a "stereotypical" base experiment used for decades that future scenarios can be derived from. I need to bookmark this article as rebuttal or "ward against trolls" in the future...

eru 18y ago
Our university even has its own lab for 'experimental economics', where they do nothing else but run such games.