Rational Herding and the Housing Bubble

Robert Shiller says information cascades can explain asset bubbles:

How a Bubble Stayed Under the Radar, by Robert Shiller, Economic View, NY Times: One great puzzle about the recent housing bubble is why even most experts didn’t recognize the bubble as it was forming. ...

Were all these people stupid? It can’t be. We have to consider the possibility that perfectly rational people can get caught up in a bubble. In this connection, it is helpful to refer to an important bit of economic theory about herd behavior.

Three economists, Sushil Bikhchandani, David Hirshleifer and Ivo Welch, in a classic 1992 article, defined what they call “information cascades” that can lead people into serious error. ...

Mr. Bikhchandani and his co-authors present this example: Suppose that a group of individuals must make an important decision, based on useful but incomplete information. Each one of them has ... information..., but the information is incomplete and “noisy” and does not always point to the right conclusion.

Let’s update the example...: The individuals in the group must each decide whether real estate is a terrific investment... Suppose that there is a 60 percent probability that any one person’s information will lead to the right decision. ...

Each person makes decisions individually, sequentially, and reveals ... decisions through actions — in this case, by entering the housing market and bidding up home prices.

Suppose houses are really of low investment value, but the first person to make a decision reaches the wrong conclusion (which happens, as we have assumed, 40 percent of the time). The first person, A, pays a high price for a home, thus signaling to others that houses are a good investment.

The second person, B, has no problem if his own data seem to confirm the information provided by A’s willingness to pay a high price. But B faces a quandary if his own information seems to contradict A’s judgment. In that case, B would conclude that he has no worthwhile information, and so he must make an arbitrary decision — say, by flipping a coin to decide whether to buy a house.

The result is that even if houses are of low investment value, we may now have two people who make purchasing decisions that reveal their conclusion that houses are a good investment.

As others make purchases at rising prices, more and more people will conclude that these buyers’ information about the market outweighs their own.

Mr. Bikhchandani and his co-authors worked out this rational herding story carefully, and their results show that the probability of the cascade leading to an incorrect assumption is 37 percent. ... Thus, we should expect to see cascades driving our thinking from time to time, even when everyone is absolutely rational and calculating.

This theory poses a major challenge to the “efficient markets” view of the world... The efficient-markets view holds that the market is wiser than any individual: in aggregate, the market will come to the correct decision. But the theory is flawed because it does not recognize that people must rely on the judgments of others. ...

It is clear that just such an information cascade helped to create the housing bubble. And it is now possible that a downward cascade will develop — in which rational individuals become excessively pessimistic as they see others bidding down home prices to abnormally low levels.

Here's another example of how this mechanism can work: Suppose you are in a long line of cars that are all headed to the same place. You have a vague, but not perfect idea of how to get to your destination. At one intersection, by chance, the first three cars get it wrong and turn left instead of right. The fourth car, though somewhat certain the correct way to go is to turn right, is not certain enough to go in a different direction and chooses to follow the first three cars. The driver assumes that since all three cars turned left, that must be the right way to go. As everyone begins to turn left, the signal strengthens and only those who are very certain of the true direction (which is a difficult condition to meet in asset markets) will choose to deviate from the group and go in the direction they know is correct. [For this to work, it is important that decisions be sequential, and that people can observe and learn from the decisions of others.]

We don't know for sure how or why bubbles occur, so this may or may not be the correct explanation. In general, it's hard for economists to admit they can occur at all since they do not appear to be rational responses to economic fundamentals and we do not have very good models of them, so this may explain some of the resistance to admitting a bubble is occurring at the time it is inflating. We prefer to explain market outcomes with economic models that we understand instead of with bubbles that occur for mysterious reasons. It is only after the fact, when it becomes clear that no fundamentals based interpretation seems possible, that we admit a bubble could be the correct explanation.  But maybe recent experience will change that and we won't be so quick to deny it when the next bubble begins to inflate, though I can't say I'm very confident in that prediction.

Comments

[esb]

How 'bout this ...

You are in a long line of cars which terminates at the edge of a cliff.

You look ahead and notice that the lead car suddenly plunges over the edge.

Then, the next car in the queue moves up and goes over.

Suddenly, you notice an off ramp immediately to your right and take it.

The car immediately behind you does not move at all.

Then, in the car behind, a "we're in the American dream business" ad comes over the airwaves, and its driver feels safe and secure, having been re-anesthetized.

He moves ahead towards his fate

as you move away towards a much different fate.

[gordon]

As I have said before, it's hard not to buy when it looks as though the market is going to rise forever and every passing month makes it harder to get into the market. In such circumstances, people will do whatever it takes to get a foot on the ladder.

[RW]

The idea of an information cascade (IC) doesn't appear unreasonable on its face and I assume there is evidence elsewhere that suggests it occurs in social systems but in this particular context it has a rather ad hoc tone, like an auxiliary hypothesis or premise in the framework proposed by Imre Lakatos where the primary function of a hypothesis is not to advance a research program but to defend a central theoretical core which in this case includes the premise of a rational actor.

So does IC theory offer opportunities to generate new knowledge and make predictions as would be the case if classical economics is a progressive program or does IC theory primarily advance the project of reconciling data that are inconsistent absent any realistic expectation it will assist in adding new knowledge or even suggest a new question that could be asked; i.e., is classical economics becoming a degenerate program?

It's not my field and I have no answer but it does seem that classical economics must either rationalize events such as bubbles or find a way to incorporate the 'irrational' in a more substantive manner.

[dirtyal]

The interesting phenomenon is when the fundamentals get so crazy that intelligent people "know" it's a bubble and go ahead and act as if it is not. Then, what you have is the "greater fool theory".

When perfectly good companies such as Cisco Systems were selling at over $90 per share, with huge p/e's, lots of people said they saw a bubble developing. But there were plenty of people willing to pay the price.

The bubble did begin to deflate. And after it had deflated by about 95%, people realized that, yes, it really was a bubble.

Then, they looked back at what people like Warren Buffett said, that intelligent investors would not buy a cash flow on such a tenuous set of facts, and now they seem like geniuses.

My wife, who is no economist, recently compared the real estate bubble to what happened in 1929 with people being wiped out by buying stocks at very low margins. So, we learned that lesson, and now you can't buy stocks with 10% down. Maybe we will learn the lesson that you can't buy houses as an investment with less than 10% down. It's too bad for some people who will have to wait to achieve the ownership of their own homes. But maybe it's a good lesson.

[Ed Barbar]

Aren't bubbles a lot like the tragedy of the commons?

Everyone wants to take advantage of a good thing, so even though the action long term is going to be very bad, people can't resist.

In fact, it may not be wise to not go along with the herd. Often bubbles are self-reinforcing, such that the more who make the decision add positive feedback to the system. That translates into shareholder and customer value.

Those who don't follow the herd provide less apparent value to their shareholders and customers, which can be damaging.

One then has to make the judgment, what is worse, being in the same boat with a bunch of people all struggling with the same problem when the bubble pops, or to have taken the high road and taken the damage during the bubble's rise.

When everyone around is making fistfuls of money, and your shareholders are demanding you make more, it's hard to take the high road.

[Bruce Webb]

It can be perfectly rational for investors to buy in a bubble market. If you have a strategy that revolves around equity extraction. If I buy a house at $300k with a 100% LTV and extract equity at each $50k interval on the way up to $600k am I really under water if the value drops 20% to $480k? Or am I in possession of some $750k of assets (after commissions and refinance fees) less the carrying charges on a $600,000 mortgage all earned using other peoples' capital? More to the point should the simple format of my loan make me eligible for a bailout?

Unless you bought in a realatively narrow window around the bubble peak being underwater on a mortgage suggests you did some equity extraction along the way. This is also true for speculators and flippers, if you scored big on five or six deals only to get stuck on the seventh are you a "homeowner losing a home" or "investor with a scoring record of 6 to 1"? Similar considerations lead smart land developers set up a new LLC for each new project, it insulates past success from potential failure. You can call this the Cynical Greater Fool theory. If you make enough money via equity extraction on the road to the cliff it can make sense to follow everyone else over the edge.

In this case the liquidity crisis stuck people in positions they couldn't exit leading to foreclosures rather than a more orderly short sale but the fact is that some of these distressed properties were held by people wearing $400 shoes and driving $75k Mercedes bought as the bubble inflated.

[Bruce Webb]

People forget that people made big money during the tech bubble.

Ask Mark Cuban. He sold 100% of his dot.com for cash right before the crash. I don't know what the company was anymore or whether it is still in business but the sale left Cuban in a position where he can buy sports teams and shrug off five and six digit fines from the league.

I don't know if Cuban was the proverbial blind pig finding the acorn or a guy that cynically rode the bubble to extract every dollar possible. It doesn't really matter much. The fact that Roubini is basking in praise now doesn't offset the fact that following his advice when he first started giving it would have left a lot of money on the table. When caught in a bubble market there is a fine line between blind and shrewd. "Didn't they see? Were they in denial?" Well some of them saw very clearly and extracted maximum advantage from the gap between the anxious that sold too early and the foolish who bought too late. Needless to say we shouldn't listen when they come hat in hand crying for bailout from their last deal.

[devoish]

Three economists, Sushil Bikhchandani, David Hirshleifer and Ivo Welch, in a classic 1992 article, defined what they call “information cascades” that can lead people into serious error.
Is that the same thing as "If you repeat a falsehood often enough they will believe it"?

[david]

"In general, it's hard for economists to admit they can occur at all since they do not appear to be rational responses to economic fundamentals and we do not have very good models of them, so this may explain some of the resistance to admitting a bubble is occurring at the time it is inflating."

This is a particular instance of the general problem -- people get suckered by the vision of money being made. Economists don't like it, anymore than anybody else does. It's a real problem.

As others have noted, information cascade is a wedge of the pie -- the other wedges include information sucking, etc, with the biggest likely being BW's people who think they are in on the con.

[Alex Tolley]

Economists talking about "rational" without defining it. The market impounds information and the various trade/investment decisions of all the players. Some are long term investors, like pension funds and Buffett. At the other extreme are noise traders. For a player with a very short term horizon, any short term "trend" is an opportunity.

The bubble can start just by random events. If the trend occurs, then it can be reinforced by more buying. If your intent is to ride the trend and exit, what is irrational about that to that player?

Housing. The trends have been generally upward for nearly 50 years. Speculators with easy access to money have just ridden and reinforced this trend, until every joe shmo was buying an "investment property", from time shares, to vacation condos to local rental property. How many informercials have you seen about the path to great wealth is via real-estate and here is how to do it with no money down?

Information cascades are just a fancy way of saying, everyone I know seems to be making money this way, so I should follow the herd. It explains little about how bubbles start (random trends, changed conditions like easy credit), nor why they are sustained and then end. No doubt information cascades are now part of "network effects" which are very sexy.

OTOH, maybe it is just better to read Charles Kindleberger, or Charles Mackay for a more historical approach:

Charles Kindleberger, Manias, Panics and Crashes: A History of Financial Crises
Charles Mackay, Extraordinary Popular Delusions and the Madness of Crowds .

[Alex Tolley]

Wikipedia on Information Cascades:
Suppose that there is a crossroad where everyone must choose whether to go left or right. If a person goes the wrong way he will be eaten by a grizzly bear, but if he goes the correct way he’ll end up in safety. Unfortunately, people have imperfect information, so they’ll only be right 2/3 of the time. If person 1 thinks left, he will go left. If person 2 thinks left he will also go left. However, if person 3 thinks that right is correct he will go left anyway. This is because the combined information of observing person 1 and person 2 go left is greater than person 3’s private information that right is correct. Even if every other person thinks that right is correct, they will all go left based on the actions of the first two individuals. In this scenario society will usually go the correct way, but at least one out of nine times everyone will go the wrong way.

"combined information of observing person 1 and person 2 go left is greater than person 3’s private information that right is correct"

Why? What is the basis of this model other than the known crowd following under uncertainty. When does it apply, and not apply? Surowecki suggested that crowds make better decisions as long as the decisions are independent, informational cascades suggests crowds are dumb because they make dependent decisions.

[Lottery Ticket]

In the case of housing, it was perfectly rational for people to buy (using other people's money) regardless of cost. Even if they knew the party would eventually stop. If prices kept going up, people could sell for a profit. If prices didn't go up, just mail back the keys, and lose nothing. Sort of like a free lottery ticket. The modern version of tulip mania, but with no downside risk for buyers. Exerting political pressure to restrict supply with zoning regs even stacked the deck in favor of game players, enabling the game to continue longer.

There really is nothing inherently rational about paying $750,000 for a broken down 2 bedroom cottage in CA. However, if you can buy using other people's money, with no risk, it becomes rational to play the game until it stops.

[Matt_in_TX]

Too bad there is no national town meeting forum where groups of differently experienced people can get together easily and discuss new information concerning the important topics of the day and gradually come to a consensus that collectively is the right decision. Why can't technology save us?

Ah well, maybe next bubble. ;)

[Jim]

As the cartoon in the last New Yorker said, I plan to get in the ground floor of the next bandwagon.

[john c. halasz]

I thought the notion of a bubble involved an at least quasi-rational reason for a boom in a given sector, (though in this case, it might have largely been just a failure of the stock market, followed by an excess of easy credit), which attracts further "investment" as prices continue to rise past their putative equilibrium point. A lot of the behavior we've seen could be accounted individually rational, though collectively irrational. That does put a considerable dent in rational expectations/efficient markets/perfect information type economic thinking, but it does speak volumes for people responding to short-run incentives and analyzing their opportunity costs in terms of diminished returns elsewhere.

The next bubble will be in alternative energy technology and infrastructure renovation, both of which are sorely needed investments, but, absent the removal of excess credit-liquidity and the reimposition of effective financial and corporate regulation, will lead to a new boom turned bubble cycle. The venture capital and private equity funds are already queuing and gearing up in anticipation of the fun.

says...

>Why? What is the basis of this model other than the known
>crowd following under uncertainty.

why is an "other than" required. IMPERFECT KNOWLEDGE is sufficient.

If a person were _100%_ certain that going 'right' was correct, they'd go right even if _everybody else_ goes left.

BUT, with imperfect knowledge, that person will _update_ their view based on what others do. the extent they are influenced depends on how certain they are to begin with.
you must be pretty damn sure of yourself if 100 people go left, and you still think right is correct, but if you are not that sure, maybe it only takes 2 cars to convince you.

THink about it, I bet every one of us were faced with this scenario while driving in a new neighbourhood, and many times took the wrong turn. :-/

when it comes to markets, I bet very few people are 100% sure of what to do (hello Buffett), and of that set, a tiny fraction are are Joe Sixpack and his wife Jane. You just don't know what lies ahead. It seems perfectly reasonable that unsophisticated 'market players' will follow the wrong turns of others ahead of them, because they aren't sure, and besides, a 100 people all turning left ALL can't be wrong, can they ?

I tell ya, living in Vancouver, even though I am pretty damn sure that we are in the midst of a huge housing bubble, it's hard not to get weak in the knees when it just keeps on going like the energizer bunny. Being 100% sure about anything is very difficult.

another reason while someone ought to jam another stake into the perfectly rational and all-knowing Homo Economicus. stick a fork in him - he's done.

[calmo]

I see anony is troubled by the crowd knowing anything...but do I know it? Or izit merely a hunch that I could know somewhat...IMPERFECTLY, he says mightily...crackpottily.
Me, I'm not sure if the informational cascade theory is an advance on my understanding of the nature of bubbles...which remains with the dispositional analysis brought to us by those modern day English philosophers. Not just Harry Frankfurt.
It's not too important what the crowd knows, only that it be capable of being moved (herded) [Check out the ObamaRamas an see if it ain't so]...and in fact there is the view that the utterance "I know..." means you can be held, (and are willing to make that stand for), accountable. (Izat a big advance, you experts? that a failure will bring ridicule on your information packagability?)[You see how the language packets make us sound silly and act clumsy?...or do you feel that this little cascade is making its mark?]
Crowds, not so much.
CR,(click on site at right) notes that the policy makers, the regulators need to be distinguished from the crowd...and maybe even their role in herding this crowd needs some examining.

[Mark Thoma]

The people standing on the corner holding signs pointing in the wrong direction - the ones with stores to the left who stand to profit from a misdirected herd - deserve some scrutiny too.

[SanFranciscoJim]

I am a little confused about all this talk of a "bubble." Isn't a bubble when an asset inflates beyond all reason and then deflates 90, 95 or 99%? Like Tulipmania or the Dot Com bubble? What we had in real estate seems just more like the typical cyclical change we have seen time and again. I guess this time, more people were caught up in it than usual.

[esb]

Mark Thoma:

Regarding your 11:30 PM ...

The "people" to whom you refer are, no doubt, those with seats on the FOMC, as they deliver their disinformational sppeches.

Wink.

[baileyman]

Isn't this like free-riding on price discovery efforts of others?

You see a price, you assume they've done the work, you conclude that's the right price, so you buy there, too. It seems in the housing bubble at least that since for the previous 90 years this kind of purchase behavior was pretty reliable and not dangerous it was reasonable to assume that it remained the right action, unless one could bring the analytical horsepower to the problem to realize the market was off track.

[ilsm]

I have been to a lot of open houses the past several years.

In the new ones the push was prices are going up.

In the older ones the owners was moving to retirement places where the prices were going up each year.

I lived vicariously through the 80's and 90's real estate down turn and never bought in to the prices go up forever appeal.

More grounding for me is I will be fixed income soon and prices going up mean very little if the mortgage is burdensome.

[robertdfeinman]

Random remarks:

1. Psychological tests have shown that people can be persuaded to select options that they know are wrong because of peer pressure. In one experiment subjects were shown a diagram with two lines on it of unequal length. The majority (who were confederates) said the lines were equal, many of the test subjects then went along, despite the evidence of their own eyes.

2. I think the correct analogy is not "bubbles", but musical chairs. Everyone knows that the music will stop, but hopes they will be quick enough to be a winner when it does. It's only at the last stages of a bubble when people get too greedy and the rubes start joining in that one sees foolish behavior. After a winning night at the roulette wheel how many people walk away with the winnings and how many bet it all on one last spin? Greenspan got it right, it's "irrational exuberance".

[ken melvin]

My son gets a big kick out of the return from lore of walking signboards. How do economists factor such symbols into their analysis? What weight given?

[Wiseman]

House buying is a sociological phenomenon.

Consider the following analogy. Take a group of 20-somethings where all or most are single. A few of the singles partner up and become couples. Now, an interesting social action takes place in the group. Most of the people in the group will begin to couple up. Those not coupling will exit the group. In the same group, this same series of actions repeats for marriage, for having children, and finally for buying a house.

There is some sociological motivation of conformity that allows a trend to continue forward with strength once it gets started. The uneconomic decision of actors are more strongly related to crowd behavior and social conformity than to any economic considerations.

These trends can also work in reverse. I suspect if we viewed married couples in China twenty years ago, we would find large groups of couples who were willing to forgo having children. Further, I suspect we will see a full generation of current 20-somethings who will forgo home ownership for a very long time.

[RW]

"The people standing on the corner holding signs pointing in the wrong direction ..."

Hmmm, predatory mimesis perhaps: Could be an interesting biological model or too there but I also seem to recall something from social psychology ...mimetic predation perhaps? Csikszentmihalyi maybe? Can't recall I'm afraid.

[maynardGkeynes]

Everybody I know says that stocks are the best for the 401K over the long run. But maybe this is just herd behavior, like housing and the belief that buying a house is the best investment.

[Arthur Cutten]

In the first sentence Dr. Shiller propagates the mythology which is also known as 'the Ken Lay defense.'

The Big Lie

http://jessescrossroadscafe.blogspot.com/

[Ed Barbar]

Aren't bubbles a lot like the tragedy of the commons?

Everyone wants to take advantage of a good thing, so even though the action long term is going to be very bad, people can't resist.

In fact, it may not be wise to not go along with the herd. Often bubbles are self-reinforcing, such that the more who make the decision add positive feedback to the system. That translates into shareholder and customer value.

Those who don't follow the herd provide less apparent value to their shareholders and customers, which can be damaging.

One then has to make the judgment, what is worse, being in the same boat with a bunch of people all struggling with the same problem when the bubble pops, or to have taken the high road and taken the damage during the bubble's rise.

When everyone around is making fistfuls of money, and your shareholders are demanding you make more, it's hard to take the high road.

[acerimusdux]

Mark Thoma wrote:

"We don't know for sure how or why bubbles occur, so this may or may not be the correct explanation. In general, it's hard for economists to admit they can occur at all since they do not appear to be rational responses to economic fundamentals and we do not have very good models of them, so this may explain some of the resistance to admitting a bubble is occurring at the time it is inflating."

Well, after all, if all the other economists are going left.... ;)

[Alex Tolley]

It would nice if someone (an economist) actually tested whether "informational cascades" were needed vs other possible candidates, like crowd behavior etc. It should be quite easy to test the difference between these two as cascades require sequential information, while crowd influences only require a single data point.

Further. whatever the proximate data, isn't this just affecting personal Bayesian estimates of likelihood of the correct decision?

[JRossi]

Bubbles are ok, but what I like is volatility. You can make or lose more money in less time. At home. In front of your computer, In your underwear.