The Joe Walker Podcast - Housing Bubble Week: Housing Bubbles As Availability Cascades - Timur Kuran

Episode Date: May 15, 2019

Timur Kuran is a Turkish American economist.See omnystudio.com/listener for privacy information.

Hello there, ladies and gentlemen, boys and girls, swagmen and swagettes. I'm your host, Joe Walker, and this is the Jolly Swagman podcast. Welcome to episode three of Housing Bubble Week. Now, I'm going to begin this episode with a story and some theory. This will take a few minutes, but it will be fun and it will help everything make sense. In mid-2018, after a recommendation from friend of the pod and former guest Jonathan Tepper, I began reading Michael Lewis's book Boomerang. When I read the chapter on the Irish housing bubble, it was like the scales had fallen from my eyes. Lewis recounts meeting the University College Dublin economist Morgan Kelly. Kelly specialised in medieval population statistics and wrote articles that were considered esoteric even by the standards of academic economists.

But he emerged from obscurity to call attention to what he rightly perceived to be a mammoth housing bubble living right under his nose in Ireland. As Kelly describes his experience to Lewis, I was in this position, sort of being a passenger on this ship, and you see a big iceberg, and so you go and ask the captain, is that an iceberg? Kelly's most strenuous warning to the ship's captain took the form of an academic article published in June of 2007 titled, On the Likely Extent of Falls in Irish House Prices. The working paper version appeared in February 2007, just one month before Irish house prices

commenced their drawnout but devastating decline. Kelly's paper strikes me as remarkable for two reasons. One, Kelly made some bold theoretical claims about how housing markets should be conceptualised. He wrote that housing markets are better modelled not as efficient markets but as information cascades, where the actions of other agents signal their private information and can cause individuals to ignore their own signals and follow the herd. Let me help you get your head around this idea briefly before I return to the other remarkable feature of Kelly's paper. The notion of an informational cascade is incredibly important for you to know.

It can be traced back to a highly influential paper published in 1992 by three economists, Sushil Bhikshandani, David Hirschleifer, and Ivo Welch. The paper is titled, A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades, and you'll find it in the show notes to this podcast. Ominously, the authors open their article with a verse from the Gospel of Matthew, the one that goes, Let them alone, they be blind leaders of the blind, and if the blind lead the blind, both shall fall into the ditch. But what is an informational cascade?

An informational cascade occurs when it's optimal for an individual, having observed the actions of those ahead of him, to follow the behavior of the preceding individual without regard to his own information. It's kind of like when you're waiting at the pedestrian lights to cross the road and someone beside you has their head buried in their iPhone. You notice there are no cars and so you cross the road. The person beside you notices you crossing and, without checking for cars, follows you over. Informational cascades are actually a subset of a broader phenomenon known as herding, but whereas herding is about behaviour only, informational cascades describe how people update their beliefs based on the observed actions of those acting ahead of them.

So what does all this have to do with bubbles? Well, in their paper, Bix and Darnie, Hirschleifer and Welch outline a specific model where each agent acting in sequence enjoys the same gain or pays the same cost of adopting some behavior. As a result, they explicitly note that although cascades may apply to bubbles or crashes, this is not directly captured by our model, since the cost of adopting, for example buying a stock, increases as a bubble forms. But there are various other models in the cascade literature that may be used to describe bubbles. Bob Schiller, the Yale economist and Nobel Prize winner, had been making this argument long before Morgan Kelly. Schiller has explicitly argued that the US housing bubble can be thought of as an informational cascade.

Here he is writing in the New York Times in March 2008, quote, Imagine the individuals in a group must decide, each, whether real estate is a terrific investment and whether to buy some property. Suppose that there is a 60% probability that any one person's information will lead to the right decision. Suppose houses are really of low investment value,

but the first person to make a decision reaches the wrong conclusion, which happens, as we've assumed, 40% of the time. The first person, A, pays a high price for a home, thus signalling 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. And that is how informational cascades can snowball. In his book Irrational Exuberance, Schiller notes that informational cascades are not a complete theory of bubbles, but they do have relevance for investor behavior. So back to Morgan Kelly. The second and perhaps most interesting aspect of his article on the likely extent of falls in Irish house prices was its reception. It was received with as much warmth and good humor

as Martin Luther's 95 theses pinned to the door of the Wittenberg Castle Church. At a conference on the 4th of July 2007 in Donegal, Ireland's then Taoiseach, or Prime Minister, Bertie A. Hearn, wondered aloud why people sitting on the sidelines cribbing and moaning about the economy did not commit suicide. The comment was interpreted as a clear swipe at Kelly and another economist, David McWilliams, who were nearly alone in raising the alarm about Ireland's housing bubble. As a prominent Irish bank analyst in Dublin, commenting on the mindset of journalists and bankers at the time, put it, you're either for us or against us. And Team Ireland was unequivocally for the bubble.

It's interesting that housing bubbles often seem to inspire this sort of vociferous conformity. Friend of the pod and former guest Jonathan Tepper was designated an enemy of Spain in 2009 for pointing out the, at that point, nose-bleedingly obvious Spanish housing bubble. According to John Malden, Tepper was added to a government blacklist containing three other names. None of them were allowed to be quoted or appear on any media programs after that. The attempted ostracisms of Kelly and Tepper make sense. Not only are housing booms bound up in nice national narratives about, for example, the Australian Dream or the Celtic Tiger, but they swell the net worth of homeowners, so almost everyone has a dog in the fight. Criticise a

housing bubble, and you might not commit literal suicide as Bertie Ahern would have it, but you're certainly at risk of committing career suicide if you inhabit an industry or mille-feuille that depends on the bubble. Which brings me to this episode. I have a secret for you, and I want to share it. I think we can do one better than Morgan Kelly or Robert Shiller. Bearing in mind informational cascades, the tense conformity surrounding housing bubbles, and the notion that prices always go up that feeds them, I think the best way to model housing bubbles yet devised is to describe them as availability cascades. Yes, one more piece of jargon for you to digest.

I need to introduce this final concept now before we throw to the interview. An availability cascade is an incredibly powerful construct advanced in 2007 by two leading US academics, the preeminent legal scholar Cass Sunstein of Nudge Theory fame and the social scientist Timur Karan, our guest. An availability cascade combines three concepts. First, an informational cascade, with which you're now familiar. Secondly, a reputational cascade. Like an informational

cascade, in a reputational cascade, late responders sometimes go along with the decisions of early responders. But in a reputational cascade, they do so not just because the late responders think the early responders are right, but also because they perceive their reputation will be damaged if they dissent from the early responders. So an availability cascade is a hybrid. It subsumes the informational and the reputational. It prioritizes to different degrees mental shortcuts and social relationships. Lastly, this combined cascade is mediated by the availability heuristic, which availability heuristic is a mental shortcut humans use, which was first described by psychologists Daniel Kahneman and Amos Tversky. It occurs when we judge the frequency of a category

by the ease with which we can call examples to mind. In other words, by how available they are. For instance, imagine there's news just in of a dramatic plane crash. The news will be salient and recent. And if you have to catch a flight that day, you will feel especially nervous doing it, even though the statistical base rate for a crash is materially unaffected. Another example is this. Imagine you're overwhelmed with good news stories about house price increases in your neighbourhood. You might be liable to jump to the conclusion that prices must always go up here, or that prices are highly likely to go up in a geographically disconnected suburb.

To tie these threads together, an availability cascade is an informational-slash-reputational cascade mediated by the availability heuristic. It generates a self-reinforcing cycle whereby a particular belief becomes relentlessly more pervasive. I'd been aware of the concept for some time and I realized it might hold a more complete way to describe housing bubbles. I decided I had to talk to Tamor Karan, the original proponent of the idea, to hear whether I was right straight from the horse's mouth. Tamor is Professor of Economics and Political Science at Duke University, and as you'll hear, he's one of my favourite social scientists. Now, this episode is an excerpt from a much longer conversation I held with Tamor.

The excerpt is specifically about how availability cascades might apply to housing bubbles, and that's why I've introduced the general concept myself, because in the way we're about to jump in, that context would have been lacking. Tamor and I take five to ten minutes to warm up, but if you can be patient with us, you will be richly intellectually rewarded. So, without much further ado, please enjoy my chat with Timur Karan. Timur Karan, thank you so much for joining me. Thanks for inviting me. Now, Timur, in an email to you, I mentioned economic bubbles. And this is a topic my listeners know me for.

It's become somewhat of an obsession. I was hoping we could try to apply the concept of availability cascades to an economic bubble, like, for example, a housing bubble. Have you thought about that? So, a little bit uh i've thought less about uh housing bubbles than about stock bubbles but they're uh bubbles that occur in the stock market in stock markets most of the players are typically interested in returns. And they're typically interested in whether a stock is going to go up or down. In the housing market, they're typically interested in when they're buying a house, among other things, in whether the house is a stock or the value of a home the probable trajectory of the value of a home you also rely on experts you also do some research you look at good people who are quite knowledgeable about, let's say, the housing market.

You look at literature on the housing market. cascades can play a role because there are are if certain neighborhoods are considered or buying a house in a certain neighborhood is considered a good investment for the future and if most people are saying that. There's a risk in saying the opposite. If you go along with the crowd and say, like most people, that this neighborhood is going to hold its value, it's going to become more and more popular, and things turn out the other way, you haven't lost very much because, after all, that's what all the experts were saying. But if you take the position, especially if you're a young realtor,

trying to, you know, advising people, and you advise people not to buy in that neighborhood to go somewhere else. And you you're wrong the neighborhood that you steered people away from actually gained in value you take your reputation takes a hit so cascades can occur there and bubbles are often driven by people's reluctance especially as home values are rising people's reluctance to say that the market is overheated gotcha so this is this is where and and it can be it's both the the availability cascade is the correct concept here because the people who are participating in promoting the notion that a particular neighborhood is a good investment are driven partly by informational motives. They're reading other experts for information

to learn from them. But at the same same time they're trying to protect their reputation they don't want to make a mistake where they can be singled out as having gotten it wrong tim or you you remind me of a quote of john maynard canes i think it I think you said that it's better for one's career to have failed conventionally than to have succeeded unconventionally. I think that's the quote. I remember reading that somewhere. I can't remember in which of his books he says that. But basically, there's more downside to

being wrong if you go against the mainstream consensus then there is downside to be wrong within the consensus yes yes that is that is correct now this isn't to say that there are some people in spite of that observation who take the risks and are contrarian. People vary in their risk tolerance and people vary in the degree to which they believe in a certain cause. And they play important roles in ending cascades, in ending the types of cascades we're talking about bubbles so bubbles sooner or later they come to an end they burst and then things start going going the other direction. They burst. Right before they burst, there are usually people who say, it's time to sell.

It's time not to invest in this market. It's time not to invest in housing, period. This housing market is overheated. It's better to rent. There are people who start saying this. Initially, they're ignored, but you have a critical mass. Again, the mix of the risk involved, the risk that we just talked about varies also. It's a lot higher in early stages of the bubble than at the very end. You get to a point where the chances, if you go against the consensus, the chances of being wrong are

much lower. Now, in your original paper with Cass, you introduce an extension of the idea, which is this lovely idea of availability entrepreneurs, people who might, even on some intuitive level, understand the dynamics of an availability cascade and stoke it and fuel it to increase the availability of certain information and generate this self-reinforcing cycle. I've been reflecting on this concept in the context of the Australian

housing market because you have a lot of people who charge for seminars and boot camps and courses where they can teach you how to become a property investor. And invariably, they'll elevate stories of people who have achieved immense capital gains through property investing. And the stories of capital gain increases become more available during a housing bubble. Thanks in part, I think, to these, what I term, availability entrepreneurs of the Australian housing market. Because one of the myths that gets perpetrated in any bubble, but taking a housing bubble in this instance, is that, you know, you can't lose, you can't, this asset class never goes down in price.

And people have literally been saying in Australia, I date our bubble as beginning in 1998. The first phase lasted until 2008. Prices went sideways. Australia escaped the Great Recession for a number of reasons. We reinflated the bubble from about 2012 to 2017. It peaked in October 2017. But one of the things people were saying to each other during both phases of the bubble was house prices in Australia double every seven to 10 years. This was almost like a mantra that was repeated. Of course, if you take it to its illogical endpoint,

we'd reach a situation in Australia, I i mean price to income ratios in sydney were over 11x at one point but if that continued we would have reached an absurd conclusion where people you know it would have been impossible to repay your mortgage during your lifetime you would have had to pass it on to your children even for middle-class suburban homes. But it was still repeated and it became, people seem to judge the prospects of success or how fail-safe real estate was as an investment based on how easily they could call to mind things like this mantra that prices always go up, success stories that they hear propagated through availability entrepreneurs. And because I've been looking at

so many property market-related literature on the internet, my Facebook newsfeed is now always full of ads. I've been targeted by the availability entrepreneurs. So I'll always get the videos in my newsfeed, them standing in front of a whiteboard um telling me about their students who've achieved immense capital gains um i just thought it was a lovely application of the idea of availability entrepreneurs i don't know if you have anything to add to that that's just my sort of two cents you've given an excellent example. I'm certainly no expert on the Australian

property market, but I've learned a great deal from you. It does ring a bell, though, because this is, we've lived through housing bubbles here. I used to before moving to North Carolina 13 years ago, we lived in Southern California, which had its own huge housing bubbles. And I home, we were being told not only by our real estate agent that if we waited even a week, we would be losing, we would be facing higher prices. And there were people who were, I remember at the time who were, who knew of course, because we were in the market for a house. And this is, this was before we had the internet, they would learn through real estate agents who is in the market looking for a house and they would you would find flyers on your door they would flyers would come through the mail inviting you to these uh to these courses that would teach you how to make a lot of money flipping houses, buying houses, and also teaching you how to buy a house wisely and so on. And invariably, what they were pushing

was that you auto it's this is this is you could never leave and i remember looking at this thing and seeing all these charts of how in the last 20 years in los angeles prices had never fallen of course they had there were times when for example in the great depression when prices did fall cleverly and self-servingly they started their series at a bottom and you know so that they would they would be able to show you exaggerate the the gains that you would that you would see and And I remember seeing how they would take out very strategically, pick out certain periods and tell you how that is exactly like the period you're living through where in one year prices rose 20%. Yeah. prices rose 20%. And so these were, again, these were availability entrepreneurs,

people who were pushing this, who were feeding this, deliberately feeding the bubble, but also in the process making money because people were paying a lot of money to join these courses, to take these courses and to learn. The first class was always free, I remember. They'd bring you in there for free and they'd give you a few juicy things about how, you know, to whet your appetite. And then, of course, you had to pay in advance. I don't know what it was a thousand dollars two thousand dollars

something like that there were these people who were actually uh benefiting from this so now of course this availability entrepreneurship works through if you're if you use Facebook if you use Twitter of course it it it can use far more sophisticated means. They can just look at what you're searching, what Google searches you did, and within five minutes, you know, the ads that pop up reflect that. Yeah. Timo, I have so many stories that resonate with that. I often attend these seminars that the property availability entrepreneurs host. I went to one in Sydney at the end of last year. It was like the Australian property market's answer to Tony Robbins. It was, you know, really wild. We walked

in, you know, happy by Pharrell Williams was playing. There were VIP seats and, you know, a stage and everything. And they brought up the people who had been really successful to tell their stories. But at the back, in between the sessions, were people waiting with, you know, the FPOS machines and the paper to sign people up to the programs. I went to another in Sydney on the 5th of March. They had a seminar and they were showing us charts of Australian house prices going back to 1986. Now, in fairness, the Australian Bureau of Statistics only began publishing national house prices in 1986. Similar story for the US, I think. Your National House Price Index wasn't compiled until

recently and then Case and Schiller did some excellent work after that. But, you know, it is misleading because two-thirds of that period overlap with an historic housing bubble. So, you're going to give a disproportionate sense of capital appreciation compared with if you took a longer run price index. So Nigel Stapleton, a fantastic academic at the University of New South Wales, did the first long run price series for Australia going back to 1890. If you use that, I don't remember Stapleton's paper exactly. It was the subject of his PhD thesis. But I think prices from 1890 to 1990 increased 0.5 percent a year in real terms i'll put a link up for people so i get it right similar story to um you know what schiller found

long-run prices in the u.s similar to what pete eichholz found in the herring grok index but this is another little image they put up in their slideshow at this seminar i went to on the 5th of march they're combining um i don't know if you can see that very well yes like i just join real estate versus shares and i think bonds um to say that you know the returns compared to shares are like just as good but the period the period cuts off in 2016 yes um and i went to this on the 5th of march 2019 and of course the australian housing bubble peaked in october the national bubble national prices that is peaked in october 2017 so we've cut off that period where prices have fallen more than 14 in syd and more than 11% in Melbourne,

to create the impression or to make the impression more available that real estate is the best investment? Favorable information, favorable to that view is made more available and information that discredits that view is made unavailable. This is, and now this is in this particular case, I doubt that, that, uh, violence is being used to keep people from reaching this, this, uh, information. It's just that people are very busy and people don't have the time to do research. You put plenty of paperwork in front of them that contains the type of information you want them to have. Most people aren't going to go much further than that. And most people don't

have the sophistication that you have to look at a time series and say wait a minute we're in this is 2018 why does this stop in 2016 now most people to most people this doesn't occur most people are not statistically that sophisticated you show people 10 examples. You bring them, you know, parade, put in front of them 10 examples who tell you anecdotes about how they've become so wealthy, flipping houses, or how they did so well buying their house, how they're able to retire early. And a lot of people think that that's very valuable information but in a large society in a large city like sydney you can find 10 people who did very well in housing you can find 10 people who did horribly in housing it doesn't tell you anything by itself it's not statistically representative but most people are not statistically statistically sophisticated which is in which is one of kahneman and teresky's points we use heuristics but as a substitute for proper statistics and for controlled experiments and so on these are

the availability heuristic is a rule of thumb and so you are availability entrepreneurs are exploiting this rule of thumb and the seminars and the literature that they show and the dazzling graphs and so on are all intended to, as you put it well, to make available certain images in people's minds. prepared greater the variety of the ways that you have presented certain information the more available that information is they've heard it through more willing to take a risk more willing to pay for the course yeah they're willing that you you reach them through anecdotes you reach them through graphs of the kind that you just showed you reach them through anecdotes. You reach them through graphs of the kind that you just showed. You reach them through famous people who come and speak and tell you about their stories. You do it in all these ways.

All this makes it more available when people think about housing. These are the things that come to their mind. And, of course, they're willing then to pay for it. Exactly. Just finally, Timur, they're willing then to pay for it. Exactly. Just finally, Timur, maybe we can finish on this. What is additionally interesting to me about availability cascades as they apply to economic bubbles is they seem to have serious implications, maybe even offer a rebuke to the efficient markets hypothesis, which from its

throne room in the University of Chicago still exerts a long and strong influence on the economics profession and financial economics in particular, of course, because efficient markets say that investors use their own private information and judgments to come to an opinion about what the intrinsic value of an asset should be. And the sort of aggregation of their opinions gives us a wisdom of the crowds, which brings that market prices closest as realistically possible to the true worth of the asset. But if investors are under an availability cascade or, you know, more banally, just an informational cascade that's causing them to decide to invest, it's not an efficient market you wouldn't expect prices to follow a random walk if people are being influenced by the herd yes and i i would go further than that

and say that you wouldn't expect uh the market to follow a random walk if there are availability entrepreneurs who are deliberately trying to affect the – deliberately trying to distort the information that is available to people. Now, having said that, let me immediately offer a counter argument to that. And the counter argument is that you can have availability entrepreneurs working at cross purposes. With regard to housing, you can have availability entrepreneurs who are also saying you can make a lot of money by shorting the market and by, you know, you can make a lot of money by investing in other areas by taking your money out of housing and renting and putting your money into other areas where the returns are going to be greater. Housing prices will fall. You'll be able to buy a house twice as large for what you're giving up now. So, you can have... You mean shorting the market just in a sort of colloquial sense?

Because, I mean, you can't really – Yes. You can't directly short a housing. Correct. Just in the colloquial sense. that as long as people are free to enter the market to defend any view concerning the market, concerning where assets are going to go, as long the efficiency is going to hold because people will, there will be entrepreneurs, if there is a housing bubble, they will see that there's a price to be made by betting against it and teaching people that uh the opposite now what they miss is that the political system

may be biased in favor of one versus another the economic system may be biased in favor of one the whole real estate industry might be biased in favor of one rather than the other. And so you're less likely to get credible availability entrepreneurs defending one position than the other. So the efficient market hypothesis assumes a way the institutions of society that can make one type of availability entrepreneur, that can give an advantage to one type of availability entrepreneur over another. Let me add a quick example just to demonstrate that. I mean, this was very true in Ireland. There was a strong conformity between the media and the elites when it came to their housing bubble. But in Australia, for instance, we have, you know, two big mastheads, the Australian newspaper,

which is owned by Rupert Murdoch's News Corp, and the Sydney Morning Herald, which is Fairfax. News Corp is sort of right leaning, Fairfax is sort of left leaning as people traditionally understand them. And yet both have interest in the property market. Fairfax owns Domain and News Corp has a stake in REA. But sorry, please continue. That's just sort of an example of the point you're making about how institutions can create parameters which might favor some availability entrepreneurs you've given a good good example and in the in the real estate sector in the in the united states again you have uh lots of people who have a stake, lots of very sort of powerful players who have a stake in keeping property values high, and including local

governments whose tax revenues are tied to housing prices. So they have an interest in perpetuating. If you're in power right now, you would like, and you know that there's a bubble underway, you sense that there's a bubble underway, you would rather keep the bubble going and have it burst after you're out of office. Because while you're in office, tax revenue is high. Therefore, you can provide all the services that you promised. Once you leave office and you go, you run for something else and you've moved on, it's somebody else who has to deal with the burst bubble and the depressed tax revenue. So that's, again, an asymmetry that favors, an institutional asymmetry that favors, in the American case, I don't know whether this resonates with you in Australia,

but at least in the American case, that favors one type of entrepreneur over another, gives an advantage to them. So the playing field is not flat. Yes, it's true. In principle, you can, availability entrepreneurs can cancel each other out in markets but in specific cases they will not cancel each other out because institutions favor one or the other and there may be a case for regulation and it might be a case of regulation but of course regulation as there's a there's for you have regulation you can have regulatory capture and certain certain groups of people can distort can can use regulations to give them particular advantages so there's no there's no perfect world in which you can level the playing field for availability entrepreneurs through and

ensure that bubbles, financial bubbles don't occur, but it's also a possible aggravator of the problem. might put in place regulations that favor the already favored availability entrepreneurs. We've certainly had our fair share of that as well in Australia. Yes. Timur Karan, it's a complex world populated by humans not econs and thank you so much for helping us to understand it in a clearer light and availability cascades i'm gonna refer everyone to your readings to your books to youtube videos and to papers you've written and thank you so much for for joining me and giving me the time thank you so much for inviting me to your program and i greatly enjoyed the conversation and learned a lot about australia yeah thank you thanks so much for listening.

I hope you enjoyed that as much as I did. Before you go, I have a quick favour to ask. If you liked this episode, please rate and review the podcast on iTunes. I know everyone asks, but it really makes a difference. I make these podcasts for free. They are bloody time-consuming, but they're important, and I couldn't do it without you. Finally, for show notes and links to everything discussed in that conversation,

you can find them on my website, josephnoelwalker.com. That's my full name, J-O-S-E-P-H-N-O-E-L-W-A-L-K-E-R.com. You can also get in touch with me there or on Twitter. My Twitter handle is at josephnwalker. Thanks again for listening. Talk to you soon. Ciao.