Theories of Bubbles: An Overview

I have created an infographic summarising the content of this blog. You can cycle through it using the arrow keys, or simply explore it in whichever order you like by clicking and dragging.





So which of these theories is correct? When it comes to recent bubbles in the housing and dot-com markets, I find behavioural explanations most convincing. Reconciling these bubbles with rational investors requires the creation of a complex model that makes a large number of assumptions about the choices made by each group. In contrast, once we assume that financial and asset markets are not uniquely immune to herding effects, the emergence of these bubbles is intuitive and self-explanatory. In accordance with Occam's Razor, the fact that the behavioural hypothesis requires so few assumptions makes it the most likely explanation; in other words, its strength is its simplicity.

The consequence of this hypothesis for identifying bubbles is outlined by Robert Shiller. When we see:

• The sharp increase in the price of an asset or share class
• Great public excitement about these price increases
• An accompanying media frenzy
• Growing interest in the class among the general public
• 'New Era' theories justifying the high price
• A decline in lending standards

It is likely that we are seeing the formation of a bubble.

Bubbles: Rational or Irrational?

In my previous post, I argued that irrational markets is a more convincing explanation of the dot-com bubble than the mechanism involving rational investors set out by Pastor and Veronesi. This conclusion should not, however, be applied generally. Peter Garber (1990) describes the Mississippi and South Sea bubbles convincingly without accounting for irrational traders. The pre-crash price, Garber argues, was indicative of the chance that these revolutionary economic 'experiments' might work. With hindsight, we know that they failed, but prior investors were perfectly rational to account for a chance of success.

Garber also argues that the tulip bubble is something of a myth, and that despite the fact that few experienced or skilled investors were involved in the trade, the observed price trends are not actually uncommon for rare or new flowers. The interesting point here is that markets may have been rational despite consisting mostly of irrational traders. This is a point argued directly by Ross (2004), who states that markets can be efficient even in the presence of a majority of irrational investors provided there is a handful of 'arbitrageurs' involved.

This theory seems perfectly plausible, but herding that is rational from an individual perspective can also result in stock prices that appear unusually inefficient. Irrational investors could produce rational markets, and rational investors could produce irrational markets.

This illustrates why the framework of Maureen O'Hara (2008), presented in the table below, is much more useful than a simple irrational/rational dichotomy.

Traders
Markets		Rational	Irrational
	Rational	Rational Traders, Rational Markets: Traders are rational and Efficient Market Hypothesis holds	Irrational Traders, Rational Markets: Efficient Market Hypothesis holds in spite of herding or overconfidence
	Irrational	Rational Traders, Irrational Markets: Traders are rational, but markets are inefficient for agency/game theory reasons	Irrational Traders, Irrational Markets: Markets are inefficient as a result of psychological flaws in investors

These four categories of theory, and their implications for the identification of bubbles, will be presented in my final post later this week.

References:

Garber, Peter, 'Famous First Bubbles', The Journal of Economics Perspectives 4 (1990), pp.35-54.

O'Hara, Maureen, 'Bubbles: Some Perspective (and Loose Talk) from History', The Review of Financial Studies 21 (2008), pp. 11-17.

Ross, Stephen, Neoclassical Finance, Princeton University Press (2004).

Possible Causes of Bubbles: Beta and New Technology

I noted in a previous post how often historical bubbles have appeared to follow the emergence of new technology. This is extremely widely observed. The Economist goes as far to say that 'Every previous technological revolution has produced a speculative bubble... America's railway boom, electricity, telephones, radio and cars.'

Why is this the case? Shiller (2005) attributes it to sociological phenomena, hypothesizing that new technology leads some investors to believe that previous rules governing the stock market no longer apply: 'new era' thinking. In this post, however, I will be focusing on a theory by Pastor and Veronesi (2009) that explains the emergence of a bubble in a market of fully rational investors.

Pastor and Veronesi develop a model in which bubbles emerge as a result of uncertainty over whether new technology will be adapted by the existing economy. If the new technology is not widely adapted, the risk associated with its stock will remain idiosyncratic: in other words, it will tend to have a very low beta. This justifies a premium on the stock price of these firms. When the technology is widely adapted, however, the risk associated with these stocks becomes systematic, and the price of the stock collapses as a result. Investors are behaving rationally, and the appearance of a bubble only emerges in hindsight when we know that the new technology was widely adapted.

Pastor and Veronesi present some empirical evidence from the dot-com bubble and railway mania to support their model. Personally, however, I do not find it very convincing. The implication that dot-com stock steeply rose in 2000 because of investors who believed the internet was not likely to be widely adapted by the economy is very counter-intuitive, and seems to contradict a lot of qualitative evidence. Popular contemporary justifications for the high price of dot-com stock in 2000 were very different to the rational explanations presented by Pastor and Veronesi. 

Furthermore, the focus of the article primarily concerns whether each action is game theory optimal, rather than whether it is what actually happened. It is assumed, for example, that existing firms will wait until the prospects of a technology become clear, rather than implement it at an early stage or discard it. This assumption is not based on empirical evidence concerning what such firms typically do in reality, but on the fact that this would be the optimal choice within their model. The result is a theory that, while interesting, explains how such a bubble might emerge in a world of fully rational people rather than how it emerged in the real world.

References:

Pástor, Ľuboš, and Pietro Veronesi, 'Technological Revolutions and Stock Prices.' American Economic Review, 99 (2009), pp.1451-83.

Shiller, Robert, Irrational Exuberance: Second Edition, Princeton University Press (2005).

The Economist, 'Bubble.com: All Technological Revolutions Carry Risks as well as Rewards' (2000), retrieved from: http://www.economist.com/node/375561

Possible Causes of Bubbles: Misaligned Incentives

Just a few days after I created this blog, Jeremy Stein, a governor at the Federal Reserve, gave a speech on how to identify overheating in financial markets. Stein presents two possible explanations for the formation of bubbles in credit markets: 'primitive preferences and beliefs' and 'institutions, agency and incentives.' Using the framework of Maureen O'Hara (2008), the first explanation can be categorised as one of 'irrational traders, irrational markets', and the second as 'rational traders, irrational markets.'

The 'primitive preferences and beliefs' explanation is strongly linked to the behavioural research I blogged about previously, implying that bubbles emerge as a result of human psychological tendencies (particularly herding). Stein, however, focuses primarily on the overheating effect of 'institutions, agency and incentives'. He argues that bubbles in financial markets often result from agency problems: institutional investors disguising low-probability risks in an effort to 'reach for yield'. The systematic understating of the risk involved in an asset ultimately leads to systematic overpricing of that asset, inevitably causing a bubble.

This appears consistent with the most recent financial crisis. From 2000 onward, sub-prime mortgages and other junk bonds were repackaged by banks into derivatives specifically designed to 'fool' the algorithms of rating agencies. As a result, the price of these derivatives rose significantly beyond their true value, and this only became clear when large numbers of borrowers defaulted.

What can this tell us about how to identify a bubble in credit markets? Stein argues that a major warning sign is 'rapid growth in a new [financial] product that is not yet fully understood'. This may be the result of innovation, a change in regulations that creates new loopholes, or by a change in risk-taking incentives. It may be the case that such a product is being used by bankers to obfuscate systematic risk. 

He warns that the next financial bubble could arise through 'collateral transformation', whereby the high credit standards of a clearinghouse are circumnavigated by borrowing Treasury securities from a third party using junk bonds as collateral. The result is not only that risks may be understated, but that an additional party is exposed to a potential default.

Another signal that the market may be overheated is the 'high-yield share', defined as issue by speculative-grade firms divided by total bond issue. Greenwood and Hansen (2012) graph the high-yield share against excess returns, finding a strong negative correlation:

There thus appears to be some empirical evidence for the theory that credit markets overheat as a result of agency problems.

While Stein's speech focuses on the exploitation of loopholes, Slate magazine point out that a credit-market bubble can also emerge from outright fraud, referencing a recent court case suggesting that JP Morgan illegally misrepresented the value of mortgage-backed securities to ratings agencies. A bubble might be compared to a naturally-occurring Ponzi scheme, but sometimes it may not be as 'naturally-occurring' as it first appears.

Another interesting aspect of Stein's explanation is that it links the emergence of a bubble to the emergence of new technology: in this case, new financial instruments. As Pastor and Veronesi (2005) point out, this seems to be a recurring theme in the history of bubbles, and the link will be examined more closely in future posts.

References:

Greenwood , Robin, and Samuel G. Hanson, "Issuer Quality and Corporate Bond Returns" unpublished paper, Harvard University (2012), retrieved from: http://www.people.hbs.edu/Shanson/Issuer_Quality_20120905_FINAL.pdf

O'Hara, Maureen, 'Bubbles: Some Perspectives (And Loose Talk) from History', The Review of Financial Studies 21 (2008), pp. 11-17

Pastor, Lubos, and Pietro Veronesi. Technological revolutions and stock prices. National Bureau of Economic Research, Working paper  No. w11876 (2005).

Stein, Jeremy C., 'Overheating in Credit Markets: Origins, Measurement and Policy Responses' (2013), retrieved from: http://federalreserve.gov/newsevents/speech/stein20130207a.htm#fn14

Yglesias, Matthew, 'Reading Fraud out of the Institutional Perspective', Slate (2013), retrieved from: http://www.slate.com/blogs/moneybox/2013/02/07/jeffrey_stein_on_credit_institutions_don_t_forget_the_fraud.html

Is Gold the next Bubble?

My previous post was based around the work of Burnside et al (2011), who modeled a housing market under the assumption that a bubble is a rise in prices not justified by a change in the underlying fundamentals. In the case of the housing market, it may be reasonable to assume that no one can tell when the underlying fundamentals of its value have changed. But in other markets, these underlying fundamentals can be very obvious: take, for example, the market for gold. The following graph shows the price of gold in USD since 1999:

Source: Moneyweek.com

By Burnside's definition, this looks like a clear bubble. The underlying usefulness of gold has presumably not changed in the past ten years, yet its price has increased by over 500%. Much of the demand for gold comes from investors rather than consumers: buyers do not want gold for its own usefulness, but because they expect to be able to sell it to someone else later for a higher price. This seems like a textbook example of a speculative bubble.

So is this the case?  Tim Harford suggests that if there is a gold bubble, it has been inflating for much longer than the past ten years. Indeed, its value has outstripped its usefulness for over 4,000 years. Its value is derived solely from the near-universal belief in its value. Unlike truly valuable commodities, such as food or electricity, if everyone in the world decided that gold was worthless, it would become worthless. 

It might, however, be more accurate to view gold as an illustration of an exception to Burnside et al's definition of a bubble. The price of a good might temporarily outstrip its true use-value as a result of excessive optimism on the part of investors, in which case its price can be expected to collapse at a later date. However, if the belief in the value of something is widespread and persistent enough, that commodity may become a widely accepted medium of exchange: a money commodity. If this is the case, the price of the commodity might remain above its 'true' value for milennia without crashing.

References

Burnside, Craig, Martin Eichenbaum and Sergio Rebelo, 'Understanding Booms and Busts in Housing Markets', National Bureau of Economic Research, Working paper no. 16734 (2011).

Harford, Tim, 'The Bundesbank Takes Back its Doughnuts' (2013), retrieved from: http://timharford.com/2013/01/the-bundesbank-takes-back-its-doughnuts/

Possible Causes of Bubbles: Infectious Enthusiasm

One of the key aims of this blog will be to identify any differences between a bubble destined to burst and a sustainable boom in prices. This may not be as straightforward as it sounds; a recurring feature of bubbles is that they are only conclusively identified in hindsight. Indeed, some of the most interesting models in the field assume that actually identifying the bubble in real time is impossible.

This is the case for Burnside et al (2011), who model the behaviour of quasi-rational investors in a housing market. Sometimes a rise in prices is sustainable, and sometimes it will not be, but the model assumes that no investors can conclusively tell which is the case before the bubble 'bursts'. The only difference between a bubble and a justifiable boom is that the optimists "happen to be right".

The distinctive feature of this model is that it compares the beliefs of investors to infectious diseases, and can therefore take advantage of the body of work originating with Daniel Bernoulli's Epidemic Modelling. When an optimist meets a pessimist, there is a chance that the pessimist will 'catch' the optimism bug, and vice-versa. Beliefs are heterogeneous, but infectious, and so periodically tend towards homogeneity until the point at which the majority are unceremoniously mugged by reality.

There is a behavioural justification for believing that this is an accurate representation of reality. Daniel Kahneman (2011) describes a psychological phenomenon called 'herding', where individuals form opinions by balancing their beliefs against those of everyone else. Surowiecki (2004) argues, and Kahneman agrees, that this is usually more rational than sticking with a belief that the rest of the population consider extreme. The formation of bubbles can thus be considered a form of prisoner's dilemma: rational behaviour on the part of individuals leads to irrational behaviour on the part of the overall group.

Alan Greenspan coined the term "irrational exuberance" to describe this phenomenon, and this phrase has since been used by Shiller (2005) as the title of an influential book on the causes of bubbles. As a means of describing the phenomenon it is very convincing, but whether it can be used to identify a bubble in real time is somewhat less clear.

References

Burnside, Craig, Martin Eichenbaum and Sergio Rebelo, 'Understanding Booms and Busts in Housing Markets', National Bureau of Economic Research, Working paper no. 16734 (2011).

Kahneman, Daniel, Thinking Fast and Slow, Penguin (2012).

Surowiecki, James, The Wisdom of Crowds: Why the Many are Smarter than the Few, Abacus (2005).

Shiller, Robert, Irrational Exuberance: Second Edition, Princeton University Press (2005).