The science of bubble spotting

Even before the collapse of the housing bubble in the United States and around the world there was a substantial debate in the field of macroeconomics concerning whether central banks should take asset prices into account when deciding on the course of their monetary policy. I would like to suggest that the appropriate answer here should be yes. The failure to do so became abundantly clear somewhere around the end of 2008, even though I can sympathize with the view that it was not the housing crash that brought down the worldwide economy, but rather failures of central banks to do their jobs by allowing nominal GDP to crash almost 2 full years after the housing bubble burst. Think about it, not much happened between 2006 and 2008. In any case, it is pointless to debate whether or even what central banks should do to try and limit the effects of asset bubbles if we don’t have a way to tell when such a bubble is present to begin with.

The most common way to go about this is to try to figure out what the price of assets “should be” and compare it to what it is. Essentially one tries to identify what the fundamentals that give an asset its value are. The concept that the price of a financial asset is nothing but the present value of (expected) future incomes that can be derived from the asset helps as well. While the whole thing is hardly rocket science, it is problematic in that it in essence involves normative statements concerning what fundamentals are supposed to influence prices of, say, real estate. Most are obvious, such as income and population growth or real interest rates. Even if one manages to capture all the major relevant factors, anything not included in the definition of fundamentals might still have a considerable influence on how prices develop. To name an example, expectations of comprehensive immigration reform in the US (yes, unlikely, I know, but just for the sake of argument) might be a very valid reason for housing prices to rise, and rise in a way that is desirable in terms of market efficiency going forward into the future. I’m still looking for better examples, but I guess the idea should be clear. Financial innovations, which seemed to have played a major role in the recent crises, could also (and sometimes justifiably) raise the value of assets, by making them more liquid, for example. A central bank focusing solely on “fundamentals” would totally miss this.

Other approaches try to look at the statistical properties of asset price movements, but remain reliant on the concept of fundamentals. One of the earliest ideas, which seems decent enough, was developed by Robert Shiller (of the S&P Case-Shiller home price index) is to look the volatility of asset prices and compare them to the volatility of the fundamentals that should be the basis for these kinds of fluctuations. Too high volatility could be a sign of a bubble. Well, most likely it isn’t, but that was the idea. Asset prices rising exponentially while “fundamentals” might only rise linearly could be another indication of a bubble forming. Based on these results some newer models try to spot bubbles, yet a study I stumbled upon claims to have identified 10 “bubbles” in Hong Kong’s residential market (.pdf) from 1994 to 2011. That might be, but it’s unlikely to be any help in devising monetary policy, as the costs of raising interest rates so often with the sole justification being a potential bubble seems to be more costly. Also, most bubbles identified with this method seem to last less than a year. Hardly the ones threatening worldwide economic mayhem.

One might also argue that it could be useful to look at some important economic indicators that could lead to an asset price bubble and compare them with their historical trend. Some of these measurements could a asset price gaps, credit gaps, investment gaps or real credit growth gaps. The idea is that if any of these measures, and potentially several, deviate from their trend above a certain threshold value, alarm bells should sound. It seems that combining a couple of these measures indeed gives us a fairly useful test for identifying bubbles while they are still forming, yet in hindsight the signal, depending on what threshold values are used, still give false positives in 18% of the cases. Certainly looks promising, yet an almost 20% miss ratio really hurts considering the huge costs of monetary tightening can have on the real economy.

Bubble spotting is hard, yet unless we learn to do it better it will prove to be difficult to prevent them from bursting in the future. And throwing our hands up while arguing we’ll just clean up after it bursts seems to be a very lazy way of going about it. And in any case, recent events show it might not even be possible to do that.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s