‘Black Swans’ or ‘Dragon Kings’?
Understanding the Predictability of Bubbles
By Jeremy Carter
August 2013
One way of looking at asset price bubbles is that they are rare and unpredictable events, like finding a black swan when you think all swans are white. Another is to see them as quantifiable and predictable characteristics of growth, what French Professor Didier Sornette describes as “Dragon Kings”. Both ways are valid.
In some sectors there is just too much uncertainty about when circumstances will change, and we take a “black swan” approach, capping our ratings or not rating assets in that sector. But avoiding the sector altogether is not always an option because of its interconnections with other areas of the global credit markets. Occasionally, quantifying the size of the bubble may work; at other times qualitative judgment has to be used. Our Risk Radar assesses the macro risks facing our ratings, including the risk of a bubble bursting and the impact it would have on our ratings.
Author and scholar Nassim Nicholas Taleb gives an illustrative example of a “black swan” event in his book of the same name. The turkey who calculates the amount of food he will receive tomorrow based on his previous years’ experience suddenly learns that it’s Thanksgiving and he’s headed for the chop. There is no way the turkey could have known this based on the data leading up to the event. Of course, access to data on other turkeys would have changed his analysis.
But some risks have to have parameters put around them. China is a prime example: macro imbalances have built up that will have to be unwound. The question is whether there will be a hard or a soft landing. This is crucial to our thinking on the Chinese sovereign and banking system, and on the world economy.
Sornette’s “Dragon King” theory has had considerable success in predicting when bubbles will burst. His research uses the strength of the positive feedback of higher return anticipation to forecast a bubble’s speed of growth and size. We agree with Sornette that in many cases the patterns of bubble formation are predictable and can be incorporated into analytical forecasts.
Taleb and Sornette’s work focuses on the largest, most dangerous bubbles, such as the rapid expansion of credit in China, but small ones that have the potential to derail particular sectors are forming all the time. Since 2005 we have systematically monitored developments in bank lending, house prices, equity prices and the real exchange rate in more than 80 countries through our macro prudential indicators. These identify potential systemic stresses sparked by a combination of rapid bank-lending growth and bubbles. This data feeds into our sovereign and bank ratings.
Constant vigilance allows us to spot new bubbles forming before most other analysts start to look for signs of overheating. The US housing market is a prime example. Most of the focus is on the recovery, but there are already early signs of bubble formation in some areas.
Our approach to the US residential market has changed, so that we now track prices against our view of sustainable house prices. Prices in some cities, especially those that never fully unwound the mid-2000s bubble, are starting to increase rapidly. For example, in Los Angeles prices are up more than 10% in the past year despite unemployment stubbornly above 10% and declines in real incomes over the past two years. Prices are now more than 75% above pre-2000 levels.
The data that we glean from our analysis of bubbles is instrumental in making our ratings forward-looking and preventing swathes of pro-cyclical upgrades as the bubble forms and downgrades as the bubble bursts.
Originally published at thewhyforum.com.
