Selecting relevant market observations

Forecasting the Improbable

“Sea levels could go up as much as three-quarters of a meter in this century, but there is a reasonable probability it could be much higher than that.”

Steven Chu, 1948-, American physicist and Secretary of Energy.

When creating a list of potential return outcomes based on events for an investment portfolio over a given time horizon, the practitioner typically assigns a probability value for each one of the “discrete” outcomes. The Gaussian “bell-curve” shaped distribution is just a more refined version of the same thing with the difference being that the outcomes are continuous rather than discrete. The “tail” ends of the curve represent the more extreme return outcomes that tend to have a very low probability of occurrence. Such things as war, bubbles and natural disasters can trigger a financial meltdown, which can produce extremely negative returns. They do occur, but with much lower frequency than the less extreme returns that are assigned a much higher probability of occurrence.

Typically, when thinking up extreme events, the practitioner draws from past experience or historical events that have had a material impact on financial markets. The subprime led crisis and the dot com bubble are but two examples of this. Things become more complicated when the event in question has never occurred in the past or occurred such a long time ago that there is little material to recollect its impact. Another issue with assessing the financial impact of events that took place a very long time ago is that we need to adjust for all the economic and financial progress that occurred between the past and the present.

A Mega Black Swan…

To illustrate this point we look at the probability of occurrence of a solar megastorm over the next decade as an example. A solar megastorm is a very large version of a solar flare, which is a sudden brightening, observed over the sun’s surface. Solar megastorms generate auroras, which are visual expression of electromagnetic radiation hitting the Earth’s atmosphere, and, given enough power, are known to knock out electrical systems.

The largest observed solar megastorm ever occurred on September 1st, 1859 and was named the “Carrington Event” after the British astronomer who discovered it. According to newspapers of the time, auroras where observed across the globe and caused significant damage to electrical systems.

A recent paper published in “Space Weather”, an international journal of research, puts the probability of a solar megastorm occurring over the next decade at a significant 12%. Taking into consideration the technological progress over the last 150 years and the degree to which our modern economy has become dependent on electronics, it doesn’t take much to realize the potential havoc that such an event would cause. Auroras are known to damage electrical power grids and may contribute to the erosion of oil and gas pipelines. They can disrupt GPS satellites and disturb or even completely black out radio communication on Earth. The National Research Council, the working arm of the U.S. National Academies in its 2008 report puts an estimate of between 1 and 2 trillion dollars in terms of collateral damage in the U.S. for a Carrington-style solar storm.

Financial Armageddon…

If the outage were to last over the longer term, it could lead to a disruption of transportation, communication, banking and finance systems, access to water and lead to a spread of diseases through the lack of refrigeration.

In summary, such an event would have catastrophic consequences on our financial system of an unprecedented nature, both in the short and longer term but even more worrying is the one in eight probability of it occurring within the next ten years. This is huge and if true, you may be wondering why it’s not in the front pages of the news. One reason may have to do with the fact that such an event is so outlandish, (it last occurred 150 years ago, damages were very limited so it is erased from our collective memories), that we prefer to ignore it. This form of “hindsight bias” is a well known feature of behavioral finance whereby we tend to put more emphasis or weight on past experiences. Whatever the reason for not taking it into account, the truth of the matter is that the most damaging form of disaster tends to be the one that is not accounted for, which means that a practitioner should always keep a very open mind when assessing potential risks.