A Very, Very, Very, Very Black Swan? by Jay Leopold, ColumbiaManagement

  • Investors should understand the risks in their portfolio, and be cognizant that black swan events can occur much more frequently than models suggest.
  • Risk models are extremely helpful when thinking about portfolio construction, but shouldn’t be relied upon exclusively.
  • It is important to avoid a sense of complacency after an extended period of stable volatility or correlations.

Nassim Taleb’s book “The Black Swan” effectively demonstrates that seemingly highly improbable events are much more common than expected, often with significant consequences. In fact, experts are often blind to these occurrences because past data is not necessarily a good predictor of the future.

Most investors are aware a black swan event hit the Swiss franc earlier this month. After capping the value of the Swiss franc (CHF) relative to the euro at 1.20 since 2011, the Swiss National Bank (SNB) surprised the markets, suddenly abandoning its policy. The franc instantly soared by around 19% versus the euro, causing tremors in various corners of the capital markets.

Currencies generally don’t move in big chunks like was experienced by the Swiss franc on January 15. In fact, during the year ending January 14, 2015, the CHF/EUR traded in the 1.20-1.24 range, and volatility averaged just 0.1% per day. A fairly small move of at least 0.2% would occur in just 5% of days and a change of 0.3% or greater would be expected less than 1% of the time. Goldman Sachs’ CFO was quoted last week that the move in the franc was “something like a 20+ standard deviation move,” while the math suggests it was actually a much rarer event. For perspective, a six standard deviation normally distributed event occurs once every 1.4 million years. If there were such a thing as a very, very, very, very (repeat “very” for a long time) black swan, this 20+ standard deviation event would be it. But previously unexpected moves like the CHF experienced occur with seemingly regularity, proving Taleb’s point that these types of events happen much more frequently than the models suggest.

So did the seemingly impossible just happen, are the risk models flawed, or are the models being relied upon incorrectly? I think it is the latter. The models are mathematically sound, but most rely on some key assumptions including stable volatility and correlations, as well as normally distributed outcomes. These models would correctly predict the incredible improbability of the move in the Swiss franc if the past year’s volatility in the CHF/EUR were stationary. Those that banked on the past to predict the likelihood of future price changes received a rude awakening.

FXCM, Inc., an online foreign currency exchange, made highly-levered margin loans to its customers based on the assumption that past volatility would continue. When the assumption failed, the appreciation of the CHF ripped through its customers’ posted margins, leaving FXCM suddenly holding the bag for its customers’ busted trades; its balance sheet had a $225 million hole to fill, overnight. The company faced sudden insolvency, but has been thrown a lifeline in the form of an extractive emergency loan by Leucadia National that apparently has wiped out most of the value for common shareholders.

While some funds benefitted, others were damaged. As a group, hedge funds were more short the CHF than at any point in the past 18 months, according to the CFTC, based on the belief that the SNB would continue to cap the Franc’s value. Some went belly-up overnight. This story is reminiscent of Long Term Capital Management’s leveraged trades in 1998 that brought down that previously very successful firm.

Investors uninvolved in trading the Swiss franc (aka: “most of us”) can learn important lessons from these events. Risk is not a bad thing. In fact, it is necessary in order to earn a return. Risk models can be extremely helpful tool to understand relative exposures in a portfolio, and used as a framework for portfolio construction in an effort to maximize risk-adjusted returns. They are particularly accurate when volatility is stable.

But it is important to remember the future is unknowable, and markets are constantly changing and very unpredictable. Risk models are extremely helpful to construct well-rounded portfolios, but shouldn’t be relied upon exclusively. Independent thinking, experience, pre-mortems (asking oneself “if I am going to be wrong, why would that be?”), and out-of-the-box scenario analysis can be very important complements to risk models.

The seeds of this “improbable” event were sown several years ago when the SNB decided to cap the franc relative to the euro. In my opinion, the unprecedented and unconventional monetary policy adopted by the United States, Japan and recently the EU will spawn other black swans in coming years.

Investors should regularly stack up their current portfolio to their long-term objectives and time horizon. Specifically, it is important to avoid a sense of complacency after an extended period of stable volatility or correlations. Understanding obvious and hidden risks in a portfolio, and being cognizant that supposed black swans can occur much more frequently than models suggest can help investors compound their returns at greater rates in the long run.

A Very, Very, Very, Very Black Swan?