How To Anticipate A Market Sell-Off

Analysing systemic market risk can come down to surprisingly few components

Reviewed by: Allan Lane
Edited by: Allan Lane

Some markets are red hot. Just look, for example, at the performance of the Shenzhen Composite Index's 150 percent plus returns in the last year –it's hard to ignore.

But with reward comes risk, and with risk comes the day of reckoning. As Greece threatens to leave the Euro and there are murmurings of an emerging market (EM) sell-off in the event of a U.S. interest rate hike, not to mention the potential bubble in Chinese equities, it is perhaps time to tackle the issue of how to anticipate systemic market risk.

Let Me Introduce Mark Kritzman

It is nigh on impossible to avoid those times when all asset classes go down together. One such time was on 22 May 2013 when Bernanke indicated quantitative easing was coming to an end, resulting in what we now affectionately refer to as the first taper tantrum.

Mitigating against such events is very hard to do, but one of my favoured techniques, which is based around the research by Mark Kritzman of the MIT Sloan School of Management and published in partnership with State Street about four years ago, aims to do just that.

Kritzman's proposition is based around an indicator that measures cross-market turbulence and keeps track of the number of independent factors that are driving returns across all key asset classes, known as the absorption ratio.

He also developed the so-called turbulence indicator. By simultaneously looking at returns of a wide set of asset classes during certain timeframes, he could extend the concept to assess the whole market and illustrate when it was acting as a single entity – in other words, one boat that rises and falls with the tide.

Read more about Kritzman's research here

Based on Kritzman's research, our in-house turbulence indicator at Twenty20 Investments combines the concepts of turbulence and absorption ration as a gauge for risk. We use a certain time series for a range of benchmark indices to model the full market, covering government, corporate and EM bonds, developed and EM equities and developed and EM currencies.

For a larger view, please click on the image above.

The above chart clearly shows that when turbulence rises and the blue line spikes, the market – the red line – goes down. Between June 2004 and April 2015, this relationship can be seen most clearly during the crash of September 2008.

But How Does It Work?

Grouping together stocks from the same sector makes sense. For example when the tech sector is booming, it's not unusual to expect those stocks to all rise and fall together. And yes, there are other factors that drive how the stocks will perform, such as their size and P/E ratios. However, for every 100 tech stocks, there are only a handful of factors that explain the majority of their returns.

Kritzman's insightful idea was to study the time series of returns for a wide range of asset classes and observe on a daily basis how many factors – or "principal components" – were needed to explain 85 percent of returns. Under normal market conditions, it transpires that one might observe up to four or five factors at work, but when the market throws a proverbial tantrum, only two or three factors seem to be driving the lion's share of the returns.

Perfect Timing: Just A Coincidence?

Here is where it gets interesting. During the last decade, there have been very few occasions when the majority of market returns were being driven by five separate factors. Likewise, very rarely has the total market acted in unity due to just two factors. In a loose sense, this 2-factor situation has served to act as a signal for the subsequent extreme sell-offs that have happened in equities and many other asset classes.

Date of Lowest Principal Component Analysis

Date of Market Event





BNP freeze collateralized debt obligation funds




JP Morgan buys Bear Stearns




GBP, USD, EUR rates all cut to 50 bps




Greece is bailed out for the first time

Looking at the above table, we see a very close relationship between those instances when the market acted as a two-factor entity and some famous market events that became the hallmark of the great financial crisis. The same four events – A, B, C and D – are shown below, and the Y axis shows how many factors are driving market turbulence.

For a larger view, please click on the image above.

Using This Rationale To Evaluate EM

Back to those red hot EM opportunities I mentioned at the start. There are signs many investors are being attracted back into EM based on today's valuations. With the market currently acting in a four-factor mode, this is an encouraging sign. That rational view, however, will go out of the window should another big blip reduce the number of factors that are driving returns.

The evidence from the last few years does seem to indicate that EM is the asset class that is first to cave in during turbulence. By the same token, this might be exactly the right time to go overweight a number of EM exposures that offer better value than their developed market counterparts.

Allan Lane is founder and CEO of Algo-Chain.