Skip to content

New tool is ‘early warning’ on systemic risk

Svetlana Borovkova
Svetlana Borovkova
Head of Quantitative Modelling at Probability & Partners and Associate Professor at Vrije University Amsterdam
Security forces and rescue teams examine the wreckage of an Indonesian military C-130 Hercules transport plane after it crashed into a residential area in the North Sumatra city of Medan, Indonesia, June 30, 2015. At least 30 people were killed when the military transport plane crashed into a residential area two minutes after take-off in northern Indonesia on Tuesday, putting a fresh spotlight on the country's woeful air safety record. REUTERS/Roni Bintang TPX IMAGES OF THE DAY - RTX1IEBG
REUTERS/Roni Bintang

Is media sentiment a more reliable indicator of financial stress? At our Crisis webinar on 18 October 2016, we profile a new risk tool that would have raised the alarm up to 12 weeks before the Lehman Brothers collapse.

Negative media attention can be a self-fulfilling prophecy. By stating that a bank is not “solid”, possibly phantom troubles can become very real ones. Time and again, we see financial markets rise and fall on the tone of news rather than factual information.

I lead a team of researchers at Vrije Universiteit Amsterdam, where, using news analytics, we have created a novel tool for quantifying systemic risk — in other words, gauging the riskiness of the entire financial system.

We have constructed a quantitative measure — called SenSR (for Sentiment Systemic Risk) — which reflects how positive or negative the media sentiment is for Systemically Important Financial Institutions: the big, significant banks that have the potential to affect the whole financial system.

A man climbs up an artificially made wall of ice near the Swiss mountain resort of Pontresina January 28, 2013. REUTERS/Arnd Wiegmann (SWITZERLAND - Tags: SPORT SOCIETY TPX IMAGES OF THE DAY) FOR BEST QUALITY IMAGE: ALSO SEE GF2E9710T0F01 - RTR3D31Y
REUTERS/Arnd Wiegmann

Computer tools for sentiment analysis have been available for several years, but this is the first time that media sentiment has been used for quantifying global financial risk, and it goes well beyond simple sentiment monitoring.

Single number

There are several key innovative elements to SenSR.

First, thousands of raw news sentiment scores, produced by our News Analytics engine, are cleaned of noise and a smooth, informative signal is extracted using sophisticated statistical techniques.

For this, news items from numerous sources are included, with the greatest weights given to the most recent and relevant news items.

Next, these sentiment signals are combined with individual bank measures which reflect each bank’s relative importance: such as size, debt or leverage.

Figure 3: Historical SenSR. Black: market-value weights; blue: leverage; red: market value and leverage; orange: debt
Figure 3: Historical SenSR. Black: market-value weights; blue: leverage; red: market value and leverage; orange: debt

These size or debt-weighted sentiment measures are then aggregated, for all systemic banks, into a single number — the SenSR — which is an explicit and quantified measure of how the entire financial system is perceived in the media.

Alarm call

When SenSR is compared to recognized tools for measuring global risk (such as VIX, called the Fear Index after the famous novel by Robert Harris, or SRISK developed by the Nobel prize winner Robert Engle), a striking picture emerges.

At stressed times, such as the collapse of Lehman Brothers or the onset of European sovereign debt crisis, SenSR raises the alarm much earlier — up to 12 weeks before — than any other established measure.


Figure 5: SenSR compared with other systemic risk indicators
Figure 5: SenSR compared with other systemic risk indicators

SenSR is the closest to SRISK as both measures start rising early in 2007, whereas the collapse of Lehman Brothers occurred in September 2008.

This is remarkable, since it is based purely on “soft” and previously unquantifiable perception-based information rather than on “hard” financial risk measures, such as a bank’s capital cushion.

Jan Willem van de End, Head Researcher of the Dutch Central Bank, said: “The innovative sentiment-based method of SenSR fits well within the complex nature of systemic risk, which is often reinforced by information contagion.”

“Wisdom of crowds”

SenSR is an invaluable early warning indicator of financial stress and a monitoring tool for central banks and policy makers.

But not just that — it can be used by individuals, much like FTSE or VIX indices are used, for gauging the health of the financial system and making investment decisions.

It seems that the “wisdom of crowds”, or collective view of the media, is a better indicator of the distress in the financial system (and often an important cause of it) than fact-based measures and analysts reports.

“Things do not pass for what they are, but for what they seem”, as the 17th century philosopher Baltasar Gracian famously said. SenSR allows us to more clearly judge how things currently seem, and decide our actions accordingly.

Discover more about Quantitative Research and Trading