AIQ speaks to the world-renowned risk expert Didier Sornette about the coronavirus pandemic and how organisations can stay resilient in the face of unexpected crises.
8 minute read
Rocket-fuel explosions, natural disasters, financial crises, pandemics. These are just some of the dangers risk expert Didier Sornette has studied during his distinguished career.
Currently professor on the Chair of Entrepreneurial Risks at the Swiss Federal Institute of Technology in Zurich, Sornette leads teams that track a range of global hazards, from bubbles in the real estate market to failures at nuclear energy facilities. He brings a scientific rigour to his predictive work: he was previously a professor of geophysics at the University of California, Los Angeles, where he specialised in monitoring earthquakes.
Sornette is perhaps best known for his theory of the “dragon-king”: a sudden, catastrophic risk cascade. Unlike Nassim Taleb’s black swan – which describes a fundamentally unpredictable event – dragon-kings announce themselves early through small changes in the workings of complex systems. Experts able to pick up on these tell-tale signs can ensure they are prepared when the disruption hits.
Sornette thinks it unlikely that COVID-19 could have been predicted in this way – but that doesn’t mean we shouldn’t have been better prepared. He recently undertook research to better understand the dynamics of the crisis and the efficacy of the various policy responses. His tentative conclusion is that the world has entered a new historical regime – “a phase of fear” – with implications for finance, economics and geopolitics.
In this Q&A, Sornette discusses his research on COVID-19 and the principles of good risk management in an increasingly complex and interconnected world.
Could COVID-19 have been predicted, in the manner of a dragon-king event?
COVID-19 was predictable in the sense that – as has been much commented upon – pandemics were at the top, or close to the top, of the list of concerns among risk managers across the world over recent decades. It was predictable in the same way an earthquake is predictable: we know another one is coming, but the timing is unclear.
My concept of the dragon-king refers to processes in which we see progressive damage, or collective behaviour or processes, that can be diagnosed, and which can be identified by those with the relevant skills. Once the pandemic began to spread, it followed a contagion process with a tree-like structure, which was to some degree predictable thanks to epidemiological models. But predicting the original case – “patient zero” – would have been impossible.
How do lockdowns compare with policy responses to sudden risk events in the past?
The extraordinary size and amplitude of the response is unprecedented in historical terms. Out of more than 200 countries, only four have not implemented lockdown measures. These measures cost between ten and 20 per cent of GDP in each case, according to some estimates. You can produce a list of epidemics in the 20th century that were probably much more severe than this one but didn’t receive anything like the same response.
COVID-19 should be considered in context and in comparison with other fatalities, and the global costs and consequences
Take the Asian flu of 1957, which killed more than one million people. In France alone, 100,000 people died. Our best estimate for COVID-19 is that the total number of French deaths would be between 20,000 and 30,000. COVID-19 is not minor – it is horrific – but it is far from the end of the world and it should be considered in context and in comparison with other fatalities, and the global costs and consequences.
Why has COVID-19 brought such a massive response compared with previous pandemics?
My tentative hypothesis is that the interaction of people through the internet and social media, and access to instantaneous news, has resulted in a global synchronisation of response. To use the language of physics, as in the phase transition between a liquid and solid, we have entered a new regime, a “phase of fear”. My premise is that the origin of this new phase can be identified around the time of 9/11. The terrorist acts revealed the changes in the way we see politics and were used to justify the Iraq intervention. It was the first time a single shock (of relatively minor amplitude, when put in the global context) synchronised large parts of the world in an extraordinary reaction with extraordinary consequences.
How can we ensure economies and societies are more resilient against these kinds of crises?
The universal emphasis so far has been on measures such as lockdowns, confinement, protective masks, tracing. What’s missing is an emphasis on the resilience of the individual. It is good we are defending ourselves against exogenous shocks by erecting protective barriers, but I would argue the first barrier should be about building a society populated by healthy individuals with healthy immune systems. The correlation between the severity of the illness and other comorbidity factors, such as obesity, diabetes and cardiovascular disease, has been well documented. We tend to be fatalistic about this, but can do something about it.
What kind of economic damage is the lockdown causing, and could it be creating new risks within the system?
COVID-19 is revealing the many – often submerged – issues societies have. It has catalysed the emergence of unspoken problems and unsolved issues, and it may exacerbate them. Take the European Union. The response to COVID-19 has illustrated the dysfunction and the lack of reality in the talk of “solidarity” within the EU. It has also drawn attention to problems with globalisation and the competition between China and the US.
With geopolitical friction increasing during the crisis – Trump pinning the blame on China, for example – could an escalating geopolitical crisis be a concern coming out of the pandemic?
COVID-19 might have triggered cooperation, collaboration and brotherhood: unfortunately, we have seen quite the opposite. Take the US, which bought a plane delivering a shipment of facemasks to France while it was on the tarmac. And even Europe has not shown unity. There has been some cooperation, but much less than we could have hoped for.
COVID-19 might have triggered cooperation, collaboration and brotherhood: unfortunately, we have seen quite the opposite
I see an 80-90 per cent probability that the confrontation between the US and China will continue and worsen. You have two superpowers, we could say two “empires”, and the world is going to develop more and more in this “bipolar” mode. My concern is that this competition will occur in a domain that is existential: imagine if there was a scarcity of some important commodity that, for example, China needed and the US blocked by using its navy. What is considered existential or threatening to a country’s way of life or identity varies between countries. In the West we tend to think too much as “Westerners”; we don’t put ourselves into the minds of our friends or competitors often enough. Look at what happened when the West backed Ukraine against Russia. We didn’t appreciate sufficiently that Ukraine was hugely important for Russia; historically it has been one of the last lines of defence that kept it from being invaded.
Does cybersecurity concern you, given the rise in remote working?
Cybersecurity is one of the biggest risks we face. Risks are characterised by a distribution, and the concept of a “fat tail” describes a distribution of returns that exhibit a tail that decays to zero much slower than the Gaussian distribution. Cyber risks have the broadest, wildest swings in the fat tail. Imagine, for example, Facebook being hacked: suddenly you have two billion ID thefts, with enormous consequences.
Another concern is greater digital integration in critical infrastructure, such as nuclear plants, which could be hacked and pushed towards criticality. These are big concerns. Stronger and stronger interconnection and “fragilization”, through optimising and just-in-time production, has made the system more efficient in the short term but left it more vulnerable to unforeseen shocks. I like to say that nature is more imaginative than mathematicians, physicists, engineers, specialists of all kinds. We are very often taken by surprise when a catastrophe occurs, as the path to it has usually not been imagined.
How does climate change compare with other global threats?
I am one of the leaders of a project supported by ETH Zurich and a consortium of organisations called Tellus, named after the Roman goddess of the Earth, and it’s all about trying to understand the sustainability of human-Earth systems. “Global warming” is too narrow a way of thinking about the problem. The planet is not in danger – it’s we who are in trouble, in that we are endangering the ecosystem that supports us. If we disappear, after a period of destruction the planet will thrive again: just look at the area around Chernobyl, which is now a paradise for animals.
We need to focus not on risks in silos, but on the whole human-Earth system
We need to steer the planet towards a more sustainable and harmonious future. There are many components to this. We don’t speak enough about water stress, which is a huge problem. We need to speak about the pollution caused by synthetic chemicals that have entered the endocrine system of the human body and disturb the hormones that allow our organs to synchronise and coordinate. We need to transform our industry, our way of life, our ecological footprint so that we transition to sustainability. We need to focus not on risks in silos, but on the whole human-Earth system.
You have written that even the most complex systems have “pockets of predictability” that enable us to anticipate future developments. How can we go about spotting these early warning signals?
My hypothesis, and we are continuously testing this with new systems, is that most crises, or transitions more generally, do not happen out of the blue but proceed through what I summarise as a “maturation” towards a tipping point – a catastrophe, using the language of mathematics, or a phase transition, using the language of physics.
Think of this analogy: you are a climber and you use a rope. The rope is made of many filaments. Suppose that due to stresses – your weight, the rubbing of the rope against the rockface – some filaments are damaged; one by one they break. Your weight is still held by the rope, until enough filaments are damaged that the rope breaks and you fall. Your fall would have been predictable if you monitored the progressive damage and if you understood the underlying mechanics through which the load is shared by the remaining filaments. If you can model this and monitor the damage, you can diagnose the progressive maturation of this instability.
What are the applications of this kind of work?
I first started working on this subject at the beginning of the 1990s, in collaboration with the company that later became the European Aeronautic Defence and Space Company [now Airbus]. We were interested in understanding the predictability of the failure of pressure tanks in the European Ariane rocket. And we did what I’ve just described with the rope and the climber: we subjected the pressure tanks to increasing pressure. Using acoustic gauges, we recorded the acoustic emissions that revealed tiny earthquakes in the matrix of the carbon fibres. These cracking sounds revealed delamination in the matrix, and the breaking of the little fibres. By monitoring the evolution of the cracking, as revealed by the acoustic emissions, we were able to develop a model that reliably predicted the failure of the pressure tanks.
In a sense, this same procedure can be applied to develop a sufficiently predictive diagnostic in a range of fields, even illnesses. People don’t develop a cancer out of the blue; they first have a recurrent inflammation induced by little stressors, which then evolves to chronic disease. Then, after 20-30 years, depending on the subject, it progresses to another severe phase, like cancer. A similar effect occurs in a financial bubble. The first stage is the development of a new technology, a nucleation phase, and then the first wave of investors arrives. More and more investors come to the market, attracted by the cumulative gain they have seen, and the market progresses as the positive feedback becomes more and more decoupled from the fundamental value. So common conceptualisations can be developed for predictions in each of these fields.
Your institute monitors the status of financial bubbles. Where could the next financial crisis come from, and how can investors ensure their portfolios are resilient to it?
Since the great financial crisis of 2007-’09, the markets have become dysfunctional. They are no longer the voting machines that encapsulate the collective intelligence of investors; they have become completely skewed by central bank and government intervention.
Do I see a big crisis coming? Yes, potentially, but of a very different type than the financial crisis. That was a “normal” crisis, with leverage and excess in the financial markets. Now the excesses are at the policy level, the debt level; the big issue is the level of public debt and the “whatever it takes” mantra policymakers are doubling-down on. There are still some localised bubbles, but they tend to be isolated in certain niches, in certain currencies or countries.
The long-term viability of the euro is being called into question
You have to look at the consequences of the policy and economic responses to COVID-19 and the geographical regions that have become more fragile as a result – think of southern Europe, perhaps even Europe as a whole. The long-term viability of the euro is being called into question. There are lots of unknowns as to political decision-making among the 27 countries. It’s quite obvious countries that are emerging as the big winners are in Asia; I’m quite bullish on Asia, in particular China. In a portfolio with a long-term view, I would also include resilient or “anti-fragile” assets: commodities such as palladium, gold and so on.
Should investors be aiming to model the precise nature of the next risk event, or is it more sensible to build portfolios that are resilient in a range of future scenarios?
Investors should look at a range of scenarios. The level of uncertainty has multiplied. We still know little, from an epidemiological perspective, about COVID-19. It is likely a second wave will come. Many experts are suggesting it will evolve over the next two to three years. But it will be with us forever, and that’s important to keep in mind.
Much depends on the lessons governments learn from the initial lockdown response
Much depends on the lessons governments learn from the initial lockdown response, and the implications of this for Europe and for globalisation. There is a lot of uncertainty about how new supply chains will develop as many people call for re-onshoring of industries, construction and so on. Then there is the global geopolitical situation, with the US becoming more adamant against China in their global superpower confrontation. Amid these huge uncertainties, investors will need portfolios that are resilient against a wide range of scenarios.