Long Read

Weak Signals: no surprises?

The world’s most profound problems are often the result of short-term thinking, probabilistic worldviews, consensus culture and lack of imagination. These failings in how we think about the future are at the root of crises from politics and science, to the limited horizons that inhibit radical, pioneering innovation. Weak signals are one of the keys to thinking differently. There should be few surprises.

Radical Uncertainty

Weak signals are ambiguous, barely defined stories that are easily dismissed, yet are a key missing link in how to navigate a world of growing complexity, radical uncertainty and above all, spee

To put this in context, the world faces unprecedented, well-known challenges. Extreme rainfall. Floods. Storm surges. Frequent, high intensity hurricanes. Drought. Wildfires. Inequality, in everything from income to health. Trade wars. Cyber weapons. Inter-generational tension. Growing concerns about automation, robotics and jobs. The potential and risks of artificial intelligence (AI). Fears about privacy, surveillance and media manipulation. Political instability.

Yet these surface-level, headline stories disguise a deeper, more complex and volatile world. In this environment the primary challenge is to identify potential shocks. As Mervyn King put it, writing after the 2008 crisis:

“In a world of radical uncertainty there is no way of identifying the probabilities of future events and no set of equations that describes people’s attempt to cope with, rather than optimize against, that uncertainty. ….”^[1]

King was bemoaning the failures of standard, probability-based risk models and consensus forecasts used by regulators, central bankers, financial institutions and economists. Few saw the crisis emerging. When the Queen of England asked economists from the London School of Economics why they had failed to see the 2008 financial crash coming they concluded it was a ‘failure of the imagination by some intelligent people’. In the financial markets, the language was and still is about ‘tail risk’ and ‘5 Sigma’ events. Statistical language and worldviews continue to frame the debates.

The problem is that we live in a world of possible futures that cannot be tested against historical and contemporary evidence, or easily quantified. The drivers of risk and opportunity are to be found not in probability models and in consensus forecasts, but amidst the volatile, often secret worlds of imagined futures, systemic innovation, pioneering invention, weak signals and the stories we live by.

More important, in the face of radical uncertainty, policy decisions must be made long in advance of events themselves, as political failures to take action on climate change illustrate.

Obscured Realities

Black Swans do not appear without warning. Most major shocks are preceded by weak signals.

Systemic failures and systemic innovation, like all events, are foreseen by someone, somewhere. The 2008 crisis was a failure of the imagination, for some, though not everyone.

Sometimes weak signals hide in plain sight, obscured by cultural worldviews, shared blind spots – unthinkable, unimaginable. More often, they do not fit prevailing mental models about how the world works, the mental models that we impose on the world that limit our ability to see the novel, the new.

Major, so-called ‘pioneering’ or ‘fundamental’ inventions, primary sources of weak signals, often begin as seemingly obscure ideas, expressed in new language before they develop, often through patents, to commercial reality. Pioneering invention by definition breaks new ground and creates new paradigms, beyond even expert knowledge.  These vital weak signals can be dismissed, for many reasons: threats to the status quo and the self-interest of established experts; cultural risk aversion; lack of comprehensive evidence; consensus seeking; rejection of mavericks; or failure to understand new ideas and language.  Some breakthroughs lack experimental proof.

In some senses, weak signals are the victims of confirmation bias, a bias so strong that we can deny reality until the evidence is overwhelming.

They are also victims of short-term thinking. Warnings can go unheeded because they can be ‘kicked down the road’. Denial of emerging, long-term crises is pervasive, as political leaders often demonstrate.

Probability vs Possibility

More important, weak signals are often rejected by probabilistic thinking. In a world of radical uncertainty, as Mervyn King pointed out, probability has limitations. The key distinction is between probability and possibility.

Frank Knight and John Maynard Keynes believed radical uncertainty was the driving force behind the behaviour of the economy. They lost the argument long ago to the seductively appealing world of rational expectations and probabilistic thinking that we now know create a false sense of certainty and security.

Since then, the limitations of the prevailing worldview are all too clear. Yet there is a slow realisation that models, statistics and probability are not enough. Necessary, but not sufficient.

Probabilistic methods are by definition backward-looking. AI does not yet integrate cause and effect with statistical models to better explore possible futures. AI and big data are fine for things you can say have reliable probabilities attached to them and that can be projected forward.

The problem is that politics, to take one example, is in the grip of populist and in some cases destructive emotion. Political leaders around the world do not deal in facts, or truth. Probability does not deal with emotion, human weakness, or strengths.

Or take climate and sea level rise. There are signs that Miami and several major cities along the Eastern seaboard of the US will be under water by about 2050. Novel forms of sea defence may not help. Retreat to higher ground, a culturally alien idea, may be the only answer.

Climate science has been reticent and cautious, relying on statistical models and language to shape recommendations. This is compounded by a culture of consensus-seeking that dismisses outliers. Extreme scenarios have long been marginalised. This approach, exemplified by the IPCC and the cornerstone of the Paris agreement, has proved inadequate, both politically and scientifically. Almost every consensus projection has proved too positive.

The challenging question is this: can probability-based systems help us understand everything from climate risk to the herd behaviour of financial markets? Or abrupt cultural change that emerges during culture wars?

The answer is: with limitations. Flawed algorithms and incomplete or inaccurate data create model risk, but equally important, mental models create blind spots to emerging realities.

The conventional thinking and caution have allowed political leaders to prevaricate, rather than develop options to deal with possible extreme scenario outcomes, which should be the foundation of strategic resilience.

Weak Signals: in practice

To recap, faced with radical uncertainty, we have to think differently about the future. This begins with a focus on weak signals, laden as they are with uncertainty. There may be competing narratives. Myriad possible outcomes and endgames. Hard evidence is often scant, or ambiguous. Yet weak signals can have momentum and ‘go viral’. They are the part-hidden drivers of disruption.

Take health care. The possible impacts of convergence of neurobiological sensors, breakthroughs in genetics and medical science, personalised care and predictive AI are beginning to emerge. The idea that ageing is not the defining feature of the human condition, but a disease to be treated, like any other chronic problem has momentum. ’Cures for everything’ may be the horizon, raising fundamental questions about what it might mean for inequalities in income and health. Questions about the possible systemic impacts of an entirely new worldview are at the margins.

Extreme longevity is another weak signal – radically uncertain and yet with growing momentum. Superhumans may emerge within a decade or two. Imagine superhuman cognition. Strength. Stamina. Creativity.

Privacy is another. There are prospects of fundamental technology breakthroughs that will give individuals control of their own data, as advocated by Tim Berners-Lee, amongst many others. Opposition to ‘Surveillance Capitalism’^[2] is growing. Regulatory intervention might emerge in the US, China and Europe. On the other hand, ‘total surveillance’ is gaining momentum as Western governments and major technology firms continue to ignore growing public concerns and demands. Amazon was recently granted a patent in the US for ‘surveillance as a service’.^[3]

The latest and barely reported weak signal in climate science is that the basic relationships between global temperatures and carbon emissions are understated. The ‘sensitivity’ of the relationships is a key model risk.

Miami may not go under water by 2050, but asset values might, sooner than many think. The structural adjustment, both short and long-term is vast in scale, with negative, but also positive outcomes.

Similarly, urban heat islands, an expression in wide circulation for decades in theory, has new momentum, as recent extreme temperatures in France reminded us, peaking at 45.8 degrees Celsius in late June 2019. Mass exodus from the world’s major cities may be a defining feature of the next two decades, as inhabitants make long-term choices about the winners and losers.

‘Too little too late’, a short narrative expression we have tracked over time, has growing momentum. Yet there are more positive narratives: the Green New Deal was an idea that emerged a decade or more ago, barely noticed and is now mainstream in US policy. From another perspective, the idea of well-being and health as key measures in political and economic policy, has growing importance, following the example set by New Zealand.

Another example: quantum computing, long on the horizon, may emerge short-term, (as we suggested in July 2017)^[4], delivering shocks in everything from cryptography and cyber risk, to politics, insurance and the legal landscape.

The recently labelled Neven’s law, which states that quantum computers are improving at a “doubly exponential” rate may mean that quantum machines overtake conventional computers within a few years, not decades. Hartmut Neven, director of Google’s Quantum Artificial Intelligence lab, was quoted recently as saying that “it looks like nothing is happening, nothing is happening, and then whoops, suddenly you are in a different world. That’s what we are experiencing here.”^[5]

Growing Optimism?

There are reasons to be optimistic, amidst the gathering storms and radical uncertainty.

One major shift is in long-term thinking, itself a weak signal. Regulators are asking financial institutions and corporations to develop long-term scenarios and resilience strategies, with the aim of creating transparency and maintaining stability against climate risk. The prevailing short-termism is slowly giving way to the long-term and the strategic.

This goes deeper: imagined futures and the stories we tell about them shape events in the present. Uncertainties clouding the long-term landscape are a primary source of short-term shocks. The long-term risk and innovation landscapes, contrary to conventional wisdom, dominate present-day politics, financial market sentiment, risk appetite and political prospects. One of our guiding principles is that imagined futures shape decisions and events in the here and now. This may soon gather momentum.

Recognising that some weak signals are important demands imagination and new ways of thinking about how to interpret uncertainty. This may not be mainstream in strategic and operational risk management, in corporate boards, or in political life.

Yet appropriately enough, there are weak signals of changing perspectives.

Take the intelligence services. About a decade ago, the US and the UK began thinking about how to embed imagination into standard practice. They talk openly about the importance of imagination in meeting the challenge of exploring potential catastrophic risks, in everything from climate to security. In dealing with cyber risk and underground, networked terrorist threats.

A more tangible example: the fundamental change in the long-term narrative about autonomous, electric vehicles led to the downgrades of the oil majors at the end of 2016. It led to investor activism and marked a turning point in climate and environmental politics. A new narrative emerged, gained momentum and then crystallised in investors’ imagination.

This was a weak signal one moment, a defining reality the next, even though the transformation itself may be a decade or more ahead. The lesson is to watch weak signals and prepare for shockwaves.

April 2020