When rare becomes routine: Why our models can’t handle today’s risks

0
When rare becomes routine: Why our models can’t handle today’s risks

Sam Sivarajan is a keynote speaker, independent wealth management consultant and author of three books on investing and decision-making.

Communities in Europe are devastated after recent flooding, with 24 dead and thousands displaced. Parts of Vienna have lost electricity and trains have been cancelled; the water level in the Elbe River in Germany is at 5.6 metres.

This story echoes a troubling trend. The continent has experienced what is described as a one-in-100-year flood not once every century, but almost every year in the past decade. Similarly, in Canada we have seen more cases of extreme weather in recent years than were expected.

If the models are to be believed, this should not be the case. The frequency of these “rare” disasters reveals a fundamental flaw in the models used to predict them. Yet we rely on modelling, not just for weather forecasts, but even more so in finance and investing.

The “one-in-100-year flood” terminology stems from risk modelling, where the likelihood of a given event is assigned a statistical probability. What we are witnessing is the breakdown of such models in the face of growing uncertainties.

Models, by their very nature, rely on historical data and established patterns. They help predict risk, but they can’t handle uncertainty. Risk can be quantified and planned for, but uncertainty is messier. It’s the unknown unknowns – the factors that aren’t even on our radar. So we are experiencing new weather patterns that our models haven’t been designed to handle.

The same problem exists in the world of finance, where reliance on outdated models has led to financial ruin. Take former Goldman Sachs chief financial officer David Viniar’s infamous comment during the financial crisis of 2007, when he remarked that the investment bank had experienced “25-standard-deviation moves, several days in a row.”

As one commentator wryly observed: “Things were happening that were only supposed to happen once in every 100,000 years. Either that … or Goldman’s models were wrong.”

Actually, a 25-standard-deviation event is comparable with the probability of winning the lottery 22 times in a row. Under current financial models, such an event would be statistically impossible. Yet it happened (the financial crisis, not the extended lottery win). Just like in Europe’s floods, the models didn’t account for the real-world complexities and uncertainties.

When facing uncertainty, the solution isn’t to build better models. After all, the key is to never lose sight of why you are creating the model, for what purpose, who will use it, what decisions will it support them to make, how much is known and how much is assumed.

Instead, we must build resilience. Communities, governments and investors need to shift from trying to predict the unpredictable to preparing for it. This means creating systems with adaptability, redundancy and slack – whether that’s reinforcing infrastructure to withstand future floods or building investment portfolios that can weather unexpected financial storms.

In practical terms, this means abandoning the assumption that extreme events will remain rare. Instead, governments should review and rework processes so resilience and slack are built in and financial institutions should stress-test portfolios against extreme scenarios.

The same mindset shift is needed for investors who, like governments, rely too heavily on financial models that assume a world of predictable risk. Here are three ways investors can build resilience in an uncertain world:

  • Question the models: As the European floods show, relying solely on past data and models can be dangerous. Investors should dig deeper into the assumptions behind their investments. What if the models are wrong? What if the future unfolds differently? How would their portfolio perform under extreme conditions? This mindset shift is crucial to protecting against outsized losses.
  • Diversify beyond the obvious: Traditional diversification spreads investments across different asset classes, but what happens when entire sectors or regions are affected by systemic events? Investors should look for uncorrelated assets that can perform well in adverse conditions, such as hard assets, infrastructure or even sectors resilient to climate risks.
  • Build slack into your portfolio: Just as cities should design infrastructure with “slack” to handle unexpected strain, investors should do the same. That means holding cash or low-risk assets provides flexibility to capitalize on opportunities when markets inevitably turn volatile. This concept of financial slack allows investors to weather the storm without being forced into panic sales. Of course, this comes at some cost but this approach is better seen as insurance rather than investment. This runs counter to investment orthodoxy but then again, according to orthodoxy we shouldn’t be having these floods.

The irony of experiencing multiple “one-in-100-year” events within a decade should serve as a wake-up call. Whether we’re talking about climate change, financial markets or technological disruptions, the world is far more uncertain than most models or commentators assume. We can no longer afford to operate on the assumption that rare events are truly rare.

Instead, the focus should shift from predicting these events to building systems – physical, social and financial – that can survive them. As Europe’s recent floods remind us, models are helpful, but they’re no substitute for resilience. It’s time we adapt to an increasingly uncertain world. As British statistician George Fox said: “All models are wrong, some are useful.”

link

Leave a Reply

Your email address will not be published. Required fields are marked *