The future is not a spreadsheet

There is a particular kind of confidence that modern institutions find difficult to resist. Put a number on the future. Add a few decimal places, construct a spreadsheet, show a pathway to 2050, and suddenly the room relaxes. The model has spoken. Uncertainty has been converted into something manageable.

The problem is that the world we are entering does not behave like a spreadsheet.

At a recent discussion that I hosted on behalf of the London Business School Alumni Sustainability Club, Dr Ron Dembo, a pioneer of modern financial risk management and founder of Riskthinking.AI, offered a beautifully simple but deeply uncomfortable critique of how climate risk is currently being modelled. The tools the financial system are using to analyse climate risk assume a level of predictability that simply does not exist.

To illustrate the point, Dembo posed a simple thought experiment. Imagine someone told you they were modelling Singapore’s financial future in 2050 and began by assuming that interest rates would be 2.5 percent. Most people in finance would immediately stop and say hold on a second. Forecasting interest rates next month is difficult enough. Forecasting them decades ahead is bordering on the absurd. Yet climate risk modelling often begins with assumptions of exactly this kind, fixing hundreds or even thousands of economic variables far into the future before calculating outcomes.

These models are what statisticians call deterministic. They assume that if we define the pathway clearly enough, the future can be projected with reasonable confidence. But climate risk does not behave in that way. The world we are entering is what Dembo describes as radically uncertain and non-stationary. In such a system, past patterns do not reliably predict future behaviour. Volatility changes, probabilities shift, and extreme events occur outside the historical range.

In that environment the real danger rarely sits at the average. It sits in the tail of the distribution.

This distinction is more than academic. Much of today’s climate discussion still focuses on average temperature increases or orderly scenarios of global warming. Yet the economic damage from climate change is rarely caused by averages. It emerges from extreme events and the cascading consequences they create: heatwaves that buckle railway lines, droughts that shut down power plants that rely on river cooling water, flooding that overwhelms urban infrastructure, or rainfall events in desert cities that were never designed to handle them.

These are precisely the kinds of events that conventional models struggle to capture.

Dembo’s argument is that in a radically uncertain world the form of the answer must change. Instead of asking for a single forecast or a small number of scenarios, risk models must produce distributions. In other words, rather than asking what the future will be, we should ask what range of futures is possible and with what probability.

The implications are significant. Much of modern portfolio theory rests on the assumption that diversification reduces risk. Yet climate risk does not respect the tidy boundaries that financial models often assume. Infrastructure failures, water shortages or extreme weather events can affect multiple sectors simultaneously, creating correlations that were never visible in traditional financial data. What appears diversified on paper may in fact share a common exposure to physical climate risk.

In other words, the financial system may be carrying risks that remain largely unpriced.

Ultimately, however, modelling alone cannot solve the problem. Even the most sophisticated analysis still leaves decision makers with difficult choices. Once the distribution of possible outcomes is understood, boards and executives must decide what risks they are willing to live with and what risks must be mitigated.

Consider a coastal city facing rising sea levels. Should the sea wall be built three feet higher, ten feet higher, or twenty? Building the highest wall imaginable may be economically impractical, while building too low may expose the city to catastrophic losses. One possible approach is to build infrastructure that can be extended later as new information emerges. In risk management terms, the decision is not simply about prediction but about designing hedges against uncertainty.

These questions will become increasingly important as the world mobilises capital to finance the energy transition. Trillions of dollars are being directed toward renewable energy, grid infrastructure, transport systems and climate resilience projects. Yet capital allocation only works well when risk is properly understood.

If climate risk continues to be measured poorly or simplistically, markets may misprice assets, underestimate vulnerabilities and allocate investment inefficiently. Infrastructure could be built in places that will become increasingly exposed. Portfolios that appear diversified may in fact share common climate exposures. And capital that should be strengthening resilience may instead flow toward investments that unknowingly deepen exposure.

Derisking the future, therefore, requires more than financial capital. It requires a new architecture of risk, one that recognises the limits of forecasting in a radically uncertain world and focuses instead on understanding probabilities, tail risks and resilience.

The future, in other words, will never reduce itself to a single forecast.

It will always be a distribution of possibilities, and derisking the future means learning to understand, price and govern the risks that live in its tail.

illuminem Voices is a democratic space presenting the thoughts of leading Sustainability & Energy writers, their opinions do not necessarily represent those of illuminem.

Track the real‑world impact behind the sustainability headlines. illuminem’s Data Hub™ offers transparent performance data and climate targets of companies driving the transition.