Forecasts are for losers Geopolitics is all about uncertainty

Sometimes, you’ve just got to admit that you don’t know the answer to a question. When will the Iran war end? I don’t know. What will it do to the global economy? I haven’t the foggiest. I may have a hunch that it is going to run a little longer than we expected. Then again, something unexpected is bound to intrude and change things. And I’ve no clue what that might be.

But what I can do — rather than offer a range of uncertain forecasts — is assess the range of plausible outcomes to this conflict and consider how each will affect countries and people in different ways. There’s no scientifically accurate way of eliminating uncertainty; so let’s embrace it. This is not a trivial exercise: I can recall only a few times in my own life when the range of possible geopolitical and economic outcomes, good or bad, has been as wild as it is today.

With his “known-unknowns” and “unknown-unknowns”, Donald Rumsfeld has actually provided us with a useful starting place to assess outcomes. The unknown-unknowns are often more important, but they are beyond our reach. “Whereof one cannot speak,” said Wittgenstein, “thereof one must be silent”. These are what economist Frank Knight was referring to when he coined the notion of “unquantifiable uncertainty”, or true uncertainty, as distinct from outcomes that carry a small probability.

So while little of value can be said about the unknown-unknowns, known-unknowns are generally worthy of deeper reflection. We all know about the truly bad scenarios stemming from the war. Airlines may run out of kerosene in six weeks; the world could run out of fertiliser and basic chemicals if the blockade continues for much longer. We know that Asia would be hit the hardest, followed by Europe, with the US affected least.

There are also potentially good outcomes. Europe has been offered a reminder that its investments in energy independence are worthwhile — and this includes both nuclear and renewables. The European Greens were mad to oppose nuclear power, but they were right about renewables, if only from the perspective of energy independence.

Beyond the war, which should end eventually, there are other positive economic scenarios coming down the track. The most interesting part of the latest World Economic Outlook by the International Monetary Fund was not the central forecast, the risk of global recession: but, rather, the wide range of possible scenarios it also weighed up. We truly live in a world of Knightian uncertainty.

I’m talking here about plausible scenarios, those that have a reasonable chance of occurring: the known-unknowns. Of particular interest are those scenarios at the extreme ends of the spectrum. Some middle-of-the-road scenarios are plausible, of course: economic forecasters always use them for their forecasts. But they have no higher probability of occurring than the good or bad ones. What you really want is information about the spread of outcomes, not the average.

If you want to understand uncertainty, after all, the mean is your enemy. Politicians should care about the voter with the median income; the voter who, when income is ranked from poorest to richest, is sitting in the middle. He would be a different person from the man with the average income, because the super-rich distort the picture. It’s the “Bill Gates walks into the pub” problem, where a single billionaire raises the average income of a group.

There’s an old joke about the statistician who died crossing a river that was three foot deep on average. Real statisticians, of course, know about this stuff. The people most at risk of such misjudgements are those with a Statistics 101 background, who have been taught some powerful techniques they then unleash on an unsuspecting public: economists, journalists, doctors, epidemiologists and unfortunately also climate scientists.

One of the most catastrophic uses of averages is represented by the communication strategy of climate scientists, when, 10 years ago this week, they set a target to keep the increase in average global temperature to below 2°C. But what kills us is not the average temperature — the 2°C extra — but the extremes: floods, heat, forest fires.

These are the metrics the climate community should have communicated. When you’re told, after all, that the temperature is going to rise by a few degrees, you get the deckchairs out. Indeed, I heard one well-known political commentator claiming that it might be nice for the weather to be a little warmer in Northern Europe. When, on the other hand, you receive a flood warning, you get the sandbags out.

In short, how we communicate risk determines how we act, and misusing statistics inevitably leads to misjudgements.

So rather than making sweeping average claims about the future, whether about the climate or war with Iran, a better way of thinking about our planet in the medium term is by reflecting on how different plausible scenarios could affect regions in different ways.

The most important positive scenario we might consider is the AI-led productivity boom. It has not happened yet — but there are data points suggesting that it may be starting to happen. If it does, it would benefit the US a lot more than the rest of the world. After all, America invented the technology; Silicon Valley’s foundational AI models are ahead of everyone else’s; and the US’s AI infrastructure is far and away the global leader. China, for its part, is next in line. The People’s Republic may not have invented AI: but it was easily the most successful second mover.

Unsurprisingly, Europe will gain least from the AI revolution. The EU is still in data-protection mode, as much a function of its cultural stance as individual government policies. From LLMs to driverless vehicles, indeed, the Europeans have decoupled from most of our century’s cutting-edge technologies. Pharmaceuticals is the one important exception, but here too China is catching up fast. China also is ahead on renewable investments, but here, at least, the Europeans do better than the Americans.

When it comes to bad outcomes, then, Europe is at greater risk than its rivals. And I don’t just mean technologically. There is a plausible scenario that the war in Ukraine could spread, for instance to the Baltic states, and the US would almost certainly not come to Europe’s rescue. Donald Trump’s line about how the Europeans “weren’t there for us” when the US attacked Iran will keep ringing in our ears as the scenario of a European war grows — in my opinion, a non-trivial tail risk.

Similarly, the Iran ceasefire could break down and the violence spread across the Middle East. Yet while there’s a lot of uncertainty about all these scenarios, we can say with some confidence that there is no realistic scenario of war hitting America’s shores.

Absent any definitive forecasts, from this brief analysis, I can make some sharp observations. First: the US is clearly in a better position than everyone else. Second: the US is more likely to benefit from any upside scenario, and less likely to be hit by a downside one. I’d therefore warn against those heralding the end of US global dominance, let alone the end of Trump’s presidency. This is mostly wishful thinking. The President has certainly made some foolish decisions, but the US is still at the head of the pack.

Things don’t look so rosy for Europe: it will not benefit from a global productivity boom and will be disproportionately exposed to the bad scenarios. And in those areas where it is strong, it is under attack by China. East and South East Asia sit somewhere in between the two polar opposites — beneficiaries of 21st-century technologies, but not as much as the US. But even though Asia is at greater risk of disruption from the blockade of the Strait of Hormuz, it is still more resilient to shocks than Europe.

Whatever happens, one single forecast, or even a focus on two or three scenarios, can’t hope to capture the deep asymmetry of the uncertainty we’re currently facing.

But people who embrace uncertainty don’t think in terms of baselines. They don’t look for polls or projections. Nor do they seek to control that uncertainty. They seek to learn from it. Perhaps we could all learn from them. Forecasts are for losers, and once you’re comfortable admitting that you just don’t know — you might end up seeing things more clearly.

---