Truth is ever to be found in simplicity
The world’s most sophisticated institutions modelled this war and said catastrophe. The market, with no model at all, said no. So far the market is winning – and the reasons why go to the heart of how I think about forecasting.
First published in Noise Cancelling
Truth is ever to be found in simplicity, and not in the multiplicity and confusion of things.
Sir Isaac Newton wrote that more than three centuries ago, and I have been thinking about it a great deal these past five months, because the war in the Gulf has staged as clean a test of it as I can remember. On one side, the world’s most sophisticated economic institutions and their models. On the other, the market. They looked at the same war and reached opposite conclusions.
Last week, Iran fired missiles and drones at American airbases across the Gulf. The United States struck Iranian targets five nights running. The President told Congress that America is back at war. And across that week, the S&P 500 rose by about 1%. Brent crude, at $84 a barrel as I write, sits roughly 2.5% above its five-year average – this during what the head of the IEA called in March the largest energy security crisis the world has ever faced. Gold, the textbook war hedge, is more than 20% below its January peak. And UK inflation was 2.8% in May, down from 3.4% in December, having come in below the Bank of England’s forecasts throughout the year.
The shrug, in four numbers
That is not consistent with the greatest energy crisis in history. Even the IEA’s own language has quietly softened: the latest hostilities, it now says, “could once again upend global supply and demand balances”. From the biggest crisis in history to could upend balances – without, as far as I can see, any acknowledgement that the model behind the first claim was wrong.
What the market did
The war began on 28 February. Within a fortnight the Strait of Hormuz was closed and Brent was through $100. By the end of March it was within a whisker of $120 – the peak close was $118.35 on 31 March – and physical cargoes, actual barrels changing hands as opposed to futures, traded near $150. Then, with the strait still closed, no ceasefire and no deal, the price began to fall. By the time the memorandum was signed in June, most of the crisis premium had already gone.
The catastrophe that priced itself away
What happened was adaptation, at a speed and on a scale that very little of the official modelling seems to have anticipated. Supply rerouted: Saudi Arabia pushed its East–West pipeline from around five million barrels a day towards seven, the UAE pushed more crude through Fujairah, which sits outside the strait entirely, tankers went dark and ran the strait anyway, and governments released strategic reserves – over 400 million barrels announced. Demand collapsed where it had to: Asia, the region most dependent on Gulf crude, cut consumption hard, through rationing, shortened working weeks and substitution. Painful and messy, but fast.
The equity market, meanwhile, did something subtler than taking a view on the war: it took thousands of views on individual businesses. American Airlines AAL$12.78▲1.4%American Airlines Group Inc.NASDAQ · AAL$12.78▲ 1.4%1W1-month▼ 5.5%12-month▲ 0.6%P/E, forward6.1×As of 29 Oct 2026 · not a recommendation is a useful illustration. Jet fuel is its single biggest input cost, and on the day the ceasefire died the market marked the shares down 4% while the index barely moved. Nobody modelled that. Thousands of people holding one stock asked one simple question – what does this war do to this company’s costs – and repriced it by lunchtime.
Not a shrug – a verdict, line by line
So when somebody tells you the market is ignoring the war, this is the answer. The index is the net; the information is in the parts. The market has priced this conflict line by line, and the verdicts have largely cancelled out.
Why the models keep failing
The world is a complex, adaptive, dynamic system, driven by billions of individual daily decisions made by people, businesses and governments. Those who attempt to model it respond to that complexity, understandably, by building labyrinthine models of their own. I understand the desire. But in my experience these models have generally been a very poor guide to reality, and they fail for two structural reasons.
First, because they are so big and unwieldy, they cannot keep pace with changes in behaviour – and the rerouting, the rationing and the substitution described above are precisely the kind of behavioural change a static assumption misses. The inaccuracies are structural, and they increase over time. Second, the assumption stack is the forecast. When the US Energy Information Administration published its April oil forecast, its own administrator listed the three things the model had to assume: how long the strait stays closed, how much production is actually shut in, and what a reopening even looks like – something, he pointed out, nobody had ever seen. Three unknowables, stacked, and out came a number with a dollar sign on it: a forecast that Brent would peak around $115 and stay elevated. By the start of July it was barely above $70.
None of this should be a surprise. In 2024 the Bank of England commissioned Ben Bernanke to review its own forecasting operation. His published verdict: significant shortcomings, an out-of-date infrastructure, inappropriate fixes, and a complicated and unwieldy system. That review is two years old. This April, the same institution published three scenarios for what this war would do to UK inflation, and all three, it appears, were wrong in the same direction. Throughout this year, both core and headline inflation have come in below the Bank’s forecasts. The ECB appears to have made the same mistake.
Complexity as comfort
My issue is not that the models are useless. The institutions must have them, and I would not suggest for a moment that they shouldn’t. The problem is the excessive reliance placed on their outputs, with little recognition of their limitations – a reliance that has crowded out the thing that actually works, which is pragmatic, informed judgement. Historically, wise people with decades of experience of the energy market, or the UK economy, or fiscal policy, would apply judgement to a model’s output before it went anywhere near a press release. Too often now, the output is the answer.
Why has this happened? For as long as I have studied economics, the subject has been pursuing respectability. It does not like being a social science; it wants to be a science. And in its attempt to become ever more mathematical and econometric, academic economists have come to gain credibility in proportion to the sophistication of the models they build – with the result that they compete on complexity, and lose sight of the fact that their models are inherently unable to capture what is going on in the real world. Physics, interestingly, is the opposite. It is a genuinely rigorous science operating at the very edge of what is knowable, and it is remarkably humble: it says, constantly, we don’t know. Economics has nowhere near that rigour, yet believes utterly in the output of its models. Policymakers become mired in complexity – indeed, I think they seek solace in it, because with complexity comes legitimacy and respectability amongst their peers.
The Bank of England has a version of this problem in its people as well as its machinery: too many academic economists, in my view, who are driven by the models they helped build, who have been almost perpetually bearish, and who have been wrong all too frequently. The spectacle of MPC members publicly trumpeting divergent views – which, as the Financial Times noted this week, has begun to undermine the institution’s credibility – does not help. It is instructive that the new Fed chairman is moving in precisely the opposite direction: retiring the dot plots, discouraging the public airing of committee disagreements, saying less, in the elliptical style Alan Greenspan made work for the best part of two decades. The Fed has models at least as sophisticated as the Bank’s. It just doesn’t let them drive.
The cascade
This would matter less if model outputs stayed inside the institutions. They don’t. The model produces a number; the number becomes the consensus; the consensus becomes what the media writes; and what the media writes becomes an accepted fact – a fact that was never a fact, just an output. People, businesses and investors then make real decisions against it. A wrong model doesn’t just mispredict the economy. It can move it.
We are watching a live example. The equity market has, in my view, priced this war broadly correctly. The bond market – which sits closer to central-bank thinking and central-bank speak – has exaggerated its inflationary consequences, because that is what the models told central bankers to expect. Energy prices are elevated, but the compounding effects the models predicted have simply not come through: last week’s US inflation numbers were well below expectations, core inflation fell, and the same has been true in the UK and in Europe. Central banks have nonetheless tightened in effect, by talking hawkishly – and in the ECB’s case by actually raising rates, a decision that already looks, in my view, like a misjudgement.
What I do instead
Accurate economic forecasting is inherently difficult, and in my experience it lends itself to keeping things simple. The founding principle is Newtonian. Strip the problem back to the variables that really matter, try to understand what will happen to those, and build the picture from there.
In the UK, the consumer is roughly two-thirds of the economy – the single most important driver of what happens next – so I spend my time trying to understand what will motivate households to save less and spend more, or the reverse.
Government spending is largely announced in advance, so there is little mystery in it. Net trade is, frankly, unforecastable, and I spend virtually no time on it. That leaves consumer spending and investment spending, and the forces acting on them. I read extensively, I lean on decades of experience, and I lean on a small group of economists whose judgement I trust – including my favourite economist, whose latest UK note we published on Noise Cancelling last week.
Here is what that looks like in practice, on the question that matters most right now – whether this war reignites inflation and forces rates up. The single number I am watching is private sector wage settlements. The latest reading was 2.9%, a five-year low, and still falling.
With productivity growth of around 1% a year, whole-economy wage growth of about 3% is consistent with inflation at the 2% target – which, in my view, is where UK inflation is heading by the middle of next year. If that is right, interest rates come down between now and then.
And I can tell you precisely what would change my mind: if private sector settlements started to rise rather than fall, that would spook the MPC, and would probably be enough to persuade them to raise rates. I attach a low probability to it. But that is the number, and I look at it before I look at anything a model says.
A healthy disrespect
It may sound arrogant, but my view these days is that if something is widely accepted by a crowded consensus in economics and financial markets, it is more than likely wrong. That is not contrarianism for its own sake. It is a healthy disrespect for those who like to sell certainty in a very uncertain world – and the forecasting record of the institutions that sell it, from the run-up to the financial crisis to the Brexit projections to the scenarios published for this war, has done nothing to soften the view.
So it pays to be aware of what policymakers are saying – and equally aware that they very frequently get things very wrong. I cannot offer a ready-made solution, but I would always suggest seeking out views that are not consensual, as an objective test of accepted wisdoms.
And when the official forecast and the market price disagree, as they have all year, it is worth remembering which of the two has money behind it. Newton had it right. Keep it simple, focus on what really matters, and treat certainty as the thing to be suspicious of.
Important information
The views expressed in this article are those of the author at the date of publication and not necessarily those of Curated Capital Ltd. The contents of this article are not intended as investment advice and will not be updated after publication unless otherwise stated.