Please see the below article from Tatton Investment Management detailing their discussions on markets over the past week. Received this morning 19/10/2026.
Bonds cracking the AI trade
Bond yields had been climbing but then the remarkable happened last week: demand for bonds jumped, and stocks sold off. Yields at attractive levels, or waning tech optimism – which was the stronger motivation?
Meanwhile, France suffered destruction through social unrest not witnessed since 1911’s “Vie chère” or – just like today – high cost of living riots. The debate about reducing the relentless growth of the government deficit has seen disaffected urbanites being goaded into acts of violence. Over the weekend, Macron’s speech to the nation was met by a lot of political opportunism, with opposition politicians failing to seek restraint clearly. Remarkably, the Euro has remained stable and European equities are opening unchanged. Investors may be viewing the events as signalling a growing political will to deal with the debt issues, as nation heads into the Presidential election on 18th April next year.
The sharp fall in US (and other developed world) yields was, counterintuitively, prompted by surprisingly strong US inflation. The initial yield reaction was up – as you’d expect – but that rapidly reversed once investors realised that the Fed will likely raise rates this month, rather than in December. Long-term inflation expectations stayed level, but growth expectations took a hit – reflected in the fall of the real yields component of total yields.
Corporate bond yields for big tech companies went the other way, and share prices dropped. The UK was one of the few positives, buoyed by a stronger-than-expected GDP number – and the fact that some international investors (perhaps incorrectly) see the UK as a hedge against the AI trade.
The drop in tech stocks made global equities negative in aggregate. This tells us that the stocks-to-bonds shift was more about waning belief in AI’s seemingly never-ending profit growth, rather than yield attraction. Stories about slowing demand for AI ‘compute’ pushed up borrowing costs for datacentre firms – with bonds of CoreWeave – the data centre construction company – reaching distressed territory. It’s unclear how accurate the slowing compute story is, since the metrics for judging AI firms’ success are opaque. But the narrative is more important than reality in the short-term; and investors may well be looking for excuses to crystallise 2026 gains.
The dollar gave up recent gains, almost in lockstep with US real yields, reflecting waning optimism in US asset returns. Currency volatility hurts international investors and it could make foreign investors repatriate capital if it carries on. That could compound stock market volatility.
The antidote, as ever, is earnings. Q3 results already look strong and should renew faith in the AI theme. The only concern is if AI firms reveal even larger debts than investors realised. That would temper, rather than destroy, investor sentiment.
Budget suffers from the government’s own gilts
Chancellor John Healey faces a difficult Autumn budget. UK government debt is higher than it has ever been during peacetime, and the Institute for Fiscal Studies (IFS) thinks fulfilling his previous defence commitments would create a £35bn fiscal hole. Its report this week also highlighted a policy that hurts both the budget and UK bonds: the government’s accounting rules on quantitative tightening (QT).
While most central banks have stopped QT, the BoE has kept selling bonds. That’s one of the reasons gilts have underperformed other bonds. But QT doesn’t just raise the government’s debt costs by pushing up gilt yields; it also costs the treasury directly, through the way they are accounted for. The BoE’s gilts are held on a separate Asset Purchase Facility (APF), which the government indemnifies against value fluctuations. This goes both ways – gains are paid to the treasury, but losses have to be covered too. Other governments apply similar accounting terms, but the UK is unique in that losses on its own debt impact actual cashflow. It looked like a blessing during QE but now it is certainly a curse: the treasury has paid about £30bn a year since 2022.
We shouldn’t overstate the yield impact. Gilts’ main issue is structural imbalance (i.e. too many long-term and inflation-linked bonds), compounded by QT. But at least those don’t impact day-to-day spending.
You could argue the APF payments deter fiscal profligacy, but that’s a fairly weak argument. In the short-term it just makes the fiscal position worse, which increases volatility and makes international investors squeamish. Changing the policy could therefore help more than hinder gilts.
There’s no sign Chancellor Healey will change this in his budget. But with the government wanting extra spending and reluctant to cut, tax or borrow, technical adjustments like aligning gilt accounting with global standards might be all that’s left. Commentators typically scoff at such fiddling, but it can be more important than you think.
The global diesel pressure point
Nowhere is 2026’s global energy shortage more apparent than in diesel prices. Wholesale diesel has more than doubled in price since the start of the US-Iran war, and the Trump administration has made it a geopolitical priority. Why are politicians, and markets, so focussed on diesel?
Diesel didn’t use to be more expensive than regular gasoline; it was often cheaper before 2004. Environmental requirements made it costlier to produce, while its fuel efficiency and the rapid expansion of global supply chains pushed demand steadily higher. Diesel’s role in transportation and industry made it a truly global market with a consistent premium. That premium widened after Russia invaded Ukraine, and the Strait of Hormuz crisis has destabilised it again. Demand is sticky, which is why the IEA expect prices to rise further.
Diesel’s global economic importance makes the shortage a geopolitical issue. Trump threatened to ban US exports, likely to force the EU to release emergency supplies – which it did, with the G7 agreeing to draw down 100mn barrels of crude and diesel. Trump also announced a sanction relief deal releasing 300,000 tonnes from Russia (less than a single day’s US consumption, ~530,000). The president talks tough on an export ban but it would be politically difficult: the biggest diesel consuming US states are also the biggest oil producers, so any relief a ban provides on prices, also takes away revenue from producers.
We wouldn’t count on reserve releases to fix the market. JP Morgan suggests the 100mn figure includes barrels already promised, and Europe is reluctant to release more. Reserves are designed for acute shortages, not price emergencies. The upshot is that we expect diesel to remain a pressure point for the global economy until the Strait of Hormuz blockades are lifted. Nations can’t keep releasing reserves forever.
Please continue to check our blog content for advice, planning issues and the latest investment, market and economic updates from leading investment houses.
Alex Clare
19/10/2026