"so far, the consequences have been far less disruptive than just about anyone expected."

Some analysts still think that’s because Trump’s threats have been mostly posturing. But the effective US tariff rate has already risen from 2.5 per cent to 15 per cent. Tariff revenue is rolling in at an annual rate above $300bn, roughly four times the pace this time last year.

Many economists had assumed that, by lowering imports, tariffs would strengthen the dollar almost automatically, as an accounting identity. Instead, it suffered its worst fall over the first half of a year since the early 1970s.

"The bigger mystery is why the stagflationary impact of tariffs has yet to materialise in the aggregate data."

Is the US really enjoying a free lunch, taking in $300bn a year in tariff revenues with none of the expected heartburn?

The likely answer is that the negative economic effect of tariffs is being countered by other forces, including the mania for artificial intelligence and more government stimulus. Since January, estimates of what the big tech companies will spend this year on building out AI infrastructure have risen $60bn to $350bn. Smaller businesses are scrambling to catch the wave too, further boosting growth. Meanwhile, the promise of tax relief makes it easier for US corporations to absorb a larger than expected share of the tariff costs, rather than pass it all on to consumers. Trump’s “big, beautiful bill” is expected to save US businesses around $100bn this year and more than that in 2026, mainly in tax breaks.

What the world got wrong, then and now, starts with its mental frameworks. The timeworn mistake of employing simple models, in which a headline-grabbing input A leads in a straight line to outcome B, has been greatly magnified by the global obsession with Trump. He is the only input anyone cares to analyse anymore. But complex economies are rarely shaped by just one factor, not even a shock as big as Trump’s tariffs.