Back in the 18th century, the cycle was seen in Britain and France, with the South Sea and Mississippi bubbles respectively. It occurred in 1920s America when investors bought stocks on margin (only paying part of the price upfront) and of course it was behind the 2007-2008 financial crisis when NINJA borrowers (those with no income, job or assets) were allowed to buy property and elaborate financial structures were built on the assumption that US house prices would never fall.

This time round, the speculation has been built on stocks in the artificial intelligence industry. As in the late 1990s when everyone expected the internet to revolutionise the economy and invested in fibreoptic cables, this time round the corporate sector is investing in data centres and computer chips. In a feedback loop that George Soros dubbed “reflexivity”, confidence in the economic benefits of a boom leads to increased demand for the products made by the booming sector, which appears to validate the rationale for continued investment.

The scale of investment is staggering with Amazon, Google, Meta and Microsoft alone set to spend $1.5 trillion in 2026 and 2027. The tech companies are in an awkward position. If they don’t invest, they may lose their dominant market position (which generates their profits and boosts their share price). But investment on this scale starts to undermine their appeal as well; Google burned through $6bn of cash in the second quarter, the first period of negative cashflow since its flotation. As a result, the tech companies are having to fund more of their capital expenditure through debt; Morgan Stanley is forecasting $570bn of AI-related debt issuance this year.

This is having two effects in the bond market. First, it is causing investors to become more concerned about defaults; the cost of insuring against default by a “hyperscaler” (i.e big AI spender) has gone from being a third of the investment grade average to 80% above it, according to Societe Generale. Second, the scale of the debt issuance may be putting upward pressure on yields more generally, especially as the US government’s deficit, at $1.4 trillion in the first half, is still trending higher. Higher bond yields weigh on economic growth (and make the AI capex more expensive to finance).

All this may not matter, of course, if AI is such a transformative technology that it pays off for the hyperscalers and boosts economic growth. But the evidence to date is inconclusive. A Fed study of earnings transcripts found that, while companies were very optimistic about the productivity-improving potential of AI in the future, very few had actually seen any benefit up till now. Maybe they are right to be optimistic or maybe they have just bought into the hype.

Considering all the messianic talk of AI’s transformative powers, it hasn’t worked yet, except by boosting capex spending. Indeed, on an encouraging note, the trade expert Richard Baldwin finds, by looking at the export statistics for “other business services”, that jobs in this area, which looked vulnerable to AI, are booming.

So on the bubble front, we have signs of speculative trades going wrong, massive investment being funded by more expensive debt and a lack of evidence of AI’s effectiveness. If we haven’t passed the peak already, we must be close to it.

But who is “the establishment” in this case? Farage’s main opponent is a comedian who stands under the moniker “Count Binface” and wears a bin on his head. The count’s manifesto includes the abolition of auto-renewals on subscriptions, a promise to build at least one affordable home, and a proposal to limit the cost of 99 ice creams to 99p (it’s a British thing).

Farage went to a private school (Dulwich College) before spending 20 years working in the City, as a commodities trader and then broker. He has been a serving politician since 1999 when he was first elected to the European Parliament. Since then he has led three political parties, UKIP, the Brexit party and now Reform. City-boy turned politician is a classic establishment career path.

As for his funders, Christopher Harborne, a billionaire businessman with interests in the crypto industry; his background includes education at the elite Westminster private school, followed by a degree at Cambridge, an MBA at Insead, the French business school, and a spell working at McKinsey, the management consultants. The other source of funding that has caused controversy is the family of George Cottrell, a long-term associate. Cottrell is known as “posh George” for a reason; his father was a landowner who went to Gordonstoun, the same school attended by King Charles, while his mother actually dated the future King. His maternal grandfather has the title Baron Manton, while his uncle is the third Baron Hesketh. These are very establishment backgrounds that are a lot more upper-class than Count Binface.

On the other side of the pond, Donald Trump is a rich businessman who was the son of another rich businessman, and who attended the Wharton business school. He has surrounded himself with other rich people such as Howard Lutnick and Scott Bessent, who made their money in finance. He briefly put the world’s richest man, Elon Musk, in charge of cutting public spending while many other billionaires were front and centre at his inauguration.

So we are not talking about working-class heroes who are planning to overthrow the class system or redistribute wealth. Indeed, one of the most significant acts of the first Trump administration was a 2017 tax plan that was heavily skewed towards the rich. Trump was once a deep sceptic abut cryptocurrencies but is now a huge enthusiast who has made a lot of money from them; Farage has called for London to “embrace” crypto.

Regular readers will know that I am not a believer in cryptocurrencies. Bitcoin was set up in 2009, two years after the introduction of the iPhone. while most of us will have some sort of smartphone in our pockets, few of us will have ever had a practical use for a cryptocurrency. They are mostly used for speculation and tax avoidance.

Indeed, there is a “magic beans” element to the whole affair. Someone creates a digital unit with no asset backing and invites others to buy it; when they buy it wealth is apparently created (in reality, it is a transfer of wealth from outsiders to insiders). Enthusiasts claim the value of cryptocurrencies stems from their limited supply but this is not true in aggregate (there are thousands of them). And a limited supply is meaningless unless the product has value; my nasal hair is also limited in supply but no one is bidding for it.

What cryptocurrencies do well is to make the founders very rich and this “new money” wants to make its mark, and to influence power. In the 18th century, Britain had “court” and “country” parties; those politicians who were close to the reigning monarch and those who were not. Occasionally, these groups swapped places but their differences were not primarily ideological. Nor were they the least bit interested in advancing the interests of the non-voting masses.

The populists are really the new wealthy who may share the same backgrounds (private schools, finance) as the old wealthy but who either feel disrespected by the old elite (a key motivation of Trump) or want the rules altered to make them even wealthier.

Since western countries are now democracies the new rich must gain power through the ballot box and they do so through advancing a nationalist message that they think will appeal (even though their own money, or their backers, may be based overseas). They don’t represent “the people” at all, just their own interests. Worse still, because their illusory appeal comes from being apparent outsiders, they want to throw out everything the mainstream believes, from the reality of climate change through the benefits of vaccination to the importance of international cooperation.

Count Binface, by contrast, has a much more sensible manifesto.

When it came to his war with the “Persians”, the main oracle consulted by Donald Trump was Benjamin Netanyahu, the Israeli leader. But as I argued back in March, Trump’s decision may have been based on overconfidence after the apparent success of his capture of Nicolas Maduro, the Venezuelan leader. He may have thought that the Iranian regime would quickly crumble.

More than four months later, the US is still bombing Iran while Iran is still sending missiles to attack shipping in the Strait or Hormuz and US bases in Middle Eastern states (even though the President said that Iran’s missiles were “just about used up or beaten” back on April 1).

In the intervening months, Trump has made repeated statements that a ceasefire was imminent and that the Iranians were desperate to make a deal. When a ceasefire finally arrived in the form of last month’s “Memorandum of Understanding”, Iran seemed to have the upper hand. William Hague, a former British foreign secretary who spent a long time negotiating with the Iranians, writes that “the concessions made to them seemed extraordinary”.

Trump clearly made that deal because he wanted the Strait of Hormuz to reopen and for oil prices to fall so that the costs of US motorists would decline in time for the mid-term elections in November. But his eagerness was too obvious to the Iranians, as was his unwillingness to escalate, despite his frequent bloodcurdling threats. The Iranians realised the power they had over oil traffic in the Strait had have continued to exercise it. As of July 12, traffic through the Strait was at a five-week low.

So Trump’s latest idea is to blockade the Strait all over again (even though this did not budge the Iranians when he previously did it) and to charge a 20% fee on the cargo of ships that want to pass through safely. Ipek Ozkardeskaya of Swissquote calculates that this will equate to a $30m plus fee on a typical oil tanker. Even if a shipping company pays it, there is no guarantee the tanker won’t be hit by an incoming Iranian missile.

Small wonder that the oil price has surged again. Some of the economic damage was contained during the spring by the ability of countries and companies to run down their oil inventories but this trick cannot be sustained forever. And the oil price may be pushed even higher by Ukraine’s success in using its drones to disrupt Russia’s oil supplies. (A further risk is a resumption of hostilities between Saudi Arabia and the Houthis of Yemen.)

It seems unlikely that Trump can actually impose his levy (see update below). But it is illustrative of his muddled thinking that, in a supposed desire to push the oil price down, his instinct is to impose a levy that pushes the costs of oil companies up. It is like trying to combat a bread shortage by taxing flour.

This muddled thinking also lay at the heart of his trade policy, which I criticised in my book “The Economic Consequences of Mr Trump”. The trade deficit is not the most serious problem facing the US economy and is not mainly the result of other countries “cheating”. Tariffs will not solve the problem since they force up the costs of US producers, which buy a lot of components and raw materials from abroad. Sure enough, the US economy grew more slowly in 2025 than in 2024, the trade deficit was unchanged and manufacturing employment fell, not rose.

The Iranian war started with Trump’s desire for a quick victory and it ended with a poorly-negotiated interim deal because of his desire for a quick exit. It has not helped that Trump relies on cronies rather than expert advisers and tends to assume that other Presidents have not followed his chosen strategy because of a lack of gumption, rather than a clear understanding of the risks involved.

His only choices now are a rumbling conflict that is not an official war, but might as well be; a full-scale invasion that would involve mass US casualties and the kind of morale-sapping long-term commitment that occurred in Vietnam or Iraq; or a humiliating climbdown that will leave Iranian’s theocratic regime in place and controlling the Strait of Hormuz while having every incentive to develop a nuclear weapon capability.

All told, it is a bungle that might not rank with Croesus’s lost empire but could mean that Trump loses control of Congress. What is also striking is how forgiving the markets have been of Trump’s mistakes.

Admittedly, excitement over the AI boom is offsetting Trump’s policy errors. The markets have also learned to discount geopolitical events as short-lived. However, I also think there is a naive belief among investors that Trump, as a rich Republican, must be good news for them in the end. He wants to cut taxes and regulation; nothing else matters.

I have quoted before this astonishing quote from Torsten Slok of Apollo Investment Management, back in April.

“Markets are overreacting to what will likely be a 4- to 6-week period of volatility, which will ultimately result in 50 years of stability in oil markets, supply chains and geopolitics.”

Such naivety seems to be all too common in financial markets. In reality, the leader of the country with the biggest economy (ignoring PPP measures), largest military and most influential financial markets is a narcissist who refuses to listen to informed opinion and whose main focus seems to be personal enrichment. That ought to make investors worried but ideology has blinded them. As in the Wizard of Oz, it remains to be seen how long it will be before the curtain is drawn back to show investors the incompetent old man behind it.

UPDATE: Since the post was written, Trump has already abandoned his 20% levy. It will be replaced, he says, with investments by the Gulf states into the US.

Those Investments will be MASSIVE but, at the same time, extraordinarily good for them, and their future. As everyone is aware, we have the largest Dollar Investment into the United States, of any Country in History, but these new Investments will make that Number even larger, and we will see Factories, Plants, and Equipment pour into the United States at Historic levels, which will create additional millions of High Paying AMERICAN Jobs!

America is WINNING again, winning like never before. The days of Iran killing hundreds of thousands of people, including 52,000 protestors, are OVER and, most importantly, IRAN WILL NEVER HAVE A NUCLEAR WEAPON!

Readers will recall that Trump has promised massive investments by other countries before. (By the way such investments, if they happen, would push up the US’s capital account surplus and would inevitably increase the current account deficit. That’s how trade accounting works.) But the Gulf states may have just made vague promises to head off the shipping levy, a tactic I dubbed the UNICORN last year (UNimportant COncession to Resolve Negotiations).

But in recent weeks, there has been a smattering of good news that gives some hope for the future. First, the oil price has declined sharply. While this may be the result of overoptimism about the state of the US-Iran peace talks, it is good for the global economy. Consumers will have less pressure on their wallets and more money to spend on other goods; central banks will be less inclined to push up interest rates. That should boost both activity and jobs.

Second, the mess that Donald Trump has made of his Iran policy is clear for all to see. J D Vance was recently trumpeting Iran’s agreement to allow nuclear arms inspectors into the country. But Iran had allowed nuclear arms inspections under the Obama deal that Trump pulled out of and the country says it will not grant inspectors access to many sites. Meanwhile, Iran has achieved sanctions relief, even though the strait of Hormuz is not fully open, and it has not agreed to surrender its enriched uranium. Meanwhile, it retains a ballistic missile programme and the theocratic regime is still in place. When Trump said the war would only end with “unconditional surrender”, we didn’t realise he meant that he would be surrendering.

All this, together with the blatant corruption of the administration and the narcissism of the President, has dented Trump’s approval rating, which is well below 40%. That makes it more likely that the Democrats can gain control of at least one chamber of Congress in November, which will put a bit of a brake on the President’s actions.

Third, Trump’s unpopularity with European voters, and his tendency to insult even sympathetic allies (such as Italy’s Giorgia Meloni) mean that even the continent’s far-right leaders are trying to disassociate themselves from him. As with the Canadian election, it may be that an association with Trump is electoral poison in Europe.

Fourth, I wrote about the “sliding doors” moment in Britain, linked to the Makerfield by-election. The Labour candidate, Andy Burnham, seized the moment and won the by-election at a canter. He will shortly become prime minister. He is very definitely not the Messiah and will face the same pressures as his predecessor; trying to balance the many demands for government spending against public dislike of higher taxes and the bond market’s limited appetite for increased borrowing.

It is still very good news that Burnham won, because it means that Reform lost its fifth significant by-election in a row. That dents the feeling of inevitability of a Reform government which could have boosted its electoral prospects; it also means that moderate voters are willing to back whichever candidate they think most likely to beat Reform whether that is Labour, Greens or Plaid Cymru. Reform might squeeze into power with under 30% of the vote but they won’t if the other 70% of the electorate vote tactically.

In short, the global economy has received a boost and the fortunes of Trump, Farage and others of their ilk have declined. Having said all that, there is still a lot to worry about. My last post mentioned the bubble signs in the market, and The Economist recently published an apocalyptic guest article on the potential for a crash. Trump continues to threaten tariff increases, may yet resume military action in the Middle East (if only out of a sense of humiliation) and is unlikely to accept electoral defeat in the mid-terms.

But after the defeat of Orban in Hungary earlier this year, we can see that the arc of history does not always bend towards bigotry and injustice, as seemed to be the trend. There is more hope than there was six months ago.

What tipped me over the edge was the other person sitting in reception. The young man concerned was on the phone boasting about all the money he had made trading in dotcom stocks and which stocks he was about to buy next. This seemed the equivalent of the story of Joe Kennedy in 1929, who realised a crash was coming because his shoeshine boy was giving him share tips. I didn’t take the job; wisely as it turned out because the company did not survive and my equity would have been worthless.

Identifying a bubble, let alone calling the top of one, is insanely difficult. There are finance professors in Chicago who think the concept is ridiculous; the market, like the heart, has its reasons. One academic even tried to justify the tulipmania of the Dutch 17th century (come on, they are flowers doomed to rot).

With artificial intelligence, there are clear elements of reflexivity, a term dreamed up by George Soros to describe how the views of participants influence the course of events. If people believe AI will be successful, they will invest lots of money in AI-related technologies. This burst of spending will only make the success of AI look even more plausible.

That said, I think there are three clear parallels with the 2000 era at the moment.

1. Advertising. In 2000, the Superbowl was marked by ads for 14 different dotcom companies in the commercial break. Since the Superbowl is a very expensive advertising opportunity, this was a sign both that dotcom companies had money to burn and that they were desperate to catch the notice of consumers. This week, just in Britain, the Times had a football-linked wraparound from ChatGPT, on the day the World Cup opened, while ITV’s World Cup coverage is sponsored by Gemini, Google’s chatbot. AI is trying to break out into the mainstream consumer market.

2. Cashing out. Shares in private companies may make you rich on paper but to cash out, you generally need the company to float. There was a boom in initial public offerings in 1999, with nearly 300 companies floating on the US market (many of them dotcom-related). This time around, there are only a few IPOs but they are huge. We have just seen the flotation of SpaceX (which was mainly raising money for its AI arm) and OpenAI and Anthropic will not be far behind. All of them will join the market with a valuation making them one of the 20 biggest US companies. This is not a case of buying a minnow and watching it grow.

   Of course, these companies do need to raise capital for their expansion plans (see next point). But by floating only a small part of their equity, they can enhance their valuations. There is no way that the investment community would be able to pony up $1.8 trillion to buy all of SpaceX. Insiders may not cash out immediately. But research shows the big benefits in recent IPOs have generally gone to insiders

   the fastest growth occurs before they go public.

3. No longer cash-generating. No one element deflated the dotcom bubble but an article in Barron’s in March 2000 pointed out how quickly many of the dotcom start-ups were burning through their cash. One element in the remarkable resilience of equity markets since the early 2000s has been a trend dubbed “de-equitisation”by Rob Buckland of Citigroup; companies spending heavily to buy back their shares.

   As the FT points out, the huge amounts spent by tech groups on AI infrastructure means this is no longer happening; new issuance outpaced buybacks in 2025. Perhaps all this investment will prove to be profitable, but past booms (railways, canals, fibreoptic cable) suggest the rush to take advantage of new trends leads to overcapacity. More than 1,500 US companies were set up to make cars in the early 20th century, but this had already consolidated to 44 by 1927.

   Investors have three big pools of liquid assets to buy; equities, bonds and cash. Since governments keep issuing lots of bonds, they can’t take a lot out of that asset class. So they may reduce their cash piles to buy these AI stocks. But presumably there is some level of cash they need to hold on to and thus some level of equity supply that is too much.

4. Valuations. SpaceX floated on a valuation of 92 times revenues (not profits, revenues) and after its post-IPO bump, is trading at 100 times revenues. Back in 2000, I used to have arguments with readers who said things like “Tech groups like Logica and Sage are good companies”. They may have been but they were trading on 100 times earnings. The SpaceX is taking Elon Musk’s promises on trust. As many people have pointed out, Musk has failed to deliver on promises in the past, notably on Twitter/X’s ad revenue.

   More broadly, the US market is trading on 41 times cyclically-adjusted earnings, according to Robert Shiller’s website. This is a smidgen under the 44 times ratio stocks reached at the end of 1999, just as the markets peaked. And back then, there was more reason to be optimistic; the economy had been growing very strongly and the US was not involved in a foreign war, nor did it have an erratic tariff policy. The higher the valuation, the more that future growth is already priced in, and the further share prices can fall if growth disappoints.

No bell rings at the top of the market. The institutional investment community is full of very smart people who no doubt fell they can justify current valuations because they feel AI will genuinely transform the economy. And perhaps it will. The problem is that the internet transformed the economy as well; it still turned out, in stock market terms, to have been a bubble.

SpaceX is set to be, by far, the biggest IPO in history, raising $75bn (and potentially as much as $86bn), in its float. That will give the company a market value of almost $1.8 trillion, or 92 times the company’s revenues. Yes, that is revenues, not profits. A valuation of 92 times profits would be regarded as pretty racy. But you can’t value SpaceX on the basis of profits as it is loss-making.

Now it is true that many a successful company has joined the stock market as a lossmaker; investors buy shares on the basis of the potential of future profits. SpaceX has been successful in lowering the cost of getting rockets into space by around 85%. It also has a very profitable subsidiary in the form of Starlink, which uses satellites to provide broadband and internet services.

But SpaceX is planning to use a lot of the launch of the proceeds to invest in its AI business, which it merged with SpaceX in February. One idea is to create data centres in space, which will be cheaper (and may be less controversial) than having them use energy on Earth. SpaceX says that AI comprises over 90% of its “total addressable market”. As the Economist points out, this depends on the unproven technology of its new, huge, Starship rocket, which needs to be both reliable and reusable, and on SpaceX’s data centres outcompeting its terrestrial rivals, which attracted an estimated $500bn of investment in 2025.

So the shares are a huge gamble. They seem like the sort of company that only professional investors have the potential to analyse. So why are retail investors being invited to participate? One reason is the immense enthusiasm shown by retail investors for stocks in recent years, on platforms such as Robinhood. JP Morgan Chase estimated that retail investment flows rose 50% between 2023 and 2025. Elon Musk, the head of SpaceX, has many US fans, particularly those who backed Tesla, his electric car company. But the offering is not confined to America. In the UK, one of my ISA providers has written to me offering the chance to back the SpaceX float.

It may be that retail investors will be less critical than the professionals. Pension funds have expressed concern at the shareholding structure of SpaceX, thanks to a dual class share structure which gives Mr Musk 79% of the votes while owning only 42% of the equity. He cannot be fired. This is a problem given his erratic personality and the fact that he also runs Tesla and is so active on X that he tweets 68 times a day.

Given how much money SpaceX is trying to raise, tapping retail investors seems like a good tactic from the company’s point of view. That doesn’t mean investors should take the bait; no need to be a sucker. Let us think about the valuation. If the float is successful, SpaceX will become the eighth most valuable company in the world (on the latest figures). This is not a case of buying a minnow and watching it grow.

The fact that SpaceX is only offering 4% of its equity is one reason it can have such a high valuation. Had it tried to sell a much bigger proportion (say 20%), investors might have balked. Scarcity value will help. But investors (including Musk himself) do have the right to sell part of their stakes later on, subject to certain conditions. And investors may have to pony up $60bn each for the anticipated floats of Anthropic and OpenAI this year (each of which have rather better AI credentials than SpaceX.

Of course, a lot of us may end up owning a bit of SpaceX by default. Index providers are rushing to let the big three IPOs into their benchmarks earlier than usual; so if you own a US or World index fund, you will have an exposure. At least, providers have moved on from the bad old days of the 1990s when a stock with a small free float could have a full weighting. Nowadays, only the free float counts. That means, according to Victor Haghani and James White of Elm Partners Management, the new IPOs will have a combined weight of just 0.4% in the US index.

Of course, given the media interest, SpaceX’s share price may pop up in the first few days of trading. But recent IPOs have underperformed the market by 3-5% a year and the omens for Space X, given its sky-high valuation and ability of insiders to sell, is a good candidate to follow suit.

If you are not convinced by the investment case, think about the moral one. Do you really want to send money Musk’s way? US readers will have their own views, given his disastrous record cutting public spending and his backing for Donald Trump. Mr Musk is an enemy of UK liberal democracy having declared that the country is heading for civil war and having promoted the views of far-right activists such as Stephen Yaxley-Lennon (aka Tommy Robinson) and Rupert Lowe. He is stirring up trouble again after a recent murder case; the Grok AI (which an investment in Space X would help to fund) falsely identified a female police officer as being involved, even though she left the force before the incident. The poor woman has now gone into hiding. That’s the kind of callous sloppiness you would expect from Musk.

So don’t be a sucker and don’t give money to SpaceX to help Musk spread division and hate and send innocent people into hiding.

The UK is facing such a moment now with a by-election in Makerfield, part of the greater Manchester area. The sitting Labour MP has deliberately stepped down to create a vacancy so that Andy Burnham, the mayor of Manchester, can become an MP and potentially replace Sir Keir Starmer as prime minister.

There isn’t really a historical parallel for this situation. In 1963, the Conservatives chose Alec Douglas-Home to be their leader and prime minister. Since Douglas-Home had previously been in the House of Lords, a sitting MP resigned to create a vacancy. But this by-election had the full support of Harold MacMillan, the sitting PM, who was stepping down because of his health (ironically, he would live for another 23 years).

This by-election is much odder, as Duncan Robinson points out in his Bagehot column for the Economist. One plausible reason to vote Labour is to get rid of the existing Labour prime minister. Despite that, much of the Labour top brass will be campaigning to get Burnham elected.

So why didn’t Starmer block Burnham from standing, as he did at a previous by-election (which Labour lost)? There would have been a plausible excuse; being mayor of Manchester is an important job and, if Burnham quits, the far-right Reform party might take over.

But after disastrous local elections for Labour, Starmer was too weak to block Burnham again. Had he attempted to do so, his existing MPs might have revolted and backed a challenger such as Wes Streeting or Angela Rayner to challenge him for the post.

Burnham’s by-election bid is quite a gamble. While Makerfield was once a safe Labour seat, in the 2024 general election, the Labour lead over Reform was just 13 percentage points. In the recent local elections, every single Makerfield ward voted Reform. Back in 2016, 66% of the constituency voted Leave, making it natural Farage territory.

So here is where the Sliding Doors come in. Should Burnham lose, the effect on Labour morale will be devastating. Burnham will no longer seem like a potential party saviour and will be unable to become PM. Starmer will still probably get replaced by Streeting or Rayner, neither of whom have Burnham’s broader appeal. Reform will look all the more like the plausible next government. As my last post argued, Britain will have taken one more step towards the darkness.

It is worth repeating one section from that piece here. (All you need to know about Reform can be summed up in one sentence. The Reform party conference gave a standing ovation to a woman who, in the midst of riots, called for people to “set fire to all the fucking (asylum) hotels full of the bastards for all I care”.)

But what if Burnham wins? Then he will have shown that he is a candidate who can beat Reform in a Leave-voting constituency. He will probably be crowned leader within months and will give the party both a boost to morale and a lift in the opinion polls. A Reform government will look less of a probability.

(By the way, I am not that great a fan of the Labour government. But remember the 2024 Presidential election. Kamala Harris wasn’t a great candidate but she was infinitely better than Trump. A Harris Presidential term would have seen no widespread tariffs, no Iran war, and no gobsmacking examples of corruption like the new legal slush fund. Labour, with all its faults, is not Reform.)

So disaster if Burnham finds the doors/polls closed, hope if they are open. It also seems as if the gamble, great as it is, might be the only way of turning the tide. Burnham deserves credit for attempting it. The current betting has him as favourite to win (for what the betting’s worth; there have been no constituency opinion polls.)

One final note. In my long FT/Economist career, one of the most likeable and thoughtful pundits I spoke to was Sushil Wadhwani, who was once on the monetary policy committee that decides UK interest rates. Sushil is not a man`for hyperbole. But in his recent Guardian piece, he wrote of the “climate of fear” that Reform would engender, a likely decline in foreign direct investment and a rise in inflation because of labour shortages. He added that

I see the impact of Reform UK policies on Britain’s growth as, in some ways, akin to the large, adverse impact on the Ugandan economy of the forced expulsions by Idi Amin in the early 1970s.

This would be darkness indeed. There is a lot riding on Burnham’s gamble.

But now opinion polls have consistently shown Reform with the largest share of the vote, a position that has been confirmed in May 7’s council election results in England. Reform is led by the blokish Nigel Farage, who previously led the UKIP and Brexit parties, and who poses as a man of the people, despite his earlier life as a commodities broker and education at Dulwich College (ironically, the same school as P G Wodehouse).

To his credit, Farage has at least kept his distance from Stephen Yaxley-Lennon (aka Tommy Robinson) but his party still has plans for mass deportations that could reach 2m people, has threatened to put detention centres in areas that vote Green, and can reasonably be described as “far right”. All you need to know about Reform can be summed up in one sentence. The Reform party conference gave a standing ovation to a woman who, in the midst of riots, called for people to “set fire to all the fucking (asylum) hotels full of the bastards for all I care”.

I am not suggesting that all the people who vote for Reform feel the same way. Undoubtedly, the main impetus behind the party is dislike of the Conservatives after 14 years in power, disappointment that Labour has been no better in its first two years, and thus a willingness to “give someone else a chance”. Most people don’t pay much attention to the news

Hope is not lost. The polling guru Sir John Curtice has calculated that, in the results to date, Reform has accumulated 26% of the vote. In other words, three-quarters of the electorate do not support it. In Britain’s first-past-the-post system, however, the party that has the plurality reaps the rewards; the other parties are polling between 15-20%. The anti-Reform is split with Labour losing votes to the Greens on its left and the Conservatives losing votes to the LibDems, as well as to Reform.

Local elections tend to be a vehicle for protest votes and for low turnout; many Labour and Conservative voters may simply have stayed at home. In a general election, they may return to the fold, if they perceive their party as having a chance of government. Indeed, Farage’s previous vehicles took the biggest share of the vote in both the 2014 and the 2019 European elections, without translating that gain into Westminster seats.

The problem, however, is perception. Reform’s advance has come at a time when both the Conservatives and Labour are doing badly. It now looks like a plausible winner of the next election, and is being treated as such by the media. So it will not be penalised by the perception that support for it in 2029 would be a wasted vote.

England has a five party system. The left is split between Labour and the Greens, which seem to have transformed themselves into a rehash of the Corbynites (even though Corbyn has his own party). It is a repeat of the 1980s, when the left was split between Labour and the Social Democrats; the difference is that the splitters are on the left rather than on the right of Labour. That divide helped Margaret Thatcher to her second and third-term victories.

This time round, the right is also split and, at the moment, Reform has a bigger chunk of the right-wing vote. So it could be that, in 2029, Greens will return to Labour to keep Reform out. Or Reform’s many scandals may see its vote drift downwards by 2029; there is a lot of difference, in terms of getting seats, between a 25% vote share and 20%.

The longer-term shift

Perhaps this political realignment should not be a surprise. Arguably, it would have happened before but the first-past-the-post system enabled the two main parties to prosper despite the declines in their underlying support; just like Wile E Coyote kept running for a while after he passed the cliff edge.

Just over a decade ago, I wrote a book “The Last Vote” about the threats to western democracy. These were triggered by the decline in voter turnout and party membership and increasing hostility towards politicians in general. The bargain of democracy is that citizens offer votes in return for higher living standards. When this does not happen (or does not happen fast enough), voters get disillusioned. They may then turn to an outsider who seems to promise a fresh approach. Donald Trump made this prediction come true in 2016.

In the UK, the big parties build their appeal on a coalition of interests. Labour’s long-term support was built on the trade union movement, which was in turn built on the manufacturing industry. UK trade union membership has halved since its 1979 peak and is now just 22% of the workforce. The party’s support was higher among those with a degree in the 2024 election than among those with just GCSEs and slightly higher among those in the highest social classes than among the lowest. They also did a lot better among ethnic minority, than white, voters, and among the young than the old. So this is a coalition rather like the Democrats in the US in which social and rights issues play a bigger part than economics.

The Conservative vote is also a coalition, made up of a combination of pro-business groups, libertarian small staters and nationalists. (This is similar to the Republicans in the US but religion plays a much smaller role in the UK.)

Both coalitions are much harder to hold together than in the 1950s and 1960s when (broadly speaking) the middle classes voted Conservative and the working classes voted Labour. Labour has lost the young to the Greens, many Asian voters to the Greens (and independent parties) and working class voters to Reform. Similarly the Conservatives have lost the nationalist vote in England to Reform. These voters may drift back to the two main parties for tactical reasons but they no longer have the same emotional connection with them (party membership numbers are much lower).

Another problem for Labour is it used to have strongholds in Wales and Scotland which offset the Conservative domination of the home counties and the shires. But Scotland has long fallen into the grip of the Scottish National Party while Plaid Cymru (the Welsh nationalists) are now the largest party in Wales, with Reform second and Labour only third. Reform were tied second in Scotland with Labour; we have come a long way from the days when the very English Farage was seen as toxic north of the border.

Three of the four nations of the UK are now run by parties who would like to leave it. Things in Scotland and Wales are not as bad as in Northern Ireland (where Michelle O’Neill of Sinn Fein is first minister). In that province, mainstream UK parties are irrelevant, a pretty clear indication that it is, in effect if not in law, a separate country.

Without Wales and Scotland, Labour will struggle to form a government, especially as it will lose some urban seats to the Greens. And the Conservatives will also struggle, without the home counties (which have drifted to the LibDems) and the small towns (which have switched to Reform).

Back in the 1970s, there was much talk of Britain becoming ungovernable as politicians struggled to asset their authority in the face of widespread strikes and terrorist incidents. Things are not quite that bad. But the most plausible election results in 2029 are now either a Reform-led government or a Parliament so hung that the next government will struggle to get anything done. The former will be met by mass protests; the latter may just increase voter cynicism.

Britain is not yet in the darkness. But it has taken a step closer to it.

The answer to this conundrum can be found in some remarkable figures outlined by Andrew Lapthorne of Société Générale, the French bank. He calculated that consensus profit forecasts have been upgraded by $400bn this year, or $600bn if one looks 12 months forward. The latter number equates to 12%, the largest ever forecast upgrade when an economy is not emerging from recovery.

Nearly half this profit upgrade comes from semiconductors, a sector which has seen a very sharp rise in share prices in recent weeks (47% in 18 trading days). Virtually all the profits upgrade comes from semiconductors, technology hardware and energy (those oil prices again).

In short, consumers may be suffering but some companies are making out like bandits. It looks as if corporate profits, already close to a recent high as a proportion of GDP, are going to take an even bigger chunk of the economy. That is “populism” for you and explains why so many billionaires were hanging around at Trump’s second inauguration.

The explanation for the surge in tech profits is the AI boom. Everybody wants to be part of it which means they need the data centres and the computers and thus the microprocessors that will make it work. Atakan Bakiskan of Berenberg, a private German bank, estimates that technology companies will invest nearly $700bn this year, or 2% of GDP. In some cases, notably OpenAI and Anthropic, companies may have capital commitments that are higher than their annual revenues, making them very dependent on good news about the effectiveness of their models. According to the Wall Street Journal, OpenAI missed its revenue and new user targets and it is now embroiled in a legal battle with Elon Musk.

Whatever the tech worries, they are nothing compared with the state of consumer confidence. The latest survey from the University of Michigan revealed the worst consumer sentiment on record. Surveys don’t always translate into spending behaviour but Americans are definitely unhappy; approval of President Trump’s economic record is down at 30%.

This contrast between consumer sentiment, economic growth and the stock market may explain why the Bank of England’s warning about overvalued share prices, while understandable, may take time to be proved right. Equity prices look very high, compared with profits. This chart, from Robert Shiller at Yale, shows the cyclically adjusted price-earnings ratio; as you can see, it has rarely been higher. This ratio averages profits over 10 years.

Now the CAPE is far from a perfect short-term indicator. Yes, the two peaks - 1929 and 2000 - coincide with the end of the 1920s boom and the bursting of the dotcom bubble. But, since the mid-1990s, the ratio has been consistently above its level in all periods bar the late 1920s.

One obvious explanation for this is the red line which shows the level of long-term interest rates (bond yields, in other words). In theory, the value of a share is equal to all the future cashflows. But since money paid in the future is worth less than money you have in your pocket today, those future flows must be discounted at some rate to arrive at a present value. The long-term bond yield is a key element in that discount rate.

So when long-term rates fell after 1981, the discount rate declined and those future cashflows were worth more today. Hence the steady rise in share valuations in the 1980s and 1990s.

While bond yields were so low, it is understandable that equity markets rebounded from shocks, even a shock as deep as the financial crisis of 2007-09. For all the talk of alternative assets, institutions tend to focus on equities or bonds as the main proportion of their portfolios (property can work for long-term investors but it is illiquid). And thus while bond yields were so low, equities seemed like the only way of generating the kind of returns investors need.

(Private equity is just unquoted equity (and illiquid) while hedge funds mostly generate their returns from equities and bonds).

However, at the end of the graph, you can see that the red line (the bond yield) has been creeping upwards in recent years. In other words, the discount rate applied to future profits has been increasing. That ought to bring down share valuations. But, as we have seen, profit estimates have been surging.

So for the Bank of England to be proved right, either those profits forecasts have to be revised down sharply or bond yields have to rise significantly further. Either could happen, of course. The returns from AI could turn out not to justify the massive investment. Or the Fed, under new chairman Kevin Warsh, could let inflation rip, pushing up the bond yields a lot more. Those are probably the conditions we need for a crash.

And we also have an uncertainty principle, in the sense that there is so little clarity about the negotiations. (For a while, it wasn’t even clear whether Iran would turn up for the next round of talks.) President Trump keeps declaring that things have been “agreed” when they clearly haven’t. On Friday, he told Reuters that Iran would cooperate in removing the “nuclear dust” created when the US bombed the country’s nuclear installations last year. “We’re going to get it together. We’re ‌going to go in with Iran, at a nice leisurely pace, and go down and start excavating with big machinery” he said. The suggestion was immediately rebutted by the Iranians.

This may be a case of what the younger generation calls “manifesting” - if you declare something to be true, it will eventually become the case. Trump clearly has a number of tics in his negotiating approach; one is to insist that a deal will occur very soon (within 2-3 weeks), another is to declare that the counterparty will deliver everything he wants, and a third is to make bloodcurdling threats if they fail to comply.

It is worth emphasising that none of these habits bear any relation to the truth. He declares that negotiations will end quickly, without knowing that they will; says the other side will give in when it is likely they won’t; and makes threats which he often retreats from carrying out.

This unreliability is at odds with the willingness of investors to react so swiftly to Trump’s pronouncements by rallying when he talks of progress. In the FT this week, Tej Parikh argued that markets were influenced by the AI boom and by the conviction that Trump will retreat from his more extreme positions if they look like causing economic damage.

By and large, investors also seem to believe two other things; that Trump has some kind of strategy (eg his wild statements are just negotiating tactics) and that the economic costs of the Iran war will be nothing like as damaging as the 1973 oil crisis or Russia’s invasion of Ukraine.

In my last post, I argued that the rambling style of Trump was an indication of his mental confusion and lack of a strategy. This is a man who never reads briefings and who has surrounded himself with sycophants. In today’s Times, Matthew Parris argues that Trump is too deranged to continue in office. He is surely right that he would be fired were he, say, a chief executive of a multinational. But there is absolutely no prospect of his removal because the Republican party cannot stand up to him (and those voters who support him).

It ought to make investors worry that a man so wilful and incoherent is in charge of the world’s largest military and biggest (on some measures) economy. They also need to consider the point that Trump’s actions make it harder to achieve a lasting peace. He has attacked Iran twice in the middle of negotiations and that has made the Iranians understandably suspicious - why should they take him at his word? In the FT today, Gideon Rachman says he thinks escalation is more likely than peace.

On the economic impact of the war, market insouciance is best expressed by Deutsche Bank who point out, in a research note, that the oil price is lower, in both nominal and real terms, that it was in 2022; that the oil intensity of the economy is much lower than it was in the 1970s; and that inflation is starting from a low level and thus central banks will not need to tighten monetary policy aggressively.

But Paul Krugman argues convincingly in his latest Substack that analysts are thinking about the Iran impact solely in terms of the oil price when they should be worrying about oil supply. People cannot easily, in the short term, switch from using oil (and gas) to using alternatives. There is not going to be a sudden switch to electric cars, for example. The easiest way to deal with an oil shortfall is to consume and produce less; ie a recession.

In March, gamblers on Polymarket thought there was a 35% chance of a US recession in 2026, but now the odds have dropped to 25%. In that sense, uncertainty has declined.

However, the fundamental nature of Trump is that he creates uncertainty wherever he goes. That is what I argued in my tariff book last year and it remains the case. He has followed up a terrible policy mistake in 2025 with another one in 2026. There are two-and-a-half years of his Presidency to go. The uncertainty principle will be with us till 2029.

As expected, the policy was a failure. In a previous post, I listed the relevant statistics; manufacturing jobs fell in 2025; the US economy grew less strongly in 2025 than it did in 2024; the trade deficit was unchanged compared with 2024; and surveys show that the tariff burden fell most heavily on Americans.

It might have been worse. The initial market reaction to the tariff package was so severe that Trump suspended them, and then watered them down. As a result, other countries (with the exception of China) did not retaliate. Overall, as Richard Baldwin points out in this excellent piece, the tariff burden was much lower than was initially feared.

But as has become clear in 2026, the economic consequences of Mr Trump are not confined to his trade policy. His attacks on the independence of the Federal Reserve make the markets more nervous about the central bank’s willingness to tackle inflation. And his recklessness in foreign policy, first by threatening the annexation of Greenland and second, by the assault on Iran, is destroying the informal American empire which has been a key part of the democratic world since 1945.

International trade, and economic relations in general, are built on trust. Mr Trump is not the kind of person whose word you can rely on; he is a deal-breaker, not a deal-maker. And, as the old saying goes, trust arrives on foot and leaves on horseback.

Katie Martin wrote an excellent Long View column in the FT this weekend on why the markets were so inclined to believe Trump’s statements. The good news is always “two to three weeks away” which is the President’s equivalent of the old excuse, “the cheque is in the post”. A lot may be explained by the experience of this generation of investors. They have learned that geopolitical risks tend only to have a short-term impact on the market; if profits are rising, and the central banks are not tightening severely, stock markets tend to rise.

But I also think there is a kind of naive belief in Trump and the idea that his main agenda is to give the corporate sector (and the rich) what they want. Take this argument from Torsten Slok of Apollo Investment Management, which is the single most astonishing piece of investment analysis I have read in 43 years of covering the markets.

“Markets are overreacting to what will likely be a 4- to 6-week period of volatility, which will ultimately result in 50 years of stability in oil markets, supply chains and geopolitics.”

50 years of stability? We would be lucky to experience 50 days. Ever since Trump returned to power, some have made the argument that he is playing a highly sophisticated game of 4-D chess or that his inconsistency is designed to catch his enemies off guard. For example, he says his allies need to clear the straits of Hormuz, then says he is not bothered whether they open or not, then makes clearing the straits his main condition for a ceasefire. I believe in Occam’s razor; the simplest answer is most likely. He makes confusing statements because he is confused.

Indeed things are worse than that. Critics like me are often accused of “Trump derangement syndrome” - that we so dislike the President we cannot grasp his genius. But instead we have reached the stage of “Deranged Trump syndrome” - the rambling non sequiturs that have long been observed in his speeches are taking a much nastier turn, and are influencing his actions.

Take this message which was posted on Truth Social during the Easter weekend. Note the reference to God during a religious festival when actual religious leaders were preaching peace

And then this message appeared on Easter Sunday (!) complete with expletive and seemingly mocking reference to the Islamic term for God.

Imagine if such a message had been written by Joe Biden. (You have to imagine it because there is no way Biden would have sent it). Fox News would have had wall-to-wall coverage of the mental deterioration of the President. The markets would have plunged for fear the President would have to be removed from office.

But Trump will just carry on. With the very occasional exception of a Supreme Court ruling there is no check on Trump’s power. He has more than 1000 days left in office!

Some may say that those messages were just bluster designed to get Iran to negotiate. It hasn’t worked so far and actions speak rather louder than words; Israel has recently bombed Iran’s largest petrochemical complex. A proposed peace plan, that has awakened some hopes in the markets, seems to have come from Pakistan rather than from either warring party.

The real problem is that the world’s greatest military superpower, and centrepiece of the global economy, is run by a man whose health and mental faculties are deteriorating, and whose anger is making him lash out in all directions, at allies as well as enemies. Deranged Trump syndrome won’t be good for anyone. As with the tariffs, the Iran war is pushing inflation higher and making growth slower around the world. It is not just costing lives. It is another enormous economic policy mistake.

That law, named after Sir Thomas Gresham, an adviser to Queen Elizabeth I, was that “bad money drives out good”. In the medieval world, the value of currency was set with regard to its precious metal content. But monarchs would often cheat, using coins where the gold and silver were adulterated with other metals; traders would also cheat, clipping or shaving coins, in order to accumulate more bullion. Rational consumers would try to get rid of these adulterated coins and keep the pure coins for themselves. Over time, this would mean that the pure coins would disappear from circulation; the bad money would have driven out the good.

We can all recall that, in the early days of email, accounts were overwhelmed with “spam” - messages promoting fake products or pornography. The result was the need to install spam filters. Google was once a genuinely useful search engine, not overwhelmed by ads. Internet banking may be spectacularly useful (when I first started work, you had to queue at bank branches on Friday to cash a cheque for the weekend) but the potential for fraud means customers now need to go through a series of security checks before the bank will authorise a payment.

Many people now use their phones for banking but as this Times journalist just found, this can be a nightmare if your phone is stolen. You need to report the theft, but you have no phone with which to report it. If you are unlucky (probably because the thief saw you using your pin) they could drain your bank accounts.

More generally, the greater connectivity of the internet also makes it more likely that we will connect with fraudsters. They may entice us to click on links via phishing approaches which risk our own, or our employers’, cybersecurity. They can target lonely people via romance frauds, or send them down a rabbit hole of extremism (far right, Islamism, incel culture etc).

Social media apps use algorithms which are intended to keep users as engaged as possible, for as long as possible. This can quickly lead to users being shown more extreme content (just as newspapers put more sensationalist stories on their front page). But this is a lot more malign than the “aliens mowed my lawn” stuff that supermarket tabloids used to print.

Take this example from the Centre for Countering Digital Hate

.In 2022, our “Deadly by Design” report showed that, within 39 seconds of a teenage girl creating an account, TikTok recommended videos about body image and mental health. Within 2.6 minutes, the platform served suicide content. And within 8 minutes, TikTok was pushing content related to eating disorders. .

There is some good news. John Burn-Murdoch in the FT notes that AI systems direct users to more centrist, consensus sources than occurs with other social media. Alas,AI has already shown it will be put to malign use in many other ways - generating fake videos on subjects such as the Iran war or UK cities, fake celebrity endorsements for products, or fake revealing images of women and children.

Now, by this stage, you may feel that my Gresham’s law wording is overstated. Bad people have not “driven out” the good. All these technologies are still being put to good uses; it is just that everyone has had to take precautions. By the same token cars are still massively useful even though they cause deaths and injuries; we have learned to impose speed limits and require seat belts.

But I think technology is having a Gresham’s law impact on politics. Across the western world, extreme politicians are using social media to put forward simplistic and toxic messages that are quickly amplified. Ironically, the impact of this shift has been enhanced by a reform that seemed more “democratic”; taking the choice of party leaders away from “elites” and putting them in the hands of party members via primaries and direct votes. Sixty years ago, Republican leaders would have shied away from nominating a character like Donald Trump; Labour MPs would never have given the leadership to a figure like Jeremy Corbyn.

As politics becomes more toxic, only the most committed people become party members and their votes go to the most ideological candidates. Social media also allows these committed members to contact each other more often, and this reinforces their enthusiasm. Few people want to join a party dominated by extremists and so the problem continues; bad leaders drive out good.

Bad people then endorse bad technologies. Regular readers will know of my contempt for cryptocurrencies as vehicles for crime and speculation. But they are endorsed by Trump and Farage because those behind the cryptocurrencies are wealthy and can make big donations. And of course holders of crypto are vulnerable to having their money stolen, and lack the protections given to bank depositors.

My law may be exaggerated but it contains much truth. New technologies spread more quickly than they used to (because, of course, new technology is around to transmit them). And that means the bad uses of those technologies spread quickly as well. By the time you have heard of a new product, the criminals and extremists will have worked out a way to exploit it.

Few independent observers can be surprised that European allies do not want to get involved in a war that was launched without any consultation with them (and a conflict they would have advised against). But the specifics of the Iranian war aren’t the real problem; it has been the tone of the second Trump administration from the start.

It all began with vice-president JD Vance’s speech to the Munich security conference in February 2025 where he denounced Europe’s stance on immigration, said the continent was retreating from its “fundamental values” and accused EU “commissars” of suppressing free speech. (This from an administration which rejoiced in the cancellation of critics such as Stephen Colbert and Jimmy Kimmel, and which would go on to threaten broadcasters’ licences over their Iran war coverage.)

In April, that was followed by Trump’s unilateral imposition of tariffs on its allies which breached past commitments of US governments to respect global trading rules. All the while, Trump was retreating from past US support for Ukraine, bullying Volodymyr Zelenskyy when he was in the Oval Office and thus undermining European efforts to support a democracy which had been invaded by a dictatorship. Furthermore, the US administration has repeatedly expressed its support for far-right European parties that form part of the opposition to current governments.

The lump of feces on this particular cake was Trump’s threats to annex Greenland, the territory of a NATO ally. Eventually he retreated from the use of force but not before European countries were forced to reassess their attitudes towards US power. It now seems, incredibly, that Denmark was prepared to fight off a US invasion.

After all this, then, the US suddenly expects Europe to lend a hand when its latest military adventure proved more complicated than expected. Suddenly allies are useful, it seems. But as Clint Eastwood said in “The Outlaw Josey Wales”, his 1976 western, “Don’t piss down my back and tell me it’s raining”. The US has exhausted its credit account.

This decline in US soft power was one of the themes of my book, “The Economic Consequences of Mr Trump”. As I wrote there “The way in which the US exercised power after 1945 created an informal empire, in which other democratic countries were happy to accept US hegemony”. The US helped European countries recover from the war with Marshall aid. It took the lead in lowering barriers to trade. In both cases, it recognised that prosperous European citizens would be good customers for American goods.

If one thinks back to the post-war period one can see how European consumers grew to like American rock music, Hollywood movies, fashion (jeans, trainers), fast food restaurants, US consumer brands (Coca-Cola), and all the rest. While Europe was not always enthusiastic about US military adventures (eg Vietnam), it was rarely outright hostile. The same went for Japan.

Nowadays, there is a risk that US brands can prove toxic. Canadians have reduced their travel to the US given Trump’s threats to turn it into the 51st state; a significant shift since Canadians were the biggest source of US tourists. They also cut their purchases of US liquor and orange juice. And Elon Musk’s support for Trump (and toxic Twitter feed) helped push Tesla’s European sales down 26% in 2025.

The US, of course, is a big economy with powerful armed forces and can go it alone. But it is facing a long-term challenge for global hegemony from China. A recent poll found that more voters in Canada, Germany, France and the UK thought that the US was now an unreliable ally than a reliable one. That gives the Chinese a lot of scope to undermine the Western alliance. Trump has scored a spectacular own goal.

The steam engine is acknowledged as a general purpose technology (GPT) - the kind of development that can transform an entire economy. Later GPTs have included the internal combustion engine and the computer. The hope today is that artificial intelligence is yet another GPT; an OECD report concluded that “generative AI appears to exhibit the defining characteristics of GPTs: i) pervasiveness, ii) continuous improvement over time and iii) innovation spawning”.

We must certainly hope so because, like the steam engine, AI is a very hungry user of energy. An IEA report last year found that the energy use of data centres had risen at an annual rate of 12% since 2017, three times the speed of general electricity usage. In 2024, data centres accounted for 1.5% of global electricity usage, but their requirements are expected to more than double by 2030 and nearly triple by 2035.

This energy use is highly concentrated in the US, where AI is expected to account for half of all electricity demand growth by 2030. The IEA concluded that “by the end of the decade, the (US) is set to consume more electricity for data centres than for the production of aluminium, steel, cement, chemicals and all other energy-intensive goods combined.”

This surge in energy use is occurring at a time when, in theory, developed nations are committed to reducing their carbon emissions in order to combat global warming. A new report suggests that global warming is speeding up; temperatures are rising at 0.35 degrees a decade, up from 0.2 degrees before 2015. In February, the planet was around 1.49 degrees warmer than the historic average, very close to the 1.5 degree limit that countries are nominally aiming for.

Now it is not all bad news. The IEA reckons that around half of the increase in data centre energy demand is met by renewables. Still that means half of the increase won’t be met by renewables and of course, if we are to reduce overall energy consumption, cuts in other sectors will need to be even greater to compensate for the impact of data centres.

So AI had better be the game-changer promised by enthusiasts. The IEA report suggested that AI could be used in the energy sector itself to optimise exploration and production, identify faults and increase the capacity of existing transmission lines. In theory, this could offset all the increased demand of data centres.

And, at least, AI is not Bitcoin. The process of generating new Bitcoins is very energy-intensive, and is around 0.5% of global energy consumption - all to produce a currency that is mainly used for speculation and tax avoidance and has not come into general usage 17 years after its creation - an eternity in modern technology terms.

Nevertheless, the rapid adoption of AI is something that should concern us all. The debate has got rather sidetracked into fears of Terminator-style machines deciding to do away with humanity. But there are a lot more practical dangers. If you read Supremacy, Parmy Olson’s book, you will be struck how the initial noble aims of pioneers like Sam Altman and Demis Hassabis were gradually diluted as they embraced the existing tech giants. They needed the giants because they had the computing and financial firepower to make the required investment.

And as the recent row between Anthropic and the US government has demonstrated, AI may be used to create weapons or conduct mass surveillance, rather than to make benign improvements to our productive capacity. Those early steam engines were the start of a technological change that eventually created the artillery and machine guns that slaughtered so many on the Western Front between 1914 and 1918. This time round, the process will be a lot quicker. Humans don’t need machines to destroy us; we can do a pretty good job ourselves.

The ease with which Donald Trump unseated Nicolas Maduro, the Venezuelan leader, now looks like a version of the gambler’s problem. Success looked easy and painless. And, as I suggested in my last post, the Supreme Court’s tariff decision limited Trump’s room for manoeuvre in economic policy; he can now tariff all countries but not pick on an individual country. Military action allows him to look decisive. Foreign affairs often obsess second-term Presidents and Trump is no exception. But the risk is a version of the Peter principle; Presidents keep intervening until one such adventure proves a disaster.

Mr Trump has also reached the “jello shots” stage of his term in his office, where like a group of men on a stag do, overconfidence leads to risky behaviour. In Iran, he probably assumed that a demonstration of US and Israeli air power would cow the Iranians in the same way it did the Venezuelans. It doesn’t seem as if the administration anticipated that the Iranians would drag much of the rest of the region of the conflict.

Hence the sharp rise in the price of oil that has been seen in the past week, accompanied by even steeper increases in natural gas prices. It is not just that the straits of Hormuz have closed, it is that oil and gas production and refining facilities have come under attack. Even though the US is self-sufficient in energy, the impact will still be felt on US citizens at the gas (petrol) pumps. MAGA voters may cheer Trump’s demonstrations of US military power but that is on the assumption that the interventions don’t cost them any money.

The markets were relatively calm at the start of last week. Most investors have learned that geopolitical risks don’t tend to have an impact for very long. Deutsche Bank noted that a big sell-off required a very large jump in the oil price (50-100%), a recession or central bank tightening. Plenty of investors may also have assumed that the intervention would be very short-lived, and thus would fail to dent the economy significantly.

Indeed, it still could be. By not reacting a lot, the market may have ironically reassured the President that all would be well. The sell-off on March 9 reflected a change in mood because of the attacks on oil installations, the announcement of the new hardline Iranian leader and Trump’s continued talk of the need for the Iranians to surrender. Now the markets have plunged, Trump may be tempted to declare victory and stop. This could be a diluted version of TACO, Trump Always Chickens Out.

But there are two dangers. The first is that Trump did not attack Iran alone but in concert with Israel. The Israelis have no desire to see the mullahs left in place; they would like to finish this off. At the weekend, Chris Wright, Trump’s energy secretary, said there were no plans to target Iran’s energy industry; it was bombed anyway.

Instead of TACO, the right acronym for Trump might be NACHO (Netanyahu Actually Controls His Operations). Marco Rubio’s remarks a week ago, that the Israelis pushed the US into war, may be revealing (Trump disavowed the remarks but he would, wouldn’t he?).

Another danger is that Trump keeps going because he wants a prize from the war. Just as in Venezuela, this may be access to cheap oil (he has hinted as much, saying the price may be up now, but will come down a lot later). Cynics have often lamented that US Presidents are too obsessed with Middle East politics because of the oil supply. The right acronym might be TOBACCO (Trump Often Bombs Any Country Containing Oil).

If Trump does call time on the intervention, with the mullahs still in place, the war will look like a mistake. And if he carries on, in the search for glory or under the influence of Netanyahu, the economic consequences will make the war an even bigger mistake,

Clearly this will be a disappointment to a President who likes pushing other countries around. This may well tempt Trump into following the well-worn path of other second-term Presidents; repeated military intervention. The latest example is the bombing of Iran in a coordinated assault with Israel, launched on February 28.

Iran is one of seven countries in which the US has launched military action in his second term of office: Iraq, Nigeria, Somalia, Sudan, Venezuela and Yemen are the others. In every case till now, the interventions have been limited in scope and short-term. This may have led the President into the belief that such attacks are essentially costless; they demonstrate American power and thus may persuade potential foes to bow to US will rather than resist.

President Trump doesn’t like any constraints on his actions. Hence his extensive use of executive orders; he has made 243 so far in his second term, more than in his entire first term, and more than any President since Richard Nixon. He dislikes having to negotiate with Congress and resents any court that rules against him. In theory, Congress should approve any war but in practice, most Presidents ignore the legislature. If Trump wants to use the military, it seems unlikely that anyone will stop him.

But a version of the Peter principle may apply here. The principle is that employees are promoted to their level of incompetence. If you are good at your job, you will get promoted; if you are not that great, you won’t.

In this case, political leaders keep making military interventions until one proves to be disastrous. Early victories are like a “gateway drug” which encourages riskier adventures. Putin had success in Georgia and with his occupation of Crimea; he has now been bogged down in Ukraine for four years, at enormous cost in blood and treasure. Tony Blair started to believe in “liberal interventionism” after limited actions in Sierra Leone and Yugoslavia; he committed British troops to the Iraq war and his reputation has never recovered.

Maybe Iran has been too weakened by sanctions and by the setbacks to Hamas and Hezbollah, its key regional allies, and so the rule won’t apply. But if Trump does unseat the Ayatollah, that may only tempt him into an adventure somewhere else (Cuba?) just as the arrest of Maduro may have emboldened him this time. If he can’t use tariffs as an arbitrary weapon, he may just use some real weapons.

By itself, that change points to the chaos in White House decision-making.* The Supreme Court decision has been coming for months and the law was pretty clear. Trump’s team should have had a policy response in place. But announcing a policy and then changing it within 24 hours is very much the Trump brand. Pity those businesses that have to cope.

At least, Trump’s powers to pick on individual countries will be limited by the ruling; he can only impose a flat rate on every country, although he will be able to pick on individual sectors. The change means that the biggest winners (in terms of export competitiveness) will be Brazil and China, while loyal ally the UK now faces a higher tariff than before.

Plenty of uncertainty remains. We don’t know whether businesses will get refunds for the tariffs they have already paid. Mind you, the argument made by the dissenters on the Supreme Court (that the tariffs should stay in place because of the mess that would result if they were ruled illegal) was bizarre. By extension, the court could not ban any action of the President, however bizarre or autocratic, if it was problematic to reverse it.

I have argued before in this Substack series that the world is suffering from PTSD or permanent tariff stress disorder. The AI boom is disguising the full impact of the trade war but the effect is still real. Atakan Bakiskan at Berenberg recently wrote that “The crippling uncertainty surrounding trade policy and its damage to economic growth will likely outweigh any “fiscal stimulus” from potential tariff refunds to businesses and the few-percentage-point reduction in the average tariff rate.”

As regular readers will know, I wrote an instant response to Trump’s tariff policy last year. Inspired by John Maynard Keynes, it took the form of a book “The Economic Consequences of Mr Trump”. Now that the legal basis of the original set of tariffs has been destroyed, it is worth pointing out what an abject failure the policy has been.

Trump’s policies are rarely coherent but the tariffs seemed to be based on three arguments; that the trade deficit was a sign of cheating by other countries and should be eliminated; that foreigners would pay the tariffs; and that the result would be a resurgence in manufacturing jobs.

How is that going? In 2024, the US trade deficit in goods and services was $903.5bn. In 2025, the deficit was (wait for it) $901.5bn. All that effort, all that fanfare, all that disruption just to knock $2bn (or less than 0.25%) off the total. Even more striking is that the tariffs just apply to goods imports. Well, the goods deficit went up and the deficit only fell because of a bigger US surplus on services. By the way, is the US services surplus a sign that the US is cheating?

And who is paying the tariffs? A New York Federal Reserve analysis found that the vast bulk of the cost (between 86 and 94 per cent) was absorbed by American producers and consumers; a study by the Kiel Institute found that they absorbed 96% of the burden. So yes, the tariffs raised money but from Americans, not foreigners. (And by the way, even with the tariff revenue, the US budget deficit was $1.8 trillion in 2025 and is expected to be even higher at $1.9 trillion in 2026.)

And there has been no resurgence in manufacturing jobs either. The numbers indicate that 88,000 manufacturing jobs were lost in 2025. This is hardly surprising. Around 45% of what the US imports is components and raw materials for things the US makes; imposing tariffs raises the costs of domestic producers. The US nowadays specialises in high-value goods requiring high-skilled labour; it makes no sense to try to return to an era when the US produced T-shirts and plastic goods. Labour costs in the US are so high that consumers would have to pay a lot more for the goods they want.

The President is still ranting and raving about the Supreme Court decision and may well lash out in new ways. Whatever the size of the hole, he will keep digging. Businesses and consumers will have to keep paying the costs of the abject failure of his policies.

Addendum*: Since the piece was written, the tariffs have gone into effect at 10%, even though Trump had said the 15% rate would happen “immediately”. There is a traditional football chant, aimed usually at the referee but sometimes at the manager, which goes “You don’t know what you’re doing”. We can all sing that today.

The latest victim of the curse is the crypto industry. The industry spent $245m on the 2024 election campaign, contributing around a half of all corporate contributions. And they certainly were successful in getting their man into the White House; Trump has championed the industry. It certainly helped that the industry offered plenty of opportunity for personal gain; both the President and the First Lady launched their own memecoins. Those loyal MAGA supporters have been hit by the curse; the FT reports that the President’s version has fallen 94% from its peak while Melania’s has fallen 99%. But the FT reckons that the coins, along with other crypto ventures, may have generated $1bn in profits for the Trump family.

It is not just the memecoins that have taken the hit. The best known cryptocurrency, Bitcoin, has lost all its gains since Trump took office and is down 44% in the last six months. In the bout of recent nervousness over Trump’s plans for Greenland and Iran, the risk-off beneficiary has been gold, not bitcoin.

The likely explanation, as Katie Martin has shrewdly pointed out, is that the bitcoin price is closely tied to confidence in the tech sector. Since October, US software stocks have dropped 27% and now trade at a discount to the rest of the market. There is a nice irony in that tech leaders tend to be very enthusiastic about artificial intelligence but now investors think AI will cannibalise the business of software companies; it is as if they have nurtured a crocodile to adulthood and are surprised to get bitten.

If you work for a tech company and are reliant on share options to build your wealth, the sell-off in tech stocks is very bad news. Small wonder that you might be inclined to sell off some of your crypto assets or be less inclined to take some risk by loading up on bitcoin.

And that brings us back to the underlying strength (and weakness) of crypto assets. They are literally created out of thin air and (except in the case of stablecoins) have no real wealth or income behind them. On the way up, this is a strength; unlike a company they can’t issue a profit warning since there are no profits in the first place. No one can say that cryptocurrencies are overvalued according to some agreed measure, since there is no sensible valuation approach. If the price is $10,000, it could just as plausibly be $20,000.

On the way down, however, the same argument applies. If the price is $10,000, it could just as plausibly be $5,000. There is no bottom at which the fundamentals kick in. Bitcoin has been around since 2009 but unlike other technologies has not found a general use; it is most associated with criminal activity and ransom demands. The demand for crypto is largely driven by speculative buyers hoping the price will rise; when the price is clearly falling, such demand disappears.

Trump’s greed has played its part in this setback. One of the bullish arguments for crypto is that it was being held back by regulation that prevented its general use. Trump has cut back all the regulation and the transformation has not occurred. Worse still, the bullish argument of future regulatory changes is no longer available.

Of course, sceptics like me have called the top in bitcoin before and been proved wrong. Largely I think this is because the speculative impulse, fulled by low interest rates and tech enthusiasm, has been so strong. But in the end, I am with Charlie Munger, the associate of Warren Buffett, who denigrated those trading crypto by saying “It’s like somebody else trading turds and you decide I can’t be left out”. Lots of rich people have spent a lot of effort polishing the crypto turd but now they have been hit by the curse of Trump.

It seems clear that this support is mainly built on frustration with traditional politicians. They obfuscate; they talk in platitudes; they repeatedly change their minds; they never get anything done. By contrast, a strong leader can cut through the bullshit and take the necessary action.

That is the theory at least and it explains why, despite Trump’s blatant attempt to overturn the 2020 election and his threats to all and sundry, American voters re-elected him in 2024. The same appeal explains the appeal of Farage’s Reform in the UK, the National Rally in France and Alternative fűr Deutschland in Germany.

But there is a fallacy at the heart of this appeal. It is that there are clear answers to national problems that politicians could implement if only they were not so selfish or useless. At the start of my book on Trump’s economic policies, I include a quote from H L Mencken “For every complex problem there is an answer that is clear, simple and wrong”.

If there really were simple solutions to our problems, leaders will generally find them. Indeed, during Covid, Trump did play his part in ensuring the development of a vaccine (although he barely mentions it now, because of the anti-science bias of his base) and even Boris Johnson managed a quick roll-out of the vaccine in the UK. In 2008 and 2009, politicians around the world figured out a combination of fiscal stimulus and bank rescues to stop the financial crisis from turning into another Great Depression.

But as I have pointed out before, western economies are not growing as fast as they used to, and so almost every decision taken by a government involves trade-offs and, thus winners and losers. Improving public spending requires higher taxes; those who pay the taxes complain while those who use the public services are insufficiently grateful. Cut taxes without cutting spending and you are punished by the bond markets, as Liz Truss discovered. Improving infrastructure takes years and triggers complaints from those who live nearby about the noise and congestion. Allow immigration and people complain about a change in culture and overcrowding; restrict immigration and key industries, like social care and agriculture, run short of workers.

We have already seen with President Trump that easy solutions don’t work. The tariffs were supposed to rebuild US manufacturing. But the amount of money being spent on manufacturing construction has been declining since Trump came into office, along with manufacturing jobs. What about all that tariff revenue? The cost is not being borne by foreigners, as Trump claims, but by Americans; about 94% of the cost has been borne by domestic producers and consumers.

And here we come to the main element of the strongman fallacy. Yes, they can make decisions quickly; Trump has already signed more executive orders in the first year of this term than he did in the whole of his first stint (and outpaced the whole of Biden’s administration as well). But just as a strongman faces no barriers to taking good decisions, he faces no constraints on his bad decisions as well.

History is littered with dictatorial leaders who make disastrous mistakes. In 1941, Stalin ignored the signs that Germany was about to invade, leaving the Russian army unprepared; Mao Zedong pushed through the Great Leap Forward (which resulted in tens of millions dying from famine); Mussolini launched costly and deadly invasions of Abyssinia and Albania, and then picked the wrong side in the war; the Hapsburg, Hohenzollern and Romanov dynasties all cheerfully signed up for the First World War which brought their rule to an end; Putin thought he could conquer Ukraine within weeks and is still at war four years later. And, of course, strongman leaders face no barriers on their ability to imprison and kill their opponents; there is usually no way of unseating them except violent revolution or conquest by another power.

Democracies may be slow and cumbersome when it comes to taking good decisions but they also provide a check on bad ones as well. That can be done by internal party opposition (Liz Truss in the UK), by an election, by constitutional checks (usually the courts), or by street protests. For all the horror of the Minnesota killings, the coverage by the media and the shift in opinion polls, means that the Trump regime cannot go as far as their Iranian counterparts.

I am not the greatest fan of Friedrich Hayek but he had a point when he argued that no individual could ever be in possession of sufficient knowledge to allocate resources efficiently; knowledge is dispersed throughout the system and is reflected in indicators such as price changes. Hayek was mainly thinking about economics but the argument holds true much more broadly.

Worse still, it is in the nature of strongmen that they will spend much of their time worrying about the possibility of being overthrown and this will devote more energy to crushing opposition than they will in ruling wisely. What is more, the nature of their rule tends towards corruption; if all decisions are taken by one man, then economic actors will seek to influence and reward that person. Only saints could withstand that temptation and, by definition, strongmen aren’t saints. Hence we have seen the Trump regime retreat from the anticorruption regulations imposed by his predecessors and openly exchange favours in return for donations.

We just have to hope that American democracy is strong enough to constrain him after the midterms and throw out his party in 2028. But it is far from guaranteed. Strongmen are like cockroaches; once you let them in the house, it is very hard to get them out again.

In a previous post, I dubbed the problem Permanent Tariff Stress Disorder (PTSD). While Trump is in office, he will use tariffs on the slightest pretext. He did so when Brazil prosecuted its former President, Jair Bolsonaro, for trying to overthrow its government; he imposed tariffs on Canada, with the supposed justification of stopping fentanyl smuggling (which is a non-existent problem on that border).

Trump keeps plugging the line that foreign countries pay the tariffs. But the latest research shows that 94% of the incidence of tariffs falls on American producers and consumers. Since around 45% of the stuff that the US imports is raw materials and components, a tariff pushes up the cost of domestic producers. Far from bringing manufacturing jobs back to the US, employment in the sector is shrinking.

None of that seems to matter. In Trump’s mind, tariffs are a hammer, and he will keep looking for nails. Never mind that the UK did a trade deal with the US last year; it will face tariffs along with EU nations. Even Nigel Farage, a shameless Trump toady, has issued a mild protest. A deal with Trump is always temporary; his word cannot be counted on. Small wonder that the European Parliament is stalling ratification of the US-EU trade deal.

The tendency to “sane wash” Trump’s policies is very strong. On Thursday, I gave a talk at the Oxford and Cambridge Club in London on my book “The Economic Consequences of Mr Trump”, which detailed the illogicality of his programme. One questioner argued that Mr Trump’s target was really China, and the tariffs on other nations were just a “smokescreen” for this purpose. But it isn’t clear why this smokescreen might be needed, or why imposing tariffs on countries that might pose a counterweight to China (such as India) will help.

One difficulty is that people don’t appreciate how globalisation has affected our world. As I write this, I sit on a wooden chair (timber imported from abroad), sipping a cup of tea (imported from India), typing on a Samsung laptop (a South Korean company, but one with 2,500 global suppliers) with occasional interruptions from messages on my iPhone (a US company with 187 suppliers across 28 countries).

Richard Baldwin, a trade expert, compares the global trading system to an omelette that cannot be unscrambled. These supply chains also mean that the conventional means for correcting a trade deficit do not work very well. In theory a country with a trade deficit should see its currency depreciate so its exports become cheaper (and imports become dearer) and the deficit corrects. But when a country’s supply chain is largely overseas, a decline in its currency simply pushes up its production costs - no help to its exporters. Hence when sterling plunged after Britain voted for Brexit, the trade deficit remained.

Trump doesn’t understand this, of course, and his flailing about simply adds to uncertainty (that permanent tariff stress disorder) and makes business less efficient. The impact on growth isn’t immediate because there are a lot of factors affecting the US economy, including the AI boom and the fiscal and monetary stimulus being applied by the government and the Fed. But the long-term trade-off between inflation and growth has worsened.

What will happen from here? Legal precedent would suggest that President Trump’s tariffs have no validity although that might not matter to the Supreme Court, which has repeatedly backed the view that Presidents can do what they want (whether they would have done so if Bernie Sanders was in the White House is another matter). But even if the Court rules that tariff-raising powers belong to Congress, that will probably not be a problem for Trump, at least until the mid-terms. At best, then, we might have only one more year of PTSD.

Meanwhile, te rest of the world is learning not to rely on the US. The eminently sensible Mark Carney, Canada’s premier, has just signed a trade deal with China, arguing that it was “more predictable” than the US. That is quite a statement from the leader of a country that has been a staunch ally of the US. As the historian Timothy Garton Ash has written, Trump is making China great again. What a dreadful legacy Trump will have.