Czech researcher lays out a business case for reducing reliance on Redmond

Comment  A sharply argued blog post warns that heavy reliance on Microsoft poses serious strategic risks for organizations – a viewpoint unlikely to win favor with Redmond or its millions of corporate customers.…

How should we debate immigration? Here's my advice: don't.

Debates don't work, at least not as they should. They don't favour the truth, but plausible liars and for those who can best appeal to prejudice and cognitive bias. Jonathan Portes, who is doing great work in trying to bring facts and rationality to the issue, describes his opponents' strategy: "Flood the zone with a mixture of lies, half-truths, misleading claims and statistics taken out of context." Against this, rational discussion is bringing a knife to a gunfight.

We needn't look far for a parallel here. It's now obvious that Brexit was a bad idea: only 31% of voters say it was the right decision. But have Brexiteers retired in shame? No. Dan Hannan - who gave us the most absurdly cretinous fantasies - still has a seat in the House Of Lords and a newspaper column. The marketplace in ideas is broken; peddlars of crap do not exit as they would in a well-functioning market. Even if immigration scaremongerers could be proven wrong, they'd not shut up. And why would they when they have much of the media on their side? As the old(ish) saying goes, "never argue with a man who buys ink by the barrel."

Put it this way. Imagine that net immigration were to fall to (say) under 200,000 by the next election, a drop of three-quarters from its 2023 peak. Would Reform 2025 Ltd then say "Labour's making great progress; we'll therefore give up"? Of course they wouldn't.

Or to put it another way, the UK has been receiving immigrants for decades - be it Jews in the late 1800s, Pakistanis and the Windrush generation in the 40s and 50s, Ugandan Asians in the 70s and so on: as one who remembers the latter, talk about immigration seems as relevant as Slade or power cuts. If the "debate" about immigration hasn't been resolved by now it's not going to be, because the evidence isn't decisive.

A second reason not to debate immigration (or at least to be very careful about doing so) lies in a distinction between two types of politics - that based on lived experience versus that based on abstractions often gleaned from the media. Some us us think of immigrants as our friends, uncles or the local shopkeeper. The right by contrast invites us to see them as an abstract threat. Hence the fact excellently described by Ben Ansell; people who live in areas with more non-UK born residents tend to be more likely to think immigration is a good thing. One reason, I suspect, why support for Trump's immigration policies has declined is that "immigrants" are changing from the demons mythologized by Fox News into actual real people.

Unless we're careful, debating immigration risks us being like a therapist who indulges his patient's fantasy thus threatening to make the psychosis worse. If we must do so, we must speak in concrete terms, about real people not abstractions.

So, what should we do instead?

Matt Goodwin inadvertently gave us the answer recently on Twitter. "London is over —it’s so over" he said, pointing to unreliable and overpriced trains, homelessness, crime, and struggling restaurants.

Which blurts out the truth. Immigration is not merely about immigration. "Concerns" about immigration are linked to a wider sense of national decline, of economic stagnation and government failure.

Sure, Goodwin might be seeing what he wants. But we now have abundant evidence that economic stagnation fuels the far right. Back in 2006, Ben Friedman showed that, historically, slower economic growth led to increased racism and intolerance. Markus Brueckner and Hans Peter Gruener have found that, across Europe "lower growth rates are associated with a significant increase in right-wing extremism." Thiemo Fetzer showed how areas hardest hit by austerity were more likely to vote for Brexit. He and his colleagues have shown that shop closures are correlated with support for rightist parties. And Diane Bolet has found that areas seeing more pub closures are more likely to support Ukip (as was).

A place in decline invites the post hoc ergo propter hoc fallacy: "since the immigrants arrived, this place has gone to the dogs". We could reverse this: if the cost of living were declining and places thriving, it would be harder for people to see immigrants as a problem - unless of course they were mere racists.

Which points to the way to fight the immigration battle: stop places going to the dogs. We should be saying to voters: "we get that many areas seem in decline, but this isn't because of immigration: here's what we should be doing instead." The better Labour MPs see this. Jeevun Sandher recently wrote:

If we want to win the next election, we need to make life affordable for our voters. Investing to get energy and housing costs down while creating good non-graduate jobs for Reform-curious voters will put more money in these voters’ pockets. If we get cash in peoples’ pockets, we win the next election. If we don’t, we lose.

This is easier said than done. One problem is that these aims are undermined by the government's plan to raise military spending; doing so requires lower spending than would otherwise be the case elsewhere - on either public services or private consumption. And cuts in the latter would exacerbate the pub and shop closures which fuel the far right.

A second obstacle is the large number of vested interests blocking growth. I'm thinking not just of nimbys opposed to infrastructure spending and housebuilding but: incumbent companies opposing tougher competition policy; lawyers, accountants and financiers opposed to simpler taxes; landlords opposing a shift from taxing incomes to taxing land; utility companies opposed to fairer prices; Brexiteers opposed to rejoining the single market; and a financial system generally that likes the low real interest rates that accompany stagnation.

A further problem is the quality of British management. It has become so accustomed to rent-seeking and bullying that it has largely lost the ability to raise productivity.

And all this is before we mention the possibility that capitalism itself might be a barrier to growth. The problem isn't just that low profitability deters investment-enhancing growth: why do so many industries from steel to nuclear require government subsidies? It's also that some investments don't happen because capitalists fear that they will get only a "minuscule fraction" (William Nordhaus's words) of the payoffs from them, whilst others are undertaken out of misplaced optimism rather than a rational assessment of their returns.

Perhaps my pessimism is misplaced. But that's the debate we should be having. And the right don't want it because they don't want to discuss the failure of British capitalism. If they do, they only show the vacuity of their own thinking, be it about Brexit or "Britannia cards". They'd prefer that we demonized migrants than looked at the root causes of stagnation - and certainly prefer that we divide people along ethnic rather than class lines.

We should ask Reform: "why should we talk about your hobby horse when we know that your last pet obsession (Brexit) was a failure, and when we have more important things to think about?"

You might object that we should stand up for the rights of migrants and indeed - given that immigration controls also reduce their liberty - the rights of citizens too. Morally yes, but we have to choose our battles. When we discuss immigration, we are fighting the battle the right wants to have. That ignores Sun Tzu's advice: "he will win who knows when to fight and when not to fight." Setting the agenda matters, and we should not concede this to the right but rather set our own.

Discussing immigration steals cognitive bandwidth. The opportunity cost of doing so is that we don't discuss more important matters - matters that, if we could at least partly resolve them, would boost economic growth and so defeat the far right.

Rachel Reeves is reportedly considering changes to the rules around “non-doms” – people who declare themselves to be “non-domiciled” in the UK in order to pay less tax – in response to a protracted media campaign against the policy. On a daily basis, newspapers from the Telegraph to the Independent are warning that the abolition of the non-dom regime is causing an exodus of millionaires. This week, estate agents have protested that the top end of the London property market is now in freefall. The mansions stand empty!

Questions to ask about these reports include: are they bullshit? And why should I care? Does it matter if the non-doms are leaving? New research suggests a surprising answer.

Firstly, these claims have at least a whiff of the farmyard in that they are being made by people who have a clear financial interest in the matter. If I was selling prime central London property, I would definitely use the media to push the narrative that the wildly overpriced houses I was pushing were in fact going for a song.

That’s not to say they’re lies, but they are claims made in the absence of the actual data on non-doms (which will be released by HMRC next month). Fortunately, there is other data that does tell us a verifiable story about what happens when you change the non-dom rules, because these rules have been changed before – by George Osborne, in 2017. The changes Osborne announced in his 2015 Budget amounted to a £1.5bn tax hike on non-doms. Last month, a study was published which shows what happened next.

Immediately after the 2015 Budget, the emigration rate of non-doms rose, peaking in 2018 – but by 2019, it was lower than it had been before the change was announced. So the policy was announced and an “exodus” appeared to follow, but it wasn’t an exodus. It was people bringing forward a change they had been planning to make anyway.

The second conclusion was even more important: the people who paid the least tax were the most likely to leave. The millionaires who left were the fiscally useless, the do-nothing rich.

A common misconception is that rich people are economically similar. But there are two very different species: people who earn loads of money and people who have loads of money. People who earn lots of money typically pay a lot of tax. A banker in London on £600,000 a year should pay about £245,000 in income tax and National Insurance (assuming maximum pension contributions), contributing as much to the state in three years as the average Brit contributes over their whole working life. This makes them very important from a fiscal point of view. But it also means they’re less likely to leave when there’s a tax change, because they have a great job that it would be hard to get elsewhere, and because they demonstrably already tolerate a high level of taxation. And after the 2015 Budget, this is what they did: they stayed.

The other species are different. They are simply parking their wealth here, paying little tax, and leaving when the rules change. Arun Advani, the economist who led this research, told me that under the non-dom regime, the UK had effectively been a tax haven for people who had money here but didn’t earn money here: “Everyone thinks, because they’re very wealthy, that means they must be paying a lot of tax,” he said, but the people who left after 2015 were recognisable as being “not very well connected” to the economy by income  – “but that’s also why they weren’t paying very much tax. So the cost of losing them is much lower.”

It’s very possible that this is what’s happening now: we are losing people whose connection to the economy was fleeting and low in value.

The stateless rich will be missed in some parts of London. Showrooms for ugly, dangerous cars will fall quiet; gaudy clothes with someone else’s name written all over them will go unsold; bad Mayfair restaurants and galleries for gaudy, laughable art will have to reconsider their value propositions. But the rest of us might benefit, thanks to the property market.

Expensive London property drags the rest of the housing market upward, imposing trickle-down inflation on homeowners. It also eats into our housebuilding capacity. When I walked through London’s wealthiest constituency with its MP, Joe Powell, recently we had to raise our voices over the sound of drills: every other property was a building site, an investment being serviced. The UK currently spends around £3bn a month on repair and maintenance of existing housing, more than six times what it spends on building new public housing. If the non-doms leave, housing might become both more affordable, and easier to build, for people who want to live here for reasons beyond their tax bill.

It’s also worth noting that leaving the country – any country – is something the stateless rich do constantly. In a normal year, with no tax changes announced, around a quarter of non-doms who have arrived recently (within the last two years) will leave anyway, and even among long-stayers (up to 15 years), one in 25 will leave each year. Maybe shopping in Harrods and driving a Lamborghini leaves you feeling empty and bored, and forces you to roam the planet seeking a fulfilment money can never buy? Who cares. They’re off to Dubai. Before the government performs another U-turn, it should ask whether anyone will miss them.

[See also: Has Rachel Reeves given in to the non-doms?]

A good columnist is never unintentionally tedious, but this week’s effort is about obsolete telephone directories, binary counter overflow, and the alternating current waveform. The boredom is the point.

Start with alternating current. As most of us once learnt and have since half-forgotten, mains electricity is supplied by an oscillating current whose direction changes rapidly. In the UK, for example, the current flips back and forth 50 times a second.

This system is highly efficient but suffers from a serious downside: if the frequency slips outside a tightly defined target range, both the system and many of the appliances plugged into it can be damaged. That almost sounds interesting, but of course it is boring after all, because electricity is generated by power stations that all but guarantee a stable, reliable waveform: heavy, rapidly spinning hunks of metal powered by steam heated by burning hydrocarbons or a nuclear reaction. Or so it used to be, but thermal generation is rapidly going out of style in favour of wind turbines that do not spin at a predictable rate, and solar panels, which produce direct current instead.

“Thermal electricity generation provided system stability so effectively and so transparently that we forgot it did that,” says Paul Domjan, one of the founders of Enoda, an electricity grid technology company. We are going to have to remember again, and quickly, because we are rapidly moving to grids that are far more prone to wobble off the target frequency, as recently happened in Spain, with dramatic consequences.

We are all familiar with renewable energy’s problem that the wind does not always blow and the sun does not always shine, but only the real nerds worry about the stability of the AC waveform. This is a problem that can be solved, but not if it is overlooked.

Just in case alternating current starts to seem too interesting, let’s move on to obsolete telephone directories, or more precisely to the diocesan directories published by the Catholic Archdiocese of Boston. An up-to-date directory is useful, if hardly riveting. A decade-old directory describing the former addresses and positions of priests seems good only for the recycling bin.

Yet in 2001, investigative reporters at the Boston Globe suspected that the sexual abuse of children by priests was widespread, and realised that the apparently useless old directories provided a vital clue. In the 1990s, after the church had begun to quietly remove offending priests, there was a sharp increase in the number of priests listed as on “sick leave”, “awaiting assignment” or at the “clergy personnel office”.

Carefully combing through the old directories, the Globe’s reporters assembled a list of priests with unexplained movements through the Archdiocese organisation. That list closely matched their growing list of accused priests, strongly suggesting the church’s complicity.

Because outdated directories turn out to be Pulitzer-prize-winning levels of not-boring, we should turn to binary counter overflows, surely a reliably tedious topic. A computer armed with a three-digit binary counter can count from zero to seven: 000, 001, 010, 011, 100, 101, 110, 111 — and then the counter overflows back to zero. A 32-digit binary counter will get you to nearly 4.3bn before overflowing — 4,294,967,295 to be precise.

This is fairly boring stuff, unless you are working at the Los Angeles Air Route Traffic Control Center and your radio system simultaneously disconnects from each of the 800 aircraft flying over southern California. This happened on September 14 2004, and whatever adjectives sprang to the minds of the air traffic controllers and the pilots, “boring” was not one of them.

The culprit? A binary counter overflow in the traffic control computer clock: it was counting down milliseconds from 4.3 billion, which takes just under 50 days to do. When it hit zero the system shut down. Standard procedure was to forestall any problem by rebooting every 30 days, but in the summer of 2004 that evidently did not happen.

As Matt Parker explains in his book Humble Pi, this wasn’t a one-off error. Windows 95 computers would crash for the same reason, while the Boeing 787 Dreamliner had a similar issue and would lose all electronics if somebody left the computers running for more than 248 days.

The boring things in life will shut down your electricity grid, identify paedophiles in the priesthood and crash your computer — or maybe even your aeroplane. Might we attempt a grand unified theory of boring things?

Perhaps. Here it is: smooth, successful operations are uninteresting, and uninteresting matters tend to be neglected. Eventually they stop working well, at which point they become interesting again.

This is certainly true of the AC waveform. It seems boring because it has felt like a solved problem. Yet, as with low inflation or herd immunity from measles, if we allow the foundations of a success story to be eaten away, we find that the problem isn’t quite as thoroughly solved as we assumed.

This is true also of archives. Even celebratory accounts of the Boston Globe team’s use of diocesan directories usually neglect to mention that if those directories hadn’t been available in library archives, the investigation would have hit a dead end.

Archives have a particular problem. If an archive fails — fails to save the right things, fails to preserve old documents or fails to maintain digital information in a format modern computers can interpret — then the fact of that failure may take years or even decades to emerge.

Success leads to boredom. Boredom leads to neglect. Neglect leads to failure. Failure is no longer boring. But if we don’t show more interest in the successful systems we have built, they may suddenly become far too interesting for comfort. By the time these boring topics start seeming interesting, it’s too late.

Written for and first published in the Financial Times on 23 May 2025.

Loyal readers might enjoy How To Make The World Add Up.

“Nobody makes the statistics of everyday life more fascinating and enjoyable than Tim Harford.”- Bill Bryson

“This entertaining, engrossing book about the power of numbers, logic and genuine curiosity”- Maria Konnikova

I’ve set up a storefront on Bookshop in the United States and the United Kingdom. Links to Bookshop and Amazon may generate referral fees.

Five days after his election victory in February, Friedrich Merz’s world collapses. That’s how he will describe it later.

That Friday evening, he steps off the stage at a large conference center in Hamburg’s port, where cruise ships usually moor. He has just been hailed as “the future federal chancellor,” and more than a thousand party supporters have cheered on their chairman at a rally of the local chapter of the Christian Democratic Union, Germany’s main center-right party. At around 8:15 p.m., he shakes a few hands in farewell, then drops into the backseat of his official car for the three-hour drive home. It is February 28, 2025.

Merz checks his phone and notices a message from his spokesperson. He should watch a video, preferably immediately. Merz pulls out his iPad, opens the link, and recognizes a room familiar to anyone who follows politics. Two armchairs upholstered in gold damask sit in front of a fireplace with no fire burning. In front of the fireplace is a table made of fine wood inlaid with an oversized seal. It’s the Oval Office in the White House.

To Donald Trump’s right sits a small, bearded man in a black military sweater embroidered with a stylized trident, the national symbol of Ukraine. It is Volodymyr Zelenskyy, the president of Ukraine, a country invaded by Russia. Merz holds him in high esteem. Merz has visited Zelenskyy twice in Kyiv and, just a few days ago, accepted Zelenskyy’s congratulations on his election victory. Ukraine has high hopes for Merz. The new chancellor is expected to finally provide the Taurus, a German cruise missile capable of penetrating bunkers, which Merz’s more liberal predecessor as chancellor, Olaf Scholz, refused to provide throughout his time in office.

In the video, Zelenskyy looks tired. Tired and helpless. Merz is dismayed as he watches the U.S. president humiliate his Ukrainian counterpart. Trump accuses him of endangering millions of lives and risking a third world war. When Zelenskyy retorts that it was Russian President Vladimir Putin who started the war, Trump interjects harshly. In front of the cameras, Zelenskyy is scolded like a naughty child for several minutes. “Did you ever say thank you?” Vice President JD Vance asks Zelenskyy, hurling this question at him several times.

“That was good television,” Trump says at the end of the meeting. The subsequent talks, which were supposed to be about security guarantees after a ceasefire, are canceled. A fully negotiated raw materials agreement is not signed. The celebratory lunch is canceled. Zelenskyy waits another 20 minutes in an adjoining room. Then, an official appears and simply sends him away.

Merz has just finished watching the nearly 40-minute scene when he posts a solidarity message to Zelenskyy in English on X: “We must never confuse the aggressor with the victim in this war!” He is on the phone nonstop in the car until he arrives in Sauerland and then for half the night. He also speaks with Scholz, who would still be chancellor for another two months.

Scholz and his designated successor agree that something historic happened that day in Washington. The Americans are threatening not only to abandon Ukraine but also all their allies. Is Article 5 of the NATO Treaty, which requires every member to come to the defense of every other member, still to be taken seriously?

Would U.S. soldiers defend Germany against a Russian attack? Are American nuclear missiles still a credible deterrent?

The two men agree that given these circumstances, Germany must rebuild its national defenses. As quickly as possible and at whatever cost. And it will cost a lot, between 1 and 1.5 trillion euros over the next 12 years — double the previous amount.

Spending that much money on defense isn’t easy. In Germany, the “Schuldenbremse” or “debt brake” is a fiscal rule enshrined in the Constitution. It is designed to limit the amount of new government debt to a maximum of 0.35 percent of gross domestic product.

Before the elections, Merz campaigned on keeping the debt brake and insisted as chancellor he could do without extra debt. But in the coming days, Merz will flip his position and agree to this new borrowing. The humiliation of Zelenskyy has changed everything.

This account of the election of Merz and his first days as Germany’s incoming chancellor is based on more than 50 conversations with sources, some close to Merz, who were granted anonymity to speak freely.

Merz’s doubts about his prior convictions had been building for weeks. A few days before the general election, Merz met with Vance in Munich. Merz wanted to dissuade the American vice president from publicly urging Germans to vote for the far-right Alternative for Germany (AfD) party. “These are not friends of America,” Merz said, “but partisans of Putin.” Vance nodded in apparent agreement.

Just a few hours later, during his speech at the Munich Security Conference, Vance stunned the audience. He declared that restrictions on freedom of speech in the EU are a greater threat than Russia or China. He called for firewalls to be torn down across Europe and for right-wing populists to be included in politics. The vice president did not mention the AfD by name. However, a few hours later, reports circulated that Vance had met with not only Merz, but also with AfD leader Alice Weidel at his hotel before the speech. He had not told Merz about this meeting.

Even then, two weeks before Zelenskyy’s humiliation in the Oval Office and one week before the Bundestag elections, Merz had begun privately considering the need for Germany to take on additional billions in debt. “What the new American president, Donald Trump, has said in Washington these last few days…” he told the audience from the campaign stage in Hamburg, “Ladies and gentlemen, we are witnessing a fundamental shift in the global political landscape.”

Following the Munich Security Conference, Merz discreetly asked former Constitutional Court judge Udo Di Fabio to explore whether it would be possible to amend Germany’s Basic Law with the votes of the outgoing Bundestag. The “Basic Law” is Germany’s equivalent of a constitution. It can only be changed by a two-thirds majority in parliament. That also applies to the debt brake. Getting a two-thirds vote would be possible with the old Bundestag, but not the new Bundestag that was expected to have a higher representation of AfD and other fringe parties.

Shortly afterwards, Di Fabio sent him his expert opinion. Amendments to the Basic Law with the votes of MPs who had already been voted out of office were possible up to 30 days after the election. That would be March 25, the same day the new Bundestag would be seated. Merz would have less than a month to execute an about-face.

On the day of the election, Merz gave the first public signal that his thinking was changing when he appeared with other candidates on the Berliner Runde, a television program in which party leaders comment on the election as soon as the polls close. “For me, it will therefore be an absolute priority to strengthen Europe as quickly as possible so that, step by step, we can really achieve independence from the USA,” he said.

Independence from the USA? Scholz, sitting right next to Merz on TV, could hardly believe it. Until now, European politicians had carefully avoided suggesting that Europe could manage its defense without the Americans. Germany, which has neither its own nuclear weapons nor a robust army, needs American troops and their nuclear umbrella more than anyone. Merz, considered a staunch transatlanticist, was giving up on the USA? “Since U.S. President Donald Trump’s statements last week, it has been clear to me that this administration is largely indifferent to the fate of Europe,” Merz continued. A summit of the transatlantic military alliance is scheduled for the end of June. “Will we even be talking about NATO in its current form then?” he asked. “Or will we then have to establish an independent European defense capability much more quickly?”

The next day, when the election results had been tallied, Merz praised the outcome in a press conference: 29 percent was much less than the Christian Democrats had hoped for, but Merz argued it was a success if you look at the number of votes rather than percentage points. The Christian Democrats gained 2.5 million votes compared to the previous Bundestag election, and the Christian Socialists gained 500,000, he noted.

What he failed to mention is that the AfD gained over 6 million votes. After an election campaign more polarizing than any in decades, more people turned out to vote than in previous years, and the AfD was the beneficiary.

Merz was genuinely outraged by the scene in the Oval Office. But he also knew he could use this indignation to his advantage. After all, he would need a credible narrative to justify the political turnaround, the astronomical increase in defense spending, that will take place under his leadership.

The election results meant that, for the first time since World War II, centrist parties no longer have a two-thirds majority in Parliament. Without a two-thirds majority, centrist parties cannot elect judges to the Federal Constitutional Court, declare war on an invader or amend the Basic Law. For example, to reform the debt brake.

The situation is reminiscent of the late phase of the Weimar Republic. At that time, the National Socialists and Communists together held over 50 percent of the seats in the Reichstag, preventing the Social Democrats, Liberals and Christian Democrats from governing effectively — thus fueling growing frustration with democracy. This created a vicious circle that led to the collapse of the first republic at the beginning of the 1930s.

Is this a bold comparison? The AfD and other fringe parties already control a blocking minority in the state parliaments of three German states: Saxony, Thuringia and Brandenburg. The same will be true in the Bundestag when the new MPs are seated March 25.

Scholz also played a role in urging Merz’s turnaround. In meetings unnoticed by the public, Scholz and Merz met several times in the chancellor’s office after the Bundestag elections, sometimes with other center-right politicians present. At one of these meetings, Scholz presented intelligence service findings on the immense scale of the Russian arms buildup. Despite the enormous losses in Ukraine, Putin would have considerably more tanks and missiles in just a few years than before the invasion. The intelligence suggested he is preparing to wage another war, this time against Europe. Scholz, who campaigned as a peace chancellor, advised his successor to do the opposite: to massively rearm.

Germany’s new government coalition joined Merz’s Christian Democrats with Scholz’s Social Democrats. In the days after the election, the coalition partners convened private negotiations to reach a spending plan they could implement before March 25. In those talks, the sums involved increased by the hour.

On March 4, when the partners reappeared in public to announce their deal, there was great astonishment. There were no longer any limits to rearmament. Merz secured special funds for a defense build-up over the next 10 years that were five times larger than an increase Scholz negotiated just three years ago. An additional special fund of 500 billion euros had been agreed upon for rebuilding the country’s infrastructure.

Why was Merz, the avowed debt hawk, now so willing to push Germany so deep into debt?

“In view of the threats to our freedom and peace on our continent, the same must now apply to our defense: Whatever it takes!” Merz said at a press conference. The saying was a quote from Mario Draghi, the former head of the European Central Bank, who used this slogan in 2012 to scare off speculators who wanted to bet on a breakup of the eurozone. Now Merz used the same quote to explain his rearmament plan.

At a parliamentary group meeting later that day, Merz reported that he would be traveling to Brussels to take part in the meeting of the heads of state and government of the EU Council. And then he said something curious: “If Trump announces his withdrawal from NATO tonight, then we, the Federal Republic of Germany, will be the first to have reacted correctly in advance.”

There was horror among the MPs. Merz was deadly serious. The total turnaround in financial policy began after the shock appearance by Vance at the Munich Security Conference. Merz justified it by pointing to the humiliation of Zelenskyy at the White House. But now he was talking about an imminent U.S. withdrawal from NATO. How did Merz get this idea?

Trump was set to give his first speech to a joint session of Congress that same night. Merz explained to close allies later that he had received information from an American source indicating that Trump would use the speech to announce a U.S. withdrawal from the Western defense alliance. He had reason to trust his source. Two weeks earlier, the source had provided him with advance information on Vance’s speech at the Munich Security Conference. Merz held a conference call the night before the speech and warned Christian Democratic leaders that Vance would shake the transatlantic friendship and launch a rhetorical attack on Europe. That is exactly what happened. Merz and his allies were prepared.

Warned once again, Merz expected the worst from Trump’s speech to Congress. During conversations and phone calls with confidants, he made it even clearer than he had in the parliamentary group meeting that if Trump announced a NATO withdrawal that night, Putin might react immediately with an attack on the Baltic states.

During those hours when he agreed Germany should take on a trillion-euro debt, Merz was acting on the belief that a new war in Europe was possible and NATO was on the brink of collapse. His vote in favor of the record debt came against this dramatic backdrop.

As we know, things turned out differently. Trump delivered his congressional speech but did not mention a withdrawal from NATO.

To this day, Merz does not believe that his Washington source misinformed him. The NATO withdrawal announcement had been prepared, he believes. Trump changed his mind at the last minute.

(POLITICO Magazine asked the White House to respond to the assertion that Trump had considered using his March 4 speech to a joint session of Congress to announce a U.S. withdrawal from NATO. In an emailed statement, White House spokeswoman Anna Kelly said, “Such an announcement was never included in any draft of any speech.”)

In the Middle East as in Ukraine, the president is discovering that simple bullying tricks don’t resolve complex international crises

It was as close as Donald Trump might get to a lucid statement of his governing doctrine. “I may do it. I may not do it,” the president said to reporters on the White House lawn. “Nobody knows what I’m going to do.”

The question was about joining Israeli air strikes on Iran’s nuclear facilities. Days later, US bombers were on their way. Some expected it to happen. Others, including Keir Starmer, had gone on record to say they didn’t. No one had known. The unpredictability doctrine wouldn’t have been violated either way.

Rafael Behr is a Guardian columnist