Bin Laden's Invoice

The clear plastic bag of liquids has a birth date. On flights from EU airports the 100-millilitre rule has been the baseline since 6 November 2006, and the European Commission says the ban was "envisaged as a temporary restriction".

Nearly everything at the checkpoint has a date like that. Behind all of it sits the attack that came first and cost most.

The 9/11 Memorial counts 2,977 people killed. On 24 August New York's medical examiner identified the 1,654th victim from the World Trade Center, the first by genetic genealogy, and some 1,099 of the 2,753 who died there are unnamed still. The fire department lost 343 members that day, and by its own count last December more than 400 have died since of illnesses from the site.

Three years later the man who sent the hijackers did the arithmetic on camera. In a video aired days before the 2004 election, bin Laden set al-Qaeda's $500,000 against America's loss of more than $500 billion, concluded that "every dollar of al-Qaida defeated a million dollars", and called the policy "bleeding America to the point of bankruptcy".

Yuval Noah Harari later made it a parable: a fly too weak to move a teacup gets into a bull's ear, and the bull wrecks the china shop. "This is what happened after 9/11", he wrote. I had it filed under his name, and when I went looking for the source I found the figures published under the attacker's, before him.

I think America overreacted, and bin Laden does not get the credit. The war that did most to make his figures look clever was Iraq, chosen in Washington and sold on weapons that were not there and on a partnership with al-Qaeda that the 9/11 Commission looked for and did not find. The attack made that choice possible; it did not make it. Harari's parable gets the mechanism right and the agency wrong, though his own prose elsewhere does not: a bull has no choice, and a government does. By his own stated aims bin Laden mostly lost. Nobody can settle the counterfactual, whether a smaller response would have meant more attacks, so every judgement here is about what was known and paid for.

His numbers, and what he expected

I checked his figures first, and they hold up better than I expected. The plot took nineteen men and nearly two years, and cost, by the Commission's rough estimate, between $400,000 and $500,000. His own number sits at the top of that range, and the estimate leaves out the Afghan training camps, so the real cost was higher. America's half he said he took from "the British diplomat and others" lecturing at Chatham House, whom nobody has identified.

Later estimates dwarf his, on a different measure: the Costs of War project puts the budgetary cost of the wars that followed at about $8 trillion, counting homeland security, war-borrowing interest and veterans' care to 2050. His own figure came with no measure attached, and later in the same speech, addressing American voters, he put the loss on "the American people and their economy". A plot's outlay and a nation's economic loss are two different measures, and they do not divide into a ratio. He had every reason to exaggerate, and a motive to inflate is not evidence about the content. The weak part of his claim lies elsewhere.

His own version of the idea is specific. The administration, he said, had made itself easy to provoke and bait, and his model, in the same speech, was the Soviet war, in which al-Qaeda and the mujahidin had bled Russia for ten years until it went bankrupt.

The record from before the event is split, and both readings deserve printing. On the first, he wanted this: the 9/11 Commission cites a source from February 2001 saying that after the attack on the USS Cole bin Laden "wanted the United States to attack, and if it did not he would launch something bigger"; a 2006 reconstruction argued that his goal for at least five years had been to goad America into invading Afghanistan, reading the 1998 embassy bombings and the Cole as stakes raised for it; and a study in International Security argued that America "was ripe for provocation, and al-Qaida appears to have understood this".

On the second, he expected America to run, as it had from Somalia, which he offered as proof that "the American soldier was just a paper tiger". The scholar who read the files from his compound in Abbottabad, 96,000 of them with two research assistants, concludes that he "never anticipated that the United States would go to war in response", and that when the invasion came he "had no plan to secure his organization's survival". That is her reading, and it is contested.

Two things tilt me away from the designed-bankruptcy story without taking me all the way to hers. One is what his commanders said in private: by July 2001, the Commission reports, several senior figures were said to oppose the attack, one of them, on a single source's account, because he "feared the U.S. response", in a record of the dissent the Commission calls "probably incomplete". The other is the calendar: the texts that present provocation as the plan, among them a 2005 book on al-Qaeda's strategy, came after the wars had begun, and the International Security study's own evidence for that reading is the 2004 video.

My reading is a third position. He wanted a reaction and expected one he could survive, perhaps a war in Afghanistan, where he had watched a superpower bleed, and the bankruptcy design looks to me like a story assembled once the wars had been chosen. So I write the tests down before measuring, as I did with a population forecast. Proportion: did a part of the response cost more than the threat it addressed? His own aims: did it advance what he asked for beforehand, which was American soldiers out of Saudi Arabia, the demand in his 1996 declaration? Bankruptcy was added in 2004. Cause: did the attack produce the decision, or make it possible for people who wanted it already?

The door, the queue and the roads

The measure I rate highest is one passengers never see. Reinforced cockpit doors were required on certain American transport aircraft by 9 April 2003. Two risk analysts put the cost at $300 to $500 million for American airlines over a decade, against about $1.2 billion a year for air marshals. The writer who popularised the phrase "security theater" argued in 2010 that "exactly two things have made airplane travel safer since 9/11", that door and passengers convinced they must fight back, and that "everything else has been a waste of money", though he would keep the screening of checked bags and airport staff. I think that is close to right.

The best-known attempt since 9/11 ended in the cabin, with people: the man who tried to light explosives in his shoes in December 2001 was held down by crew and passengers. The door carries a cost nobody chose, as Germanwings Flight 9525 showed in 2015, when the co-pilot locked his captain out and flew into a mountainside, killing all 150: French investigators found the requirements "made it impossible to enter the flight compartment" in time. They still judged an attack from the cabin the greater risk, and so do I.

Everything visible at the checkpoint is younger and, to me, harder to defend. The liquids rule is the clearest case I know of an emergency becoming furniture. It answers the transatlantic plot of August 2006, whose plotters, investigators conceded, had not yet made a viable bomb or bought tickets when surveillance and arrests stopped them nowhere near a checkpoint. Newer scanners that let passengers keep larger bottles were quietly retiring the limit, until in 2024 the European Commission ordered airports using them back to 100 millilitres. A measure nobody re-decides becomes a default, paid for in queue time by everyone who flies, which is friction added on purpose.

What the checkpoints catch is classified, and the leaks are poor: officials briefed on a 2015 report told ABC News that undercover testers got mock explosives or banned weapons through in 67 of 70 tests. The same two analysts put the decade's rise in American homeland-security spending at more than $1 trillion, which to pay for itself needed to save nearly 11,500 lives a year.

The fear landed on people, not only on budgets: 52 per cent of Americans told a pollster last month that Muslims in the country have been treated worse than people of other religions since the attacks. Its deadliest domestic effect may have been on the roads. One estimate puts the extra American road deaths in the twelve months after the attacks, as people drove instead of flying, at 1,595, more than six times the 246 passengers and crew, hijackers aside, killed on the four planes. Its author calls it a nonexperimental estimate, and a re-analysis of the same data for October to December 2001 found the rise only in the north-east. I read it as a real effect of uncertain size, and I count it.

The law moved faster than the doors

The USA PATRIOT Act, whose section 215 later carried the bulk collection of Americans' phone records, was signed on 26 October 2001, six and a half weeks after the attacks. The authorisation for the use of force had passed on 14 September, 98 to 0 and 420 to 1, with Barbara Lee the only "no" in either chamber. Its sixty words authorise "all necessary and appropriate force" against whoever the President determines planned, authorised, committed or aided the attacks, or harboured them.

My own job includes answering a lawful request for data that happens to live on a machine I administer, so I look for whether anyone could show the collection helped. The government's own oversight board reviewed the bulk phone records and in 2014 found no instance involving a threat to the United States in which the programme "made a concrete difference in the outcome of a counterterrorism investigation", though two of its five members wrote separately, one disputing that analysis. The authority lapsed in March 2020, with a grandfather clause that still covers investigations under way and offences already committed, and Section 702, the larger foreign-intelligence power, lapsed as a statute this June, with collection continuing under certifications to March 2027. Neither lapse came from a vote to end anything, which I find telling. Its defenders point to the Senate Judiciary chairman's claim that 702 let Austria stop the 2024 plot against Taylor Swift's Vienna concerts: the strongest thing they have, and still one plot, claimed by a senator arguing for renewal rather than shown in a published finding.

Torture is where I think a choice most plainly spoiled the case it was meant to serve. The Senate Intelligence Committee's study found that the CIA had held at least 119 people, at least 26 of them wrongfully, that its techniques were "not an effective means of acquiring intelligence or gaining cooperation from detainees", and that its own records show Khalid Sheikh Mohammed, accused of planning the attacks, waterboarded at least 183 times. The committee's Republican members disputed that, counting 32 of the 39 men so treated as having produced some intelligence, while conceding that the count says nothing about quality. The techniques then followed the prosecution into court: a military medical panel, AP reported, linked a co-defendant's psychosis to "his torture and solitary confinement", and in 2023 Ramzi bin al-Shibh was ruled unfit for trial.

What I find hardest to defend is that the 2001 authorisation is still law. Congress repealed the 2002 Iraq authorisation in December 2025 and left the older one in place, and the White House's latest war-powers report cites it, alongside the President's constitutional powers, for the detention at Guantánamo, for operations in Iraq and Syria, and for strikes on two Islamic State sites in Nigeria last Christmas Day.

The war that answered the attack

On cause, Afghanistan is the straightforward case: a response to the attack, against the organisation that carried it out and the government that sheltered it, with the Security Council recognising the right of self-defence the next day, without naming Afghanistan or authorising force. The strikes began on 7 October, the Taliban had left Kabul by 13 November, and on proportion I think that opening campaign passes too: it was quick and, next to what followed, cheap.

It also let bin Laden go. A Senate Foreign Relations Committee majority staff report found fewer than a hundred American commandos at Tora Bora, requests for reinforcements refused, and bin Laden walking into Pakistan around 16 December 2001, while General Franks and Vice-President Cheney maintained the intelligence on his location was inconclusive. The same report records that on 21 November, in Bob Woodward's account, the President asked Donald Rumsfeld "What kind of a war plan do you have for Iraq?", and that Franks wrote in his memoir of being told to "dust it off" while he worked on air support for Tora Bora. I cannot read that sequence as anything but attention moving elsewhere.

Then came two decades. America appropriated more for reconstruction than the Marshall Plan cost in real terms, by its own inspector general's reckoning, and then withdrew and watched the Taliban enter Kabul on 15 August 2021. This August the UN's sanctions monitors reported that "Al-Qaida's status and strength in Afghanistan remained unchanged".

My view is that the war was justified at the start and lost that justification somewhere in its second decade, and that it comes nearest to bin Laden's script. About a week before he was killed he wrote that the jihad in Afghanistan had been "bleeding" America until Muslims regained "some self confidence and daring".

Iraq was already on the desk

The case that Iraq was chosen does not, as I read it, rest on hindsight. In the Iraq Liberation Act of October 1998 it was the sense of Congress that American policy should be to "support efforts to remove" the regime headed by Saddam Hussein, though the law did not authorise force. The two political scientists behind that International Security study, who argue that the invasion served al-Qaeda's purposes, still note the administration's eagerness to overthrow Saddam Hussein, "a desire that predated the September 11 attacks".

What the attack supplied was the sale. The authorisation of October 2002 cites 9/11 in its preamble and says "members of al Qaida" are "known to be in Iraq"; in February 2003 Colin Powell told the Security Council of "the potentially much more sinister nexus between Iraq and the Al Qaeda terrorist network", centred on a network he said Iraq harboured, "headed by Abu Musab al-Zarqawi". The 9/11 Commission found "friendly contacts" but no evidence of "a collaborative operational relationship". The Iraq Survey Group found the undeclared chemical stockpile destroyed in 1991 and the nuclear programme ended that year, but also that Saddam wanted to recreate the capability once sanctions were lifted, the one finding a defender of the war can still point to, a judgement about intentions with "no formal written strategy or plan" behind it.

The sale worked. In August 2003 69 per cent of Americans thought it at least likely that Saddam had been involved in the attacks, and a political scientist who rates terrorism's record a failure calls them the "critical impetus" for public support of the invasion. That is what I mean by enabled: the attack gave the argument its force.

The best case for the other side starts here. Without 9/11 there is no Iraq war, so the attack was a condition without which none of it happens, and on that reading bin Laden earns everything the war cost. My answer is that a deliberate decision, taken by people who wanted it beforehand, is not something the attacker authored, however much the attack helped them. Panic is different: it is the reaction itself, and I have counted it above. The decision, the target and the evidence were Washington's, the desire was on the record in 1998, and the war plan was back on the desk ten weeks after the attack. Nor do his own papers point at Iraq: on the same scholar's contested reading, he never anticipated a war in response at all. "Critical impetus" describes permission, not the choice of target.

Which war cost more depends on the allocation rule: the Congressional Research Service put Iraq at 51 per cent of war appropriations through 2014, while the Costs of War project's rough split through 2022 puts Afghanistan and Pakistan slightly ahead. In lives the problem is worse: the project's 550,000 to 580,000 direct deaths covers Iraq since 2003 together with Syria since 2014, out of 905,000 to 940,000 across Afghanistan, Pakistan, Iraq, Syria and Yemen through 2023, so it is no clean Iraq total and no comparable Afghanistan figure sits in the record I am using.

What grew out of it was worse than what went in: Zarqawi himself founded al-Qaeda in Iraq in 2004, after the invasion, and that group became the Islamic State. I have argued before that Iraq is no answer to what Russia has done in Ukraine, and nothing here offers it as one. On its own terms the war fails the two tests that matter to America: the threat it addressed was smaller than the price, and the decision was taken in Washington.

The strangest witness is bin Laden, who in that 2004 speech told Americans that those who said al-Qaeda had won "have not been precise", because "one cannot say that al-Qaida is the sole factor in achieving those spectacular gains" when a White House policy "that demands the opening of war fronts" had helped. That is the author of the invoice conceding he did not write all of it, days before an American election, and I would not lean on it harder than that.

Article 5, invoked once

The wars I have been pricing were not only America's. NATO says it has invoked its mutual-defence clause once, for the United States. The North Atlantic Council agreed the next day, conditionally, that an attack directed from abroad would count, and confirmed on 2 October that this one had been.

European soldiers went to Afghanistan and died there. Casualty compilations count about 3,600 coalition military deaths, roughly 2,460 of them American, which leaves more than 1,100 allied and partner troops, 457 of them British. Estonia, where I live, sent rotations to Helmand "without caveats limiting combat exposure" and lost nine soldiers, which one analyst puts among NATO's highest losses per head of population, inside the European range rather than outside it.

In January this year, amid the Greenland dispute, President Trump said of the allies "We've never needed them" and that their troops in Afghanistan "stayed a little back, a little off the frontlines." Keir Starmer, then prime minister, called that "insulting and frankly appalling", and two days later Trump posted praise for Britain's soldiers alone, naming no other ally and stopping short of an apology. I put weight on that exchange because the clause works only while it is believed.

Who else borrowed the vocabulary

Other governments took the words. Thirteen days after 9/11, in the brief honeymoon with the West I have written about, Putin said that "what is happening in Chechnya cannot be viewed out of the context of the fight against international terrorism", and after the Beslan school siege, in which more than 330 people died, 186 of them children, he ended the direct election of Russia's regional governors on the same grounds, until 2012. In Xinjiang the UN human rights office found serious violations in the context of the government's counter-terrorism and counter-"extremism" strategies, findings China rejected as groundless.

Europe built its own apparatus, and my reading of the documents is that most of it was built for Europe's own dead. After 191 people were killed in Madrid in March 2004 and 52 more in London in July the following year, the EU harmonised the retention of telecoms data, citing declarations on terrorism alongside organised crime, and its court struck the directive down in 2014 as a "particularly serious interference" with privacy beyond what was strictly necessary. I think that pruning is the better half of Europe's record, and the building continues: since April the Entry/Exit System has replaced passport stamps with fingerprints and facial images for non-EU travellers on short stays.

Measured by his own aims

Bin Laden set his own test. The 1998 statement he signed made killing Americans "an individual duty for every Muslim" and named the aims: to liberate al-Aqsa and the mosque in Mecca, and for American armies "to move out of all the lands of Islam, defeated and unable to threaten any Muslim." Its first grievance was the American presence in Arabia.

On that one he got something. On 29 April 2003 in Riyadh, Rumsfeld said the American aircraft there would be able to leave "by mutual agreement", citing the change of regime in Baghdad. Most troops and all aircraft left that June from a presence of about 5,000, the Saudi government denied asking, and troops began returning in 2019. The political scientist who coded terrorist groups' records calls it a "limited success", small beside the roughly 150,000 additional American troops sent to the same theatre. I price it lower still: a concession handed over by a war he neither asked for nor foresaw shows nothing about whether his method works.

On the rest he lost. I know of no regime he opposed that fell to his movement. The letters recovered at Abbottabad show a man whose hold on his own affiliates the reviewers called a seeming inability to exercise control, urging them to abort attacks that killed Muslim civilians and to focus on the United States, "our desired goal". One letter, probably his or a senior aide's, concedes that America could still "bring down any state that we might establish". When governments did fall, in 2011, it was to uprisings, and in his last letter he claimed a share anyway, which I read as the same move as the 2004 video: credit taken for events other people made.

America did not go bankrupt. At their peak in 2008 the wars cost 1.2 per cent of GDP in military operations, against 2.3 per cent for Vietnam. Federal debt has risen almost sevenfold since the attacks, from $5.77 trillion in September 2001 to $40.07 trillion this month, and the wars are part of that. Yet even if the whole $8 trillion the Costs of War project counts, spent and projected, had been borrowed, it would be about a quarter of that increase, by my deliberately generous arithmetic, so the debt has other authors.

His organisation broke. The same scholar concludes, on a reading other analysts dispute, that it never recovered the ability to launch attacks abroad in his lifetime. He was killed at Abbottabad in May 2011, and this August the UN's monitors reported the senior leadership "remained marginalized", its capability and intent to attack abroad "unclear". The brand outlived him. The same monitors put al-Qaeda's Sahel affiliate at 7,000 to 8,000 members, and more than half the world's terrorism deaths last year were in the Sahel.

Concealment on the record

The inside-job claim comes back every anniversary, and the towers carry it. The reconstruction by NIST, the American federal standards laboratory, has the impacts stripping the fireproofing and the fires weakening the frame until the perimeter columns buckled, and finds no corroborating evidence that planted explosives brought them down. For the third tower, 7 World Trade Center, which no plane hit, NIST traced the collapse to a fire that burned for almost seven hours until a girder walked off its seat at Column 79, and calls it "the first known instance of the total collapse of a tall building primarily due to fires".

That report is also where the strongest objections sit. It concedes about 2.25 seconds of free fall in the north face, a stage its own August 2008 draft had not identified, and it says the building, its records and the steel were gone before Congress funded the investigation, so some facts could not be established. Neither admission, as I read it, supplies the explosives: by NIST's own calculation, the smallest charge able to fail that column would have been heard at 130 to 140 decibels half a mile away, and no witness reported it. The President's account of watching the first plane hit on television is, as I have written before, an ordinary false memory: no footage of the first impact is known to have reached the air that day.

The theories also flatter the government they accuse. An inside job needs a state that can plan a mass murder and keep it quiet for twenty-five years, and the state in this record learned in January 2000 that a future hijacker held an American visa and in March that another had flown to Los Angeles, registered neither man on the State Department's watchlist until 24 August 2001, and by then, the Justice Department's inspector general found, searched for them with no "urgency or priority". That competence is not in the file, which is why I suspect the story appeals: a story with a villain is easier to hold than one about institutions failing.

There was concealment, in places the theories rarely look. The President's Daily Brief of 6 August 2001, headed "Bin Ladin Determined To Strike in US", reported FBI information showing "patterns of suspicious activity in this country consistent with preparations for hijackings or other types of attacks". The White House declassified it only in April 2004, after the Commission asked, saying then that the item did not discuss using planes as weapons, and the Commission reported the President calling it historical in nature: what was withheld was the paper, not a specific warning. Part four of the 2002 congressional inquiry, the "28 pages", was held back until July 2016, and found that some hijackers "were in contact with, and received support or assistance from, individuals who may be connected to the Saudi Government", while noting that the inquiry had not itself verified that material.

The intelligence committee's ranking member said on publication that the agencies and the 9/11 Commission had investigated those leads and never found sufficient evidence for them. On Omar al-Bayoumi, the man at the centre of them, the 9/11 Commission wrote that it had seen no credible evidence that he believed in violent extremism or knowingly aided extremist groups. Later FBI records cut the other way: a memorandum released in 2021 says, as Al Jazeera reported, that his help to two hijackers in San Diego "included translation, travel, lodging and financing", while offering no evidence of Saudi government complicity, and Riyadh has long denied that he acted for the kingdom. The families' case against Saudi Arabia survived a motion to dismiss in August last year, on jurisdiction rather than liability and with the disputed facts left for trial, and the appeal has not yet been argued. The 28 pages read thinner to me published than they did secret, and the memorandum leaves a question for the court.

All of that came out through a congressional study, declassification orders, court filings and reporting, which is what being capable of being caught looks like. The biggest public misrepresentation of the period needed no conspiracy. Iraq was sold in the open, and a Senate committee later found that statements by the President and the Secretary of State suggesting an Iraq and al-Qaeda partnership "were not substantiated by the intelligence". The vice chairman's office answered that flawed intelligence, not administration deception, was the basis for it, and the WMD Commission, appointed by the President, found no evidence of political pressure on the analysts, while stressing that it was not authorised to examine how policymakers used what they were given. I still read a pattern of certainty nobody had, and the correcting was public too, which mostly works, if slowly.

Where the parable breaks

The mechanism in the image is real. A weak group does most of its damage through the reaction it provokes, the logic I described when writing about Russia's pressure on Estonia. The 1975 essay that made the argument first also drew the corollary that tends to get dropped: "You can always refuse to do what they want you to do." Where the image fails is that it takes the decision away from the only party that had one. Harari's own 2015 essay put it nearer the way I would, that the American debacles in Iraq and Afghanistan "were the result of the Americans misusing their immense power, rather than of al-Qaida flexing its miniature muscles".

Terrorist groups are not flies, and that is the fair criticism of the image. They have aims of their own, and sometimes they get them. Three days after the Madrid bombings, Spanish voters ousted the government that had sent troops to Iraq and the new prime minister brought them home, although how far the bombings moved the vote is contested. The Taliban, who sheltered bin Laden, govern Afghanistan again.

A verdict, and a trial still waiting

Put through the three tests, the response splits in a way the parable hides. The cheap measures at home worked and much of the expensive apparatus has never been shown to. Afghanistan followed from the attack and kept its justification into its second decade before losing it. Iraq cost more than any threat it addressed and was chosen in Washington. On bin Laden's own aims he mostly lost: his organisation broken, himself killed, no regime toppled, no bankruptcy, and one real concession, from Saudi Arabia, out of a war he had not caused. America overreacted, and the overreaction that mattered most was a choice, which is why I do not give him the credit.

Many Americans do not read it that way. Asked last month, 48 per cent said the operations abroad and the security at home had been worth the cost, against 25 per cent who said not. That is a judgement about values as much as about evidence, and I cannot settle it by counting.

The other limit is the counterfactual. Jihadists have killed 121 people inside the United States since 9/11 by one research group's count, which calls that similar to the 137 killed by far-right attackers, and its 2021 analysis found one deadly attack there directed by a foreign jihadist organisation. Whether that is the apparatus working, the organisation broken abroad, or a threat smaller than it looked in 2001, the numbers cannot tell those three apart. I lean toward the second and cannot prove it.

The case against the man the Commission says provided nearly all the plot's money is still before a military judge. Khalid Sheikh Mohammed was captured in Pakistan in March 2003 and held about three and a half years in secret CIA detention before Guantánamo. Plea agreements signed in July 2024 were withdrawn two days later, and the withdrawal enforced by a divided appeals court last year. On 28 August a military judge ruled the statements he gave FBI agents at Guantánamo in January 2007 involuntary and unusable at trial. That decides nothing about guilt, and the prosecutors chose not to appeal, citing the risk to the trial date. That date, set two days earlier, is Monday 5 June 2028, provided the deadlines are met.