The Strange Death of Keir Starmer’s Brain
Why Morgan McSweeney's Departure Matters
For years, Westminster talked about “Starmerism” as if it were a philosophy.
It was not. It was a management system. And the system has just lost its architect.
Morgan McSweeney’s resignation matters not because he held high office, but because he held something more important. He held the operational logic of the project. He was described as the Prime Minister’s brain. The odd part is that the brain never really belonged to the leader. The leader belonged to the project, and the project belonged to the brain.
To understand why his departure matters, you have to understand the world that built him: Labour Together, the secret projects in Room 216, and the version of Britain that was constructed around a single imaginary voter while real voters quietly walked away.
Labour Together and the Restoration of Control
Labour Together was never just a loose faction gathered around a charismatic figure or political philosophy. It was an answer to a structural shock.
In 2015 Jeremy Corbyn won the Labour leadership and power inside the party slipped. Away from MPs, advisers and donors. Towards members. Hundreds of thousands of people joined. Conference votes began to matter again. Constituency parties started to influence policy and select candidates who looked like the membership rather than the leader’s office.
For the first time in decades Labour began to resemble a mass democratic organisation instead of a parliamentary election machine.
For the people whose political lives had been built in committee rooms, TV studios and think tank offices this was not a gentle disagreement about the correct shade of social democracy. It felt like losing control of the machine.
Labour Together was the answer to that loss. It brought together a very recognisable coalition. MPs and their staff. Policy thinkers like Jonathan Rutherford and Jon Cruddas. Local government leaders. Campaign professionals. A small ring of wealthy Labour-aligned donors such as Trevor Chinn and Martin Taylor who had been close to the party in the past.
What united them was not a detailed manifesto. It was a shared understanding of what a “serious” party is supposed to look like. In that worldview, authority rests with elected representatives, permanent party staff, and established institutions, not with a mass membership who can be swayed by enthusiasm, activist groups and independent media that you cannot control using traditional tactics.
In public, Labour Together styled itself as a place for reflection. Conferences about patriotism, belonging, Labour’s loss of connection with England. Some people inside it initially took that at face value. They thought they were doing cultural analysis and renewal.
Alongside that, something more targeted developed. Labour Together became a coordination network. If you could not be sure of winning internal ballots, you moved the fight outside formal party democracy. Journalists and opinion-formers were cultivated. Donors were kept aligned. Potential parliamentary candidates were quietly identified and encouraged. The goal was not to win votes on the conference floor. It was to decide what would ever be allowed onto the floor.
The 2017 general election raised the stakes. Labour did not win, but it came close enough to prove that a membership-directed programme could plausibly form a government. That frightened people who had assumed that the left could never get near real power again.
Labour Together’s internal SWOT analysis of Corbynism, later described by journalists, made two linked points. A government led by the Labour left, if it collapsed in office, could poison the well for a generation. At the same time Corbyn’s personal position and the party’s rules made it difficult to attack him frontally. The opportunity they saw was precise. Find a leader who could present as “competent Corbynism” to the members, while offering something quite different to the institutions that mattered.
That is how Keir Starmer became “their man”.
He did not arrive as a roaring anti-Corbyn figure. He had sat in the shadow cabinet. He had backed much of the manifesto. He was a senior lawyer, measured in public, plausible to the establishment and, crucially, amenable. Ideal if your sales pitch to the members was continuity, while your real intention was to end it.
At the same time, Labour Together figures were looking hard at the information environment. Members did not only form their views at CLP meetings. They formed them while reading independent pro-Corbyn and left-leaning mainstream outlets on their phones.
In internal discussions Labour Together MPs talked about the need to “destroy The Canary”, the most prominent left-wing site read by ordinary members. That phrase matters. It shows that the battle was not only over who led Labour, but over whose account of reality would be treated as legitimate.
Here the party’s disciplinary machinery became a political tool. Allegations around antisemitism and Israel could lead to suspension or expulsion. That determined who could stand for selection, who could organise, who could speak with the party logo next to their name. The antisemitism issue was real and serious, and it harmed real people. It was also politically decisive. Once you frame disputes as misconduct rather than disagreement, you can transform the internal balance of power without a single formal vote.
By the time the 2020 leadership contest opened, the infrastructure and positioning were in place.
The campaign spoke two languages. To members it offered ten pledges that read like continuity with the 2017 and 2019 manifestos: public ownership, workers’ rights, tax justice, radical climate investment. To the wider establishment it offered stability, respectability, and a clear end to the experiment of party democracy.
Those two lines were not held honestly in parallel. They were held sequentially. First you promise continuity to secure the ballot of the members whose votes you need. Then, once the votes are counted and the keys are in your hand, you quietly lock the door on the promises.
That is what happened. The pledges were dropped one by one. Candidate selections tightened. Dissent narrowed. The new parliamentary party began to resemble the leadership rather than the membership.
None of this was improvised. It was the plan.
And the individual most closely associated with turning that plan into something that existed in the real world was Morgan McSweeney.
McSweeney, The Fraud and the Secret Projects
If Labour Together explains the structure, McSweeney explains the method.
He did not enter politics to write theory. He entered to run operations. During the 2001 general election he worked in Peter Mandelson’s rebuttal unit, using the Excalibur system to track hostile stories and help coordinate rapid responses. If you want a picture of the education that produced the later project, it starts there: in a room where politics is treated as message management, not public argument.
Later he worked in Lambeth with Steve Reed, and on campaigns across east London and Essex. The experience that appears to have hardened his outlook came in Barking and Dagenham, where Labour was losing working-class support to the BNP. The successful fightback was built, not on standing outside housing estates shouting “no to racism”, but on talking about allocation rules, neighbourhood change and local services. The aim was to make Labour look attentive and reassuring rather than morally superior.
The conclusion he seems to have drawn travelled back with him into national politics. A party that looks like it is driven from below by activists and social movements will struggle to be trusted with the state. It will always look volatile. From there, his reading of Corbynism followed.
The failed Liz Kendall leadership campaign in 2015 clarified the constraint. A direct anti-Corbyn candidacy was crushed by the membership. You could not simply tell members they were wrong and expect them to thank you. They would vote you down.
So the strategy changed. You did not confront the membership head on. You reshaped the environment in which they made choices.
After 2017, as McSweeney became central to Labour Together, the project took two linked forms. Succession planning and delegitimation.
Succession planning meant identifying a leader who could win a members’ ballot while reassuring MPs, donors, the media and the state. Delegitimation meant stripping authority away from the culture that had allowed Corbynism to arise in the first place.
That required money, information work and secrecy.
Holden’s book pulls back the curtain. Using internal party documents, Companies House records and correspondence, he describes what he calls the “secret projects” run from Room 216 at the China Works building in south London, Labour Together’s base.
Only a handful of people were allowed to work there. Alongside McSweeney, Holden names staffers Hannah O’Rourke and Will Prescott, plus Imran Ahmed, a Labour operative with a long record of combative, anti-Corbyn spin. Steve Reed, rising through the party from local government to the parliamentary front bench, is ever-present in the background.
Holden describes three strands.
First, an intervention in the internal debate on antisemitism. McSweeney and Ahmed trawled through about twenty pro-Corbyn Facebook groups, with a combined membership of roughly four hundred thousand and an estimated four million posts, harvesting examples of antisemitism, racism and violent language. They produced a dossier of around two thousand “incidents”. According to later accounts by friendly journalists, McSweeney ensured that the most disturbing examples were placed with the Sunday Times.
The resulting splash, headlined “Jeremy Corbyn’s hate factory”, treated these large, often open Facebook groups as if they were official arms of the leadership. The existence of the worst posts was presented as proof of a uniquely hateful culture stamped with Corbyn’s name.
The scale and context were barely mentioned. One administrator later pointed out that the “hate” posts were a tiny fraction of total communications, that volunteer moderators had spent years developing “Corbyn standards” of zero tolerance for racism, antisemitism, sexism and homophobia, that there were meta-groups where administrators from different Facebook spaces shared best practice. That did not make the cut. It would have complicated the story.
Second, Holden shows that Labour Together was centrally involved in setting up the Center for Countering Digital Hate (CCDH) and its anonymous campaign Stop Funding Fake News (SFFN). Internal briefing notes prepared for Reed boasted that Labour Together had “set up” what they mis-named the “Campaign for Countering Digital Hate”, raising start-up funds and providing office space.
Corporate records show that CCDH began life as Brixton Endeavours, a company registered at Labour Together’s address, with McSweeney as sole director. Another company, Labour Campaigns, directed by Ahmed, used the same address. The domain for SFFN was registered to Ahmed personally.
Outwardly, SFFN presented itself as a plucky band of anonymous activists inspired by the US Sleeping Giants campaign, asking advertisers not to fund “fake news”. In reality it was a classic astroturf operation: created by insiders, funded from undeclared Labour Together donations, aimed directly at the pro-Corbyn media ecosystem.
The target list tells its own story. At launch in March 2019 SFFN highlighted four sites. Two were right-wing platforms, Westmonster and Politicalite. Two were pro-Corbyn sites, The Canary and Evolve Politics, both regulated by Impress, with independent assessments that rated their factual content as high even while calling their politics left-wing.
The method was simple but brutal. Take a screenshot of a mainstream advert appearing next to a controversial article. Tag the brand on Twitter. Accuse them of funding hate and lies. Repeat until the brand pulls its ads.
Macmillan Cancer Support was one of the early trophies. The charity was lobbied over an ad that appeared on a Canary piece about Joan Ryan. The key tweet came from a small account calling itself “Dave Gordstein”. Holden shows that “Dave Gordstein” was a fake Jewish-sounding persona run by a non-Jewish activist from Labour Against Antisemitism. Within hours Macmillan announced that it was “taking action” to remove the ad placement. SFFN thanked “Dave” and celebrated publicly. Behind that victory were real journalists facing lost income and job cuts.
The campaign continued. Later, they boasted that The Canary’s monthly traffic had fallen from roughly six million views in 2017 to around 1.4 million by the time of the 2019 election, and that Evolve Politics had tumbled from about two million views to 170,000. It branded that collapse a triumph over fake news. In practice it meant that two of the only significant pro-Labour, pro-Corbyn platforms were badly weakened by the time members went back to the polls.
Third, Holden details work with the Jewish Labour Movement and allied groups to “engineer” the Equality and Human Rights Commission investigation into Labour. Complaints were curated, packaged and pressed. The concept of “denialism” was deployed so that questioning the way antisemitism was being politicised could itself be classed as antisemitism. Reports that drew on weak or tendentious allegations were folded back into the media narrative as further proof of the severity of the crisis.
This does not erase the existence of genuine antisemitism. It does something more uncomfortable. It reveals that, alongside real pain and real prejudice, there was at least one organised hidden hand inflaming and steering events for factional ends, using undisclosed money and front organisations that posed as independent.
When Holden dug into all this, the reaction from those involved was revealing. Instead of addressing the evidence, they tried to wreck the messenger. According to his account, around £30,000 was spent on private investigators. Sources were hunted. Friendly journalists were briefed that he was working for Russia.
After years running an astroturf “fake news” operation that labelled others as disinformation, they treated scrutiny of their own methods as more “fake news” to be neutralised.
Seen in that light, the 2020 leadership contest looks very different.
By then, the succession plan existed on paper and in practice. Starmer was put forward as the vehicle. The ten pledges were the sales pitch to the members. The real manifesto lived in strategy documents and private assurances to donors, journalists and state actors that the programme would change once control had been recentralised. Crucially, the scale of the operation behind him was hidden in plain sight. Under McSweeney’s watch, Labour Together failed to declare hundreds of thousands of pounds in donations to the Electoral Commission until months after the legal deadline and after the leadership contest had finished.
Whatever the internal justification, the effect was simple. Members voted for their new leader without knowing how much money, and which interests, had already been assembled behind “their man”. Members voted for what they thought was competent continuity. The people running the project knew that continuity would not survive contact with power.
That was not an accident. It was the point.
Hero Voters, Gaza and Reactionary Centrism in Office
On paper, McSweeney’s strategy worked. Labour won a huge parliamentary majority in 2024. Four hundred and eleven seats, a majority of 174, off just 33.7 per cent of the vote, the lowest vote share ever to deliver such dominance. A landslide in seats, a shrug in the country.
The nature of that victory mattered. Voters had removed the Conservatives after fourteen years of scandal, stagnation and administrative decay. Nigel Farage’s return had split the right-wing vote and helped Labour over the line. What Labour had assembled was not a mass movement. It was a thin, cautious coalition based on reluctant permission.
The leadership did not treat this as a fragile mandate for serious change. It treated it as a warning against it.
Here the logic sometimes called the “Hero Voter” comes into view. The central character in the story is not the activist, the member, the traditional Labour supporter, or even the left-leaning centrist. It is a hypothetical moderate “middle” voter, usually imagined as slightly socially conservative, Brexit supporting, economically anxious, anti-immigration and easily spooked by anything recognisably left-wing.
Everything is done in the shadow of this imaginary figure. Policy is weighed less on its social impact than on whether it might frighten the Hero Voter. If something looks too redistributive, too pro-union, too sympathetic to refugees, too tolerant of protest movements, it is treated as radioactive, even after you have already won the election.
The result is what you could call reactionary centrism. Instead of balancing between left and right, you define yourself against your own side. It is not enough to oppose the Conservatives. You have to demonstrate to the Hero Voter that you are also willing to stand up to the people who put you in office.
That logic explains a pattern that otherwise looks incoherent.
Harsh rhetoric on migration and asylum, including talk of an “island of strangers”, followed by partial retreats once backbenchers, members and the wider public react with disgust.
An initial response to the Hamas atrocities and Israel’s assault on Gaza that echoed the language of collective punishment, defending measures that cut off water, power and fuel to an entire civilian population, followed by months of resistance to calls for a ceasefire, followed again by staggered and limited shifts in position once the domestic political costs grew too high. Local councillors, many of them Muslim, resigned in waves. Labour’s support haemorrhaged in Muslim-majority wards. The party’s standing among British Muslim voters collapsed, and it was not hard to see why.
The decision to ban Palestine Action under terror legislation sent a similar signal. Protest movements rooted in the left would be treated first as public order problems, not as legitimate political actors.
Inside Parliament, MPs who rebelled on issues from welfare to foreign policy faced discipline. Some saw the whip removed for voting with their conscience, only for the government to move towards their position later. The message was not about policy detail. It was about authority. The party had to look “changed”.
The problem was the Labour voters didn’t like the change. The polling picture followed.
Within eighteen months of the “landslide”, Labour’s vote had not consolidated, it had fractured. An Ipsos poll at the start of February put Reform on 30 per cent, Labour on 22 per cent, the Conservatives on 19 and the Greens on 12. A More in Common MRP projected a Reform majority if an election were held now, with Labour and the Conservatives fighting for second place.
The party that had designed itself to neutralise far-right populist challengers had helped to normalise them. By validating the language and framing of its opponents in order to reassure the Hero Voter, it strengthened those opponents’ legitimacy and weakened its own.
What began as an electoral operating model became the method of government. A strategy built to reassure sceptical observers of your discipline is very effective at sterilising real enthusiasm. It is much less effective once you have to govern through crises rather than commentary.
Mandelson, Epstein and the Moment the Machine Broke
Every political system eventually meets the moment that shows you what it really is.
For the Labour Together system, that moment arrived with Peter Mandelson.
For decades Mandelson carried a nickname that was hardly affectionate. The Prince of Darkness. Twice forced out of office, forever associated with spin, dirty tricks, back-channels and elite networking, he embodied the older New Labour world in which Morgan McSweeney learned his trade. He was also close enough to be described as a mentor to key Labour Together figures, including McSweeney himself and Wes Streeting.
Within that mindset, none of this made him a liability. It made him an asset. He had international connections. He was taken seriously by business and media. He was implacably hostile to the Labour left. He had been involved, formally and informally, in shaping candidate selections and internal alignments. To a leadership whose method relied on reassuring powerful institutions that Labour was once again a known quantity, he looked like a human guarantee.
His association with Jeffrey Epstein was not some late surprise. They knew that it continued after Epstein’s conviction for child sexual offences. In any political culture organised around clear moral red lines, that alone would have been disqualifying for a state role. For the Labour Together mentality, the calculation was different. The risk was reputational, not moral, and reputations could be managed. If you had the right people in the room and the right comms strategy, you could tame any beast.
This is the key to understanding why they thought they could get away with it.
The world they imagined themselves operating in was brutal and transactional. Across the Atlantic they expected a hostile White House dominated by right-wing populism. Diplomacy, in that frame, would rely on private networks and informal leverage rather than open politics. Mandelson, they told themselves, could work that terrain. He spoke the language of donors, lobbyists and professional operators they instinctively treated as the serious people in the room.
So when the decision was taken to lobby for him to become Britain’s ambassador to the United States, it was not a careless aberration. It was a textbook application of their method. You select the figure who best signals continuity and “seriousness” to the people you most want to impress and you trust your ability to message your way through anything that follows. Voters and members are not partners in that story. They are the ball being kicked up and down the pitch while the grown-ups play the game.
This time, the game got away from them.
When documents emerged showing Mandelson sharing confidential government information with Epstein, the problem stopped being one of optics. It became a direct test of judgement and of what the government actually valued. The same traits that had once made Mandelson attractive to the leadership suddenly looked like open vulnerabilities: private channels, dodgy personal networks, and a lifetime spent edging along the line between influence and impropriety.
The scandal landed harder than it might have under a different kind of administration because the government had already stripped away its own political defences.
For two years the governing strategy had been to discipline activists, marginalise internal critics and treat dissent as risk. It had alienated large parts of its base over Gaza. It had frozen out figures like Jeremy Corbyn and Diane Abbott who symbolised earlier eras of Labour support, sending a very clear message to the people who admired them. It had helped legitimise narratives about “fake news” and “extremism” that were easy to turn back on itself. It had bent and twisted policy to please the imagined Hero Voter and still failed to win that voter’s long-term loyalty, while driving away people who had once believed Labour stood for something recognisable.
When the Mandelson story broke, there was no protective layer.
There was no mass of activists instinctively inclined to defend the government in the pub, at work or online. They had been treated as a problem to be managed. There was no deep reservoir of trust among supporters who could say, “I may not like this, but I know what they stand for.” They no longer knew. There were no grateful factions on the left willing to say, “This is bad, but we must hold the line for a Labour government.” Those people had been kicked once too often. The Hero Voters, meanwhile, did what they always do. They looked at the mess, shrugged, and drifted away.
A government built to reassure the “right” people discovered that it had very few people left who wanted to defend it.
And that brings us back to McSweeney. Starmer’s premiership has never been anchored in clear values. It has been anchored in technique. The theory was that you could build a value-free electoral machine, staffed by clever operators, guided by polling and message discipline, and that this machine could master a chaotic world. Donors, lobbyists and senior officials were treated as the serious adults whose anxieties needed to be soothed. Members and voters were treated as a problem to be managed with the right line to take.
Without the operator who designed that machine, what is left?
That is why McSweeney’s resignation matters. Not because a clever backroom strategist has left the building, but because it exposes the emptiness at the centre of a project that confused managing politics with doing it. If your premiership is built on comms rather than conviction, removing the person who ran the comms does not reveal the values underneath. It reveals that there were none.
The strange death of Keir Starmer’s brain is not just a personnel story. It is the story of a government that built its house on the shifting sands of perception, discovered that the tide was coming in, and realised, too late, that it had dismantled the scaffolding it would have needed to survive. Even if it tried to find a moral spine now, after years of punishing those who still had one, it would face the electorate with a question it cannot answer.
Why would anyone trust you when you claim to have changed?