This is part 7 of Crisis Money, a Money Outsider series about what financial crises do to our minds, and what our minds do back. It stands on its own, but part 6 covered the digital herd, and when crowds can be trusted.
On the morning of 14 September 2007, a queue formed outside a branch of Northern Rock in Kingston upon Thames. By mid-morning, queues had appeared outside branches across the country. Over the following three days, depositors withdrew approximately £4.6 billion, roughly a quarter of the bank's retail deposits.1 It was Britain's first bank run in 150 years.
The queue was the signal.
The facts of Northern Rock's situation were, at that point, publicly available. The Bank of England had announced an emergency lending facility. Deposits up to £35,000 were covered by the Financial Services Compensation Scheme. The bank was still solvent in the technical sense. None of this mattered to the people standing in the rain outside branches in Kingston, Golders Green, and Cheltenham. What mattered was the queue. The queue was the signal.
Paul Goldsmith-Pinkham and Tanju Yorulmazer analysed the spillover effects in a 2010 paper for the Journal of Financial Services Research.2 They found that both the bank run and the subsequent bailout announcement had statistically significant effects on the wider UK banking system, measured by abnormal returns on bank stocks. Banks that relied heavily on wholesale market funding, as Northern Rock had, were disproportionately affected. The contagion wasn't irrational panic. It was a rational response to new information about the fragility of a particular funding model, amplified and accelerated by the physical visibility of the run. Alliance & Leicester, Bradford & Bingley, and HBOS all saw their share prices fall sharply in the days following the Northern Rock queues. The visible queue had become a signal that triggered wholesale-market herding, as institutional investors reassessed their exposure to banks with similar funding structures.
There is something very British about the Northern Rock story. Not the bank run itself, which followed classical patterns. But the mechanism of contagion: the orderly queue. People joined the queue not because they had analysed Northern Rock's balance sheet, but because other people were queueing. In Britain, a queue is both a social signal and a social obligation. If enough people are standing in line, the assumption is that there's a good reason. The Northern Rock queue was, in Bikhchandani's framework, a perfect information cascade made physical. Each person who joined confirmed the signal for the next.
The same dynamic, though less photogenic, played out during the pandemic panic buying of March 2020. Lorry Taylor, writing in the Journal of Contingencies and Crisis Management in 2021, found that perceived scarcity drove purchasing decisions more than actual scarcity.3 Social media amplified the signal: photographs of empty shelves circulated on WhatsApp and Twitter, creating a feedback loop in which the fear of running out caused the running out. Moinuddin Naeem, in a 2021 study, identified a specific social-media cycle in which images of empty shelves generated anxious posts, which generated more buying, which generated more images of empty shelves.4 The actual supply of toilet paper, pasta, and flour was, in most cases, adequate. The perceived supply, shaped by photographs that spread faster than delivery trucks could restock, was not. Once the herd started buying, not buying felt like negligence.
And then there is property. The British relationship with buy-to-let investment has all the hallmarks of herd behaviour. Research has found that the growth in private landlordism was driven not by professional investors with diversified strategies, but by small portfolio owners with little experience, attracted by paper profits and an almost religious cultural belief that house prices only go in one direction. Over half of landlords surveyed cited interest rates as a deciding factor, and 40 per cent said they were influenced by recommendations from intermediaries, brokers, and the same social networks that amplified every other herd signal described in this post.5 When prices rose, more people bought. When more people bought, prices rose. The circularity was self-evident to anyone standing outside it and invisible to anyone inside it.
The Investment Association's fund flow data tells the same story from a different angle. In 2020 and 2021, when UK equity markets were rallying, retail investors poured £18.4 billion into equity funds. In 2022, when markets fell, they pulled out £18.2 billion.6 The symmetry is almost comic. Money in at the top, money out at the bottom, driven by the same herding impulse that sent Newton back into South Sea Company stock and depositors into Northern Rock queues. The faces change. The pattern doesn't.
What makes the British version of herding distinctive is its social register. In America, herding is loud. The WallStreetBets crowd celebrated their trades publicly, performatively, with rocket emojis and screenshots of six-figure gains. In Britain, herding is quieter but no less powerful. It happens over dinner when a colleague mentions their buy-to-let yields. It happens at the school gate when another parent asks which platform you use for your ISA. It happens in the silent observation of what other people seem to be doing with their money, filtered through a culture that finds talking about money vulgar but finds keeping up with the neighbours essential. The British herd moves discreetly, which makes it harder to see, and harder to resist.
What this means
The tempting conclusion from all of this is that herding is a flaw, a cognitive weakness that education and self-awareness can correct. If we just understood the mechanics of information cascades, recognised the brain's conformity signal, and remembered the lessons of every bubble from the South Sea to GameStop, we could resist.
That's too easy, and it probably isn't true.
Herding is not a character flaw. It is a species-level cognitive pattern with deep evolutionary roots. For most of human history, copying the behaviour of others was an excellent survival strategy. If everyone in your group started running, the smart move was to run first and ask questions later. The people who stopped to conduct independent analysis of whether the threat was real were, on average, the people who got eaten. Our brains are built to follow the crowd because, for the vast majority of our evolutionary history, following the crowd kept us alive.
The problem is that financial markets are an environment our brains did not evolve to navigate. The signals that were reliable on the savanna (if everyone is doing it, there's probably a good reason) are unreliable in a market (if everyone is buying it, the price is probably too high). The error-detection signal that Klucharev identified, the one that fires when you disagree with the group, doesn't distinguish between "the group has spotted a predator you haven't" and "the group is piling into a cryptocurrency they don't understand." The neural machinery is the same. The context is completely different.
And here is the part that makes this a systemic problem rather than a personal one. The system actively produces herding. Fund managers are benchmarked against each other, which rewards following the crowd and punishes deviating from it. Financial regulation assumes that individual investors make independent decisions, an assumption that every part of this series will challenge. Media coverage amplifies consensus and gives less airtime to dissenting views. Social media accelerates the feedback loop to the point where cascades form and collapse in hours rather than months.
The people who pay the highest price for herding are retail investors, the individuals with the least information, the strongest social-proof signals, and the fewest institutional protections. When a fund manager herds and loses money, they lose a performance bonus. When a pension holder herds and sells their equity holdings at the bottom of a crash, they lose years of retirement income that they may never recover. The career incentives that drive professional herding are a problem. The absence of any equivalent protection for retail investors is a bigger one.
Think about what this means in practice. A 58-year-old teacher in Leeds with a defined-contribution pension has no career incentive to herd. She has no career in investment management. She is not benchmarked against other teachers' pension decisions. She has, in theory, every reason to ignore what the crowd is doing and focus on her own time horizon, her own risk tolerance, her own retirement date. But she is also human. She watches the news. She sees colleagues moving their pensions to "safer" options after a market fall. She feels the brain's conformity signal telling her that disagreeing with the crowd is an error. And the pension platform she uses, with its daily valuation updates and its red-and-green performance indicators, is feeding her exactly the kind of real-time social information that makes herding irresistible. The system is designed as though she is a rational, independent decision-maker. She is a social animal watching the herd.
Newton, for what it's worth, never fully recovered from the South Sea Bubble. He forbade anyone from mentioning the company's name in his presence for the rest of his life. He had calculated the motions of the planets. He could not calculate the pull of the crowd. Three hundred years later, neither can we. The question is whether we can build financial systems that account for that fact, instead of pretending it doesn't exist.
The herd is one half of the story of how crises move through a population. The other half is the deal we believe we have with the institutions that hold our money, and what happens when that deal breaks. That is where this series goes next, starting with wartime rationing and the fairness instinct.
Next in this series: Why people accept some sacrifices but not others. If this was useful, subscribing gets you the rest of the series as it lands.
Further reading:
Savills (2024) UK Housing Market Research. Over-65s generating GBP 10.1 billion annually from buy-to-let; total housing wealth held by over-65s exceeding GBP 2.6 trillion.
Shiller, R. (2000, 2005, 2015) Irrational Exuberance. Princeton University Press. Three editions tracking the dot-com bubble, the housing bubble, and the bond market bubble. A masterclass in how herd behaviour creates asset price distortions.
Shin, H.S. (2009) 'Reflections on Northern Rock', Journal of Economic Perspectives, 23(1), pp. 101-119. Approximately £4.6 billion withdrawn, roughly 25% of retail deposits.
Goldsmith-Pinkham, P. & Yorulmazer, T. (2010) Spillovers from firms. Journal of Financial Services Research, 36(2/3), pp. 145-161.
Taylor, L. (2021) Social media and scarcity perceptions. Journal of Contingencies and Crisis Management.
Naeem, M. (2021) 'Social media role in panic buying amplification during COVID-19', Journal of Retailing and Consumer Services. Social media feedback loops created perceptions of scarcity that produced real shortages.
English Private Landlord Survey (2021). 56% of landlords describe buy-to-let as long-term pension investment.
Investment Association (2022) Fund flow data. GBP 18.4 billion equity inflows 2020-21; GBP 18.2 billion outflows 2022.

