A few weeks ago, I went to dinner in Manhattan with friends who work in finance on America’s East and West coasts. Nothing odd about that, you might think. But this gathering was memorable: over wine, my companions traded tales about the tactics they and their colleagues had used the previous day to yank deposits from troubled banks such as Silicon Valley Bank (SVB) and First Republic.
Some had done this on their laptops or smartphones, sitting in taxis and meetings, or while attending the South by Southwest tech conference in Austin, Texas; others had dispatched emails to their assistants instead. Either way, as the tales piled up, I kept surreptitiously glancing at my own phone for updates on the panic. Physically we were in a sushi restaurant; but in cyberspace we had a ringside, real-time view of a modern-day bank run.
In some senses, it felt wearily familiar. I have watched two financial crises unfold before: once in 1997 and 1998 in Tokyo, as an FT correspondent, when Japanese banks imploded after the 1980s bubble; then in 2007 and 2008, when I was capital markets editor in London during the global financial crisis. I wrote books on both.
Those events taught me a truth about finance that we often ignore. Even if banking appears to be about complex numbers, it rests on the slippery and all-too-human concept of “credit”, in the sense of the Latin credere, meaning “to trust” — and nowhere more than in relation to the “fractional banking” concept that emerged in medieval and early Renaissance Italy and now shapes modern finance.
The fractional banking idea posits that banks need to retain only a small proportion of the deposits they collect from customers, since depositors will very rarely try to get all their money back at the same time. That works brilliantly well in normal conditions, recycling funds into growth-boosting loans and bonds. But should anything prompt depositors to grab their money en masse, fractional banking implodes. Which is what happened in 1997 and 2007 — and what I saw unfold in the sushi restaurant last month.
However, in another respect, this latest panic was different — and more startling — than I have seen before, for reasons that matter for the future. The key issue is information. During the 1997-98 Japanese turmoil, I would meet government officials to swap notes, often over onigiri rice balls. But it was a fog: there was little hard information on the (then nascent) internet and the media community was in such an isolated bubble that the kisha (or press) club of Japanese journalists had different information from foreigners. To track the bank runs, I had to physically roam the pavements of Tokyo.
A decade later, during the global financial crisis, there was more transparency: when banks such as Northern Rock or Lehman Brothers failed, scenes of panic were seen on TV screens. But fog also lingered: if I wanted to know the price of credit default swaps (or CDS, a financial product that shows, crucially, whether investors fear a bank is about to go bust), I had to call bankers for a quote; the individual numbers did not appear on the internet.
No longer. Some aspects of March’s drama remain murky; there is no timely data on individual bank outflows, say. Yet CDS prices are now displayed online (which mattered enormously when Deutsche Bank wobbled). We can use YouTube on our phones, anywhere, to watch Jay Powell, chair of the US Federal Reserve, give a speech (which I recently did while driving through Colorado) or track fevered debates via social media about troubled lenders. Bank runs have become imbued with a tinge of reality TV.
This feels empowering for non-bankers. But it also fuels contagion risks. Take Silicon Valley Bank. One pivotal moment in its downfall occurred on Thursday 9 March when chief executive Greg Becker held a conference call with his biggest investors and depositors. “Greg told everyone we should not panic, because the bank will not fail if we all stick together,” one of SVB’s big depositors told me.
Similar conversations took place in Japan in 1997, physically, in smoke-filled rooms. But few customers knew. Not so in 2023: reports of Becker’s words leaked into the internet, fuelling a stampede. In a few hours, some $42bn — or a quarter of SVB’s funds — departed. Back in 1984, by way of comparison, it took depositors an entire week to withdraw half their funds from Continental Illinois — in person — when that giant lender failed.
The SVB managers asked the Federal Reserve for help in meeting depositors’ claims. But unlike mobile banking, the Fed facilities are open for only a few hours a day. By Friday morning “a total of $100bn was scheduled to go out the door”, Michael Barr, Fed vice-chair for supervision, later told Congress. The bank was dead. Or as Jane Fraser, chief executive of Citigroup, noted: “There were a couple of tweets and then [SVB] went down faster than we have seen before.”
And the panic did not end there: as rumours snowballed, cyber-herds targeted groups that were seen as vulnerable, be that Signature Bank (which was perceived to have mismanaged interest rate risk and had big exposures to real estate), First Republic (which, like SVB, had a high proportion of rich customers whose accounts exceeded the official $250,000 bank insurance limit, making them a potential flight risk), or Credit Suisse (which was so scandal-tainted and poorly managed that depositors were already withdrawing their funds.)
Such contagion had erupted before in finance; think of the crowd panic in the streets of London during the South Sea Bubble of 1720. But as Powell ruefully observed: “The speed of the run [is] very different from what we’ve seen in the past.” Or to cite Fraser again: social media and mobile banking today are a “game-changer” for finance — as in many other areas of our lives.
So how should investors, regulators and bankers respond? One obvious answer would be to drag central banking processes into the 21st century, and keep them operating 24/7 in a crisis. Regulators could also bolster capital reserves, protect more deposits, or make it harder to withdraw money at such times. But ultimately, banks and investors will become more risk-aware — and risk-averse — only by doing what airline pilots do: prepare for future shocks by studying past accidents or near-disasters. And from my observations over three decades in Tokyo, London and New York, there are five key lessons to ponder.
1. No bank is an island
The first lesson is that when a bank implodes, this is almost always a symptom — not a cause — of something askew in the wider financial world, affecting other institutions. Financiers rarely want to admit this. When Fraser of Citi recently appeared before the Economic Club of Washington DC, she insisted in a chirpy, no-nonsense manner that the problems at SVB and Credit Suisse were “idiosyncratic”.
Perhaps so. “Idiosyncratic sounds like idiotic — and Silicon Valley was that,” says Lawrence J White, a finance professor at New York University who formerly worked in government during the savings and loan crisis. Or as Powell observed: “At a basic level, Silicon Valley Bank management failed badly [because] they grew the bank very quickly, they exposed the bank to significant liquidity risk and interest rate risk, didn’t hedge that risk.” In plain English, the core reason a panic erupted was that SVB’s balance sheet was stuffed with long-term Treasury bonds whose value has plunged in the last year as the Fed has raised rates, creating losses.
But SVB was not entirely alone. “Other banks have substantial unrecognised losses on investments and high levels of uninsured deposits,” says White. That stems from the most crucial problem: after 15 years of ultra-loose monetary policy, many financial institutions have strategies that are designed for a low-rate world, and are ill-prepared for higher rates.
I have seen this before. When banks such as Japan’s Long-Term Credit Bank failed in the late 1990s, critics wailed about idiosyncratic “scandals”, such as the fact that some LTCB managers stuffed embarrassing records of loan losses into manholes (yes, really).
But that was a symptom of a bigger problem: the Japanese banks were so flush with extra cash in the 1980s that they threw money into real estate deals that went wrong. Similarly, when Lehman Brothers failed in 2008, and politicians vilified its former head, Dick Fuld, this was part of a bigger pattern: a decade of financial engineering by banks had encouraged dangerous risk-taking. Cheap money always carries costs.
2. Don’t fight the last war
The second lesson is that investors and regulators often miss these bigger structural flaws because they — like the proverbial generals — stay focused on the last war.
Take interest rate risks. These “flew under the supervisory system’s radar” in recent years, says Patrick Honohan, former central bank governor of Ireland; so much so that “the Fed’s recent bank stress tests used scenarios with little variation [and] none examined higher interest rates” — even amid a cycle of rising rates. Why? The events of 2008 left investors obsessively worried about credit risk, because of widespread mortgage defaults in that debacle. But interest rate risk was downplayed, probably because it had not caused problems since 1994.
The global financial crisis was similar: when I asked bankers at entities such as UBS in late 2008 why they had missed mortgage default risks in earlier years, they told me that their risk managers were too busy worrying about hedge funds and corporate loans instead. That was because a big hedge fund (Long-Term Capital Management) imploded in 1998 and the dotcom bubble burst in 2000, creating corporate loan losses. The past is not always a good guide to future risks.
3. Safety is a state of mind
A third, associated, lesson is that items considered “safe” can be particularly dangerous because they seem easy to ignore. In the late 1990s, Japanese bankers told me that they made property loans because this seemed “safer” than corporate loans, because house prices always went up. Similarly, bankers at UBS, Citi and Merrill Lynch told me in 2008 that one reason why the dangers around repackaged subprime mortgage loans were ignored was that these instruments had supposedly safe triple-A credit ratings — so risk managers paid scant attention.
So, too, with SVB: its Achilles heel was its portfolio of long-term Treasury bonds that are supposed to be the safest asset of all; so much so that regulators have encouraged (if not forced) banks to buy them. Or as Jamie Dimon, head of JPMorgan, noted in his annual shareholders’ letter, “ironically banks were incented to own very safe government securities because they were considered highly liquid by regulators and carried very low capital requirements”. Rules to fix the last crisis — and create “safety” — sometimes create new risks.
4. Beware blind spots
Fourth: bankers need to recognise that cultural patterns matter. They often ignore this — in themselves and others — because they are trained to focus on hard numbers. But it mattered hugely with SVB. Its culture emulated its client base, which was mostly from the tech and start-up worlds, which tend to have a “skew” in their concept of risk: they are willing to take bold bets, knowing that there is a small chance of a massive payout (say, if their brilliant idea goes viral), while thinking that they can always reinvent themselves after a downside risk (ie, their company fails). This, as behavioural economist Colin Camerer notes, is different from finance. “Risk management culture, as it’s usually practised [in banks], is antithetical to the Silicon Valley culture.”
Bankers have their own tribal patterns. Another reason why banks failed to see the looming mortgage risks before 2008 was they were often detached from “real” life (ie, what subprime borrowers were doing with their loans) and different teams inside big investment banks were often fighting each other to protect their bonuses and did not share information.
And right now there is another crucial cultural issue that potentially matters even more: consumer behaviour. Torsten Slok, an economist at Apollo, notes that “the share of [US] households using mobile banking or online banking increased from 39 per cent in 2013 to 66 per cent in 2021”.
Until now, the models used in finance do not seem to have taken account of the fact that consumer behaviour online might be different from that in the old-fashioned, physical banking world. But one striking feature about American banks, even before the March panic, was that consumers were moving money out of low-paying deposit accounts into better-yielding money market funds at a dramatically faster pace than at similar points before in history.
That might imply that greater information transparency accelerates consumer reaction to news, even outside crises, increasing the risk of “herding”. Either way, we urgently need some behavioural finance analysis, since American banks will stay healthy only if they hang on to deposits — and digital herding could increase the risks of turmoil in other markets, such as Treasury bonds, if shocks emerge there too.
5. Don’t bet against bailouts
The fifth lesson is that banks are never “just” businesses. In calm times, bankers dress themselves up in free-market language and talk about their profits and business plans as if they were selling hamburgers, laptops or holidays. But that free-market mantra vanishes when panic erupts, since governments almost always step in to protect some depositors, buy bad assets or even nationalise entire banks. That happened in 1990s Japan and around the world during the global financial crisis.
So, too, last month: although deposit insurance was supposed to cover only the first $250,000 of SVB and Signature accounts, the government protected them all, at a cost of more than $20bn. And the Swiss regulators not only protected depositors when Credit Suisse imploded but — controversially — gave some (very small) value to shareholders too. On both sides of the Atlantic central banks have offered liquidity lines to banks (and in America, the Fed is letting banks exchange their holdings of Treasuries for cash at face value, as if rate rises never happened).
Governments do this partly because banking is essential to the wider economy. But also because of contagion. The dangerous weakness of fractional banking is that if nobody has a reason to panic, banks are safe; but if everyone runs, a bank can collapse, even if it previously passed tests on issues such as capital adequacy — unless a government steps in. And while the government never used to worry about smaller banks collapsing, now they fear the digital domino effect.
Maybe governments can contain such risks. After all, the “March madness” — as some journalists and traders now call it — has died down, and the losses have been relatively small to date compared with the previous bank shocks. I can go to dinner without constantly feeling the need to check my phone.
But when I consider the last month, another lesson I learnt from Tokyo and London keeps coming to mind: the trajectory of financial crises can be lengthy, with ebbs and flows. In Japan in the 1990s, the moment of most panic (the collapse of LTCB) came months after the first ructions around Nippon Credit Bank. In the global financial crisis, Lehman Brothers collapsed more than a year after the first subprime mortgage dramas.
I desperately hope we will buck history this time — and ensure that investors and regulators around the world quickly learn from the SVB debacle and improve risk management skills. But I also fear that the past decade of quantitative easing has distorted finance so deeply that there will be unexpected chain reactions, if not in banks, then other corners of finance.
SVB might now have a place in the history books. Sadly, this story is unlikely to end here.
Gillian Tett is chair of the FT’s editorial board and US editor-at-large
Find out about our latest stories first — follow @ftweekend on Twitter