Symposium probes cost of bank failures

Facebook
X
WhatsApp
Telegram
Email

LET’S READ SUARA SARAWAK/ NEW SARAWAK TRIBUNE E-PAPER FOR FREE AS ​​EARLY AS 2 AM EVERY DAY. CLICK LINK

BANK failures may be uncommon, but when they happen the fallout is deep and lasting. 

Research shows that systemic bank runs can shrink an economy by almost a tenth, leaving scars long after deposits return.

This was discussed during the ‘Failures are Rare – But Can We Afford One?’ session at the National Resolution Symposium (NRS) 2025, hosted by Perbadanan Insurans Deposit Malaysia (PIDM) in Kuala Lumpur. 

The discussion was moderated by Deputy Director of Bank Negara Malaysia’s Monetary Policy Department Amanda Chong, and featured National University of Singapore’s Dr Karsten Mueller and Ruth Walters from the Bank for International Settlements.

Historic patterns, modern risks 

Presenting a dataset covering more than two centuries and 184 countries, Mueller said his team had identified 316 run events by combining 503 historical sources with cross-country banking deposit statistics from 50 sources. 

Of these, he said, more than 300 episodes qualified as bank runs, and roughly half showed systemic characteristics where overall deposits contracted.

He noted that runs have become more frequent since the 1970s, matching levels seen in the 1920s and 1930s.

“When a systemic run happens, there is a permanent contraction. On average, your economy is 9 per cent smaller. 

“Even when there are no solvency issues, runs can still cut GDP by about 5 per cent,” he said.

While traditional thinking attributes runs mainly to solvency problems, he stressed that liquidity shocks alone can destabilise economies. 

He stressed that three pieces of evidence show solvency is insufficient: countries still lost about 5 per cent of GDP when runs occurred without solvency problems; they suffered the same contraction even when no banks actually failed; and even well-capitalised banking systems shrank by 5 per cent when systemic runs hit.

“Bank runs can be damaging to the real economy even if there are no failures. 

“The depth of recession is strongly correlated with the size of deposit contractions. The sharper the outflow, the deeper the GDP decline,” he added.

The impact is not confined to historical cases. 

He pointed out that since deregulation in the 1970s and 80s, runs have become more common, and today withdrawals can spread in seconds through wholesale funding rather than queues at bank branches.

“The likelihood of any kind of run materialising has never been higher than today.

“The unconditional probability of a run is about 2 per cent, but during credit booms with unusually low risk premia, the chance can surge to nearly 50 per cent,” he added

Safeguards and their limits  

Deposit insurance has helped prevent runs from escalating, especially when there are no fundamental problems. 

Mueller pointed out that with insurance, the chance of a run turning systemic is significantly reduced. 

“For runs without a fundamental reason, deposit insurance all but eliminates the likelihood that they become systemic,” he said.

But he also stressed that once a run has already turned systemic, the cushioning effect is limited. 

“After the bad event has materialised, deposit insurance matters much less,” he said, adding that evidence beyond the 2007–08 crisis shows little proof that it shields GDP from loss.

Yet once panic spreads across the system, insurance offers little protection. 

He cautioned that safety nets can also encourage banks to rely more heavily on wholesale funding, which is far more volatile. 

“We have preliminary evidence suggesting that overall funding stability can actually worsen when banks lean on wholesale markets,” he explained.

At the same time, he added, policy intervention can soften the blow. 

“Liquidity injections and liability guarantees by central banks are the only policies shown in the data to reduce the depth of recessions after major runs. 

“Slides also showed that when regulators step in with liability guarantees, the economic fallout is consistently less severe,” he said.

Readiness, stigma and lessons from 2023  

Walters stressed that preparation is as important as policy. 

She described how the Bank of England requires banks to ‘pre-position’ collateral, a process that can take months. 

“Commercial banks must submit detailed data tapes and questionnaires on loan portfolios, undergo on-site visits, and clear extensive legal reviews before collateral is approved. 

“Once accepted, that collateral remains under the bank’s control but can be drawn down within a day in a crisis. 

“Some central banks, such as Switzerland’s, even incentivise pre-positioning by allowing it to count towards prudential liquidity requirements,” she said.

She pointed to the March 2023 US bank failures, when institutions could not access the Federal Reserve’s discount window in time because they had not practised collateral procedures. 

“SVB hadn’t tested access for a year and didn’t know how to post collateral. Signature Bank hadn’t done so for five years and even tried to pledge assets that were not eligible.

“Operational readiness is about speed. If banks and central banks pre-position collateral, they can draw liquidity within a day instead of waiting weeks or months for due diligence,” she said.

But stigma remains a barrier. 

She noted that banks often hesitate to borrow from central banks for fear of appearing weak. 

She distinguished between anticipatory stigma, which discourages banks from including central bank borrowing in contingency plans, and in-the-moment stigma, which delays actual usage in a crisis. 

“Stigma can delay action even by days, and those delays themselves can cause problems,” she said. 

She cited that Credit Suisse’s initial reluctance to borrow from the Swiss National Bank in 2023 was one example. 

“Some central banks are trying to reduce the stigma by normalising such borrowing or disguising it through collateral upgrades.”

The new triggers  

Looking ahead, Mueller warned that technology could amplify future risks. 

He said panics often start with rumours or misinformation, and the ease of generating false information today raises concerns. 

“You can get panics without any underlying reason, and now anyone can spark them,” he said.

He added that overheated credit markets remain a red flag. 

When investment-grade spreads are unusually low, the risk of runs is heightened. 

“Bank runs are not a relic of history, and they are particularly likely in an environment of overheated credit markets,” he said.

Historical US data reinforces this risk. 

He noted that banks hit by outflows during systemic runs sharply curtailed lending, triggering a wider credit crunch and amplifying recessions.

Meanwhile, Walters added that while large systemic banks are forced to prepare, medium-sized banks also need to strengthen their contingency planning. 

“It is essential that liquidity and resolution planning are integrated,” she said.

The message  

Both speakers agreed that resilience is not only about having safeguards in place but about knowing how to use them.

Chong said that failures may be rare, but the consequences are never trivial. 

“Resilience is not just about the tools we have, but about the readiness to activate them when needed,” she said.

Related News

Most Viewed Last 2 Days