Updating Your Liquidity Risk Management Framework: Lessons from the 2023 Bank Failures

Liquidity risk management sits at the heart of a stable bank, and the events of 2023 put it firmly back on the agenda. On 28 May, UBA UK hosted the first of a three-part treasury management series, with Professor Moorad Choudhry leading a session on updating the liquidity risk management framework in light of those failures. With over 120 senior treasury and business leaders in the session, the discussion offered a clear-eyed look at what the collapses should teach the rest of us about deposit concentration, asset-liability management and the role of Basel III.

A familiar problem in a new disguise

The session opened with the collapses of Silicon Valley Bank (SVB), Signature Bank and First Republic. It would be comforting to file these away as something exotic, but Professor Choudhry made the opposite case: the underlying fault was not a new species of risk at all. The same dynamics drove the savings and loan crisis of the 1980s. What changed was the speed at which the failure played out.

SVB is the instructive example. Over two to three years, exceptional deposit growth doubled its balance sheet. Around 85 per cent of those deposits were instant-access operational balances from large corporates and non-bank financial institutions, while 60 to 65 per cent of assets sat in long-dated US treasuries and agency bonds. In practice, the bank had started to behave more like a bond fund than a traditional lender. The trigger was not the Federal Reserve raising rates; it was the concentration baked into how the bank funded itself.

Interest rate risk, left unhedged

SVB held long-dated, fixed-coupon bonds and did not hedge the bulk of them. US accounting rules played a part here: securities classified as Hold to Maturity cannot be hedged, so only the smaller Fair Value through Other Comprehensive Income portfolio carried any protection.

The warning signs were visible internally. The bank breached its Economic Value of Equity limits repeatedly across 2021 and 2022. Rather than acting on those breaches, it restored the limits by stretching the behavioural tenor assumptions applied to non-maturing deposits, an accounting adjustment that flattered the numbers without addressing the exposure beneath them.

When a bank run takes hours, not weeks

The most striking feature of the SVB failure was its pace. Roughly $85 billion left the bank in a very short window. Mobile and online banking meant depositors could move funds instantly, and once commentary about unrealised mark-to-market losses began to circulate, confidence evaporated. A concentrated funding base, lacking the granularity of a broad retail deposit franchise, meant there was little to slow the exit. Speed of outflow, Professor Choudhry argued, is itself a risk factor that frameworks now need to take seriously.

Looking beyond the LCR

A recurring theme was the limit of any single metric. The Liquidity Coverage Ratio (LCR) guards against Pillar One risks, but it was never designed to capture everything, and nor was the Net Stable Funding Ratio that sits alongside it. The session set out the Pillar Two risks that deserve closer attention, with deposit concentration sitting at the top of the list, whether by depositor type, contractual maturity or product. Alongside it came franchise viability, intraday liquidity risk, cash flow mismatch and liquid asset management risk. This is where genuine liquidity stress testing earns its keep.

That said, the existing toolkit is not broken. SVB’s LCR sat well below 100 per cent for a full year before it failed, which suggests the ratio works rather well as an early warning indicator when people pay attention to it. The Basel III framework, on this reading, remains fit for purpose.

Practical recommendations

Several suggestions emerged for strengthening a framework, rather than rebuilding it from scratch.

On metrics, Professor Choudhry proposed extending the LCR into a 90-day Stress Liquidity Ratio, complemented by a seven-day survival metric to capture immediate-term stress.

On governance, the contrast with SVB was pointed. Its treasury function reportedly sat four levels below the board. A healthier structure puts the CEO, CFO and CRO at the ALCO table as mandatory attendees, alongside business line heads, the treasurer and a representative from the general counsel’s office. Strong governance and oversight, monthly meetings, a culture that welcomes open debate, and a direct escalation route to the board risk committee all matter more than any single ratio. A credible contingency funding plan should sit behind all of it.

On funding, the advice centred on diversification and self-sufficiency. That means reducing reliance on US treasuries where practical and considering alternatives such as deposits at globally systemic banks, treasury-backed money market funds, and term deposits with early-breakage restrictions. Gold and silver were mentioned as options too, with the obvious caveat about price volatility. Notice deposit accounts can help extend the contractual tenor of funding, and each entity should be able to stand on its own.

What comes next

This was Part 1 of three. Part 2, on Interest Rate Risk in the Banking Book, follows on 25 June, and Part 3 turns to AI applications for bank balance sheet management in July.

The closing thought was a reassuring one. The 2023 failures were not evidence that the rules had stopped working. They were a reminder that frameworks only protect you if the governance, the culture and the attention behind them are real.

Updating Your Liquidity Risk Management Framework: Lessons from the 2023 Bank Failures