ALM banking after the crisis: stress-testing for more robust liquidity management practices

A panel of industry experts discusses a new age of depositor behaviour and the expected evolution of regulations in the wake of the asset-liability management (ALM) banking crisis. They share insights on achieving integrated approaches to ALM, as well as dynamic hedging strategies in the current high interest rate environment and best practice in accounting classification of securities in light of Silicon Valley Bank’s (SVB’s) collapse

What are the key learnings for risk managers around depositor behaviour, and what will be the impact on models?

Christopher Jones, Moody’s Analytics: Years of low interest rates have resulted in a sense of complacency among financial institutions and risk managers regarding deposit balance behaviour. Systemically, deposit balances grew during the past two rising rate cycles. This was further complicated by unusual systemic deposit inflows during 2020–21 due to Covid-19 pandemic-related government fiscal stimulus. Many financial institutions believed that deposit balances would remain stable as rates began to rise in 2022. However, the spectacular bank failures in the US that occurred earlier this year clearly demonstrated that depositor behaviour in this cycle is different from what we have become accustomed to in the past two decades.

Consequently, deposit modelling assumptions will need to change to reflect shifts in balance behaviour. Shorter-weighted average lives should probably be built into models for non-maturity deposits. Additionally, behavioural models will need to be linked to macroeconomic variables to produce more robust behavioural assumptions. Finally, all relevant deposit modelling assumptions should be regularly stress-tested to provide a better understanding of their impact on the overall interest rate risk profile of an organisation.

Adrian Docherty, BNP Paribas: It is easy to forget that 99.9% of banks have had remarkably stable deposit bases this year. The stressful shenanigans in California and Zurich were isolated – idiosyncratic. Nonetheless, risk managers will have noted the unprecedented velocity and ferocity of deposit withdrawals from failing banks. Crises can unfold in a matter of hours. Banks need to be superbly prepared for crisis management. Far more importantly, they need to anchor their strategies in stable, resilient and sustainable business models.

Dimitrios Papathanasiou, Credit Suisse: One of the most important learnings is around concentration risk. SVB had a lot of concentration on its deposit base. Until recently, the market saw deposits as one of the safest products in the liability structure of a bank, but the recent crisis revealed that some types of depositors are much more sensitive than others. Large concentrations in these types of clients create a higher risk of deposit flight, especially if these clients talk to each other – as was the case with SVB. It is therefore more important than ever to categorise the various types of depositors and try to understand better how they might behave in a crisis.

Banks should also reconsider how to diversify their funding bases. Traditionally, banks use deposits as their main funding source and, on some occasions, they rely on them completely. But the deposits are usually very short term and can be withdrawn in large quantities – as we saw with the collapse of SVB. So banks must look not only at the product mix, but at the duration or tenor mix of their liabilities’ structures to ensure they do not have too much concentration in the very short term.

Jacek Rzeznik, mBank: In this period, statistical modelling has not been a good predictor of stability as the historical data was not adequate to feed the models, which needed overriding. Banks must approach the determination of core balances prudently, taking into account both a backward-modelled perspective and a forward business and risk perspective to build up and maintain adequate buffers. This also encompasses making assumptions on migration when rates shift, to set the non-core balances.

Furthermore, it is important the asset-liability committee carefully manages its price strategy to respond to the competitive landscape and balance the stability of volumes and margins. As part of that, the assumptions on pass-through rates need to be frequently reviewed to ensure they are suited to the current environment.

Will changes in depositor behaviour be permanent in a new digital age or will a new generation of risk managers need to learn to manage rate-tightening cycles?

Christopher Jones: In today’s environment, depositors can move money with a tap on a mobile phone screen. Gone are the days of relationship banking (especially at the consumer level). Therefore, risk managers must be prepared to assume that deposits will not stick around for any of the traditional reasons they may have in the past.

Furthermore, many of the current generation of risk managers have never experienced a rate cycle during their careers. They must be willing to shift their modelling approaches to account for changing depositor behaviour. Note that, since deposit models represent the single largest assumption in the interest rate risk framework, changing these assumptions could potentially have a very large impact on the economic value and funds transfer pricing (FTP) risk profile results. Again, assumption stress-testing is critical to the model governance process to understand the potential implications of being wrong.

Adrian Docherty: Conventional wisdom suggests that ‘digital’ changes everything, lubricating the financial system and reducing the stickiness of customer deposits. Indeed, deposit flight from failing banks has accelerated – that is no surprise. And yes, the ‘hot money’ from rate-chasers is of ever more dubious quality as a source of funding for a bank’s lending activities. But, post-digital, we haven’t really seen any significant change in mainstream depositor behaviour, despite record-low pass-through of rate increases to depositors. I wonder how much the friction of branch banking contributed to franchise stability versus other factors such as institutional stability and trust. Maybe we will see digitally enabled shifts in customer loyalty – but it is certainly not yet an empirical fact. You never know, it could even be the case that customers actually increase loyalty and stickiness towards trusted providers because the digital ecosystem gives rise to new concerns such as scams and hacks.

Dimitrios Papathanasiou: Changes in depositor behaviour are permanent and have brought in a new era in treasury and liquidity risk. If banks want to act defensively, they should reflect this in their tenor structures. They should also reconsider how they price deposits. Since depositors can move their funds in a matter of seconds, banks need to be careful how they price short-term liabilities. They should also consider how to incentivise the right mix of clients to provide a more stable liability structure.

This tightening cycle in rates behaviour is the most aggressive we have seen in the past 40 years. However, there have always been loosening and tightening cycles – I put it down to bad interest rate risk management by SVB, which had an impact on capital and a knock-on effect on liquidity.

Jacek Rzeznik: This is the first time in this new digital age we have experienced such an abrupt tightening cycle. Risk managers have more data and are equipped with better tools to analyse it, including artificial intelligence (AI) applications. So they should collect and analyse data from this period and learn from it to verify assumptions, and build and calibrate models.

The recent collapse of US medium-sized banks and that of Credit Suisse provide lessons on how concentration risks can materialise at a time when information and fake news can spread rapidly across social media and trigger sudden mobile bank runs. In response, banks need to focus on managing reputational risk, engaging and managing key stakeholders, developing social media monitoring, and building scenarios and response strategies. These contingency plans need to be prompt and nimble so banks can react swiftly to rapidly developing situations and launch timely countermeasures.

How will the regulations evolve and how will risk managers adapt?

Christopher Jones: Historically, the regulatory focus on liquidity tends to be much higher internationally – particularly in the UK and Europe. Given the bank failures that occurred earlier this year, the expectation is that US regulators will increase their expectations for enhanced analysis and reporting around liquidity for banks of all sizes – but particularly for regional and smaller banks that might be at risk during liquidity events that could have an adverse impact on their solvency. This may include higher expectations around liquidity stress-testing and enhanced monitoring of daily liquidity (FR 2052a report) and/or liquidity coverage ratio (LCR).

Risk managers will need to adapt by staying up to date on liquidity regulations – even ones that may currently apply only to larger institutions. They should also be willing to communicate proactively with regulators to gauge what the expectations of them might be. Finally, risk managers will need to be prepared to adapt processes to accommodate more robust liquidity management practices. This may involve developing additional modelling capabilities.

Adrian Docherty: The obvious start is for the US to adopt Basel Committee on Banking Supervision regulations on interest rate risk in the banking book (IRRBB), liquidity and funding – that’s a no-brainer. Many US banks will need to upgrade their ALM capabilities. Regulators worldwide are making rumbling noises about revisiting regulations and calibrations, which would be a mistake. The ALM problems in California do not reflect imperfections elsewhere. Recalibrating LCR and the net stable funding ratio (NSFR) would needlessly reduce the maturity transformation of the banking system and increase the cost of capital to the real economy. The Australian way of introducing a Pillar 1 capital requirement for IRRBB, which is much admired by some international regulators, would be conceptually wrong: why would you capitalise the variability risk in franchise value when it is not solvency risk? It just doesn’t make sense.

Disappointingly, the events in California have galvanised some in the regulatory community and changes to regulations are probably inevitable. This will increase the incentive to invest in superior ALM capabilities and structures. I expect to see changes in product features, higher levels of wholesale term funding, more hedging and increased use of options to cover uncertainties.

Dimitrios Papathanasiou: Regulators will need to investigate the credit default swaps (CDS) market further. During the crisis, just after Credit Suisse’s collapse, there was significant fluctuation across all markets following Deutsche Bank’s CDS widening. An investigation found it was a single CDS trade of only €5 million that created this panic, which reveals a weakness in the market.

In the early 2000s, when there were very large volumes in structured credit, the CDS market was very liquid, but it seems this is no longer the case. The credit spread of a bank is far more important relative to a corporate. A big widening in a bank’s CDS shows a lack of trust. Trust is everything in the banking world. All stakeholders, including counterparties and depositors, follow the credit spread of a bank closely to transact with it.

Given the Deutsche Bank CDS episode, it seems easy for a short-seller to buy CDS protection to breach the trust of the stakeholders and then sell their stock price, which can create systemic risks. Regulators should be very concerned about whether the CDS market reflects the actual credit risk of a bank, and whether this could be exploited by speculators and cause systemic risks.

In terms of concentration risk, regulators generally expect large, systemic banks to have a framework for this. But they have never been clear on exactly what they expect and put this into regulations or defined the criteria. Sometimes it is quite appealing for banks to have large concentrations and target large corporates since they benefit from the relationship overall, but they can run into big problems if these clients decide to go.

Furthermore, I expect more scrutiny on smaller US banks now some of their fundamental weaknesses have been revealed.

Regarding modelling, the LCR does not cover some key areas, such as initial margin, prime business or intraday. But these weaknesses did not cause the issues in this crisis. There could be some refinement in terms of how the LCR classifies some types of depositors. But, overall, the LCR and the NSFR are doing their job.

Jacek Rzeznik: As evidenced by the recent collapses, no regulation can be a substitute for a good risk culture, where the entire firm understands that proper risk management is essential for its survival.

We are seeing a further increase in regulatory requirements for more frequent and more granular reporting. Furthermore, we can expect a revision of stress scenarios. Most scenarios were credit risk focused until now. Assuming there will be a recession followed by cuts in interest rates and negative impact on net interest income (NII) – which requires income stabilisation hedging strategies – there will be more focus on the impact of rates increases on economic and accounting values, including higher scrutiny of accounting classifications.

The sudden liquidity crisis also highlights that managing liquidity with a 30-day horizon of LCR is not sufficient and there should be increased focus on intraday liquidity management.

The European Banking Authority's (EBA’s) Supervisory Outlier Test may give false signals that the high margins banks have had on their unstable or non-core balances are transitory side effects of the ultra-loose environment and over-liquidity. In downshift scenarios, due to floor effects, those assumptions would show lost income and might encourage banks to stabilise this income with a fixed receiver position. However, when rates go up, those high margins may not be sustainable and banks may be caught out as they need to replace those balances with new money at prevailing market rates (in effect, at zero margin), while their income has been locked at low levels for a long time. This is essentially the trap SVB fell into, and its collapse is a lesson that regulations should avoid pushing banks to stabilise NII on unstable funds.

What are the wider implications of the rate hikes on balance sheet management – and how can banks achieve an integrated approach to ALM?

Christopher Jones: Due to the historically low-rate environment financial institutions have operated in for the past 15 years, the focus on balance sheet management has been on NII. Most financial institutions have purposely positioned themselves to benefit from rising rates, which has resulted in a bit of tunnel vision when it comes to risk measures. However, rising rates have revealed market value exposures, especially for organisations that have utilised excess deposit balances to increase holdings of fixed income securities in search of higher margins. This has once again demonstrated the importance of valuation analysis.

Breaking down the components of market value of equity, it represents the present value of both the principal and interest cashflows of assets and liabilities. The present value of principal cashflows represents the carrying value of equity of the organisation (not the stock price). The present value of the interest cashflows represents the NII, or future earnings potential of the organisation.

Using NII and market value risk profile metrics provides a more complete long-term view of interest rate risk. In addition to these fundamental components, financial institutions should also utilise dynamic balance sheet forecasting – optimally in conjunction with a well-designed FTP programme, as well as key rate duration – to achieve a truly integrated approach to ALM.

Adrian Docherty: Higher rates are a kind of ‘back to the future’, but the real shift is the degree of uncertainty on the path of rates. Today’s rates environment challenges our assumptions of customer behaviour. Customers have a lot of options: they can accept or decline an offer of finance at an agreed rate, giving rise to pipeline risk. They can choose between prepaying, rolling onto a new rate or refinancing, which gives rise to uncertainties around what rate they will pay and for how long.

Banks generally have good information systems to monitor and model changes in customer behaviour, but their ALM strategies are too often focused around a single, base-case view of the future. Recent experience is a reminder that less likely scenarios can often play out. A good ALM approach seeks to deal with this uncertainty. I think banks will reduce their reluctance to spend money on option-based strategies to reduce ALM risk.

Dimitrios Papathanasiou: I think we are coming close to the end of the rate hiking cycle. Banks should therefore start asking forward-looking questions about staying stable at these levels or loosening.

In a rate hiking cycle, depositors gradually start moving from floating to fixed and looking for a better rate, which can have implications both on liquidity risk and on overall ALM. For example, if banks have replicated portfolios and they are hedging a specific percentage of their non-maturing deposits, they might be forced to unwind trades in a non-favourable environment.

Jacek Rzeznik: Recent history has proved that dealing with ALM in silos is likely to lead to bad outcomes. In a more complex world, a wider and more integrated perspective needs to be supported by an appropriate organisational setup and empowerment.

Recent experiences have proved that, when running ALM strategies of hedging NII on modelled deposits, it is key to manage not only the economic value perspective but also the accounting view and its potential impact on capital. It is advisable to stress-test it and see the interaction of IRRBB with capital and liquidity. Firms should consider liquidity and interest rate perspectives when modelling deposits.

ALM strategies should be reviewed frequently – as part of organisations’ discussions around risk appetite – and adjusted to changing environments. Clearly, income stabilisation is not always the most prudent strategy and, if mismanaged, may lead to serious problems. Hence, both income and economic value of equity (Eve) perspectives need to be considered.

Lately, we are also faced with new challenges, as we see record-high inversion of the yield curve. In normal environments, income stabilisation strategies are neutral or bring extra yield. However, currently, these come with negative carry, which brings the strategy into doubt.

With interest rates still potentially rising, large banks may bet that the liquidity crisis is over, classifying their securities as held-to-maturity. Is this the right bet or should they designate bonds as available-for-sale, marking-to-market through other comprehensive income?

Christopher Jones: The accounting treatment applied to an investment portfolio can have massive implications for both income and capital. There is no universal answer to the question of which accounting treatment is correct. Each organisation is different – with a different balance sheet composition, different investment portfolio size and concentrations, different capital and liquidity needs, and so on. In a rising rate environment, financial institutions may consider maintaining a mix of held-to-maturity and available-for-sale to maintain a source of liquidity should deposit balances run off unexpectedly. This is a decision that should be thoroughly analysed, discussed and stress-tested before being undertaken.

Adrian Docherty: It shouldn’t be a bet – banks should follow the accounting rules and book their assets accordingly. A buy-and-hold security funded by stable funding should have a matched interest rate position and a stable value, so it is wrong to look at the fair-value change of the asset alone. Imbalances should show up in the IRRBB stress test, where the accounting treatment does not change the Eve metric. It is a bit easier for International Financial Reporting Standards banks – which can hedge hold-to-collect securities under hedge accounting with no accounting noise – than for US banks, for which US Generally Accepted Accounting Principles can generate lots of profit-and-loss (P&L) noise for hedged securities.

Dimitrios Papathanasiou: I expect we are close to the end of this hiking cycle since inflation has a downward trend in some key economies. In general, I don’t see a problem with banks having some securities as held-to-maturity. Problems arise when they classify too many securities as held-to-maturity as a liquidity buffer because, as with SVB, they might have to suddenly unwind some of these positions. Banks must be aware of this risk.

I am not criticising the held-to-maturity approach. Banks might have a lot of depositors coming in and a lot of liquidity to classify. Therefore they might buy securities that offer a much higher yield than they pay depositors and still make a margin, even if market yields move higher later. On a mark-to-market basis, they may be making losses. But it is still okay as long as they have a healthy margin on what they pay to the depositors.

Problems would arise if they had to sell the bonds when yields were higher and record losses. But, if banks address this risk, they should choose the accounting treatment of the securities that is best for them.

Jacek Rzeznik: Banks need to decide carefully on the accounting classification of their securities portfolios as part of their risk appetite discussions. Accounting treatment gives benefits – for example, not marking-to-market reduces P&L and capital volatility – but also puts some constraints on their disposal. Here again, both the liquidity perspective and IRRBB need to be reviewed. If securities are booked as held-to-maturity, banks should ensure they look at the liquidity constraints from a liquidity management perspective and include that in their scenarios (for example, in-depth and access to repo markets).

In my view, in the risk management framework, the economic perspective should prevail, and the risks from accounting effects on capital should be understood and managed. Despite accounting of those portfolios being at amortised costs, losses in economic value of those securities reflect actual economic loss in the form of opportunity cost – in effect, lower NII, which also has an impact on the value of high-quality liquid assets. In the case of SVB’s liquidity crisis, those losses needed to be crystallised, and what followed was the scrutiny of those portfolios by the market. The EBA has recently published an ad hoc analysis of unrealised losses on EU banks’ bond holdings and the value of the portfolios of banks in their sample. As of February 2023, it showed €73 billion in unrealised losses. Clearly, the duration of those portfolios plays a key role here, and a key element of the risk appetite is the duration of modelled liabilities that drives decisions in those portfolios.

What impact are rate rises having on hedging strategies? What types of dynamic hedging are banks engaging in and what are the more successful instruments they are using?

Christopher Jones: Because rates were low for so long, many financial institutions purposely positioned themselves to be asset-sensitive from an NII perspective. As rates rose, these institutions benefited from higher rates. However, many banks also went out on the curve (for example, purchasing long-term investment securities) in search of yield when rates were low, and the valuation of long-term securities and loans would have been adversely impacted by rising rates resulting in increased valuation and, in some cases, liquidity risk.

Organisations that purchased protection in the form of pay-fix swaps when rates were relatively low likely benefited from rising rates. If properly executed, those hedges would offset valuation decreases on fixed-rate securities as rates increased. Buying hedges is sometimes viewed as a gamble but, when done judiciously, can provide a decent amount of protection against rate changes.

As we reach the peak of the rate cycle, financial institutions may be looking for ways to protect their income from falling rates. This can be done synthetically by transforming variable cashflows to fixed using received fixed swaps. Again, timing is important to minimise the cost of this type of protection, and extensive analysis should be undertaken to determine the optimal mix of notional and terms.

Adrian Docherty: We have seen a rapid change in the level and volatility of rates, so it is no surprise hedging strategies need a makeover. Banks’ dynamic structural hedges exist primarily to monetise the behavioural duration of non-maturity deposits and smooth income. When hedged items outperform, as is the case at present, banks receive low rates from these hedges. They are repricing higher but slowly, dampening NII growth. Deposit duration has shortened and asset-liability managers have shortened the hedge duration accordingly, resulting in a mark-to-market loss. We saw billions of dollars’ worth of losses through fair value through other comprehensive income in 2022 – in some cases decimating the net income reported in P&L. Some banks have reduced the size of their programmes because of increased uncertainty about deposit stickiness. It could be regrettable or even painful if rates were to move lower in the coming years.

We have seen some banks addressing the issue by using swaptions. These allow banks to retain their existing hedges: the swaps cancel out if rates go higher. And, if rates decrease, the bank benefits from a larger swap notional.

On the asset side, banks have used swaptions to address prepayment risks in fixed-rate mortgage portfolios, especially for new production at higher rates, which has a greater risk of being remortgaged. Swaptions mitigate uncertainty in customer behaviour and, if realised rates volatility remains high, it can lead to savings against the rebalancing costs associated with vanilla hedges.

Dimitrios Papathanasiou: Depositor behaviour can force banks to unwind swaps at a loss. Banks need to be aware of this risk if they follow a swaps-only strategy. It would be good to have the flexibility of using some options.

Managers that follow a more passive strategy should take this into consideration in periods where rates are extreme – for example, during the pandemic when the 10-year US Treasury had a yield of less than 1%.

Jacek Rzeznik: In Poland the mortgage market has historically been dominated by floating-rate loans, meaning banks have traditionally been exposed to higher NII volatility. To manage it, banks have been employing hedging strategies with receiver interest rate swaps and fixed-rate bonds to stabilise NII. In using those instruments, they have considered various accounting approaches to minimise the impact of valuation on their capital.

Natural hedges in the form of fixed-rate loans have only recently started to pick up, having attracted attention following the rates increases and encouragement from regulators. However, the big challenge to the development of this product in Poland is the lack of appropriate regulations around the ability to charge prepayment fees. Without this ability, this product is very risky for banks as clients can essentially prepay at any point without any cost.

What ALM risks are likely to emerge in the coming years that banks should consider?

Adrian Docherty: The interest rate curve is not steep – but it is volatile. Banks need to consider how they might fare under a further sharp rise in rates but also a precipitous decline in rates. Economists and markets sometimes indicate that the risks are to the downside – in effect, a collapse in rates. That risk might mean locking in high rates on the asset side. Or banks might choose to hedge upside and downside simultaneously with options. Diversification is, as ever, the most efficient risk management tool to boost resilience without impacting earnings. In times of uncertainty, it is good to have lots of tools in the ALM toolbox.

Are stress tests working as a risk management tool?

Adrian Docherty: Yes – internal stress-testing is an essential tool. Good stress-testing capability helps risk managers identify vulnerabilities and consider remedies. Public, single-scenario tests, on the other hand, are gaining regulatory prominence but are of limited use. Other than keeping banks on their toes, I don’t think they’re of any use for risk management – they’re mostly just used to justify capital add-ons.

Sponsored content