Emerging Threats to Financial Stability – New Challenges for the Next Decade | Speeches (2024)

Watch video: Speech by Bradley Jones, Assistant Governor Financial Systems, AFIA Conference, Sydney 31 October

Introduction

‘May you live in interesting times’ is a sardonic curse with uncertain origin.^[1]Similarly uncertain is what the next decade might have in store for policymakers and participants in thefinancial system – what new risks and opportunities might emerge, what their impact might be andhow they might interact. In the spirit of today’s conference, I will step back from the conjuncturaloutlook to discuss some of the emerging challenges to global and domestic financial stability that couldbe with us for many years.^[2]

To spare any suspense, I have two punchlines.

First, the nature of the emerging risks we are likely to confront over the next decade have a differentcomplexion to those of recent decades (Table1). Some of these risks could originate fromwithin the financial system, such as the risk of rapid-fire deposit runs in the digital age,spillovers from entities that (individually) are not systemically important and higher interest ratevolatility. Importantly, other emerging risks are emanating from outside the system –geopolitics, operational risk and climate change – and for which there is no historical precedentto guide us.^[3]

If we are entering a decade laden with new types of risks, some of which could interact, the second mainmessage is that we need to rethink the concept of resilience. It’s entirely possible, even likely,that regulation and supervision will need to adapt as the threat environment evolves. But the ultimateresponsibility here lies with industry. There are limits to what even the most finely calibratedregulatory and supervisory regime can achieve when industry governance and risk management practices fallshort. We have a shared interest in getting our collective response right.

Inside risk #1 – Rapid bank deposit runs in the digital age

Bank deposit runs are as old as banking itself. This is an artefact of liquid deposits funding less-liquidassets, like loans. Much ink was spilled in the aftermath of the global financial crisis (GFC) to reducethis inherent vulnerability, but international events over the past year suggest there is more work todo.

Concerns over profitability and poor risk management have long been the catalyst for deposit runs. In thissense, the banking turmoil in the United States and Switzerland earlier this year was no different.^[4] Whatwas unprecedented was the speed of deposit runs in affected banks (Graph1). As a casein point, Silicon Valley Bank (SVB) lost 30percent of its deposit base in a matter of hours,with a further 50percent poised to be withdrawn the following day.^[5] In ourcurrent system, any bank would struggle to survive a run of this magnitude.^[6] It wasfar beyond the provisioning required under Basel III,^[7] and more severe than the fastest runsexperienced during the GFC.^[8]

So why was this time different? My reading is the answer lay with the interaction of new technology and adifferent set of balance sheet vulnerabilities to those prevailing in the GFC.

Herding effects associated with social media present a new challenge for financial regulators. And it isclearly easier to withdraw deposits at the stroke of a keyboard than it is to stand for hours in the rainoutside a bank branch and bury cash in the yard. But I see the role of technology here more as anamplification mechanism, not a root cause.^[9]

The share of deposits that were uninsured had been steadily rising in the US banking system, to thehighest level since 1947. For the three banks that failed in March, however, the share of uninsureddeposits was more than double the system-wide average, at 89–94percent for SVB, Silvergate and Signature Bank(Graph2).^[10] This might not have been an issue had theseuninsured deposits been widely distributed among unconnected parties. The problem was they wereconcentrated in a small number of hands that were tightly connected via social media and industrylinkages. More than half of SVB’s deposit base was sourced from companies in the technology and lifesciences sectors backed by venture capital firms that were exposed to the same macro shock (i.e. higherinterest rates lowering valuations), and key account holders were closely connected in socialnetworks.^[11] Similarly, crypto-asset clients accounted for almost90percent of the deposits of Silvergate Bank, almost all of which were uninsured and exposedto the same risk – a ‘crypto-winter’ induced by higher interest rates. The10largest depositors at Silvergate accounted for almost half of its deposit base.^[12] Andpart of the fear that spread through corporate deposit holders in the failed banks was not just relatedto the safety of their deposits, but also access to them – businesses needed to make payroll andother time-sensitive obligations.

Meanwhile in the case of Credit Suisse, key indicators of liquidity risk introduced under Basel III failedto signal impending trouble ahead of its deposit run. This dog didn’t bark, a point that is alsoprompting reflection among policymakers.^[13]

These events raise no shortage of policy issues:

• How should deposit flight assumptions be recalibrated for the digital age?
• How should interaction effects between uninsured deposits and concentrated deposit bases be accountedfor?
• Should all banks be required to mark-to-market their liquid asset buffers?
• Should deposit insurance caps be lifted when uninsured deposits are held by so few depositors, andwhat would be the moral hazard implications?
• How best can continuous access to deposits be enabled in banks experiencing severe stress?
• If banks choose to fund themselves with a high proportion of uninsured deposits, should they berequired to pre-position collateral at the central bank so they are better placed to handle runs ofany size?

It is impossible to do justice to these questions here, other than to say that they are animatingdiscussions among policymakers everywhere.

Before moving on, I should note that the Australian banking system did not experience the types ofstresses observed elsewhere (Graph3; Graph4). As the APRA Chair recently observed, wedon’t have the same level of concentration risk or uninsured deposits in the deposit bases of keyAustralian banks, and our liquidity and capital requirements are more onerous.^[14] Thatsaid, there is no sense of complacency here – our colleagues at APRA will soon be consulting onoptions for strengthening liquidity and interest rate risk management requirements.^[15]

Inside risk #2 – Systemic risks arising from entities that individually are not systemicallyimportant

Rapid-fire deposit runs are arguably a special case of a broader vulnerability where systemic risks arisefrom entities that, in the normal course, are not systemically important themselves. This means contagionrisk is taking on a new, more challenging complexion.

In asking ‘what’s different now?’, my sense is that two amplification mechanisms have beendialled up:

1. the speed of money and information flows, which can magnify herding effects
2. procyclicality in parts of the international asset management industry that have grown strongly.

First, money can now flow out of institutions and markets with unprecedented speed in response to therapid spread of information and misinformation, including that amplified through social media. In normaltimes, fewer frictions in the flow of money and information should lead to better economic outcomes– few advocate for the frictions and stale information of yesteryear. But the issue is more complexin periods of heightened stress. As noted by the Financial Stability Board in its recent review, theubiquity of social media, combined with 24/7access to banking andpayment services and more globally connected trading platforms, raises the possibility that a shocksomewhere in the financial system metastasizes into a broader self-fulfilling crisis of confidence.^[16]

In banking, concerns about rapid-fire outflows enabled by technology have been building ever since thecrisis at Continental Illinois in 1984, which was an early adopter of the systems that enabled its largecorporate clients to engage in rapid cash transfers inside and outside the United States.^[17]Since then, money has moved within and across borders with increasing velocity. And in recent years,social media has emerged as a new enabler of herding behaviour, including when depositors are in closecontact. Researchers have linked the unusual spike in social media activity not only to the rapid fall ofregional US banks in March, but to the stress experienced by other banks.^[18] InEurope, Credit Suisse was immediately caught in the cross-fire, having only narrowly survived a depositrun in October 2022 around the time a journalist tweeted that a major investment bank was ‘on thebrink’; this post had more than 6,000retweets in less than 24hours, a time when CreditSuisse was losing CHF12–15billion per day in deposits.^[19]Likewise, rumours about the imminent collapse of Deutsche Bank in March saw its credit spreads tradewider than during the peaks of the GFC and the European debt crisis, despite its revenues andprofitability having recently hit a multi-year high.^[20]

Institutions that might have little direct connection to the original source of stress, but are perceivedto have broadly similar business models, seem particularly vulnerable in an environment of ‘shootfirst, ask questions later’ – especially when the ‘shooting’ can happen faster thanever.

A different amplification mechanism relates to procyclicality in parts of the fast-growing internationalasset management industry. As they don’t accept deposits, investment funds operate with lessintensive regulation and oversight of their liquidity management and use of leverage compared with banks.Nor do they benefit from public backstops that can dampen liquidity stress, such as central bankliquidity access or deposit insurance. Herding effects can also be amplified when funds are tied tosimilar benchmarks. Where they offer redemption terms to investors that cannot be easily met in stressedmarket conditions, asset fire sales can result. The periods of market dysfunction in key bond markets inearly 2020 and 2021 were associated with this dynamic and saw a number of central banks respond. The useof leverage can be an accelerant of these ‘illiquidity spirals’. This was evident in the giltmarket meltdown of September 2022, when an initial increase in yields set off a self-reinforcing cycle ofasset sales by pension funds and required intervention by the Bank of England to restore marketfunctioning.^[21] While the make-up of the Australian financial systemmeans we are not as directly exposed to these fire sale dynamics, our funding markets have beensignificantly disrupted by international events (Graph5).^[22]

The wider spillovers amplified by small banks and investment funds in recent times raises a variety ofissues. One is whether the bar for banks to be designated as systemic should be lowered. Another is howfar regulators should go in imposing more onerous obligations on smaller banks, recognising thatproportionality is central to regulatory design and that a regime that is unduly onerous could curtailthe sort of competition and innovation that would otherwise benefit society. A more general issue is howresilience and recovery planning should better account for increased contagion risk. And in the case ofinvestment funds, an ongoing issue is how they can continue to play an important role in intermediatingsavings while minimising the liquidity risks that can severely destabilise the functioning of criticalfinancial markets. All of these issues are now under active international review.

Inside risk #3 – Structurally higher interest rate volatility

Much of the prior two decades saw long-term interest rate volatility either decline or languish atunusually subdued levels (Graph6, left panel). There are, however, a number of factors that, in theyears ahead, could contribute to a regime shift characterized by structurally higher interest ratevolatility:

• The great moderation is behind us. The global economy now appears more vulnerable tostagflationary supply shocks, including from the rewiring of globalisation, geopolitical andpolitical economy tensions, and energy shocks from climate change (and related policies). Thiscontrasts with recent decades, where developments on the supply side of the economy were typicallyfavourable for growth and inflation and so dampened financial market volatility.
• A higher and more volatile bond term premium.^[23] For the better part ofthree decades, the term premium on bonds declined relentlessly (Graph6, right panel). Thisreflected factors that are unlikely to continue, including declining uncertainty over futureinflation outcomes and real policy rates, coupled with structural supply–demand imbalances in bonds that meant governments were able toissue debt with little or even negative term premia.
• Reduced smoothing from price-insensitive buyers. To varying degrees over the past twodecades, volatility in key bond markets has been supressed by the bid from foreign exchange (FX)reserve managers and domestic asset purchase programs by central banks. Again, this era now seemsbehind us. In recent years, reserve managers have become net sellers of US Treasuries and, on currentplans at least, central banks will be reducing their domestic bond holdings for years to come. Thisis occurring against a backdrop where broker-dealers have reduced ability and willingness to bid forbonds during periods of heightened volatility.

A period of structurally higher interest rate volatility would present challenges to financial systemstability, a few of which I will mention here:

• Bank duration risk. Banks of all sizes will have to manage overall duration risk moreintently. Inadequate management of interest rate risk was an important contributor to the US bankingstresses in March (Graph7). In Australia, banks tend to have modest exposure to securitiescompared with their peers, large banks are required to hold capital for interest rate risk in thebanking book, and residual interest rate risk tends to be hedged (Graph8). Nevertheless, thisis an area of ongoing supervisory focus in many jurisdictions.
• The interaction of interest rate and credit risk. Over much of the past few decades,growth in incomes consistently exceeded the level of interest rates (‘g> r’). Thisdampened credit risks for lenders. But debt servicing among borrowers with high debt levels will bemore challenging – and credit risks will increase as a result – in an environment whereinterest rates periodically exceed the growth in incomes.
• Financial market functioning. The post-GFC era of exceptionally low interest ratevolatility gave rise to a new generation of investment business models and strategies. This includesstrategies that: (i) involve the procyclical leveraging up of bond positions when volatilitydips;^[24] and (ii) assume a negative correlation betweenbond and equity returns (as tends to occur in low and stable inflation regimes). It remains to beseen how market functioning could be affected in a regime of higher inflation and interest ratevolatility, including where open-end fixed-income funds encounter liquidity stress, margining is moretightly enforced than in the past,^[25] and broker-dealers are unable or unwilling tosupply liquidity.

Outside risk #1 – Geopolitical risk

War may well be ‘the father of all things’, as the philosopher Heracl*tus once observed. Since1945, however, the absence of direct conflict between economically integrated powers meant twogenerations of economic policymakers and industry participants could turn their attention elsewhere. Butrecent events suggest this is no longer tenable. For a start, geoeconomic fragmentation will render theglobal economy more prone to supply shocks. The International Monetary Fund (IMF) has estimated it couldresult in global output losses of up to 7percent, rising to 12percent for somecountries.^[26] And as a more strained geopolitical environmentcould also have material implications for financial stability – not limited to the cross-borderallocation of capital, international payment systems, key financial institutions, and global funding andcommodity markets – a considerable body of work is now emerging internationally to determine wherefinancial resilience needs to be bolstered.^[27]

As an organising framework, I find it useful to decipher the related risks to financial stability alongtwo dimensions: slow-burn fragmentation, and fast-burn kinetic risk.

Slow-burn fragmentation is arguably no longer a risk – it is already underway. Trade and financialsanctions (including counter-sanctions) were increasing even prior to the Russian invasion of Ukraine(Graph9). ‘Just in time’ efficiency is being traded off for strategic resilience insupply chains. Alternative payment messaging systems are being developed, and FX reserves are beingshifted to countries and assets deemed less vulnerable to seizure.^[28] A smaller share of foreigndirect investment, portfolio investment and bank credit is flowing between countries that are lessaligned on foreign policy issues (Graph10). It is now common to hear some countries characterisedas ‘uninvestable’ in a way that was not the case just a few years ago. And fraying support forinternational cooperation risks hampering operation of the global financial safety net.^[29]

The second (and even more troubling) dimension is fast-burn kinetic risk – the risk that tensionsescalate into conflict among countries critical to the global economy. Global financial stability risksincrease along a continuum in this respect, depending on the scale, location and countries involved.Events over the past two years give a sense as to the underlying stress transmission channels. Globalshipping could become prohibitively costly or uninsurable along certain routes. Global financialconditions could tighten abruptly, with key funding markets, risk asset prices and cross-border lendingseverely affected. Cross-border assets could be frozen or seized. Payment systems and financial marketinfrastructures could be affected by sanctions and counter-sanctions. Cyber-attacks on key institutionscould become more prominent. Disruptions to global commodity markets would be most significant whereglobal production is concentrated in a small number of countries and where key trade routes areaffected.^[30] More generally, the longer any conflagrationpersisted, the stronger the feedback loop between real and financial channels, including deterioratingasset quality for banks.

The policy implications and trade-offs associated with heightened geopolitical risk are daunting andcomplex, so I will simply note here the importance of building financial and operational resilience alongvarious dimensions – at both the domestic and international level. At the international level, onekey concern is to prevent the global financial safety net from splintering into smaller liquidity poolsbased on geopolitical blocs. As the IMF has noted, a danger here is that the substantial benefits thatcome with international risk sharing are compromised at a time when the call on international resourcescould increase substantially.^[31]

Outside risk #2 – Operational risk

The increasing digitalisation of financial services has been a key source of innovation. But it also meansoperational risk is now a first order enterprise-wide risk that demands the highest standards ofmanagement and governance. As some operational risks could also have systemic implications, this isemerging as a key regulatory priority the world over.^[32]

I will focus here on three risks, the most immediate and challenging of which is cyber. The scope for, andconsequences of, cyber-attacks continue to increase. Challenges are building from third-party technologydependencies, geopolitical tensions and resourcing constraints among smaller institutions that could bekey points of vulnerability for the wider system. Attacks can result in direct financial losses (fromtheft and ransoms) and indirect losses arising from reputational damage. The past year alone has bornethis out in Australia and elsewhere.

A multi-faceted policy response is rolling out in response. In Australia, the Government established therole of National Cyber Security Coordinator in May 2023 to coordinate and strengthen cybersecuritypolicy, preparedness and response capability. APRA has recently conducted a cybersecurity stocktake andfinalised a new broader operational risk standard for its regulated entities, which will incorporatestronger requirements for maintaining and testing internal controls, improved business continuityplanning and enhancing institutions’ oversight of external service providers.^[33]Regulators are also signifying a clear intent to undertake enforcement actions where firms fail to meetexpected standards of conduct. And the Cyber Operational Resilience Intelligence-led Exercises Framework,developed by the Council of Financial Regulators (CFR) and led by the Reserve Bank, continues to test theresilience of key financial institutions to cyber-attacks. All that said, this is an area where no amountof supervisory focus can substitute for robust planning and preparedness by financial institutions.

A second operational risk relates to critical technology service providers that have been largelyoperating outside the financial regulatory perimeter. Banks and financial market infrastructures areconsidering moving some services to the cloud, and for sound reasons. But exposure is concentrated in a select number of service providersand an outage could leave many institutions unable to perform critical functions. Concernsabout governance, risk management, platform concentration risk and responsiveness in a crisis have alsobeen raised in respect to the use of emerging technologies by financial institutions, such as distributedledgers. And in some jurisdictions, Australia included, BigTech has sought to move into payment services(including through digital wallets) without the regulatory oversight typically imposed on other financialservice providers.^[34]

A third operational risk – artificial intelligence (AI) – seems more distant but no lesschallenging. AI offers enormous opportunities for the financial system and wider economy, but managingthe associated risks will not be straightforward. As Gary Gensler (Chair of the US Securities ExchangeCommission) has noted for some time, the financial stability risks relate to the potential for shocks inthe financial system to be magnified.^[35] Contagion and herding risk is one potential channel,where parties could come to rely on similar models and data aggregators that are trained on similar dataand so generate similar actions – including in a crisis. And because the underlying models arecomplex and typically developed outside of the financial regulatory perimeter, system-widevulnerabilities could be brewing in a way that is not obvious to financial institutions or supervisors.

Outside risk #3 – Climate change

Risks to financial stability resulting from climate change and related policies will be with us for yearsto come. Physical risks from more extreme weather patterns and higher average temperatures can reduce thevalue of assets and income streams in sectors and parts of the country. Transition risks, includingunexpected changes to regulations and consumer preferences, can also lower the value of assets orbusinesses in emissions-intensive industries. These risks can result in unexpected losses for lenders,increased claims on insurers and write-downs for investors – as we are now observing.Internationally, direct exposures to emissions-intensive businesses are largest for investment funds,followed by insurers, and then banks.

As the Bank has previously set out the financial stability implications in Australia from climate change,I will briefly recap our main takeaways to date.^[36] The Bank and APRA have applied complementaryapproaches to examining the possible effects of physical and transition risks for the Australian bankingsystem.^[37] This preliminary analysis has found that banks wouldexperience losses under different scenarios but not of a magnitude that would cause significant stress,as lending losses appear concentrated in specific regions and industries that represent a smallproportion of banks’ overall assets. Nevertheless, these early results need to be treated withcaution as our understanding of climate-related risks to the financial system is still developing; lossescould be larger and occur faster than currently anticipated. More advanced modelling techniques and moregranular data are needed to produce more robust estimates of possible credit losses.

A particular domestic focus in the period ahead will be the household insurance sector. Payouts have beenrising in real terms for some time, and more frequent and severe weather events are expected to furtherincrease claims on damaged property and other assets. Insurers have the ability to reset insurancepremiums quickly to recoup climate-related losses, or they could withdraw coverage from high-riskregions. This would pass risk back to households, businesses and governments, or to the lenders in thecase of loan defaults where affected assets are used as collateral. To better understand the risks to thefinancial system and broader community that could occur due to changes in the cost and coverage ofinsurance, APRA, on behalf of the CFR, will soon conduct a climate scenario analysis with insurers. Morebroadly, the Bank, in conjunction with other CFR agencies, is coordinating on a set of priorities tosupport industry participants to manage their climate-related risks and opportunities.^[38]

Concluding remarks

There are two common threads to the themes I have discussed today.

The first is that there are a number of emerging risks to financial stability that look quite different tothose of recent decades and that will likely be with us for some time. Some have no historical precedent,and could interact, and so will require particularly careful attention. Governance and risk managementpractices that have a strong forward-looking orientation will have a central role here – think morescenario analysis and reverse stress testing, and less historical backtesting.

The second is that building resilience along various dimensions will be critical, recognising of coursethat there are always trade-offs to be managed. Strengthened resilience was the north star for thereforms for large global and domestic banks following the GFC. But for the financial system at large,there is more to do here. We can’t know which risks will be realised or when. But we can be betterprepared for whatever might come our way.