In this short blog, we highlight three sets of questions for policymakers on the Bank’s Financial Policy Committee ahead of their July FSR.
Our questions are focused in three separate areas:
• Implications of the rise in interest rates for deleveraging and default risk;
• Lessons from the March banking turmoil for the effectiveness of bank resolution frameworks and implications for bank capital requirements;
• The recently announced System-Wide Exploratory Scenario.
We discuss each in turn.

Author: Professor David Aikman

Published: 11 July 2023

In this short blog, I highlight three sets of questions for policymakers on the Bank’s Financial Policy Committee ahead of their July FSR.

My questions are focused in three separate areas:

  • Implications of the rise in interest rates for deleveraging and default risk;
  • Lessons from the March banking turmoil for the effectiveness of bank resolution frameworks and implications for bank capital requirements;
  • The recently announced System-Wide Exploratory Scenario.

I discuss each in turn.

The Bank of England and Royal Exchange, London by dynasoar, Getty Images

The Bank of England and Royal Exchange, London by dynasoar, Getty Images

Implications of the rise in interest rates for deleveraging and default risk

The steep rise in mortgage interest rates and its implications for the resilience of household finances and the wider financial system are clearly the central questions in the macroprudential policy arena at this juncture.  The FPC will no doubt have spent a good deal of time exploring this issue over recent weeks.  The following questions seem especially salient.

First, what metric or set of metrics is most useful for gauging the potential for higher mortgage rates to lead to financial instability, either via higher arrears and defaults by mortgagors or via macro-critical debt deleveraging in the household sector?  A lot of the discussion of this topic in previous FSRs has tended to focus on one or other of two polar extremes – either the burden of mortgage debt across the entire UK household population or the relatively small number of people in the extreme tail of the distribution with very high debt-servicing burdens.[1]  Arguably the right focus for gauging the macro-significance of these channels lies in between.  After all, most people do not have a mortgage at all, and income levels vary considerably across the mortgaged population.  Moreover, many households also have other debts (personal loans, credit cards) and some may have liquid savings which can act as a buffer.  What is needed is a distributional assessment of the burden of debt and debt servicing, including other debt liabilities, relative to resources (disposable income and savings) across the household population and how that burden is being affected by rising interest rates.  Such an analysis would also need to take into account the effects of rising interest rates on the rental market.  That information will provide a clearer impression of the outlook for arrears and defaults and would also be useful for the MPC to help it gauge the potential hit to consumer spending.

Second, while current market pricing suggests Bank Rate could rise above 6%, what is a reasonable stressed estimate for how high Bank Rate and hence mortgage rates might rise at this juncture?  As discussed in a previous blog,[2] it would make sense for the MPC to provide an answer to that question, with support from Bank staff.  The MPC already produces a probabilistic assessment of the macroeconomic outlook in the form of fan charts, so it should be possible to produce an assessment of the policy paths that are consistent with the upper tails of those fan charts and an approximate estimate of how high Bank Rate could go in a bad case scenario.  That information would be useful to the FPC, to help it reach an informed assessment of the risk to household finances, beyond the current base case outlook for mortgage rates.  But this would also be useful information to provide to the public, including those whose fixed-rate deals will expire over the next few years and those contemplating switching from a fix- to a floating rate mortgage, so that they can prepare now for possible bad outcomes in the future.  Obviously, the Bank will be concerned about the risk that any statement about the future path of rates will get over-taken by events.  But it is not obvious that leaving mortgagors to work it out for themselves instead is a better outcome for financial stability.

Third, how resilient are bank balance sheets and household finances to a severe macroeconomic stress?  It seems possible, if not probable, that any severe interest rate stress of the kind discussed above would be accompanied by a material slowdown in the economy.  After all, the current mantra from central banks is that they need to raise interest rates to tackle inflation because otherwise they risk losing control of inflation and it will then be much more painful to dis-inflate the economy.  The stress scenario we have in mind is that “much more painful” scenario where unemployment must rise – and potentially materially – to drive wage and price inflation down, reducing disposable incomes alongside the squeeze from higher mortgage rates.  It is not inconceivable that there could be a material correction in property prices too – especially as some landlords will struggle to service their debts and may choose instead to sell properties, putting downward pressure on valuations.  The stress tests previously carried out by the Bank suggest that lenders can withstand this kind of stress.  We’ll of course see the results of the latest stress test alongside the publication of the July FSR.  Is the FPC confident that the stress scenario it chose back in September for this year’s test encompasses the stress the UK economy could now face?  One reason for doubt is that, at the time of writing, the market curve is already above the path for Bank Rate assumed in the stress.

Fourth, how does the Committee feel about the proliferation of ultra-long-term mortgages and other efforts to forebear?  We know that the incidence of mortgages beyond the traditional 25-year term has been on the rise, particularly among first-time buyers.  One possible outcome of the looming mortgage stress might be for more borrowers to pursue that option.  If the debt is repaid over a much longer term than the monthly repayment can fall.  Of course, the obvious catch is that you pay more over the lifetime of the loan because the principal is re-paid so slowly.  There are broader issues here too: not least, how comfortable the authorities should be about people extending mortgage repayments beyond the conventional retirement age.

Effectiveness of resolution frameworks and implications for bank capital requirements

Is the UK banking system sufficiently resilient to stress?  While some regulators and industry groups were quick to point to the behaviour of banks during Covid episode as validation of the post-crisis regime, the March banking turmoil has put the issue of bank resilience back on the table.

One touchpoint specific to the UK is the interplay between the resolution regime and the bank capital framework.  In December 2015, the Bank’s FPC published an analysis of the optimal level of capital requirements for UK banks.  A key judgement the Carney-led FPC made at the time was that UK banks’ capital requirements should be reduced by 5% points from the level that would otherwise be warranted to reflect the benefits of an effective bank resolution regime.  Or as the FPC put it:

“The FPC judges that effective arrangements for resolving banks that fail will materially reduce both the probability and costs of financial crises. In the updated Bank of England analysis, these arrangements are assessed to reduce the appropriate equity requirement for the banking system by about 5% of risk-weighted assets.”

“These changes [the internationally agreed TLAC and MREL standards, plus the introduction of the UK resolution authority] will help to support a credible and effective resolution regime for banks in the United Kingdom, allowing individual banks to be recapitalised in resolution, without the need for public solvency support. Orderly resolution will minimise the damage to the real economy caused by bank failure and avoid unnecessary interruption to the critical functions those banks provide to the real economy.”  (Capital Supplement to the December 2015 FSR)

The events of March brought conflicting evidence on the validity of this judgement.  On the one hand, we had the seemingly successful resolution of the UK subsidiary of Silicon Valley Bank, which was taken over by HSBC at zero cost to the UK taxpayer after the failure of its US parent.  On the other, we had the failure of Credit Suisse.  While the AT1 holders of Credit Suisse were wiped out and its equity holders took significant losses, there was no attempt to bail-in TLAC bondholders.

Speaking shortly after the event, Swiss National Bank president Thomas Jordan is reported by the FT as having explained that putting Credit Suisse into resolution would have risked a systemic crisis: “resolution in theory is possible under normal circumstances, but we were in an extremely fragile environment with enormous nervousness in financial markets in general…Resolution in those circumstances would have triggered a bigger financial crisis, not just in Switzerland but globally.” “[It] would not have worked to stabilise the situation but, on the contrary, created enormous uncertainty …It was clear that we should avoid it if there was any other possibility.”[3]  Similarly, both the Swiss finance minister and the head of Finma are reported as having said that resolution was unsuitable because Credit Suisse involved “loss of trust” and a run.[4]

In the light of events in Switzerland, does the FPC have confidence that the Bank could use its resolution toolkit were a UK bank of the size and systemic importance of Credit Suisse to become distressed in disorderly market conditions?  If not, is it time to revisit its judgement that capital requirements for UK banks should be reduced by 5% points thanks to the presence of a resolution regime?

Gauging vulnerabilities in the non-bank financial system

On the 19 June, the Bank announced that it would be launching the first system-wide exploratory scenario exercise.  The exercise, which we are told will involve large banks, insurers, central counterparties, and a variety of funds including pension funds, hedge funds, and funds managed by asset managers, has two aims.  First, to “enhance understanding of the risks to and from NBFIs, and the behaviour of NBFIs and banks in stress, including what drives that behaviour”.  Second, to “investigate how these behaviours and market dynamics can amplify shocks in markets and potentially bring about risks to UK financial stability.”  We are also told that the focus of the exercise will be the gilt market, gilt repo market, sterling corporate bond market and associated derivative markets.

Overall, the Bank should be commended for embarking on this ambitious exercise, which chimes very much with previous blogs on this site that have expressed concern about vulnerabilities in the non-bank financial system.  Moreover, given the significance of the LDI meltdown in September after the Liz Truss mini budget – an event that was missed by the FPC in its risk scanning work – it’s vital that the Bank and FPC improve its understanding of where such fault-lines lie.

In the spirit of being constructive, we offer some thoughts about various design aspects of this exercise.

First, the launch document is clear that the Bank will not use this exercise as a test of the resilience of the individual firms it has asked to participate.  This seems a curious decision; isn’t the FPC interested to learn what the scenario will mean for individual firms’ resilience?  How can it hope to uncover whether there is another LDI-type event lurking if it is not gauging the impact on firms’ capital and liquidity?[5]

Second, how is the exercise going to account for the impact of past BoE emergency liquidity policies having put out the flames whenever risks in this sector have crystalised in the past?

Third, the launch document tells us the Bank anticipates publishing a report on this exercise “in 2024”.  It also expects to publish interim results in the FSR.  Is this time horizon sufficiently ambitious given the scale of the risk here?  After all, we are told in the December FSR that “in 2023, there is a need for urgent international action to reduce risks in non-bank finance”.  If a risk-scanning exercise will not yield results until an unspecified time in 2024, presumably policy action will not take place until a significant period after that.

Fourth, the exercise as envisaged is highly complex.  It will comprise two rounds, enabling the Bank to “account for system-wide interactions and amplification effects”.   That is, the Bank will take firms’ initial responses to the scenario and then attempt to add them up to assess their impact on asset prices and market liquidity.  It will then play these system-wide responses back to participating firms.  While this seems interesting in principle, might this not be a situation where a simpler exercise is preferable?

Fifth, the exercise will be based on a single “severe but plausible” scenario, as per bank stress tests.  While a single scenario has advantages in terms of the narrative it provides and the sharpness of any policy implications that flow from it, it could be severely misleading in these circumstances where shocks may have a highly non-linear impact on the financial system.  We saw this nonlinearity play out in the September LDI crisis, where it was reported that LDI funds had been stress tested to withstand a 100-basis point move in bond yields (ie lower than the actual move that materialised and led to systemic stress).  While the Bank will presumably look to make the stress scenario sufficiently severe to capture any such tipping points, it’s unclear how easily this can be done in practice.

Taking stock of some of these points, an alternative way this exercise could have been designed is via a simple data template, which would be sent to participating firms seeking to elicit their sensitivities to a set of risk factors and liquidity scenarios.[6]  For example, firms could be asked to report the gain or loss in their capital and how their liquidity position changes if, say, commercial real estate prices fall by 15%, 20%, 25% etc, or given a range of specific moves in the yield curve.  That is, rather than focusing on a single risk scenario, the exercise would seek to glean firms’ sensitivities to a set of shocks.  The reporting template would also need to include both single factor scenarios and cross-scenarios that combine moves in several factors simultaneously.  If the set of risk factor shocks and liquidity scenarios could be kept constant, this would create a panel of responses over time which could be a useful input into understanding the dynamic evolution of risk, while in addition reducing the burden on firms participating in the scenario.   

Endnotes

[1] For example, recent FSRs have focused on the proportion of households with cost-of-living adjusted mortgage debt-servicing ratios above 70%.

[2] “Reactions to the Bank of England’s December 2021 Financial Stability Report”, Richard Barwell and David Aikman.  https://macroprudentialmatters.com/reactions-to-the-bank-of-englands-december-2021-financial-stability-report/

[3] https://www.ft.com/content/4f0c9cc8-192c-4b2b-bd78-4943d23b17a3

[4]  As reported by Admati, Hellwig, Portes (2023).

[5] This could perhaps be “Bank speak” for there being no clear pass/fail hurdle rate in this exercise akin to what it uses in the ACS stress test.

[6] Readers will recognise this as essentially the Risk Topography approach advocated by Marcus Brunnermeier, Gary Gorton and Arvind Krishnamurthy in 2011.