Over a three week period following 23 September 2022, UK gilt markets experienced dramatic market activity as systemic risks began to dominate.

During this period the Bank of England stepped in to manage market volatility and purchase c£20bn of gilts, giving pension schemes the breathing space needed to adjust their portfolios to ensure they can weather further storms.  These headlines are now well known. 

What will now inevitably and appropriately follow is debate and analysis on what if anything “went wrong”, whose responsibility it was, and what should now be done to mitigate the risk of a repeat.  This will be considered by at least three parliamentary committees in the coming weeks, and the Bank of England, Financial Conduct Authority (“FCA”) and The Pensions Regulator (“TPR”) have already written to their respective committees to set out their position.  There will also no doubt be further guidance and statements from the various regulators setting out their new expectations in the coming weeks and months. 

Clearly, lessons need to be learned.  No doubt investment managers, consultants and the regulators will (rightly) get some tough questions from MPs, peers and clients.  But in all the noise, it will be important to distinguish between “individual pension scheme risk” and “systemic risk”, and this is what this paper is all about. Otherwise, poor regulatory decisions may be made that could create further risks in the system, leading to potentially worse negative impacts in the future, including on individual pensioner benefits.

In this paper I first explain the difference between these two risks, using a “run on the bank” as an analogy. I then explain how even if individual pension schemes were being well managed, this would still leave residual systemic risk. I then consider the roles of different regulators in overseeing and regulating systemic risk, considering the Bank of England, FCA and TPR in turn.  Finally, I set out what actions can be taken now, by regulators and market participants, to reduce those risks in the future, in most cases within the existing regulatory framework. 

This paper concentrates on systemic risk rather than individual scheme risk.  Of course, there are also lessons to be learned by investment managers, consultants and trustees about managing individual scheme risk. This includes the need to improve (in some cases) the governance around quick access to sufficient liquid asset buffers where leveraged LDI is being used by schemes.  This has already happened quickly across the pension scheme universe in recent weeks, but new guidance and potentially new regulatory constraints in this area may also be helpful, and I would expect this from regulators in the coming weeks. I am not wanting to underplay the importance of good management of individual pension scheme risk, but that’s a separate debate.

What is systemic risk?

Systemic risk has no single recognised definition but is generally defined as being the risk of the collapse of an entire financial system, or entire market, as opposed to the risk associated with any one individual market participant. 

Even with the simplest commonly understood market systems, systemic risk exists.  Take banking as an example. For good capital efficiency reasons, banks never hold all our deposits as readily accessible cash. Instead, they lend out most of the money deposited with them to other people.  This creates the well understood systemic risk of the possibility of a “run on the bank”.  In particular, each individual participant in that market (you and me) can be acting very sensibly and managing our finances very well, but if we all get nervous about the strength of the bank, or all need our money on a particular day, the bank may fail since it is holding as ready cash only a small fraction of the money originally deposited with it – the rest is tied up in less liquid loans to other customers.

It is widely recognised that the responsibility for managing systemic risks cannot fall primarily on the market participants. It falls on governments and regulators to regulate the banking system to minimise the risk of a bank run happening. They do this through a variety of means including: closely monitoring the cash being withdrawn from banks, requiring banks to hold strong levels of capital, supporting confidence in the banking system, and setting up central protection funds to guarantee cash protection for consumers up to a reasonably high level (currently £85,000 in the UK). It is then generally considered reasonable for consumers to assume that they are operating in a reasonably stable system, governed sensibly by central regulators. And as part of this consumers generally also assume that, in a worst case scenario, the Bank of England will step in and take action, evidenced by historic central bank actions. And indeed this is recognised as a rational and helpful role of central banks.

What is systemic risk for leveraged LDI?

There are many parallels between bank runs and the systemic risks that played out with pension scheme leveraged LDI arrangements following 23 September. 

Simplistically, the “run on the bank” for LDI happened something like this:

• Gilt prices had already fallen over the course of 2022, and available liquidity in pension schemes was therefore lower than historically, but pension schemes remained generally well capitalised – they had mostly managed their individual scheme risks well
• In response to the mini-budget, gilt prices fell sharply on 23 September
• In response, some LDI investment managers asked pension schemes to provide additional cash to top up leveraged LDI portfolios (this process was well understood, and had already been happening throughout 2022)
• From the evidence that I am aware of, most schemes were able to and did provide the necessary cash
• A few schemes could not do so, or could not do so quick enough, or chose not to do so
• To shore up their positions, the LDI managers who didn’t receive the necessary cash reduced hedges for such schemes, which in effect involves selling gilts - this was a process deliberately set up by these LDI arrangements to ensure that leverage levels did not become excessive
• The level of gilt sales was on such a scale that, when combined with other sellers taking fright at the budget (and perhaps short-selling by speculative investors looking to profit), gilt prices further fell sharply on the next trading day
• And the cycle started again, leading to a “negative feedback loop” (TPR’s words) that was only halted by Bank of England intervention (there were further repeats of this in the following weeks as regulators, investment managers and schemes further reduced the risk in the system by acting to increase available collateral)

It is important to note that, even if we were to make the extreme assumption that all pension schemes are very well managed, and they all make very sensible risk-based financial decisions based on their own financial position, systemic risk will still exist. In relation to leverage, systematic risk can only be avoided by insisting no pension scheme uses leverage in their portfolio. However, leverage itself is not a fundamental problem if used sensibly. Look at houses and mortgages – the vast majority of people would never be able to own a home if they weren’t able to sensibly borrow through a mortgage.  

As many other commentators on LDI have explained, eliminating the possibility of leverage within pension scheme risk management solutions would be capital inefficient, increase other risks in the pension system, and result in additional £100s of billions of extra pension costs for UK PLC (who sponsor the schemes). In turn, this would be at the cost to investment in their business and the economy, and reductions in dividends to shareholders. 

Worst still, in a minority of cases, eliminating leverage from some pension schemes would lead to excessive risk being put onto some sponsors, a few of whom may then become insolvent, the pension scheme would be tipped into the Pension Protection Fund, and members could have their pensions cut.  So LDI clearly has a very helpful role to play and a regulatory outcome that bans leverage is clearly to be avoided – in my simple analogy, it would be equivalent to requiring all banks to always hold everyone’s cash ready to withdraw.

An individual pension scheme’s primary focus should be on its own risks: doing what is best for the scheme to protect the pensions of its own members. In many cases, this will involve using some leveraged LDI. In coming to decisions about how much risk should be taken, the trustees and their advisers should be well aware of systemic risk (just as an individual should be careful about not investing more than £85,000 in one bank). But schemes can only meaningfully quantify systemic risk if they have good information about the nature and size of systemic risks. And that involves having information about what all other pension schemes are doing. Only regulators can collect, monitor and publish that information – more on that later in this paper.

It should also be noted that it will always be rational for some market participants to take more risk than others, in the absence of any regulatory constraints. For example, one pension scheme may choose to increase leverage and hold considerable amounts of growth assets such as equities, if this is the only way they can see a reasonable path to providing their pension scheme benefits over the next 40 years without putting crippling risk on their sponsor. The scheme may be very well managed, and its trustees may do this with their eyes open about the risks, but such a scheme will necessarily increase the systemic risk for everyone else. This is because (as the trustees will be fully aware) they will need to unwind their position by selling gilts if gilt prices fall sharply, which in turn will exacerbate the market challenges for all other pension schemes. But it is important to note that they have added to systemic risk not because the scheme is necessarily poorly managed (although no doubt some such schemes were) but because the scheme was making entirely rational decisions for its own circumstances.  

What is the role of various regulators in managing systemic risks?

Whilst individual market participants should be aware of systemic risk and should clearly bear it in mind in their risk management decisions, systemic risk is primarily the responsibility of regulators. 

Regulators can choose to address the risk in different ways, but typically should be identifying risks, collecting and publishing data, monitoring risks and communicating about the risks as a minimum.  In many markets, law makers and regulators go further by putting constraints on the behaviours of individual market participants to further mitigate the risks.

As highlighted in the 5 October Bank of England’s letter on LDI to the House of Commons Treasury Committee, the Bank of England’s 2018 Financial Stability Report included a study into the risks of leverage in the “non-bank” financial system, which is primarily leveraged LDI in pension schemes.  This study (see pages 51 to 57) makes interesting reading. The key points I took from it are as follows.

First, it is positive that the Bank of England was working with the other relevant regulators (the FCA and TPR) to consider systemic risk arising from leveraged LDI back in 2018. Considering the statutory objectives of the regulators, it is worth noting that TPR does not have a market based objective, and is generally unable to intervene directly in the investment strategies of schemes as a whole, and therefore cannot lead on systemic risk considerations. However, the FCA has an objective of making sure “the relevant markets function well” and the Bank has a financial stability objective.  It is therefore helpful and appropriate that they see the importance of working closely together in this area and see it as their collective responsibility.

Second, in the report the Bank describes a scenario test it conducted based on a 1% pa upward movement in gilt yields, concluding that pension schemes could cope with that impact. This is described as a 1-in-1000 event over one month. It is interesting to note that, ignoring systemic risk, that may well be a reasonable assumption. But, in hindsight, systemic risk (amongst other things) contributed to movements much greater than this in late September 2022, and it would be helpful for all market participants to note this. 

Third, they helpfully concluded that “Data currently reported to the supervisors [ie TPR] of non-banks [ie pension schemes] do not include all the information needed to monitor the risks appropriately. The Bank will work with other domestic supervisors — the PRA, FCA and [in the case of pension schemes] The Pensions Regulator (TPR) — to enhance the monitoring of these risks”.  They also conclude that “If it is found that risks reach systemic levels, further action should be considered [presumably action by them and their fellow regulators]”.

Turning to TPR, in their 10 October letter to the Work and Pensions Committee, TPR stated “We do not record in-depth data on the scale of collateral or leverage agreed to by DB schemes, and we do not ask every scheme to provide this data.” TPR does have a natural place to systematically collect this data from all schemes (which in my view is what is needed to effectively assess systemic risk) which is the annual scheme return, legally required to be completed by all schemes.  It is interesting that they have so far chosen not to collect complete data in this way, and my view is that this should change in the future. 

It is also worth considering the role of the FCA in collecting information.  In their 20 October letter to the House of Lords’ Industry and Regulators and Economic Affairs Committees, the FCA state:

“In March this year, the FCA contacted the largest LDI fund managers asking them what plans they had in place to deal with increased volatility and to absorb volatility in excess of the previous scenario analysis. We also probed large managers on the speed with which they could call money from underlying pensions funds in the event of stress, in some cases leading to a shortening of this period and an increase in their resilience.

In May and June, we saw these funds subject to significant margin calls, which they met according to their plans without systemic issues or concerns.

We remained in close contact with fund managers through the summer and early autumn. By Thursday 22 September firms had buffers in place to deal with greater market stress than they had experienced previously, in accordance with what they saw as extreme but plausible in terms of sharp moves in the value of UK gilts.”

This work with investment managers was no doubt helpful and presumably helped reduce risks through the first 9 months of the year. However, it is interesting to note that this was only focussed on the “largest” managers, and that the FCA could necessarily only “see” the money that those managers had direct access to and sight of, and could only rely on the manager’s reports of their own governance systems. Note that many larger schemes appoint multiple managers for different purposes (one for LDI, another for equities etc) and the FCA cannot have direct line of sight to pension schemes, as the FCA does not regulate pension schemes. The FCA is also not responsible for the oversight of the governance of pension schemes.

In a published speech dated 7 November, Sarah Breeden, Executive Director, Financial Stability Strategy and Risk, of the Bank of England discussed learnings from the experiences in the LDI and gilt markets since 23 September. 

In her speech, she says that “The onus for building resilience in the non-bank system sits first and foremost with the firms themselves” (ie on pension schemes) and that this includes the risks arising from “exposures to amplification mechanisms from the wider system” (ie systemic risks).  

Whilst I agree that pension schemes need to be aware of systemic risks in their decision making, as I have explained in this paper my view differs in emphasis, in that I believe it is the regulators (including the Bank of England) that should be leading on oversight of the resilience of the pensions system and the risks of amplification mechanisms.

The speech helpfully goes on to re-iterate that regulators and market participants need information (“transparency”) to be able to do this effectively, and concludes:

“Lessons must be learned from these episodes, most importantly by non-banks themselves, but also by: their counterparties; market infrastructure; their regulators; bank supervisors; central banks; and the global regulatory community as we continue our global efforts to ensure the resilience of the system of market-based finance.

Transparency is an important first step. That enables the necessary next step of ensuring non-banks’ positions and interlinkages with the rest of the financial system can be comprehensively stress tested and understood in a system-wide manner.”

I agree with this conclusion. Better data collection and more transparency is vital to managing systemic risk, from the Bank’s perspective, from the perspective of the FCA, TPR, and for all market participants. 

In conclusion

There are clearly lessons to be learned by all participants from the extreme market volatility, exacerbated by systemic risk, that occurred in gilt markets following 23 September.  In this paper I have concentrated on the primary role of regulators in regulating systemic risk: the Bank of England, FCA and TPR.

The complex interaction of regulatory responsibilities when it comes to pension schemes, and specifically LDI, can cause confusion, but my view of what now needs to happen to mitigate systemic risk, by all three regulators and the pensions industry working closely together, is:

1. TPR should work with market participants to collect relevant systematic information from all pension schemes on the size of LDI exposures and the associated leverage risks (if the lawyers say this is not possible because of TPR’s limited statutory objectives in this area, this should be addressed by Parliament)
2. Regular monitoring (weekly or daily in times of stress) of exposures should be extended by the FCA to cover the whole of the market, with risk trigger points for further actions
3. Information and data should be shared between the regulators, and analysed to better understand the nature and extent of the systemic risks for the gilt and related markets
4. Regulators should use that analysis to help market participants understand whether they are using leveraged vehicles sensibly or not
5. Regulators and policy makers should work closely together with the industry to identify if the introduction of new constraints on pension scheme leverage (eg requiring managers to have line of sight to certain levels of liquid collateral) might be helpful, but this needs to be considered carefully in the round given the potential for unintended consequences.

Note: This analysis was submitted as written evidence to the Work and Pensions Committee ahead of its oral evidence session on 23rd November 2022 at which Jonathan Camfield gave evidence.