Written evidence from WTW LDI0038

Who are we?

This response has been drafted by WTW’s Investments business in the UK. Our Investments business works with a diverse range of institutional investors, including pension funds, insurance companies, sovereign wealth funds, government funds, wealth management companies, endowments and foundations. We provide fiduciary management and advisory investment services to our clients as well as the management of specialist portfolios through our internally managed pooled funds. We advise on >£3 trillion of assets globally (as at 31/12/2020) and manage £137 billion of discretionary assets globally (as at 31/12/2021).

Introduction

We welcome this opportunity to respond to your call for evidence into the use of Liability-Driven Investment (“LDI”) by UK defined benefit (“DB”) pension schemes. LDI has played a key role in helping trustees to manage the risk/return trade-off of the investment strategies adopted by DB schemes and we believe it will continue to do so. LDI often forms the foundation of risk management for UK DB schemes and thereby underpins the security of pension benefits for millions of members. However, we recognise that recent market events have not just impacted LDI portfolios; they have impacted government bond markets and the millions of savers and borrowers whose interest rates are linked to these markets.

Summary of our view

-          LDI addresses real risks faced by pension schemes and is likely, in concept, to remain appropriate

-          Details of how LDI strategies operate are important and will need to change in some cases

-          Additional oversight of LDI leverage levels is likely appropriate – given that it has been demonstrated that the “weakest” holders of LDI assets can cause a yield spiral, the details of this require care.

Why use LDI?

LDI has been a very successful strategy historically. In terms of financial outcomes, the value of LDI was most evident during the significant yield falls from the early 2000s to 2020. Over this period LDI helped UK pension schemes stabilise funding levels, reduce deficits and reduce company contributions – relative to a counterfactual where clients had similar portfolios but less protection against yield falls.

DB pension schemes are largely mature and maturing so will inevitably hold a large quantity of interest rate (and inflation) sensitive assets to manage the economic risks they face, which include:

• The need for a low-risk source of income to pay benefits (or hedging the need to purchase assets of this type in the future)
• Inflation-sensitivity to match pension increases
• Hedging the likely cost of any future buy-ins/buy-outs as the ultimate way of securing the pension liabilities.

As such, whilst “accounting” measures of pension scheme funding can increase pension schemes’ desire to enter into LDI type strategies – this is not their sole driver.

The use of some leverage within LDI has benefits in terms of being able to hold the required interest rate (and inflation) sensitivity whilst also meeting other investment objectives (such as expected return and risk). The amount of leverage to use (or the amount of liquidity to ‘reserve’ against yield increases, which is called “collateral headroom” in the jargon) has always had to be balanced against other considerations such as taking more risk elsewhere in a portfolio if less leverage is used. The calibration of these competing objectives was severely tested by recent market moves in September and October that were multiples of prior experience and revealed the relevant part of the gilt market to be more fragile and less liquid/deep than prior evidence indicated.

Even those trustee boards that were prepared enough for an outcome materially worse than anything seen in history may still not have been sufficiently prepared for the size of the daily changes in gilt yields experienced in recent months – particularly if they had been left to continue without intervention from the Bank of England.  Particular challenges were faced by investors in pooled LDI arrangements given the lack of timely information on the exposures in these funds by the LDI managers concerned and issues relating to the dealing/settlement cycles of these funds and the short notice given to investors by those managers to respond to calls for additional collateral.

How LDI might change?

This evolution in the risk/return trade-off for LDI was rapidly recognised by WTW and many of our clients during the crisis. Significant asset sales (mostly in non-Gilt/LDI assets) were undertaken to meet margin calls and, in many cases, modest, pro-active (i.e. not enforced) reductions in LDI exposure were carried out to reflect the increased riskiness of leveraged LDI in light of the very high volatility environment. 

Markets remain volatile and it is too soon to say for certain how the industry might evolve in response to the crisis. However, some of our tentative observations include:

• We think we will see lower leverage levels used in LDI strategies. The “worst-case” outcomes for gilt yield changes over specific periods have now changed as a result of the unforeseen extreme market moves. This means that what was generally considered as “prudent” in relation to target LDI leverage levels, and collateral headroom, for pension schemes will now change to reflect the unprecedented risks which became evident during the recent extraordinary market conditions.
• Some pension schemes may need to consider faster acting liquidity management processes around the use of LDI. For example, in the use of collateral “waterfalls” (pre-defined disinvestment plans in the event of yield increases). These “waterfalls” are not only relevant in “crisis” scenarios – but for ensuing liquidity is maintained in ‘normal time’ so it is then available in crisis times. Pension schemes that did not start rebalancing early and with urgency in 2022, through the yield increases seen before September, will have been worse placed for the crisis seen in September. The perception of the need for urgency around this rebalancing is likely to have changed.
• We believe that the hurdle to pension schemes using segregated LDI mandates, relative to pooled LDI mandates, may now be lower given the additional flexibility of, and clarity of reporting of, segregated structures in periods of market stress.
• The trade-off between risk and return for LDI has also changed. Given current market conditions, liquidity risk is greater than it was previously, meaning that the overall risk reduction from using LDI is lower than it once was. While this does not mean LDI is now unattractive, these risks will need to be considered in more detail on a scheme specific basis when deciding on the use of LDI and level of interest rate and inflation hedging to target going forwards.
• Pension schemes with greater resources, or access to resource (for example, larger internal teams or outsourced management) have found it easier to respond in real time to market events.

How regulatory changes might help

A pension scheme using leverage in their LDI portfolio cannot guarantee that yields will not rise to such a level that they are forced to deleverage to be able to meet margin calls. In a fragile and illiquid gilt market, the actions of pension schemes with lower governance, or with more aggressive leverage levels, could cause issues similar to what was seen in September and adversely impact other pension schemes. Regulations that limit this would benefit the system as a whole and protect those that have been more prudent.

However; any regulatory changes would need to be done thoughtfully. A “simple” requirement to hold higher levels of minimum-collateral headroom could be pro-cyclical. Once deleveraging commences and gilt yields spiral, schemes are required to either sell assets to increase the levels of eligible collateral available, or else sell hedging exposure to reduce the size of potential collateral calls, which acts to depress gilt yields further.  The pro-cyclicality of regulation has been explicitly recognised in bank capital requirements, with a counter-cyclical capital buffer that can be increased in good times and reduced in bad times (e.g. during Covid).  A counter-cyclical approach to collateral headroom would have seen requirements raised as gilt yields fell and reduced as gilt yields rose, thereby reducing the severity of the deleveraging cycle. Additionally; there is a material difference in “risk” if the assets outside of LDI are daily dealt vs illiquid assets that may take months or years to sell.

Another relevant area for further regulatory focus is LDI pooled funds. These suffer from similar “collateral headroom” issues as set out above. However they also face unique challenges due to (1) their “limited liability” nature which may mean that bank counterparties may then treat LDI pooled funds differently than segregated LDI accounts, and (2) the settlement/dealing cycle of pooled funds which means, in stress scenarios, they are less able to recapitalise quickly – and are thus more susceptible to becoming forced sellers. This makes existing Pooled Fund structures more exposed to the pro-cyclicality point than segregated structures.

As a general point, we caution that introducing new regulations for LDI could push risk elsewhere. For example, introducing a minimum level of collateral headroom could mean that interest rate and inflation hedging within pension schemes is reduced – effectively trading one risk for another. Alternatively – leverage could move to less tightly regulated jurisdictions or asset classes. This would need to be considered carefully ahead of any policy changes being implemented.

Broader implications

-          The use of lower leverage levels in LDI strategies will have consequences for other parts of a pension scheme’s investment strategy. For example, all else equal it will be necessary to target a lower level of expected return or accept greater risk exposures elsewhere. This may well have consequences for the sponsoring company that could have to make additional deficit contributions, be more exposed to a funding deficit widening, or need to support the pension scheme for longer before a buy-out is affordable.

-          Similarly the appetite of pension schemes to invest in illiquid assets is likely to reduce. This may well have an impact on the willingness of pension schemes to make investments that support the real economy such as infrastructure investments.

-          We suggest that the investigation of root causes here should go beyond LDI and seek to identify other causes of why the relevant parts of the gilt market were fragile. In this regard we would note:

• There are “real world” consequences of this fragility – such as on mortgage rates
• To support ‘market functioning’, we would suggest that consideration is given to how backstop liquidity can be provided to the non-banking sector, such as pension schemes, without having to go through the banking sector.

Summary

Based on what we’ve seen so far in the market, we continue to believe that LDI is an appropriate risk management strategy for pension schemes, provided there is appropriate evolution in how it is used to reflect the changed market conditions in which the industry now finds itself.  We have identified some regulatory interventions that could potentially improve the risk/return trade-off for LDI further.

We trust that our observations on the LDI market in this letter are helpful to the Work and Pensions Committee’s inquiry, and we would be happy to respond to any follow up questions that you may have on our response.

Please note that the observations in this letter are provided to the Work and Pensions Committee for general information purposes only and should not be considered a substitute for specific professional advice. We also note that some of the observations in this letter may not apply to all pension schemes.

November 2022