This article has kindly been written for this blog by Con Keating and Iain Clacher

Iain Clacher and Con Keating

TPR Leveraged LDI survey

This year, TPR’s annual DB and hybrid (mixed benefit) scheme return template contained new questions on liquidity and leverage. There are 23 new questions, listed below as appendix A.

The Times refers to these as a ‘survey’ in its article of September 24, 2024, entitled “Pensions watchdog assesses risks to defined benefit schemes”, but as yet, the results have not been published by TPR:

“While this data has given us an important insight into the use of LDI, it is not data we have published on our website.  We intend to publish a technical note in the first half on (sic) 2025 once we have finalised the analysis of the data.”

The first thing that is noteworthy is the future looking nature of the survey. None of the questions seek to address the issue of what was experienced by schemes in the 2022/23 period. As such, they are not the forensic post-mortem examination which has been widely called for, including by the Work and Pensions Select Committee, and TPR’s attempt to give an answer to the Committee was a “modelled” answer, not a definitive one. By not looking backwards, the current survey will not be able to offer evidence of any change in behaviour driven by that experience.

When the first question is: “1.  Does the scheme use leveraged LDI?”, (emphasis added), we wonder just how informative this survey might be as to the current use of LDI rather than purely leveraged LDI by schemes. Surveys are only too often biased by the questions posed.

Surely, even in a survey that only seeks to look to the future, the most fundamental question to ask would have been:

“To what extent does your asset allocation seek to hedge or offset the interest rate sensitivities of scheme liabilities?”

This addresses the extent to which LDI, including leveraged LDI, was being pursued.

It is important to separate the hedges of real risks to scheme liabilities, such as longevity of members and inflation, from the arbitrary, unreal ‘risk’ of changes to discount or interest rates. It would also have been useful to have estimates of the extent to which LDI strategies failed to hedge real changes in liabilities. For example, the 2068 ILG fell by 87% in price during 2022 while retail price inflation rose from 4.1% in 2021, to 11.6% in 2002 and 9.7% in 2023.

Question 3 asks: “Is the LDI mandate pooled or segregated?”, no further questions are asked concerning leverage within segregated mandates, even though these are precisely the mandates which can and did result in substantial collateral and margin calls. This once again plays to the view of the Bank of England, and their often quoted statement that it was only a small number of pooled funds that led to the LDI-induced gilt market crisis, which was based on a misstatement of the size of pooled funds, as the BoE only reported the equity in pooled funds, not their exposure. An error that is repeated in this survey.

Given that it is liabilities which are being hedged by these strategies, the question then posed is: “What is the net asset value of this LDI mandate as a proportion of the scheme’s total assets?”  (Emphasis added) and is clearly misconceived.

Leverage should then be properly addressed by questions beginning with: What is the value of scheme assets pursuing this LDI strategy, as a proportion of scheme pension liabilities? This captures the gearing of scheme assets used in the LDI hedge strategy.

To illustrate this point, a scheme may choose to hedge its liability interest rate sensitivity by purchasing assets with a far longer duration than the duration of liabilities. If the liabilities have a value of £100 and duration of 15 years, the scheme may have chosen to invest in £50 of assets with a duration of 30 years; the scheme is 100% hedged, but twice geared. The liquidity of this strategy is immaterial as there are no inherent collateral calls associated with any decline in the assets held, though in response to large market moves some rebalancing of hedge exposures may be advisable.

The scheme might alternately have used highly geared derivatives, such as interest rate swaps and call options, to achieve its hedging. Some, but not all, derivatives carry margin (or collateral) call liabilities and should be identified for this property.

The next question would be: What borrowing has the scheme undertaken, including by the use of repo? with the further question: “What assets serve as collateral for these borrowings?”

The survey asks if a scheme’s LDI mandates are pooled or segregated. It should be noted that pooled funds cannot impose liquidity or collateral calls on the scheme holding them, though pooled funds may make requests for further subscriptions. Any borrowing by a fund and collateral calls arising from them are obligations of the fund, not the scheme holding units in that fund. A scheme’s exposure is limited to the initial purchase price. Pooled funds cannot impose collateral or margin calls on their unit-holders.

In this regard, question 6 is redundant for pooled fund holdings from a liquidity standpoint: “What is the yield buffer to zero net asset value (NAV) of this LDI mandate?” From the standpoint of a segregated mandate, it is nonsensical as it involves an arbitrary allocation of assets within the overall scheme fund, a distinction which would not be legally enforceable.

TPR has provided some guidance for schemes completing their return in the form of help files. For example:

Standalone synthetic equity mandate

9. A synthetic equity mandate uses derivatives to gain all or substantially all of the equity market exposure.

Does the scheme use a standalone synthetic equity mandate?

Answer ‘Yes’ regardless of whether or not it is leveraged (more than £100 equity exposure per £100 invested capital).” (Emphasis added)

A call option, which of course has no recourse to its scheme’s owner, and no risk of loss beyond its initial purchase price, would be captured here.

Waiting for TPR

Rather than continuing with an exhaustive critique of the questions posed by TPR and their utility, we have drawn attention to some of the main issues with survey as we see it today. Of most concern is there is nothing here that will help us learn the lessons of the crisis nor benchmark progress as to actions undertaken to reduce the systemic risk posed by LDI. The Appendix lists all 23 questions below and we await publication of the promised technical note.

Appendix: The questions.

1. Does the scheme use leveraged LDI?

If you answer ‘Yes’, you will need to complete the rest of the leverage and liquidity section.

Section 1: LDI mandate

For each individual LDI mandate you add, you will be asked the following questions.

2. Asset manager name
3. Is the LDI mandate pooled or segregated?
4. (If you answered ‘Pooled’ to question 3) What is the name of the fund?
5. What is the net asset value of this LDI mandate as a proportion of the scheme’s total assets?

Answer as a percentage.

6. What is the yield buffer to zero net asset value (NAV) of this LDI mandate?

Answer in basis points (BPS).

7. Does this LDI mandate give the scheme exposure to any other asset classes beyond (nominal or real) interest rates?
8. Select the exposure beyond (nominal or real) interest rates provided by this mandate:

Select all that apply: Equity, Credit, Currency, Other.

Section 2: Standalone synthetic equity mandate

9. Does the scheme use a standalone synthetic equity mandate?

Section 3: Availability of liquidity

10. What is the breakdown of scheme assets by the frequency on which they regularly trade?

Give percentages for: Daily, Weekly, Monthly, Quarterly, Less than quarterly.

Section 4: Delivery of liquidity

11. [If an LDI manager issues a routine request for additional capital] In how many business days are the trustees capable of carrying out the necessary transfers and fulfil the request for capital?

Select from: 0 to 10 days (and enter the exact number of days), More than 10 days.

Section 5: Efficient governance of transactions

12. What proportion of the assets (that the trustees envisage selling in response to an LDI manager’s collateral call) have the trustees delegated unlimited authority to a third-party to sell?

Select one from: All of the assets, Some of the assets, None of the assets.

13. (If you selected ‘Some’ or ‘None’ to question 12) Do the trustees have a single authorised signatory list or multiple authorised signatory lists?
14. (If your answer to question 13 was ‘Single’) How many individuals are listed as authorised signatories?

Select from 1 to 10 (enter the exact number of authorised signatories), More than 10.

15. (If your answer to question 13 was ‘Multiple’) How many individuals are listed as authorised signatories on the list with the fewest authorised signatories?
16. When was this list of authorised signatories last reviewed?

Section 6: Systemic liquidity impacts

17. Do the trustees have a pre-agreed plan of asset sales to fulfil a request for capital from an LDI manager?
18. (If you answered ‘Yes’ to question 17) Which asset class is on the first step of the waterfall?

Select from Bonds, Equity, Property, Diversified growth funds, Absolute return funds, Other or Credit facility.

19. (If you answered ‘No’ to question 17) Which is the scheme’s largest holding in a liquid asset class?

Select from Bonds, Equity, Property, Diversified growth funds, Absolute return funds, Other or Credit facility.

20. (If your answer to question 18 or 19 was ‘Bonds’, ‘Equity’, ‘Property’, ‘Diversified growth funds’, ‘Absolute return fund’ or ‘Other’) Is the asset class from which the trustees would source the required moneys (from a request from an LDI manager) accessed by a pooled fund?
21. What are the names of the asset manager and pooled fund? Name of asset manager, Name of pooled fund.
22. (If your answer to question 18 or 19 was ‘Credit facility’) When was a legal review last carried out on the credit facility?
23. Does the scheme’s LDI manager have discretion to disinvest from the relevant asset class without direct intervention from the trustees?