Written evidence from Bernard H Casey LDI0008

Liability Driven Investment – a whale or a minnow, a minnow or a fugu?

My name is Bernard H Casey.  I am a retired academic, a former senior economist at the OECD, and I was also a frequent consultant to the European Commission.  I maintain a website updating my activities at www.soceconres.eu.

My concern is with the importance of LDI to the UK defined benefit (DB) pension system.  This relates to the issue of disclosure – what do we know about LDI both at the macro and the micro level? And if we do not know, we do not know, either as a government, or regulator, or as an investor, how to respond.

LDI has been the subject of extensive reporting and commentary since late September.  Numbers are bandied about.  What they mean is unclear.  My concern is with what is being talked about at all, whether the extent of LDI is large or small and whether, even if it is small, it is, nonetheless, dangerous.[1] I have invested a considerable amount of time an effort into researching the matter and end up still in the dark.  Nevertheless, I trust my contribution might shed some light upon how those who have expressed concerns might go forward.

As well as presenting this written evidence, I should be happy to make myself available to answer questions from the Committee.

How much LDI is there in total?

Some of the estimates used will be familiar to the Committee.  They are listed in the table below.

The numbers here appear to relate to “exposure” as opposed to value of holdings – although this is not always made clear.  The size of the exposure needs to be set in relation to something, too.

Best estimates of the size of UK DB pension assets in autumn 2021 were some £2.3tn.[7]  On this basis, exposure might have been as great as 65 per cent of assets or as little as 43 per cent of assets.   However, a lower denominator has also been frequently used - £1.5tn.  This is the sum of assets recorded by schemes listed with the Pension Protection Fund.  On this basis, exposure is between 66 and 100 per cent of assets.   In fact, the ONS denominator might be too high and the PPF one too low.  The ONS covers all funded DB pension schemes, whilst the PPF excludes local government pension schemes.  The latter have assets in the order of £0.4tn.  However, there are also suggestions that local government pension schemes are much less intensive users of LDI than those sponsored by private corporations.[8]  The PPF also excludes schemes subject to “crown guarantee”, although these are neither many nor large. 

The numbers reported in Table 1 appear to have more “whale-like” than “minnow-like” qualities.  On the other hand, the latest estimates from the BoE suggest that the problems lie in only a fraction of the exposure – that contained in the £200bn (£0.2tn) in pooled funds.  These make up only one seventh of total exposure.[9]  These funds are, however, often highly leveraged, and it is investments in these which have been most likely to have had to sell off assets to meet collateral requirements when gilt rates rose.[10]  Thus, they might be small, but they are highly dangerous – more “fugu-like” than “minnow-like”.

Obtaining further information has proved yet more difficult.  I have contacted the ONS for details on the extent to which participation in LDI is recorded in its series Funded occupational pension schemes in the UK. It does not contain the relevant data simply because this is not collected via the survey behind it – the quarterly Financial Survey of Pension Schemes.[11]  The latter does include an item “Pooled investment vehicles”, but this is much wider ranging than merely pooled LDI arrangements.  With respect to “Gross liabilities other than pension liabilities, excluding derivatives”, the ONS conducted a one-time special survey (September 2021) that showed that “Repurchase agreements accounted for most [88 per cent] of pension schemes’ non-pension liabilities”.  It assumes (personal communication with author) that this share has remained relatively constant in recent times.[12]  Last, the ONS also reports a figure for “Swaps (interest rate and inflation swaps but not currency swaps)”.  In each case, the values quoted are the assumed market values of the asset, not the exposure. 

How much LDI is there fund-by-fund?

Understanding exposure at the level of individual pension schemes is equally fraught.  Annual reports, where available, do publish some details, but the funds do this under the heading of assets, and they quote values at the current value of the asset.  Where they do have exposure, this is usually buried in footnotes to the main accounts, and how the reports present this is not consistent – it is necessary to search for terms such as “leverage”, “derivatives”, “swaps”, and “pooled (investment vehicles)”.  What is required can be found under different headings and depends upon the report in question.  For example, BT, which protested that it had been unaffected by the LDI crisis of September/October 2022,[13] However, BT states that it has £3.2bn on swap contracts and a longevity insurance contract .[14]

SAUL, which provides pensions to administrative staff in a consortium of (mainly London-based) universities, invests in a Legal & General Synthetic Equity Portfolio that uses exchange traded equity futures.  This gives it exposure to £294mn worth of equities against an asset allocation of only £28mn. It also has an interest rate exposure of £153mn against an asset allocation of £5mn.  On top of this, swaps on inflation and interest rates give it an exposure of £4,674mn (£4.7tn), but as an asset this is marked at only £36mn.  Net assets in SAUL, against which exposures can be measured, are £4,564mn (£4.6tn).[15]`

The Universities Superannuation Scheme (USS), which provides pensions to academic staff in the pre-1992 universities, proudly proclaims its “leverage” – to the extent of 27 per cent.[16]   The assets of the scheme are in the order of £90bn.  Table 2 provides a partial examination of the USS’s position.

The table illustrates the scheme’s use of leverage to build “growth assets”.  However, little details are given as to how this is done.  The report mentions the use of “repos” as a 5bn liability under a footnote labelled “other investment balances”.

The BoE pension fund holds £4.1bn of its £5.2bn assets in a “Legal & General LDI Portfolio”.  This invests in UK fixed-interest gilts, UK index-linked gilts, UK other index‑linked securities, derivatives, and cash.[19]  The fund also has exposure of £4bn within one year against fixed income derivatives against an asset value of £587mn.  It also has an exposure of £1bn of interest and inflation swaps with a value as an asset of £76mn.[20]

Shortcomings become apparent when the case of Sainsbury’s is investigated.  The company was reported as having offered its DB pension fund a temporary loan of £500mn (one third of the company’s then gross cash balance) at the height of the gilt market crisis that followed the September “mini Budget”.[21]  Because the scheme is “closed”, it needs to report no more than its FRS 102 position (a table of assets and liabilities), and it does this as a note to the company’s main annual accounts.[22] All this note says is that the scheme “adopts a liability driven investment framework to generate favourable asset returns with reference to its liabilities by largely removing its interest and inflation uncertainties”.  It then shows the scheme had derivatives to the value of £745mn.  The extent of exposure is unrecorded, and so what the scheme might have needed to meet collateral calls is unknown.

Most corporate DB pension are already closed.  Thus, information about them might be as limited as it is for Sainsbury’s.

Summary

In order to understand the nature of LDI, far better reporting is required.  There might well be fugu out there. Major financial regulators – from the Bank of England to the Pensions Regulator, appear to be doing little more than making “guestimates”.  The Office of National Statistics has, so far, not collected the necessary information – however assiduous its current surveys of pension schemes are.  Financial reporting by individual firms is woefully lacking.  There is no consistent use of pertinent terms.  Information that allows proper exposure to be put up front is not mandated. Thus, accounts of companies sponsoring DB pension schemes – whether or not these are closed – fail to provide investors with critical information.

November 2022