Volumes of articles have already been written on the subject of the recent gilt market disruption, the role of Liability Driven Investment (LDI) and the unprecedented intervention of the Bank of England, and doubtless, there are books in preparation. The central role of LDI has been correctly identified and there has been some focus on these investment management strategies as risk management operations, but the question largely unasked and certainly unanswered is what motivated pension schemes to adopt LDI in the first place.
The answer to that question is: a misconceived accounting standard.
As the FT has observedi, the first LDI construction was introduced in 2002 response to a change in the accounting standard for pensions
The corporate accounting standards, starting with SSAP 24, then FRS17, and latterly IAS 19 and FRS 102, prescribe the use of market-based interest rates as discount rates derived from AA corporate bonds or the equivalent in the estimation of present values for defined benefit pension scheme liability projections. Similarly, the funding regulations under which schemes must operate revolve around market rates but with the option to use the expected return on the scheme’s investments or expected long term interest rates (or a mix). It became almost universal practice for actuaries and almost all others to describe the discount rates employed in gilt yield relative terms – a practice which the Pension Regulator adopted, which likely served to encourage others to follow suit.
The discount rate employed is applied to the projected values of liabilities and defines a trajectory of liability values for the scheme; this trajectory might alternatively be described as an amortisation schedule. There is no “correct” discount rate; there is a plethora of rates which are feasible and plausible. Using market-based rates, such as gilt or corporate bond yields, introduces discontinuities and market volatility into the trajectory or “journey” plan. Unless these yields are unchanged over time, a most unlikely situation with market rates or prices, the associated assets required for funding purposes will require rebalancing with further funding needed for those where rates have declined.
There is an optimal discount rate for pensions; the compound rate of return determined by equalisation of the contribution made and the benefits projected. We have referred to this as the Contractual Accrual Rate and noted that it is used by HMRC for some long-term contracts. Unless there are revisions to the projected benefits, this rate is invariant and fixed, which is a desirable property in a discount rate.
One of the most fundamental aspects of the current debates that is missed is that variability of the discount rate is not a risk of the benefits ultimately payable to members. Rather it is concerned with the path to meeting those liabilities. In and of itself, an inappropriate or highly variable discount rate is inconsequential unless and until decisions and actions are based on it. Unfortunately, most of the major decisions and actions of schemes are.
All that hedging of the discount rate does is to fix the trajectory defined by the discount rate prevailing at the time of introducing the hedge. This can be seen as attempt to recover the invariance of the optimal discount rate. Crucially, it is not a risk to pensioners that is being hedged. A kind interpretation of this would be that hedging reflects a desire to avoid the errors and costs of decisions and actions
We have raised these issues on many occasions over many years with many institutions, officials, and others. Most recently, in November 2021, in a letterii to the Chair of the UK Endorsement Board, the body charged with responsibility for the adoption of European standards into the UK post Brexit. That letter was headed: “Time to review IAS19 (Accounting for pension costs) as it does not meet the criteria for its adoption and retention including the legal requirement of being conducive to the long term public good in the United Kingdom”. We have received an acknowledgment of receipt, but nothing more. It gives us no pleasure that our fears have proved well-founded.
The hedging of interest rates, of the discount rate, is the central element of LDI, common to every scheme we have examined. The scale of this activity is huge, a fact which has been known to the authorities for many years. In November 2018, the Bank of England’s Financial Stability Reportiii has a chapter devoted to “The FPC’s [Financial Policy Committee] assessment of the risks from leverage in the non-bank financial system”. Box 5 of that report contains chart C, which displays the magnitude of exposure to interest rate swaps and the sensitivity of those swaps to changes in interest rates of the four major segments of the non-bank financial system. The measure of this sensitivity, a risk measure, is the DV 01 of a fixed receiver interest rate swap; its price change for a one basis point change in yield.
Funds and Hedge Funds had inconsequential amounts of exposure and very low sensitivities of those exposures to interest rate changes. Insurance companies had exposure of a little under £600 million with a risk of around £12 million. Pension Funds had exposure of some £900 million and risk exposure of slightly above £100 million. More recent, but pre-crisis, estimates of the magnitude of pension fund use of swaps put this at around £1.5 trillion. The interest rate swaps employed in LDI by pension funds are long-term and extremely sensitive to changes in interest rates.
What is also missed in the current analysis we see in the press and in other places is that the practice of “hedging” interest rates has actually introduced interest rate risk into pension schemes and done so in extremely large amounts.
Many commentators and trustees have taken comfort in the fact that though scheme assets have fallen in value, the present value of their liabilities has declined by as much or more, and their scheme funding ratios improved, and that places them closer to the price of buyout required by insurance companies. This is the stated objective of many schemes. However, it is a perverse or paradoxical aspect of actuarial practice that a scheme with fewer assets can be considered financially superior to one with more assets. This risk analysis by the Bank of England should raise questions as to the feasibility and indeed desirability of such a wholesale shift of liabilities to the insurance sector. Moreover, it is also worth stating clearly and without ambiguity - there has been significant destruction capital through this process, capital that comes from member and employer contributions as well as tax relief from the Exchequer. The fact that the “accounting” of this gives the veneer of schemes being stronger, they are not, they are weaker as a consequence of all of this.
Taking false comfort seems to be rather common currently. For example, a poll taken during a recent webinar, with an extremely large audience, showed that only 3.6% of schemes had failed to meet the collateral calls made upon them; that is, had defaulted on some of their financial contracts. To put this in context, a 3.6% annual default rate for corporate bonds is firmly in the domain of sub-investment grade corporate bonds. Put another way, in three or four days LDI strategies had lowered the quality and security of the DB pension system to speculative status. In addition, 5.8% reported difficulties with their pooled LDI funds. It appears, from press commentary, that most or all of these problems have now been cured, either by the infusion of new cash from the sponsor employer or by restructuring of the contracts or are in the process of effecting this. Time will tell if this is correct, but we are not optimistic.
Another poll in the same webinar reported that just 7% of schemes had seen their funding ratios decline. These schemes had to have been over-hedging, that is speculating imprudently and recklessly on further declines in interest rates.
There are far wider concerns with prudence and fiduciary duty. The instruments being used in these strategies, such as derivatives and “repo” contracts, carry very substantial liability to provide additional collateral. The counterparty may call for collateral payments reflecting the change in price of those contracts. In some cases, we have heard that fund managers have direct, unhindered access to the other assets of a scheme, with no trustee intervention. The question to be asked is how can any of this be considered compatible with a trustee’s fiduciary responsibilities?
There are significant concerns as to the legality of many elements of LDI strategies for trustees. As we have written and published extensively4 on this, we shall provide only a synopsis of those issues here.
The Pensions Regulator does not consider any of these activities to be borrowing and has actively promoted the hedging of liabilities. In our opinion, it has failed to enforce the law or to understand what it was regulating by encouraging the use of leverage by pension funds when it was well aware of the prohibition on borrowing by pension funds other than for temporary liquidity purposes (a restriction to prevent leverage or gearing).
It appears that the Bank of England takes our view as is evident from the following quotations taken from the response5 of Sir Jon Cunliffe, Deputy Governor of the Bank of England, to the Rt Hon Mel Stride MP., Chair of the Treasury Committee of the House of Commons.
“LDI strategies enable DB pension funds to use leverage (i.e., to borrow) to increase their exposure to long-term gilts, while also holding riskier and higher-yielding assets such as equities in order to boost their returns.” And, “If gilt prices fell further, it would risk eroding the entire cushion, leaving the LDI fund with zero net asset value and leading to default on the repo borrowing.”
That letter provides a balanced and accurate description of the gilt market disruption and its timeline. It describes in simple and accessible terms the mechanisms by which the crisis came about. At the risk of oversimplifying, a large external shock caused gilt prices to decline by 4% - 5% and that decline triggered cash collateral calls on the LDI repo and derivatives exposure requiring pension funds to sell assets.
These were sales into a marketplace which was already febrile and skittish. The weakness of the Sterling exchange rate discouraged foreign buyers (who currently own around 30% of gilts outstanding) from participating. The Bank of England was not a buyer, quite the opposite, it was planning to begin selling down (£80 billion over a year) the 30% of the gilt market which it owns, commencing in mid-October. Dealers, Gilt Edged Market Makers (GEMMs), have greatly reduced their trading-book inventories since the 2007/9 financial crisis (GFC). The cause of this decline has been partly the result of the increased capital requirements imposed after the GFC and partly the very low price they would receive for the liquidity provided (gilt and money market rate decreases). We should remember that, through quantitative easing, the Bank of England has, in effect, been rigging the gilt market. Investing in a market where the price is being rigged to reduce yields, does raise issues of judgment for those investing (or advising their clients to invest) in assets with a rigged price.
The letter also makes clear that the Bank’s motivation for the intervention was market integrity in the gilt market and mitigation of the broader harms possible, not the rescue of pension funds engaged in LDI. This was the first gilt purchase announcement that the Bank has justified on financial stability grounds. It has proved successful this far and has involved very few actual purchases; at the time of writing less than £4 billion.
There are many concerns over what may happen after this intervention ceases on October 14th. It is seen as a form of cliff-edge. In another recent webinar poll, 54% of the audience believed that the intervention will be extended beyond that date. Given the strength of the Bank’s statements, we believe they will be disappointed; another case of taking false comfort in wishful thinking.
There are also sound theoretical reasons which favour rapid termination of support – a distorted free market as fundamental as the gilt market carries many costs for financial services companies and the broader economy.
Could it happen again? We have pointed out that this is a crisis deferred, not resolved. Resolution will be determined by the form of action taken by the authorities. Outright banning of LDI strategies would be effective, but 90% of respondents in yet-another webinar poll stated that they believe LDI has a continuing role in pension scheme management. Like the 1920s prohibition, we ought then to expect circumvention on a grand scale.
It is widely reported that schemes are lowering their leverage and increasing their provisions of highly liquid securities, restructuring their implementation of their LDI strategies. It has also been reported that some schemes are being advised to sell significant proportions of their assets (circa 20%), despite huge uncertainty about their true funding position.6 This approach seems likely to find favour with the Pensions Regulator, as it is face-saving. Clearly, such an approach would require supervision as it is simply too important for the economy and financial system to introduce reliance upon voluntary arrangements, but there is no regulatory mechanism or regulator with the mandate and ability to do so. The scale of the value destroyed can be calculated in due course.
Moreover, this ad hoc approach allows the possibility of recurrence of such a crisis. The endogenous spiral would be reduced but still present. To use an analogy, the fire once lit would burn more slowly. The question that needs also to be considered is what is the size of the shock, the price decline, needed to ignite these coals. Unfortunately, we have little knowledge about that. Earlier this year, smaller price declines led to reports of some distressed selling and market price gap behaviour supports that. We do know that the 4% - 5% fall after the mini-budget was enough. Indeed, it appears to have been more than enough.
But with external shocks we are in the land of unknown unknowns; we have no control over such events. It is to be expected that the day will come when an external event triggers a recurrence. Some seem to believe that reversing some or all of the changes introduced in the mini budget will restore the status quo and return us to the market that prevailed before the crisis. While we do not approve of most of those policies, such reversals of policy will not restore the market to its prior levels. Trust has been lost. Resurrection is not possible.
There is a solution which works and is by far the cheapest and most efficient. It is to correct the errors of the accounting standards and remove the motivation for schemes to indulge in LDI, and that should appeal to every economist and politician. It would open the door to sound risk management and allow and indeed encourage schemes to pursue the much-discussed productive long-term investment strategies the economy needs.
1 How bond market mayhem set off a pension “time bomb”, Financial Times, 8th October 2022.
2 The authors of this letter and its supporting paper were Philip Bennett, Professor of Pensions Law. Durham University, Iain Clacher, Professor of Pensions and Finance, Leeds University and Con Keating, Chair of the Bond Commission of the European Federation of Financial Analysts Societies. The letter and supporting paper are available on request from the authors.
3 https://www.bankofengland.co.uk/-/media/boe/files/financial-stability-report/2018/november-2018.pdf Attention is drawn to Box 5, Chart C in particular.
4 See, for example: https://henrytapper.com/2022/09/29/anatomy-of-a-crisis/ and https://henrytapper.com/2022/10/03/tprs-role-in-the-current-crisis-keating-and-clacher/
5 The letter and its context are available here: Bank of England confirms details of emergency bond buying in correspondence with Treasury Committee - Committees - UK Parliament
6 Op. cit. endnote i above