Iain Clacher and Con Keating say the costs of LDI strategies amounted to £175bn by the end of September alone
The Office for National Statistics has just published its Financial Survey of Pension Schemes (FPFS) as of 30 September 2022. At the end of December 2021, the ONS reported net pension assets of £1,821bn, which was essentially the same as the PPF 7800 Index’s value for scheme assets of £1,818bn, and our own estimate of £1,810bn.
At the end of September 2022, the PPF 7800 index reported scheme assets of £1,451bn, a decline of £367bn (20.2%). The latest figures from the ONS reports scheme assets of £1,276bn, a decline of £545bn (30%). Our own estimate to the end of September 2022 was £500bn, which we made shortly after the Bank of England intervention had ended. This number was roundly challenged as being far too high.
To put these declines in context, the conventional gilt market fell by 27.3% in market value to £1,295bn, and the index linked gilt market declined by 28.4% to £619bn over this period. With PPF liabilities estimated to be £1,076 billion, the PPF funding ratio of 135% declines to 118.6% under the ONS data. It would appear that buyout still remains much further away for most schemes than much of the trade press would have us believe (ignoring the capacity of insurance sector to absorb this).
The difference between the PPF asset estimate and that of the ONS, £175 billion, may be interpreted as a direct cost of liability-driven investment (LDI) from the beginning of the year until the end of September arising from derivatives, repo leverage, and portfolio rebalancing. If we compare the performance of private sector schemes with public sector schemes, given that public sector schemes largely eschewed LDI, this gives some insight into the cost of maintaining LDI strategies. Public sector schemes are reported as holding just 6.4% of their assets in long dated bonds, of which 3.5% is in gilts. According to the latest FSPS, public sector schemes lost just 9% of their December value 2021 over the same period. had private sector schemes followed their example, they would have lost just £163 billion, and not the £545 billion reported by the ONS.
Gilt markets and portfolio rebalancing
It would seem that most rebalancing of LDI funds, and anecdotally of segregated accounts, took place after the September month-end but before the end of the Bank of England intervention in the gilt market on the 14 October 2022. One indication of this is the extremely low take-up of the Bank of England's extended repo facility, which was available after the end of the gilt market intervention.
In this nine-month period, the holdings of conventional gilts in own and segregated accounts fell by 29.6% to £126bn suggesting that very few conventional gilts were sold (on balance only about £4bn). The holdings of index linked gilts declined by 29% to £264bn. That decline is in line with the overall market movement for index linked gilts, and would imply no net selling by schemes in this period.
While we do know that in its intervention the Bank of England bought £19.3bn of gilts, of which £12.1bn were long-dated conventionals, and £7.2bn were index linked gilts, less than £5bn, all long-dated conventionals, had been bought by the close of 30 September.
The declines in the value of gilt holdings from the beginning of the year, at 18.3% for the 15-25 year maturity bucket and 33.6% for over 25 years, are actually less than the declines in value due to interest rate movements, implying that the pension fund sector had been a net buyer of long-dated conventionals over this nine month period. It appears that in the June to September quarter, schemes were net buyers of around £10 billion of ultra-long conventionals. We have insufficient information to determine whether schemes were net buyers or sellers over the period for the 15-25 year bucket - more information would be needed on the precise maturity structure of scheme holdings.
Pooled LDI, repo, and swaps
Pooled LDI funds declined in value from £230bn to £127bn. On the face of this, the 45% decline implies leverage of only about 1.7 times. However, this does not account for additional contributions, the recapitalisations made over the nine months from the beginning of the year. If the leverage was actually 3.5 - 4.0 times as was widely reported, those additional contributions would account for all of the £127bn net asset value of funds reported by the ONS at end of September.
As well as this, the small increase in the market value of repo from £192bn to £202bn masks a serious increase in the notional amount of bonds under repo. If there were £192bn nominal at the beginning of the year, it would be around £340 billion at the end of September. This highlights one of the issues with the use of gilt collateral rather than cash - further demands for extra top-up collateral.
The net position of Interest rate swaps, an element not captured by the PPF asset algorithm but captured in the FSPS is shown in the table below. The balance has moved from £11 billion in surplus to £30 in deficit, a net loss amd collateral call of £41bn.
From the data, it would also appear that schemes were net sellers of around £18bn of corporate bonds.
Net swap gains and losses
Source: Financial Survey of Pension Schemes, Office for National Statistics, March 2023.
Net Statistics vs Activity
These ONS figures, as aggregates, show the net behaviour of schemes - where things ended up. However, the gilt market turnover figures suggest quite exceptional activity which is masked by those net changes.
Gilt edged market makers' weekly turnover (2 September 2022 to 28 October 2022)
Finally, and worryingly, there is one reported statistic that we believe deserves closer scrutiny, is unquoted private equity and alternatives are reported to have increased in value from £79bn to £87bn, 10%. These valuations need further investigation and explanation as the flows of new money cannot account for this aberrant behaviour.
Final thoughts
Unfortunately, we may never have a full picture of or accounting for what took place, but the end December 2022 ONS figures will likely give us a close approximation of the total final cost of LDI strategies, over and above cash securities price and yield movements, which had already amounted to £175bn by end September.
Perhaps the most troubling aspect of all of this sits outside of the world of pensions. How such relatively small imbalances, a feature which is also evident in the Bank of England's December Financial Stability Report, could have overwhelmed the gilt market and resulted in such cataclysmic price movements is extremely concerning. We would suggest that a detailed study of the gilt market microstructure is urgently needed and that steps should be taken to improve its resilience.
By way of ending, we cannot help but wonder if the gains from leveraged LDI achieved by schemes under a regime of falling interest rates over the past decades have not been more than fully offset by the losses incurred in just 9 months.
Iain Clacher is professor of pensions & finance and head of the Centre for Financial Technology and Innovation at the Leeds University Business School. Con Keating is head of research at Brighton Rock Group.
