LDI more good than harm? LDI more harm than good?

DB liability funding levels over the past 15 years


I’ve been wondering why I’m interested in LDI. It’s not out of academic interest – I don’t think academically and I’m not a practitioner. I have no skin in the game. But LDI interests me as someone who’s interested in how we turn pots into pensions, which is my central pre-occupation and why I set up AgeWage.

This blog is about the questions I keep asking myself, as a 61 year old with 30 years of life expectancy in front of him!

Why did we give up on true Liability Driven Investment?

It seems to me that the guarantees that are offered by defined benefit schemes aren’t greatly valued by those who get them. This is evidenced by the demand for DB transfers which swap guaranteed income for a set amount of cash paid into a personal pension.

Even the limited guarantees of inflation protection in corporate DB plans aren’t valued very much. Consultants soon worked out that swapping these increases for “sexy cash” (Steve Webb’s phrase) was a no-brainer for people for whom the impact of inflation was a thing of the past. So billions have been wiped off DB liabilities and very little scrutiny paid to whether “Pension Increase Exchange” was offering long-term value or turning DB pensions into a financial irrelevance over the decades most pensioners have in retirement.

The easiest way to “de-risk” a scheme’s liabilities is to incentivise people to give them up and that’s how those advising sponsoring employers on pensions – have occupied their time over the past 15 years. While sponsors craved the stability that LDI offered their corporate balance sheets, they embraced de-risking programs as their first step to buy-out.

The next stage to buy-out is to have all interest and inflation liabilities hedged. The full cost version of LDI offers the employer a high certainty of successfully offloading the pension scheme altogether. Like I say, nobody wants these guarantees. Liability Driven investment has not been embraced as a means to meet pension scheme liabilities but as a means to get rid of guarantees that companies never wanted to offer.

In retrospect, it was when – around the turn of the century, the political imperative – post Maxwell – was to write pension promises in stone, that DB  pensions turned from an employee benefit to a “risk-management exercise”. True LDI saw its moment of greatness flicker with the Boots pension scheme, but that experiment was unwound within a couple of years , when the sponsor baulked at LDI’s expense to its P/L. Since then – true LDI has been practiced only by those with the deepest pockets, notably the Bank of England’s funded staff scheme – which sucks in over 50% of pensionable payroll as an ongoing funding rate. We have – for the most part – given up on “true LDI”.

Is Leveraged LDI – “costume jewelry”?

Instead of properly matching liabilities, the market moved on to something we might consider “costume” jewelry.

Costume jewelry has its value, it is not the real thing (like the BOE’s diamond pension scheme) but it allows schemes to scrub up well for the accountants and TPR can point to fewer schemes failing into the PPF than predicted and a healthy funding position of schemes right now (see graph at top).

Among both academics and fundamental de-riskers , Leveraged Liability Driven Investment (LLDI) is seen as doing more harm than good. John Ralfe , Con Keating and Iain Clacher can agree on that.

But on the other side of the argument are TPR, the consultants and the trustees who bought into leveraging their long dated gilts to buy more. This leverage was in line with other means of “de-risking”,  indeed it incentivised transfers and increase exchanges because it drove down discount rates which pushed up transfer values. Throughout the last decade, the nominal discount rates  in low single digits for most schemes, the  “plus” in the gilts + formulation was created by borrowing against gilts using derivatives and repo.

This worked for as long as it worked and when it came to an end, there was a loud bang as sharp rise in yields burst the balloon. Was LLDI fake or did it offer a cheap way for schemes to dress up to their accounting responsibilities to a point that we can say – LLDI is still worth it?

I puzzle over that. I can see a number – £500bn – which is Clacher and Keating’s estimate of the damage of the balloon bursting in 2021. (This needn’t of course have happened, but it’s hard to deflate a balloon)

I don’t see any number which points to the benefit to schemes of LDI between -say 2010 and the start of 2022. I suspect this could be done by taking the asset mixes of schemes before embarking on LDI ( a proxy being those who don’t use LDI) and plotting their funding positions over the 12 years. What would have been the end point if 60% of schemes had not used LDI?  What would have been the impact on sponsors ? Would there have been schemes having to join the PPF and would corporate failures have increased if LDI had not been in place?

We don’t have any academic research to show us the counter factual and without it, the benefit counterweighing the £500bn loss occasioned by the bursting of the balloon, can’t be assessed.

There are of course , other things to bear in mind. Schemes liabilities over the period 2010 to 2020 will have changed – their duration decreased as closed schemes don’t balance more people becoming pensioners by more people joining schemes as youngsters. Did the steady closure of schemes to new entrants and future accrual (which continues to this day), make LLDI essential, or was there another way?

Is there another way to LDI and LLDI?

I started this blog thinking about why I think about LDI and LLDI and hinting that it’s because I see another way of turning pots into pensions. That “other way” is of course to stop guaranteeing pensions and moving (back) to a system of best endeavours where a promise is only as good as the fund behind it. Which is essentially what CDC is.

Modelling tells us that the need for cuts in nominal pensions  (assuming prudent pension promises and the buffer of conditional indexation) is very small. CDC pensions will rarely fall in payment but they could at any time.  Most of the wins and losses will be measured by the level of increases CDC pensions can afford to pay over time.

If you release those offering pensions (however you organise them) from having to guarantee benefits, you dispense with the need for “de-risking”, you run with risk at steady state and you plan your scheme or fund to pay the pensions as they come due, without the need to link funding valuations  to a “risk-free” rate.

My interest in LDI and LLDI is in trying to understand the advantages of both. I am concluding that neither is offering much advantage as both are trying to solve a problem that need not be there.

If most people are happy to live without guaranteed second tier pensions (and most people seem to be happy – even without a DB plan) then we should be focussing on ways of turning these DC pots into pensions and running DB plans on a best endeavours basis. That may be the world I live in when I’m in my 70s and 80s.

What if DB had stayed as CDC?

What if we had decided to dispense with guarantees altogether and not closed all those DB schemes? What if our DB system had continued to invest in the open scheme sweet spot paying pensions with asset income and adjusting pensions according to the success of “best endeavours”?

I suspect that the real cause of harm for DB pension schemes in the past thirty years, is that they have shackled themselves to guarantees that they should have not taken on. Only those schemes guaranteed by the tax-payer (the public sector unfunded schemes) can really offer such guarantees (and the cost of those guarantees is at the cost of personal and corporate taxation).

If we work in the private sector, we have a different employer covenant – both at work and looking beyond work. I suspect that if we took the long view, we would answer the question in this blog’s title –

“both LDI and LLDI are ultimately harmful to pensions and to those who fund them”.



About henry tapper

Founder of the Pension PlayPen,, partner of Stella, father of Olly . I am the Pension Plowman
This entry was posted in pensions and tagged , , , , , , . Bookmark the permalink.

6 Responses to LDI more good than harm? LDI more harm than good?

  1. John Mather says:

    Have you noticed that there are still no accurate figures for the cost of LDI just the best educated guess of £500bn repeated so many times that the guess becomes truth just like other statements like “Gold Plated”.

    If they were so wonderful then there would be no need for a rescue fund

    Now a call for even more sticking plaster and bolting of stable doors

    Follow the money, the rules favour investment into Government debt

  2. SuffolkBoy says:

    I’d always assumed that the reason you’re interested in LDI is so that the rest of us don’t have to be 😎.

  3. Did I miss something in the article, or did you dismiss the buy-in/buy-out/bulk annuity route to hedging the liabilities. That would seem to be the second easiest route to de-risking (after buying out members directly).

    • henry tapper says:

      I hope I didn’t diss or dismiss it. But market capacity for buy-out/in is thought to be around £30bn, maybe 1-2% of insurable liabilities, which suggests that most schemes are not going to be able to go down this route any time soon.

  4. Jnamdoc says:

    It was of interest that when presented with an open goal at the WPC last week, Mr Fairs did not actually dispute the £500bn – rather he offered a meek retort along the lines of “well it depends when one measures such things, etc etc”.

    Back to the central theme of this blog – it really starts with the maxim to prioritise pension-in-payments above all other calls on the economy, and playing on the fear of old age poverty.

    The moral arguments runs along the lines that pensioners are less able to work to replace a loss in retirement, and so we look to ensure that (DB) pensioners have a fixed / secure source of income that grows with some measure of inflation. Fair enough. This leads to the modelling assumption to use a fixed / secure source of assets to pay those pensions, which leads to gilts and LDI. LDI with Leveraged is borne of the TPR’s drive for schemes to seek sufficiently high funding levels so as to de-risk all pensions from reliance on the sponsoring employer, with can later be shifted to buy-out when sufficiently funded. The leverage part (which creates a need to collateralise i.e. pledge pension scheme assets because banks do not take counter-party risk – they simply broker for a ‘modest’ margin between lender and borrower, if we accept use of the b word) came in with a well-intended notion to fund schemes as if, so as to be, buy-out-ready in terms of hedging against the main controllable risks as perceived by insurers – inflation and the liability measurement (using risk-free assets – gilts). But as the cost of actual gilts was prohibitively high (because of QE, which also led to a scarcity of bonds), meaning hedging had to be accessed through synthetic gilts via derivatives and repos (i.e. for no initial cost, other than increased borrowing – the liability side of the swap). The effect of this was akin to accelerating the hedging part of the buy-out solution by at least a generation, some 10-15 years (the point of significant maturity in DB technical language). A sort of “hedge now/pay later” type arrangement. The near catastrophic LDI fallout witnessed over the autumn was, just as Con Keating and Prof Iain Clacher have been trying to educate us these last couple of years, a natural consequence just waiting for timing and markets to crash into (or, out of the modelled but predictable) alignment. The leveraging of LDI simply multiplied and accelerated, bringing forward the consequences of over-borrowing (to use the b word again). You could tell it’s risky when the banks demand collateral!

    But as noted at the start of this blog response, the genesis of this is the notion to fix or hypothecate a section of our (dynamic – meaning up or down) economic output, which became encapsulated in strong regulatory powers driving private sector DB schemes on a quest to de-risk (so members’ benefits were not ‘at risk’, a mortal sin)

    To simplify, let’s consider a thought exercise at the extremes. If an entire population and economic capacity (so with no savings or investments) is comprised of 19 productive workers, and 1 non-working retired person, then the 19 can agree to support the retired person providing a fixed share from their collective output, and they can even agree to codify that into a set of rules. Roll-forward a number of years, till we have 1 productive worker and 19 retired persons. Clearly at that extreme, something about the model has to change! The issue we are faced with now is understanding where along that demographic journey we are as an economy. Demographic and longevity changes mean that there are fewer productive workers proportionately to deliver the economic output needed to provide a non-working income for an increasing number of years to be enjoyed in retirement. And the sheer enormous scale now of the DB savings pots give evidence to the proportion of current and future economic output we collectively have agreed to provide to the non-working elderly – the pension needs and share is a long way from the 1:19 ratio it may once have been.
    The problem is routed in group-think, in not re-considering, not updating the thinking behind the model for different times, and sticking with the desire to hypothecate that allocation of income in a fixed or secure way – somewhere along our collective DB journey we rolled out the de-risking model to such an extent we forgot about and have lost the critical feedback loop (ie investment) into the economy that is so crucial for generating the income to pay the pensions.
    This years’ Reith Lecture by Dr Fiona Hill is a brilliant essay and listen on the seductive and corrosive power of fear and how it can grip the psyche of a population (The Reith Lectures – The Four Freedoms – 4. Freedom from Fear – BBC Sounds). A most primal fear for us all is the fear of poverty and destitution and abandonment in old age. And it is not an unreasonable step, indeed it is perfectly laudable aim, to want to provide for ourselves lessening the burden on others in our old age and to want to do that in the most secure or risk-free way possible, and so convinced we have been on the moral basis for this that we have even allowed introduced laws to criminalise those who (with judgemental hindsight) could be judged to have put at risk (using the double negative, who lessened the de-risking, or sought to invest) members’ benefits.

    Taking from Dr Hill’s concluding comments in her December 2022 Reith Lecture, what is needed to overcome fear is courage, courage to invest in ourselves and in the younger generation, and having the belief they will create the wealth to support us in our old age. At a fiscal and societally level we make tax rules giving tax benefits to encourage behaviours we consider will benefit the whole of society – that is the reason pension contributions are tax deductible and the income in the accumulation phase until retirement is also tax free – investing is ‘risky’ and somewhere in our history we were aware that investment by pension schemes should to be encouraged. As I have commented in earlier blogs, the (£3trn) of DB savings should be treasured as a national store of wealth capable of being invested for the greater and common good, providing the drivers for economic growth we all need to pay a living wage for the elderly.

  5. Pingback: “DB savings should be treasured as a national store of wealth!” | AgeWage: Making your money work as hard as you do

Leave a Reply