This blog is written in response to an FT article entitled “A different perspective on LDI”. We quote it in its entirety and respond after each quotation.
LDI is very far from the light entertainment of an Ealing comedy.
“It is still early to talk about the crisis in the UK pension fund industry in the past tense, often a prerequisite for rational discourse. Nevertheless, I feel compelled to advance an alternative perspective on recent events.”
Far from being an ‘alternative perspective’ this article is the standard pitch of promoters of LDI – note it is concerned with the crisis in the UK DB pension fund industry and not the harm done to systemic financial stability, and indeed ignores completely the very real losses suffered by DC pension savers nearing retirement.
“Polemics against the “liability driven investment” approach by pension funds seems to have overlooked some fundamental points. The objective of a defined benefit scheme is to secure members’ benefits.”
No, the objective of a DB scheme fund is to help to secure benefits, to pay them on time and in full; the sponsor employer carries the obligation to ensure pensions are paid. The fund may be viewed as defraying the employer’s cost of producing the promised. The scheme has wider concerns.
“On this measure, such funds have never been in better shape.”
It is unstated but the only measure by which schemes might be considered in better shape is the funding ratio – the ratio of scheme assets at mark-to-market values relative to the present value of liabilities derived using a market-based discount rate, usually gilts. In fact, scheme assets will now cover around 45% of the projected ultimate liabilities compared with about 80% at the beginning of the year. It is rather more than a paradox that schemes can have lost some £500 billion of asset value and now have people suggesting that schemes are stronger than they were at the beginning of the year.
There is also a question of the degree of confidence that we may have in the funding ratio statistic. Under normal market volatility conditions, for a fully funded scheme, the one standard deviation confidence interval ranges from 96.2% to 104.1%. Under the market conditions seen recently, that confidence interval has expanded to 89.1% – 112.6%. There are very few, if any, schemes employing LDI with improvements large enough to qualify as statistically significant.
There is also no mention of the fact that the schemes which have performed the best, by some considerable distance, using this measure are precisely those not employing LDI and for the majority of those schemes, the results are statistically significant.
“In aggregate, DB pension schemes now have a funding surplus partly due to the discipline imposed by LDI. UK trustees and the Pensions Regulator have done a good job, better than almost anywhere else, in protecting the interest of members.”
LDI came about because the accounting and regulatory standards imported the level and volatility of gilt market prices into scheme valuations. LDI was designed to counter this – and it did so at considerable cost to schemes and their sponsor employers. These standards and their associated costs have been the largest contributor to the demise of DB schemes. This is the principal reason for closure to new members and future accrual, leaving us with perhaps only 11% of scheme open.
The Pensions Regulator has adopted policies which encourage such closures and indeed now is concerned with ‘the end game’, the final wind down of DB schemes. We do not consider that a ‘good job’, though they may have done it well in the sense of having succeeded in overseeing the closure of the overwhelming majority of schemes.
“It is no coincidence that countries adopting a liability focused approach (the UK and Netherlands) have among the healthiest DB plans in the world.”
Only if the measure of health is the looking glass measure of ‘scheme funding’ covered earlier. But this is to obfuscate the bigger point that schemes have suffered significant value destruction. If our estimate is correct, this is around 30% of DB pension assets, we have not yet had a chance to work out the losses in DC. Put another way, the asset management industry in the UK has just lost circa 5%-10% of all assets under management. None of the this feels like a “healthy” pension system or a stable one.
According to the FT, the newspaper in which this article appeared: “LDI strategies are used by only a tenth of Dutch pension funds — and those are mostly smaller schemes.” Are we really to believe that such a small group are capable of producing the results claimed?
“In its essence, LDI is simply about putting the liabilities of a DB scheme — paying pensioners — at the heart of its investment strategy.”
This is simply untrue. LDI strategies are predominantly concerned with eliminating the volatility of valuations arising from the market-based discount rates – they are concerned with today, and not the long-term future of the pension to members. This is obvious when one recognises that no discount rate figures in the determination of any scheme member’s pension.
In this crisis we have seen scheme’s concerns reduced to a matter of a few days or hours not the many decades of the pension promises that they are obligated to pay.
“This is a shift from traditional approaches that focus exclusively on the returns on the asset side of the balance sheet.”
It is a shift from the long-term of traditional approaches, which were not hamstrung by inappropriate introduction of market volatility, to the short and indeed sometime immediate timescales of markets.
“A scheme would hedge the inflation and interest rate risk of its liabilities using leverage to free up capital to invest in high-quality securities that pay a premium over gilts, such as corporate bonds or asset-backed securities.”
This is a description of what is being done as LDI strategies. It is being done in total disregard of the law. Both borrowing and the use of derivatives to hedge liabilities are prohibited as we have discussed extensively elsewhere. The Pensions Regulator has chosen not to enforce the law and indeed has actively encouraged schemes to adopt these unlawful strategies.
It is also worth noting that it is necessary to augment the income of the fund as borrowing through repo or the use of fixed receiver interest rate swaps pay less than the yield on gilts – they pay a long-term fixed rate but must pay short-term (usually six month) borrowing costs.
The income of the scheme is now coming from the credit spread of corporate bonds and asset-backed securities. There is no recognition that this is highly variable and represents compensation for the security specific illiquidity and both the general and idiosyncratic default risks of these securities.
Where previously schemes bought equities, now we have LDI managers who are effectively playing with options on equity, but as far as we can see without any recognition of that. The relatively lack-lustre performance of the London Stock market in the period from 2002 may be attributed in large part to the persistent and substantial selling of UK equities by pension funds in pursuit of LDI strategies. UK DB pension schemes now own less than 3% of London listed stocks. This has major consequences for governance, and the fact that the level of the FTSE has barely changed since the late 90’s has to be acknowledged as a significant failure on a number of fronts.
“This allows the scheme to generate extra income and improve its funding position with a great degree of confidence.”
This requires more than a minor leap of faith. The income for covering scheme liabilities in fact comes not from gilts but from the securities used to generate extra income. There is no sound reason for any scheme member to have any greater confidence in this rather than traditional strategies.
“Even accounting for recent events, LDI has substantially reduced the volatility of schemes’ funding statuses, improving the certainty of meeting pension promises.”
It is true that LDI has offset much of the volatility of funding ratios (derived of course under a flawed accounting practice) – a case of introduce a problem and then resolve it at great cost. It has not improved the certainty of meeting promises, and given the costs of the strategy, it has almost surely made the pension promise harder to meet.
“If DB schemes are in good health, what is the problem? The current crisis stems from a mismatch between the liquidity of the assets of funds and the collateral requirements of the hedges.”
No, it does not stem from a liquidity mismatch, the liquidity is a symptom of the disease, not its cause – that was the accounting practices which in turn gave rise to LDI. Having to post collateral is itself problematic, it not only shortens the management horizon to the immediate, but it also places a creditor ahead of pensioners.
“Disposing of assets may take days (or weeks for the more illiquid). In contrast, collateral requirements, known as margin, on the hedges must be settled daily and mostly in cash.”
This is recognition of a basic flaw in the LDI strategy.
“This is a side effect of regulations intended to bolster banks after the financial crisis.”
It was a well-known issue at the time these repo and swaps transactions were originated and was not so much to bolster banks as to avoid the prospect of a bank suffering catastrophic loss through the default of a counterparty.
“If margins could have been posted in a broad range of high-quality bonds, much of this problem would not have come to the fore.”
If they were, this would resurrect the problems and disputes over the valuation of securities that we saw at the time of the GFC. The Bank of England’s collateral haircuts table is a fair indication of the value of securities as substitutes for cash, and they range from 0.5% up to 42% of the market price.
“To manage the liquidity mismatch, pension schemes hold some of their assets in a liquidity buffer.”
This, of course, has a cost to the fund, and indeed the fund manager since it impairs return performance and that provides an incentive for the manager to minimise this. The currently discussed buffers as being of sufficient size to withstand a 3% instantaneous rise in interest rates would have approximately one third of a scheme’s assets held in these near cash instruments. A very substantial cost for the fund to bear.
In most applications, the liquidity buffer is a mixture of bonds and cash, and the bonds require a market in which to liquidate them – for all schemes employing LDI. This seems to be even greater reliance than was seen at the beginning of the recent crisis.
“This is designed to bridge the time involved in converting assets into cash.”
If these are to bridge a gap, then the assets being sold are in fact to replenish a now depleted cash element of the buffer. The buffer realisation is dependent upon the smooth and continuous functioning of markets but should be expected to be called upon precisely when those markets are under strain and unlikely to possess those characteristics.
“However, if the buffer is eroded too fast and schemes need to restore balance, they must sell assets or reduce their hedges. The latter involves selling the gilts linked to the hedges.”
This is an acknowledgement that the endogenous market-disrupting cycle is a property of the strategy.
“Regardless of how conservatively the buffers were calibrated, it was impossible to deal with the speed and the magnitude of the recent sell-off in gilts.”
This fails to mention that the speed and magnitude of the recent sell-off was the collective result of those employing the LDI strategies and some of the considerable leverage that some of these strategies entailed – the strategy is self-destroying.
“The longest gilt with returns linked to inflation (one that matures in 2073) fell from a price of just over 115 to 50 in a few trading days. That is a performance reminiscent of bitcoin, not a flagship security of the UK.”
It is worth noting that the actions of pension funds in bond markets have long distorted their price function. If we take the 1/8th % ILG 2068, we see a high-low yield range from RPI – 213 basis points to RPI +212 basis points – a range of 4.25% in yield. The equivalent conventional gilt, proxied by the 40-year spot rate, has varied by just 214 basis points over the same period. Pension funds and their LDI strategies dominate the ILG market; this is the result of their actions. This point is also to suggest that LDI and the extent to which pension funds were holding ILGs were hit by some exogenous and unpredictable shock, rather than having embedded a risk into the system.
“The depletion of these buffers led to selling of gilts in what was already a dysfunctional market creating a self-feeding loop, as price falls triggered more sales.”
The market was febrile prior to the mini-budget; interest rates have been rising since the beginning of the year. In fact, there were some reports of distressed selling by pension funds in March/April of this year, at which time these markets might reasonably have been described as being briefly dysfunctional.
“It needed the Bank of England to break the cycle and stabilise the market — providing time for these buffers to be replenished and strengthened.”
It is true that Bank of England intervention was needed to break the LDI cycle and that prompts the obvious question: what sort of investment strategy is it that depends upon external and costly intervention for its viability?
“While the extreme stress created by the fiscal crisis has exposed a vulnerability, it is not all doom and gloom as has been portrayed. In fact, the problem is eminently solvable.”
The stress which triggered the LDI ‘doom-loop’ was in fact rather modest. The 15-year conventional gilt fell just 2.5% in price on the day of the mini-budget, and this was a market absorbing the prior day’s increase in bank rate, which may have led to some pressure from LDI strategies. On that Friday, the 15-year conventional fell by just 3% and there were reports of liquidation of gilts by pension funds.
It will be most interesting to hear the solution to this problem.
“The strong funding position of schemes means that the returns required to achieve full funding has declined significantly.”
This simply is not true. The value of assets has fallen significantly. That means that they must earn higher returns in order to deliver the pensions promised.
“This allows further de-risking, reducing partially funded hedges and exposure to risk assets.”
This sounds eerily reminiscent of the promises made to trustees by the promoters of LDI. Reducing partially funded seems to be a reference to reducing borrowing…so if not more of the same, more of the similar. .
“To prevent fire sale of good assets in the meantime, facilities in the market to transform collateral need to be put in place.”
This statement is one that is being heard more and more but misses a key point – a such facilities already exist; they are known as markets. The Bank of England also publishes a summary table of ‘haircuts’ for eligible collateral as part of its monetary policy operations, which provide a guide to the relative of assets as collateral; they vary from 0.5% to 42% of market price. It is LDI strategies which, collectively, have disrupted those
“This is happening organically for high-quality bonds, but we need to ensure that adequate capacity exists.”
We are unsure what this sentence refers to.
“However, as far as illiquid investments are concerned, these facilities are not yet in place.”
This is special pleading. It can be rephrased as:
‘We want a facility created that transforms the illiquid securities, we knowingly bought, into liquid securities.’
As liquidity has a cost, it is interesting that there is no mention of any price for this service. The moral hazard implicit in this and its impact on price discovery and the efficiency of markets is also considerable.
“In the short term, it is essential to reinforce the resilience of the system: schemes’ governance models need to be enhanced for faster decision-making and cash buffers need to be increased especially for pooled funds.”
It is only too likely that faster decision-making is merely code for giving fund managers powers to act, that is make free with other scheme assets, without trustee intervention. The Pensions Regulator has also promoted this in its most recent missive on scheme funding.
“LDI has been a force for good for nearly two decades.”
No, this is untrue. For two decades LDI had been operating as a positive rather than negative feed-back loop, which has driven yields to ludicrous levels on occasions. This is most evident in the index linked gilt market where expected rates of return as low as RPI -320 basis points were seen. The index linked market is dominated by pension funds.
“It has focused attention on what matters most, namely funding the retirement of its members.”
It has not. It has focussed attention on a flawed measure, the funding ratio. Funding for the payments of retirement benefits has fallen by hundreds of billions of pounds this year, the coverage has fallen from around 80% of those benefits to around 45%.
“It has reduced risks embedded in liabilities and imposed a discipline to the way assets are managed.”
It is not possible to reduce the risks in a liability without the explicit agreement of the owner of the benefit that liability represents. The activities that constitute LDI do not operate on liabilities; they cannot. The most that might be said is that they operate in respect of risks perceived as arising from liabilities.
“The result has been improvements in the funding status and funding surpluses, measured on a conservative basis, for the first time.”
Using a measure as badly flawed as the funding ratio is far from conservative, it is reckless. We have seen the damage it may do; the higher mortgage costs and now impaired DC savings of many millions are just two of the more visible harms. It is also worth noting that the yield on gilts is around where it was prior to the LDI crisis taking hold, the extent to which the “funding” of schemes is as healthy as many are saying seems like quite a debatable proposition. As Pink Floyd had it: “I’m alright, Jack, keep your hands off of my stack”
Right of reply
If the author of the FT article discussed, would like to exercise her/his right to reply, please contact me email@example.com
 As we have noted elsewhere, a pension liability is a function of the promise made e.g., ½ final salary, wage growth, inflation, and longevity. There is no interest rate risk in this. Interest rate risk is only introduced by the discounting of future pension obligations into a present value using a market rate.