Liability-Driven Investment: Let's get physical
It is fair to say that the market volatility of 2008-09 revolutionised the way liability-driven investment (LDI) is practised in Europe. Before the financial crisis, government bonds were held as collateral to back a portfolio of interest rate swaps, the default asset for matching the liability discount rate; this achieved leverage that freed-up capital to allocate to excess return-seeking assets. The only active decisions concerned the initial amount of hedging and the ongoing steps to de-risk.
Then two things happened. Super-liquid interest rate swaps became a windfall asset, their rates plummeting through government bond yields. Many investors booked profits by rotating into the higher-yielding bonds. Secondly, credit spreads widened dramatically. Pension funds loaded-up on strategic credit assets that, once spreads had recovered, would deliver a whole new slug of duration into portfolios, arguably reducing the level of LDI hedging necessary.
A third effect of the financial crisis was less immediate. While neither interest rate swaps nor pension funds played any real part in that crisis they inevitably got caught up in the resulting wave of regulation. Basel III is set to change banks’ appetite to offer balance sheet leverage or act as derivative counterparties, and the European Market Infrastructure Regulation (EMIR) will push interest rate swaps into central clearing, with expensive new protocols around operations and collateralisation.
All of this would appear to make physical, funded cash-market assets like bonds more attractive for liability matching than unfunded derivatives like swaps. Are we seeing that – and if so, how far might it go?
There is no doubt that 2009-10 saw a significant move from swaps to bonds exposure in LDI. Practitioners say that the typical mandate in 2006 would have hedged 25-50% of matched liabilities with bonds; that is now more like 50-75%.
But, of course this is about relative value and that can reverse. This is precisely what has happened: at the long end of the GBP inflation-linked curve, for example, the negative swap spread has narrowed from a peak of 100 basis points to 20 basis points and counting.
It is also important to recognise that only a part of the bonds exposure pension funds rushed into came in the form of physical assets. Apart from positions at the ultra-long end of the curve, most was done either with total return swaps (TRS) or repurchase agreements (repo).
“The main reason for the move to bonds exposure has been the extra yield you could get relative to swaps,” as Alex Soulsby, head of LDI at F&C Investments, puts it. “The innovation has been to access that synthetically via total return swaps and repo, to maintain the capital efficiency we associate with interest rate swaps. Leveraged approaches are still fundamental for pension funds, and will be until they are sat on surpluses under relatively conservative valuations assumptions.”
At a glance
• Market volatility in 2008-09 encouraged investors to think dynamically about the role of swaps, bonds and credit-spread in LDI.
• While a lot of government bond exposure is unfunded, in the form of repo trades, the trend for long-dated illiquid credit assets can reduce the need for unfunded duration hedges.
• Incentives to deploy these strategies are implied in the use of an ultimate forward rate, in Solvency II, and in the shift to buyout portfolios in the UK.
• But the picture is complicated by the fact that most investors need at least some leverage to meet return targets – and therefore liquid risk-free assets for collateral.
• Similarly, regulation that will make derivatives more costly will actually make liquidity more essential in matching portfolios.
• Longer term, however, better funding levels may be accompanied by a growing appetite to de-risk into physical rather than derivative exposures.
The fact is that any scheme that wants to hedge more than about 70% of its interest rate and inflation risk without enjoying a funding ratio above 90% will find itself with a required return high enough to constrain the amount of leverage it can forego in favour of physical-asset hedging.
Even at PenSam, which, like many Danish pension providers, has been significantly shortening the duration of its liabilities by moving customers into products that offer greater upside potential in exchange for lower guarantees, derivatives remain an essential part of the toolkit.
“Liability hedging is still very important to us and a significant part of the balance sheet, but clearly it isn’t of the same magnitude as it has been historically,” says Morten Hemmingsen, head of equities and fixed income. “But, the extent to which one uses cash instruments or unfunded instruments is a matter of how much balance sheet leverage you want or need to run. In practice, in a Danish framework it hasn’t been possible to hedge liabilities without using swaps or repo – you simply cannot buy enough funded instruments to do the required hedging.”
Being leveraged positions, TRS and repo incur financing costs just as swaps do, but they are cheaper: with a swap, a pension fund pays LIBOR or EURIBOR, with TRS and repo it pays LIBOR or EURIBOR minus a spread. While in recent years that spread has grown quite wide, those costs are now converging again: all things considered, if anything the case for a move back into swaps exposure is improving.
How do physical credit assets fit into this? At first glance it seems obvious: they have the potential not only to provide liability-matching cash flows and offsetting duration for the portfolio, but also some deficit-closing excess return.
“Is LDI moving back towards physical assets?” asks Tom McCartan a vice president in manager research at Redington. “Yes, massively yes, and especially to those that have a foot in both the return-seeking and hedging camps, offering credit and illiquidity risk.”
John Dewey, managing director in BlackRock’s multi-asset client solutions group, agrees that deploying quasi-matching assets like infrastructure and real estate debt, social housing or long-lease real estate, has been one of the “big themes” of recent years.
But as with bonds-over-swaps, there is an argument that this only made sense when there was a clear value opportunity. With credit spreads much tighter and the GBP credit curve in particular quite flat, can you justify taking spread duration risk unless you are a true believer in these assets as “quasi-matching” – a source of interest rate and inflation risk that reduces your need to hedge with leveraged positions?
Even at Axa Investment Management, where head of LDI Jonathan Crowther talks up these “midfield” assets as a “genuine bridge” between the return-seeking and liability-matching portfolios, his colleague, solutions strategist Shajahan Alam, sounds a note of caution.
“The return will generally be lower than you could get from pure return-seeking assets, and you don’t want to drag it down so far that you push out the date at which you achieve full funding,” he says.
Why give up the ability to hold the punchier return-seeking assets that using swaps, TRS or repo gives you? Regulation might suggest some of the answer.
Use of buy-and-hold non-government assets tends to be more common in markets like the Netherlands. This is partly because the euro credit curve is more upwardly-sloping. But the introduction in Denmark and the Netherlands of the ultimate forward rate (UFR) – an extrapolation of an interest rate at the ultra-long end of the curve where liquid market prices are unavailable – might also have an effect.
“This doesn’t necessarily bias you towards corporates, but corporates might be the only way to hedge if the market swap rate you are being offered is nowhere near the UFR,” says Phil Page, client manager at Cardano. “It certainly incentivises less swap hedging, because the fixed rate beyond a certain point means 20-40% less sensitivity to rates in longer-dated liabilities.”
We see a similar dynamic in the UK, where the lack of swap liquidity beyond the 30-year point informs those discussions about infrastructure and social housing. However, for various reasons there is little evidence of the UFR really changing LDI structures – reasons that include the fact that its muddying of the return-seeking and liability-matching functions tends to crowd out higher-yielding return-seeking assets.
“We draw a clear distinction between liability-matching and return-seeking assets,” says Hemmingsen at PenSam. “And not only has there been no trend to increase the allocation to low-yielding fixed income assets that are not used for liability hedging – in fact, you could say the opposite: there is a continuing need to allocate non-liability matching assets to equities and higher yielding alternatives, simply to produce the returns we need.”
Insurance companies already do less derivatives-based hedging than pension funds because they do not take investment risk to fund deficits and compensate for that with leveraged interest-rate exposure. In addition, the introduction of the matching-adjustment concept via Solvency II recognises, in the discount-rate curve, the credit spread that long-term buy-and-hold investors earn.
How does that affect pension funds that are not regulated as insurance companies?
First of all, it may simply change the general macroprudential mood music. “For example, The Pensions Regulator in the UK has moved away from ‘Read my lips: Gilts is the long-term funding standard’ to something more like ‘Gilts plus a sensible margin, particularly if that margin reflects some illiquidity premium’,” says Rupert Brindley, managing director in the global multi-asset group at JPMorgan Asset Management.
Longer term, in the UK in particular, many pension schemes will look more like insurance companies as they approach full funding and become cash flow negative; or as they prepare their portfolios for a bulk annuity buyout.
“For some very well funded clients we have started to put in place optimised matching credit portfolios, using swaps very much at the margins,” says Andrew Giles, solutions CIO at Insight Investment. “The driver has not really been derivatives-aversion but a desire to put together a portfolio that’s easily fungible with insurers to give them more clarity around buyout pricing.”
In this situation, a pension scheme may begin to re-think where it spends its illiquidity risk budget. Currently, much of that is spent in the return-seeking portfolio but, if anything, this portfolio needs to become more liquid to enable the investor to de-risk dynamically. The real function of the super-liquid matching portfolio, by contrast, is to throw off regular cash flows or reflect the underlying portfolio of a buyout provider – which is often tilted towards illiquid credit assets.
But again, the capacity of a pension scheme to adopt this thinking is determined by its funding level, which dictates its leverage needs. This constrains the potential to hold physical assets that are ineligible to be used as collateral to fund derivatives.
Those who prefer to think in terms of separation portfolios tend to articulate this by saying that anything ineligible as collateral should not be seen as a liability-matching asset at all; practitioners that think of certain assets as “bridging the gap” say instead that investors should “think about liquidity needs through the whole portfolio”, as Alam at Axa IM puts it, “including collateral requirements for derivatives”.
“You cannot allow yourself to run out of eligible collateral and have to unwind your swap overlay,” agrees Kenny Nicoll, a director in manager research at Redington. “There is definitely a trend away from derivatives and towards more illiquid physical assets, but it can’t go all the way.”
Winds of change
But what happens if derivatives become permanently, prohibitively costly? This is the threat posed by EMIR, which aims to bring standardised derivatives into a centrally-cleared trading environment. Pension funds enjoy an exemption until August 2015, which may be extended.
Trying to stick to bilateral derivatives probably won’t work – bank counterparties will pass on the much higher Basel III capital charges they will incur for non-cleared derivatives. Betting on an extension of the exemption would be risky, too. Page at Cardano, who puts the odds at 50/50, explains that the decision will be revealed so close to the deadline that it will not leave enough time to put the necessary infrastructure in place if the exemption is lifted. “The authorities are telling you that, either way, you need to set yourself up to do it,” he says.
Hemmingsen makes a similar point about operational readiness. PenSam is building a central-clearing infrastructure because it will need it to pursue its plan to be more active with equity derivatives, anyway.
“We are getting ready for EMIR for ‘free’, so to speak, although EMIR has had an impact on the time line,” he says. “In an uncertain world, I think that basically we have to put ourselves in a position where we are able to use the full range of instruments. And if we tried to get out of our swap positions and replace them with repo we would run into other risks. Of course, if these regulations were not coming up we would not face the issues of getting ready for central clearing: but they are, and we just need to get on with it.”
Although trade reporting is proving tricky, many of EMIR’s operational protocols agree with standard best practice in existing bilateral trades. The real additional cost is associated with the requirements for higher levels of initial margin, and for cash as variation margin.
“The main driver for clients redesigning their ALM framework to allow greater use of physical assets is really the serious worry about expensive collateral requirements under EMIR,” says McCartan at Redington.
Crowther at Axa IM agrees: “EMIR is definitely going to have a large effect because pension funds have enjoyed a lot of latitude on collateral in the bilateral environment.”
Today, a standard three-month liquidity stress test recommends a collateral pool of 20-25% of the hedged liability. Under EMIR, that will be more like 40-50%.
“If you mingle together your corporate bonds, Gilts and cash as LDI, so you just don’t have as much Gilts or cash to post as additional margin, EMIR will be a much, much bigger problem than if you tend to keep them separate,” argues Page. “LDI managers that use corporate bonds have been very concerned.”
The requirement to post cash as variation margin could be even more of a headache.
“If the EMIR exemption was lifted in 2015 the costs associated with the regulation certainly become much more significant in your considerations,” says Giles at Insight. “And if you are going to incur the big opportunity cost associated with cash collateral that’s a bit of a show-stopper.”
The standard way to turn bonds into cash is repo. But, while there has been some relief that Basel III will allow banks to continue with limited repo netting, the bigger de-leveraging trend will inevitably cramp liquidity-provision activities.
“Repo is one of the areas where there could be a material change in bank risk appetite, so we are focusing a lot of attention on making sure that, if there is a contraction, clients can maintain their strategies,” says Dewey at BlackRock. “This balance sheet-intensive activity just isn’t as profitable for banks as it used to be.”
Even if repo liquidity remains available and cost-effective, there is little doubt that EMIR will make collateral management more intensive for derivative users as they go from one type of variation margining to two: repo and bilateral ISDA-based contracts like inflation swaps or CDS, alongside new centrally-cleared contracts.
“Management of your collateral pool will have to become more dynamic as you try to limit the drag from the cash you hold to meet some of your variation margin calls,” says David Rae, head of LDI solutions, EMEA at Russell Investments. “It will be more complex than simply sitting on a single pool of liquid collateral.”
The irony is that as long as under-funded schemes require ambitious returns, penal regulation like EMIR, far from discouraging the use of derivatives, might actually discourage the use of physical “midfield” assets. Any scheme in that position needs some derivative exposure, and more liquidity to back those up will constrain the amount of non-eligible matching assets you can carry.
The alternative, of course, is to achieve full funding. Even in the UK, after a year like 2013, this no longer seems like a pipe dream. Throw in a fast-evolving bulk annuity buyout market and a meaningful reduction in the amount of derivatives-based liability hedging could be on the cards sooner than many think.
“As liability profiles get shorter in duration, and trustees start to recognise the impact of collateral requirements under central clearing of swaps, we see them feeling more comfortable if they can move towards holding physical bonds and contractual cash flows,” suggests Brindley at JPMAM. “It’s at the margins, but it’s what the winds of change are indicating.”
Practitioners at work in LDI should not lose sight of the bigger picture here. In the short term, liquidity will probably become more important in matching portfolios; but in the medium term, we may see the opposite.