There are several ways to make LDI implementation smarter, but practitioners differ significantly over whether or not these are tactical moves - and the extent to which they should deviate from the strategic de-risking journey plan. Martin Steward reports

"Every risk you take should be deliberate, diversified, scaled and flexible," says Robert Hayes, head of client strategy EMEA-Pacific for BlackRock's Multi Asset Client Solutions group. The key risk for pension schemes are funding levels, so let's consider what events might lead those to drop, say, 20%. Even with riskily-invested 60/40 funds, equity markets would have to crash by 30%. But long-term interest rates or inflation would only have to change by 0.7% to cause the same damage. The un-hedged scheme is taking a big bet; all that work on its cutting-edge portfolio diversification is swamped by the economic sensitivities of its liabilities.

Reducing those risks is clearly sound strategy. But while hedging 100% with swaps and transforming the asset portfolio into an efficient LIBOR-plus machine might seem simple, in reality it is an impractical blunt instrument for most schemes. Practicalities force schemes to compromise, so why not compromise in the smartest ways possible? If they take active risk against benchmarks with other assets, why not also with liability-hedging assets? If they have views on rates, why not express them via liability hedges as well as via equity and bond portfolios?

"As risk-reducing becomes a more significant part of portfolio allocation, it will become just as valuable to spend time and governance budget refining the hedge portfolio as on refining strategic asset allocation," says Shalin Bhagwan from the structured solutions group at Legal & General Investment Management.

Relative value
The first tactical decision - over which instruments to use to hedge liability risks - is the least controversial because it is a basic relative-value decision (between sovereign and corporate bonds, bond futures, swaps, swaptions and inflation caps). Indeed, some decisions are so structural as to be strategic: if you are looking to hedge real rates and inflation-linked bonds look expensive, you are likely to split up your interest rate and inflation hedges simply because separate swaps tend to be cheaper than real-rate swaps. More tactically, it helps in taking separate timing decisions for hedging the two risks - many would argue that with both rates at their current low levels, buying inflation makes a lot more sense than buying rates.

Since swap spreads turned negative at the longer end of curves early in 2009, some schemes coming into the financial crisis with swaps in place have unwound them to buy the illiquidity premium in the cash bond markets or, if they wished to preserve their leveraged overlay structure, switched to bond futures or repo-funded bond positions.
These vertical relative-value decisions may be different at different points on the curve. By extension, there are lateral opportunities across the curve: it is possible to overweight your hedge at cheaper points - either in such a way as to maintain benchmark duration, or to take bets against it.

"We may find that the cost of hedging 50-year inflation is 3.5% and the cost of hedging 30-year inflation is 3.8%," observes Andrew Giles, solutions group CIO at Insight Investment. "Typically the reason for the curve inversion is supply-demand dynamics, which causes us to move some of the hedge at 30 years, which we would put in place in a perfect world, out towards the 50-year point to achieve better value. Then, as the curve reverts, we roll back to the perfect hedge position, or to the new best-value curve. Once the liability modelling team has translated the client's liability cash flows into a minimum risk portfolio representing the best hedge we can build, we have a team in the LDI business called the Market Strategy Group that figures out where it makes sense to move away from that minimum risk portfolio to take advantage of market distortions. Typically, we would have a risk budget: the client can say that we can deviate from the minimum risk portfolio to the extent of, say, 10% of the overall or 10% of the inflation PV01 or the interest rate PV01."

Before a scheme makes these relative-value decisions, it needs to settle on its hedge ratio or delta - how much of the liability risk does it want to match? The strategic case for hedging 100%, or at least 70-80%, is strong. For a scheme with a weak sponsor or covenant it seems like a no-brainer: "If a scheme that's 15-times larger than its sponsor sneezes, that sponsor catches pneumonia," as Hayes puts it. "It can't afford to take mark-to-market risk against its liabilities." But it is pretty strong even when the sponsor is considered robust: Kambiz Deljouie, liability and multi-asset solutions director at Aviva Investors, observes that this sponsor will only default on its promise if there is broad market failure - precisely the point at which all of the scheme's growth assets will correlate on the downside and active risk against liabilities will be punished. "In a perfect world, everyone would want to hedge their liabilities," he reasons. "But that sends the cost of hedging up - and then the tactical side of it comes in. Tactical positioning is inconsistent with the needs of the market, but reality is what it is."

When a weak sponsor ill-disposed to further cash contributions is combined with a funding deficit that can only be plugged by taking active risk, the structural costs of hedging (interest rate hedging will usually generate carry, but matching inflation for 30 years costs an insurance premium of perhaps 20%) are compounded by opportunity cost. A balance has to be struck between the strategic objective of an 80% delta and the tactics of timing your journey towards 80% to maximise the chances of closing the funding gap. That can be done systematically. As the market rewards your unhedged position - because your risk assets have a good run or your discount rate has risen (or both) - you can hedge a portion of your liabilities and lock in your newly-improved funding ratio. It's a bit like trading the gamma of your liability hedge. "You are constantly moving your downside risk level up without giving up the upside potential you need to get back to full-funding," says Joe Moody, global head of LDI at State Street Global Advisors, which calls it "dynamic de-risking".

Is ‘dynamic' synonymous with ‘tactical'? None of this involves a tactical market call. Indeed, as Redington founding partner Robert Gardner notes: "Even with just 30% of your risk off the table, you still benefit in a higher-yield environment, and [at that point], even if real yields look incredibly low, a pension fund should still de-risk because the funding level has increased from 75% to 85%, and you can bank that." This is more about governance than gambling. "The important thing - whether you have a view on where rates are going or not - is to have a plan in place ready for when they reach your target level," says Pierre Couture, senior actuary for ING Investment Management Americas. Sorca Kelly-Scholte, managing director in consulting and advisory services EMEA with Russell Investments, agrees: "We're working with a number of clients around building trigger mechanisms purely around interest rate hedging," she says. "Liquidity can dry up quickly. You can't afford to find yourself thinking: ‘Ah, interest rates look quite attractive, now what do I do?'"

Most LDI managers and consultants are implementing these kinds of dynamic or trigger-based journey plans. Some of the more adventurous are adding an extra systematic layer with the use of swaptions (and because these pay a premium while you wait for rates to hit your strike price, they pay for you to shave a couple of basis points off that strike price). "We see less of the situation where someone has put a hedge in place with an active manager and asked him to add value to the benchmark by playing duration long or short," says Crispin Lace, principal and senior investment consultant with Mercer.

"I'd suggest that, once LDI is in place, you wouldn't want to phase it out again," agrees Kelly-Scholte. "Lots of clients say that they don't want LDI now because of where rates are - and I sympathise with that. But then I ask them to conduct a thought experiment: if you were hedged right now, would you be taking that hedge off? Usually the answer is no."

But in your asset portfolio, you may well want tactically to underweight equity risk at certain points in the business cycle. Why should strategy always trump tactical market views on the liabilities side?

As Guus Boender, chairman of Ortec Finance, observes, wherever you have a true sensitivity to real rates but an accounting sensitivity to nominal rates, you are faced with an awkward dilemma (one also highlighted by Wim Barentsen of APG in his Ahead of the Curve piece on page 70 this month). "Minimising interest rate risk might increase your exposure to inflation risk, and vice versa," he says. "This is the dilemma of almost all of our clients. They would like to manage real interest rate risk, but if you have a funding ratio of 101%, in nominal terms, you cannot afford to increase your nominal interest rate risk because it could push you under the 100% nominal funding ratio. We have given a lot of thought to the question of whether LDI is strategic or tactical, because the worst-case scenario for Dutch funds is that we move into a period of deflation followed by a period of high inflation: suppose a pension fund does nothing on interest rates, anticipating a period of inflation, they can get killed if we see a fall into Japan-style interest rates. But suppose they completely close the interest rate risk - then they get killed if we see rates of 10% because governments decide that inflation is the only way to get rid of their debts."

From that perspective, perhaps the only action that makes sense is dialling your hedge delta up as high as you can afford, across all liability risks - and not to try and gamble on rates or inflation at all. And even if we do sanction some tactical risk-taking (as opposed to risk-taking forced by the affordability of hedging), we should bear in mind Hayes's dictum about those risks being ‘scaled'. Few schemes would move tactically between a zero and 100% equities asset allocation and, as we've seen, most have funding levels that are far more sensitive to moves in rates and inflation than to equity markets. Scale the risk budget for your hedge ratio to the risk budget of your tactical asset allocation, and it severely limits any tactical range around that strategic hedge ratio. "I can see some of our clients varying their hedge ratio by 15% or 20%," says Hayes, "but going from 100% or 70% to zero seems extreme to me."

Bhagwan at LGIM agrees. "We are having conversations with clients about varying the hedge ratio," he says. "But we are certainly not going from 40% to zero. I haven't come across clients comfortable making that kind of brave, binary call."

Well, Ben Clissold, head of risk management solutions at PSolve Asset Solutions, has. The majority of his clients unwound interest rate hedges at the beginning of 2009 - some going from 100% to less than 50%. Antony Barker from JLT's Pension Capital Strategies group goes even further, acknowledging that his is a controversial view not only among other consultant groups but even among some colleagues: "We've gone very short duration, and if a scheme had a hedging programme in place, particularly if the sponsor is supportive, we've taken that hedge off completely," he says. "We've taken the view that even if equities and bonds go nowhere, we're likely to win on the liabilities side."

As we have noted already, this is really a question of diversification. Anyone who took a ‘tactical' 100% equities position with their assets would be regarded as abandoning a sensible, diversified strategy. We should think of the economic sensitivities of our liabilities in that context. "Diversification is about controlling asset volatility but doesn't necessarily significantly reduce risk versus liabilities," says Charles Marandu, director of European advice for institutional business at SEI. "Trustees need to understand the structure of their liabilities and interaction with assets when building strategies to mitigate risk." (Although see Con Keating's views, page 48, for a different perspective).

The economic sensitivities of growth assets can correlate with liabilities. Equity generally has a low correlation with 10-year yields (about 0.20), but it has been known to swing up as high as 0.60-0.80 as recessions take hold. Similarly, while corporate bonds are generally regarded as a duration asset, in times of credit expansion or contraction spread volatility becomes a much more significant driver of performance (and negative correlation with equity decreases). If a corporate bond yield is being used to discount liabilities, this could also cause liabilities to become more correlated with equity. In short, at some points through the business and credit cycles, growth assets exhibit enhanced hedging utility against mark-to-market liabilities. Likewise, liability-risk hedges could be eroding some of the diversification benefits of growth assets.
"The first time this discussion about correlations really emerged was when we were talking with a big pension fund in Connecticut," says LGIM's Bhagwan. "They were concerned that not enough time was being spent trying to understand that risk."

One person who has given it a lot of thought is John Fox, head of fixed income factor modelling at MSCI Barra, which sees a niche for its factor attribution models in the LDI space to address these issues. Taking an example of a $25bn fund, invested 60/40, with liabilities of $19bn discounted with a corporate bond yield, he finds that although adding interest rate swaps increases the volatility of the asset portfolio by almost 14% (by introducing net duration and credit spread risks), it reduces funding level volatility by almost 20% (by reducing duration risk almost tenfold). However, when a CDS position is added on top of the swaps, not only does asset portfolio risk go up by a further 20% (as credit risk increases), the funding level volatility also goes up by more than 13%. Although the CDS has cut the contribution of credit risk to funding level volatility fourfold, the effect is more than cancelled out by correlation between the liabilities and the asset portfolio. "The asset portfolio was already hedging the change in the value of the liabilities when credit spreads close," says Fox. "We just added another source of risk to the total surplus risk."

These are just average correlations. At certain points in the credit cycle, those correlations are likely to spike. Should we, therefore, tactically asset allocate to minimise these overlaps, and if so, should those tactics be focused on the growth assets, the matching assets (and therefore the hedge ratio), or both?

The actuarial profession gives that idea short shrift, pointing out that equity is not an explicit liability-matching asset. One can simply point to the fact that equity markets discount on a time horizon of about 18 months - not 20 years. Even corporate bond curves don't stretch much further out than 10 years. "While in theory these assets triangulate, you are looking at very different timeframes and therefore huge anomalies," says JLT's Barker. "The derivatives market is there partly to correct those anomalies, after all." Jeroen van Bezooijen, product manager responsible for LDI strategies at PIMCO, is similarly sceptical: "Credit spreads clearly play a part, but through most normal cycles durations only move away from one another by perhaps half a year - and that's manageable."

For Gardner at Redington it is all about cash flows. He contrasts an equity with a corporate bond: "Even if spreads have blown out on that bond, if it doesn't default and you held on to it, it's continued to give you matching cash flows and the strategic rationale - to hedge liabilities with cash flows - remains in place even though you've taken a mark-to-market loss."

Alongside bonds, he suggests that pension schemes consider ‘Third Way' assets, hybrids of growth and matching characteristics with inflation-linked cash flows offering a yield pick-up - infrastructure debt, secured leases, equity release mortgages. The fact that the extra yield is an illiquidity premium reinforces that these are strategic liability-matching assets, unsuitable for tactical allocation.

Still, there are surely good reasons at least to monitor changing mark-to-market correlations between liabilities and growth assets, if not to respond with tactical re-weighting. First of all, if they affect funding level volatility as much as Fox's numbers suggest, then that is a serious balance sheet issue. More importantly, these correlations - equities falling, credit spreads widening and interest rates falling - tend to go up most sharply during severe credit contractions: you can end up selling growth assets at the bottom of the market to fund collateral calls for your swaps or, worse still, your sponsor could be one of those stressed credits.

So while there is significant disagreement over the extent to which short duration or inflation bets should be allowed to dominate scheme risk, there is greater consensus against varying the hedge tactically in response to changing correlations between liabilities and assets. Those who favour properly-scaled bets on duration and inflation via the hedge show consistency with their view that assets should be strategically diversified. But while many encourage scenario testing, and some recommend tactical downside-risk protection overlays for the growth portfolio to mitigate correlation risk, do they show consistency with their pro-diversification positions when they stop short of varying the hedge delta for the same function? It is a difficult question, still to be fully interrogated.