Liability-Driven Investing: Best foot forward
As short rates collapsed and curves steepened over the past three years, pension funds have become more attuned to forward rates when deciding to de-risk or to take active risk with their LDI portfolios. Martin Steward finds out why
For most investors, 2011 was all about getting the big call right: if you sold duration in the spring, you regretted it bitterly when falling yields made history in late summer. But if you found some reason to hang on to bonds, you ended the year looking like a genius.
The bond portfolio managers at Ignis Asset Management launched their Absolute Return Government Bond fund in April - just in time to make the perfect call on the market. So what was their reason for clinging on to UK gilts?
"Everyone was convinced that government bonds were in a bubble," recalls Stuart Thomson, Ignis's chief economist and co-manager of the fund. "We were seen as complete madmen when we said, ‘Bonds are not in a bubble - it's only base rates that are at their lowest level since 1694'."
Looking at spot yield curves - which show the rates offered for lending out to various maturities starting today - it seemed that all rates, nominal, inflation-linked, euro, sterling, dollar, bond or swap - were at all-time lows. But remember, every spot rate beyond, say, five years, includes those ultra-low one, two, three, four and five-year rates, and is dragged down by them. So what if we stripped them out? What if, instead of looking at the 20-year spot rate, we decided to look at the rate offered to lend for 15 years, but starting in five years' time?
This is a forward interest rate - in this case, the ‘5Y15Y' - and it represents the rate required on the re-invested capital and proceeds of the first five years of lending, for the period beginning in year five and ending in year 20, to deliver the same return as the 20-year spot rate. Figure 1 shows how the picture looked for spot rates versus one-year and five-year forwards in euro interest rate swaps in May 2011. While the 20-year spot swap rate was 3.96%, the 5Y15Y forward-starting swap rate was 4.25%. Stripping out the first five years gives us 29 basis points of ‘extra' yield on a swap with the same maturity date. The 5Y1Y forward was offering 69 basis points over the six-year spot. Even stripping out just the first year of rates had an effect; the 1Y1Y forward - lending for one year, starting in one year - was 41 basis points higher than two-year spot.
This is what Thompson means when he says that only base rates were at all-time lows. As those short-end rates were slashed, the curve steepened sharply, and the steeper the spot curve, the higher forwards go - even as the entire spot curve is dragged lower.
"Medium and longer-dated forwards had actually risen since 2008," says Thompson. "By the time we got to April we saw mind-boggling carry on the longer-dated forwards, which were safe as long as there were no short-rate cuts. Our big outperformance in 2011 came when those forward rates dropped sharply in the summer."
Because it saw higher rates in the forward curves, Ignis felt confident holding on to longer-dated bonds as long as it was also selling shorter-dated cash flows: and those positions helped it generate a 5.74% return in its first nine months - or 7.7% annualised.
While forward rates are hardly news to fixed-income professionals, Thompson credits Ignis's ‘ClearCurve' model - which not only derives forward curves from the UK gilt market but also generates the most efficient ‘packages' of trades to maximise exposure to the most keenly valued forward rates - for clarifying the portfolio managers' ideas and making them tradable.
Developed by Russ Oxley, the process now informs the investment for more than £20bn of Ignis's assets under management. But it is also used to manage another £25bn of liability-matching assets for Ignis's insurance-company parent, and, since 2011, a £1.5bn LDI mandate for the Ignis Asset Management pension scheme. Laura Brown, head of Ignis Solutions, is busy showing the process to other UK pension schemes.
So how are these forward rates relevant to a pension fund? Just as Thompson felt confident holding on to bonds despite such low spot rates, so forward rates might help pension funds overcome their distaste for buying apparently ‘expensive' liability hedges. Many have put off de-risking as spot rates have fallen further and further below the trigger rates they set before the financial crisis - leaving them behind their flightplan schedules and exposed to those rates.
"Pension funds that have been looking at average spot rates to find a trigger for their hedging missed a golden opportunity to hedge last spring," says Thompson. "Why not disaggregate the forwards and focus on the segment of the curve that a pension fund is really interested in? Those are the trigger rates you should be managing your de-risking against."
Given the steepness of the curve and the increasing imperative to de-risk, most consultants and LDI managers have broached this issue with their pension fund clients. Head of LDI product research Marc Pautz says that Mercer has been citing five-year forwards against client's triggers. At Insight Investment they call this the ‘forward rate framework', and solutions CIO Andrew Giles says that his team maintains an entire matrix, live, of every forward rate in the market.
Reinhold Hafner, CEO of Risklab, adds that, because forward rates tend to move more idiosyncratically than spot rates, they also enable funds to weigh up the relative value of different trigger rates with the same maturity: the 2Y18Y might not only be cheaper than the 20-year spot, but also meaningfully cheaper than the 1Y19Y, for example.
And Gary Knapp, managing director for insurance & liability-driven strategies with Pramerica, offers a kind of whole-scheme forward-rates play: if you really don't like short rates, why not get your sponsor to sell five-year corporate bonds, use the cash for a pension contribution, and buy 20-year bonds with it for the carry? "I haven't seen that in Europe, but sponsors in the US are definitely looking into it," he says.
"There is a big message here," says Yves Yosseaume, a principal with Aon Hewitt. "Rates are higher than you think. If you set a trigger at 3.5% and you only looked at the spot rate, you wouldn't be hedging today. But if we expect rates to rise to the level we need in five years, why not buy that now? Spot rates are dangerous - you could miss opportunities to hedge by ignoring what the market is telling you."
Back in May 2011, a euro-based pension fund with a 4% trigger would not have de-risked based on the swap-spot curve. But the one-year forward curve offered 4%-plus from 1Y14Y through 1Y19Y and five-year forwards would have triggered from 5Y3Y through 5Y20Y. As figure 2 shows, by the end of March 2012, there were no swap rates above 4%. The best I could find was the 11Y1Y at 3.35% - but that still beats 12-year spot by 88 basis points.
But before we get carried away, let's remember that those 88 basis points are only the return the market estimates you will need as compensation for selling away those first 11 years, ultimately to deliver the same return as 12-year spot.
"This is not alchemy," as Robert Gall, head of marketing strategy in Insight's financial solutions group, puts it. "You are not creating bigger rates. You are simply looking at them differently." As Pete Drewienkiewicz, head of manager research a Redington, describes it: "You're creating a more visually appealing rate of 4% so that a client doesn't say ‘I can't believe I'm hedging at 3.8%'."
All other things being equal, this is emphatically an active risk. Stripping short rates out of your hedging portfolio doesn't strip them out of your liabilities - you will be under-hedged in the parts of the curve where you choose to sell rates. Of course, if you sell the first five years of rates, you can offset the resulting interest-rate sensitivity mismatch by buying a multiple of the notional exposure required for the six-year point (and every other point thereafter) - but that is merely trading a cash-flow match for a rate sensitivity match in order to get the forward rate.
So, beyond making the hedging decision more "visually appealing", why would anyone want to take that forward rate? As Drewienkiewicz observes, there is no new information to suggest that the five-year rate, for example, is actually too low, so exposing a fund in this way is effectively a bet that rates out to five years will rise further and faster than what is implied in the current forward rates. "If you are taking that view, fine. But you must be aware that that is what you are doing," he warns.
So there is a decision to be made. After all, just because you use ‘visually appealing' forward rates to help you take your de-risking decision, it doesn't follow that you must implement the resulting hedges as forward-rate positions.
"Where clients have a view that rates at the front end are due to rise faster than what is priced into the market, that points to a forward position," says Pautz at Mercer. "If they think that cash rates will increase in line with what is implied by the forward-rate curve, then they would trade the spot rate and simply accept that for the first five years they will pay a low financing cost for a low fixed leg. It's difficult to have a strong view that that part of the curve is either over or underpriced. Mercer's view is that it actually represents fair value."
It's important not to miss the wood for the trees here. For starters, while short rates might not be ‘too low', there is no escaping the fact that they are low in an absolute sense - and they would have to fall substantially below zero to hurt you if you sold them. Even more importantly, if showing clients the forward-rate matrix is about persuading them to hedge, any active risk against liabilities they take on with a forward is insignificant compared with the active risk associated with not hedging at all. "The financial implications of not hedging the second-year key rate is much less significant versus the financial implications of choosing not to hedge at the back end of the curve, where the duration sensitivity is," as SEI's Charles Marandu puts it.
And there are other reasons why you might choose to implement a forward position. Brown notes that many funds have set similar de-risking triggers: as spot rates rise they all rush to lock-in rates together - exerting a counterproductive downward pressure. A broader range of rates should help diversify demand along the curve. "It could enable some pension schemes to gain significant advantage in the market as they will be acting at different times to the herd," she adds.
So far we have only looked at the decision to hedge or not hedge. What information do forward curves bring to the active management of a hedge portfolio that is in place?
As we have discussed, the decision to implement forward-rate positions that strip out rates at the front-end of the curve should be informed by a view that those rates are ‘artificially' low and set to rise faster than the forward curves suggest. Why would anyone assume that? Right now, the argument would be that central banks are extraordinarily active buyers. Do forwards reveal similar supply-and-demand-related ‘distortions' further out on the curve?
Certainly, forward rates offer a much more ‘granular' picture of how curves are moving and twisting than spot rates do - their co-variation is much more idiosyncratic than that of spot rates. Remember the point made by Hafner at Risklab - it can be meaningfully cheaper to hedge using one forward rate against another that, on the face of it, should look very similar.
Brown at Ignis says that ClearCurve's highly-precise breakdown of forward rates shows that "they change very differently from the average [spot] rates over periods of time". According to Ignis' analysis, between the end of July 2011 and the end of October 2011, while the 30-year UK gilt spot rate fell by about 0.7 percentage points, the 1Y11Y forward rate fell by more than 2.5 percentage points but the 1Y1Y, 1Y21Y, 1Y27Y and 1Y28Y forward rates actually rose.
How to explain these movements and kinks in the curve? There are two main theories: ‘pure expectations', which holds that each forward rate is an expression of the market views on the future level of interest rates; and ‘market segmentation', which holds that these are rather anomalies driven by supply-and-demand mismatches. Very few practitioners go with ‘pure expectations'.
Phil Page, client manager with Cardano, notes that it is not unusual to fund the UK 10Y5Y rate some 50-70 basis points higher than the 30Y5Y rate. "Who really knows what's going to happen in 10 years' time - let alone in 30 years' time?" he asks. "Surely your view on Bank of England short rates at 10 years can't fundamentally be different than your view for it at 30 years, and yet you are being offered extra yield."
He points out that pension funds have need to hedge at the 30 and 40-year points, but the supply of instruments is constrained; and while there is much more paper available at the 10 and 15-year points, pension funds generally have more liabilities stretching beyond those points.
"That implies a lot of alpha opportunity," he suggests. "These discrepancies can persist for years, but if you are very patient and decide not to hedge the 30 and 40-year stuff because it's too expensive, but hedge the 10 and 15-year stuff, in the long term you will probably be better off."
Just by breaking a pension fund's curve into five-year rate sensitivity buckets and comparing the forward rates for each of those buckets (5Y10Y, 5Y15Y, 5Y20Y, 5Y25Y, 5Y30Y, and so on), decisions about which buckets to over or under-hedge can be optimised for cheapness. "We've added quite a bit of value by doing that sort of analysis, because the majority of the market is looking at spot rates and not thinking at the more granular level," says Page.
PIMCO's European solutions product manager Jeroen Van Bezooijen agrees that these opportunities "should be exploited in LDI programmes", but adds a couple of caveats. Just as the forward position that involves selling short-dated rates is an active bet on the level and direction of rates, so is a forward position on any other part of the curve. If an apparent anomaly is a reflection of a ‘structural' supply-and-demand imbalance at a particular point on the curve, there is no reason to regard it as a mispricing.
"There's never a guarantee attached to those extra basis points, but if you don't think it is structural, it could be a trading opportunity," he says. "But you have to weigh up the basis risk and transaction costs you are taking on, and not every anomaly you see on your Bloomberg screen can be picked off easily."
So is there really any appetite from pension funds to exploit these opportunities?
At ING, they don't see much sign of it. With funding ratios so low in the Netherlands, appetite for active risk has collapsed and clients just want matching portfolios that "do what they promise", according to LDI specialist Arjen Monster.
Loranne van Lieshout, a senior investment consultant at Ortec Finance, sees a more varied picture. Immature funds with longer-dated liabilities (and a lot of interest-rate risk) prefer to stay precisely matched; others might see themselves as having more of a risk budget. And funds that deploy active management in their broader portfolios because they consider markets to be relatively inefficient are more likely to extend that belief into LDI. Marandu at SEI and Van Bezooijen at PIMCO also say that they have implemented these positions for clients, and that experience with other successful active LDI tactics has opened minds to various possibilities.
"Investors are becoming more comfortable with the idea that the LDI portfolio needn't be a straightforward ‘swap-and-forget' portfolio, because they have seen demonstrable value-add from active management and deploying different instruments," as Tarik Ben-Saud, head of LDI at BlackRock, puts it.
This change in perspective is perhaps the most important aspect of discovering and discussing forward-rate curves. As we have seen, those rates can inform spot-based implementation of hedges - they don't have to lead to forward-based implementation. Thinking about them makes investors consider in more detail the true level of short-dated interest rate sensitivity they may be carrying in their asset portfolios; and leads them into discussions about different and contrasting forms of liability-matching. As Gall at Insight puts it: "Since forwards came onto the agenda when short rates really collapsed, it has greatly improved the quality of the debate between practitioners, consultants and trustees."
Whatever the end results, that has to be a good thing.