Asset Allocation: Raincoat and sunglasses
Today's distorted and correlated markets confound asset allocation. Martin Steward asks when normalisation will be restored, and how to manage the transition
"Somehow I've become very cautious," reflects the ruthless drugs smuggler in Wong Kar-Wai's 1994 movie Chunking Express. "When I put on a raincoat, I put on sunglasses too. Who knows when it will rain, or when it will turn out sunny?"
At first glance, it looks like the philosophy behind investment diversification. But the idea there is to benefit from your coat in the rain and then swap it for your shades in the sun - in the movie, Brigitte Lin always wears both at the same time. The effect on her is a sinister chic. I suspect that the rest of us would either suffocate in the heat or bump into things in the dark.
That has been the problem recently. The rainclouds are still heavy (bond yields have plummeted); and yet the sun has been shining (equities are up). Why complain? We made money on everything! But we cannot go on like this: to benefit from gains, we need to recycle them into cheaper assets. But which are cheaper? And what are the dangers that the expensive assets could get more expensive still?
Our starting point must be to analyse the reasons for the correlations. They are not that unusual in a transition from one part of the investment cycle to the next, but the persistence of the phenomenon this time is. "I don't know when this correlation will break," as Horacio Valeiras CIO at Allianz Global Investors Capital puts it, "but it's gone on a lot longer than I would have guessed." He is in no doubt about the cause, however: it is the bond markets that are being ‘manipulated' by liquidity-pumping authorities. Similarly, American Century Investments' CIO Enrique Chang suggests that asking whether bonds or equities are priced wrongly misses the point. "The bonds could be right from an asset perspective and the riskier assets from a flow perspective," he says. "With QE it's not as though we're borrowing money from Martians - it's being created within the system, which leads to asset inflation."
Indeed, all assets potentially get inflated: the bonds that are the primary target, and the risk assets targeted by the carry trades funded by those low yields. This is the topsy-turvy world of the ‘Bernanke put', in which gathering rainclouds make you reach for your sunglasses, and it makes perfect sense for investors to say things like, "We have seen some softness in the US housing market over the past few months, and if that continues it would be a pro-risk scenario."
What does it all mean for pension funds? Denmark's Unipension has had an interest rate swap overlay in place for some time, but that was unwound at the end of August because of low yields and expectations of more QE, chief strategist Jøren Christian Hansen explains. "QE2 is the last resort for the Fed. We are seeing the last leg down in bond yields and things are going to change in a more fundamental sense."
He concedes that there is a risk that QE2 could be much larger than the market expects and push yields down even further, but Unipension has an advantage over pension schemes that use variable discount rates to value liabilities: its nominal pension guarantees were recently lowered from 4.5% to 1.5%, giving them much greater flexibility to risk shortening duration. Schemes in jurisdictions with similar liability structures have found 2010 less arduous than others. Günther Schiendl, CIO of Austria's VBV Pensionskasse, notes that fixed discount rates that are higher than market rates have enabled VBV clients with more risk budget to get ahead of their funding plans this year. Ilmarinen in Finland, with no guaranteed return rate, is also able to take more of a conviction view on rising yields.
"At such low absolute levels, even a small increase in yields will have a big and sudden negative impact on your return if you are long-duration," says deputy CEO and head of investments Timo Ritakallio. "Our strategic duration target is four years, but today it's much shorter - which should tell you how worried we are. The timescale of yield-curve steepening is a big question… [but] start this process too late and you will already have suffered losses."
But others might suffer funding losses if yields continue to buckle. Swedish schemes get to choose their discount rate from the government bond, mortgage bond and swap markets, and PP Pension's choice of the two former has sheltered it from the worst effects of falling yields this year. CIO Cecilia Thomasson Blomquist knows that the negative real rates at the short end of the Swedish curve are simply not sustainable - but she has to recognise that such anomalies can persist. "You need your investment strategy and views but you also have to control your interest rate risks because you never know what's going to happen," she says. For PP Pension that means swaptions. "We think paying the premiums is worthwhile to avoid locking-in these rates."
There are some technical indications that the latest ‘risk-on' trade - set in motion at the end of August by a dovish speech from Fed president Ben Bernanke, and suggestions that the Fed might be changing its thinking about inflation - might not be exhibiting the same correlations as previous regimes this year. At end-October the US 30-year yield has moved above 4% - it was at 3.66% at end-September. The 10-year is up 23 basis points to 2.69%. German and UK 10 and 30-year yields have also edged up. Breakevens have been rising across all three markets, and on October 25 $10bn of five-year US TIPS attracted a negative yield at auction of -0.55%. Elsewhere, the risk picture falls into place: CDS spreads tightened, the euro bounced against the Swiss franc, gold and copper hit post-crisis highs. In the UK, huge lumps would have been gouged out of your deficit by September and October's steepening yield curve and rising stock markets, and the greater your asset-liability mismatch the bigger the lumps.
But the risks remain considerable. There are still plenty of reasons to buy bonds. "Look at Ireland for an extreme example," suggests Neil Williams, chief economist at Hermes Fund Managers and a convinced disinflationist. "Six percent for an eight-year bond sold in late September. Factor in Irish deflation and we're talking over 7% of yield if you believe they're going to pay it all back. Even in the US, 10-year yields at 2.66%, if you're expecting deflation of 1%, isn't too bad. On a short-term view it's difficult to see institutional investors wanting to miss out on further disinflation and deflation trades."
Deciding what it is yield curves are telling us is not easy in this environment. As Union Investment Institutional portfolio manager Thomas Bossert asks: "How would we know if the market is pricing inflation in at 10 years? You won't see it because of all this artificial buying."
Valentijn van Nieuwenhuijzen, head of fixed income strategy and economics at ING Investment Management, makes a similar observation from the disinflationists' point of view. "We need to put recent moves in perspective," he says. "I don't think that the US or Europe is Japan. Having said that, Japan-style elements can come into play, like the short-end of the curve becoming fixed and movements being dominated at the 10-year, most liquid part of the curve. From time to time you could see steepening between two and 10 years but flattening at 10 to 30 years. One of the lessons I took from Japan was to look at curve dynamics a little differently in these situations."
Basically, it is difficult to extrapolate from market indicators caught-up in expectations of QE. We need to see indicators from the real economy that can tell us something about the likelihood and size of QE. But then we get into politics and the psychology of central bankers, which sends out equally problematic signals.
"Don't expect anything special from the markets in response to weak data for the rest of the year," suggests William de Vijlder, global CIO for BNP Paribas Investment Partners. "In fact, if you get weak payroll numbers in November and December the markets might think, ‘Well, the Fed has anticipated this weak data, so if anything they will respond more forcefully'."
Sure enough, the weak US Q3 GDP growth print weighed down on the US dollar and steepened the US curve. UK yields also fell, which is significant given that the UK's Q3 GDP growth number, which far outstripped market expectations, barely interrupted their two-month drift upwards.
As we move into the New Year, however, De Vijlder expects markets to look for more solid indications that the real economy is benefiting from low rates, such as improvements in mortgage refinancing and National Association of Housebuilders data, increased bank lending, stabilising Institute for Supply Management numbers (the 1 November data was encouraging) - and, of course, some respite from this dreadful ‘jobless recovery'.
Of course there is an outside chance that the Fed will see that as evidence that its monetary policy is working and therefore justification to do more - but the likelihood is that it will ease off and finally allow a sustainable bear flattening of the curve. And if we don't see any good news? The conventional expectation is that we continue to see the ‘Bernanke put'. Williams at Hermes is about as bearish on growth as any - he points to the fact that the US GDP deflator is not only sub-zero again, but lower than it was even at the lowest point of the recession - but he still describes himself as an ‘equity bull'.
"Corporate margins are being maintained not because of extra demand but because of cost control," he says. "That leads me to feel that there is an awful lot of bad news yet to come from US payroll numbers. More QE will mean a lot more cash in the system: it's no accident that when QE1 was announced in March 2009, growth assets such as equities and property were supported and the same could happen with QE2."
But you do not have to be a rampant inflationist to put a different spin on ‘bad news'. "There is a lot of attention from policymakers on the very high levels of unemployment, and the view is that this reflects a cyclical lack of demand," observes Philip Poole, global head of macro and investment strategy at HSBC Global Asset Management. "But there is an argument that the crisis has exposed an increase in the level of structural unemployment for a given level of inflation in the developed world. If that is the case, some US unemployment can be addressed with monetary policy, but not the structural rate."
If we accept that the Taylor Rule needs to be re-calibrated we might also assume our post-crisis growth expectations need similar re-calibration. Indeed, when the Riksbank hiked rates in September it also mused on the permanence or otherwise of Sweden's post-crisis output gap. If there has been a permanent loss of GDP, ‘looser for longer' starts to look like a dangerous striving for an unrealistic level of economic activity. And in the US the fact is that the Richmond and Philadelphia Fed presidents have already added their scepticism against QE2 to established hawks in Kansas City and Minneapolis, while Dallas' Fed president has explicitly said that, "while unemployment is a terrible problem, it is less obvious to me that it is amenable to monetary policy solutions at this point".
Still, Unipension's Hansen is right to point out that this remains "a minority view" in the Fed - and moreover one diametrically opposed to Bernanke, who would consider the worst case scenario of money leaking into US Treasuries and emerging markets a risk well worth taking. Giancarlo Perasso, chief economist at Redux Research and visiting professor at the Università Carlo Cattaneo in Italy, agrees. He makes no bones about the structural issues in the US jobs market - he sees this rupture as "one of those global events like the fall of the Roman Empire". No amount of QE can re-train a 50 year-old assembly-line worker as a retirement home nurse or chemical engineer. "But what are the alternatives? If for each dollar of QE, 50¢ went to Brazilian and Chinese assets, 30¢ went to US Treasuries and 20¢ come out as more lending they'd be happy with that."
Allocating to BRICs
Which suggests that one should allocate some risk to riding that 50¢ flowing into Brazil, China and every other risk asset under the sun. It helps that most people seem to think that fundamental valuations support the move, too.
So Unipension has been unwinding its rates exposure and moving into high yield and emerging markets, for example. "We see liquidity flowing out of the US and into these markets as one consequence of QE2 - and we find valuations to be quite attractive at the moment," says Hansen. "There's probably never been a better time to be long equities," enthuses VBV's Schiendl.
But the fundamentals are slippery. If you think discount rates are artificially depressed, what does that mean for valuations? So while QE2 will float all boats, investors may be tempted to take profits quicker, pushing up volatility.
"Longer-term, it's a lot easier to identify themes that you can apply with conviction," as HSBC's Poole puts it. "In the short-term it's difficult to have conviction on anything. QE will leach to areas where investors perceive that there is some topline growth on offer, potentially blowing up bubbles in the emerging world."
This would be ideal for dynamic de-risking, were it not for the fact that taking profits from the growth portfolio would involve locking-in rates at unsustainable levels. In recent conversations with Dutch pension funds, BNPIP's De Vijlder says that they do indeed want to avoid this dilemma, seeing three basic choices: simply taking risk against liabilities; using swaptions in the liability-matching portfolio; or achieving positive convexity in the growth portfolio.
We have seen already that the swaptions route works for PP Pension. But many investors balk at the premiums, high even before bond volatility started rising. "You have to work with it dynamically," says Blomquist. "You always have to ask, ‘Do I need the protection against falling rates, and if so, at what price?' Our strategy is to buy at different strike levels for each different expiry date, rather than focusing on one fixed price point. Good returns on the growth side means that we can tolerate a little more interest rate risk and go for lower strikes and therefore pay lower premiums."
Those without liability-matching obligations may wish to focus purely on convexity in their growth portfolios. Convertibles are great in this context, of course: we explore the asset class further in this month's ‘Briefing' (p53). But they do present problems. Union Investment is one of Germany's biggest managers of convertibles, thanks to its business in dynamic protection strategies, but Bossert concedes: "Often the assets that you would like don't have enough liquidity, and the liquid or more attractively priced markets are not necessarily in the assets that you want."
Ritakallio says that Ilmarinen does allocate to convertibles, but emphasises the importance of looking for convexity across the entire portfolio, from increased FX exposure through the high yields still available in some real estate markets ("If we didn't have so much real estate we might have considered convertibles," says PP Pension's Blomquist). This logic can lead to an overweight in something as straightforward as high-yielding equities. "We have had a lot of focus on high-dividend paying stocks," says Ritakallio. "You can get dividend yields that are clearly higher than corporate bond yields on some companies."
One consequence of moving towards maximising convexity across the whole portfolio is a blurring of the line between liability-matching and growth portfolios. The huge appetite for credit spread - and increasing recognition of the diverse spectrum available - is an indication that investors are making this move.
"Low bond yields put us in a situation where re-investing cash flows became more difficult," says VBV's Schiendl. "In response we pursued diversified bond strategies: running perhaps half of our fixed income in credit - corporate bonds, high yield and emerging market bonds."
If you worry that these have had too good a run lately, there are still plenty of other forms of credit out there. Mexico's 100-year bond yielding 6.1% and Goldman Sachs' 50-year yielding 6.25% have little to do with meeting demand for longer duration - it's all about credit. AGIC's Valeiras sees a good scalable opportunity in the illiquidity premium available on senior bank loans. Perhaps you could sell CDS on your sovereign.
And then of course there is the peripheral euro-zone. Here our pension fund managers are not so keen, seeing much better risk-reward prospects in emerging market and corporate bonds. Ilmarinen's discussion around the so-called ‘PIIGS' has been about whether to reduce its strategic allocation, if anything. Similarly, Hansen reveals that Unipension briefly thought that Greek government bonds might be "one of the great investments of the year" 12 months ago. "But we went there in January and came back with a very negative view," he says. "We sold all of our southern Europe exposure, although it was not very much - we don't see that the upside justifies the risks."
Overall, then, optionality on falling yields is good where available and optionality on growth essential, which leads to a whole-portfolio focus on convexity. This changes the LDI paradigm to one of rebalancing between different sources of convexity rather than rebalancing between growth and duration assets. If and when yields (and correlations) begin to ‘normalise', investors will need to decide the extent to which they continue down this asset allocation path, rather than turning back to a more rigid segregation of liability-matching and growth portfolios.