Martin Steward tries to pin down the nature of the commodity risk premium

In November 2006 Knut Kjær, then executive director of the Norwegian Government Pension Fund, gave a speech about its ongoing project to turn North Sea oil into a diversified investment portfolio that contained a startling fact: one dollar invested in equities in 1900 was worth $376 by the end of 2005, whereas one dollar's worth of oil bought in 1900 was worth just two dollars.

"Oil shortages may be greater than they have been," Kjær conceded. "Nevertheless, I argue that a far larger share of wealth should be held in equities than in petroleum. By buying equities we exchange oil barrels for interests in production in more than 3,000 companies worldwide."

Companies create value by making and selling stuff. The better they do it, the more money they make - and that is the source of investment risk. It is not so clear that there should be a long-term return from raw materials. For pension funds whose default position is passive buy-and-hold, that is a crucial ambiguity.

So where do commodity returns come from? Most obviously, the spot price changes. In the futures market, trading on margin allows a risk-free yield to be harvested from collateral. And then there is the ‘roll yield': the value of a future contract converges on the spot price as its expiry nears and when a trader sells it and buys a new one further from expiry to maintain his position he gains (or loses) the spread between them (commodity-indices roll during a few designated days each month - the ‘roll period'). Do any of these represent a long-term risk premium?

Let's take the spot price. Much discussion focuses on supply-and-demand dynamics. Populations are rising and getting wealthier while arable land is shrinking; and one does not have to be a ‘Peak Oil' fanatic to accept that fossil fuels and metals must be a finite resource. "Fundamentally, commodities are finite, whereas a company can issue as many shares as it likes," notes Nick Brooks, head of research and investment strategy at the commodities specialist ETF Securities. "People will always get over-excited by the long-term story and then get corrected heavily when everyone focuses back on the short-term demand weakness but, ultimately, there is a medium-to-long-term upward trend, and my view is that pension funds should not be getting too tactical."

But this involves long-term extrapolation of unstable dynamics. Whenever we face expensive challenges to extraction, technological solutions can evolve to bring prices down again; and if supply really does start to get tight, the incentive to find alternative solutions to energy and industrial needs grows. Similarly, it is not hard to imagine diets changing in response to food prices; crop yields improving through better seeds, husbandry, irrigation; or food wastage being reduced by more efficient transportation and storage. To get exposure to the wealth-creation of wind, sunlight, tides, GM seeds, hi-tech tractors or refrigerated lorries one has to turn, once again, to equities.

So what about the roll yield? This is positive if the futures curve is downward-sloping (‘backwardation'); but it is negative if the curve is upward-sloping (‘contango'). John Maynard Keynes suggested that backwardation is ‘normal' because buyers of futures are paid a premium for taking price-fluctuation risk away from commodity producers.
That may be, but that premium is pushed down by the cost of storing (and insuring) commodities until delivery: when inventories are well stocked, consumers are not so anxious to secure their share immediately, and the futures curve goes into contango; when physical markets are tight, they tend towards backwardation. Contango in the first quarter of 2009 saw the S&P GSI index give up a staggering 14% in rolling costs.

"Reluctance to invest in broad indices due to the current contango suggests an expectation that it will persist," says Mike McGlone, head of commodities at Standard & Poor's Index Services. "But historically this has had a history of mean reversion: periods of low prices and ample supplies generally lead to production reductions and increases in demand, thus higher prices."

 The fact that tightening physical supply pushes the slope of the curve down at the same time as it drives spot prices up brings the roll yield question back to the long-term uncertainty surrounding structural demand - although spot is directional and backwardation is relative-value, so the latter would appear to be the more robust ‘premium'. 

The more serious challenge to normal backwardation may come from the increasing volume of financial participants in futures. There is still a strong correlation between inventories and futures spreads. But it is also indisputable that backwardations get smaller and contangos wider during index roll periods. If that weight of money has an effect during this concentrated period of time, is it not fundamentally safe to assume a similar effect should enough money roll enough contracts at any point on the curve at any time of the month? Indeed, increased financial participation could ultimately break down the fundamentals of the curve completely: "The hedging premium that is supposed to lead to normal backwardation is less evident in a market where market participants hedge both up and downside movements in price," observes Alex Koriath, principal consultant in investment advisory at KPMG.

Perhaps looking for the logic for long-term returns misses the point about commodities.
"The base case for commodity investing is diversification," says Martin Woodhams, head of commodities structuring with Barclays Capital. "You can't lose sight of that as you drill down into the question of what the returns have been."

The academic evidence is compelling, the best-known paper being Gorton and Rouwenhurt's from 2005 that found that "over all horizons - except monthly - the equally weighted commodity futures, total return is negatively correlated with the return on the S&P500 and long-term bonds". Even Kjær argued that the tangency portfolio of oil and equities would retain about 15% in the black stuff. Most commodities fell through the floor in the wholesale flight to safety in the final quarter of 2008, but one cannot deny the year-long dislocation from financial assets between the beginning of the credit crisis in summer 2007 and the final blow-up. Many would argue that this also demonstrated the inflation-hedging qualities of real assets.

"There is never any guarantee that you will get the non-correlation just when you need it, but I think it's true to say that even during the recent cycle you were better off with commodities in your portfolio than without," says David Nahmanovici, commodity structurer with BarCap.

Correlations within commodities and between commodities and other growth-sensitive assets have broken down again very quickly this year, as markets have refocused on fundamentals. BarCap's latest commodities research quarterly singles out copper: despite the downward pressures on global GDP, prices have soared in recognition of a big move up in Chinese import demand that has led to a spike in material booked for shipment out of London Metal Exchange (LME) stockpiles. The sceptics might point to that dynamic and say that this is functional rather than structural with respect to the economic cycle, and commodity fundamentals are therefore not a systematic hedge. 

"We believe the diversification benefits of most passive commodities may not be as large as expected for long-term investors like pension funds," says Koriath. "While commodities (in particular oil) are probably a good hedge against short-term exogenous shocks, they offer little protection against a prolonged recession, which is much more risky for pension funds. The inflation hedging properties of commodities are also not perfect for a pension fund that is more exposed to 20 or 30-year expected, rather than current, inflation."

"They provide diversification when you want it most," says Graeme Johnstone of Hymans Robertson. "But there are still long spells when the asset class gives you nothing except volatility. From our perspective that argues against it as a buy-and-hold asset class."

Michael Gran, co-deputy head of quantitative asset management with TOBAM - which has just launched a commodity-futures version of its ‘anti-benchmark' strategy - agrees: "The argument for going long-only is that it's uncorrelated with other asset classes - but if you have a zero expected return from that is it worthwhile?"

Pension funds are asking the same question. A year ago Russell Investments found that, of those allocating to commodities, almost twice as many North American pension funds did so passively than actively. But when it came to the future, the interest in active management overtook passive. And that was before the cliff edge last autumn. 

Active solutions to commodity exposure form a spectrum. An investor may want a buy-and-hold long-only strategy that dampens or optimises roll yield. For those outsourcing their allocation, there are a number of exchange-traded products tackling the issue. For example, Deutsche Bank's db x-trackers' proprietary benchmarks boast a rules-based ‘optimum yield' overlay that rolls its one-year futures at the best time of the month to minimise negative or maximise positive spreads. "We regard that as a smart beta strategy that delivers better performance than the plain beta," says Manooj Mistry, head of db x-trackers UK. Of course, if contango persists throughout the month the spread remains negative, but the damage can be limited.

Note also that this product is rolling one-year contracts rather than front-month.
"At one, two or three years you get less negative effect from contango and less positive effect from backwardation," says Brooks at ETF Securities, which provides WTI and Brent crude products that replicate these longer maturities.

For those managing exposures in-house, the simplest solution would be to exclude certain commodities. Generally, energy and industrial metals are best for positive roll yield and agricultural commodities much less so, while precious metals always yield negative. History suggests that an investor seeking to optimise long-term roll yield should overweight copper, oil, nickel and sugar and underweight US natural gas, aluminium, gold and cocoa futures.

Taking a more active view on weights across the complex makes more sense from the spot-price perspective, too, now that the broad-based trend of the last couple of years has given way to the more complex story of fundamental drivers.

"A passive exposure that worked well from 2002 until late 2007 to capture that broad upswing may no longer be best now that each component is acting under its own steam," says Woodhams. BarCap reckons commodity index products saw only $3bn (€2.1bn) of inflows during the first quarter, with "a significant amount" of that going to "enhanced index products that allow passive investors in broad-based indices to add, subtract or overlay specific exposures".  

Institutional investors are recognising that commodities are a cyclical momentum trade; that hunting for a risk premium is a fool's errand and settling for passive diversification benefits represents significant opportunity loss. That means timing is everything. While a passive long-only allocation to the Dow Jones-AIG Index lost 1.4% between June 2000 and March 2009, a bet long or short once every six months - if it was right each time - would have returned almost 500%. The same move every quarter would have netted a staggering 800%. BarCap notes that a five-year buy-and-hold position put on whenever prices have fallen 10% or more in a month has always generated an average annualised return of 11%.

The current opportunity to buy at the bottom is enticing pension funds like FRR and CalPERS to increase exposure. BarCap's recent survey of institutional clients found four-fifths looking to follow suit over the next three years. Investors that traded out at the top of the market last year and now look to ease back in include Total Pensioenfonds, the Port of Cork Pension Fund and Finland's KEVA.

Again, for those outsourcing their allocation, there are long/short index products available. BarCap estimates that their assets are around $14bn, up some 25% from 2008. Examples include TOBAM's new Anti-Benchmark product, which maintains an equal-weighted portfolio of S&P GSCI contracts and takes long or short positions depending on which way the spot price has moved over the course of a full year - effectively maintaining a diversified exposure to momentum as the ‘risk premium' of the complex. Further along the spectrum are products like S&P's Commodity Trends Indicator, essentially a systematic long/short managed futures index of 24 contracts; and Harewood Solutions' BNP Paribas Comac Fund, tracking a long/short index that incorporates a discretionary element from managed futures specialist Tiberius Group, based on technical and fundamental indicators.  

But moving along that alpha spectrum raises a question: if ‘the base case for commodities is diversification', is it not important to try to identify the nature of commodity beta and get exposure to it?

"Some [investors] have been doing commodities for a decade now and are actively pursuing every segment of the available risk spectrum," says Woodhams at BarCap, which responds to this demand with five product categories: passive, front-month index structures; "static enhancement", which involves some active management of roll dates or curve positioning; "dynamic enhancement", which manages curve positioning to optimise roll yield and risk-adjusted return; quantitative long/short investment strategies; and pure alpha. "Each investor is different, but as a base case it's fair to say that they will almost all maintain some exposure to a benchmark, simply because that's how they are measured for a large part of their commodity investment."

Ultimately, the aim must be to maintain the commodity ‘flavour' while improving the skewness of returns. 

"The further out on the alpha spectrum you go, the more the return becomes idiosyncratic," says Ewen Cameron Watt, managing director in the multi-asset portfolio strategies group at BlackRock. "But it isn't entirely separable from the beta. Even CTAs will tend to make their highest returns during periods of rising rather than falling prices."
Furthermore, there are funds of funds that state that as their objective, implicitly or explicitly.

"We aim to give a kind of structural long exposure to the long-term commodities story while providing management of the downside risk," says Adam Taylor, lead analyst on a new Commodities fund from fund of hedge fund boutique Liongate Capital Partners, which allocates to futures traders as well as investors in resource companies' equity and credit. Importantly, among traders the firm will avoid systematic, ‘black box' managed futures in favour of more fundamentals-based discretionary programmes - the commodities theme is at the heart of the return stream. "Typically our managers would come from Cargill or Louis Dreyfus if they run an agricultural fund, or Shell or Exxon Mobil if it's an energy fund. These people have a much better understanding of the fundamental drivers of commodity prices."

This could be the template for pension funds' allocation to the theme - diversified across physicals, futures, equity and credit; with downside risk management delivering both diversification and positive skew for long-term returns. Demand is growing for an escape from inefficient, passive long-only commodities exposure; and the product suite is rapidly evolving to meet it.