controling investment risk in the commodity space · ral gas is a blatant example of the ine ciency...

OSSIAM RESEARCH TEAM

Controling investment risk in the commodity space

April, 4, 2014

WHITE PAPER

1

Controling investment risk in thecommodity space

Bruno Monnier

April, 4, 2014,This is the submitted version of the following article : Controling investment risk in the commodity

space, B. Monnier, November 2013, Investment Pension Europe

Abstract

Bruno Monnier, CFAQuantitative [email protected] Commodity futures have evolved from trading instruments

to vectors of long-term investment for this new asset class.As uses and users have changed, new ways of representingthe performance of commodities class and investing in ithave to be devised. In this article, we outline the draw-backs of the traditional commodity indexing solutions andwe expose solutions realigning commodity investing withlong term investor goals. In particular, we detail risk man-agement techniques, at the single commodity level as well asat the portfolio level, that can be used to build a balancedinvestment benchmark.


2

Formerly a market for hedging purposes -used mainly by commodity producers or users- the commodity space has gradually openedto investors. Commodity indexation has gonea long way towards popularizing this marketproviding standard, cost-effective access prod-ucts. It simplifies the specificities of this mar-ket - its term structure, the futures contractsthat previously required a high level of sophis-tication. While commodity indices as bench-marks have existed since 1957 (the CRB in-dex), the first investable index (the S&P GSCIIndex ) was launched in 1991. A wide ar-ray of similar indices are now available, cov-ering the whole asset class, slicing it into var-ious subsectors or even into single commodi-ties. A survey conducted by consultancy firmETFGI, inventoried as much as 783 ETFs andETPs totaling $151bn AUMs in this mar-ket, as of the end of August 2013. Traditionalcommodity indices were designed on princi-ples that make them useful market indicatorsbut poor benchmarks for investors intentingto hold long term exposure to the asset class.These indices overlook key features of the termstructure of the commodities futures contracts,such as contango and backwardation. This re-sults in higher than needed maintenance costswhen the index rolls over the futures positionsto replace the expiring contracts. Moreover,built as broad economic benchmarks, they al-locate the weights in the index basket propor-tionally to the economic weight of each com-modity. This results in an overconcentrationon some market sectors that is at odds withthe best practices in portfolio allocation. Re-cent evolutions in the financial industry reg-ulation, emphasizing portfolio diversification,make the traditional indices even less suitablefor investing purposes. We present here newapproaches to commodity index design thataddress the issues of the traditional commod-ity indices and allow to invest in this market ina risk-controlled way. We show how contract

maturity management gives ways to manageefficiently each single commodity investmentposition. Finally, we address how portfolio al-location can be redesigned in order to make thecommodity index more risk-efficient

1 The case for a new paradigm

The commodity asset class contains but a fewdifferent underlyings when compared to themore traditional equity or fixed income assetclasses: there are less than 25 different com-modities in the major broad commodity in-vestment sets. This lack of breadth in the in-vestment universe means that the commodityportfolios will tend to be less diversified thanequity or fixed income portfolios. The classi-cal weighting scheme within traditional broadcommodity indices by production weight andliquidity - heightens the concentration prob-lem. The S&P GSCI index for example has anexposure skewed towards one specific sector:energy. At the end of August 2013 It repre-sented 72% of the total index weight crude oilis 30% by itself while corresponding to only 1

4of the index constituents. This concentrationissue is even more acute when we consider riskcontribution of each sector in the total portfo-lio risk instead of the nominal exposure. Thepercentage of total volatility explained by theallocation to this sector is 82%. Figure 1 dis-plays the breakdown of risk accross the differ-ent sectors within the S&P GSCI .

Regulators have identified this concentra-tion issue. The new European guidelines e.g.UCITS IV - impose stricter rules that maketraditional commodity indices uneligible in-vestment benchmarks. To respond to thischange in regulation landscape, ”capped” ver-sion of the traditional indices have emerged toensure compliance. But the capped indices,using ad-hoc constraints, do not offer a re-ally new risk-aware allocation principle. Ethi-


3

Figure 1: Risk contribution by Sector in theS&P GSCI

Source: S&P , calculation by Ossiam

cal considerations should also be added on topof the regulatory constraints. With increasedcontroversy about investment in food relatedcommodities, investors are shying away frominvestment in this sector or even fully disal-locating. This shrinking of the eligible assetbase will mechanically increase the concentra-tion on other commodities even further. In thiscontext, traditional allocation principle has tobe redesigned in order to provide risk-efficientand diversified benchmarks, compatible with along term investment.

2 New building blocks

One specificity of commodity market is theneed to use futures contract to implement com-modity investments. Each underlying com-modity is represented by a number of futurescontracts of different maturity. The traditionalindices were based on the nearest-month con-tracts that are rolled to the next available ma-turity before they expire. One can refer tothese sequences of futures positions as mono-commodity indices. They provide investableunderlyings that can be used as building blocksin the design a broad commodity portfolio.

The choice of the nearest-term futures con-tracts inside the traditional mono-indices wasdetermined with the objective of reproducingas close as possible the spot commodity pricesmovements. This is suitable if one has touse a single-commodity benchmark for short-term hedging or tactical purposes. Still, overthe long run the traditional mono-indices suf-fer from high roll costs. Investment in natu-ral gas is a blatant example of the inefficiencyof the traditional roll technique: the mono-commodity index lost 99% over the period2003-2013 whereas the spot price was downonly 32% over the same period. It is possi-ble to reduce the roll cost by taking into ac-count the shape of the futures price curve. Themost common curve shape is ”contango” withprices increasing with contract maturity at adecaying rate. Such a situation prompts thatthe less costly roll occurs between contractsthat are at the far end of the forward curve(the ones with the longer maturities). Theuse of such a technique when indexing natu-ral gas yields a performance of −67% on theperiod 2003-2013, more aligned with the spotprice own performance. This rolling techniquewas used to build the so-called second genera-tion mono-commodity indices, that are calleddynamic or enhanced. Performance enhance-ment in the second generation mono indicesis also accompanied by a decrease in risk mak-ing them truly more efficient investment under-lyings. The longer term contracts often usedinside the enhanced contract selection processare also those that exhibit the lower volatili-ties. Figure 2 shows this decreasing pattern.

A simple economic reasoning can explainthis behavior. Commodity prices are directlyaffected by supply and demand e.g. a sup-ply shortage will trigger soaring prices. De-mand can vary greatly in a short amount oftime while producers need time to adjust theirproduction ouput to the new level of demand.This creates tensions in short term prices and


4

Figure 2: Volatility of S&P GSCI for differentfutures maturities for the period 2000-2013Source: Bloomberg, calculation by Ossiam

overreaction to new economic data or restate-ments of production estimates. These tensionsare smoothed out when the contract term islonger than the time needed to adjust the sup-ply: the future price only accounts for the eco-nomic equilibrium price. Natural gas is onesuch commodity whose volatility is highly de-pendent on the maturity stocking gas is alengthy and complicated process. The volatil-ity of the 1 month contract is 48% over thelast ten years while for the 3 month contract itis down by one forth, with 36% volatility. Asthey improve on both performance and risk,second generation mono indices are suitablebuilding blocks to use in redesigning broadcommodity indexing.

3 Portfolio allocation

The solutions combining the classic produc-tion weight allocation with the new genera-tion of mono indices are more efficient thanthe traditional indices, yet they fail to addressthe core problem: the lack of diversification.This issue should be addressed explicitly atthe portfolio construction level. In an assetclass with such a low number of assets, apply-

ing a ”Risk Parity”1 allocation will provide abalanced portfolio which contains all the com-modities within the investment universe. RiskParity allocates the same risk budget to eachportfolio constituent. The weight of each com-modity in such a portfolio is inversely propor-tional to its volatility. Such a constructionevens out the risk contributions of the assetswithin the portfolio and aligns the weight ofdifferents sectors to the number of constituentsthey contain. Figure 3 shows the risk contri-bution breakdown that results from applyingrisk-parity approach to S&P GSCI investmentset.

Figure 3: Risk contribution by Sector for aRisk Parity Portfolio

Source: Bloomberg, calculation by Ossiam

This allocation can be described in anintuitive manner when compared to themost straightforward portfolio constructionthe Equal Weighting all assets have the sameweight. The risk parity goes one step fur-ther than Equal Weighting in the diversifica-tion process. While Equal Weighting wouldhave sector weights proportional to the num-bers of assets that compose each sector, RiskParity will also factor in the relative riski-ness of each sector. For instance, energies aregenerally more volatile than the other com-

1Performance simultated by Ossiam on the period2003-2013


5

modities, their risk parity weight will thus beslightly lower than what equal weighting wouldprompt. Livestocks on the contrary will beoverweighted because of their relatively lowerrisk. The two innovations, enhanced maturitymanagement within mono-commodity indicesand risk-aware portfolio allocation, lead to alarge reduction in the volatility of commodityindices. When applied on the S&P GSCI uni-verse, it gives a commodity portfolio, the ”RiskParity Portfolio ”, with volatility reduced byaround 45% with respect to the traditionalS&P GSCI Index (over the period 2003-2013).The majority of the reduction (83%) resultsfrom the change in the weighting scheme whilethe remainder comes from the enhancement inmaturity management. Portfolios that are de-signed around risk management are also betteraligned with the new regulations that targetdiversification. Additional capping mechanismmay be embedded in the portfolio constructionprocess as safeguard to prevent any breach ofthe regulatory constraints. Below we give anexample of a new generation commodity index:Risk Weighted Enhanced Commodity Index(ex Grains), that implements the portfolio con-struction principles described above and repre-sents a risk-efficient investment benchmark.

4 Proof of concept

The RiskWeighted Enhanced Commodity exGrains Index an index of new generationcreated by Societ Generale, combines the en-hanced construction of mono indices with riskparity portfolio construction. In addition,the index methodology forbids investments ingrains and ensures that the weight of eachcommodity sector does not exceed regulatorylimits 2. This provides us with an exam-ple of a readily available investment solution

2UCITS IV impose a maximum of 20% on the ex-posure of each group of highly correlated commodities

on commodities that embodies the conceptswe described throughout this article. Fig 4,that plots the rolling volatilities of the S&PGSCI , the Risk Parity portfolio and the Risk-Weighted Enhanced index, shows that the riskreduction achieved using the allocation man-agement techniques is present and consistentover time.

Figure 4: Comparison of historical 1 yearrolling volatility


It can also be noted also that removinggrains from the portfolio and introducing sec-tor weight limits does not alter significantlythe risk reduction potential: the Risk Parityportfolio has roughly the same volatility asthe Risk-Weighted Enhanced ex Grains index.As is shown in Table 1, the volatility reduc-tion is not the only benefit of the new indexdesign. Over the period 2003-2013 the alter-native commodity portfolios generated consis-tent overperfromance with respect to the tra-ditional S&P GSCI Index . This extra perfor-mance came from contract selection for 80% ofit, the rest resulted from the allocation.

5 Conclusion

Traditional commodity indices are valid eco-nomic indicators but inefficient investment


6

Risk Parity

RiskWeightedEnhancedCommodity

Perf. over S&PGSCI (Annual-ized)

1.62% 1.61%

Risk Reduction 45.27% 43.98%

Table 1: Simulated Statistics on the period2003-2013


benchmarks for long term investors. The im-balance of production-based weights leads tosevere concentration of the related index prod-ucts on specific commodity sectors. This goesagainst sound allocation rules, such as the onesembodied by the asset management regula-tions. A new generation of broad commodityindices has emerged. They offer investors sim-ple and efficient ways to access the asset classwhile controlling the investment risk and op-timize futures rolling cost. This evolution issimilar to what was witnessed in equities overthe past years under the name of ”Smart Beta”Indexing. In the commodities market as well,systematicaly applying risk management tech-niques adds value to the investment process,both on an absolute and on a risk adjusted ba-sis.

Ossiam, a subsidiary of Natixis Global AssetManagement, is a French asset managerauthorized by the Autorit des Marchs Fi-nanciers (Agreement No. GP-10000016).Although information contained herein is fromsources believed to be reliable, Ossiam makesno representation or warranty regarding theaccuracy of any information of which it is notthe source. The information presented in thisdocument is based on market data at a givenmoment and may change from time to time.This material has been prepared solely forinformational purposes only and it is notintended to be and should not be considered asan offer, or a solicitation of an offer, or aninvitation or a personal recommendation tobuy or sell participating shares in any OssiamFund, or any security or financial instrument,or to participate in any investment strategy,directly or indirectly.This material reflects the views and opinionsof the individual authors at this date and inno way the official position or advices of anykind of these authors or of Ossiam and thusdoes not engage the responsibility of Ossiamnor of any of its officers or employees. Pastperformance is not necessarily a guide tofuture performance. Any opinions expressedherein are statements of our judgment on thisdate and are subject to change without notice.Ossiam assume no fiduciary responsibilityor liability for any consequences, financialor otherwise, arising from, an investment inany security or financial instrument describedherein or in any other security, or from theimplementation of any investment strategy.This material may not be distributed, pub-lished, or reproduced, in whole or in part.

The Risk Weighted Enhanced CommodityIndex (ex-grains) (the ”Index”) is the propertyof Socit Gnrale (”SG”) and is used underlicense. The Ossiam Risk Weighted EnhancedCommotidy ETF UCITS are not sponsored,


7

endorsed, or promoted by SG. Neither SGnor any of its affiliates have passed on theOssiam Risk Weighted Enhanced CommotidyETF UCITS as to their legality or suitability,and such parties make no warranties nor bearany liability with respect to the Ossiam RiskWeighted Enhanced Commotidy ETF UCITS.Neither SG nor any of its affiliates acts as aninvestment adviser with respect to the OssiamRisk Weighted Enhanced Commotidy ETFUCITS, Ossiam or Index. Neither SG norany of its affiliates have been involved in thewriting of the article. SG and its affiliatesshall bear no liability whatsoever with respectto the opinion(s) or any kind of informationexpressed by the author(s) of this article.


controling investment risk in the commodity space · ral gas is a blatant example of the ine ciency...

Documents