Commodities as an Asset Class

Transcription

1 Commodities as an Asset Class Delivering Beta & Beyond Dr. David-Michael Lincke, CFA, FRM Continuing Education Seminar CFA Society Switzerland Zurich, 14 October 2016

2 Contents Commodities - State of the Market and Asset Class 3 Portfolio Contribution Diversification, Inflation Sensitivity, Risk Premium 10 Evolution of Commodity Beta Strategies 17 Alternative Risk Premia in Commodities 31 Summary 36

3 State of the Asset Class Where do we find ourselves in the commodity cycle? The upswing of the China-driven commodities supercycle peaked in 2011 and has given way to a prolonged downswing Global growth has disappointed since the global financial crises. The transition of the Chinese economy from an investment-focused and export-led growth model towards a domestic consumption driven economy implies reduced rates of growth In recent years, vastly expanded production capacity incentivized by the commodities boom has met a marked slowdown in commodities demand growth Historical analysis suggests an average duration of cyclical commodity downswings of approximately 7 years with concomitant price declines amounting to -40% on average (cf. chart p.4) The current downswing has persisted for 5 years and is exhibiting tentative signs of bottoming Price declines observed on leading commodities have already reached significantly above average magnitude, e.g. copper -55%, crude oil -65% Analysis of the total return of the broad-based Bloomberg Commodity Index indicates a complete reversal of the recent supercycle upswing with the benchmark retreating to levels last seen in Divergence with spot returns has been exacerbated by the compression of collateral yield as a function of monetary policy response since 2008 Current market pricing put in relation to marginal production cost curves across sectors suggests an impending acceleration of measures to curtail unprofitable excess capacity clearing the path towards a rebalancing of supply with demand and the commencement of a new cyclical upswing, the initial recovery phase of which is being driven by a rationalization of supply

4 Historical Analysis of Downswings in Commodities Markets Duration of downswings and extent of price declines Analysis based on data from 1950 to Source: International Monetary Fund (IMF), Capital Economics

5 Commodity Benchmark Total Return and Spot Return Depressed collateral yield and negative roll yield driving divergence with spot returns since 2008 On a total return basis, the China driven supercycle upswing has been nearly fully reversed Source:, Bloomberg LP

6 Valuation of Commodities Relative to Equities Ratio of Bloomberg Commodity TR Index and S&P 500 TR 1990s bear market China supercycle upswing Quelle:, Bloomberg LP

7 Commodity Assets Under Management A niche asset class emerging from a cyclical downturn Total commodity assets by type USD billion Total commodity assets by sector USD billion Data as of end of August Source: Bloomberg LP, MTN-i, ETP issuer data, Barclays Capital Research

8 Commodity Index Investments Exposure concentrated in major beta benchmarks, systematic alpha strategies gradually gaining ground Commodity index-linked assets by index USD billion Commodity index-linked assets by strategy type USD billion Data by index as of end of August 2016, data by strategy type until end of Source: Bloomberg LP, MTN-i, ETP issuer data, Barclays Capital Research

9 Contents Commodities - State of the Market and Asset Class 3 Portfolio Contribution Diversification, Inflation Sensitivity, Risk Premium 10 Evolution of Commodity Beta Strategies 17 Alternative Risk Premia in Commodities 31 Summary 36

10 Why Consider Commodities in the Strategic Asset Allocation? Benefit to the portfolio Diversification Low average correlation with major traditional asset classes equities and bonds Historically, the addition of commodities to an equity/bond portfolio has resulted in a sustained increase of risk-adjusted returns Inflation Protection Among all asset classes, commodities exhibit the highest degree of sensitivity with respect to price pressures Return Earning the asset class risk premium (beta) The cyclicality and lack of cash flow yield from commodities investments raises questions regarding the existence of an asset class risk premium Historically, provided a suitable investment strategy that adequately accounts for the cyclicality of the asset class is chosen, a broadly diversified commodity portfolio has achieved equity-like risk-adjusted returns over the long term

11 Diversification Benefits with Respect to Equities Average 12-month rolling correlation of 0.14 but with significant excursions Rolling correlation analysis using monthly data from to for the S&P 500 and the S&P GSCI Commodity Index. Efficient frontier analysis using monthly data from to using the S&P 500 TR and the UBS Bloomberg CMCI ER Index with S&P 500 returns converted to excess return using 1M USD LIBOR rates

12 Diversification Benefits with Respect to Bonds Average 12-month rolling correlation of 0.17 Rolling correlation analysis using monthly data from to for the JP Morgan Global Aggregate Bond TR Index (USD unhedged) and the S&P GSCI Commodity Index. Efficient frontier analysis using monthly data from to using the JP Morgan Global Aggregate Bond TR Index and the UBS Bloomberg CMCI ER Index with JP Morgan Global Aggregate Bond TR Index returns converted to excess return using 1M USD LIBOR rates

13 Inflation Sensitivity across Asset Classes Inflation betas with respect to US consumer prices (US CPI) years 9 years Commodities Global Real Estate US Equities US TIPS US Bonds US Dollar Calculations based on monthly year-over-year returns of the following indices: US CPI Urban Consumers YoY NSA, Bloomberg Commodity TR Index (commodities), FTSE EPRA/NAREIT Developed USD (global real estate), Russell 3000 (US equities), Barclays US Inflation-Linked Bonds TR (US TIPS), Barclays US Aggregate TR Value USD (US bonds), ICE US Dollar Index (US dollar). Period: January 1999 (17 year span) and January 2007 (9 year span), respectively until December Source: Picard Angst Ltd., Bloomberg LP

14 Long-Term Asset Class Risk Premia Inflation-adjusted returns from equities, bonds and commodity futures 1959 to 2014 Commodities US Equities US Bonds Return p.a. 4.95% 5.91% 2.93% Volatilit p.a % 14.81% 8.97% Sharpe Ratio Source: Bhardwaj, Gorton, Rouwenhorst, «Facts and Fantasies about Commodity Futures Ten Years Later», Yale YCF Working Paper No , May Commodities represented by an equal-weighted total return index composed of 36 commodity futures with monthly rebalancing, US equities represented by the S&P 500, US Bonds represented by the Long-term US Government Bonds TR Index. Period: July 1959 until December 2014

15 The Commodity Beta Risk Premium Theories regarding the source of the commodity asset class risk premium Keynes' theory of «normal backwardation» Risk premium arising out of the insurance function of futures markets: F(t) = E[S(T)] P(T) «Hedging pressure» hypothesis Generalization of the theory of normal backwardation to take into account the consumer in addition to the producer perspective with respect to hedging behavior and its impact on roll yield Explanation of the shape of the term structure as a function of producer and consumer hedging behavior Theory of storage Focus on convenience yield as a risk premium and measure of the value of prompt availability of a commodity and its role in determining commodity prices in addition to storage and financing costs Diversification return Periodic rebalancing generates returns exceeding the weighted sum of portfolio component returns The size of diversification return arises as an inverse function of the covariance matrix

16 Contents Commodities - State of the Market and Asset Class 3 Portfolio Contribution Diversification, Inflation Sensitivity, Risk Premium 10 Evolution of Commodity Beta Strategies 17 Alternative Risk Premia in Commodities 31 Summary 36

17 Return Decomposition of Commodity Index Strategies Spot price return Commodity selection Portfolio component weights Rebalancing strategy ( diversification return) Collateral yield Management of the underlying funding of an unleveraged investment Roll yield Impact of the commodity futures term structure Contango upward sloping forward curve negative roll yield Backwardation declining forward curve positive roll yield

18 Commodity Futures Term Structure Contango and Backwardation Indicator of the supply/demand balance and determining factor of roll returns Price positive roll yield Backwardation Futures Price = Spot Price + Financing/Storage Costs Convenience Yield Contango negative roll yield nearby contract next contract Maturity

19 Commodity Beta Return Drivers Roll yield as the primary long-term driver of commodity beta returns Correlations of 10-Year Returns with S&P GSCI Return Drivers Analysis based on monthly S&P GSCI Commodity index returns between January 1970 and June Source: Erb/Harvey: Conquering Misperceptions about Commodity Futures Investing. Financial Analysts Journal, Volume 72, Number 4, July/August 2016.

20 Commodity Index Strategies From first to second generation beta strategies First generation index strategies Holding contracts situated at the front end of the term structure Rolling exposure to the next nearby contract on a frequent basis (as often as monthly) Second generation index strategies Introduced with the objective of mitigating the detrimental impact of contango, which due to the typical curvature of forward curves is most pronounced at the front end of the term structure Common approaches Forward roll (F1 through F12 adaptations of first generation indices) Enhanced seasonal roll (e.g. S&P GSCI Enhanced Commodity Index) Constant maturity roll (e.g. UBS Bloomberg CMCI, JP Morgan CCI) Optimized implied roll yield (e.g. S&P GSCI Dynamic Roll Commodity Index, Bloomberg Roll Select Commodity Index)

21 Second vs First Generation Commodity Index Strategies Static approaches like F3 Forward Roll keeping pace with more complex dynamic strategies Source: Bloomberg LP

22 Roll Yield Optimized Index Strategies Dissipation of market inefficiencies: Backtest results vs out of sample performance backtest period out of sample performance Ratio of roll yield optimized and standard versions of commodity benchmark indices based on weekly returns of the S&P GSCI Dynamic Roll TR and S&P GSCI TR indices as well as the Bloomberg Roll Select Commodity TR and Bloomberg Commodity TR indices. Source:, Bloomberg LP

23 Compositional Deficits of Common Commodity Benchmarks Eliminating structural drag and aligning basis risks In terms of portfolio composition, standard commodity benchmarks, such as the S&P GSCI and Bloomberg index families, suffer from inadequate construction methodologies focused on production and liquidity statistics Disparate sector weightings, e.g. extreme energy overweight in the S&P GSCI due to exclusively taking into account production volume data Inclusion of commodity contracts, e.g. US natural gas, whose economic characteristics guarantee losses if held over the long term Disregard for the correlation structure among as well as within commodity sectors resulting in unnecessarily high volatility levels or overly expansive portfolio size The US dollar based nature of commodities markets raises challenges for European investors looking to take commodity exposure as an inflation hedge Basis risks abound with standard commodity benchmarks (e.g. US natural gas, US livestock) A dynamic approach to sector allocation can improve significantly on the inflation hedging suitability of commodity investments

24 Persistent Negative Long-Term Excess Returns from Natural Gas Highest weighted component in the Bloomberg Commodity Index Source: Bloomberg LP

25 Optimizing Commodity Exposure for Inflation Hedging Taking into account time-varying inflation sensitivity across commodity sectors Inflation sensitivity varies across commodity sectors Over the long term, energy and industrial metals exhibit the highest inflation betas However, betas do not remain stable over time and are subject to substantial volatility as a function of the economic cycle Thus, if exposure to commodities is sought primarily as a defensive investment to take advantage of the pronounced inflation sensitivity of the asset class, portfolio construction should be driven by the time-varying inflation betas of the investment universe, e.g. Weighting of sectors energy, industrial metals, precious metals and agriculture (grains, softs) as a function of their respective dynamic inflation betas (subject to a sector weight cap of 70%) Monthly rebalancing of sector allocations Exclusion of commodity futures contracts from the investment universe that bear no relation to the European economy (notably NYMEX natural gas and the entire livestock sector)

26 Inflation Sensitivity of Commodity Sectors Sector betas: Energy and industrial metals dominate over the long term years 12 years Energy Industrial Metals Precious Metals Agriculture Beta calculations based on monthly year-over-year returns of the following indices: US CPI Urban Consumers YoY NSA, PACI Energy TR Sub-Index (Energy), PACI Industrial Metals TR Sub-Index (Industrial Metals), PACI Precious Metals TR Sub-Index (Precious Metals), PACI Grains TR Sub-Index (Agriculture). Period: January 1994 (22 years) and January 2004 (12 years), respectively, until December Source: Picard Angst Ltd., Bloomberg LP.

27 Inflation Sensitivity of Commodity Sectors Significant variation of sector betas over time Energy Industrial Metals Precious Metals Agriculture Beta calculations based on a rolling four year windows of monthly year-over-year returns of the following indices: US CPI Urban Consumers YoY NSA, PACI Energy TR Sub-Index (Energy), PACI Industrial Metals TR Sub-Index (Industrial Metals), PACI Precious Metals TR Sub-Index (Precious Metals), PACI Grains TR Sub-Index (Agriculture). Period: January 1997 to December Source: Picard Angst Ltd., Bloomberg LP.

28 Inflation Sensitivity Driven Commodity Allocation Strategy Dynamic weighting of sectors by inflation beta 1' Inflation-Linked Commodity TR BCOM TR Strategy portfolio rebalanced monthly to new sector weights (capped at a maximum of 70% per sector) based on the four year rolling betas of the following indices vs US CPI Urban Consumers YoY NSA: PACI Energy TR Sub-Index, PACI Industrial Metals TR Sub-Index, PACI Precious Metals TR Sub-Index, PACI Grains TR Sub-Index. Comparison strategies: Picard Angst Commodity TR Index (PACI TR), Bloomberg Commodity TR Index (BCOM TR). Period: January 1997 to May Source:, Bloomberg LP.

29 Inflation Sensitivity Driven Commodity Allocation Strategy Evolution of sector weights over time 100% 90% 80% 70% Energy Industrial Metals Precious Metals Agriculture 60% 50% 40% 30% 20% 10% 0% Strategy sector weights capped at a maximum of 70%. Period: January 1997 to May Source:

30 Contents Commodities - State of the Market and Asset Class 3 Portfolio Contribution Diversification, Inflation Sensitivity, Risk Premium 10 Evolution of Commodity Beta Strategies 17 Alternative Risk Premia in Commodities 31 Summary 36

31 Commodity Beta vs Alternative Risk Premia Characteristics, return drivers, implementation Commodity Beta Alternative Risk Premia Characteristics cyclical return profile inflation protection diversification due to on average low correlation with traditional asset classes (equities, bonds) Return Drivers risk premium compensating for the assumption of general asset class risk (beta) absolute returns moderate to low volatility diversification due to persistently low correlation with traditional asset classes and commodity beta momentum value carry volatility etc. Implementation long only long/short market-neutral 31

32 Categories of Alternative Risk Premia in Commodities Sources of systematic commodity alpha The rich structure of commodity derivatives markets enables investors to access targeted alternative risk premiums by accepting exposure to specific risk factors. Such risk factors typically exhibit behaviour that is largely independent of the general trend in commodity markets (beta) Momentum Markets in commodities are highly cyclical. The economic cycle produces sustained divergences between supply and demand dynamics resulting in persistent trending behaviour of commodity prices during the up- and downswing phases of cycles Value Value strategies are intended to take advantage of situations of scarcity or excess of supply relative to demand. To that end, the inverse relationship between inventories and prices is exploited Carry / Seasonality Given the pronounced influence of weather and seasons on the demand and supply of energy and agricultural commodities strong seasonal elements are evident in their price dynamics. By arbitraging the term structure investors can take advantage and profit from this phenomenon Volatility Volatility premium strategies aim to earn the premium at which implied volatility is traded in option markets relative to realized volatility. In doing so they accept the risk of unanticipated peaks in realized market volatility 32

33 Example Risk Premium Strategy Combination of momentum and value factors vs commodity beta Return p.a. Volatility Sharpe Momentum - Value LS TR 13.09% 10.34% 1.06 Picard Angst Commodity TR 4.89% 17.02% 0.24 Bloomberg Commodity TR 1.24% 16.78% 0.03 Period: to Source:, Bloomberg LP 33

34 Benefits and Contribution at the Portfolio Level Diversification benefits to the commodities allocation and the overall portfolio Blending market-neutral commodity risk premium strategies with a core commodity beta portfolio presents a compelling value proposition considering the marginal degree of correlation with diversified commodity beta strategies such as the Bloomberg or Picard Angst Commodity Index (PACI) Furthermore, consistently uncorrelated behaviour relative to global benchmarks of traditional asset classes such as equities (as represented by the MSCI World Developed Markets Index) and bonds (as represented by the JP Morgan Global Aggregate Bond Index) in combination with an attractive risk-return profile naturally predisposes such systematic strategies to be used as portfolio diversifiers and return enhancer in its own right across asset classes Cmdty Alpha Cmdty Beta Equities Bonds Momentum - Value LS TR Commodity Beta - PACI Equities - MSCI World DM Bonds - JPM GABI 1.00 Calculation period: to Source:, Bloomberg LP 34

35 Blending with Commodity Beta Allocations Pronounced diversification benefits as the blend rate increases Return p.a. Volatility Sharpe Ratio 50% PACI / 50% CMV 9.52% 10.44% % PACI / 20% CMV 6.88% 13.97% % PACI 4.89% 17.02% 0.24 Period: to Quarterly rebalanced returns of a portfolio blending exposure to commodity beta as represented by the Picard Angst Commodity TR Index (PACI) and a market-neutral commodity risk premium strategy (CMV) combining momentum and value factor exposure. Source: 35

36 Summary From a price perspective commodities markets may be near a cycle low. Market prices of the majority of commodities have declined deep into marginal production cost curves. Given the tepid global growth outlook, however, any sustained recovery will be driven initially by rationalization of excess supply The recent cycle downswing is not out of line with the historical record. Factors such as diversification potential and inflation sensitivity that originally prompted pension funds to include commodities in their strategic asset allocation retain validity The overwhelming majority of institutional asset invested in commodities continue to track standard benchmarks such as the Bloomberg or S&P GSCI Commodity indexes or enhanced variations thereof despite obvious deficiencies in their construction In order to mitigate the pronounced return cyclicality inherent to the asset class, institutional investors are well-advised to look beyond pure beta investment solutions to commodities and consider adding exposure to targeted alternative risk premia offered by the asset class

37 Thank you for your attention Contact Information Dr. David-Michael Lincke, CFA, FRM Head of Portfolio Management Bahnhofstrasse Pfäffikon SZ, Switzerland Web: 37