September 2017 Risk parity allocation to major asset classes IN BRIEF Last winter, Finaltis published a trilogy of research letters focused on anomalies, risk premia and risk factors within European equities Based on our background as European equities investors and systematic futures traders, we extended the research and built a sample of 16 investable risk premia on main asset classes: equities, short-term rates, bonds, currencies or commodities AUTHORS: RÉMY CROISILLE Head of Research Senior Fund Manager CHRISTOPHE OLIVIER CIO In this research letter, we describe the construction of these risk premia and their characteristics in comparison with a classic universe of 70 traditional assets (futures contracts on equity indices, short-term rates, long-term rates, currencies, commodities) Finally, we allocate these risk premia using a risk parity approach to build a balanced portfolio We conclude that: Diversification across 16 risk premia is better than diversification across 70 traditional assets Although correlations between risk premia increased in 2007/2008, the overall levels remained lower than correlations between traditional assets From 2005 to September 2017, a systematic risk parity approach to allocate these risk premia leads to a portfolio delivering +144% annualized return with 100% volatility and a -257% maximum when: a traditional 50% equities / 50% bonds composite index yields +55% annualized return with 102% volatility and -39% maximum over the period; a risk parity allocation of traditional assets exhibits +80% annualized return, 91% annualized volatility and -267% maximum NICOLAS RENAUD Head of Risk
Building 16 investable risk premia on different asset classes Note: This sample of 16 risk premia is not exhaustive It is based on a qualitative analysis by the authors for the purpose of building a balanced portfolio Equity Risk Premia World Equity Market premium: This risk premium corresponds to the world equity market excess return above risk-free rate Academic research regularly deals with this topic: according to Dimson-Marsh Staunton (2010) quoted by Ilmanen (2011), the equity market risk premium yields around + 52% for the United States and + 44% for the World, with a range of + 25% to +68% per market between 1900 and 2009 To invest on this premium, we use future contracts on main equity indices (S&P 500, Euro Stoxx, Topix, FTSE, SMI, SP60, SPI200 and Hang Seng) to replicate the return of the MSCI World, adjusted for the US 3-Month yield Our risk premium delivered a +61% annualized performance between 2004 and 2017 European Equity - Size: the influence of the size on equity returns has been theoretically demonstrated by Fama-French (1993) To capture it, we quarterly build an equally weighted portfolio of the 25% smallest equities of the Stoxx 600, hedged by a Beta adjusted Stoxx 600 short position European Equity - Valuation: As for size (see above), the influence of valuation on equity returns has been theoretically validated by Fama-French study (1993) To capture it, we quarterly build an equally weighted portfolio of the stocks that are simultaneously among the 50% cheapest in terms of PER and the 50% cheapest in terms of Price / Book This portfolio is hedged by a Beta adjusted Stoxx 600 short position European Equity - Momemtum: Jegadeesh & Titman (1996), then Carhart (1997) demonstrated the influence of the Momentum factor on equity returns: recent winners tend to outperform losers in the short run To capture it, we quarterly build an equally weighted portfolio of the 25% last 12 month best perfomers hedged by a Beta adjusted Stoxx 600 short position European Equity - Quality: Piotroski (2000) demonstrated the influence of Quality factors on equity returns To capture this risk premium, we quarterly build an equally weighted portfolio of equities that are simultaneously above the median in terms of return on invested capital (ROIC) and profit margin (Profit Margin) This portfolio is hedged by a Beta adjusted Stoxx 60 short position European Equity - Low Volatility: We have already written several research letters about the low volatility anomaly To capture this risk premium, we quarterly build an equally weighted portfolio of the equities that belong to the 25% lowest in terms of volatility hedged by a Beta adjusted Stoxx 600 short position Global Equity Trends: This risk premium captures the trending behavior of equity markets To calculate this risk premimum, we applied a proprietary trend following program combined with a bubble risk indicator on a sample of 18 equity index futures contracts 2
Fixed Income & Credit Risk Premia Global Bond Market: This risk premium corresponds to the excess return above the risk-free rate of worldwide bond markets Ilmanen (2011) quoted figures from Bank of America / Merrill Lynch which suggested an average +18% pa premium between 1952 and 2009 However, it is worth noticing that the premium is negative between 1952 and 1980 and higher than 5% between 1981 and 2009, a period of overall general decrease in interest rates To invest on this premium, we use futures contracts on major sovereign bonds (10-year US, 20-year US, 10-year Japan, 10-year UK, 10-year Germany) to replicate the return of the Bank of America / Merrill Lynch 10-year Bonds Index, generating a +34% annualized return between 2004 and 2017 Short-Term rates Trends: This risk premium captures the trending behavior of short-term rates markets (from 3 month to 21 month maturities) We applied a proprietary trend following program combined with a bubble risk indicator on a sample of 16 short-term rates future contracts Long-Term rates Trends: This risk premium captures the trending behavior of long-term rates markets (maturities above 2 years) We applied a proprietary trend following program combined with a bubble risk indicator on a sample of 14 long-term rates future contracts Short-Term rates Spreads: This risk premium captures the slope of the short-term yield curve resulting from the overestimated forward yields compared to spot yields We built this premium by systematically being long in the fourth maturity of Euribor, Eurodollar and Short Sterling and being short in the second maturity of the same contracts Yield Curve: This risk premium benefits from the better risk-return ratio of the short-end of the yield curve compared to its long-end It can partially be explained by aversion to leverage : investors prefer or are forced to increase duration rather than apply leverage We build this premium by systematically taking long positions in 2-year or 5-year contracts hedged by sales of 10, 20 or 30-year contracts Exposures are adjusted on order to keep a neutral duration on each curve Currency Risk Premia Currency - Carry: As Ilmanen (2011) pointed out, buying high-yielding currencies and selling low-yielding currencies has been profitable for the past 50 years The strategy has been particularly profitable hence popular between 2000 and 2007 The 2008 losses were only made greater by players selling their positions simultaneously The bad timing of these losses associated with a strong negative skewness justifies a high risk premium To calculate it, we simulated a 50/50 program that purchases the top 25% yielding currencies and sells the last quartile through listed futures contracts to minimize trading costs Currency - Trends: This risk premium captures the trending behavior of currencies We applied a proprietary trend following program combined with a bubble detection system on a sample of 6 currency future contracts 3
Commodity Risk Premia Commodity Markets: This risk premium corresponds to the excess return above the risk-free rate of commodity markets Ilmanen (2011) quoted figures from S&P GSCI and mentioned a +38% pa risk premium between 1970 and 2009 (+137% pa in the 70s, +22% pa in the 2000s) As the S&P GSCI index is investable, it is simple to replicate these returns The performance is negative from 2004 to 2017 (-49% annualized) Commodity - Trends: This risk premium captures the trending behavior of commodity markets We applied a proprietary trend following program combined with a bubble risk indicator on a sample of 10 commodity future contracts Reference Universe & Composite Index For the purpose of this letter, our reference is a composite index composed of 50% MSCI World and 50% of the Bank of America / Merrill Lynch +10-year Government Bond Index (W9G1 Index) Our reference universe of traditional assets includes 70 futures contracts: 19 stock index futures contracts from all geographic areas, 10 bond futures contracts, 16 short-rate futures contracts, 17 commodity futures contracts and 8 currency futures Statistical analysis of our 16 risk premia From 2004 to 2017, the risk premia exhibit the following characteristics: Return Volatility Maximum Return/Vol Ratio Return/Max ratio World Equity Market premium 61% 170% -587% 036 010 European Equity Size 32% 77% -228% 042 014 European Equity Valuation 17% 60% -222% 028 008 European Equity Momentum 48% 59% -137% 082 035 European Equity Quality 59% 47% -110% 126 054 European Equity Low Volatility 52% 38% -117% 135 044 Global Equity Trends 34% 62% -126% 054 027 Global Bond Market 34% 61% -111% 055 030 Short-Term Rates Trends 37% 48% -93% 076 039 Long-Term Rates Trends 21% 52% -104% 041 021 Short-Term Rates Spreads 04% 16% -48% 027 009 4
Return Volatility Maximum Return/Vol Ratio Return/Max ratio Yield Curve 04% 24% -82% 018 005 Currency - Carry 37% 98% -240% 038 016 Currency - Trends 24% 47% -79% 052 031 Commodity Markets -49% 238% -829% -021-006 Commodity - Trends 38% 96% -321% 039 012 Composite index 50/50 55% 102% -393% 054 014 We computed the correlation matrix of these risk premia over the period 2004-2017; a strong positive correlation is shown in red, a low correlation is white and a strong negative correlation is blue ρ = +1 ρ = 0 ρ = -1 5
As shown by the correlation matrix, the risk premium universe exhibits a very good level of diversification, especially when compared with the correlation matrix of traditional assets ρ = +1 ρ = 0 ρ = -1 We are now focusing on the 2007-2008 period Although correlations increased, the overall levels between risk premia remained lower than between traditional assets 6
ρ = +1 ρ = 0 ρ = -1 Note: The correlation matrix for traditional assets over the 2007 2008 period is very similar on a color basis to the one for 2004-2017 A Risk Parity Allocation of our 16 Risk Premia The low average correlation level between our risk premia suggests to test a risk parity approach, ie weighting each risk premium according to the inverse of its volatility (with a quarterly rebalancing) The resulting returns are compared to those of the previously defined composite index (50% equities / 50% bonds) over the 2005 2017 period (2004 data are used to build the first portfolio) Compared with returns based on traditional assets or a 50/50 composite index, our risk premia portfolio clearly outperforms 7
Return Volatility Maximum Return/Vol Ratio Return/Max ratio Risk Premia - Risk parity 144% 100% -257% 144 056 Traditional Assets - Risk parity 80% 91% -267% 087 030 50/50 Composite index 55% 102% -393% 054 014 Conclusions We built a universe of 16 risk premia with a better level of diversification than a universe of 70 traditional assets Although correlations between risk premia increased over 2007 2008, the overall levels remained lower than the traditional assets correlations Between 2005 and 2017, a systematic allocation of these risk premia according to a risk parity approach produced a +144% annualized return with 100% volatility and -257% maximum These figures compare favorably with: 1 a traditional 50/50 composite index: +55% annualized return with a 102% annualized volatility and - 39% maximum ; 2 a risk parity allocation of traditional assets: +80% annualized return with 91% annualized volatility and -267% maximum 8
CONTACTS 63, Avenue des Champs Elysées 75008 Paris France wwwfinaltiscom Denis Beaudoin Tél: +33 (0)1 55 27 27 01 dbeaudoin@finaltiscom Thierry Rigoulet Tél: +33 (0)1 55 27 27 07 trigoulet@finaltiscom DISCLAIMER This document does not constitute a recommendation or investment proposal It was issued for information purposes only It has no contractual value and may contain errors and/or omissions Nothing contained herein shall in any way constitute an offer by Finaltis to provide any service or product, or an offer or solicitation of an offer to buy or sell any securities or other investment product Finaltis will not be liable for the content of these pages, nor for any use thereof or reliance placed thereupon by any person Product(s) described herein is/are not available to all persons in all geographic locations It is/ they are of a speculative nature so that its/their success may be affected by a fall in assets value as well as a total loss It/they may invest in OTC instruments which can be volatile and render difficult to predict or anticipate any fluctuation The institutions with which it/they will contract may encounter financial difficulties impairing the value of the product(s) There is no guarantee of capital preservation and minimum return There can be no assurance that the investment objectives shown are achieved Despite Finaltis EfficientBeta TM Euro not being restricted to professional investors as per the MIFID, this document contains information dedicated to professional investors As such, if you are not an investment professional, Finaltis recommends you stop reading this document and turn to your financial advisor for proper analysis and advice 9