ASSET ALLOCATION. In defence of complexity. Tommaso Mancuso, Head of Hermes Multi Asset. OUTCOME #14

ASSET ALLOCATION In defence of complexity Tommaso Mancuso, Head of Hermes Multi Asset OUTCOME #14 Our championing of shareholder rights led to greater transparency of a global carmaker's remuneration practices, shifting its corporate governance into a higher gear. For professional investors only www.hermes-investment.com

2 ASSET ALLOCATION INTRODUCTION The multi-asset category covers a of different approaches, from static balanced strategies to diversified growth strategies and hedge-fund-like absolute return strategies. While these approaches can differ substantially, they all share a common root: the well-established 60:40 equity-bond portfolio. For decades, professional investors have attempted to develop strategies to beat this simple approach by both expanding the universe to include new assets (e.g. credit, commodities, risk factors) and by asset timing through developing ever more sophisticated investment processes and portfolio construction methodologies 1. Our research suggests that neither the value of active management nor the benefits of Universes are stable through time. In fact, they appear to be cyclical. KEY POINTS uufor the past five years, the value provided by active and sophisticated asset-allocation strategies relative to passive and simple ones seem to have vanished uuour analysis suggests such that value is cyclical and tends to be higher during market inflection points uutherefore, in the context of asset allocation, a comprehensive shift to passive strategies would carry the risk of being short-sighted Historically, such added complexity has often been rewarded by higher or more consistent returns, which has spurred the rapid growth of multi-asset strategies. Yet, since the global financial crisis (GFC), it is quite likely that investors with an allocation to a traditional 60:40 portfolio will have outperformed more sophisticated and complex alternatives at least on an unlevered basis. The industry continues to debate the relative merits of active and passive approaches and at a general level an increasing number of investors are switching from active to passive strategies. In the context of multi-asset investing, the crucial question is whether active asset allocation and complex investment universes have had their day, or whether their relatively poor performance in recent years has just been temporary. If it s a case of the former, moving from a broadly active to passive allocation and from complex to simple universes should continue to pay off. But if it s the latter, such a decision could be short sighted because it would involve significant timing risk. In this paper we investigate the benefits of active relative to passive asset allocation, and of simple relative to expanded investment universes. Rather than evaluating the track record of selected funds, we opted to create simple generic quantitative strategies and backtested their performance from January 1996 to June 2017. Such an approach has the benefit of eliminating any selection bias, increasing the consistency of results through time, and allowing us to differentiate between the value of expanding the investment universe from the value of active asset allocation. We consider 10 strategies applying five different asset-allocation approaches separately to and Universes Passive, Risk Parity, Mean Reversion, Momentum and Mixed Active allocation (see the appendix for further details). Across these strategies we analyse overall performance, the isolated value generated from either an Universe or active allocation approach, and finally assess the value from combining an active approach with an Universe. These factors are evaluated first over the full test period and then by dividing the period into three cycles: pre-gfc, GFC to quantitative easing (QE) II and post-qe II. RESULTS FROM FULL TEST PERIOD: JANUARY 1996 TO JUNE 2017 Tables 1 and 2 show the risk and return statistics between January 1996 and June 2017 for the various strategies applied to the and Universes. The tables show the results across the five asset allocation methodologies: Passive, Risk Parity, Momentum, Mean Reversion and Mixed Active Allocation. Asset universe: we define a Universe (s) as consisting of two asset classes: equities and government bonds. We define an Universe (e) as consisting of 26 assets across four broad asset classes: equities, government bonds, corporate bonds and commodities. We use common market indices as proxies for these asset classes (see tables 10 and 11 in the appendix). 1 A review of the characteristics of different portfolio construction methodologies was covered in Portfolio construction methodologies: looking beyond the good, the bad and the ugly (Hermes Multi Asset, August 2015)

HERMES MULTI ASSET WHITE PAPER Q3 2017 3 Investment style: we define a Passive Allocation (PA) as a strategy that maintains a fixed strategic allocation to each asset over time. We define a Mixed Active Allocation (MAA) as an equal-weighted combination of three generic active asset allocation strategies: Risk Parity (RISK), Momentum (MOM), and Mean Reversion (MR). We describe these methodologies in Table 1 below. Table 1. Universe performance Jan 96 to Jun 17 Passive PA(s) Risk Parity RISK(s) Momentum MOM(s) Mean Reversion MR(s) Mixed Active Alloc. MAA(s) Ann. Return 5.5% 5.0% 6.5% 6.1% 5.9% Ann. 8.2% 3.0% 6.6% 8.3% 5.6% Volatility Absolute IR 1 0.68 1.65 0.99 0.74 1.06 Max Drawdown -31.5% -5.0% -16.8% -27.6% -16.9% Table 2. Universe performance Jan 96 to Jun 17 Passive PA(e) Risk Parity RISK(e) Momentum MOM(e) Mean Reversion MR(e) Mixed Active Alloc. MAA(e) Ann. Return 6.6% 5.9% 7.3% 7.4% 6.9% Ann. 6.1% 4.2% 5.8% 6.8% 5.3% Volatility Absolute IR 1 1.08 1.39 1.25 1.09 1.31 Max Drawdown -21.7% -12.5% -14.9% -17.9% -13.7% Comparing the results across asset universes, we note that the strategies investing in the Universe produced higher returns across all asset allocation methodologies combined with, in most cases, a lower level of risk. Interestingly, even for the passive approach, the Universe would have delivered significantly higher riskadjusted returns than the Universe. This benefit cannot only be simply ascribed to the increased diversification benefits that one would expect from investing in a broader universe (i.e. two v 26) in fact, the average cross-asset correlation of the Universe is 0.17, compared with -0.26 between equities and government bonds in the Universe. One possible explanation for the better results is the reduction of the market-cap bias inherent in the broad indices used as a proxy for the Universe. Comparing the results across investment styles, we note that Mean Reversion and Momentum deliver higher returns than Risk Parity and Passive. From a risk (i.e. volatility and drawdown) perspective, Risk Parity and Momentum produce better results, with Mean Reversion and Passive being the worst scenarios. Chart 1 shows the cumulative performance of each of the five construction methodologies applied to both the and Universes. Over this period, combining an Universe and active allocation approaches generated higher performance than a passive approach on a Universe, which is demonstrated by the two bold lines, MAA(e) and PA(s) respectively. Chart 1. Cumulative returns of each of the simulated portfolios during the test period NAV 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 Jan-96 Jan-97 Jan-98 Jan-99 Jan-00 Jan-01 Jan-02 Jan-03 Jan-04 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Jan-10 Jan-11 Jan-12 Jan-13 Jan-14 Jan-15 Jan-16 PA(s) PA(e) RISK(s) RISK(e) MOM(s) MOM(e) MR(s) MR(e) MAA(s) MAA(e) Source: Hermes and Bloomberg as of July 2017. We go on to use regression analysis to quantify the value added by investing in an Universe relative to a simple one, and of adopting an active asset allocation methodology relative to a passive one, first separately and then combined. Universe alpha: We define the value added by investing in an expanded relative to a simple investment universe as the average alpha that investing in an Universe would have generated across all investment styles (including passive). Universe alpha = AVG[ α(pa e ), α(risk e RISK s ), α(mom e MOM s ), α(mr e MR s ) ] Table 3. Regression of Universe alpha relative to PA(s) Jan 96 to Jun 17 Passive PA(e) Risk Parity RISK(e) Momentum MOM(e) Mean Reversion MR(e) Average ( Universe ) 2 2.61% 0.45% 2.00% 2.67 1.93% T-Stat 4 5.41 0.71 3.64 4.70 3.62 Beta 3 0.70 1.08 0.80 0.76 0.84 T-Stat 4 42.47 19.86 35.12 39.81 34.32 R-Squared 5 0.88 0.61 0.83 0.86 0.79 In all cases the Universe would have generated alpha relative to the Universe, although this effect was less pronounced for the Risk Parity approach than for the others. T-stat 4 scores greater than two indicate statistically significant results. On average, the Universe would have generated an annualised alpha of 1.93%. The average positive alpha combined with an average beta of less than one indicates that the excess returns delivered by the Universe is not caused by an increased level of risk. In fact, as we can see from comparing tables 1 and 2, this positive alpha from an Universe comes with the added benefit of lower volatility and drawdown.

4 ASSET ALLOCATION Active Asset Allocation : We define the value added by using an active asset allocation methodology as the average alpha of investing in an active relative to a passive strategy across both and Universes i.e. the average alpha from Risk Parity, Momentum and Mean Reversion approaches relative to the passive weighting. See table 4 for the results. Active Asset Allocation = AVG[α(RISK s ), α(risk e PA e ), α(mom s ), α(mom e PA e ), α(mr s ), α(mr e PA e )] Active Allocation : Finally, we determine the value of expanded active allocation as the alpha produced relative to the passive simple portfolio by combining an expanded universe with the MAA methodology, i.e. Mean Reversion, Momentum and Risk Parity combined (see table 4 for the results). Unlike Active Asset Allocation, this only incorporates the Universe and does not consider the active asset allocation methodologies applied to the Universe. It is as follows: Active Allocation = α(maa e ) Figure 4 shows the average statistics across and Universes for the three active approaches: Risk Parity, Momentum and Mean Reversion. It also shows the results for Active Asset Allocation and Active Allocation as defined above. Table 4. Regression of Active Styles, Active Asset Allocation and Active Asset Allocation to PA(s) Jan 96 to Jun 17 Risk Parity Momentum Mean Reversion Active Asset Alloc. Exp. Active Asset Alloc. 2 2.84% 1.87% 0.76% 1.82% 3.75% T-Stat 4 5.47 3.42 1.29 3.40 6.10 Beta 3 0.41 0.81 0.99 0.74 0.55 T-Stat 4 18.94 39.37 43.53 33.95 26.28 R-Squared 5 0.57 0.85 0.87 0.76 0.73 But the key question now is: does this hold true over all time periods? To determine this, we split the full timeframe that we considered previously into three shorter periods: 1: pre-gfc (January 1996 to June 2007) 2: GFC to Quantitative Easing (QE) II (July 2007 to October 2010) 3: QE II to date (November 2010 to June 2017). Firstly we revisit some of the performance results shown in tables 1 and 2 that were split into the three periods but focusing only on the portfolio that employs the MAA strategies within an Universe, or MAA(e), and the Passive Allocation within a Universe, or PA(s). Table 5 shows that MAA(e) would have generated markedly better risk-adjusted returns than PA(s) in both the pre-gfc and the GFC to post QE II periods. However, in the post-qe II period, the opposite has been the case. It is therefore unsurprising that we are currently seeing mounting discontent among investors in active strategies. Table 5. and passive v expanded and active approach pre-gfc Jan-96 to Jun-07 GFC Jul-07 to Oct-10 post-qe II Nov-10 to Jun-17 Jan 96 to Jun 17 Passive PA(s) Ann. Return 6.4% 8.0% Ann. Volatility 8.0% 4.6% Absolute IR 1 0.80 1.73 Max Drawdown -26.2% -4.7% Ann. Return -1.2% 6.0% Ann. Volatility 11.8% 7.6% Absolute IR 1-0.10 0.79 Max Drawdown -31.5% -13.7% Ann. Return 7.6% 5.5% Ann. Volatility 5.9% 5.0% Absolute IR 1 1.27 1.11 Max Drawdown -8.0% -6.9% Mixed Active Alloc. MAA(e) In all cases, active styles would have generated alpha relative to the passive approach. On average, active styles irrespective of their universe would have generated an annualised alpha of 1.82%, with a significantly lower beta of 0.74 relative to a passive simple approach. In other words, active styles generate alpha in part through risk reduction (i.e. beta less than one). A mix of the active styles combined with an Universe would have generated an impressive 3.75% annualised alpha. As we can see from Tables 1 and 2, the alpha from MAA(e) would have been generated with substantially lower levels of both volatility (5.3% v 8.2%) and drawdowns (-13.7% v -31.5%) than a passive portfolio investing in a Universe, or PA(s). To explore this further, we carry out a regression analysis considering the active and Universe elements of asset allocation against the passive Universe, similar to that shown above in tables 3 and 4. In this, we again calculate the Universe, the Active Asset Allocation and the Active Asset Allocation but this time split the dataset into the same three periods. The results and observations are shown below in table 6 and chart 2. The results so far suggest that between 1996 and June 2017, adopting an active approach to asset allocation and investing in an Universe would have produced superior risk-adjusted returns than using a fixed 60:40 allocation to equities and government bonds.

HERMES MULTI ASSET WHITE PAPER Q3 2017 5 Table 6. Regression of Universe, Active Asset Allocation and Active Asset Allocation to PA(s) pre-gfc Jan-96 to Jun-07 GFC Jul-07 to Oct-10 post-qe II Nov-10 to Jun-17 Jan 96 to Jun 17 Active Asset Alloc. Exp. Active Asset Alloc. 2 2.96% 3.53% 4.78% T-Stat 4 4.29 4.22 6.04 Beta 3 0.75 0.57 0.48 T-Stat 4 24.32 32.99 17.51 R-Squared 5 0.79 0.66 0.69 2 3.39% 4.92% 6.66% T-Stat 4 2.31 2.34 3.91 Beta 3 0.99 0.58 0.59 T-Stat 4 17.99 11.51 14.49 R-Squared 5 0.84 0.75 0.85 2-0.14% 0.40% 0.11% T-Stat 4-0.08 0.60 0.10 Beta 3 0.89 0.74 0.72 T-Stat 4 18.59 16.02 15.28 R-Squared 5 0.78 0.70 0.75 Chart 2. Three year rolling Universe, Active Asset Allocation and Active Asset Allocation (3 year rolling) 9.0% 8.0% 7.0% 6.0% 5.0% 4.0% 3.0% 2.0% 1.0% 0.0% -1.0% -2.0% -3.0% Dec-96 Apr-99 Aug-01 Dec-03 Apr-06 Aug-08 Dec-10 Apr-13 Aug-15 Active Asset Alloc. Universe Active Asset Alloc. associated with investing in an Universe. If QE was indeed the driver of the reduction of alpha associated with an Universe, that would also explain why alpha started to pick up again following the gradual normalisation of Federal Reserve policies that began with tapering in 2013. Active Asset Allocation : The alpha associated with active relative to passive asset allocation has been fairly variable through time. We can see that alpha is higher around major market inflection points and lower in the interim periods, as indicated by the troughs in chart 2. In other words, for those investors who are shifting from active to passive strategies, the cost of that decision may not be felt until the next crisis hits. The level of alpha fell significantly after 2010, although it appears to be rising from the lows of 2013. Why should this be the case? Chart 3 shows the cross-sectional standard deviation of annual returns for the 26 assets in the Universe. The average dispersion was 16.3% in the pre-gfc years, 17.2% during the GFC, and 12.1% since QE II, as indicated by the lighter blue line. We can see how the dispersion of returns has remained within a much narrower range since QE II. On the assumption that high dispersion between asset class returns represents fertile ground for active asset allocators, the reduction in asset class return dispersion since 2010 may help explain the reduction in alpha associated with active asset allocation in recent years. Chart 3. Universe: asset class return dispersion Cross asset dispersion of year-on-year returns 40% 35% 30% 25% 20% 15% 10% 5% 0% Dec-96 Jan-99 Feb-01 Mar-03 Apr-05 May-07 Jun-09 Jul-11 Aug-13 Sep-15 Universe : Chart 2 shows that the alpha associated with the Universe was fairly stable until the beginning of 2011, when it collapsed. The alpha generated seems to have bottomed in 2014 although it remains below its long-term average. Why should this be the case? It is quite possible that the additional liquidity generated by the expansion of QE (QE II was launched in late 2010) has had a rising tide lifts all boats effect, reducing the benefits Active Asset Allocation : We can see from chart 2 that the alpha associated with active asset allocation in an Universe is a direct result of both the alpha associated with investing in an Universe and with active asset allocation. This alpha is not stable over time, but is generally positive. It tends to be at its highest around major market inflection points. 1. Absolute Information Ratio (IR): IR is a measure of a portfolio s risk-adjusted return relative to a benchmark. In this paper we use absolute IR, which does not use a benchmark and instead represents the portfolio s annualised return over its annualised volatility. Regression statistics: 2. : Represents the intercept of the regression line and provides the expected return of a portfolio that corresponds to a zero return in the benchmark index, the passive portfolio investing in a simple universe. The excess return generated by the active portfolios over the simple passive portfolio is also dependent on the beta between the two portfolios. 3. Beta: Represents the gradient of the regression line. It is a measure of how much the variability of risk of the active portfolios can be explained by the variability of risk of the passive portfolio. A beta of 1 indicates the variability is fully explained by the passive portfolio while zero indicates no explanation. 4. T-stat: Used as a measure of the statistical significance of both the specified alpha and beta regression values between the two portfolios. The greater the T, the higher the confidence level and scores greater than 2 indicate a statistically significant relationship. 5. R-Squared: A statistical measure that represents the percentage of the active portfolio s movements that can be explained by movements in the passive simple method. R-squared values range from 0 to 1, and an R-squared of 1 would mean all movements of the active portfolio are completely explained by movements in the passive portfolio.

6 ASSET ALLOCATION CONCLUSION Our analysis suggests that neither the value of active asset allocation nor the benefits of investing in an Universe are stable through time. In fact, our results suggest that they are cyclical. Since hitting a low point in 2013, the benefits associated with an expanded asset universe appear to be resurfacing. Should the normalisation of monetary policies continue, we would expect them to increase further. The value associated with active asset allocation appears to be highly cyclical. In short, active managers seem to have most scope to produce significant alpha during market inflection points. We would also expect to see the benefits of active allocation rise somewhat should asset class return dispersion return to pre-qe II levels. One could draw the conclusion that more sophisticated multi-asset approaches had the potential to outperform in the pre-qe II years, while after this period the 60:40 portfolio has performed better albeit this trend has started to show signs of abating. Consequently, for those investors evaluating the benefits of passive relative to more complex approaches to asset allocation i.e. incorporating either active allocation and/or expanded universes we would not recommend making wholesale movements between the two as their cyclicality can make it challenging. Instead, a more prudent strategy would be to maintain a balanced allocation to both passive and active approaches. This would be most effective if the allocation to active managers focused on those who were truly active (i.e. they are unconstrained with a high level of flexibility). Furthermore, where seeking to change an approach, our suggestion is to keep it simple and to adopt a gradual approach to change. NOTES Investment universe: We aimed to select a universe that was simple but still representative of the choice that a typical asset allocator faces. Further work shows that the conclusions would stand even if we had expanded the universe s depth by increasing the amount of underlying assets or its breadth by including other assets such as risk factors. Trading costs: Our analysis does not take into account trading costs, but our conclusions would not be altered if they did. Even a passive strategy requires portfolio rebalancing. We estimate the average annual gross turnover (i.e. buy and sell) of the Passive strategy applied to the Universe to be 26% p.a. By comparison, we estimate the average annual turnover of the active styles to range from 39% p.a. for Risk Parity to 117% p.a. for Momentum. Assuming an average trading cost of 20 bps, the average impact to the returns caused by trading costs across each of the three active styles is 0.16% p.a. (and 0.23% p.a. for the most trading-intensive strategy) compared to 0.05% for the passive simple strategy. These costs were calculated by averaging the differences in the trading costs between each of the styles. Active management fees: The results do not take into account the fees associated with active management. In some cases, the effects of fees could alter our conclusions. In fact, it is possible that fees could eat up all of the alpha produced by active management. On the other hand, the active styles illustrated here are simplistic. Talented managers would argue that they deliver additional alpha over time to compensate for the added cost. Also, active allocation fees are declining throughout the industry. Systematic bias: We chose to use simple systematic strategies as a proxy for active management styles for the purpose of consistency through time and to avoid manager selection biases. It is possible that the results carry a systematic premium systematic approaches tend to work well over the long run and at major inflection points as they avoid behavioural biases. On the flipside, there is scope for talented discretionary managers to add substantial value over a mechanical process. The answer is manager-specific and we leave the reader to choose how to interpret our findings. Leverage and other constraints: Our analysis focuses on long-only, diversified, fully invested strategies. The ability to short or hedge, to be completely unconstrained by the strategic benchmark, and to use leverage or cash, represent additional sources of alpha relative to a passive approach. APPENDIX Overview of strategies Table 1 describes the generic strategies we have used as proxies for real-life strategies: Passive, Risk Parity, Momentum and Mean Reversion. These are intended to be simplistic. Our goal is to broadly replicate the key characteristics of active asset allocation strategies. While in practice real investment strategies tend to be more complex and often time-varying, most of them tend to incorporate elements of mean reversion, momentum or risk parity either singularly or combined.

HERMES MULTI ASSET WHITE PAPER Q3 2017 7 Table 7 Generic strategies Name Passive (PA) Risk Parity (RISK) Momentum (MOM) Mean Reversion (MR) Mixed Active Allocation (MAA) Methodology Fixed strategic weights Equal volatility weights Momentum driven Mean reverting weights Average active weights based on past returns performance universe Asset class allocation Asset class allocation weights are adjusted based on historical proportionally based Asset class allocation Asset class allocation momentum. If on each asset s past based on strategic based on equal risk momentum is positive return. Overweight Average performance of weights defined as: weighting using volatility. (negative) then the (underweight) if the RISK, MOM and MR Equities 60%, Portfolio rebased to target allocation is twice return in the previous Government Bonds 40% 100% (half) the fixed strategic year has been negative weights. Portfolio rebased (positive). Portfolio to 100% rebased to 100% universe Asset class allocation based on strategic weights defined as: Equities 40%, Government Bonds 40%, Corporate Bonds 15%, Commodities 5%. Sub-asset class components are equal weighted within each asset class Asset class allocation based on risk weights defined for the simple universe. Equal volatility weight across broader asset classes Same as above but with broader asset classses Same as above but with broader asset classses Rebalancing Monthly Monthly Monthly Calendar Year (January) Table 8. asset universe and average allocation Same as above but with broader asset classses Universe Passive Risk Momentum Mean Reversion Equities MSIC World Net TR Local Index 60% 19% 52% 58% Government Bonds Citigroup Group of 7 Global Bonds 40% 81% 48% 42% Total 100% 100% 100% 100% Table 9 asset universe and average allocation Universe Passive Risk Momentum Mean Reversion Equities MSIC World Energy Net TR Local Index 40.0% 4.0% 11.0% 0.9% 42.8% 4.2% 41.3% 3.8% MSIC World Materials Net TR Local Index 4.0% 0.9% 4.1% 4.5% MSIC World Industrials Net TR Local Index 4.0% 1.1% 4.2% 3.9% MSIC World Consumer Discretionary Net TR Local Index 4.0% 1.0% 4.3% 4.1% MSIC World Consumer Staples Net TR Local Index 4.0% 1.5% 4.6% 3.7% MSIC World Health Care Net TR Local Index 4.0% 1.3% 4.4% 3.6% MSIC World Financials Net TR Local Index 4.0% 1.0% 4.1% 3.7% MSIC World Information Technology Net TR Local Index 4.0% 0.8% 4.3% 5.9% MSIC World Telecommunication Services Net TR Local Index 4.0% 1.1% 4.4% 4.2% MSIC World Utilities Net TR Local Index 4.0% 1.4% 4.2% 3.8% Government Bonds Bloomberg/EFFAS Bond Indices Canada Govt 7-10yrs TR 40.0% 4.0% 45.9% 3.8% 39.9% 4.1% 39.7% 3.9% Bloomberg/EFFAS Bond Indices France Govt 7-10yrs TR 4.0% 4.1% 4.0% 4.1% Bloomberg/EFFAS Bond Indices Germany Govt 3-5yrs TR 4.0% 4.1% 4.0% 4.0% Bloomberg/EFFAS Bond Indices Germany Govt 7-10yrs TR 4.0% 3.8% 3.9% 4.2% Bloomberg/EFFAS Bond Indices Italy Govt 7-10yrs TR 4.0% 6.5% 4.0% 3.7% Bloomberg/EFFAS Bond Indices Japan Govt 7-10yrs TR 4.0% 3.9% 4.0% 3.9% Bloomberg/EFFAS Bond Indices UK Govt 7-10yrs TR 4.0% 2.1% 3.8% 4.2% Bloomberg/EFFAS Bond Indices US Govt 10+yrs TR 4.0% 6.1% 4.1% 3.8% Bloomberg/EFFAS Bond Indices US Govt 3-5yrs TR 4.0% 3.2% 3.9% 4.0% Bloomberg/EFFAS Bond Indices US Govt 7-10yrs TR 4.0% 8.3% 4.1% 3.8% Corporate Credit Barclays Global Investment Grade Corporate 15.0% 7.5% 33.8% 19.0% 13.7% 7.1% 14.1% 6.8% Barclays Global High Yield Corporate 7.5% 14.9% 6.5% 7.2% Commodities Bloomberg Agriculture Subindex Total Return 5.0% 1.3% 9.3% 2.6% 3.6% 0.9% 4.9% 1.2% Bloomberg Energy Subindex Total Return 1.3% 1.7% 0.9% 1.4% Bloomberg Industrial Metals Subindex Total Return 1.3% 2.4% 0.9% 1.2% Bloomberg Precious Metals Subindex Total Return 1.3% 2.6% 1.0% 1.2% Total 100.0% 100.0% 100.0% 100.0%

HERMES INVESTMENT MANAGEMENT We are an asset manager with a difference. We believe that, while our primary purpose is to help savers and beneficiaries by providing world class active investment management and stewardship services, our role goes further. We believe we have a duty to deliver holistic returns outcomes for our clients that go far beyond the financial and consider the impact our decisions have on society, the environment and the wider world. Our goal is to help people invest better, retire better and create a better society for all. Our investment solutions include: Private markets Infrastructure, Private Debt, Private Equity, Commercial and residential real estate High active share equities Asia, global emerging markets, Europe, US, global, and small and mid cap Credit Absolute return, global high yield, multi strategy, global investment grade, real estate debt and direct lending Multi asset Multi asset inflation Stewardship Active engagement, advocacy, intelligent voting and sustainable development Why Hermes Multi Asset? Hedge fund selection sophistication A focus on greater predictability of outcomes. Risk control Factor-based analyses improves diversification and reduces tracking error. Risk premia expertise Early adopter within alternative multi asset portfolios Asset selection expertise Systematic approach with pragmatic discretionary overlay. Leveraging off Hermes asset-specific expertise Combining the inflation expertise of our Inflation-Linked, Credit, Real Estate and Infrastructure teams. Offices London New York Singapore Contact information Business Development United Kingdom +44 (0)20 7680 2121 Africa +44 (0)20 7680 2205 Asia Pacific +65 6850 0670 Australia +44 (0)20 7680 2121 Canada +44 (0)20 7680 2205 Europe +44 (0)20 7680 2121 Middle East +44 (0)20 7680 2205 United States +44 (0)20 7680 2205 Enquiries marketing@hermes-investment.com This document is for Professional Investors only. The views and opinions contained herein are those of Hermes Multi Asset, and may not necessarily represent views expressed or reflected in other Hermes communications, strategies or products. The information herein is believed to be reliable but Hermes Fund Managers does not warrant its completeness or accuracy. No responsibility can be accepted for errors of fact or opinion. This material is not intended to provide and should not be relied on for accounting, legal or tax advice, or investment recommendations. This document has no regard to the specific investment objectives, financial situation or particular needs of any specific recipient. This document is published solely for informational purposes and is not to be construed as a solicitation or an offer to buy or sell any securities or related financial instruments. Figures, unless otherwise indicated, are sourced from Hermes. The distribution of the information contained in this document in certain jurisdictions may be restricted and, accordingly, persons into whose possession this document comes are required to make themselves aware of and to observe such restrictions. Issued and approved by Hermes Investment Management Limited ( HIML ) which is authorised and regulated by the Financial Conduct Authority. Registered address: Lloyds Chambers, 1 Portsoken Street, London E1 8HZ. HIML is a registered investment adviser with the United States Securities and Exchange Commission ( SEC ). BD00702 08/17 0001432 Certified ISO 14001 Environmental Management www.hermes-investment.com