Fundamental Indexation Usually, but not always, a value play July 2017 Prepared by Aon Hewitt Retirement & Investment
Summary This paper is the third in a series on alternative indexation so called smart beta investing. The intention to develop a framework to provide medium term views on the major smart beta factors. Fundamental indexation, a term coined by Research Affiliates, is not strictly a way of creating a value biased portfolio of stocks at relatively low cost, but it tends to do this in practice. As with other alternative indices, the core idea is to break the index s weighting scheme from the stock price, in this case, using information on a company s balance sheet and earnings strength, as well as its value relative to assets on the books. Value is probably the first factor to be identified, and has a large body of empirical evidence to support its existence. But value has not provided a premium over the past decade. As with all factors, the performance of value and, hence of the fundamental indices, is highly cyclical, with very long periods of underperformance. Value, and the RAFI index, especially is higher beta than the market cap index. This means that the indices are higher risk and will tend to underperform in times of market stress. This is an important issue to consider when deciding on investing. Returns have been driven by a few specific sectors and countries, such as energy and materials stocks, meaning that the indices can be overly exposed to stock or country specific developments that are not related to the value factor. This is a risk with most alternative indices and should be kept in mind by investors. Our medium term economic and market outlook indicates a moderate preference for the value factor relative to low volatility and quality focused indices. The ride will not be smooth, however, with variable headwinds to contend with, hence the small tilt. Fundamental indices, such as RAFI, are also preferred for their value and cyclical exposures. 2
What is Fundamental Indexation? Fundamental indexation is commonly classified as an alternative indexation (or smart beta) approach, which uses a weighting scheme that breaks the link between past performance and index weight. In contrast to standard market cap indices, where the largest companies have the largest weight in the index, companies with the highest rank in terms of a selection of other company and balance sheet metrics get the biggest weight in these indices. The unifying theme across all smart beta approaches is that they focus on factors, such as low volatility, quality, momentum or small companies, that are thought to offer a long term premium relative to market cap indices. In other words, portfolios weighted towards these factors have proven to outperform the market cap index over long periods of time and across different markets. Fundamental indices are normally thought to be an effective way to invest in so-called value stocks or those companies whose stock prices are lower than is justified by their underlying value. However, this assertion is controversial. We will test this assumption on pages 8 and 9. As the name suggests, the weighting criteria are related to the fundamentals of companies. These can vary a little across the major providers but generally include strong earnings or sales, along with valuable balance sheets at a reasonable price. In terms of the last criterion, the most common metric is the book value to share price ratio. There are many different providers of fundamental indices, but this note will concentrate on easily the most well-known of these, the Research Affiliates Fundamental Index or RAFI. The note will also contrast this, where appropriate, with the MSCI Value Weighted index, which we use to represent a more explicit value tilted index. We will be using the most global versions of both. Fundamental indexation looks for profitable and valuable companies with low stock prices. The main criteria for selection 3
How are the indices constructed? RAFI They start with a parent index and use an average of four metrics or "fundamental factors" sales, cash flow, book value and dividends to rank all the companies in descending order of this averaged fundamental measure (see flow chart). Apart from book value, the other three metrics are 5 year averages. At this point, the bottom most stocks are excluded, depending on the version and the weights of the companies in this new index are linked to the fundamental measure. Each version has its own parent index FTSE Global All Cap for the RAFI All World 3000 or the FTSE US All Cap for the RAFI US 1000, for example and the number of excluded stocks varies according to the number in the title. MSCI Value Weighted Index The MSCI Value Weighted index uses the three year averages of sales, earnings and cash earnings, along with the latest book value reading to determine its weights. There is no exclusion of stocks relative to the parent, market cap index. This contrasts with the confusingly named MSCI Value index, in that the latter still uses price related ratios, such as Price to Book Value and Price to Sales. Both the Value Weighted index and RAFI break the link to stock price. There is substantial overlap between the construction criteria of RAFI and the MSCI Value Weighted index. At face value, this hints that RAFI might have a bias towards value stocks. As we will see, however, the small differences do add up over time. RAFI construction 4
Where does RAFI fit amongst the other smart beta strategies? Statistics for selected MSCI passive smart beta indices compared with the MSCI AC World index MSCI AC World Minimum Volatility Quality Momentum Value RAFI No. of stocks 2476 363 498 492 2474 3043 Beta 1.00 0.62 0.88 0.95 1.03 1.10 10yr Sharpe Ratio 0.09 0.14 0.15 0.12 0.07 0.09 Tracking Error 0.00 7.36 4.31 7.83 3.17 2.57 Turnover % 2.74 20.66 23.79 147.96 16.27 18.35 Price/Earnings Ratio 20.72 21.14 21.53 25.20 17.10 14.66 Price/Book Value Ratio 2.22 2.74 5.27 2.85 1.49 1.59 Dividend Yield 2.41 2.58 2.05 2.23 2.83 2.98 Source: MSCI and Research Affiliates. All indices are the All Country World index versions, except RAFI 3000, which is also the most global version of the index All figures correct as at end May 2017 As the above table shows, both the MSCI All Country Value Weighted index ( Value ) and the FTSE RAFI All World 3000 index ( RAFI ) contain a large amount of stocks and certainly suffer less from concentration problems in comparison with the other prominent smart beta indices. They also both have a beta, or sensitivity to the market, of greater than one. This means that upwards or downwards movements in the wider market would tend to translate into a greater movement in the same direction. The key implication is that investors will be taking on extra risk by switching from the market cap index to either RAFI or Value. At the same time, we can see that the 10 year risk adjusted monthly returns, or Sharpe ratio, are broadly in line with the market cap index there certainly does not seem to be a clear outperformance on this measure at least. As with all smart beta strategies that employ a method of altering allocations and excluding stocks, the portfolio turnover figures are higher versus the market cap index. Regular buying and selling of stocks (the indices rebalance every 6 months) incurs transactions costs, which will eat into the performance of these indices. This is not excessive relative to active managers with typical turnover rates of 85% or more, but investors should be aware that this cost must be paid before returns can be earned. Finally, we can see from the valuation measures that RAFI is currently invested in cheap stocks by Price/Earnings and Price/Book Value standards, thus further supporting the view that RAFI gains the investor exposure to value stocks. RAFI is value oriented but is riskier than the market cap index, as evidenced by a higher tracking error and a higher beta. 5
How has RAFI performed? Since 2002, the RAFI 3000 has outperformed relative to the market cap index (here the MSCI All Country World Index, but results are similar with the FTSE All World index). $100 invested in 2002 in the market cap index would be worth $279 now, whereas the equivalent in the RAFI All World 3000 index would be worth $402 (see top chart). By contrast, the MSCI Value Weighted index, whilst also an outperformer, has not performed quite as well as the RAFI the same $100 would be worth $315 now. However, outperformance has been much harder to come by since the global financial crisis (see bottom chart). Over this 9 year period, the FTSE RAFI All World 3000 index has outperformed the MSCI All Country World Index by 7% points, which translates into an annual outperformance of only 76bps, without also accounting for the extra risk inherent in the RAFI. The Value index has actually underperformed over this period. Since the start of the year, performance has been even worse, with marked underperformance of 2.6% and 3.7% for RAFI and Value respectively. Indeed, we can see from many measures that the past decade has seen the longest period of underperformance by value stocks since data collection began over a century ago. The degree of outperformance relative to market cap for both RAFI and Value has weakened markedly since the financial crisis. RAFI and Value are long term outperformers but have struggled since the financial crisis 6
What are the drivers of RAFI? Value and small cap recently As the top chart of a returns attributions analysis shows, RAFI is currently heavily weighted towards value stocks, where value stocks are defined as those with low price to book value ratios. There is also a weighting towards small companies but this is a little misleading because only 10% of the parent index can be classed as small companies. Large companies still dominate the RAFI and the parent FTSE index. The theoretical underpinnings of both the value and small cap premiums are strong, having been identified over very long periods of time and across different regional markets. The value factor premium was also the first to be identified and value investors are very prevalent in the active management industry (the appendix sets out the theory behind both factors). Furthermore, looking at the performance of the RAFI index relative to the Value weighted index since 2014, we can see a very close correlation, thus supporting the view that RAFI has been biased sharply towards value stocks over this period (see bottom chart). This is in contrast to the data on pages 6 and 7 that clearly show a differential in performance between the two over a 10 or 15 year time horizon. There is clearly more going on here and it would be incorrect to think that all value focused indices are the same. The next page delves deeper into this issue. RAFI performance has been driven by an exposure to value and small cap stocks in recent times but this does not explain performance over longer periods. RAFI is strongly weighted towards value currently and performance over the last few years indicate a value bias 7
What are the drivers of RAFI? Why do RAFI and Value weighted indices perform differently? The purveyors of RAFI often say that they are not strictly a value tilted index, but the evidence is quite clear that the value factor is the leading driver of returns, both recently and over the longer term. However, there is clearly a difference in returns between RAFI and other, more explicitly value focused indices, such as our choice of the MSCI Value Weighed Index. The key here is that there are small, but important differences in the degree to which indices lean into their selection criteria. We can see this by comparing the exposure to stocks with high book value to price ratios and high sales to price ratios of RAFI versus the Value weighted index. As the top chart shows, the difference over time between the exposure of RAFI and Value Weighted to stocks with high book value to price is very similar over the past 10 years. Remember that book value is a core metric to select value stocks. But, the bottom chart reveals that the RAFI has a consistently higher exposure to stocks with high company sales than the Value Weighted exposure. Of course, these relative exposures may well be different if we compare the RAFI with other value focused indices, but the point remains the same. No value index will be the same and small differences can make a material difference over time. RAFI selection criteria focuses more on sales than Value Weighted but has a similar exposure to book value. RAFI and Value Weighted have very similar exposures to book value to price over time but there is a small and material difference in terms of sales to price 8
What are the drivers of RAFI? Digging deeper is it worth it to be more exposed to high sales companies? Given that RAFI is more heavily weighted to companies with high sales to price ratios, we have to ask whether this approach really has been the underlying driver of returns compared with the Value Weighted index. Historical data compiled by Bank of America Merrill Lynch for the S&P 500 index certainly suggests that it has. To start with, the top chart shows clearly that stocks with high book value to price ratios (or conversely, low price to book value ratios) took a sharp dive during the global financial crisis and have been tracking sideways ever since. Most financial sector companies fall into the category of high book value to price stocks. In other words, value indices were hurt by being overweight in financial stocks during the financial crisis and have faced headwinds since then. In contrast, stocks with high sales to price ratios started to outperform strongly after the financial crisis and have continued to make progress up to around 2014. RAFI has benefitted more from this than the Value Weighted index and, given the structurally higher weighting to companies with higher sales (using 5 year averages helps), we can assume that this will continue to be an important driver of returns. RAFI s larger bias to high sales companies has supported returns in the past and is an important differentiator to other value focused indices. High book value to price stocks have not been helpful but high sales to price stocks have done well 9
What are the drivers of RAFI? Small differences add up over time Another way of showing the impact of slightly different selection criteria is by comparing the sector weightings of the RAFI and Value Weighted indices. As the top chart shows, the RAFI has a higher weighting to energy and materials sectors and a lower weighting to financials and IT stocks. This means that RAFI will be more closely linked to commodities than the Value Weighted index. This was not always the case, however. As the bottom chart shows, in 2010, the RAFI had a significantly higher weight to financials stocks than the Value Weighted index, and a lower weight to energy stocks. Back then, the RAFI would not have been driven by commodity cycle factors relative to the Value Weighted index Also note the behaviour of the value exposures during the financial crisis that we can see in the chart on the previous page. The RAFI remained much more exposed to both high book value and high sales companies than the Value Weighted index over 2008. This is due to the longer period used when averaging the selection criteria (5 years versus 3 years). This has an impact on performance differences too. So which should you pick? This depends on whether a pure exposure to value stocks is required, because all the evidence shows clearly that the RAFI might not be a reliable, long term way of achieving this. RAFI weights are different enough to other value indices to become material over time, which will require regular review by investors. Bigger overweights to energy and materials in RAFI compared with value weighted at the moment Relative stock weightings vary over time 10
The risks of RAFI investing long periods of underperformance The first risk is applicable to all alternative indexation or smart beta approaches. This is that the performance of the factors being used as weighting criteria are highly cyclical in nature. Furthermore, these cycles can be very long, so although it can mean that outperformance relative to the market cap index can endure, it also means that underperformance can last several years too. For example, RAFI has underperformed the market cap index for the past 3 years (see top chart). However, as the table shows, RAFI has both outperformed the market cap index and the other alternative indices considered here over the past 12 month period. The performance of the others has been poorer, with 3 of them quality, minimum volatility and momentum actually underperforming. The key implication is that the timing of initial investments can be highly important in terms of relative performance, even over quite long periods of time. It also means that tilts to a portfolio of differently focused smart beta indices (on value, quality, low volatility etc) can enhance returns further. The highly cyclical nature of the relative performance of alternative indices increases the importance of timing initial strategic investments and that of medium term asset allocation to dynamically tilt exposures over time. It also means that combining different factors in a portfolio of alternative indices has merit. Timing can matter even over long time horizons RAFI performance moves around a lot, as it does for all factor indices Total returns for selected alternative indices, % 1 year 5 year 10 year 15 year MSCI (market cap) 22.3 76.2 53.2 213.6 RAFI 25.4 79.7 61.0 305.5 Value Weighted 22.8 73.2 42.4 224.0 Quality 21.7 84.2 110.8 285.5 Minimum Volatility 13.3 75.3 80.7 274.1 Momentum 19.1 90.1 85.3 308.4 Source: DataStream Note: All indices are the All Country World versions, except RAFI, w hich is the All World 3000 version 11
The risks of RAFI investing unintended exposures A second risk is the issue of unintended exposures. When investing in smart beta products, such as RAFI, the investor is consciously investing in an altered weighting scheme to the market cap index. There is nothing inherently wrong with this. The problem is that particularly large weightings to specific sectors or countries can increase risks beyond those related directly to the factor. For example, the RAFI index has a large overweight to energy, materials and financial companies, and an underweight to IT and Healthcare (see top chart). This would naturally mean that performance will be driven by commodity prices, which could change due to things completely unrelated to value factor specific developments. Examples include geopolitical tensions and supply disruptions. The same argument applies to country exposures. The bottom chart shows a large underweight to US stocks (of over 12%), whilst the largest overweight relative to market cap is in Japan. Again, it may well be the case that there is an overabundance of stocks meeting the RAFI criteria in the Japanese market, but these are unlikely to be accounting for country specific risk. So, if there is a change in Japanese legislation that improves the prospects and stock prices of the country s financial sector, can we really say that this is due to the value factor? Probably not! Equally, a Japanese recession could disproportionately affect the RAFI, which would be unintentional. Beware large relative weights in RAFI that expose the investor to sector and country specific risk. RAFI is overweight in energy, materials and financials and overweight in Japanese, UK and Brazilian stocks 12
The medium term outlook conditions point to moderate RAFI outperformance but this may take some time to become apparent The stage of the macroeconomic cycle, along with relative valuations, provide the best indication of medium term outlook for value stocks and, by extension, RAFI. In terms of the cycle, RAFI tends to outperform the market cap index when longer term interest rates are rising (see top chart). This is because interest rates normally rise when economic growth is picking up and company earnings are rising, or future expectations are improving. Value stocks outperform in this environment as earnings improvement is greater for them relative to the average and investors notice. An improving cycle also tends to mean higher commodity prices, and energy stocks tend to be cheap as we have seen earlier. Our view is that developed interest rates are likely to trend higher from here, and oil prices have some moderate upside over the medium term, both of which are supportive for value stocks and RAFI at the current time. But we must remember that macro risks are especially high currently, so our view is tempered. As for valuations, the gap between measures for the cheapest and most expensive stocks can provide a useful guide to turning points in relative performance. When the gap between cheap stocks and expensive stocks is especially large, value tends to perform better (see bottom chart). The gap was very large 12 months ago, over which time value stocks outperformed, but we do not see a large valuation gap at the moment. This means that relative valuations are not particularly supportive. Bond yields and RAFI relative performance are closely linked Value stocks tend to outperform when the gap between the cheapest and most expensive sectors is very wide 13
Implementation advice While we think there is a clear benefit to applying Medium Term Views to dynamically alter allocations over time, the key starting point has to be the strategic positioning of alternative indices and the underlying aims of clients. A combination of factor (or alternative) indices could be used to replace part or all of a passive portfolio. This is in order to mitigate the burdensome cyclicality of the returns of individual factors, whilst still achieving superior returns to the market cap index. Remember that relative return patterns are not synchronised across the key factors, such as value, low volatility and quality. From this perspective, we believe that RAFI warrants a place within such a portfolio of factors. Of course, some clients may be considering replacing an active manager with alternative indices in order to lower costs while maintaining a broad exposure to certain factors. This is feasible but clients should be aware that no passive index will provide a completely pure exposure to a factor. For example, a mechanical weighting process would not account for the myriad of qualitative aspects involved in picking value stocks some stocks are cheap for a reason! Furthermore, one should not consider investing in value indices or RAFI with the view of reducing portfolio risk since these styles are actually riskier than the market cap index in market downturns, the underperformance would be greater, not smaller. Downside risk mitigation would more effectively be achieved through investments in low volatility factor focused indices, such as the MSCI Minimum Volatility index. Investment timing can also play an important role, given the fluctuating relative performance of smart beta indices. This applies to both the initial strategic investment and altering allocations over time using medium term views. Asset allocation views are a crucial element in this decision making process. Finally, it should be understood that the true benefit of alternative indices will be borne out over long periods of time and so investments should be undertaken with long time horizons in mind. Please speak to your Aon Hewitt consultant for more details. 14
Appendix The theory behind the value and small cap premiums Value The value premium was the first to be identified, back in 1928 by Ben Graham and David Dodd, and has been found across several decades and in many different country markets. It was further identified and applied in the context of passive investing by Eugene Fama and Kenneth French in 1992. More recently, however, this return premium has fallen into negative territory, implying that the market cap index has been a consistently better choice over the past 10 year (see chart on next page). There are two broad theories one which supports the investment industry orthodoxy that markets are efficient and that prices reflect all available information, and another that challenges that orthodoxy. The former theory is that cheaper stocks are inherently riskier than average and this higher risk is rewarded with higher returns. The heightened risk of value stocks comes from large fixed assets, for example, that cannot be sold quickly in an economic downturn and that put the company in financial risk. Conversely, when economic demand picks up, these previously struggling companies are the quickest to respond because previously unproductive capital, such as mothballed factories or offices, can be put back on line with little extra cost. This is also why value companies tend to perform better when economic activity starts to pick up after a downturn. This theory has merit but we do not think it is a sufficient explanation for a number of reasons. Firstly, returns seem to be greater than can be explained simply by risk or volatility because the returns adjusted for risk of value stocks are also higher than market cap stocks. Secondly, a phenomenon such as the superior risk adjusted returns of low volatility stocks runs contrary to the theory that higher risk is always rewarded with higher returns. This is also supported by the historical data that shows a long term underperformance of the highest volatility stocks. The second theory, which uses investor behaviour as the key explanation, is perhaps more convincing. Investors routinely overestimate the prospects of growth stocks companies that are forecast to grow strongly and attach too much weight on earnings forecasts as opposed to current earnings and balance sheet strength. This so-called "overreaction effect" is in evidence as stock prices adjust sharply when earnings miss forecasts. If markets were efficient and stock prices reflected all available information accurately, the change in prices would not be so great. This theory is controversial too and empirical evidence is hard to come by. 15
Appendix The theory behind the value and small cap premiums Small Cap The small cap premium was identified in the 1970s by Rolf Banz in his paper, The relationship between return and market value of common stocks, and empirical evidence has shown that a portfolio of the smallest companies in the stock market has delivered superior risk adjusted returns relative to the market cap index over the long term. But, along with value, this outperformance has declined markedly over the past decade (see chart). In terms of the theory, smaller stocks are generally less well researched, are more difficult to trade (less liquid) and sometimes have lower reporting responsibilities to the market. They are therefore considered riskier investments. More risk means better expected return, at least according to traditional theory. Smaller firms are also more likely to be able to respond to rapid changes in demand, normally as the economy recovers from a downturn, and their earnings may grow more sharply. However, recent performance hints that there has been a change, at least in terms of research coverage, meaning that there may be less of a premium. Where has the return premium gone? 16
Contact List Koray Yesildag CFA Principal, Asset Allocation Specialist Consulting Global Investment Practice +44 (0)207 086 9605 koray.yesildag@aonhewitt.com 17
Disclaimer Investment advice and consulting services provided by Aon Hewitt Investment Consulting, Inc. ( AHIC ). The information contained herein is given as of the date hereof and does not purport to give information as of any other date. The delivery at any time shall not, under any circumstances, create any implication that there has been a change in the information set forth herein since the date hereof or any obligation to update or provide amendments hereto. This document is not intended to provide, and shall not be relied upon for, accounting, legal or tax advice or investment recommendations. Any accounting, legal, or taxation position described in this presentation is a general statement and shall only be used as a guide. It does not constitute accounting, legal, and tax advice and is based on AHIC s understanding of current laws and interpretation. This document is intended for general information purposes only and should not be construed as advice or opinions on any specific facts or circumstances. The comments in this summary are based upon AHIC s preliminary analysis of publicly available information. The content of this document is made available on an as is basis, without warranty of any kind. AHIC disclaims any legal liability to any person or organization for loss or damage caused by or resulting from any reliance placed on that content. AHIC. reserves all rights to the content of this document. No part of this document may be reproduced, stored, or transmitted by any means without the express written consent of AHIC. Aon Hewitt Investment Consulting, Inc. is a federally registered investment advisor with the U.S. Securities and Exchange Commission. AHIC is also registered with the Commodity Futures Trade Commission as a commodity pool operator and a commodity trading advisor, and is a member of the National Futures Association. The AHIC ADV Form Part 2A disclosure statement is available upon written request to: Aon Hewitt Investment Consulting, Inc. 200 E. Randolph Street Suite 1500 Chicago, IL 60601 ATTN: AHIC Compliance Officer Aon plc 2017. All rights reserved 18