A Fundamental Shift to Fundamental Indexing. Abstract

Transcription

1 Dr. William G. Droms, CFA Powers Professor of Finance McDonough School of Business Georgetown University Washington, DC A Fundamental Shift to Fundamental Indexing January 24, 2010 Revised 3/29/10 Forthcoming, Journal of Financial Services Professionals, July 2010 Abstract A new approach to constructing index funds is gaining increasing prominence in the neverending debate over active versus passive equity management. As total assets invested in fundamental index mutual funds approaches eight billion dollars, personal financial planners and private wealth mangers can expect to receive more inquiries from their clients about this relatively new phenomenon. In contrast to conventional index funds, fundamental index funds weight the percentage allocation of each stock in the index based on fundamental factors, such as total dollars of earnings or total dollars of dividends paid, rather than simply market capitalization. Whether or not fundamental indexing is a revolution in theory, the intuitively appealing logic in favor of fundamental indexing and the excellent performance of fundamental index funds to date should command the attention of individual investors and their advisors. Since the first FTSE RAFI 1000 Fundamental Index Fund was launched in December of 2005, FTSE RAFI and Wisdom Tree fundamental index funds have attracted over 7.5 billion dollars of investments in mutual funds and ETFs dedicated to the fundamental index strategy. On the institutional side, institutional investors have over 22 billion dollars invested that follow the Research Affiliates fundamental index models. The extraordinary success of this new investment concept most likely flows from the fact that fundamental indexing makes a lot of intuitive sense as an opportunity for risk reduction and return enhancement. Fundamental indexing is likely to prove to be a viable investment option for many years to come.

2 A Fundamental Shift to Fundamental Indexing A new approach to constructing index funds is gaining increasing prominence in the neverending debate over active versus passive equity management. As total assets invested in fundamental index mutual funds approaches eight billion dollars, personal financial planners and private wealth mangers can expect to receive more inquiries from their clients about this relatively new approach to indexing. Although eight billion dollars in open-ended and exchange-traded fundamental index funds is still relatively small in comparison to the estimated $604 billion held in conventional stock index funds (Investment Company Institute, 2010), the wide dissemination of information related to fundamental indexing is bringing increased visibility to the fundamental approach. The fundamental indexing story starts with a comparison to conventional indexing. Conventional indexing has been widely used as an investment strategy for over 40 years and involves constructing an index fund that is designed to exactly mimic some market index. The largest stock index funds are the funds that are constructed to exactly mimic the returns on the Standard and Poor s 500 Index. Conventional index funds are weighted by market capitalization, that is, they are structured so that they invest the fund s assets in exact proportion to the outstanding market capitalization of each stock in the index. In contrast to conventional index funds, fundamental index funds weight the percentage allocation of each stock in the index based on fundamental factors, such as total dollars of earnings or total dollars of dividends paid, rather than simply market capitalization. Proponents of fundamental indexing say that fundamental index funds weight their allocations based on Main Street factors rather than Wall Street factors. The logic underlying fundamental 1

3 indexing rests on the premise that because capitalization weighted indexes are weighted by dollars of outstanding market capitalization, simple arithmetic will result in cap weighted funds overweighting overpriced stocks and underweighting underpriced stocks. This logic says that stocks trading at prices above their underlying intrinsic values will have be allocated a larger percentage of the index fund s investment than should be allocated based on what the total outstanding fair market value of the company s equity should be. Similarly, stocks trading below fair value will be underweighted because their outstanding market capitalization weights will be less than they should be. The ambivalence of the investment community toward the efficacy of fundamental indexing is reflected clearly in the title of the article that won the prestigious Graham and Dodd award in 2008 for the best article published in the Financial Analysts Journal: Why Fundamental index Might or Might Not Work authored by Paul Kaplan, CFA, Director of Quantitative Research at Morningstar (Kaplan 2008) Opponents of fundamental indexing maintain that since fair value can only be estimated and not determined exactly, fundamental indexing rests on a faulty premise that attempts to take advantage of deviations from the unknown fair value of each stock in the index. Since fair value is unknown, the extent of deviation from fair value must also be unknown. Furthermore, critics contend, overpriced stocks and underpriced stocks will cancel each other out and in the aggregate, the market portfolio will be fairly priced. Hence, investors who own index funds that vary from market cap weighted indexing will of necessity underperform the cap weighted indexes. The intellectual father of the fundamental index fund movement is widely considered to be Robert Arnott, founder of Research Affiliates, the institutional investment firm that sponsors 2

4 the Research Affiliates Fundamental Index TM (RAFI). The Fundamental Index methodology involves selecting and weighting individual stocks in a portfolio using a composite of four fundamental factors: book value, income, sales and dividends (Arnott 2005). These four factors aim to capture the economic size of component companies as the appropriate way of weighting stocks in an index portfolio, rather than simply weighting by market capitalization. The back tested data presented by Arnott show that over the 1962 to 2004 time period, an index constructed from a variety of fundamental factors would have outperformed the market cap weighted S&P 500 Index by approximately 1.9 percent per year (compounded). Research Affiliates partners with FTSE TM (FTSE is a joint venture of the Financial Times and the London Stock Exchange) to publish FTSE RAFI indexes globally using the universe of FTSE companies. In addition to Arnott s original article, the theoretical case in favor of fundamental indexing is laid out in an article by Jack Treynor, president and CEO of Treynor Capital Management and former long time editor of the Financial Analysts Journal (note: Treynor serves on the Advisory Board of Research Affiliates; Nobel Prize winner Harry Markowitz also serves on the board). Treynor s argument essentially states that stocks market prices deviate from their actual fair values and these valuation errors are symmetrical: overpriced stocks in a given stock index are counter-balanced by underpriced stocks and these valuation errors are symmetrical (Treynor 2005). Because of these offsetting errors, the market cap weights of stocks in a given index are above fair value and market cap weights of underpriced stock are below fair value. Hence, an index fund that weights stocks in an index by their market caps will overweight overpriced stocks and underweight underpriced stocks. Using Treynor s term, a market value indifferent weighting scheme should outperform market cap weighting. 3

5 The case against fundamental indexing has been articulated by Harvard Business School Professor Andre Perold (Perold 2007) (note: Professor Perold is a member of the Board of Directors of the Vanguard Group and also a founder of High Vista Strategies, a firm specializing in endowment management that offers a fund that is broadly diversified across marketable asset classes, hedge funds and private investments). Perold s argument is that fair values for stocks are unknown and that since the collective holdings of all investors must aggregate to the market portfolio, using market value indifferent weighting is a zero-sum game: for every investor who is underweight a stock, another investor is overweight. In the aggregate, then, after deducting fees and transaction costs, the average investor who deviates from market capitalization weights must by definition underperform the market portfolio. The crux of the issue, according to Perold is that because market capitalization does not reveal whether a stock is over-valued or undervalues, the random mispricing of stocks does not systematically shift the portfolio weights toward overvalued stocks. The back tested data from Research Affiliates offer strong support for the fundamental index argument. Research Affiliates has simulated the performance of seven RAFI strategies by back testing over periods of ten to 47 years (all periods end as of 6/30/09). The results of these simulated returns are summarized in Table 1. As can be seen in the table, these strategies show that simulated fundamental index funds have produced higher returns at approximately the same volatility level for all simulated strategies. Research Affiliates maintains that as markets become less efficient, e.g., as one moves from the market for large cap U.S. stocks to emerging market stocks, the out performance of the fundamental index approach compared to market benchmarks will become more pronounced. The data in Table 1 support this contention. 4

6 Table 1 shows that for the 47.5 year period ending 6/30/09, the simulated RAFI US Large Cap portfolio outperformed the S&P 500 Index by 2.1% per year, with a standard deviation of 15.2% compared to 15.1% for the Index. By contrast, in what would presumably be the least efficient market, the RAFI Simulated Emerging Markets portfolio outperformed the MSCI Emerging Market Index by 10.2% per year, with a standard deviation of 25.3% compared to 24.9% for the Index. The table shows that this return spread increases as one progress from the RAFI US Large through the Developed 1000, Europe, Japan, US Small Cap, and MSCI EAFE to the Emerging Market portfolio. Standard deviations of return are very close in all cases so that the Sharpe Ratios of the RAFI strategies are consistently higher than the underlying indexes. Table 1 Here The most common criticism of the fundamental index approach is the the performance differential between fundamental index funds and cap weighted funds is due primarily to the value tilt of the RAFI approach. Research Affiliates maintain that their approach has a value tilt, but that the value tilt is unique in that the RAFI approach uses a dynamic value tilt. During periods of strong growth performance relative to value, as high P/E growth stocks increase their market capitalization relative to value stocks, RAFI s value tilt increases and when the market turns such that value begins to outperform growth, RAFI s value exposure decreases. In support of their contention that the fundamental index results are not due solely to value, Research Affiliates points to their simulated results for , which show RAFI US Large Index outperforming the Russell 1000 Value Index by 1.79% per year (13.85% with 15.49% standard 5

7 deviation compared to 12.06% with 14.89% standard deviation; for the same period, the S&P 500 returned 11.47% with 15.44% standard deviation). Since live funds using fundamental indexation have been in operation only since December 2005, the data are limited, but so far show excellent returns. The oldest publically offered fundamental index fund, the FTSE RAFI US 1000 Index (ticker PRF), first offered by Invesco Powershares in December 2005, is an exchange-traded fundamental index fund of the 1,000 largest U.S. stocks based on a composite of the four fundamental factors (book value, income, sales and dividends). Comparing PRF to the Russell 1000 Index (based on market cap) shows that the performance of the FTSE RAFI 1000 relative to a conventional market cap weighted index of 1,000 stocks has been mixed. PRF outperformed the Russell 1000 in 2006 and 2009, underperformed in 2007 and 2008, but had better total returns for the entire four-year period. The year just ended has the most dramatic performance gap: PRF returned a remarkable 41.75% in 2009, compared to 28.43% for the Russell For the four years ended 12/31/09, PRF returned an average annual compound return of 0.74% per year versus a loss of 0.54% per year for the Russell This return advantage of 1.28% per year is in line with what fundamental index proponents expect to see over multi-year holding periods. In total, Invesco Powershares offers nine fundamental index exchange-traded funds that are indexed to the FTSE RAFI indexes. In the mutual fund universe, Schwab offers four FTSE RAFI funds: a U.S. large cap 1000, a U.S. small/midcap, an international large cap, and an emerging markets index fund. As of the end of 2009, there was approximately $1.1 billion invested in the Powershares Fundamental ETFs and approximately $1.4 billion invested in the Schwab Fundamental Funds. 6

8 The main competitors for the FTSE RAFI funds are the exchange-traded funds offered by Wisdom Tree. Wisdom Tree lists Jeremy Siegel, the prominent Wharton finance professor, as the fund group s Senior Investment Strategy Advisor, but their website notes that Professor Siegel is not involved in the day-to-day operation of the Wisdom Tree Fundamental ETFs. Wisdom Tree offers an array of fundamental index funds that allocate the weights to funds in the index according to earnings or dividends, rather than using a composite approach like FTSE RAFI. Based on Wisdom Tree s research of fundamental factors, they have concluded that earnings and dividends offer the most effective metrics of fiscal fitness for companies in their fundamental index funds. Wisdom Tree earnings ETFs only include companies with positive earnings, thereby, according to Wisdom Tree, reducing index volatility by screening out money-losing and speculative companies. Wisdom Tree cites research by Jeremy Siegel showing that weighting by earnings can reduce a portfolio s overall P/E ratio. This approach also relies on literally decades of research showing that low P/E stocks within the S&P 500 Index have outperformed the S&P 500 Index as a whole over most long-term time periods. The dividend ETF rests on the premise that cash dividends provide an objective measure of a company s value and profitability that cannot be manipulated. The dividend ETF also rests on the premise that weighting by dividends should raise the ETF s dividend yield relative to its underlying index and that many research studies have shown that the highest dividend-yielding stocks within the S&P 500 have outperformed the index. Wisdom Tree also cites Siegel s research that from 1926 through 2007, reinvestment of dividend accounted for 96 percent of the stock market s total return after inflation. 7

9 Wisdom Tree offers a much larger number of funds than are available using the FTSE RAFI approach, seemingly slicing and dicing various combination and permutations of earnings and dividend weighting schemes, albeit with a much heavier reliance on dividends than earnings. All told, Wisdom tree offers 7 domestic earnings based ETFs, 6 domestic dividend based ETFs, 19 earnings or dividend based international ETFs, and 10 international dividend sector ETFs. Assets under management in the 32 domestic and international earnings and dividends funds totaled $4.7 billion at the end of Two of Wisdom Tree s largest funds use the stocks in the S&P 500 Index as their investment universe and weight each fund s stock allocation according to either earnings or dividends. The Wisdom Tree Earnings 500 Fund (EPS) allocates weights based on total dollars of company earnings (hence the fund s ticker symbol, EPS). The Wisdom Tree Large Cap Dividend Fund (DLN) weights allocations according to total dollars of dividends paid by S&P 500 companies. Wisdom Tree also offers earnings and dividend weighted funds indexed to the Russell 3000, Russell 2000 and Russell 1000 Indexes, the S&P 400 (Midcap) Index and several other indexes. Looking at the S&P 500 Earnings and Dividend Funds, one would expect to see different performance than the large cap Schwab and Power Shares funds both because the Wisdom Tree funds select from a narrower investment universe (the S&P 500 rather than the FTSE RAFI 1000) and because these two funds weight according to only one variable at a time rather than a composite of five variables EPS outperformed the S&P 500 Index by a razor thin 0.13% in 2008 and finished % ahead of the S&P 500. The large cap dividend ETF outperformed the S&P 500 Index by approximately 1.8% in 2008, but underperformed the S&P by over nine 8

10 percentage points in 2009 (17.31% for DLN compared to 26.46% for the S&P). Table 2 provides a summary of the investment performance of the ETFs and mutual funds indexed to RAFI and the Wisdom tree ETFs (excluding the Wisdom Tree international dividend sector ETFs). Differences in the performances among the funds obviously are driven by the choice of index method and to date the fundamental indexing approach appears to be living up to it proponents expectations. For the two Powershares FTSE RAFI funds that have been in operation for the last three full years, performance has been above their benchmark indexes. The US 1000 fund averaged a loss of 4.71 percent for the three years ended on 12/31/09, compared to a 5.63 average annual loss for the Russell 1000 Index. The FTSE RAFI 1500 Small Mid Cap Index lost 1.46 percent per year compared to a 1.83 percent loss for the S&P Midcap 400 Index and a 4.79 percent loss for the S&P Smallcap 600 Index. Table 2 Here For calendar year 2009, the outperformance of the FTSE RAFI funds is truly remarkable and is not likely to be repeated absent another 100 year flood debacle like the equity markets experienced in The FTSE RAFI 1000 returned percent and the FTSE RAFI 1500 Small Mid Index returned percent compared to 28.43, and percent, respectively, for the Russell 1000, S&P Midcap and S&P Smallcap Indexes. The RAFI funds generally are rebalanced once a year in mid to late March. In 2009, the equity markets bottomed out on March 9, so the RAFI March rebalance heavily weighted the RAFI funds in the most beaten down stocks in the most beaten down market in decades. Hence the RAFI funds 9

11 benefited greatly when the markets rebounded from the March 9 bottom (the S&P 500 Index bottomed at on 3/9/09 and recovered to close at on 12/31/09, an increase of 64.8%). This type of equity market plunge followed by a spectacular recovery is unlikely to reoccur, and if it does, the timing and impact of rebalancing is totally unpredictable. For practitioners, Kaplan s conclusion (Kaplan 2008) in regard to fundamental indexing may be the most relevant: They may have a successful investment strategy, but they have not produced a revolution in investment theory. Whether or not fundamental indexing is a revolution in theory, the intuitively appealing logic in favor of fundamental indexing and the excellent performance of fundamental index funds to date should command the attention of individual investors and their advisors. Since the first FTSE RAFI 1000 Fundamental Index Fund was launched in December of 2005, FTSE RAFI and Wisdom Tree fundamental index funds have attracted over 7.5 billion dollars of investments in mutual funds and ETFs dedicated to the fundamental index strategy. On the institutional side, institutional investors have over 22 billion dollars invested that follow the Research Affiliates fundamental index models. Fundamental indexing offers an intuitively appealing option to further diversify investment assets dedicated to indexing strategies and is likely to be of interest to personal wealth mangers for many years to come. Both the back tested and live performance results support the position that private wealth managers should at least consider the possibility of using the fundamental indexing approach for the passively managed portions of client portfolios. Of course, if stock markets of the future evolve to a higher level of efficiency than the apparent level of efficiency during the back tested periods, then one would expect to see the value of fundamental indexing compared to market cap indexing decline. 10

12 References Arnott, Robert, Jason Hsu and Phillip Moore, Fundamental Indexation, Financial Analysts Journal, March/April 2005, (Vol. 61, No. 2), pp Investment Company Institute, 2009 Investment Company Fact Book, Investment Company Institute, 2010). Kaplan, Paul, Why Fundamental Indexation Might - or Might Not - Work, Financial Analysts Journal, January/February 2008, (Vol. 64, No. 1), pp Perold, Andre, Fundamentally Flawed Indexing, Financial Analysts Journal, November/December 2007, (Vol. 63, No. 6), pp Siegel, Jeremy, The Future for Investors, Crown Publishing Group, Treynor, Jack, Why Market Valuation Indifferent Indexing Works, Financial Analysts Journal, September/October 2005, (Vol. 61, No. 5), pp

13 Table 1: Simulated RAFI Fundamental Indexes vs. Market Benchmarks Simulated RAFI Data Start Std RAFI - % 3-YR Total Return Data as of 6/30/09 Date Return % Dev % Bench Wins Simulated RAFI US Large 1/1/ % 15.2% 73.6% S&P % 15.1% 2.1% Simulated RAFI Developed /1/ % 15.0% 83.4% MSCI Global 9.2% 15.5% 2.8% Simulated RAFI Europe 1/1/ % 18.0% 94.1% MSCI Europe 11.5% 17.7% 2.9% Simulated RAFI Japan 1/1/84 5.4% 19.7% 90.8% MSCI Japan 2.2% 20.1% 3.2% Simulated RAFI US Small 1/1/ % 19.7% 99.7% Russell % 19.9% 3.8% Simulated RAFI International Small 1/1/ % 17.4% 95.6% MSCI EAFE Small Cap 6.5% 19.4% 5.6% Simulated RAFI Emerging Markets 1/1/ % 25.3% 100.0% MSCI Emerging Markets 4.7% 24.9% 10.4% % 3-YR Wins: percent of three year periods where RAFI outperformed benchmark Source: Research Affiliates 12

14 Table 2: Fundamental Index Funds Performance Data Total Return Data as of 12/31/09 $ MM Total Return % Inception Wisdom Tree U.S. Earnings ETFs Ticker AUM Date 3-YR 4-YR Wisdom Tree Earnings 500 Fund NAV EPS (36.64) n.a. n.a. 2/23/07 n.a. n.a. Wisdom Tree Earnings Top 100 EEZ (45.04) n.a. n.a. 2/23/07 n.a. n.a. Wisdom Tree Total Earnings EXT (36.82) n.a. n.a. 2/23/07 n.a. n.a. Wisdom Tree Large Cap Value EZY (41.35) n.a. n.a. 2/23/07 n.a. n.a. Wisdom Tree Midcap Earnings EZM (35.10) n.a. n.a. 2/23/07 n.a. n.a. Wisdom Tree Small Cap Earnings EES (32.39) n.a. n.a. 2/23/07 n.a. n.a. Wisdom Tree Large Cap Growth ROI n.a. n.a. n.a. 12/4/08 n.a. n.a. Wisdom Tree U.S. Dividend ETFs Wisdom Tree Total Dividend DTD (34.76) 1.16 n.a. 6/16/06 n.a. n.a. Wisdom Tree Large Cap Dividend DLN (35.04) 2.62 n.a. 6/16/06 n.a. n.a. Wisdom Tree Midcap Dividend DON (32.25) (5.88) n.a. 6/16/06 n.a. n.a. Wisdom Tree Small Cap Dividend DES (28.13) (12.88) n.a. 6/16/06 n.a. n.a. Wisdom Tree Equity Income DHS (38.84) (4.61) n.a. 6/16/06 n.a. n.a. Wisdom Tree Dividend ex-financials DTN (36.21) (0.05) n.a. 6/16/06 n.a. n.a. Wisdom Tree International Div/Earn ETFs Wisdom Tree International Small Cap Dividend DLS (44.37) 3.98 n.a. 6/16/06 n.a. n.a. Wisdom Tree Europe Total Dividend DEB (42.60) n.a. 6/16/06 n.a. n.a. Wisdom Tree Emerging Markets Small Cap Div DGS (45.75) n.a. n.a. 10/30/07 n.a. n.a. Wisdom Tree DEFA Equity Income DTH (43.99) n.a. 6/16/06 n.a. n.a. Wisdom Tree International Div ex-financials DOO (45.12) n.a. 6/16/06 n.a. n.a. Wisdom Tree Pacific ex Japan Total Dividend DND (45.68) n.a. 6/16/06 n.a. n.a. Wisdom Tree International Midcap Dividend DIM (41.05) 7.64 n.a. 6/16/06 n.a. n.a. Wisdom Tree International Large Cap Div DOL (39.78) n.a. 6/16/06 n.a. n.a. Wisdom Tree Pacific ex Japan Equity Income DNH (47.17) n.a. 6/16/06 n.a. n.a. Wisdom Tree Japan Small Cap Dividend DFJ (12.55) (10.61) n.a. 6/16/06 n.a. n.a. Wisdom Tree Japan Total Dividend Fund DXJ (20.81) (7.33) n.a. 6/16/06 n.a. n.a. Wisdom Tree International Real Estate DRW (56.82) n.a. n.a. 6/5/07 n.a. n.a. Wisdom Tree Global Equity Income DEW (43.39) 8.31 n.a. 6/16/06 n.a. n.a. Wisdom Tree Europe Small Cap Dividend DFE (52.37) (2.85) n.a. 6/16/06 n.a. n.a. Wisdom Tree World ex US Growth DNL (12.71) (7.77) n.a. 6/16/06 n.a. n.a. Wisdom Tree Emerging Markets Equity Income DEM (34.56) n.a. n.a. 7/13/07 n.a. n.a. Wisdom Tree India Earnings EPI n.a. n.a. n.a. 2/22/08 n.a. n.a. Wisdom Tree Middle East Dividend GULF n.a. n.a. n.a. 7/16/08 n.a. n.a. Wisdom Tree Subtotal 4,706 Powershares FTSE/RAFI ETFs Powershares FTSE RAFI US 1000 PRF (39.97) /19/05 (4.71) 0.74 Powershares FTSE RAFI US 1500 Small Mid PRFZ (39.05) (1.05) n.a. 9/20/06 (1.46) n.a. Powershares FTSE RAFI NASDAQ Small Cap PQSC n.a. n.a. n.a. 4/3/08 n.a. n.a. Powershares FTSE RAFI Asia Pacific ex-jp PAF (48.20) n.a. n.a. 6/25/07 n.a. n.a. Powershares FTSE RAFI Dev Mkts ex-us PXF (42.39) n.a. n.a. 6/25/07 n.a. n.a. Powershares FTSE RAFI Dev Mkts ex-us Sm-Mid PDN (41.24) n.a. n.a. 9/27/07 n.a. n.a. Powershares FTSE RAFI Emerging Markets PXH (45.28) n.a. n.a. 9/27/07 n.a. n.a. Powershares FTSE RAFI Europe PEF (48.73) n.a. n.a. 6/25/07 n.a. n.a. Powershares FTSE RAFI Japan PJO (25.98) n.a. n.a. 6/25/07 n.a. n.a. Powershares Subtotal 1,126 Schwab FTSE/RAFI Mutual Funds Schwab Fundamental Idx Large Cap (I shares) SFLNX (40.07) n.a. n.a. 4/2/07 n.a. n.a. Schwab Fundamental Idx Small Mid (I shares) SFSNX (38.83) n.a. n.a. 4/2/07 n.a. n.a. Schwab Fundamental Int'l Idx Large Cap (I shares SFNNX (43.05) n.a. n.a. 4/2/07 n.a. n.a. Schwab Fund Idx Emerging Markets (I shares) SFENX n.a. n.a. n.a. 1/31/08 n.a. n.a. Schwab Subtotal 1,440 Grand Total 7,272 Market Index Benchmarks 13 Standard and Poor's 500 Index n.a. n.a (37.00) n.a. (5.63) 0.68 Russell 1000 n.a. n.a (37.60) n.a. (5.36) 0.54 Standard and Poor's Midcap 400 n.a. n.a (36.23) n.a. (1.83) 1.07 Standard and Poor's Smallcap 600 n.a. n.a (31.07) (0.30) n.a. (4.79) 0.16 MSCI EAFE (Europe, Australia, Ear East US$) n.a. n.a (45.09) n.a. (8.66) 1.51