Quantitative Investment: From indexing to factor investing For institutional use only. Not for distribution to retail investors. 1
What s the prudent portfolio mix? It depends Objective Investment approach Total return Obligationdriven Principal protection Absolute return Market-cap-weighted index vehicles + Static tilts + Traditional active vehicles ƒ + Alternative investments Passive Active Other Global equity index Global equity index (tilt) Global fixed income index Global fixed income index (tilt) Global equity active Global equity active (tilt) Global fixed income active Global fixed income active (tilt) Cash Alternatives Notes: The allocations in each pie chart are for purposes of illustration only and are not intended as specific recommendations. Any actual portfolio recommendations would be determined using investor-specific criteria. For more on this topic, see the Vanguard research paper A Framework for Institutional Portfolio Construction, 2016. Source: Vanguard. For institutional use only. Not for distribution to retail investors. 2
A framework for portfolio construction Broad cap-weighted indexes Static tilts Security selection + + + Timing = Portfolio Strategic asset allocation Potential alpha sources Note: For more on this topic, see the Vanguard research paper A Framework for Institutional Portfolio Construction, 2016. Source: Vanguard. For institutional use only. Not for distribution to retail investors. 3
Quantitative investing: Indexing For institutional use only. Not for distribution to retail investors. 4
Number of funds Number of funds The key to indexing: Low costs Investing is a zero-sum game before costs Investing is a negative-sum game after costs Market benchmark Market benchmark Costs Distribution of fund returns Distribution of fund returns Fund returns before costs Fund returns after costs Funds above benchmarks The holdings of all investors aggregate to form a market. Outperformance by one investor necessarily means underperformance by another. After costs, a majority of investors will underperform the market benchmark. The key to long-term, consistent success is low-cost investing. For institutional use only. Not for distribution to retail investors. 5
Relative difference in median ten-year annualized alpha Is there an ex-ante metric that can improve the odds? Cost is a powerful indicator of future alpha 1.5% 1.0% 0.5% 1.18% 0.66% 0.58% 0.26% Not statistically significant 0.0% -0.5% Expense ratio Turnover Tracking error Fund concentration -0.15% -0.15% -0.20% -0.21% Fund size Past alpha Number of holdings Sharpe ratio Median of lowest-quartile metric funds outperformed Median of highest-quartile metric funds outperformed Notes: The bars represent the difference in the median annual alpha from January 1, 2006, through December 31, 2016, of mutual funds in the lowest and highest quartiles, based on the above metric measurements from January 1, 2004, through December 31, 2006. Fund concentration is measured as the percentage of assets in the top ten holdings. The expense ratio and turnover were the only variables significant at the 99% confidence level. For more information, see the Vanguard research note Shopping for Alpha: You Get What You Don't Pay For, 2015. Sources: Vanguard calculations, based on data from Morningstar, Inc., and Fama-French factor data. For institutional use only. Not for distribution to retail investors. 6
Tracking the indexes Steps to keep our portfolios on track with benchmarks Proprietary system designed to neutralize the major risk factors Limit orders help minimize transaction costs Futures contracts help provide liquidity Portfolio holdings rebalanced daily Cash positions equitized daily For institutional use only. Not for distribution to retail investors. 7
Comprehensive risk assessment Vanguard exercises ongoing rigorous risk control Individual stock level Daily evaluation of security weights relative to the index Factor level Daily monitoring of sector weights, market capitalization, volatility, and style relative to the index Portfolio level Daily analysis of portfolio characteristics and ongoing rebalancing to match benchmark characteristics Sample imbalance report for individual holdings Ticker Shares held Recent price Market value Portfolio wt. Index wt. % Imbalance Dollar imbalance Share imbalance SPG 2,425,936 36.780 $89,225,926 0.12977 0.12808 0.00169 1,162,840 31,700 MCO 1,838,772 23.410 $43,045,653 0.06260 0.06167 0.00093 639,270 27,400 RHI 1,468,728 18.900 $27,758,959 0.04037 0.03956 0.00081 558,419 29,600 GWW 612,161 73.600 $45,055,050 0.06553 0.06476 0.00077 527,711 7,200 ALTR 2,815,015 18.400 $51,796,276 0.07533 0.07463 0.00070 484,199 26,400 CI 2,603,008 18.640 $48,520,069 0.07057 0.06986 0.00070 482,218 25,900 DRI 1,313,981 37.010 $48,630,437 0.07073 0.07011 0.00062 423,549 11,500 For institutional use only. Not for distribution to retail investors. 8
Quantitative investing: factors For institutional use only. Not for distribution to retail investors. 9
Broad capweighted index Everything else active Active investing is any investment strategy that is not broad market-cap-weighted. For institutional use only. Not for distribution to retail investors. 10
ƒ Factors are the DNA of an investment portfolio. They are the underlying characteristics that explain and influence an investment s risk. Equity factor-based investing is a systematic, diversified, and disciplined form of active management that intentionally targets consistent exposure to one or more well-known factors that have reasonable odds of helping an investor meet an objective. For institutional use only. Not for distribution to retail investors. 11
Equity exposure should be evaluated through different lenses L M S V B G ƒ Style Location Sectors Factor For institutional use only. Not for distribution to retail investors. 12
Sources of variation Factors are raising the bar for traditional active management Advances in research and technology and low-cost factor vehicles are raising the bar for traditional active management 100% Manager skill + noise Alpha + noise Pure alpha + noise Pure alpha + noise Size and value factors Style factors (f 1 + f k) Beta Market factor Market factor 0% Attribution models are evolving to distinguish factors from true alpha Note: The scenario shown is hypothetical and is for purposes of illustration only. Source: Vanguard. For institutional use only. Not for distribution to retail investors. 13
Equity factors that supposedly generate excess return have proliferated Drivers Outcomes New factors documented in academic literature Advances in technology 18 Academic acclaim 1 ƒ 5 ƒ ƒ ƒ ƒ ƒ Financial incentive 1980 1991 1991 2003 2003 2012 Average per year Source: For the outcomes and time periods only, Campbell R. Harvey, Yan Liu, and Heqing Zhu, 2016, And the Cross-Section of Expected Returns, The Review of Financial Studies 29 (1), 5-68. For institutional use only. Not for distribution to retail investors. 14
Which, if any, may benefit investors in the future? 300+ documented characteristics today Enduring logical rationale Risk-based and/or behavioral reason(s) Extensive empirical evidence Out-of-sample tests, breadth of academic studies, etc. Investability Real-world implementation costs ƒ For institutional use only. Not for distribution to retail investors. 15
Academic results don t equal real-world portfolio results Academic papers often... Test long/short, not long-only strategies Ignore management and oversight expenses Disregard transactionrelated costs For institutional use only. Not for distribution to retail investors. 16
Calendar year excess returns (versus the broad market index) Relative performance has been inconsistent over the short term 25% 20% 15% 10% 5% 0% -5% -10% -15% -20% 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Value Quality Volatility Momentum Size Liquidity Notes: Data cover January 1, 2002, through December 31, 2015. Excess returns are calculated relative to the MSCI World Total Return Index (USD). This figure includes back-tested index performance. For information regarding index inception and index-live dates, please see the table in the Appendix. Sources: Vanguard calculations, using data from Thomson Reuters Datastream, MSCI, Bloomberg, and FTSE. For institutional use only. Not for distribution to retail investors. 17
Average pairwise correlation Diversifying factor exposure globally can yield benefits 1.0 0.9 Correlation of equity factor tilt excess returns versus local market indexes (in USD) 0.8 0.7 0.6 0.61 0.5 0.4 0.39 0.3 0.2 0.19 0.19 0.19 0.1 0.0 Value Size Quality Momentum Volatility Notes: All calculations cover monthly excess returns from December 31, 2001, through May 31, 2016, except for the volatility factor, which covers November 30, 2002, through May 31, 2016. Country factor indexes included in the analysis are for the United States, United Kingdom, and Japan. Sources: Vanguard calculations, based on data from Thomson Reuters Datastream and MSCI. For institutional use only. Not for distribution to retail investors. 18
Single or multifactor? Low correlation of excess returns can diversify active risk Correlation of monthly excess returns, 2002 2016 Value Quality Volatility Momentum Size Liquidity Value Quality 0.59 Volatility 0.38 0.49 Momentum 0.11 0.37 0.37 Size 0.43 0.30 0.15 0.10 Liquidity 0.57 0.44 0.13 0.08 0.63 High: > 0.7 Medium: 0.3 0.7 Low: < 0.3 Notes: Data cover December 31, 2001, through September 30, 2016. Excess returns are calculated relative to the MSCI World Total Return Index (USD). This figure includes back-tested index performance. For information regarding index inception and index-live dates, please see the table in the Appendix. Sources: Vanguard calculations, using data from Thomson Reuters Datastream, MSCI, Bloomberg. and FTSE. For institutional use only. Not for distribution to retail investors. 19
What are your goals? Your path forward Enhance return Generate excess returns using a static tilt to a particular factor Reduce risk Lessen total volatility with a minimal reduction in return Substitution Represent a high-cost active manager s returns with a broad market index vehicle paired with one or more low-cost, factor-tilted equity products Completion Adjust a factor over/underweight within the portfolio by allotting assets to a factor-tilted strategy For institutional use only. Not for distribution to retail investors. 20
Annualized excess returns Role option 1: Static tilt Enhance return Numerous equity factor tilts have outperformed historically 5% 4.5% 4% 3% 2.6% 2.7% 2.7% 2% 1.3% 1.5% 1% 0% Quality Volatility Momentum Value Liquidity Size Notes: All results are as of September 30, 2016. Excess returns are calculated relative to the MSCI World Total Return Index (USD). MSCI World Momentum Index (USD) history begins May 31, 1973; MSCI World Value Index (USD) is from December 31, 1974, to November 30, 1997, and MSCI World Enhanced Value Index thereafter; MSCI World Quality Index (USD) begins November 30, 1975; MSCI World Minimum Volatility Index (USD) begins May 31, 1988; MSCI World Small Cap Index (USD) begins December 31, 2000; and FTSE Developed Illiquidity Factor Index (USD) begins September 30, 2001. This figure includes back-tested index performance. For information regarding index inception and index-live dates, please see the table in the Appendix. Sources: Vanguard calculations, using data from Thomson Reuters Datastream, MSCI, Bloomberg, and FTSE. For institutional use only. Not for distribution to retail investors. 21
Gross monthly returns Role option 3: High-cost traditional active manager substitution Step 2: Try to represent the manager s performance while reducing costs 20% 20 10% 0% 0-10% -20% Time High-cost traditional active equity fund Broad market index fund + value factor fund Note: The example shown is hypothetical, representing a period of several decades, and does not represent the exposures for a specific traditional active manager. Source: Vanguard. For institutional use only. Not for distribution to retail investors. 22
Conclusion A form of active management Greater transparency More control Lower cost Key considerations Persistence rationale Portfolio construction and implementation trade-offs Cost Suitability depends on: Investment objectives Risk tolerance Due diligence capabilities For institutional use only. Not for distribution to retail investors. 23