optimizing value · 2019. 2. 26. · value and smart beta 5 • value is the textbook example of...
Post on 22-Sep-2020
1 Views
Preview:
TRANSCRIPT
Aperio Group Three Harbor Drive, Suite 315 Sausalito, CA 94965 (415) 339-4300 www.aperiogroup.com Copyright © 2015 Aperio Group LLC
Aperio v. [Latin] to make clear, to reveal the truth
Optimizing Value
References to returns, risks, performance, tracking error, and other such characteristics describing portfolios in this paper are based on hypothetical analysis techniques (also known as back-testing) and are not actual portfolios. Since returns included herein are hypothetical and based on back-testing, it is important to note that they are for illustrative purposes only. Past performance, whether illustrative or actual, is not a guarantee of future performance. Please refer to the important disclosures within and at the end of this paper.
Value Investing
2
• The orderly pursuit of underpriced securities • Based on valuation ratios like book-to-price (B/P)
and earnings-to-price (E/P) to determine whether a company is value or growth
• Popular with investors since Graham and Dodd (1934)
• Technology has taken us beyond fundamental
Behavior: Investors Love Glamour
3
Behavior: Investors Love Glamour
4
82
86
3
167
2
10
P/E P/E
Value and Smart Beta
5
• Value is the textbook example of Smart Beta
• The historical outperformance of value relative to the market is at the heart of the smart beta fad
• But there is more than one way to invest in value… • …and some ways are smarter than others
Value Strategy Implementation Choices
6
• Which value metric to use? • Book-to-price (B/P) is the academic standard • Earnings-to-price (E/P) is preferred by many
practitioners • What about a blend?
• Portfolio construction: How to select and weight stocks • Rank & chop methodology • Optimization and multi-factor models
B/P Versus E/P: Academic Perspective
7
• Fama and French examined various value measures in their seminal 1992 paper
• They found that over the period 1963–1990, B/P was the best performer and included it in their three-factor model
• We extend their study to include more data
Portfolio Construction: Rank & Chop
8
• Begin with a universe: US market
• Rank stocks by B/P or E/P
• Chop off the bottom of the list (bottom 70%)
• Capitalization-weight the remaining stocks (top 30%)
• Rebalance once a year
• No transaction costs
B/P Versus E/P: 1963–1990
9
Portfolio E/P B/P MarketGeometric Return 13.75% 14.93% 9.99%Standard Deviation 16.45% 16.53% 15.83%Sharpe Ratio 0.46 0.52 0.26
Portfolio E/P B/P MarketGeometric Return 13.75% 14.93% 9.99%Standard Deviation 16.45% 16.53% 15.83%Sharpe Ratio 0.46 0.52 0.26
Source: Ken French’s website. Past performance is not a guarantee of future returns. Returns are gross of transaction costs and fees. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
B/P Versus E/P: 1991–2013
10
Portfolio E/P B/P MarketGeometric Return 14.56% 12.39% 10.46%Standard Deviation 15.64% 16.29% 15.12%Sharpe Ratio 0.77 0.62 0.54
Source: Ken French’s website. Past performance is not a guarantee of future returns. Returns are gross of transaction costs and fees. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
B/P Versus E/P: 1951–2013
11
Source: Ken French’s website. Past performance is not a guarantee of future returns. Returns are gross of transaction costs and fees. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Portfolio E/P B/P MarketGeometric Return 15.80% 14.39% 11.01%Standard Deviation 15.72% 16.09% 14.93%Sharpe Ratio 0.73 0.64 0.48
Metric Matters: Conclusions
12
• E/P outperformed B/P with less risk over the study period (1951- 2013)
• B/P outperformed E/P over the period featured in the landmark study that established B/P as the academic standard
• Had Fama and French waited a few years to embark on their famous study, the academic literature might look different today
Academia à Real World
13
Source: Chris Madden. Reprinted with permission.
Toward a More Realistic Experiment
14
*In this presentation, S&P 500 refers to the S&P 500 Total Return Index, which, unlike the S&P 500 Index itself, is an investable universe.
• Construct strategies in a liquid universe (S&P 500*) and penalize turnover with a round-trip trading cost of 0.12%
• Test different signals: E/P, B/P, blend of E/P and B/P
• Test different construction methodologies • Optimization • Rank and chop: Top 30%
• Quarterly rebalancing (annual in Ken French’s data)
Top 30% Value Strategies: 1973–2013
15
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Top 30% Value Strategies: 1973–2013
16
Top 30% Value Strategies: 1973–2013
17
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Top 30% Value Strategies: 1973–2013
18
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Top 30% Value Strategies: 1973–2013
19
Top 30% Value Strategies: 1973–2013 (continued)
20
Strategy Top 30%: B/P
Top 30%: E/P
Top 30 %: Blend S&P 500
Annualized Return (%) 12.79 13.48 14.21 10.30
Standard Deviation 18.79 16.41 17.36 15.49
Tracking Error (%) 9.27 7.03 8.13 0.00
Sharpe Ratio 0.44 0.52 0.53 0.35
Information Ratio 0.30 0.43 0.47 -
Turnover (%) 114.5 110.0 119.4 -
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Top 30% Value Strategies: 1973–2013 (continued)
21
Strategy Top 30%: B/P
Top 30%: E/P
Top 30 %: Blend S&P 500
Annualized Return (%) 12.79 13.48 14.21 10.30
Standard Deviation 18.79 16.41 17.36 15.49
Tracking Error (%) 9.27 7.03 8.13 0.00
Sharpe Ratio 0.44 0.52 0.53 0.35
Information Ratio 0.30 0.43 0.47 -
Turnover (%) 114.5 110.0 119.4 -
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Top 30% Value Strategies: 1973–2013 (continued)
22
Strategy Top 30%: B/P
Top 30%: E/P
Top 30 %: Blend S&P 500
Annualized Return (%) 12.79 13.48 14.21 10.30
Standard Deviation 18.79 16.41 17.36 15.49
Tracking Error (%) 9.27 7.03 8.13 0.00
Sharpe Ratio 0.44 0.52 0.53 0.35
Information Ratio 0.30 0.43 0.47 -
Turnover (%) 114.5 110.0 119.4 -
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Top 30% Value Strategies: 1973–2013 (continued)
23
Strategy Top 30%: B/P
Top 30%: E/P
Top 30 %: Blend S&P 500
Annualized Return (%) 12.79 13.48 14.21 10.30
Standard Deviation 18.79 16.41 17.36 15.49
Tracking Error (%) 9.27 7.03 8.13 0.00
Sharpe Ratio 0.44 0.52 0.53 0.35
Information Ratio 0.30 0.43 0.47 -
Turnover (%) 114.5 110.0 119.4 -
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Top 30% Value Strategies: 1973–2013 (continued)
24
Strategy Top 30%: B/P
Top 30%: E/P
Top 30 %: Blend S&P 500
Annualized Return (%) 12.79 13.48 14.21 10.30
Standard Deviation 18.79 16.41 17.36 15.49
Tracking Error (%) 9.27 7.03 8.13 0.00
Sharpe Ratio 0.44 0.52 0.53 0.35
Information Ratio 0.30 0.43 0.47 -
Turnover (%) 114.5 110.0 119.4 -
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
High Turnover à Taxes à Lower Return
25
• In California top bracket tax payers are taxed at 52%* on short term capital gains
• Strategies with over 100% turnover can have very
large penalties
• If all the return is realized S/T capital gain, then a 14.2% annualized return can quickly turn into 6.8%!
* Assumes the highest federal bracket of 39.60% plus the 3.80% Medicare rate applied to investment income plus the itemized deduction phase out impact (Pease) of 1.19%. The phase out impact is calculated as 39.60% times the 3.00% of disallowed itemized deductions. For the combined state and federal rates, state taxes of 13.3% are assumed to be fully deductible for both conventional federal tax and Medicare tax. To calculate the combined rate, the state tax rate is reduced by the deduction at the federal ordinary rate plus the Medicare rate.
Optimized Portfolio Construction
26
• Using a risk model and an optimizer, construct a portfolio that: • Minimizes tracking error against the S&P 500 • Tilts the portfolio towards value stocks (E/P, B/P,
Blend) • Controls turnover
• Target approximately 3% tracking error
Optimized Value Strategies:1973–2013
27
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Top 30% versus Optimized: Rolling 10-Year Return
28
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Top 30% versus Optimized: Rolling 10-Year Return
29
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Top 30% versus Optimized: Rolling 10-Year Return
30
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Top 30% versus Optimized: Rolling 10-Year Return
31
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Top 30% Versus Optimized Portfolios
32
Strategy Top 30 %: Blend
Optimized: Blend
S&P 500
Annualized Return (%) 14.21 12.81 10.30
Standard Deviation 17.36 15.66 15.49
Tracking Error (%) 8.13 3.48 0.00
Sharpe Ratio 0.53 0.49 0.35
Information Ratio 0.47 0.66 0.00
Turnover (%) 119.4 30.0 -
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Top 30% Versus Optimized Portfolios (continued)
33
Strategy Top 30 %: Blend
Optimized: Blend
S&P 500
Annualized Return (%) 14.21 12.81 10.30
Standard Deviation 17.36 15.66 15.49
Tracking Error (%) 8.13 3.48 0.00
Sharpe Ratio 0.53 0.49 0.35
Information Ratio 0.47 0.66 0.00
Turnover (%) 119.4 30.0 -
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Top 30% Versus Optimized Portfolios (continued)
34
Strategy Top 30 %: Blend
Optimized: Blend
S&P 500
Annualized Return (%) 14.21 12.81 10.30
Standard Deviation 17.36 15.66 15.49
Tracking Error (%) 8.13 3.48 0.00
Sharpe Ratio 0.53 0.49 0.35
Information Ratio 0.47 0.66 0.00
Turnover (%) 119.4 30.0 -
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Top 30% Versus Optimized Portfolios (continued)
35
Strategy Top 30 %: Blend
Optimized: Blend
S&P 500
Annualized Return (%) 14.21 12.81 10.30
Standard Deviation 17.36 15.66 15.49
Tracking Error (%) 8.13 3.48 0.00
Sharpe Ratio 0.53 0.49 0.35
Information Ratio 0.47 0.66 0.00
Turnover (%) 119.4 30.0 -
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
The Tracking Error Roller Coaster of Top 30%
36
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Top 30% Versus Optimized Strategies
37
• Top 30% strategies had slightly higher risk-adjusted return
• But their tracking errors and turnover were much greater than their Optimized counterparts • Tracking error was more than double • Turnover was greater by a factor of four
Active Sector Exposures: 1973–2013
38
Average sector exposures in value portfolios based on blended signals relative to the S&P 500, January 1973–December 2013. Source: Aperio Group.
Sector-Constrained Optimized Strategies
39
Source: Aperio Group. Past performance is not a guarantee of future returns. Returns are gross of fees and we assume a round-trip trading cost of 0.12%. Performance is hypothetical, and is not based on an actual portfolio or account. Please refer to important disclosures at the end of this paper.
Summary: The Time Period Matters
40
• The relative performance of the academic standard (B/P) and the preferred practitioner metric (E/P) depends on when you look
• E/P outperformed in the most recent period we considered (1991-2013)
• And it was the overall winner (1951-2013)
Summary: The Metric Matters
41
• A value strategy based on a blend of the two metrics outperformed either single-metric strategy during the period 1973–2014 and in most 10-year sub-periods
• For both rank & chop and optimized strategies
Summary: Portfolio Construction Matters
42
• Rank & chop strategies had higher returns and higher risk than optimized strategies
• The two approaches had similar risk-adjusted returns
• However, optimized strategies resulted in materially lower turnover and tracking error than rank & chop strategies …
• … leading to less career risk and a substantial advantage for taxable investors
Thank You
43
References
44
Eugene F. Fama and Kenneth R. French. “The Cross-Section of Expected Stock Returns.” Journal of Finance, 47(2):427–465, 1992. Available here: http://faculty.som.yale.edu/zhiwuchen/Investments/Fama-92.pdf
Josef Lakonishok, Andrei Shleifer, and Robert W. Vishny. “Contrarian Investment, Extrapolation and Risk.” The Journal of Finance, 49(5):1541–1578, 1994. Available here: http://scholar.harvard.edu/files/shleifer/files/contrarianinvestment.pdf
Ran Leshem, Lisa R. Goldberg, and Alan Cummings. “Optimizing Value.” Forthcoming in Journal of Portfolio Management. Aperio Group. 2015. Available here: https://www.aperiogroup.com/resource/429/node/download
Ken French’s Data Library: http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html
Disclosures
45
The information contained within this presentation was carefully compiled from sources Aperio believes to be reliable, but we cannot guarantee accuracy. We provide this information with the understanding that we are not engaged in rendering legal, accounting, or tax services. In particular, none of the examples should be considered advice tailored to the needs of any specific investor. We recommend that all investors seek out the services of competent professionals in any of the aforementioned areas.
With respect to the description of any investment strategies, simulations, or investment recommendations, we cannot provide any assurances that they will perform as expected and as described in our materials. Past performance is not indicative of future results. Every investment program has the potential for loss as well as gain.
The performance reflected in the tables and charts in this report are hypothetical, shown for illustrative purposes only, and not based on actual investments. Furthermore, they do not reflect the deduction of any management fees or transaction costs, which would lower performance returns. The use of hypothetical performance has significant limitations, some of which are described below. Back-testing involves simulation of a quantitative investment model by applying all rules, thresholds, and strategies to a hypothetical portfolio during a specific market period and measuring the changes in value of the hypothetical portfolio based on the actual market prices of portfolio securities. Investors should be aware of the following: 1) Back-tested performance does not represent actual trading in an account and should not be interpreted as such, 2) back-tested performance does not reflect the impact that material economic and market factors might have had on the manager’s decision-making process if the manager were actually managing client’s assets, and 3) there is no indication that the back-tested performance would have been achieved by a manager had the program been activated during the periods presented above. For back-tested performance comparisons, the benchmark returns are simulated using historical constituents’ weights and total returns. The S&P 500 Total Return Index is an unmanaged group of equities representing the large-cap sector of the US domestic market. Index returns reflect reinvestment of dividends but do not reflect fees, brokerage commissions, or other expenses of investing.
Disclosures (continued)
46
Information Ratio (IR): Measures a portfolio manager's ability to consistently generate excess returns relative to a benchmark (average portfolio return relative to a benchmark divided by the volatility of relative return).
Sharpe Ratio: The average portfolio return in excess of the risk-free rate divided by the volatility of portfolio excess return. The Sharpe Ratio is a type of Information Ratio in which the benchmark is the risk-free rate. Standard Deviation: A statistical measure that is used to quantify the level of dispersion in a variable data set, such as portfolio return.
Tracking Error: A measure of how closely a portfolio tracks its benchmark. In technical language, tracking error is the standard deviation of the return difference between a portfolio and a benchmark.
top related