Assessing the Risks of a Yield-Tilted Equity Portfolio

Transcription

1 Engineered Portfolio Solutions RESEARCH BRIEF Summer 2011 Update 2014: This Parametric study from 2011 is intended to illustrate common risks and characteristics associated with dividendtilted equity portfolios, and the relationship between portfolio tracking error, systematic exposure, and the degree of factor tilt implemented. This is presented for illustrative purposes. Market conditions presented may have changed since the time of original publication. Rey Santodomingo, CFA Director of Investment Strategy - Tax Managed Equities Parametric Portfolio Associates LLC Gordon Wotherspoon Director - Advisor Channel Portfolio Management Parametric Portfolio Associates LLC Jackson Wang, FRM Vice President MSCI Assessing the Risks of a Yield-Tilted Equity Portfolio Recently, investors have shown an increased preference towards income versus capital appreciation. One option to increase income is to favor dividend-paying stocks over nondividend paying stocks within an index-based portfolio strategy known as yield tilting. For example, a yield-tilted portfolio can be constructed to deliver 1.5x or 2.0x the yield of the index. In other words, when the index yield is 2.0%, the 1.5x tilted portfolio seeks to deliver a yield of 3.0%. While this additional yield is attractive, it comes at a related cost. The process of tilting a portfolio toward dividend paying stocks introduces systematic risks relative to the target index. In this paper we explore the risk characteristics of the yield-tilted portfolio. Specifically, we examine the historical tracking error of 1.5x and 2.0x yield-tilted portfolios versus a market cap-weighted index. In addition, we decompose the tracking error to uncover the persistent and dynamic biases resulting from this strategy. Finally, we examine the current risk characteristics of yield-tilted portfolios managed by Parametric. Parametric 1918 Eighth Avenue Suite 3100 Seattle, WA T F Parametric Portfolio Associates LLC. For Informational Purposes Only. Not an Offer to Buy or Sell Securities.

2 OBSERVATIONS Through a monthly back test spanning 12 years, our results show that the 1.5x and 2.0x yield-tilted portfolios exhibit a realized tracking error of % and % respectively, when compared against the S&P 500 Index. The tracking error varies over time and is higher during periods of high cross-sectional volatility. 1 Related to the tracking error, we find that the magnitude of the systematic biases varies over time as well. Some systematic biases, such as exposure to the financial sector, changed quite dramatically over the test period. During most of the period, the financial sector bias was positive. However, more recently, the bias has become negative. By contrast, the Barra Size factor bias of the dividend-tilted portfolios remained slightly negative and relatively consistent during the same period. THE RESULTS We used a 12-year history of the S&P 500 as our base index. To tilt the index towards yield, we used an optimizer and the Barra USE3L Risk Model to construct portfolios with a constant positive bias, or overweight, to the Barra Yield factor, while simultaneously working to minimize the tracking error. The portfolios were rebalanced monthly and the constant bias to this factor on average resulted in portfolios with yields of 1.5x and 2.0x the yield of the S&P In Exhibit 1 we show the rolling 36-month annualized realized tracking error (standard deviation of excess returns) of the portfolios compared to the S&P 500. As expected, the more extreme the tilt, the higher the realized tracking error. The tracking error for the 1.5x yield portfolio is on average 1.60%, while the 2.0x portfolio shows an average closer to 3.7%. Tracking error increases substantially as we increase the yield tilt from 1.5x to 2.0x because the number of eligible securities is reduced substantially. Exhibit 1: Realized Tracking Error of the Yield-Tilted Portfolio Compared to the S&P Month Annualized Rolling Realized TE 1 Cross-sectional volatility is a measure of security return dispersion. When securities returns are similar, cross-sectional volatility is low. When security returns differ, cross-sectional volatility is high. 2 To facilitate our research, the optimizer was configured to seek a constant bias to the Barra Yield factor. The two test cases studied were a 0.5 and a 1.0 constant bias to this factor. The units of the exposures are standard deviations. The Barra Dividend Yield factor exposure is a standardized score of a combination of the firm s last four quarterly dividend payments and future dividend announcements. The actual yield multiple for each portfolio varied over time around 1.5x and 2.0x,and the actual average multiple for each case over the test period was 1.55x and 2.10x respectively. 7% 1.5X Yield 2.0X Yield 6% 5% 4% 3% 2% 1% 0% Feb-2002 May-2002 Aug-2002 Nov-2002 Feb-2003 May-2003 Aug-2003 Nov-2003 Feb-2004 May-2004 Aug-2004 Nov-2004 Feb-2005 May-2005 Aug-2005 Nov-2005 Feb-2006 May-2006 Aug-2006 Nov-2006 Feb-2007 May-2007 Aug-2007 Nov-2007 Feb-2008 May-2008 Aug-2008 Nov-2008 Feb-2009 May-2009 Aug-2009 Nov-2009 Feb-2010 May-2010 Aug-2010 Nov-2010 Feb Parametric Portfolio Associates LLC. For Informational Purposes Only. Not an Offer to Buy or Sell Securities. 2

3 The average number of securities in the 1.5x yield portfolio was 255, compared to 124 securities in the 2.0x yield portfolio. We also note that the tracking error varies during different market regimes. From the dataset, we see three distinct time periods: a period of high tracking error from , a period of low tracking error from , and a return to high tracking error from The periods of high and low tracking error coincide with periods of high and low cross-sectional volatility in the market. During periods of low cross sectional volatility, the tracking error of the 2.0x tilt portfolio is about double that of the 1.5x tilt portfolio. However, during periods of high cross-sectional volatility, the gap increases to as high as 2.5 times that of the 1.5x tilt portfolio. Exhibit 2: Rolling 12-Month Excess Total Returns of 1.5X and 2.0X Yield Strategies Relative to the S&P % 15.00% 1.5X Yield 2.0X Yield 10.00% 5.00% 0.00% -5.00% % % % Feb-2000 May-2000 Aug-2000 Nov-2000 Feb-2001 May-2001 Aug-2001 Nov-2001 Feb-2002 May-2002 Aug-2002 Nov-2002 Feb-2003 May-2003 Aug-2003 Nov-2003 Feb-2004 May-2004 Aug-2004 Nov-2004 Feb-2005 May-2005 Aug-2005 Nov-2005 Feb-2006 May-2006 Aug-2006 Nov-2006 Feb-2007 May-2007 Aug-2007 Nov-2007 Feb-2008 May-2008 Aug-2008 Nov-2008 Feb-2009 May-2009 Aug-2009 Nov-2009 Feb-2010 May-2010 Aug-2010 Nov-2010 Feb-2011 Hypothetical returns are presented net of 30 bps annual fee. From a performance perspective Exhibit 2 shows that the strategy did well during the bear market. It also did well from The strategy underperformed during the global financial crisis of 2008 and outperformed during the recovery. To gain more insight into the strategy s excess return patterns, we examine the historical risk factor exposures of the portfolio Parametric Portfolio Associates LLC. For Informational Purposes Only. Not an Offer to Buy or Sell Securities. 3

4 SYSTEMATIC BIASES WITHIN THE YIELD-TITLED PORTFOLIO The yield tilted portfolio is constructed using an optimizer, which is configured to minimize the predicted tracking error while maintaining a constant positive bias to the Barra Yield factor. Selection of higher yielding securities results in systematic biases to certain risk factors. As of June 17, 2011, Parametric s composite of S&P 500 yield-tilted portfolios exhibits a negative bias to: The size, growth and volatility fundamental risk factors 3 The health, financial, transportation and technology sectors and a positive bias to: The value and earnings yield fundamental factors The utilities, basic materials, industrials and telecommunications sectors Our research shows that the systematic biases to these factors are dynamic. More recently, dividend portfolios exhibit a negative bias to financials, whereas during the study period the portfolio was typically overweight financials to varying degrees: Exhibit 3: Historical Relative Financial Sector Exposure Historical Relative Financial Sector Exposure X Yield 2.0X Yield Feb-1999 May-1999 Aug-1999 Nov-1999 Feb-2000 May-2000 Aug-2000 Nov-2000 Feb-2001 May-2001 Aug-2001 Nov-2001 Feb-2002 May-2002 Aug-2002 Nov-2002 Feb-2003 May-2003 Aug-2003 Nov-2003 Feb-2004 May-2004 Aug-2004 Nov-2004 Feb-2005 May-2005 Aug-2005 Nov-2005 Feb-2006 May-2006 Aug-2006 Nov-2006 Feb-2007 May-2007 Aug-2007 Nov-2007 Feb-2008 May-2008 Aug-2008 Nov-2008 Feb-2009 May-2009 Aug-2009 Nov-2009 Feb-2010 May-2010 Aug-2010 Nov-2010 Feb We examined the portfolios using the Barra USE3L risk model. The risk model is made up of industry and fundamental risk factors such as growth, value, size, volatility and momentum. For more information see the MSCI website at msci.com From , the 1.5x yield-tilted portfolio showed a 4% overweight to financials, while the 2.0x yield portfolio showed an overweight as high as 10%. Starting in November 2008, the portfolios began to exhibit a negative bias to financial stocks. The negative bias to financials coincides with the global financial crisis when many financial firms cut their dividends. In general, the larger the bias is to any given sector, the larger the expected tracking error will be. While the magnitude of the portfolio bias to financials varied over time, we find that the company size bias was relatively stable: 2014 Parametric Portfolio Associates LLC. For Informational Purposes Only. Not an Offer to Buy or Sell Securities. 4

5 Figure 4: Historical Size Factor Exposure Historical Size Factor Exposure Large Cap Bias X Yield 2.0X Yield 0.20 Neutral Small Cap Bias Feb-1999 May-1999 Aug-1999 Nov-1999 Feb-2000 May-2000 Aug-2000 Nov-2000 Feb-2001 May-2001 Aug-2001 Nov-2001 Feb-2002 May-2002 Aug-2002 Nov-2002 Feb-2003 May-2003 Aug-2003 Nov-2003 Feb-2004 May-2004 Aug-2004 Nov-2004 Feb-2005 May-2005 Aug-2005 Nov-2005 Feb-2006 May-2006 Aug-2006 Nov-2006 Feb-2007 May-2007 Aug-2007 Nov-2007 Feb-2008 May-2008 Aug-2008 Nov-2008 Feb-2009 May-2009 Aug-2009 Nov-2009 Feb-2010 May-2010 Aug-2010 Nov-2010 Feb-2011 It is interesting to note that there are more high dividend-yield small companies than large companies: Figure 5: S&P 500 Small Company Yields 120 S&P 500 Small Company Yields 100 # of Companies >8 3 This data is for illustrative purposes only. Indexes are unmanaged, cannot be invested in directly, and are not subject to fees or expenses. Each index will have different historical returns. This illustration is not intended to reflect the result of any investor. Any historical results presented herein should not and cannot be viewed as an indicator of future performance. 4 week returns are observed daily. Yield (%) 2014 Parametric Portfolio Associates LLC. For Informational Purposes Only. Not an Offer to Buy or Sell Securities. 5

6 Figure 6: S&P 500 Large Company Yields 120 S&P 500 Small Company Yields 100 # of Companies >8 Yield (%) Exhibit 5 and 6 show the frequency of dividend paying large and small companies in the S&P 500. Currently the index yield is 2%, so a 1.5x yield-tilted strategy would seek securities with yield greater than 3%. In the S&P 500 there are 71 small companies and only 18 large companies fitting this criterion. The larger high yield opportunity set in smaller companies combined with the objectives of higher yield and broad diversification result in portfolio biased slightly toward smaller companies. IMPLEMENTING A YIELD-TILTED PORTFOLIO Parametric currently manages portfolios with a 1.5x yield tilt to the S&P 500. To construct these dividend-tilted portfolios, we use an optimizer and constrain the portfolio indicated dividend yield to 1.5x the benchmark value. We also constrain the stock-level, industry-level, sector-level and Barra risk factor biases in the portfolios. Exhibits 7 and 8 show the current factor and sector biases of a representative 1.5x dividend yield-tilted portfolio: Figure 7: Current Barra Factor Bias Current Barra Factor Bias 4 For illustrative purposes only. Indexes are unmanaged, cannot be invested in directly, and are not subject to fees or expenses. NONESTU CURRSEN LEVERAGE SIZENONL MOMENTUM VOLTILTY VALUE EARNYLD EARNVAR SIZE TRADEACT GROWTH YIELD Source: Parametric Bias 2014 Parametric Portfolio Associates LLC. For Informational Purposes Only. Not an Offer to Buy or Sell Securities. 6

7 Figure 8: Current Barra Factor Bias Current Barra Sector Bias Telecommunications Utility Basic Materials Industrials Commercial Services Energy Consumer Services Consumer Cyclicals Consumer Noncyclicals Technology Health Transport Financial Bias -5% -4% -3% -2% -1% 0% 1% 2% 3% 4% 5% Source: Parametric % Overweight/Underweight S&P Sector Weight Given these biases, the current predicted tracking error of this portfolio is ~1.5%. SUMMARY Using an optimizer and a risk model we constructed dividend yield-tilted portfolios exhibiting a 1.5x and 2.0x yield compared to the S&P 500. Backtesting over this 12-year test period shows that these portfolios delivered excess yield over the benchmark and that they exhibit certain dynamic systematic biases, resulting in higher tracking error. The tracking error for these strategies ranged from % and % for the 1.5x and 2.0x yield-tilted portfolios. The tracking error varies depending on the degree of the yield tilt and the cross-sectional volatility of returns. Over the test period, we see that the the magnitude of systematic biases varied over time. For the 2.0x yield case, the financial sector bias ranged from +10% to neutral in more recent years. The portfolios also exhibited a small company bias which was relatively stable during the test period. Tilting towards dividends may lead to excess returns, but certainly creates active risk exposures Parametric Portfolio Associates LLC. For Informational Purposes Only. Not an Offer to Buy or Sell Securities. 7

8 Disclosures Parametric Portfolio Associates LLC ( Parametric ), headquartered in Seattle, Washington, is registered as an investment adviser under the United States Securities and Exchange Commission Investment Advisers Act of This material is intended for use with investment professionals. This information is intended solely to report on investment strategies and opportunities identified by Parametric. Opinions and estimates offered constitute our judgment and are subject to change without notice, as are statements of financial market trends, which are based on current market conditions. We believe the information provided here is reliable, but do not warrant its accuracy or completeness. This material is not intended as an offer or solicitation for the purchase or sale of any financial instrument. Past performance is not indicative of future results. The views and strategies described may not be suitable for all investors. Investing entails risks and there can be no assurance that Parametric will achieve profits or avoid incurring losses. Parametric does not provide legal, tax and/or accounting advice or services. Clients should consult with their own tax or legal advisor prior to entering into any transaction or strategy described herein. Charts, graphs and other visual presentations and text information were derived from internal, proprietary, and/ or service vendor technology sources and/or may have been extracted from other firm data bases. As a result, the tabulation of certain reports may not precisely match other published data. Data may have originated from various sources including, but not limited to, Bloomberg, MSCI/Barra, FactSet, and/or other systems and programs. Parametric makes no representation or endorsement concerning the accuracy or propriety of information received from any other third party. This material contains hypothetical, back-tested and/or model performance data, which may not be relied upon for investment decisions. Hypothetical, back-tested and/or model performance results have many inherent limitations, some of which are described below. Hypothetical returns are unaudited, are calculated in U.S. dollars using the internal rate of return, reflect the reinvestment of dividends, income and other distributions, include management fees, but exclude transaction costs, and do not take individual investor taxes into consideration. The deduction of such fees would reduce the results shown. No representation is being made that any client account will or is likely to achieve profits or losses similar to those shown. In fact, there are frequently sharp differences between hypothetical performance results and the actual results subsequently achieved by any particular trading program. One of the limitations of hypothetical performance results is that they are generally prepared with the benefit of hindsight. In addition, simulated trading does not involve financial risk, and no simulated trading record can completely account for the impact of financial risk in actual trading. For example, the ability to withstand losses or to adhere to a particular trading program in spite of trading losses are material points which can also adversely affect actual trading results. There are numerous other factors related to the markets in general or to the implementation of any specific trading program which cannot be fully accounted for in the preparation of hypothetical performance results and all of which can adversely affect actual trading results. Because there are no actual trading results to compare to the hypothetical, back-tested and/ or model performance results, clients should be particularly wary of placing undue reliance on these hypothetical results. Perspectives, opinions and testing data may change without notice. Detailed back-tested and/or model portfolio data is available upon request. No security, discipline or process is profitable all of the time. There is always the possibility of loss of investment. Benchmark/index information provided is for illustrative purposes only. Indexes are unmanaged and cannot be invested in directly. Deviations from the benchmarks provided herein may include, but are not limited to, factors such as: the purchase of higher risk securities, over/under-weighting specific sectors and countries, limitations in market capitalization, company revenue sources, and/or client restrictions. Parametric s proprietary investment process considers factors such as additional guidelines, restrictions, weightings, allocations, market conditions and other investment characteristics. Thus returns may at times materially differ from the stated benchmark and/or other disciplines provided for comparison. The S&P 500 Index represents the top 500 publicly traded companies in the U.S. Standard & Poor s and S&P are registered trademarks of Standard & Poor s Financial Services LLC ( S&P ), a subsidiary of The McGraw-Hill Companies, Inc. This strategy is not sponsored or endorsed by S&P, and S&P makes no representation regarding the content of this material. Please refer to the specific service provider s website for complete details on all indices. Parametric is located at th Avenue, Suite 3100, Seattle, WA For more information regarding Parametric and its investment strategies, or to request a copy of Parametric s Form ADV, please contact us at or visit our website, Parametric Portfolio Associates LLC. For Informational Purposes Only. Not an Offer to Buy or Sell Securities. 8