Evaluating S&P 500 Sector ETFs Using Risk-Adjusted Performance Measures

Transcription

1 Journal of Finance, Accounting and Management, 5(1), 48-62, Jan Evaluating S&P 500 Sector ETFs Using Risk-Adjusted Performance Measures Onur Arugaslan Associate Professor of Finance State Farm Professor of Financial Services Department of Finance and Commercial Law Haworth College of Business Western Michigan University Kalamazoo, MI USA Ajay Samant Dean and Professor of Finance Coggin College of Business, Building 42/2010 University of North Florida Jacksonville, FL USA Abstract This study examines the nature and performance of S&P 500 Sector Exchange Traded Funds (ETFs) with a view to providing empirical documentation which can be used as input in decision making by stock market investors. The performance evaluation techniques used in the study utilize state of the art statistical measures grounded in modern portfolio theory. Returns are adjusted for the degree of total risk and systematic risk inherent in each ETF, and the securities are then ranked on the basis of risk-adjusted performance. A relatively new evaluation metric, the Modigliani measure, is used for ranking these ETFs. The study also demonstrates how financial leverage can be used to lower the risk of a sector ETF and maintain a desired rate of return. The results of the study should be of interest to business academicians, investors, bankers, and investment fund managers. Keywords: risk-adjusted performance, exchange-traded funds

2 Journal of Finance, Accounting and Management, 5(1), 48-62, Jan Introduction This study examines the characteristics and performance of S&P 500 Sector Exchange Traded Funds (ETFs). An ETF is listed on a stock exchange and holds a portfolio of equities or bonds. One of the first ETFs created in 1993 was the Standard and Poor s Depositary Receipt (SPDR) which was listed under the ticker symbol SPY on the New York Stock Exchange, and commonly referred to as the Spider. Subsequently, in 1998 State Street Global Advisors created the Sector Spiders which track all the sectors of the S&P 500 index. Since then the market for ETFs has proliferated. It is estimated that more than 1000 ETFs are listed on US stock markets, with over $990 trillion in assets. Investors can trade ETFs in the same manner as any listed stock, which includes the ability to short sell or buy on margin. The popularity of ETFs can be attributed to many factors. Unlike individual stocks, they provide a means of diversifying firm-specific risk at a very low cost. Unlike mutual funds, they can be bought or sold instantaneously. Also, unlike mutual funds, they can be shorted or purchased on margin. Further, a variety of buy or sell orders can be placed on ETFs, including market orders, limit orders and stop loss orders. Finally, investors can use ETFs to hold a range of financial securities including domestic stocks, international stocks, bonds, currencies and commodities, including gold. ETFs are created by investment companies. The largest companies in the business are State Street Global Advisors, Bank of New York, Merrill Lynch, Morgan Stanley, Barclays, Blackrock and Vanguard.

3 Journal of Finance, Accounting and Management, 5(1), 48-62, Jan The sectors of the S&P 500 Index are the following: Consumer Discretionary, Consumer Staples, Energy, Financial Services, Healthcare, Industrial, Materials, Technology, Utilities and Telecom. ETFs are available for investment in each of these sectors. A convenient way for an investor to hold these sectors in a portfolio is to purchase a Sector Spider. This study evaluates the performance of these Sector Spiders with a view to providing empirical documentation to an investor who is interested in holding these securities, but is not sure of their risk and return. The rest of the paper is structured as follows. Literature Review section reviews the literature on ETFs and summarizes relevant studies in the area of modern portfolio theory. Characteristics of S&P 500 Sector ETFs section describes the number of holdings, dividend yield, price/book ratio, price/earnings ratio, and weighted average market capitalization. Performance of S&P 500 Sector ETFs section evaluates the performance of these ETFs on a risk-adjusted basis, using the S&P 500 Index as a benchmark for comparison purposes. Summary section summarizes the paper and Conclusion section concludes. Literature Review Treynor (1965), Sharpe (1966), and Jensen (1968) pioneered the evaluation of the performance of investment portfolios. They developed statistical techniques that are the most commonly used portfolio performance measures even today. Treynor (1965) suggested a way of evaluating the performance of a portfolio by adjusting the mean excess return for the degree of market risk and thus calculating the performance of the portfolio. Sharpe (1966) computed mean excess return and adjusted for the degree of total risk involved in the portfolio. Jensen (1968) devised a method of determining whether the deviation of portfolio returns from market returns was statistically

4 Journal of Finance, Accounting and Management, 5(1), 48-62, Jan significant, and, therefore, determining whether the excess return could be attributed to superior management, or purely to chance. The techniques used in these three pioneering studies were further refined by Kon and Jen (1979), Henrikkson and Merton (1981), and Chang and Lewellen (1984). Different techniques have been used in portfolio performance measurement over time. Recently, Modigliani and Modigliani (1997) did some pioneering work in the area of financial reward and risk. They proposed a new risk-adjusted performance measure (hereafter referred to as, M Squared), which is intuitively quite appealing to investors. The idea that underlies their methodology is to adjust the returns of a portfolio to the level of risk in an unmanaged stock market index and then measure the returns on the risk-matched portfolio. Exchange traded funds have become an increasingly popular research topic in the new millennium. In one of the earlier studies, Hasbrouck (2001) focuses on the intraday price formation in the U.S. equity markets. Among other things, he finds that the sector ETFs can closely replicate the S&P 500 ETF. He also reports that the technology sector is the most actively traded and contributes a modest amount to price discovery in the overall index whereas the other sector ETFs play only a minor role. In a separate vein, Huang and Guedj (2009) develop an equilibrium model to explore whether an ETF is a more efficient indexing vehicle than an open-ended mutual fund. They derive empirical predictions that ETFs are better suited for narrower and less liquid underlying indices, and for longer term investors.

5 Journal of Finance, Accounting and Management, 5(1), 48-62, Jan Gleason, Mathur, and Peterson (2004) use intraday data to examine herding behavior during periods of extreme market movements using nine sector ETFs traded on the American Stock Exchange. Similarly, Chen, Ho, Lai, and Morales-Camargo (2011) study the trading behavior of institutional investors in the ETF market from 1993 to 2007 and find that herding behavior is influenced by experience in the ETF market and by certain ETF characteristics ( liquidity, size, etc.) In contrast, Hamm (2011) studies the effect of ETFs on stock liquidity and reports that highly diversified ETFs benefit more from liquidity inflow than sector ETFs. To the knowledge of the authors, this is the first study of the characteristics and performance of S&P 500 Sector ETFs. The results of this study should be of interest to investors and mutual fund managers who are contemplating investing in the sector ETF market. Characteristics of S&P 500 Sector ETFs There are ten S&P 500 Sector ETFs as of January 31, The characteristics of these ETFs are summarized in Table 1. The ETF with the highest number of holdings is Consumer Discretionary with 84 stocks. Utilities Select ranks first in terms of dividend yield (3.95%), whereas Consumer Discretionary displays the best price-to-book ratio (3.43). Regarding the priceto-earnings (P/E) ratio, Telecom Services is at the top of the list with a P/E ratio of Finally, the ETF with the largest weighted average market capitalization is Technology Select with a capitalization of $165.4 billion. The table also provides the same characteristics of the S&P 500 ETF.

6 Journal of Finance, Accounting and Management, 5(1), 48-62, Jan Table 1 Characteristics of S&P 500 Sector ETFs Sector ETF Ticker No. of Dividend Price/Book P/E Ratio Weighted Avg holdings Yield Ratio Mkt Cap ($b) SPDR Consumer Discretionary XLY SPDR Consumer Staples XLP SPDR Energy Select XLE SPDR Financial Services Select XLF SPDR Healthcare XLV SPDR Industrial Select XLI SPDR Materials XLB SPDR Technology Select XLK SPDR Utilities Select XLU SPDR Telecom Services XTL SPDR S&P 500 SPY Performance of S&P 500 Sector ETFs Data and Methodology Monthly return data for the three-year period January December 2012 are obtained from CRSP. CRSP has full return data for 9 Sector ETFs. Therefore, the final sample in this study for the performance analysis consists of 9 ETFs. The return on U.S. 4-week Treasury Bills is used as the proxy for the risk-free rate. The S&P 500 Index is utilized as the market benchmark. Monthly returns are averaged over the three-year period to obtain the Mean return. Risk-free rate of return is subtracted from the mean return to compute the Mean excess return. Mean excess return of each ETF is divided by its standard deviation to compute the Sharpe measure: = Si Ri - R f i where Ri = mean return on ETF i, Rf = mean risk-free rate of return,

7 Journal of Finance, Accounting and Management, 5(1), 48-62, Jan = Ti - Ri σi = standard deviation of returns for ETF i. Mean excess return of each ETF is divided by its beta to obtain the Treynor measure: i R f where βi is estimated from the market model: R it = + R i i mt +e it where Rmt = market return during period t, eit = error term. Sharpe measure is multiplied by the market standard deviation and then the risk-free rate added to calculate the M Squared measure: 2 M i = Ri - R f i m+ R f Benchmark standard deviation is divided by the ETF standard deviation to obtain the Leverage Factor: m Li = i Leverage Factor reports a comparison of the total risk in the ETF with the total risk in the market portfolio. For example, a Leverage Factor less than one implies that the risk of the ETF is greater than the risk of the market index, and that the investor should consider unlevering the ETF by selling off part of the holding in the ETF and investing the proceeds in a risk-free security, such as a Treasury Bill. On the other hand, a Leverage Factor greater than one implies that the standard deviation of the ETF is less than the standard deviation of the market index, and that the investor

8 Journal of Finance, Accounting and Management, 5(1), 48-62, Jan should consider levering the ETF by borrowing money (if possible, at the risk-free rate of return) and investing in that particular ETF. The significance of the Leverage Factor is that we use it to form a portfolio that has the same total risk (standard deviation) as the market portfolio using the ETF and the risk-free asset. By forming this portfolio, we limit our risk exposure to the market risk without sacrificing too much on the return. These Leverage Factors are used to compute the adjusted returns in Table 4. First, the Mean Monthly Adjusted Return is computed using the Leverage Factor: MARi Li Ri ( 1 Li) R f Finally, Mean Annual Adjusted Return is calculated by compounding over 12 months: 12 AARi ( 1 MARi) 1 Results The 9 ETFs with full monthly return data are identified in Table 2 along with their risk, return, and performance statistics. The ETFs are ranked in alphabetical order. The ETF with the highest mean return is Consumer Discretionary with an average monthly return of percent. In comparison, the monthly mean return of the benchmark S&P 500 Index is percent. The ETF with the highest total risk (measured by the standard deviation of returns) is Energy Select with a monthly standard deviation of percent. In comparison, the standard deviation of the benchmark S&P 500 Index is percent. Further, Table 2 reports the numerical values of the Sharpe measure, which is used to rank the ETFs in Table 3. The highest Sharpe measure obtained (1.32) is by Consumer Staples. In comparison, the Sharpe measure of the benchmark S&P 500 Index is 0.71.

9 Journal of Finance, Accounting and Management, 5(1), 48-62, Jan Table 2 3-Year Performance on a Monthly Basis ( ) ETFs Avg (%) Std (%) Sharpe Beta M Squared Treynor 1 SPDR Consumer Discretionary SPDR Consumer Staples SPDR Energy Select SPDR Financial Services Select SPDR Healthcare SPDR Industrial Select SPDR Materials SPDR Technology Select SPDR Utilities Select S&P 500 Index US 4-Week Treasury Bill N/A Table 2 also reports the values of ETF Betas, M Squared measure, and Treynor measure, all of which are computed using the benchmark S&P 500 Index. The ETF with the highest systematic risk (Beta=1.28) is Financial Services Select. In comparison, the Beta of the benchmark S&P 500 Index is, by definition, exactly The ETF with the highest M Squared measure (19.99) is Consumer Staples. In comparison, the benchmark S&P 500 Index has an M Squared measure of Finally, the ETF with the highest Treynor measure (21.06) is also Consumer Staples. In comparison, the Treynor measure for the MSCI EAFE Index is Table 3 reports the rankings of all the ETFs. The Sharpe ranking indicates that only four ETFs have returns (adjusted for total risk) that exceed the risk-adjusted return of the S&P 500 Index. The Treynor ranking in Table 3 indicates that five ETFs have returns (adjusted for systematic risk)

10 Journal of Finance, Accounting and Management, 5(1), 48-62, Jan that exceed the risk-adjusted return of the S&P 500 Index. Industrial Select is the additional outperformer. The ranking based on the M Squared measure is identical to the ranking based on the Sharpe measure. However, the M Squared measure enables us to draw some inferences, which cannot be drawn from the Sharpe measure and these are detailed at the end of this section. Table 3 Three-Year Ranking ( ) Sharpe Rank Treynor ETFs (M Squared Rank) Rank SPDR Consumer Staples 1 1 SPDR Consumer Discretionary 2 3 SPDR Healthcare 3 4 SPDR Utilities Select 4 2 S&P 500 Index 5 6 SPDR Industrial Select 6 5 SPDR Technology Select 7 7 SPDR Energy Select 8 8 SPDR Materials 9 9 SPDR Financial Services Select Table 4 reports the average returns that accrue to the whole sample of ETFs with and without riskadjustment. The risk-adjustment is performed by using the S&P 500 Index as the benchmark. The returns are annualized for the convenience of investors. This is done by compounding the monthly mean returns over twelve periods. In that table, Industrial Select, which ranks second based on unadjusted returns, falls back to rank six on the basis of returns adjusted for risk. On the other hand, Consumer Staples, which ranks third on an unadjusted basis, ranks first when the returns are adjusted for risk. More strikingly, Utilities Select ranks eighth on the basis of unadjusted returns, but ranks fourth based on returns adjusted for risk. The leverage factor for this ETF is 1.57, which implies that an investor, who is comfortable with bearing the same level of risk as in the benchmark S&P 500 index, could lever the ETF (borrow 57 percent, if possible, at the risk-free rate of interest and invest all in the ETF) and thereby attain an annual return level of percent. The example below details how this return is obtained.

11 Journal of Finance, Accounting and Management, 5(1), 48-62, Jan Table 4 Three-Year Annualized Performance: Unadjusted and Adjusted for Risk ( ) Unadjusted Unadjusted Adjusted Adjusted Leverage Annualized Rank Annualized Rank Factor ETFs Returns (%) Returns (%) SPDR Consumer Staples SPDR Consumer Discretionary SPDR Healthcare SPDR Utilities Select S&P SPDR Industrial Select SPDR Technology Select SPDR Energy Select SPDR Materials SPDR Financial Services Select Consider an investor who would like to earn superior returns on an ETF and, at the same time, bear only an average level of risk. In this context, the average level of risk is measured as the standard deviation of the benchmark S&P 500 index, which is percent on a monthly basis. Now consider the following investment strategy: Suppose that the investor has $1,000 to invest. The investor could borrow $570 and invest $1,570 in Utilities Select. The end of month return from the ETF portion of the portfolio will be $1,570 x = $ Suppose that the borrowed funds were loaned at the monthly risk-free rate of percent. In that case, the borrowed funds will cost $570 x = $0.03. The portfolio return is $ $0.03 = $130.91, which is a return of percent on a monthly basis or percent (slightly off the percent in Table 4 due to rounding) on an annual basis. Note that the monthly risk of the portfolio is 1.57 x 9.62 = percent (again slightly off the percent in Table 1 due to rounding), which is the same as the monthly standard deviation of the benchmark S&P 500 Index.

12 Journal of Finance, Accounting and Management, 5(1), 48-62, Jan This investment strategy, therefore, enables the investor to earn superior returns for an average level of risk. It may be noted that the above example assumes that the returns on risk-free US Treasury Bills are not correlated with the returns on the ETF. Summary Investors who want to invest in exchange traded funds (ETFs) need to make, at least, two important decisions: one, which funds to hold, and two, how much money to invest in each. Regarding the first decision, this study evaluates the performance of the nine S&P 500 Sector ETFs and uses the M Squared measure to identify the funds that yield the highest return per unit of risk. The contribution of this study to the existing body of literature is that it is one of the first studies that apply the new M Squared measure to evaluate the performance of ETFs. Various performance measures are utilized including the Sharpe and Treynor, and the results are reported in percentage terms in a manner, which is easily comprehensible to an average investor. This is in contrast to most studies on funds, which report risk-adjusted returns using absolute performance measures that are difficult to interpret. The performance evaluation analysis in this paper is done over a three-year investment horizon ( ). Monthly returns are used to compute the performance measures. Regarding the second decision on how much money to invest, the study highlights the role of leverage in attaining a desired level of risk. For the benefit of an investor who is comfortable with an average level of risk (for example, the risk that exists in a portfolio indexed to the benchmark S&P 500), this study presents the leverage factor that an investor needs to attain. Once again, this is in contrast to the existing literature on the subject, which does not provide any guidance at all

13 Journal of Finance, Accounting and Management, 5(1), 48-62, Jan on the degree of leverage that would be needed to attain a desired level of return by investing in ETFs. Conclusion S&P 500 Sector Exchange Traded Funds (ETFs) represent a convenient investment vehicle to invest in various industries without having to worry about stock selection. There are ten Sector ETFs and each offers a diversified portfolio of companies from each sector. This study examines the characteristics of these ETFs with emphasis on the number of holdings, dividend yield, price/book ratio, price/earnings ratio, and weighted average market capitalization. This paper also provides documentation on the risk-adjusted performance of ETFs. The evaluation is based on objective performance measures grounded in modern portfolio theory. Using the methodology developed by Modigliani and Modigliani in 1997, the study reports the returns that would have accrued to these ETFs if they had the same degree of risk as that which prevails in the benchmark S&P 500 Index for a three-year holding period. It is evident from the empirical results of this study that the ETFs with the highest average returns may lose their attractiveness to investors once the degree of risk embedded in the fund has been factored into the analysis. Conversely, some ETFs, whose average (unadjusted) returns do not stand out, may look very attractive once their low risk is factored into their performance. This study also demonstrates how financial leverage can be used to raise the returns on ETFs with low risk. Alternatively, the risk of some funds can be lowered by unlevering the investor s holding.

14 Journal of Finance, Accounting and Management, 5(1), 48-62, Jan The empirical evidence presented in this study can be used as input in decision making by investors who are exploring the possibility of participating in the stock market via ETFs, but are not sure of what selection criteria to employ. On a final note, future researchers may want to update the information presented in this study on a regular basis, for the benefit of investors who are evaluating investment opportunities in ETFs.

15 Journal of Finance, Accounting and Management, 5(1), 48-62, Jan References Chang, E. C., and Lewellen, W. G. (1984). Market timing and mutual fund investment performance. Journal of Business, 57(1), Chen, H., Ho, J., Lai, C. W., and Morales-Camargo, E. (2011). Institutional Trading Behavior in the ETF Market. Midwest Finance Association 2012 Annual Meetings Paper. Available at SSRN: Gleason, K. C., Mathur, I., and Peterson, M. A. (2004). Analysis of intraday herding behavior among the sector ETFs. Journal of Empirical Finance, 11(5), Hamm, S. J. W. (2011). The Effect of ETFs on Stock Liquidity. Available at SSRN: Hasbrouck, J. (2001). Intraday Price Formation in US Equity Index Markets. Available at SSRN: Henrikkson, R. D., and Merton, R. C. (1981). On market timing and investment performance II: Statistical procedures for evaluating forecasting skills. Journal of Business, 54(4), Huang, J. C. and Guedj, I. (2009). Are ETFs Replacing Index Mutual Funds? American Finance Association San Francisco Meetings Paper. Available at SSRN: Jensen, M. C. (1968). The performance of mutual funds in the period Journal of Finance, 23(2), Kon, S. J., and Jen, F. C. (1979). The investment performance of mutual funds: An empirical investigation of timing, selectivity, and market efficiency. Journal of Business, 52(2), Modigliani, F., and Modigliani, L. (1997). Risk-adjusted performance. Journal of Portfolio Management, 23(2), Sharpe, W. F. (1966). Mutual fund performance. Journal of Business, 39(1), Treynor, J. L. (1965). How to rate management of investment funds. Harvard Business Review, 43(1), Author Biographies Dr. Onur Arugaslan, Associate Professor of Finance at Western Michigan University, received his Ph.D. in Finance from the University of Texas at Dallas. His research includes the separation of cash flow rights and voting rights, the endogeneity of liquidity, the risk-adjusted performance of American Depository Receipts, and the market reaction to the acquisitions by unified dual class firms. Arugaslan has published articles in various journals including the Journal of Finance and the Journal of Corporate Finance. Dr. Ajay Samant, Ph.D., Indiana University, serves as Dean and Professor of Finance at the University of North Florida. His published research includes the areas of international financial markets, performance evaluation of mutual funds and interest rate swaps. He has served as keynote speaker at international business conferences and has been interviewed on national television for expert opinions in the areas of banking and financial markets.