CENTER FOR HOSPITALITY RESEARCH The Role of REIT Preferred and Common Stock in Diversified Portfolios By Walter I. Boudry, Jan A. deRoos, and Andrey D. Ukhov EXECUTIVE SUMMARY While “maximizing returns” is a stated goal of many investors, it is clear that some are more willing than others to embrace risk in their pursuit of those returns. An analysis of risk-return profiles finds that investors see different purposes for real estate investment trust (REIT) common stock and preferred stock depending on their tolerance for risk. Using a utility-based approach and imposing realistic constraints on the investor’s portfolio, this report shows that REIT preferred and common stock provide diversification benefits, but to different sets of investors. Risk tolerant investors find REIT common stock beneficial, while risk averse investors find the preferred stock more favorable. The key highlight from the study is that investors, especially those who have investment grade bonds, should consider adding REIT preferred stock to their portfolios. Cornell Hospitality Report • December 2016 • www.chr.cornell.edu • Vol. 16, No. 28 1 ABOUT THE AUTHORS Walter I Boudry, Ph.D.,is an assistant professor of real estate in the School of Hotel Administration at Cornell University. Prior to joining the School of Hotel Administration, he taught both undergraduate and graduate real estate courses at the University of North Carolina at Chapel Hill’s Kenan-Flagler Business School and New York University’s Leonard N. Stern School of Business. His research interests include real estate and general finance. His most recent works have focused on Real Estate Investment Trusts (REITs) and have examined their repurchase decisions, dividend payout policy, security issuance decisions and price dynamics. A regular presenter and discussant at the annual AREUEA meetings, his papers have been published in Real Estate Economics, the Journal of Real Estate Finance and Economics, and the Journal of Business Finance and Accounting. He received his B.A. in Commerce with first class honors and a B.A. in Economics from the University of Queensland, Australia, and his M.Phil. and Ph.D. from the Leonard N. Stern School of Business at New York University. Jan A. deRoos, Ph.D., is the HVS Professor of Hotel Finance and Real Estate in the School of Hotel Administration at Cornell University, where he has taught since 1988. He has devoted his career to teaching and research related to hospitality real estate, with a focus on the valuation, financing, development, and control of lodging, timeshare, and restaurant assets. He co-developed a free tool, the Hotel Valuation Software, with Stephen Rushmore of HVS International, and has developed a respected online executive education curriculum for hotel real estate professionals. His book on hotel management agreements, co-authored with the late James Eyster, is the seminal academic publication on the topic. Prior to joining Cornell, deRoos worked extensively in the hotel industry as a construction and engineering manager. He received his B.S., M.S. and Ph.D. degrees from the School of Hotel Administration at Cornell University. Andrey D. Ukhov, Ph.D., is an assistant professor of finance in the School of Hotel Administration at Cornell University. He is an authority on investments, including preferred stocks, warrants, derivative securities, and convertibles. His research papers have been published in Management Science, the Journal of Financial and Quantitative Analysis, the Review of Finance, Quantitative Finance, the Economic History Review, the Journal of Real Estate Research, and other academic journals. Prior to joining the School of Hotel Administration, he taught both undergraduate and graduate finance courses at the Kelley School of Business at Indiana University and the Kellogg School of Management at Northwestern University. He has received numerous teaching awards at Cornell, Indiana, and Northwestern for undergraduate-, master’s-, and Ph.D.-level courses. He received two U.S. patents for technology inventions. He received his B.A. in Economics with distinction, and his M.A., M.Phil., and Ph.D. in Financial Economics, all from Yale University. 2 The Center for Hospitality Research • Cornell University CORNELL HOSPITALITY REPORT The Role of REIT Preferred and Common Stock in Diversified Portfolios By Walter I. Boudry, Jan A. deRoos, and Andrey D. Ukhov People planning for retirement must eventually decide how much of their savings to put in stocks, bonds, or other assets. This report sheds some light on the diversification benefits of the real estate investment trust (REIT) market. There are two primary reasons for exploring this issue. First, little is known about the risk-return characteristics of REIT preferred stock. This is a significant omission, because preferred stock constitutes 20 percent of all the public equity that REITs issue.1 Second, using standard mean-variance tools, studies show that investors with well diversified portfolios receive little benefit from access to REIT common stock. Our research examines whether this is also true for REIT preferred stock. 1 http://www.reit.com/DataAndResearch/REIT-Capital-Offerings/Detailed-Data.aspx. Cornell Hospitality Report • December 2016 • www.chr.cornell.edu • Vol. 16, No. 28 3 Exhibit 1 REIT security issuances ($millions) Year Debt IPO Preferred Seasoned 1992 310 693 46 808 1993 2,348 8,485 666 2,609 1994 3,173 6,714 155 3,337 1995 3,324 827 1,678 4,727 1996 4,327 1,108 1,550 8,561 1997 9,785 4,776 4,795 19,381 1998 13,941 1,269 4,879 12,006 1999 9,555 292 2,150 1,966 2000 6,045 — 365 1,171 2001 8,650 — 679 1,769 2002 8,353 517 1,067 3,342 2003 9,958 2,325 4,905 4,484 2004 16,956 4,581 4,822 6,698 2005 15,515 1,726 2,735 6,805 2006 24,322 1,824 3,751 13,554 2007 15,765 737 2,998 7,243 2008 4,343 — 947 7,492 2009 10,193 633 — 18,172 2010 18,444 1,987 1,631 18,742 2011 13,525 1,766 4,108 14,782 2012 19,400 1,049 8,466 23,815 Total 218,229 41,309 52,390 181,463 % of Total 44% 8% 11% 37% Notes. The figures above are the U.S. dollar amounts of securities issued by U.S. REITs each year, expressed in millions of dollars. The methodology employed in this study is the same as that risk-free asset. The slope of the red line is determined by the described in a recent article we prepared for Real Estate Econom- Sharpe ratio, and any asset that increases the Sharpe ratio helps ics.2 While the full derivation and analytics can be found in that all investors.4 If an asset that is part of the tangency portfolio paper, the intuition behind their framework is straightforward. helps improve the slope of the red line (increases the Sharpe In the standard mean-variance realm introduced by Markowitz, ratio), it is valuable to all investors regardless of risk preferences. investors are unconstrained in their portfolio formation and may Among the assumptions behind the standard mean-vari- borrow at the risk-free rate.3 In this framework, all investors hold ance setting is that the investors have no limitations in forming the same collection of risky assets and adjust their allocation to their portfolios. This means that the investors can short any the risk-free asset based on their risk preferences. This can be asset in any amount. They can also borrow any amount needed seen in the standard mean-variance frontier in Exhibit 2. All to finance this investment at the same rate as the U.S. govern- optimal portfolios lie on the red line, and all portfolios on the ment. In practice, both of these assumptions are unlikely to be red line are a combination of the tangency portfolio and the true. While some assets can be shorted, it is typically costly to do so, and no investor can borrow money as cheaply as the govern- 2 Boudry, W.I., J.A. deRoos, and A.D.Ukhov. 2016. “Diversification Benefits of REIT Preferred and Common Stock: New Evidence from a Utility 4 The Sharpe ratio indicates the average return minus the risk-free based Framework.” Real Estate Economics, forthcoming. return divided by the standard deviation of return on an investment. See: 3 Markowitz, H.M. 1952. Portfolio Selection. Journal of Finance 7: pp. William F. Sharpe (1966). “Mutual Fund Performance.” Journal of Business 39 77-91. (S1): pp. 119–138. 4 The Center for Hospitality Research • Cornell University Exhibit 2 Exhibit 3 Classical mean-variance frontier Constrained mean-variance frontier Mean Mean-Variance Mean Return Efficient Portfolios Return Highest Return Portfolio Tangency Portfolio Tangency Portfolio Rf Rf Standard Deviation Standard Deviation Notes. The horizontal axis is the portfolio standard deviation (or risk) and the Notes. The horizontal axis is the portfolio standard deviation (or risk) and the vertical axis is the annual portfolio return. The parabola traces the maximum return vertical axis is the annual portfolio return. The line shows the maximum return at possible at each level of risk. The red line shows the maximum return at each level each level of risk using a constrained (realistic) investment environment. of risk using a set of stylized, unrealistic assumptions. ment. We show that when short sales and borrowing constraints Data are imposed on the investor’s portfolio, the results from the In considering the diversification benefits of REITs, a standard mean-variance setting regarding diversification no well done study should allow investors to have a large menu longer hold.5 of securities in which to invest. To create this list we collected Analyzing the investor’s constrained portfolio problem, we monthly returns for 13 indices for the period November 1992 show that imposing short sales and borrowing restrictions on to November 2012. From Datastream we obtained returns on the investor leads to a new efficient frontier, as seen in Exhibit the Barclays Investment Grade Corporate Bond Index, the 3.6 The frontier is no longer represented by a straight line; rather Barclays High Yield Corporate Bond Index, the MSCI World it now has segments (from the risk free rate [rf], to the tangency Ex-US index, the Russell 2000 index, the Russell 2000 growth portfolio, tangency portfolio to the highest return portfolio, and index, the Russell 2000 value index, the Russell Mid Cap index, the highest return portfolio itself). There are two key points to the Russell Mid Cap growth index, and the Russell Mid Cap note about these segments. First, the risky assets that constitute value index. We obtained the SNL US Equity REIT index each segment are different. This means that optimal portfolios from SNL Financial. From MSCI we obtained the MSCI REIT in different parts of the frontier have different risky assets. Sec- Preferred Index, and from CRSP we obtained returns on the ond, investors with different levels of risk aversion will invest in 30-day T-bill and the returns on the S&P500. As a result, our different segments on the frontier. Combining these two points investor has access to international stocks, U.S. large-, mid-, and results in the following outcome: investors with different levels small-cap stocks, value and growth portfolios, and high yield of risk aversion will hold different portfolios of risky assets. Thus, and investment grade bonds in addition to the REIT common we can no longer define diversification benefits in the aggregate and preferred indices. using summary measures such as Sharpe ratios. When consider- All of the indices used in the analysis are standard, apart ing the diversification benefits of various assets, it is important to from the MSCI REIT Preferred Index. The terms of the pre- identify the risk preferences of the investor. ferred stock issued by REITs are standardized. REIT preferred 5 Boudry et al., 2016. stock has a fixed dividend rate, is typically issued at par, and is 6 Ibid. callable after five years. Preferred stock has priority to common Cornell Hospitality Report • December 2016 • www.chr.cornell.edu • Vol. 16, No. 28 5 Exhibit 4 Return and correlation percentages Pref REIT World SP500 MidCap MidCap MidCap Rus2000 Rus2000 Rus2000 IG HY Ex-US Growth Value Growth Value Bonds Bonds Panel A: Annualized Mean and Standard Deviation Mean 10.26 12.89 7.85 9.20 11.47 10.75 11.80 10.35 9.10 11.46 7.06 8.33 Std 11.35 20.11 16.95 15.10 16.84 21.34 16.02 19.62 23.26 17.40 5.55 8.98 Panel B: Correlation Pref 100.00 62.50 39.63 38.49 46.54 34.18 52.10 41.67 33.69 49.50 53.89 67.62 REIT 62.50 100.00 54.37 55.99 65.36 46.84 74.61 64.96 52.04 77.30 30.73 60.85 World Ex-US 39.63 54.37 100.00 81.51 81.94 75.13 77.31 74.51 71.55 71.88 28.63 62.84 SP500 38.49 55.99 81.51 100.00 92.50 85.46 88.12 80.92 77.98 78.95 26.60 61.84 MidCap 46.54 65.36 81.94 92.50 100.00 92.81 93.31 93.05 89.43 89.89 28.00 68.03 MidCap 34.18 46.84 75.13 85.46 92.81 100.00 73.77 90.34 94.58 75.67 20.56 60.11 Growth MidCap 52.10 74.61 77.31 88.12 93.31 73.77 100.00 82.19 71.68 90.94 30.63 66.03 Value Rus2000 41.67 64.96 74.51 80.92 93.05 90.34 82.19 100.00 97.18 93.81 19.16 62.78 Rus2000g 33.69 52.04 71.55 77.98 89.43 94.58 71.68 97.18 100.00 83.24 15.59 58.77 Rus2000v 49.50 77.30 71.88 78.95 89.89 75.67 90.94 93.81 83.24 100.00 22.10 62.95 IG Bonds 53.89 30.73 28.63 26.60 28.00 20.56 30.63 19.16 15.59 22.10 100.00 54.13 HY Bonds 67.62 60.85 62.84 61.84 68.03 60.11 66.03 62.78 58.77 62.95 54.13 100.00 Notes. This exhibit shows descriptive statistics of monthly returns on 13 indices for the period November 1992 to November 2012. Panel A shows annualized mean returns and standard deviations expressed as a percentage, while Panel B shows return correlations expressed as a percentage. Pref is the MSCI REIT Preferred index, REIT is the SNL Equity REIT index, World Ex-US is the SCI World Ex-US index, SP500 is the S&P 500 index, MidCap is the Russell MidCap index, MidCap Growth is the Russell MidCap Growth index, MidCap Value is the Russell MidCap Value index, Rus2000 is the Russell 2000 index, Rus2000 Growth is the Russell 2000 Growth index, Rus2000 Value is the Russell 2000 value index, IG Bond is the Barclays Investment Grade Corporate Bond index, and HY Bond is the Barclays High Yield Corporate Bond index. This is the same sample as Boudry, deRoos, and Ukhov (2016). stock in the payment of dividends and unpaid dividend sums. stock, evident in their 67.62 percent and 62.50 percent correla- The MSCI REIT Preferred Index is a broad-based fund consist- tions with high yield bonds and REIT common stock. Second, ing of non-convertible preferred stock traded on the NYSE, the 74.61 percent and 77.30 percent correlations between AMEX, or NASDAQ that are issued by public U.S. equity and REIT common stock and the MidCap Value and Russell 2000 hybrid REITs. Value indices highlight the small and midcap value nature of Exhibit 4 provides statistics for the indices used in our analy- the REIT realm. Finally, investment grade bonds are shown to sis. Panel A shows mean return percentages and standard devia- be the asset class least correlated with the other asset classes in tions, while Panel B shows return correlations. As shown in Panel our analysis. A of Exhibit 4, the average annualized return for REIT preferred Exhibit 5 shows the current value of a dollar invested in stock is 10.26 percent with a standard deviation of 11.35 percent. each asset class in 1992. Some clear patterns emerge regarding This compares to an average return of 12.89 percent for REIT the time series behavior of the asset classes. Consistent with common stock, which has the highest average return in our the average returns observed in Exhibit 4, REIT common sample. The S&P 500 had a mean return of 9.20 percent, the stock was the best performing asset class during the sample lowest of the domestic equity asset classes. As expected, invest- period. This was not always the case, however. Prior to 2000 ment grade bonds had the lowest average return of 7.06 percent, REIT common shares were nearly always the worst perform- while high yield bonds did slightly better with an average return ing equity index, and up to that time had a total return nearly of 8.33 percent. identical to high yield corporate bonds. In fact, prior to 2002, The return correlations in Panel B show some noteworthy REIT common stock and preferred stock had similar results. characteristics about the asset classes. First, preferred stock shares From 2002 to 2007 we observed a marked increase in the a high correlation with high yield bonds and REIT common performance of small and midcap value, and REIT common 6 The Center for Hospitality Research • Cornell University Exhibit 5 Value of a dollar invested $9.00 $8.00 $7.00 $6.00 $5.00 $4.00 $3.00 $2.00 $1.00 Prefs REIT Common Mid Cap Value S&P 500 MSCI World MSCI World Ex US Mid Cap Mid Cap Growth Russell 2000 Russell 2000 Growth Russell 2000 Value T-Bill Investment Grade Bonds High Yield Bonds Exhibit 6 Value of a dollar invested (select asset classes) $9.00 $8.00 $7.00 $6.00 $5.00 $4.00 $3.00 $2.00 $1.00 Prefs REIT Common Mid Cap Value S&P 500 shares. The close relationship between REIT common shares Analysis and midcap value stocks is apparent in Exhibit 6, where we plot As a starting point for understanding the diversification just the REIT common and preferred shares, the S&P 500, and benefits of REIT preferred and common stock, we first calcu- the midcap value index. The effects of the global financial crisis, late improvements in the Sharpe ratio from giving an investor subsequent recession, and recovery are evident in the latter part access to the REIT market. In this sense we gauge the diversifi- of the sample. All risky asset classes declined significantly during cation benefits of the market using the classical mean-variance the crisis and rebounded during the recovery. This V-shaped framework. The Sharpe ratio is the slope of the red line in pattern is pronounced in both the REIT common and preferred Exhibit 2. The higher this slope, the better off all investors are. shares. Cornell Hospitality Report • December 2016 • www.chr.cornell.edu • Vol. 16, No. 28 7 10/1/1992 10/1/1992 7/1/1993 5/1/1993 2/1/1994 4/1/1994 10/1/1994 1/1/1995 2/1/1995 10/1/1995 10/1/1995 7/1/1996 5/1/1997 4/1/1997 2/1/1998 1/1/1998 10/1/1998 10/1/1998 5/1/1999 7/1/1999 2/1/2000 4/1/2000 10/1/2001 1/1/2001 5/1/2001 10/1/2001 2/1/2002 7/1/2002 10/1/2002 5/1/2003 4/1/2003 2/1/2004 1/1/2004 10/1/2004 10/1/2004 2/1/2005 7/1/2005 10/1/2005 4/1/2006 5/1/2007 1/1/2007 2/1/2008 10/1/2007 10/1/2007 7/1/2008 10/1/2008 4/1/2009 4/1/2009 1/1/2010 2/1/2010 10/1/2010 10/1/2010 7/1/2011 5/1/2011 2/1/2012 4/1/2012 10/1/2012 Exhibit 7 Classical Sharpe ratio analysis Equity Only Equity and Bonds Market Market Sharpe Increase Market Market Sharpe Increase Expected St.Dev Ratio in Sharpe Expected St.Dev. Ratio in Sharpe Return (%) Return (%) Panel A: All Equity Indices Including Prefs & 0.158 0.140 0.908 8.258 0.107 0.076 0.998 0.383 REITS Including REITs, No 0.206 0.207 0.845 0.661 0.106 0.075 0.995 0.070 Prefs Including Prefs, No 0.159 0.142 0.907 8.134 0.107 0.077 0.997 0.244 REITs No REITs, No Prefs 0.207 0.211 0.839 0.106 0.076 0.994 Panel B: S&P 500 and World Ex-US only Including Prefs & 0.107 0.112 0.683 64.143 0.082 0.062 0.823 3.846 REITS Including REITs, No 0.123 0.173 0.536 28.619 0.080 0.060 0.812 2.404 Prefs Including Prefs, No 0.105 0.109 0.680 63.394 0.081 0.061 0.817 3.059 REITs No REITs, No Prefs 0.096 0.158 0.416 0.076 0.057 0.793 Notes. This exhibit shows annualized mean return (expressed as a percentage), annualized standard deviation (expressed as a percentage), and the increase in the Sharpe ratio for optimal portfolios constructed using different investment opportunity sets. Panel A considers the case where the investor has access to all of the equity indices, while Panel B considers the case where the investor only has access to the S&P 500 and the World Ex-US. Lefthand columns consider the case where the investor has no access to bonds, while the righthand columns consider the case where the investor has access to bonds. In each panel, statistics for portfolios including REIT preferred stock, REIT common stock, and both REIT preferred and common stock are reported in addition to the improvement in the Sharpe ratio from the case where no REIT access is allowed. Equity indices are MSCI World Ex-US Index, S&P 500 Index, Russell Mid Cap Index, Russell Mid Cap Growth Index, Russell Mid Cap Value Index, Russell 2000, Russell 2000 Growth, and Russell 2000 Value. Bond indices are Barclays Investment Grade Bond Index and Barclay’s High Yield Bond Index. This is why it is used as a summary measure of diversification still helps investors, improving Sharpe ratios by 8.1 percent, but benefits in the classical mean variance setting. moving across to the righthand columns of Panel A we observe Exhibit 7 shows the improvements in Sharpe ratios when that this effect disappears in the presence of bonds. the investor is given access to the REIT market. In Panel A we The conclusion to draw from Exhibit 7 is that the REIT consider the case of a well-diversified investor who has access to market provides no benefits to a well-diversified investor. Still, all of the equity indices, while in Panel B we consider the case using Sharpe ratios to measure diversification benefits is predi- of a less-diversified investor who only has access to the S&P 500 cated on the unconstrained portfolio formation of the classical and the World Ex-US. Lefthand columns exclude bonds, while mean-variance setting. When normal constraints are imposed righthand columns include access to investment grade and high on the investor’s portfolio, it is important to consider alternatives yield bonds. regarding the assessment of diversification benefits. Starting in the lefthand columns of Panel B, we see that To further examine the role that REIT preferred and access to the REIT market is valuable for a poorly diversified common stock play in an investor’s portfolio, the constrained investor. Being able to invest in REIT preferred stock improves portfolio issue is estimated using historic data. In this setting, the Sharpe ratios by 63.4 percent, while access to REIT common investor can no longer borrow at the risk-free rate or short sell stock improves Sharpe ratios by 28.6 percent. To a poorly diver- the risky assets. This study shows optimal portfolio allocations sified investor, then, the REIT market is quite valuable. However, across the different asset classes for investors with different levels if we look at Panel A, we see that REIT common stock provides of risk aversion. In any portfolio allocation experiment, the port- no benefit to investors. Access to REIT common stock only folio weights obtained are a function of the time period used in improves Sharpe ratios by 0.66 percent. REIT preferred stock the estimation. To avoid the possibility that any results are an ar- 8 The Center for Hospitality Research • Cornell University Exhibit 8 Optimal portfolio weights Portfolio Weights (All Assets) 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0 1 2 3 4 5 6 7 8 9 10 11 12 13 Risk Aversion Pref REIT IG Debt HY Debt Equity Rf Notes: The lines trace the weight (vertical axis) expressed as a decimal for each of the asset classes in a portfolio. For any level of risk aversion, the sum of the weight totals 100%. tifact of the chosen sample period, we conducted a Monte Carlo The numbers from left to right reflect an increased aversion to simulation to generate 10,000 possible histories, incorporating risk. Several noteworthy results are evident in Exhibit 8. Very a wide variety of random values and outcomes.7 Each history risk tolerant investors (risk aversion of 1-4) have portfolios that is created by sampling 241 months from the actual sample of are dominated by equity securities and REIT common stock. return using replacement. This sampling procedure means that This is because REIT common stock allows the investor to form inside the Monte Carlo histories, there may be an environment high-return portfolios. As risk aversion increases, we observe that is in permanent recession, another that is permanently in a that the investor quickly moves out of REIT common stock and boom, and another that replicates the historic sample. Calcu- into REIT preferred stock. At the same time the investor also lating the average portfolio weights allows us to determine the reduces his allocation to the other equity asset classes and moves stability of any diversification benefits. toward investment grade and high yield bonds. Finally, for the Exhibit 8 shows the mean portfolio allocations for each risk-averse investor (risk aversion of 10-14) we see risk-free assets asset class, averaged across the 10,000 histories. Blue is the dominating the portfolio. Over a wide range of risk aversions, allocation to REIT preferred stock, red is the allocation to REIT preferred stock forms a material part of the investor’s REIT common stock, green is the allocation to the other equity portfolio. However, for REIT common stock, we observe that asset classes, brown is the allocation to investment grade bonds, this asset class is mainly valuable to very risk-tolerant investors. orange is the allocation to high yield bonds, and black is the al- So the REIT market does provide diversification benefits, but it location to the risk-free asset. The horizontal axis measures risk provides them to different sets of investors. aversion, with higher numbers reflecting a greater risk aversion. A classic investment perspective on preferred stock pre- sented by Graham and Dodd in their value investing text shows 7 Monte Carlo simulations are used to model the probability of different outcomes in a process that cannot be easily predicted due to the intervention of random variables. Cornell Hospitality Report • December 2016 • www.chr.cornell.edu • Vol. 16, No. 28 9 Portfolio Weight Exhibit 9 Optimal portfolio weights (excluding preferred stock) Portfolio Weights (Without Pref) 0.70 0.65 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0 1 2 3 4 5 6 7 8 9 10 11 12 13 Risk Aversion REIT IG Debt HY Debt Equity Rf Notes: The lines trace the weight (vertical axis) expressed as a decimal for each of the asset classes in a portfolio. For any level of risk aversion, the sum of the weight adds to 100%. that preferred stock is a hybrid of debt and equity.8 To examine in allocation to given asset classes when REIT preferred stock, this issue, we repeat the analysis from above, but remove REIT common stock, and both preferred stock and common stock are preferred stock from the investment menu. In this sense, a added to the investors’ choices. comparison of Exhibit 8 to the weights reported in Exhibit 9 Exhibit 10 also shows that the addition of REIT preferred shows which assets displace REIT preferred stock in the inves- stock results in lower allocations to REIT common stock over tor’s portfolio. all levels of risk aversion, but most markedly for low risk- A comparison of Exhibit 8 to Exhibit 9 shows that REIT aversion investors. For moderate risk-aversion investors, there is preferred stock displaces allocations to REIT common stock a dramatic decline in the allocation to investment grade bonds. and investment grade and high yield bonds, supporting Gra- Regarding REIT common stock, there is a large displacement ham and Dodd’s thesis. However, when investors have access to of allocation to the other equity asset classes, particularly for low REIT preferred stock, they choose to invest in it. This implies risk-aversion investors, who reduce their allocation to the other that, although REIT preferred stock behaves like a hybrid of equity classes by 26.1 percentage points. These investors also REIT common stock and investment grade and high yield move a substantial amount of their funds to REIT preferred bonds, investors prefer to invest in REIT preferred stock than to stock. REIT common stock doesn’t displace allocations signifi- form a replicating portfolio using those other asset classes. This cantly for high risk-aversion investors, because these investors shows that REIT preferred stock has a risk-return profile that find little benefit from REIT common stock. is not easily replicated by other asset classes. The allocation of Access to both REIT common and preferred stock shows 20 to 30 percent of the investors’ portfolio to REIT preferred significant movement of allocations across all levels of risk aver- stock also shows that investors make a big mistake by ignoring sion. Once again, this is driven by REIT common stock for less the asset class. risk-averse investors and by REIT preferred stock for more risk- To further examine the role REIT stock plays in the averse investors. These changes are most pronounced among investors’ optimal portfolio, in Exhibit 10 we show the change the other equity asset classes, with low risk aversion investors reducing their allocation by 43.1 percent. 8 Graham, B., and D. Dodd. 1934. Security Analysis. McGraw-Hill, New To provide a different view of the results from Exhibits 8 York. and 9, and to highlight the results from Exhibits 2 and 3, we 10 The Center for Hospitality Research • Cornell University Portfolio Weight Exhibit 10 Changes in portfolio weights when real estate stocks are added to the investible universe REIT Preferred Added REIT Common Added REIT Common and to the Investment Univerise to the Investment Universe Preferred Added to the Investment Universe Risk Rf REIT Equity IG HY Rf REIT- Equity IG HY Rf Equity IG HY Aversion Com- Debt Debt Pref Debt Debt Debt Debt mon 1 0.0 -5.8 -5.7 -3.9 -2.3 0.0 -10.2 -26.1 -0.4 -0.5 0.0 -43.1 -6.4 -5.4 2 0.0 -6.3 -6.9 -7.0 -4.2 0.0 -9.2 -18.3 -0.3 -0.6 0.0 -35.0 -10.0 -7.8 3 0.0 -6.1 -6.4 -10.7 -5.8 0.0 -7.7 -12.8 -0.1 -0.7 0.0 -27.0 -13.6 -9.6 4 -0.1 -5.5 -5.1 -13.4 -6.7 0.0 -6.3 -9.4 0.2 -0.7 0.0 -20.5 -16.1 -10.3 5 -0.1 -4.8 -3.8 -14.9 -6.7 0.0 -5.2 -7.3 0.4 -0.6 -0.1 -15.8 -17.2 -10.0 6 -0.1 -4.1 -2.9 -15.3 -6.3 0.0 -4.4 -5.9 0.6 -0.6 0.0 -12.5 -17.2 -9.2 7 0.0 -3.5 -2.2 -14.8 -5.7 0.0 -3.7 -5.0 0.6 -0.6 0.1 -10.3 -16.5 -8.2 8 0.1 -3.0 -1.8 -14.0 -5.1 0.0 -3.2 -4.3 0.6 -0.6 0.2 -8.7 -15.4 -7.3 9 0.1 -2.7 -1.5 -13.0 -4.5 0.1 -2.8 -3.8 0.6 -0.6 0.3 -7.6 -14.3 -6.4 10 0.2 -2.3 -1.4 -12.0 -4.0 0.1 -2.4 -3.4 0.5 -0.5 0.4 -6.7 -13.2 -5.7 11 0.3 -2.1 -1.2 -11.1 -3.6 0.1 -2.2 -3.0 0.5 -0.5 0.6 -6.0 -12.3 -5.2 12 0.4 -1.8 -1.1 -10.3 -3.3 0.1 -1.9 -2.8 0.4 -0.5 0.7 -5.5 -11.4 -4.7 13 0.4 -1.7 -1.0 -9.6 -3.1 0.1 -1.8 -2.5 0.4 -0.4 0.8 -5.0 -10.7 -4.3 This exhibit shows changes in portfolio weights (expressed as percentages) when REIT common stock, REIT preferred stock, and both REIT common and preferred stock are added to the investor’s investment universe. For each case the remaining investment universe includes MSCI World Ex-US Index, S&P 500 Index, Russell Mid Cap Index, Russell Mid Cap Growth Index, Russell Mid Cap Value Index, Russell 2000, Russell 2000 Growth, Russell 2000 Value, Barclays Investment Grade Bond Index, and Barclay’s High Yield Bond Index. report changes in the mean Sharpe ratio, portfolio return, and classes. The significant portfolio allocations to REIT common portfolio standard deviation in Exhibit 11, when REIT common and preferred stock observed in Exhibit 8 show that these asset stock, preferred stock, and both common and preferred stock classes are important to investors. are added to the investors’ portfolios. The striking result from Exhibit 11 is that when risk-toler- Conclusions and Implications ant investors have access to REIT common stock, their optimal In an environment where investors face constraints on their portfolio Sharpe ratio declines. They form a portfolio that has ability to short risky assets and to borrow at the risk-free rate, a higher mean return (by 7.4 percent), but the portfolio also the diversification benefits of an asset class are dependent on has a high standard deviation (by 10.8 percent,) and the Sharpe risk aversion. An asset may provide diversification benefits to ratio falls by 0.58 percent. This is highly counterintuitive, given one set of investors, but might not be of use to another. This is that these individuals invest most heavily in REIT common fundamentally different from the standard Markowitz setting, in stock. If investing in REIT common stock is not the optimal which all investors hold the same portfolio of risk assets, and as choice, they could choose to not invest in it. While this appears such, if an asset provides diversification benefits to any investor, to suggest that adding REIT common stock to a portfolio is a it provides those benefits to all investors. This also means that bad option, it is in fact a limitation of using Sharpe ratios to standard metrics of investment performance, such as Sharpe ra- analyze portfolio performance in the constrained setting. This is tios, are no longer valid measures of performance. Confronted an application of the previous discussion related to the differ- with these constraints, investors must be aware of the limitations ences between Exhibits 2 and 3. When there are constraints of such commonly used metrics. on the portfolio, Sharpe ratios cannot be used to judge the best In examining the constrained portfolio problem, we find investment choices. Similarly, it doesn’t make sense to compare that REIT common stock benefits risk-tolerant investors by returns or standard deviations. Portfolio weights are the correct allowing them to form high return portfolios. While measuring way for investors to judge the economic significance of the asset risk aversion is difficult, it is possible to see what the portfolios Cornell Hospitality Report • December 2016 • www.chr.cornell.edu • Vol. 16, No. 28 11 Exhibit 11 Changes in Sharpe Ratios, portfolio returns, and portfolio standard deviation REIT Common REIT Preferred REIT Common and Preferred Risk Sharpe Portfolio Portfolio Sharpe Portfolio Portfolio Sharpe Portfolio Portfolio Aversion Ratio Mean Std Ratio Mean Std Ratio Mean Std 1 -0.58 7.37 10.79 7.61 0.94 -3.71 9.91 9.94 6.21 2 -0.30 6.93 9.59 6.86 1.16 -3.45 8.83 9.92 6.40 3 0.19 6.07 7.79 5.73 2.10 -1.48 7.58 10.30 7.94 4 0.59 5.10 6.08 4.72 3.38 1.06 6.55 10.95 9.90 5 0.85 4.27 4.77 4.02 4.61 3.37 5.91 11.45 11.46 6 1.01 3.61 3.81 3.61 5.47 4.93 5.60 11.62 12.14 7 1.10 3.10 3.12 3.41 5.89 5.71 5.49 11.41 12.05 8 1.16 2.72 2.64 3.34 5.94 5.87 5.51 10.91 11.42 9 1.20 2.44 2.30 3.33 5.78 5.69 5.57 10.28 10.56 10 1.22 2.23 2.05 3.35 5.52 5.36 5.65 9.62 9.68 11 1.24 2.06 1.86 3.37 5.22 4.98 5.72 8.98 8.84 12 1.25 1.92 1.71 3.40 4.94 4.63 5.78 8.39 8.10 13 1.26 1.80 1.59 3.43 4.67 4.33 5.82 7.87 7.48 Notes: This shows percentage changes in Sharpe Ratios, Portfolio Returns, and Portfolio Standard Deviations when REIT Common stock, REIT Preferred stock, and both REIT Common and Preferred stock are added to the investor’s investable universe. For each case the remaining investable universe includes MSCI World Ex-US Index, S&P 500 Index, Russell Mid Cap Index, Russell Mid Cap Growth Index, Russell Mid Cap Value Index, Russell 2000, Russell 2000 Growth, Russell 2000 Value, Barclays Investment Grade Bond Index and Barclay’s High Yield Bond Index. of these investors look like. A low risk-aversion investor, who such as investment grade and high yield bonds. In our analysis, cares more about returns, holds a portfolio that is dominated the risk-return profile of REIT preferred stock is extremely by stocks, with fewer alternative asset classes. These individuals valuable to these investors as it allows them to reduce risk with- would be well served by examining the REIT market. This is out sacrificing as much return. Investors who like investment now easier with S&P separating real estate from financials, and grade bonds should also consider the REIT preferred stock designating the real estate sector as the 11th Global Industry because it likely has a risk-return profile that suits their invest- Classification Standard (GICS) in its indices. ment profiles. In contrast to REIT common stock, REIT preferred stock The results of our analysis also have a practical implica- is valued by moderately risk-averse investors, because it provides tion for issuers of REIT preferred stock. We show that REIT a venue for risk reduction. As risk aversion increases, investors preferred stock is a valuable diversifying asset for investors pay more attention to the risk-return tradeoff, and no longer because of its risk-return profile. Issuers can reduce the yield on focus only on more balanced portfolios, with this action coming their preferred stock and still have it be a valuable asset class to at the expense of their stock allocations. While they reduce their investors. In this sense, issuers may have been giving preferred interest in stocks, they start investing in fixed income securities stock investors too good of a deal. n 12 The Center for Hospitality Research • Cornell University Center for Hospitality Research Publication Index chr.cornell.edu 2016 Reports Vol. 16 No. 17 Highlights from the 2016 Vol. 16 No. 7 Instructions for the Food Sustainable and Social Entrepreneurship Preparation Scheduling Tool v2015, by Gary Vol. 16 No. 27 Do You Look Like Me? How Enterprises Roundtable, by Jeanne Varney Thompson, Ph.D. Bias Affects Affirmative Action in Hiring, by Ozias Moore, Ph.D., Alex M. Susskind, Vol. 16 No. 16 Hotel Sustainability Vol. 16 No. 6 Compendium 2016 Ph.D., and Beth Livingston, Ph.D. Benchmarking Index 2016: Energy, Water, and Carbon, by Eric Ricaurte Vol. 16 No. 5 Executive Insights on Leader Vol. 16 No. 26 The Effect of Rise in Integrity: The Credibility Challenge, by Interest Rates on Hotel Capitalization Rates, Vol. 16 No. 15 Hotel Profit Implications Tony Simons, Ph.D., with Kurt Schnaubelt, by John B. Corgel, Ph.D. from Rising Wages and Inflation in the U.S., John Longstreet, Michele Sarkisian, Heather by Jack Corgel, Ph.D. Allen, and Charles Feltman Vol. 16 No. 25 High-Tech, High Touch: Highlights from the 2016 Entrepreneurship Vol. 16 No. 14 The Business Case for (and Vol. 16 No. 4 Authenticity in Scaling the Roundtable, by Mona Anita K. Olsen, Ph.D. Against) Restaurant Tipping, by Michael Vision: Defining Boundaries in the Food and Lynn, Ph.D. Beverage Entrepreneurship Development Vol. 16 No. 24 Differential Evolution: A Cycle, by Mona Anita K. Olsen, Ph.D., and Tool for Global Optimization, by Andrey D. Vol. 16 No. 13 The Changing Relationship Cheryl Stanley Ukhov, Ph.D. between Supervisors and Subordinates: How Managing This Relationship Evolves Vol. 16 No. 3 Communication Planning: Vol. 16 No. 23 Short-term Trading in over Time, by Michael Sturman, Ph.D. and A Template for Organizational Change, by Long-term Funds: Implications for Financial Sanghee Park, Ph.D. Amy Newman Managers, by Pamela Moulton, Ph.D. Vol. 16 No. 12 Environmental Implications Vol. 16 No. 2 What Guests Really Think Vol. 16 No. 22 The Influence of Table of Hotel Growth in China: Integrating of Your Hotel: Text Analytics of Online Top Technology in Full-service Restaurants, Sustainability with Hotel Development, Customer Reviews, by Hyun Jeong “Spring” by Alex M. Susskind, Ph.D., and Benjamin by Gert Noordzy, Eric Ricaurte, Georgette Han, Ph.D., Shawn Mankad, Ph.D., Nagesh Curry, ,Ph.D. James, and Meng Wu Gavirneni, Ph.D., and Rohit Verma, Ph.D. Vol. 16 No. 21 FRESH: A Food-service Vol. 16 No. 11 The International Hotel Vol. 16 No. 1 The Role of Service Sustainability Rating for Hospitality Sector Management Agreement: Origins, Evolution, Improvisation in Improving Hotel Customer Events, by Sanaa I. Pirani, Ph.D., Hassan and Status, by Michael Evanoff Satisfaction, by Enrico Secchi, Ph.D., Aleda A. Arafat, Ph.D., and Gary M. Thompson, Roth, Ph.D., and Rohit Verma, Ph.D. Ph.D. Vol. 16 No. 10 Performance Impact of Socially Engaging with Consumers, by Chris CREF Cornell Hotel Indices Vol. 16 No. 20 Instructions for the Anderson, Ph.D., and Saram Han Early Bird & Night Owl Evaluation Tool Vol. 5 No. 4 Third Quarter 2016: Hotels (EBNOET) v2015, by Gary M. Thompson, Exhibit Positive Momentum, by Crocker Liu, Vol. 16 No. 9 Fitting Restaurant Service Ph.D. Ph.D., Adam D. Novak, Ph.D., and Robert Style to Brand Image for Greater Customer M. White, Jr.Satisfaction, by Michael Giebelhausen, Vol. 16 No. 19 Experimental Evidence Ph.D., Evelyn Chan, and Nancy J. Sirianni, that Retaliation Claims Are Unlike Other Vol. 5 No. 3 Second Quarter 2016: Ph.D. Employment Discrimination Claims, by Slowdown for Large Hotels Continues: David Sherwyn, J.D., and Zev J. Eigen, J.D. Small Hotels Have Now Slowed as Well, by Vol. 16 No. 8 Revenue Management Crocker Liu, Ph.D., Adam D. Novak, Ph.D., in Restaurants: Unbundling Pricing for Vol. 16 No. 18 CIHLER Roundtable: and Robert M. White, Jr.Reservations from the Core Service, by Dealing with Shifting Labor Employment Sheryl Kimes, Ph.D., and Jochen Wirtz, Sands, by David Sherwyn, J.D. Ph.D. Cornell Hospitality Report • December 2016 • www.chr.cornell.edu • Vol. 16, No. 28 13 Advisory Board Cornell Hospitality Report Vol. 16, No. 23 (October 2016) © 2016 Cornell University. This report may not be Syed Mansoor Ahmad, Vice President, Global Business reproduced or distributed without the express permission Head for Energy Management Services, Wipro EcoEnergy of the publisher. Marco Benvenuti MMH ’05, Cofounder, Chief Analytics and Product Officer, Duetto Cornell Hospitality Report is produced for the benefit Scott Berman ’84, Principal, Real Estate Business Advisory of the hospitality industry by Services, Industry Leader, Hospitality & Leisure, PwC The Center for Hospitality Research at Cornell University. Erik Browning ’96, Vice President of Business Consulting, The Rainmaker Group Christopher K. Anderson, Director Bhanu Chopra, Founder and Chief Executive Officer, Carol Zhe, Program Manager RateGain Jay Wrolstad, Editor Susan Devine ’85, Senior Vice President–Strategic Glenn Withiam, Executive Editor Development, Preferred Hotels & Resorts Kate Walsh, Interim Dean, School of Hotel Administration Ed Evans ’74, MBA ’75, Executive Vice President & Chief Human Resources Officer, Four Seasons Hotels and Center for Hospitality Research Resorts Cornell University Kevin Fliess, Vice President of Product Marketing, CVENT, School of Hotel Administration Inc. 389 Statler Hall Chuck Floyd, P ’15, P ’18 Global President of Operations, Ithaca, NY 14853 Hyatt 607-254-4504 R.J. Friedlander, Founder and CEO, ReviewPro Gregg Gilman ILR ’85, Partner, Co-Chair, Labor & Employment Practices, Davis & Gilbert LLP Dario Gonzalez, Vice President—Enterprise Architecture, Carolyn D. Richmond ILR ’91, Partner, Hospitality Practice, DerbySoft Fox Rothschild LLP Linda Hatfield, Vice President, Knowledge Management, David Roberts ENG ’87, MS ENG ’88, Senior Vice President, IDeaS—SAS Consumer Insight and Revenue Strategy, Marriott International, Inc. Bob Highland, Head of Partnership Development, Barclaycard US Rakesh Sarna, Managing Director and CEO, Indian Hotels Company Ltd. Steve Hood, Senior Vice President of Research, STR Berry van Weelden, MMH ’08, Director, Reporting and Sanjeev Khanna, Vice President and Head of Business Unit, Analysis, priceline.com’s hotel group Tata Consultancy Services Adam Weissenberg ’85, Global Sector Leader Travel, Josh Lesnick ’87, Executive Vice President and Chief Hospitality, and Leisure, Deloitte Marketing Officer, Wyndham Hotel Group Rick Werber ’83, Senior Vice President, Engineering and Faith Marshall, Director, Business Development, NTT DATA Sustainability, Development, Design, and Construction, Host David Mei ’94, Vice President, Owner and Franchise Hotels & Resorts, Inc. Services, InterContinental Hotels Group Dexter Wood, Jr. ’87, Senior Vice President, Global Head— David Meltzer MMH ’96, Chief Commercial Officer, Sabre Business and Investment Analysis, Hilton Worldwide Hospitality Solutions Jon S. Wright, President and Chief Executive Officer, Access Nabil Ramadhan, Group Chief Real Estate & Asset Point Financial Management Officer, Jumeirah Group Umar Riaz, Managing Director—Hospitality, North American Lead, Accenture 14 The Center for Hospitality Research • Cornell University