Thesis
Reference
Behavior and predictability of securitized real estate returns
SERRANO MORENO, Camilo
Abstract
Consisting of three essays, the dissertation is motivated by some existing gaps in the financial real estate literature. First, we exploit the fact that no research has used the well documented hybrid nature of real estate securities to forecast its returns. Second, we propose the first 'true' comparison of the predictability of securitized real estate and stock returns, as we argue that the previous comparisons done in the literature are biased. Third, we examine the long-run dynamics that govern the relationships between securitized real estate and the three models most commonly used to explain and forecast its returns. Our main findings suggest that the hybrid nature of real estate securities can be used to predict its returns, that securitized real estate returns are more predictable than stock returns in countries with well established REIT regimes, and that securitized real estate is fractionally cointegrated with both macroeconomic variables and the hybrid factors.
SERRANO MORENO, Camilo. Behavior and predictability of securitized real estate returns. Thèse de doctorat : Univ. Genève, 2009, no. SES 704
URN : urn:nbn:ch:unige-45728
DOI : 10.13097/archive-ouverte/unige:4572
Available at:
http://archive-ouverte.unige.ch/unige:4572
Disclaimer: layout of this document may differ from the published version.
1 / 1
Securitized Real Estate Returns
Th` ese pr´ esent´ ee ` a la Facult´ e des Sciences ´ Economiques et Sociales de l’Universit´ e de Gen` eve
Par Camilo SERRANO MORENO
pour l’obtention du grade de
Docteur ` es sciences ´ economiques et sociales Mention finance
Th` ese N
◦704 Gen` eve, Novembre 2009
Membres du jury de th`ese
Prof. Dr. Pierre-Andr´e DUMONT, Universit´e de Gen`eve Prof. Dr. Piet EICHHOLTZ, Maastricht University
Prof. Dr. Martin HOESLI, Universit´e de Gen`eve, Directeur de th`ese
Prof. Dr. Jaya KRISHNAKUMAR, Universit´e de Gen`eve, Pr´esidente du jury
qui s’y trouvent ´enonc´ees et qui n’engagent que la responsabilit´e de leur auteur.
Gen`eve, le 3 novembre 2009
Le doyen Bernard MORARD
Impression d’apr`es le manuscrit de l’auteur
It would be impossible to overstate my gratitude to my Ph.D. advisor, Professor Dr. Mar- tin Hoesli. Working with him for the last four years has been a privilege and a capital experience for my development and my career. I could hardly imagine anyone doing their Ph.D. in better conditions than I did with him. A real mentor, always available, with great ideas, sound advice, and always in a good mood. Infinite thanks!
I would also like to thank Professor Dr. Pierre-Andr´e Dumont for his valuable remarks and for participating in the thesis jury, Professor Dr. Jaya Krishnakumar for presiding this jury, and Professor Dr. Piet Eichholtz for participating as the external member of the jury.
Four papers have emerged from this dissertation. I am grateful to the anonymous referees from the Journal of Real Estate Finance and Economics and the Journal of Real Estate Portfolio Management who have reviewed these papers and provided helpful comments.
Additionally, I wish to thank the participants of the ARES 2007 meeting in San Francisco, CA, the AREUEA 2008 meeting in New Orleans, LA, the ERES 2008 conference in Krakow, Poland, the 2008 Real Estate Research Symposium in Rotterdam, the Netherlands, the ARES 2009 meeting in Monterey, CA, and the ERES 2009 conference in Stockholm, Swe- den. In particular, Professor Dr. Graeme Newell and Professor Dr. Dirk Brounen have provided insightful remarks.
I am also indebted to the great group of Ph.D. candidates and friends at the Business School (HEC) of the University of Geneva for providing a stimulating and fun environ- ment to work in. I am especially grateful to Georgios Gatopoulos, Charlotte Beauchamp, Dejan Munjin, Thibaut Bardon, Milana Finyutina, Sandro Lorini, Jussi Rouhento, and Ramona Westermann. Also to Dr. Elias Oikarinen our visiting Post-Doc and friend with whom I have some research projects in the pipeline.
To those who have indirectly contributed to this thesis by sharing their lives and experi- ences with me, I express my deepest gratitude. Namely, my girlfriend Annina Odermatt and my brothers Andr´es P´erez, Lucas Echavarr´ıa, Pablo Mej´ıa, Santiago Vel´asquez, and Eric Schober.
Finally, and most importantly, I wish to thank my dad and my mom for their unconditional love and support. I owe them everything, to them I dedicate this thesis.
1 Introduction 1
2 The Institutional Set-up 9
2.1 Overview of the Market: Direct and Indirect Real Estate . . . 9
2.2 REITs . . . 11
2.3 Market Size and Composition . . . 14
2.3.1 Market Size . . . 14
2.3.2 Market Composition . . . 17
2.4 Securitized Real Estate Benchmarks . . . 23
2.4.1 Global Securitized Real Estate Indices . . . 23
2.4.2 U.S. Securitized Real Estate Indices . . . 33
2.4.3 Return, Risk, and Correlations . . . 36
2.5 Why Invest in Real Estate Securities? . . . 36
2.5.1 Performance . . . 36
2.5.2 Dividends . . . 39
2.5.3 Diversification . . . 40
2.5.4 Liquidity . . . 42
2.6 Conclusion . . . 43
3 Forecasting EREIT Returns 47 3.1 Introduction . . . 48
3.2 Literature Review . . . 50
3.3 Data . . . 54
3.4 Methodology . . . 56
3.4.1 Models Employed . . . 57
3.4.2 Forecasting Techniques . . . 58
3.5 Empirical Results . . . 64
3.5.1 Models’ Results . . . 64
3.5.2 Forecasting Results . . . 67
3.5.3 Trading Strategies on the Forecasts . . . 70
3.6 Conclusion . . . 72 i
4 Predictability: Securitized Real Estate vs. Stocks 75
4.1 Introduction . . . 76
4.2 Literature Review . . . 78
4.3 Data . . . 83
4.4 Methodology . . . 85
4.4.1 Forecasting Techniques . . . 86
4.4.2 Predictability Comparisons . . . 89
4.5 Predictability Results . . . 93
4.5.1 Comparisons based on Prediction Errors and Excess Returns . . . . 93
4.5.2 Trading Strategy Comparisons . . . 99
4.6 Conclusion . . . 102
5 Fractional Cointegration Analysis 105 5.1 Introduction . . . 106
5.2 Literature Review . . . 108
5.3 Data . . . 111
5.4 Methodology . . . 113
5.4.1 Fractional Cointegration . . . 115
5.4.2 Fractionally Integrated Error Correction Model . . . 120
5.5 Empirical Results . . . 121
5.5.1 Fractional Cointegration Results . . . 121
5.5.2 Forecasting Results . . . 122
5.6 Conclusion . . . 128
6 Main Conclusions 131
Introduction
Real estate investments worldwide (excluding housing) are estimated to be worth $17 trillion (EPRA, 2008), while the size of stock markets is around $54 trillion and that of the bond market approximately $67 trillion as of December 2006 (SIFMA, 2007).1 If residential real estate is included in the estimation, the real estate market is thought to be as large as the stock market. The asset class thus constitutes a very significant share of worldwide wealth. Research in the area, however, is often hampered by data measurement issues.
Indeed, two main features make real estate valuation far more complicated than stock or bond valuation. First, the lack of a centralized real estate market reduces the transparency of this asset class as the access to information is limited and information asymmetries exist between market participants. Second, the high unitary value of real estate assets, as well as their heterogeneous characteristics, result in infrequent trading or illiquidity of the asset class.
Tracking real estate prices on real estate markets has been the subject of much research.
For commercial markets, appraisal-based indices are often used, but these have been shown
1The size of each market evolves significantly through time. Our figures date from December 2006 since a more recent estimate of real estate investments worldwide is not available. However, it is important to note that the current market turmoil has decreased, for instance, the size of stock markets to $40 trillion as of September 2008.
1
to suffer from smoothing and lagging (Geltner, 1993; Fisher, Geltner and Webb, 1994;
Hoesli, 2008). For residential markets, simple methods include using average or median house prices. However, such indices are not constant-quality which can render their usage problematic (Crone and Voith, 1992; Gatzlaff and Ling, 1994; Wang and Zorn, 1997).
Quality control is achieved either by means of the hedonic or repeat sales approaches (Case and Shiller, 1989; Thibodeau, 1989; Shiller, 1991; Clapp and Giaccotto, 1992; Englund, Quigley and Redfearn, 1998; Bourassa, Hoesli and Sun, 2006), but great care has to be exercised when implementing such models.
An initial attempt to overcome these problems was to consider the price evolution of indirect real estate as a proxy for real estate prices. Indirect real estate refers to funds that invest in real estate. Therefore, investors do not buy property directly, but acquire shares of a fund that holds a portfolio of property. These funds can be listed or non-listed. If such funds are listed, information is more transparent and the liquidity of the investments is increased. Most of these funds have a Real Estate Investment Trust (REIT) status that gives them special tax considerations comparable to those accorded to mutual funds.
Therefore, listed funds constitute homogenous, liquid vehicles whose value should follow the underlying asset. However, several studies (Ambrose, Ancel and Griffiths, 1992; Khoo, Hartzell and Hoesli, 1993; Ghosh, Miles and Sirmans, 1996; Brounen and Eicholtz, 2003) find that the behavior of real estate securities and real estate differs, in the short term anyway, and that real estate securities constitute an independent asset class.
Securitized real estate can be more accurately described as a hybrid asset, a hybrid of stocks, bonds and real estate. This results from the fact that it is publicly traded, that the generally long term fixed leases generate a fixed income, and that the underlying asset is real estate. The linkages between securitized real estate and the underlying asset were the first subject of study. Giliberto (1990), Gyourko and Keim (1992), and Mei and Lee (1994) find the presence of a common real estate factor linking the performance of securitized and
direct real estate. Shortly after, the relation of securitized real estate with financial assets is analyzed and the evidence found about REIT returns being correlated with both stock and bond returns is compelling (Peterson and Hsieh, 1997; Karolyi and Sanders, 1998;
Ling and Naranjo, 1999). Clayton and MacKinnon (2001, 2003) were the first to describe REITs as a hybrid asset and examine their linkage with stocks, bonds and real estate.
Recently, and at an international level, Hoesli and Serrano (2007) cover 16 countries and conclude that securitized real estate returns are generally positively associated with stock and direct real estate returns, but negatively related to bond returns.
Much research has also examined the time-varying nature of these linkages. Glascock, Lu and So (2000) find that before 1992, REITs were cointegrated with bonds and inflation, while after 1992 they were cointegrated with stocks and even more so with small caps.
The similarity of securitized real estate with small caps is also acknowledged by Clayton and MacKinnon (2003). They report that REITs went from being driven by the same economic factors as large caps in the 1970s and 1980s, to being more strongly related to both small caps and real estate related factors in the 1990s. Anderson, Clayton, MacKinnon and Sharma (2005) distinguish between value and growth small cap stocks and find that REITs behave more like small cap value stocks than like small cap growth stocks or large cap stocks.
Studying the behavior, the risk and return characteristics, the diversification benefits, the driving factors, and the predictability of publicly traded real estate has become of outmost importance due to the increased interest of investors in this asset class. Securitized real estate’s global market capitalization as measured by the GPR General Global Index has grown from $28 billion in 1984, to $234 billion in 1995, to $1.14 trillion in 2007. This index is based on investment companies, but if we consider the whole securitized real estate universe, the market capitalization is even larger. This tendency continues worldwide as more countries adopt REIT structures. REITs already exist in more than 30 countries and
several others are considering such legislation. The rising importance of this asset class in the investment universe constitutes an opportunity for investors to further diversify their portfolios and constitute more efficient risk and return generating patterns. A review of the financial economics literature on the environment, performance, and diversification potential of securitized real estate is provided in Corgel, McIntosh and Ott (1995), Glascock and Ghosh (2000), Worzala and Sirmans (2003), and Zietz, Sirmans and Friday (2003).
As for any other asset class, there are many important reasons for studying the return generating process of securitized real estate, such as to ascertain the diversification bene- fits of the asset class (Eichholtz, 1996; Mull and Soenen, 1997; Gordon, Canter and Webb, 1998; Conover, Friday and Sirmans, 2002; Hoesli, Lekander and Witkiewicz, 2004), to iden- tify the intra-asset diversification potential of investing across property types or geographic regions (Gyourko and Nelling, 1996; Boer, Brounen and Op ’t Veld, 2005; Glascock and Kelly, 2007), to analyze the inflation-hedging effectiveness of such vehicles (Gyourko and Linneman, 1988; Yobaccio, Rubens and Ketcham, 1995; Liu, Hartzell and Hoesli, 1997;
Chatrath and Liang, 1998; Adrangi, Chatrath and Raffiee, 2004; Hoesli, Lizieri and Mac- Gregor, 2008), and to examine whether returns can be forecasted (Bharati and Gupta, 1992; Liu and Mei, 1992; Mei and Liu, 1994; Brooks and Tsolacos, 2001, 2003). This dissertation falls into the latter category. In sum, its main objectives are to examine and model the behavior of securitized real estate returns and then build upon these results to establish which models and forecasting techniques are more useful for predicting purposes.
Our interest in forecasting securitized real estate returns is motivated by the follow- ing reasons. Most importantly, returns forecasting is the cornerstone of any investment strategy. Establishing a clear, methodological, and objective picture of the possible fu- ture behavior of returns is of outmost importance for any portfolio manager. It is in this spirit that this dissertation is motivated by some existing gaps in the financial real estate literature. First, we exploit the fact that no research has used the well documented hy-
brid nature of real estate securities to forecast their returns. Second, we propose what we believe is the first ’true’ comparison of the predictability of securitized real estate and stock returns, as we argue that the framework used in the previous comparisons done in the literature may bias the results. Third and last, we build upon the main contribution of our first essay (i.e., that the hybrid nature of real estate securities, can in fact, be used to predict their returns) to analyze whether financial and real estate factors are better predictors than economic variables for securitized real estate return forecasting.
Predictability is tightly linked to market efficiency.2 That is, if markets are efficient, no information or analysis can be expected to result in outperformance of an appropriate benchmark. This means that if markets are efficient, the usefulness of forecasting returns might be limited as it becomes more difficult to profit from the forecasts. The evidence concerning the efficiency of the securitized real estate market is mixed. Kleiman, Payne and Sahu (2002) find that the market is efficient, while Kuhle and Alvayay (2000) conclude that it is not. Jirasakuldech and Knight (2005) on the other hand, find signs that the efficiency of the market is increasing. This dissertation does not focus on formally testing for market efficiency by means of autocorrelation tests, variance ratio tests, non-parametric runs tests, or other formal tests. Rather, we address this issue indirectly by examining if abnormal returns may be achieved through forecasting. Indirect evidence on market efficiency is provided by Brooks and Tsolacos (2001) and Nelling and Gyourko (1998) who support the weak-form market efficiency hypothesis by examining the predictability of EREIT returns and finding no evidence of unexploited arbitrage opportunities once transaction costs have been taken into account. Contrarily, Cooper, Downs and Patterson (2000) examine the predictability of REIT returns for evidence of information-based trading and their results contradict the strong-form market efficiency hypothesis. Overall, the results on market efficiency are mixed, and hence predictability in returns could be expected.
2For a review on market efficiency see Fama (1970).
The dissertation is composed of four main chapters. A first chapter presenting the institutional set-up of international real estate securities is followed by three linked essays;
each representing a chapter of this dissertation. Each chapter includes the relevant review of the literature. The first essay analyzes diverse forecasting methodologies and forecasting variables. The second essay compares the predictability of securitized real estate returns to that of common stocks. The final essay compares the forecasting ability of financial and real estate factors to that of economic variables. The dissertation ends with some concluding remarks where the major contributions and results are highlighted.
The institutional set-up of international real estate securities is presented in Chapter 2. The description of the global securitized real estate market is important to understand the differences between REIT regimes across countries, as well as the nuances between the available indices. First, an overview of the market in terms of market size and investment types is presented. An in-depth examination of REITs follows in which specific attention is placed on the differences of these tax-transparent structures across countries. We continue by discussing the most used global benchmarks and highlight the particularities of these in- dices by reviewing their construction methodologies. Finally, we examine the performance of global real estate securities and compare it with the performance of other asset classes.
Chapter 3 analyzes the role played by financial assets, direct real estate, and the Fama and French factors in explaining EREIT returns and examines the usefulness of these variables in forecasting returns. Four models are analyzed and their predictive poten- tial is assessed by comparing three forecasting methods: time varying coefficient (TVC) regressions, vector autoregressive (VAR) systems, and neural networks models. Trading strategies on these forecasts are compared to a passive buy-and-hold strategy. Although securitized real estate has often been described as a hybrid of stocks, bonds, and real es- tate, no research has to date attempted to use this hybrid nature for predicting returns on this asset class. This chapter contributes to filling this void in the literature. The results
show that EREIT returns are better explained by models including the Fama and French factors. The VAR forecasts are better than the TVC forecasts, but the best predictions are obtained with neural networks and especially when they are applied to the model using stock, bond, real estate, size, and book-to-market factors.
In Chapter 4, we examine whether the predictability of securitized real estate returns differs from that of stock returns. It also provides a cross-country comparison of securitized real estate return predictability. In contrast to most of the literature on this issue, the analysis is not based on a multifactor asset pricing framework. Such analyses may bias the results as the differences or similarities in predictability may be due to differences in the explanatory power of the specified model for the two asset classes. By using a time series approach, the contribution of this chapter to the literature is thus to create a level playing field to compare the predictability of the two asset classes. Forecasts are performed with ARMA and ARMA-EGARCH models. Such forecasts are evaluated by comparing the entire empirical distributions of prediction errors, as well as with a trading strategy. The results suggest that the maturity of the securitized real estate market plays an important role in the predictability of its returns. That is, in countries with mature and well established REIT regimes, securitized real estate returns are found to be more predictable than stock returns.
Chapter 5 uses fractional cointegration analysis to examine whether long-run relations exist between securitized real estate returns and the three sets of variables most frequently used in the literature as the factors driving securitized real estate returns. That is, we examine whether such relationships are characterized by long memory (long-range depen- dence), short memory (short-range dependence), mean reversion (no long-run effects) or no mean reversion (no long-run equilibrium). The forecasting implications are also considered.
The characterization of the nonlinear linkages between securitized real estate and each of the three sets of variables is important as most of the research studying the explanatory
factors and the predictability of securitized real estate returns uses variations of these three models. The results show strong evidence of fractional cointegration between securitized real estate and the three sets of variables. Such relationships are mainly characterized by short memory although long memory is sometimes present. Therefore, the analysis shows that securitized real estate returns should be predictable to some extent.
A final chapter provides the main findings and contributions of this dissertation. A summary of the main results of the three essays and their implications for academics and the investment community are presented. Finally, we propose some topics that would warrant further research.
The Institutional Set-up of
International Real Estate Securities
A modified version of this chapter has been published as Global Securitized Real Estate Benchmarks and Performance. Journal of Real Estate Portfolio Management, 2009, 15(1), 1-19, with M. Hoesli.
2.1 Overview of the Market: Direct and Indirect Real Estate
Real estate investments may be performed directly or indirectly. Direct real estate refers to investments in real and tangible assets. It is an illiquid and heterogeneous asset class that is uncorrelated with financial assets (Eichholtz and Hartzell, 1996; Quan and Titman, 1997; Hoesli, Lekander and Witkiewicz, 2004) and has good inflation hedging character- istics (Fama and Schwert, 1977; Hartzell, Hekman and Miles, 1987; Hoesli, Lizieri and MacGregor, 2008). Such investments may fall into one of the following three categories:
core, value-added or opportunistic. Core funds are generally the least leveraged (between 9
0 and 30%) and present the lowest risk and return potential. These are investments in stable income-producing assets with expected returns of 8-10%. Value-added funds are more aggressive and risk tolerant, with debt levels around 50-65%. Expected returns are around 11-16%. Investors seek a prospective growth in rents or possible transformations or renovations of the assets. Opportunistic funds have the highest levels of debt (over 70%), risk, and expected return. These include construction and development activities, as well as investments in depressed or emerging markets.
Indirect real estate refers to pooled investments in real estate. These pooled investment vehicles can be non-listed or listed. Non-listed vehicles offer better diversification opportu- nities than the listed ones, but they are less liquid and almost exclusively for institutional investors (Hoesli and Lekander, 2008). Listed securitized real estate vehicles are the focus of this dissertation. They constitute homogeneous, liquid, and diversified investments with low transaction costs whose value should follow the direct real estate market, in the long- run anyway, although stock market fluctuations may have an impact in the short-term.
Such securities are referred to as Real Estate Operating Companies (REOCs) in the U.S.
and as property companies in the U.K. when they are listed as any other company would be, and as Real Estate Investment Trusts (REITs) when they meet certain conditions to gain tax-transparency.
The chapter is structured as follows. Since most real estate securities worldwide benefit from REIT status, we first explain what REITs are and how they differ across countries.
Then, a discussion of the size, growth, and composition of global securitized real estate markets follows. Third, we compare the construction methodology, as well as the returns and risk, of the most commonly used global securitized real estate benchmarks, and an- alyzes the degree of correlation between the indices. An analysis of the performance of global real estate securities follows, and the chapter ends with some concluding remarks.
2.2 REITs
Real Estate Investment Trusts (REITs) are companies that own, finance, or operate income-producing real estate. REITs are tax-transparent instruments that trade on ex- changes just as any other stock. Tax-transparency refers to the fact that they are not taxed at a corporate level provided the income be distributed to the shareholders. This important characteristic puts REIT shareholders in a comparable position with an investor who owns a direct property portfolio. REITs were initially introduced in the U.S. in 1960 and they were classified in three categories: Equity REITs, Mortgage REITs, and Hybrid REITs. Equity REITs (EREITs) own and operate income-producing real estate, Mortgage REITs (MREITs) finance real estate investments either by lending money directly to real estate owners and operators or by acquiring loans or mortgage-backed securities, and Hy- brid REITs (HREITs) are companies that combine both activities. REIT status can be reached by meeting a number of organizational, operational, distribution, and compliance requirements. The main requirements in the United States are as follows: (1) a REIT must have at least 100 different shareholders without having 5 or fewer shareholders own more than 50% of the stocks, (2) at least 75% of a REITs assets and of its gross income must be derived from real estate assets such as real property or mortgages on real property, (3) a REIT must distribute in dividends at least 90% of its taxable yearly income, and (4) to qualify as a REIT, companies must apply for REIT eligibility.
An interesting structure that appeared in 1992 to significantly reduce the cost of going public for private real estate companies was the Umbrella Partnership REIT (UPREIT).
Basically, an UPREIT is a structure in which private real estate companies exchange property assets for partnership interests in a REIT without generating capital gains tax on the properties transferred. In an UPREIT, the REIT does not own the properties themselves, but a controlling interest in a limited partnership that owns the properties.
As a result of this tax efficient manner of transferring property from a private real estate company to a public one, the REIT market went through an (Initial Public Offering) IPO boom in 1993-1994 (Geltner, Miller, Clayton and Eichholtz, 2007) that contributed greatly to the growth and consolidation of this market.
From an international perspective, the global property industry has been transformed in the last decade by the proliferation of REIT legislation in an increasing number of coun- tries. The U.S. REIT served as a model for these legislations, but each country created their own set of rules. In a general manner, REITs worldwide have to comply with certain operational restrictions, financing limitations, and shareholder requirements. Operational restrictions refer to regulations concerning the type of assets that can be invested in, the possibility of engaging in property development activities, diversification requirements, and portfolio trading activities. Financing limitations make reference to the maximum autho- rized levels of debt in the capital structure of the real estate companies. Lastly, shareholder requirements refer to the minimum number of investors required, the maximum percentage of ownership allowed for a single investor, and whether listing is mandatory. Albeit these requirements are set to ensure the stability of the real estate investment companies and to protect the shareholders, substantial differences exist in REIT-like structures across coun- tries. Tax-transparent legislation does not ensure that all countries treat listed real estate companies in the same way. In fact, tax-transparency is more or less the only thing these regimes have in common.
A detailed description of the existing REIT regimes worldwide is provided in EPRA (2007). Overall, operational restrictions are relatively homogeneous across countries con- cerning the percentage of equity to be invested in real estate, generally standing at around 75%. However, regarding the type of assets, development activities and investments in real estate securities, there is no consensus. In 22 of the 31 countries where there is a REIT status, at least 80% of income must be distributed as dividends. As for leverage, it
is constrained in 25 of the countries examined. It ranges from 20% of income generating assets in Bulgaria, to twice the equity in South Korea. The minimum number of investors and ownership limits vary greatly across countries. The highest minimum initial capital requirements are generally observed in Europe (40 million in Italy, 29 million in Greece, and 15 million in France and Germany). Twelve countries including the U.S. and Australia have no such requirements. Finally, in 17 of the 31 countries, listing is mandatory.
Eichholtz and Kok (2007) argue that the different REIT-regimes hinder geographical diversification, hampering the optimal diversification of property investors and thereby in- creasing systematic risk. Small countries being the most affected, this could explain why the REIT market in the United States is more important than the REIT market of all European countries combined. The United States is the largest property share market operating under one institutional regime, whereas the European Union (EU) has 13 coun- tries with 13 different REIT-legislations and 14 countries without tax-transparent regimes.
Being able to invest across regions, United States REITs are mainly sector specialists.
Specialization in property market niches not only increases the property sectors in which the capital market can invest in, but it has also been documented to improve performance through economies of scale (Eichholtz, Hoesli, MacGregor and Nanthakumaran, 1995). On the other hand, European REITs are usually constrained to invest in a single country, limiting the scale and forcing to invest in several property types. The major obstacles for cross-border investments in REITs concern the different treatment that is given in various countries to private and institutional investors, as well as to domestic and foreign investors with respect to withholding tax, situation which leads to an uneven playing field for both property companies and investors.
With the goal of creating the largest and most efficient property capital market in the world, Eichholtz and Kok (2007) propose the creation of the EU REIT and provide a blueprint with its key components. They suggest that the EU REIT should not have
any restrictions regarding operations or assets. Therefore, developing activities should be allowed but taxed, as to have the same treatment as “pure” developers. Regarding the capital structure, they propose that no debt limits should be imposed as the decisions on the capital structure of a company should be dictated by its managers and not the government.
For the same reason, public listing of the shares should not be mandatory but a close-ended structure that is internally managed should be imposed as recommended by the following studies (Cannon and Vogt, 1995; Capozza and Seguin, 2000; Ambrose and Linneman, 2001). It is hard to define the optimal REIT legislation, but an EU REIT would increase transparency, promote cross-border diversification, and allow for sector specialization and a more efficient allocation of capital. In terms of tax collection, governments would not be renouncing to considerable revenues because investors have already been eluding double taxation by making use of tax havens.
2.3 Market Size and Composition
This section contains a discussion of the size and growth of global real estate securities markets (2.3.1). We also analyze the composition of international securitized markets in terms of geographical areas, tax-transparency status and investment focus (2.3.2).
2.3.1 Market Size
The last decade has seen a proliferation of real estate securities around the world. By the end of 2007, REITs had been introduced in 31 countries while several others were considering such legislation. Figure 2.1 depicts the growth of global real estate securities in terms of market capitalization and number of stocks from 1984-2007. Securitized real estate’s global market capitalization as measured by the GPR General Global Index has grown from $28 billion in 1984, to $234 billion in 1995, to $1.14 trillion in 2007. This
Figure 2.1: Growth of the Global Securitized Real Estate Market, 1984-2007
Source: Global Property Research.
index is based solely on investment companies but if we consider the whole securitized real estate universe (i.e., if developers and construction companies are included), the market capitalization has grown to $1.5 trillion in 2006.1
During the last decade, the number of stocks has not increased as vigorously as has the market capitalization, hence, evidencing a change in average company size. Indeed, the average company size grew from $206 million in 1984 to $2.4 billion in 2007 (Figure 2.2). This means that in a 20-year period, real estate securities went from being small caps to mid caps. Although the classification into small or mid caps can vary among brokers, small caps are generally defined as companies with a market capitalization of between $200 million and $2 billion, and mid caps as companies with a market capitalization of between
$2 billion and $10 billion.
1To our knowledge, this is the most recent evaluation of the whole securitized real estate universe.
Figure 2.2: Average Company Size Per Year
Source: Global Property Research and author’s calculations.
This growth has likely been driven by real estate appreciations, as well as by an increas- ing demand for real estate securities during the period. These last two factors, which are interrelated, are likely to exert opposite effects in the future, at least in the largest market which is that of the U.S. On the one hand, real estate prices in the U.S. have been decreas- ing since 2007, but demand for real estate securities is likely to continue increasing. Figure 2.3 shows that pension funds’ strong demand for real estate assets should continue. Their target and current allocations to real estate vary across countries, but a clear intention of increasing their exposure to real estate in most markets is evident. Some new investments might be allocated directly, but the bulk of them, especially those done abroad, are likely to be done by means of securitization. The recent adoption of REIT legislation in many countries is certainly a result too of pressures from the demand side.
Figure 2.3: Pension Fund Allocations to Real Estate
Source: Topintzi, Chin, and Hobbs (2008).
2.3.2 Market Composition
We consider various dimensions when analyzing the composition of global securitized real estate markets. Such dimensions include regional and country weights, as well as the per- centage of REITs, investment focus, and sector breakdown in each region. To illustrate how the conclusions drawn depend on the chosen benchmark, we compare a geographical breakdown done with the GPR General Index to one done with the FTSE EPRA/NAREIT Global Real Estate Index. The regional breakdown of global real estate securities is pre- sented in Figure 2.4. According to the GPR General Index, the Americas have the biggest share with 42%, followed by Europe with 31%, and Asia with 27%. With the FTSE EPRA/NAREIT Index, the Americas represent a similar share of the global market (39%) but Asia is significantly larger (42%) and Europe merely represents 19% of the world mar- ket. The discrepancies result from the fact that the GPR General Index includes developing
activities in all the continents, while the FTSE EPRA/NAREIT Index only includes them in Asia.
Figure 2.4: Global Securitized Real Estate Regional Breakdown
Source: Global Property Research and FTSE EPRA/NAREIT.
The importance of individual countries in the global market also depends on the bench- mark being used. Figure 2.5 shows that the largest markets with the GPR General Index
are those of the U.S. (39%), Germany (12%), Hong Kong (10%), the U.K. (8%), and Aus- tralia (7%). This ranking changes significantly if the breakdown is performed with the FTSE EPRA/NAREIT Index. With such a benchmark, the largest markets include the U.S. (36%), Hong Kong (14%), Australia (13%), Japan (13%), and the U.K. (8%). As explained above, the increase in share of the Asian countries is due to the exclusion of development companies in the Americas and Europe with the latter index. As for the contrasting size of the German market, it is due to the inclusion of bank-funds (German open-ended funds) in the GPR General Index and not in the FTSE EPRA/NAREIT Index.
For the tax-transparency status, investment focus, and sector breakdowns, we use the FTSE EPRA/NAREIT Global Real Estate Index due to data availability issues. As with the regional breakdown, the interpretations should be put into perspective by fully under- standing the type of constituents that compose the index (this issue is addressed thoroughly in the following section). The percentage of REITs and Non-REITs varies greatly across continents and also countries. As shown in Figure 2.6, REITs represent 95% of the securi- tized real estate market in North America, 66% in Europe, and only 40% in Asia. Investing in REITs is not exactly the same as investing in Non-REITs. For starters, REITs dispose of a tax waiver at the corporate level that Non-REITs do not have, and secondly, REIT returns have been found to be more predictable than Non-REIT returns (Serrano and Hoesli, Forthcoming). The latter result is likely a consequence of the tax status giving REITs more real estate-like features. An example of this would be the stable cash flows that direct real estate and REITs have in common.
Another factor differentiating global real estate securities is their investment focus. A regional breakdown by investment focus is depicted in Figure 2.7. In North America and Europe, approximately 90% of securitized real estate investments are focused on rental investments, i.e., income producing real estate, while in Asia, rental investments amount to only 39%. In fact, a major factor differencing Asian markets is that they are mainly
Figure 2.5: Global Securitized Real Estate Geographic Breakdown
Source: Global Property Research and FTSE EPRA/NAREIT.
Figure 2.6: REIT/Non-REIT Breakdown by Region
Source: European Public Real Estate Association.
Figure 2.7: Investment Focus Breakdown by Region
Source: European Public Real Estate Association.
dominated by property developers (Newell and Chau, 1996; Liow, 1997). Non-rental activ- ities such as construction and development are riskier than rental investments; therefore, it is not surprising that real estate securities in Asia are more volatile.
Finally, real estate securities in the various regions invest differently across sectors (Fig- ure 2.8). The reason for this could be related to the aforementioned differences in legislation across countries. In North America, real estate securities tend to concentrate their assets in a specific sector as they can obtain economies of scale by diversifying geographically within the country with the same REIT legislation. Therefore, all the sectors are relatively well represented in the market as such economies of scale will be sought for. In Europe and Asia, the generally smaller size of the countries and the differences in REIT legislation across countries make geographical diversification more difficult, so real estate securities generally diversify across sectors. As a result, 55% of real estate companies diversify across sectors in Europe and as much as 61% do it in Asia.
Figure 2.8: Sector Breakdown by Region
Source: European Public Real Estate Association.
2.4 Securitized Real Estate Benchmarks
A benchmark is a standard used for comparison. Indices are commonly used in finance to measure the performance of individual or aggregated assets. An index becomes a bench- mark when it is used as a reference point to quantify the relative performance of an asset or a portfolio of assets. Therefore, benchmarks are useful to evaluate the performance of active investment strategies or to summarize the performance of any given segment of the market. Such segmentation may be geographical, by asset class, industrial sector, size, or any other criteria. Thus, the importance of having a suitable benchmark that reflects the specificities of the market or portfolio to be analyzed.
Choosing an appropriate benchmark is not unproblematic for academics or practition- ers. Index construction methodologies vary from index to index as tradeoffs are made between the breadth of market coverage and the investability of the securities in the index.
This section is devoted to understanding the nuances between the major securitized real estate benchmarks. The principal indices used to measure the performance of securitized real estate globally are the FTSE EPRA/NAREIT Global Real Estate Index, the Global Property Research General Property Share Index (GPR General), the Global Property Re- search 250 Property Share Index (GPR 250), the S&P/Citigroup Global Property Index, and the Datastream Real Estate World Index.
2.4.1 Global Securitized Real Estate Indices
Table 2.1 provides a comparison of the main global securitized real estate benchmarks.2 With the exception of the GPR General Index that is available since 1984, the remaining indices analyzed exist since around 1990. Their market capitalization as of the end of 2007, ranges from $660 billion for the GPR 250 Index to $1,138 billion for the GPR General Index.
2Datastream’s Real Estate World Index is not included in the table due to limited information on its construction methodology.
The index covering the largest number of stocks (more than 500) and countries (more than 35) is the S&P/Citigroup World Property Index. All indices are free float adjusted and have size and liquidity screens, except for the GPR General Index that is only screened for size. Concerning the property sectors included in the indices, all the benchmarks include office, industrial, retail, residential, and diversified sectors. The GPR Indices also include hotels and the GPR 250 Index includes healthcare. The FTSE EPRA/NAREIT Global Real Estate Index and the S&P/Citigroup World Property Index cover a broader range of property sectors such as the storage, resorts, healthcare, and specialty sectors.
The criteria for inclusion of companies in the indices varies somewhat. First of all, the required real estate activity for inclusion in the index is a little less stringent for the S&P/Citigroup World Property Index as well as for the FTSE/EPRA NAREIT Asia Real Estate Index. Second, the minimum size of the companies included in the indices is lowest for the GPR General Index and highest for the FTSE EPRA/NAREIT Global Real Estate Index. Third, the liquidity constraints for the constituent companies are stronger for the GPR 250 Index and for the FTSE EPRA/NAREIT Global Real Estate Index. Overall, we conclude that the GPR General Index and the S&P/Citigroup World Property Index are more appropriate to examine the performance of the market as a whole.
That is because the former index has the largest market coverage in terms of market capitalization, and the latter in terms of number of countries and stocks covered. However, to evaluate the performance of active investment strategies, the GPR 250 Index and the FTSE EPRA/NAREIT Global Real Estate Index are better suited, as the investability of their constituents allows for the replication of such indices. To better understand the existing differences, we continue by describing the construction methodologies of each of these indices.
Table 2.1: Summary of Global Real Estate Securities Indices
Family FTSE Global Property Research S&P/Citigroup
EPRA/NAREIT Global Property
Available Indices Regional and GPR General, GPR General World Property,
Country, Quoted, GPR 250, GPR Emerging
Capped, REITs, and GPR 15 Property,
Dividend+, Global REIT,
Global Sectors, World REIT
Investment Focus, and REITs and Non-REITs
FTSE GPR General GPR 250 S&P/Citigroup
EPRA/NAREIT World
Global Real Property
Estate
Index Construction
Inception Date Dec.29.1989 Dec.31.1983 Dec.31.1989 Jul.30.1989 Base Date Dec.31.1999 Dec.31.1983 Dec.31.1989 Dec.31.1997
Frequency Daily Monthly Monthly Daily
(Daily since Dec.31.1999
Weighting Free Float Market Cap Free Float Free Float
Market Cap as of 794 1138 660 844
Dec.31.2007 ($ Billion)
Number of Stocks 313 405 250 >500
Number of Countries 22 27 25 >35
Investability Free Float Size screen Free Float Free Float
adjusted, size adjusted, size adjusted, size
and liquidity and liquidity and liquidity
screens screens screens
Foreign Exchange Rates WM/Reuters WM/Reuters WM/Reuters WM/Reuters
Closing Spot London Close London Close Mid-Market
Rates Rates Rates Fixings
Property Sectors Included in the Indices
Office √ √ √ √
Industrial √ √ √ √
Industrial/Office Mixed √
X X X
Retail √ √ √ √
Storage √
X X √
Residential √ √ √ √
Hotel X √ √
X
Hotel/Resorts and X X X √
Entertainment
Lodging/Resorts √
X X X
Healthcare √
X √ √
Specialty √
X X √
Diversified √ √ √ √
Table 2.1 (continued): Summary of Global Real Estate Securities Indices
FTSE GPR General GPR 250 S&P/Citigroup
EPRA/NAREIT World Property
Global Real Estate World Property
Inclusion Criteria
Real Estate >75% of EBITDA At least 75% of At least 75% of >60% of revenue Activity in NA and Europe, operational turnover operational turnover (includes
>60% in Asia derived from derived from development
(includes investment investment activities)
construction of activities or from activities or from
residential homes investment and investment and
for sale) development development
activities combined. activities combined
Both listed property (case in which 25%
companies and of operational
bank-funds are turnover must come
included from investment
activities)
Min. Market Free Float Mkt Cap Mkt Cap>$50M Free Float Mkt Cap Float-adjusted Mkt
Cap >0.10%, 0.10%, >$50M Cap>$100M
and 0.30% of their respective regional index for NA, Europe, and Asia, respectively
Liquidity Median daily trade No liquidity 250 most liquid Min. $25M traded in
per month must constraints property stocks last 12 months
exceed 0.05% of worldwide
shares for more than 10 of the 12 months prior to the review
IPOs Free Float Mkt Cap Free Float Mkt Cap Free Float Mkt Cap IPOs that rank
>0.15%, 0.20%, ranks among the ranks among the among the top 5
and 0.40% for NA, top 150 companies top 150 companies companies of a
Europe, and Asia, particular country
respectively
Attributed At least 75% of At least 60% of At least 60% of At least 60% of Sector gross invested operational turnover operational turnover revenue comes
assets come from comes from one comes from one from one specific
one specific sector specific sector specific sector sector
Attributed Country of listing Country of primary Country of primary Case-by-case basis Country where it is more stock listing. If more stock listing. If more depending on the
liquid. However, than 75% of than 75% of country of
residence for tax operational turnover operational turnover incorporation,
purposes and is derived from a is derived from a primary exchange,
investor protection different country, different country, liquidity, geographic
regulations, market company attributed company attributed source of revenues,
perception, to country where to country where geographic location
currency trading, assets are located assets are located of assets, and
and other factors headquarters of the
could be taken into company
account
Note: Datastream’s Real Estate World Index is not included in this table as its construction methodology is not available.
FTSE EPRA/NAREIT Global Real Estate Index
The FTSE EPRA/NAREIT Global Real Estate Index aims to represent general trends in listed real estate stocks worldwide. The index provides a complementary set of indices that include regional and country indices, Capped indices (i.e., with a 10% weighting cap placed on individual constituents), Dividend+ indices (i.e., higher yielding stocks; those with a one-year dividend yield forecast greater than 2%), Global Sectors, Investment Focus, and REIT and Non-REIT indices. Based on worldwide real estate stocks, it is a free float weighted index to ensure that only the investable opportunity set is included. Companies are eligible for inclusion in the index if they meet specific geographic financial standards that demonstrate that the majority of earnings or total assets are the result of relevant real estate activity, defined as the ownership, trading, or development of income-producing real estate. The universe covered by this index is limited to companies providing an audited annual report in English. For North America and Europe, companies must derive at least 75% of their EBITDA from relevant real estate activities in their respective region. For Asia, the threshold is 60% and relevant real estate activities also include the construction of residential homes for sale.
Non-constituents are considered for inclusion at a quarterly review if their investable market capitalizations are equal to or greater than their respective regional index by 0.10%, 0.10%, and 0.30% for North America, Europe, and Asia, respectively.3 Newly listed real estate companies (IPOs) are included in the index after their first day of trading only if their free float market capitalization is equal to or greater than 0.15%, 0.20%, and 0.40%
for North America, Europe, and Asia, respectively. In order to provide stability to the composition of constituents of the index, the required thresholds for inclusion are higher than those needed for exclusion. Liquidity, calculated as the median daily trading per month, is assessed on an annual basis to ensure that the index is tradable. That is, non-
3The respective amounts in millions of dollars as of the end of 2007 are: $302, $149, and $1,026.
constituents are eligible for inclusion if their median daily trading per month exceeds 0.05%
of their shares for more than 10 of the 12 months prior to the review.4 On the other hand, existing constituents failing to trade at least 0.04% of their shares in issue for more than 4 of the 12 months prior to the review will be dropped.
The FTSE EPRA/NAREIT Global Real Estate Index is available as a capital return, a total return, and a net total return index calculated at a daily frequency.5 It is calcu- lated using the chained Paasche methodology. The index has a base value of 1000 as of December 31, 1999. The inception date however is December 29, 1989. The stock prices used are the actual closing mid-market or last trade prices where available. The number of shares outstanding for each constituent security is amended only when the total shares outstanding held within the index system change by more than 1% on a cumulative basis.6 Free float, cross-holdings, and foreign ownership limits are adjusted on a quarterly basis.
Dividends paid by a company are reinvested in the index at the ex-dividend date. The main calculation currency is the Euro although other currencies are also employed. Foreign exchange rates used in end-of-day calculations are the WM/Reuters closing spot rates. For the indices disseminated in real-time, Reuters’ real time foreign exchange rates are used.
The property sectors included in the FTSE EPRA/NAREIT Global Real Estate Index are: healthcare, self storage, industrial, office, industrial/office mixed, residential, retail, lodging/resorts, specialty, and diversified. A REIT is assigned to a particular property sector when at least 75% of its gross invested assets are invested in one specific property type. The selection and base weighting of stocks and countries will be adjusted on a quarterly basis. The sum of the weighted individual constituents of each country constitutes
4Considering a mean company size of $2.4 billion as of the end of 2007, the median daily trading per month should exceed $1.2 million.
5The net total return index uses a net dividend calculated by deduction of withholding taxes (and other relevant taxes) at a rate applicable to non-resident individuals who do not benefit from double taxation treaties. The withholding tax rates are those applicable to Luxembourg holding companies.
6If accumulated changes in the number of shares outstanding are greater than or equal to 10%, or if they represent $2 billion of a company’s total market capitalization, the amendment takes place between quarters.
the country weights of the index. A company will normally be allocated to the country of listing where it is most liquid. However, other factors taken into account may include, but are not limited to, the investor protection regulations under which the company is governed, the country where the company is resident for tax purposes, market perception, and currency trading.
GPR General Property Shares Index
The aim of the GPR General Index is to represent the behavior of the securitized real estate market worldwide. It is a value-weighted index based on worldwide real estate stocks.
It includes listed property companies, as well as bank-funds.7 Companies are eligible for inclusion in the index if they derive at least 75% of their operational turnover from investment activities (property investment companies) or from investment and development activities combined (hybrid property companies). Additionally, the companies included in the index must have a market capitalization over $50 million for two consecutive months.
In case there is an initial public offering (IPO), the new company is included in the index after its first day of trading only if its free float market capitalization ranks among the top 150 companies.
The GPR General Index is a total return index calculated at a monthly frequency. Total returns are composed of price (capital) returns and dividend (income) returns. Dividends paid by a company are reinvested in the index at the ex-dividend date. The index has a base value of 100 at its base date (December 31, 1983). The stock prices used to calculate the index are the most recent end of month closing trade prices of the stock exchange of primary listing. Foreign exchange rates used are the WM/Reuters London close rates.
The property sectors included in the GPR General Index are: office, residential, retail, industrial, hotel, and diversified. A company is assigned to a particular property sector
7The GPR General Quoted Index is a sub-index of the GPR General Index in which the bank-funds are excluded.
when it derives at least 60% of its operational turnover from one specific property type.
The selection and base weighting of stocks and countries is adjusted on a monthly basis.
The country weights of the index are constituted by the sum of the weighted individual constituents of each country. However, if more than 75% of a company’s operational turnover is derived from a country different than the country of the company’s primary stock listing, the company is attributed to the index of the country where the assets are located.
GPR 250 Property Shares Index
The GPR 250 Index is a value-weighted index based on the 250 real estate stocks with highest average monthly trading volume in U.S. dollars over the previous year. The aim of the GPR 250 Index is to represent the behavior of the leading, most liquid securitized real estate companies worldwide.8 Companies eligible for inclusion in the index are those deriving at least 75% of their operational turnover from investment activities or from invest- ment and development activities combined, case in which, at least 25% of the operational turnover must come from investment activities. In addition, inclusion is only considered if the companies’ free float, calculated as the stock price times the available amount of shares outstanding, is greater than $50 million for two consecutive months. IPOs are included in the index after its first day of trading only if its free float market capitalization ranks among the top 150 companies.
The GPR 250 Index is a total return index calculated at a daily frequency since Decem- ber 31, 1999. Previous to this date it was calculated at a monthly frequency. The index has a base value of 100 at its base date that is December 31, 1989. The stock prices used are the most recent closing trade prices of the stock exchange of primary listing. Changes in free float are implemented on the first trading day of each month. Dividends paid by a
8The GPR 15 Real Time Index is a sub-index of the GPR 250 Index that reflects the most liquid property companies in Europe.
company are reinvested in the index at the ex-dividend date. Foreign exchange rates used are the WM/Reuters London close rates.
The property sectors included in the GPR 250 Index are: office, residential, retail, in- dustrial, hotel, healthcare, and diversified. A company is assigned to a particular property sector when it derives at least 60% of its operational turnover from one specific property type. The selection and base weighting of stocks and countries is adjusted on a quarterly basis. The country weights of the index are constituted by the sum of the weighted indi- vidual constituents of each country. However, if more than 75% of a company’s operational turnover is derived from a country different than the country of the company’s primary stock listing, the company is attributed to the index of the country where the assets are located.
S&P/Citigroup Global Property Index
The aim of the S&P/Citigroup Global Property Index is to reflect the risk and return char- acteristics of the investable universe of publicly traded property companies. It only includes equity property companies, where at least 75% of the revenue comes from equity property- related activities, and excludes mortgage and hybrid property companies. With over 500 constituents from more than 35 countries, the index provides four sub-indices covering developed, emerging, global REIT, and developed REIT markets. The S&P/Citigroup Global Broad Market Index (BMI) constitutes the universe of stocks from which the con- stituents of the S&P/Citigroup Global Property Index are drawn.9 Therefore, companies must be domiciled in one of the 53 countries covered by the BMI. Companies are eligible for inclusion in the index if they derive at least 60% of their revenue from real estate devel- opment, management, rental, and/or direct investment in physical property. Additionally, the companies’ float-adjusted market capitalization must be over $100 million and it must
9The BMI aims to include all the institutionally investable stocks in each country. This includes all equity share classes provided they meet market capitalization and liquidity requirements.
post a minimum value traded of $25 million for the 12 months preceding the annual re- constitution date. An index constituent whose float-adjusted market capitalization falls beyond $75 million or whose trading value is less than $15 million during the preceding 12 months will be excluded from the index at the annual reconstitution date. However, if a constituent’s float-adjusted market capitalization falls below $25 million, deletion from the index will occur after a five day notice period. IPOs that meet size requirements are added to the index at the next quarter rebalancing. If an IPO ranks among the top five companies of a particular country, it may be added sooner.
The S&P/Citigroup Global Property Index is available in price returns, total returns, net returns, and hedged indices at a daily frequency.10 It is calculated using a base-weighted aggregate methodology. The index has a base value of 100 at its base date that is December 31, 1997. However, the inception date is different to the base date, so this index starts on July 30, 1989.11 The stock prices used are closing prices. Shares are adjusted for corporate actions on their ex-dates. By adjusting for corporate actions such as dividends, splits, spin-offs, share issuance, etc., the index is able to represent stock market performance and not reflect the corporate actions of its constituents. The indices are available in six currencies (USD, GBP, JPY, CAD, EUR, and AUD) although other currency calculations are available upon request. Foreign exchange rates used in end-of-day calculations are the WM/Reuters mid-market fixings. The selection and base weighting of stocks are fully reconstituted on a yearly basis unless there is a 5% or greater change in shares, case in which the adjustment is made after a five day notice period.
The property sectors included in the S&P/Citigroup Global Property Index are: ho- tel/resort and entertainment, industrial property, office space, healthcare property, retail property, storage property, specialty, residential, and diversified. The sum of the weighted
10The net total return index uses a net dividend calculated by deduction of withholding taxes (and other relevant taxes) at a rate applicable to non-resident individuals who do not benefit from double taxation treaties. The withholding tax rates are those applicable to Luxembourg holding companies.
11The sub-indices may have different base and inception dates.
individual constituents of each country constitutes the country weights of the index. A company will be allocated to a country on a case-by-case basis including the country of incorporation, primary exchange, liquidity, geographic source of revenues, geographic loca- tion of assets, and headquarters of the company. The index is available on world, regional, and country levels. It may also be customized depending on the investor’s needs.
Datastream Real Estate World Index
Datastream’s Real Estate World Index aims to represent securitized real estate markets worldwide. The Thomson Datastream database constitutes the universe from which the index is drawn. Companies included in the index represent around 75-80% of the total market capitalization. Suitability for inclusion in the index is determined by market value and availability of the data. There are no liquidity requirements, as well as no adjustments for non-public holdings of shares or for cross-holdings. The index constituents are reviewed at a quarterly basis and re-set to represent the new top group of stocks by market value.
2.4.2 U.S. Securitized Real Estate Indices
Since the United States represents around 40% of the global securitized real estate market, we find it useful to briefly describe the most important benchmarks tracking this market.
As documented by Taylor (2005) in a study covering 119 real estate funds with investments in REITs of $86 billion, approximately 40% of the real estate funds are benchmarked to the Dow Jones Wilshire Real Estate Securities Index, 30% to the NAREIT Index, and 30%
to the Morgan Stanley REIT Index. In addition to these indices, we also present the well known S&P REIT Composite Index.
