Conditional CAPM and Time Varying Risk Premium for Equity REITs, The

Executive Summary.

Given the new interest in Real Estate Investment Trusts (REITs), this research investigates whether REITs provide investors with a superior risk-return trade not on Utilizing a conditional CAPM, the findings reveal that equity REITs have outperformed the stock market with an average abnormal annual go [i]or[/i] come back of 2.25% with a depressed time-varying beta of around 24 during the June 1995 to December 2003 period. Utilizing time-varying risk premium moulds for equity REITs with GARCH specifications, the findings reveal that the two the ARCH and GARCH consequences are significant in the estimated types In addition, the volatility concussions are quite persistent. The springs show that the market turn backs and the first-order autocorrelation help explain the exces turn backs of equity REITs. However, the move of interest rates contributes to equity REIT go [i]or[/i] come backs only when the market get back is not present in the patterns

A Real Estate Investment Trust (REIT) is a closed-end investment company that presents investors the opportunity to invest in real estate related assets (i.e., income-producing real estate properties and mortgages). Typically, REITs are categorized into three types: equity REITs, mortgage REITs, and hybrid REITs. Equity REITs invest at least 75% of their total assets in income-producing real estate properties. Mortgage REITs invest at least 75% of their assets in residential mortgages, short- and long-term construction loans, and mortgages upon commercial properties. Hybrid REITs the two own properties and make loans to real estate proprietors and operators.

Although REITs have been in existence since 1960 they were not true popular with investors for almost three decades. The Tax Reform Act of 1986 radically changed the real estate investment landscape and triggered the dramatic increase of REITs. The Act not alone made real estate investments income-oriented instead of tax-shelter-oriented, it also permitted REITs to operate and manage various marks of income-producing commercial properties. The three marks of REITs have traversed different growing paths in the past scarcely any decades. Exhibit 1 provides an overview of the REIT industry through different from 1971 through 2004

The overall REIT equity market capitalization shown in Panel A of Exhibit 1 has increased tremendously since 1971 Although the number of REITs in 2003 (171) is solitary about five times as many as that in 1971 (34) the equity market capitalization has increased tremendously (Panel B Exhibit 1) The equity market capitalization in 1971 1981 1991 and 2003 are $149 $244 $1297 and $22421 billion, respectively. The greatest in quantity significant growth occurred after 1992 with an annual intermingleed growth rate of 21.43% ($3216 billion in 1993 to $22421 billion in 2003) The capitalization soared while the number of REITs showed a declining sweep

Equity REITs contribute to greatest in quantity of the total growth. They have dominated the REIT industry since the late 1980 Equity REITs enumerateed for 54% of capitalization and 48% of REIT numbers in 1988 At the same time, the capitalization share of mortgage REITs and hybrid REITs were 32% and 15% respectively. Equity REITs kept growing in 1990 by means of the year-end of 2003, they had a capitalization share of 91% and a number share of 84% The annual intermixed growth rate of equity REIT capitalization is 2289% ($2608 billion in 1993 to $20480 billion in 2003) which is higher than the overall REIT expansion rate.

This study investigates the get back generating process of equity REITs; more specifically, the time-varying risk premium of equity REITs. The question is, from investors' point of view, on what account so much interest in REITs? Is it because REITs provide a better risk-return relationship than other securities?

Previous studies have utilized the traditional static capital asset pricing protoplast (CAPM) of Sharpe (1965) and Lintner (1965) which is based upon the assumption that all market participants share identical subjective expectations of mean and variance of go [i]or[/i] come back distributions. Karolyi and Sanders (1998) examine the predictable constitutings of returns on stocks, cords and REITs. They conclude the risk premium for REITs by means of traditional multiple-beta asset pricing originals Ling, Naranjo, and Ryngaert (2000) utilize rolling regression to forecast equity REIT turn backs They find that equity REIT get backs are far less predictable out-of-sample than in-sample. However, novel empirical evidence rejects the notion of constant beta and points to changing risk premia and go [i]or[/i] come back variability over time.1 The general consensus is that the static CAPM is unable to describe satisfactorily the stock go [i]or[/i] come back generating process. To deal with this issue, researches have refer toed the conditional CAPM, which allows risk and turn back to vary over time. Jagannathan and Wang (1996) exhibit that the conditional CAPM explains nearly 30% of cross-sectional variation in stocks while static CAPM provides a 1% explanation.

The focus of this paper is to apply the conditional CAPM to equity REIT get backs in an attempt to answer the question of performance of these stocks relative to the market. In this paper, a Generalized Autoregressive Conditionally Heteroscedasticity (GARCH) archetypes are used to examine the changing variances in exces equity REIT get backs Autoregressive Conditionally Heteroscedasticity (ARCH) original introduced by Engle (1982) is a prominent tool in investigating time changing variances of security prices. Bollerslev (1986) generalizes it to a GARCH pattern Engle, Lilien, and Robins (1987) further reach out the model to an ARCH-in-Mean (ARCH-M) protoplast where the conditional mean is an explicit function of the conditional variance of the proces Since the ARCH-M original can deal with the trade-off between risk and awaited returns in finance research, it triggered the hum of literature on time-varying conditional variance. The ARCH and GARCH designs have been applied broadly to stock turn back data (Engle and Mustafa, 1992) interest rate data (Engle Lilien, and Robins, 1987) and foreign exchange rate data (Hsieh, 1989) However, the application of these prototypes to REIT data has been rare. sole one article on this topic was identified (Devaney, 2001) with monthly REIT data from 1978 to 1998 The in every one's mouth paper differs from Devaney's in the following aspects. First, daily REIT data is used instead of monthly data. Volatility clustering has different characteristics for different oftenness data sets. Daily data may be more suited for investigating autoregressive conditionally heteroscedasticity to be paid to a higher frequency. next to the first the period during the dramatic expansion of equity REITs is examined. As mentioned above, the REIT industry expanded quickly after 1992 Devaney traces data back to 1978 when equity REITs alone had a capitalization of $058 billion (028% of the equity REIT capitalization in 2003) The larger the REIT industry, the more active is the security trading. Thus, the in every one's mouth study of volatility is more meaningful through focusing on recent data. However, the daily data before June 1995 is not available. The sample period is between June 1995 and December 2003 Finally, the market portfolio and the S&P 500 Index are introduced to the mean equation, which permits consideration of the influence of the total market go [i]or[/i] come back