|
Article Excerpt
It is well documented that REITs in the 1990s experienced significant changes in their structure and attracted greater institutional participation. This article finds that REIT stocks with higher institutional holdings perform better on Monday than REITs with lower institutional holdings during the 1990s, but not in the 1980s. Furthermore, REITs that went public in the 1990s are the ones associated with the shift in the Monday return pattern. Our study supports the claim that the change in REIT structure and the increase in institutional participation in the REIT market in the 1990s make REIT stocks behave more like other equities in the stock market.
**********
The most significant changes in the real estate investment trust (REIT) industry during the past four decades are probably the two structural changes that started in the late 1980s: the switching away from the use of external advisors and the change to a more operating-oriented company. This evolution started in 1986 with the tax law changes that allow REITs to use internal advisors. (1) However, the impact of those changes on the REIT market only became detectable in the early 1990s, at which time a critical mass of new REIT IPOs adopted the new REIT structure.
While the most important change in the 1990s was the switching away from the use of external advisors, REITs in the 1990s were also significantly different from their earlier peers in many aspects. For example, when compared to REITs in the 1980s, REITs in the 1990s were, on average, more liquid and larger in size (see, e.g., Clayton and McKinnon 2000, Beneveniste, Capozza and Seguin 2001, Chan, Erickson and Wang 2003) and more focused by property types in their investment portfolios (see, e.g., Capozza and Seguin 1999, 2001a). In addition, they had a significantly higher inside ownership (see Capozza and Seguin 2003) and used different capital structures and management strategies (see Capozza and Seguin 1998, 2001b). In other words, in the 1990s, REITs switched away from their original fund-like structure and had characteristics more like other operating firms traded in the stock market.
As we know, the original rules governing REIT operations were designed by Congress to attract small investors. Consequently, there were fewer institutional investors and financial analysts participating in the REIT market than in the general stock market. We also know that, in general, institutions prefer to invest in large, liquid, transparent companies with focused business lines and predictable earnings. It might be fair to say that the structural changes in the REIT market attracted the participation of institutional investors. (Alternatively, we can also argue that in order for REITs to sell IPO shares to institutional investors, they were forced to make the changes in order to meet the desires of institutional investors.) Chan, Leung and Wang (1998) point out that, while the institutional holdings level of REIT stocks is significantly lower than that of other stocks in the market before 1990, the level in the late 1990s is about the same as that of other stocks in the market.
A natural question to ask is, after the REIT structural change and the increase in institutional participation in the REIT market, whether REIT stocks behave more in line with other stocks. Are the structural changes in the REIT market correlated with a change in the stock performance pattern of REITs? Ling and Ryngaert (1997) are the first to point out clearly in the academic literature that the new REITs established after 1990 could perform differently from their earlier peers. They document two important changes in the REIT market during the 1990s. First, the average institutional holdings level of REIT IPO stocks in the 1990s was significantly increased from that before 1990. Second, REIT IPOs in the early 1990s were no longer overpriced on the first trading day, and IPO activities increased significantly after the structural change in the REIT market. These important results contrast with Wang, Chan and Gau's (1992) finding that REIT IPOs during the 1971-1988 period were overpriced and had low institutional participation. (It should be noted that IPOs of stocks in the general market are, on average, significantly underpriced.)
However, although Ling and Ryngaert (1997) point out that the average initial-day return of REIT IPOs is no longer negative in the 1990s, this does not mean that REIT IPOs now behave similarly to the IPOs of other stocks in the market. Indeed, an examination of Table 9.1 in Chan, Erickson and Wang (2003) indicates that the gap between the initial-day return of REIT IPOs and non-REIT IPOs was actually widening (not narrowing) in the 1990s. They report that the mean initial-day return of non-REIT IPOs was about 12.1%, 18.4% and 21.4% higher than that of REIT IPOs in the 1970-1979 period, 1980-1989 period and 1990-2000 period, respectively. Clearly, the change in REIT structure and the increase in institutional participation have not helped the performance of REIT IPOs. In addition, the surge in REIT IPO activities described by Ling and Ryngaert (1997) dissipated quite quickly; there were no REIT IPOs in 2000 and 2001. It should be noted that the last time we observed zero REIT IPO activity for longer than a 2-year period was in the 1974-1978 period. Given this, we have yet to see conclusive evidence indicating that the change in the REIT market and the increase in institutional investors make REIT stocks behave more similarly to other equities in the market.
This article examines whether REIT stocks behave more in line with non-REIT stocks after the structural changes in the 1990s by studying the behavior of their Monday returns before and after the changes. (2) Monday return is selected for the study because the negative Monday return phenomenon has been well documented in the finance literature. Since the finance literature indicates that institutional investors affect the Monday return and that the direct result of the change in REIT structure in the 1990s is the increase in institutional participation, we focus our attention on the impact institutional investors might have on the Monday returns of REITs before and after the structural changes. Our hypothesis is that the structural changes and the increase in institutional investors in the REIT market make REIT stocks behave more like other operating company stocks in the stock market.
Our finding supports our hypothesis. We find that the Monday return pattern of REITs is different from that of the general stock market before the structural change in the REIT market. That is, the institutional holdings level did not affect REIT Monday returns prior to 1990, but affected return patterns in the 1990s when the increase in institutional participation in REIT stocks occurred. We also find that the impact of the institutional holdings level on the Monday anomaly is significant only for REIT stocks that went public in the 1990s. This evidence gives more credence to the claim that the REITs that went public in the 1990s might be different from their earlier peers.
The next section briefly discusses the role of institutional investors and its relationship to the Monday anomaly. Following the discussion, we describe our sample selection and data collection procedure. The last two sections report our empirical results and present our conclusions.
Monday Seasonal and Institutional Investors
One of the most noticeable seasonal anomalies documented in the finance literature is the Monday anomaly. That is, the Monday stock return is significantly negative and is significantly lower than the average Tuesday to Friday return. Among the numerous explanations developed to explain this anomaly, the trading pattern of individual investors (or the lack of institutional participation) might be one of the most plausible explanations. (3)
Lakonishok and Maberly (1990) argue that individual investors typically do not have time during the weekday trading hours and therefore process information and make investment decisions only during the weekend. They support this view by documenting that, when the market reopens on Monday, individual investors tend to increase trading activity (especially sell transactions) on that day. The results from early empirical studies on the relationship between Monday return and institutional holdings are mixed. Sias and Starks (1995) find essentially no evidence of differences in Monday returns between NYSE stocks with high institutional ownership and stocks with low institutional ownership. However, Kamara (1997) reports that the Monday seasonal declined over the 1962-1993 period for stocks in the S & P 500 index and believes that the increased role of institutions in the stock market can explain this phenomenon. Recently, Chan, Leung and Wang (2005) provide direct evidence supporting Lakonishok and Maberly's argument. They find that NYSE-AMEX-NASDAQ stocks with higher institutional holdings are less affected by the Monday effect and this observation holds true for most years during the 1981-1998 period. (4)
Indeed, institutional investors have been shown to be important players in the stock market. First, they constantly gather pricing information about stocks in the market and probably act as price setters in the market. Second, once institutional investors hold a major chunk of stocks in a particular company, they will monitor the performance of the firm and its management team. It is well documented that in the REIT market there are certain advantages to following institutional investors' leads in the selection of stocks. For example, Wang et al. (1995) report that REIT stocks with low institutional holdings perform worse than other stocks on a risk-adjusted basis. Downs (1998) indicates that institutional ownership contributes to REIT value by documenting an industry-wide creation and distribution of value caused by the increase in institutional investors (due to the change in the 5/50 rule) in REIT stocks. Since there is strong empirical evidence that the level of institutional holdings affects Monday stock returns, an examination of the impact the change in institutional holdings in REIT stocks might have on their Monday returns seems to be warranted.
Sample Selection and Data Gathering Process
We use 1981 to 1999 as our sample period. This period is selected primarily because one of our main data sources only starts to report detailed information after 1980. To identify publicly traded REITs during this period, we rely primarily on NAREIT (National Association of Real Estate Investment Trusts) publications and the CRSP (Center for Research in Security Prices) tapes. NAREIT publications are used as the starting source for identifying the list of REITs. (5) However, because the format of the publications changes over time and there are several years for which the NAREIT publications do not report information on nonmember REITs, the list of REITs based solely on NAREIT publications may not be complete (especially in the earlier years). Given this, we also use the CRSP tapes as another means to identify REITs. (6)
A search of the CRSP tapes using the SIC codes for REITs yields a list of publicly traded firms that are candidates for possible inclusion in our REIT list. However, a careful examination of the firms classified as REITs by the CRSP tapes reveals that some of those firms are, in fact, not REITs. At the same time, there are also a significant number of firms that are classified as REITs by the CRSP tapes but are not listed in the NAREIT publications. In addition, some of the REITs listed in the NAREIT publications are not classified under the SIC codes for REITs by the CRSP tapes. To overcome the problems of incomplete information on REITs provided by NAREIT in certain years and the possible exclusion by NAREIT of some of the REITs listed in the CRSP database, we use additional sources such as the firms' Annual Reports, Standard and Poor's Stock Reports and the LEXIS-NEXIS database for further information. This process helps us verify the tax-qualified status of each questionable REIT for every year during the 1981-1999 period.
The firms that are included in the final REIT sample meet the following two criteria: They are confirmed to be a REIT and they have stock trading information in the CRSP tapes. A total of 435 REITs meet the criteria and have stock return data reported in the CRSP NYSE-AMEX-NASDAQ tapes at any given time during the years from 1981 to 1999. It should be noted that this process ensures that our REIT sample is free of survivorship bias.
We use Spectrum 3:13 (f) Institutional Stock Holdings Survey tapes (Spectrum) as the source for gathering information on the level of institutional holdings for the identified REITs in the sample period. (7) To do this, we first obtain from the CRSP tapes the permanent numbers, CUSIP numbers and names of all 435 REITs traded on the NYSE, AMEX and NASDAQ during the 1981-1999 period. We then obtain the names and CUSIP numbers of firms in the Spectrum tape during the same period. Finally, we combine data from the Spectrum tapes with data in the CRSP tapes using the eight-digit CUSIP number to create the final sample for this study.
From this file, we are able to calculate the percentage of institutional holdings of each REIT stock in each quarter during the 1981-1999 period. Since the Spectrum tape reports the number of shares outstanding figure rounded to the nearest one million,...
|
