Initial Public Offerings (IPOs) of equity securities often exhibit underpricing; common stock values tend to rise significantly from the offer price on the first day of trading. The evidence for Real Estate Investment Trust (REIT) IPOs is less consistent. Early studies found evidence of overpricing; a 1990s study found underpricing. Several theories as to why this underpricing might be a rational strategy for a firm are outlined in the paper along with a discussion of their likely applicability to REITs as an asset class. The author's preliminary study examines the IPOs of the 13 Canadian REITs listed on the TSE in mid-1998, and shows evidence of underpricing of IPO units over the first 10 and 20 days of trading.
Real Estate Investment Trusts (REITs) have been an available investment in the U.S. since the 1960s. REITs are publicly traded unit funds which invest primarily in income-producing real estate assets.1 These trusts allow investors to add real estate as an asset class to their investment portfolio. Relatively small amounts of money can provide ownership in a portfolio that is diversified by geography and property type. Because the units are publicly traded, the investment is substantially more liquid than a standard real estate investment. Daily trading data are available for REITs, making them a subject of intense study in the real estate finance literature.
REITs are relatively new phenomena in Canada,2 emerging largely as a result of a change to the Canadian Income Tax Act in 1995. This change allowed REITs to qualify as closed-end trusts, benefitting from more favourable tax treatment. The income earned in the REIT is not taxed at the trust level; it is passed as taxable income to the unit holders, along with applicable CCA deductions. This avoids the double taxation inherent in dividend payments from corporations, as well as providing some tax shelter for the cash flow received. The legislation also removed the 21-year deemed disposition rule for REITs, which was a drawback to investing in open-ended real estate funds. A further attraction is that REITs now automatically qualify as eligible investments for retirement savings plans. As the implications of these changes became apparent and as the economy improved, a flurry of new REITs appeared. The most recently developed REITs are closed-end. Three are reincarnations of open-ended funds that ran into difficulty providing redemptions in the soft real estate market of the late 1980s.
As of November 1994, there were only three REITs trading in Canada, and just five by March 1997. At the time of the Initial Public Offering (IPO) pilot study reported here (May 1998), there were 13 REITs trading on the Toronto Stock Exchange (TSE), seven of which were launched in 1997. Total assets in Canadian REITs grew from Cdn$80 million in 1993 to $4 billion in 1998. While this growth is dramatic, the market is still in its infancy. REIT assets in the U.S. in 1997 were US$100 billion (Clayton & MacKinnon, 2000). Since 1998 the Canadian REIT market has been relatively quiet. Riocan acquired REALFUND and Summit acquired Avista, both on unsolicited offerings. Three new funds have issued IPOs: Cominar in September of 1998, O&Y REIT and Retirement Residences REIT (RETREIT) in 2001. RETREIT and CPL REIT announced in 2002 that they would merge.
The surge of new REIT investment in Canada occurred for various reasons. The impressive performance of the stock market recently is part of the explanation, particularly when contrasted with the low interest rate returns available on alternative investments. REITs allow those with real estate portfolios to liquidate a portion of their holdings, with the opportunity to retain some ownership and possibly management control. For small investors, REITs allow access to a liquid investment in a diversified real estate portfolio with relatively little cash.
Institutional investors have also been attracted to these aspects of REIT investment when compared to the management, diversification, and liquidity risks associated with direct investment in real estate assets. The ability to frequently and accurately assess the value of this type of real estate holding is also attractive to institutional investors. Direct real estate investments have to be valued by an appraiser to estimate the appreciation portion of portfolio returns. Since large institutional-grade real estate changes hands infrequently, these value changes are difficult to measure accurately. The fact that REITs trade daily and thus the value of an institutional real estate portfolio, consisting of REIT units, can be tracked frequently and accurately is appealing to advisors and management who are asked to report on portfolio returns.
The Canadian REIT taxation rules are similar to those in the U.S. At least 95% of each year's operating income must be distributed to unit holders and realized capital gains must be distributed annually. Tax is not paid on the REIT income at the fund or trust level, but by the individual unit holder. However, Canadian trusts can invest in shares, bonds, mortgages, marketable securities, cash and/or real property as long as at least 80% of its investments are situated in Canada. In practice, Canadian REITs invest primarily in real property and would be classed as equity REITs in the U.S. vernacular.3
Equity REIT structures changed in the U.S. in the 1990s as the following quote from Ling and Ryngaert (1997) indicates:
Most of the recent REIT IPOs are fully integrated operating companies that can be characterized as 11 management plays" rather than as passive conduits for investors' capital. Their managements usually have substantial equity positions in the company (Ross & Klein, 1994) and all are infinite life REITs. Property management is either done internally or by management that works solely for the benefit of the REIT shareholders (McMahan, 1994). (p. 437)
All of the Canadian REIT IPOs date from the 1990s. The structure of these funds can be said to be similar to that described in the paragraph above. Each Canadian REIT used its IPO proceeds to purchase an existing portfolio of real estate assets; these were identified in the prospectus for the IPO. The original owners) of these properties retained some interest in the REIT through ownership of units and/or a management role.
The Residential Equities REIT (RESREIT) is a typical example of this. Greenwin Properties Group and Lehndorff Tandem Properties Group each owned and managed a significant portfolio of apartment buildings in the greater Toronto area and elsewhere in Canada. Together they developed the concept for RESREIT. The initial RESREIT portfolio consisted entirely of properties bought from either Greenwin or Lehndorff. The advisor' and property management arm for the REIT are both managed by senior personnel from Greenwin and Lehndorff and are jointly owned by these two entities. Greenwin and Lehndorff together owned approximately 25% of the REIT units after the initial public offering; they paid for their units through equity in the properties transferred. This arrangement, with partial ownership of the REIT by the advisor and property management arm, mirrors the 1990s version of REIT structure in the U.S.
This paper summarizes finance literature related to IPOs. There is considerable evidence that shares issued in the process of taking industrial firms public are, on average, underpriced. This evidence, and theories that purport to explain this underpricing, are examined in the next section. A discussion of how these theories of IPO underpricing might or might not relate to REITs follows, along with a summary of the literature related to IPOs for REITs in the U.S. A small pilot study analyzing IPOs for Canadian REITs is also reported.
Firms issuing shares for the first time could be assumed to have a goal of raising as much capital as possible through the process. Therefore, it has puzzled researchers over time to find evidence of significant underpricing of IPOs. Beatty and Ritter (1986), Chalk and Peavy (1987), and Ritter (1984), among others, all found evidence of significant positive returns for the first day of trading of IPO stocks. Smith (1986) summarizes the literature which documents underpricing in IPOs; for new issues, average underpricing appears to exceed 15%, a significant amount of money for firms to "leave on the table."
Several theories have evolved to explain why this underpricing might be rational. The most commonly offered explanation was developed by Rock (1986) and involves the "winner's curse." His model requires uncertainty about the potential performance of a company after it goes public and groups of investors interested in the offering who have heterogeneous information about the value of the firm. Those who are better informed will bid on what they perceive to be underpriced offerings and leave the others. The less informed investor is just as likely to bid on a "properly" priced offering as an underpriced one. Most new issues are rationed, so that the poorly informed investor ends up with a disproportionately large share of the IPOs that do not appreciate after they go on the market. This produces an adverse selection problem for uninformed investors. Deliberate underpricing of an average IPO by the issuer compensates the poorly informed investor for facing the winner's curse; if this were not done, the non-institutional investor would likely stay completely out of the IPO market and raising new funds would prove difficult.
Beatty and Ritter (1986) extended and tested Rock's model. They argued that one firm alone cannot provide a credible signal that it has underpriced its IPO in expected terms, since it will only go public once. The investment bank is required to vouch for the fact that the stock is priced below its expected value in the market. Since the investment bank relies on its reputation to gain future business, this signal from them should be credible. The empirical results identify a loss of market share for investment bankers who do not adequately underprice IPOs in the first time period of their study. They also find that greater uncertainty about the expected results for the firm requires greater underpricing of its IPO to attract investors. It is therefore in the firm's interest to release as much information as possible about what projects it is funding with the IPO proceeds.
This raises an interesting distinction between REIT IPOs and taking a private operating company public. Companies have two reasons for not wanting to be too specific about their future plans with the proceeds of a public offering. They do not want to be sued by investors if things do not go according to plan. And, they do not want their competitors to gain an advantage through increased knowledge about their operations. REITs, on the other hand, so long as they have appropriate purchase options on the properties they intend to buy, can reveal as much detail as they know about the assets they intend to acquire with the IPO proceeds, including type of property, size, location, major tenants, previous year's rent, and so on. For the Canadian REITs in this study, the significant amount of information revealed about their intended property acquisitions should have decreased the required underpricing for their IPO, if Beatty and Ritter are correct.
Rock's model was substantiated by the results of Chalk and Peavy (1987) who found that returns differed by initial price per share of the offering and that shares which sold for less than $1 initially accounted for much of the excess returns. There is likely to be considerable uncertainty about the outcome for the companies issuing these penny stocks. As well, there is likely to be a broader mix of potential investors the lower the initial offering price of the stock.
There is also evidence that the returns on IPOs are cyclical; that is, there are "hot markets" for IPOs. Ritter (1984) concluded that the mean return for IPO stocks bought at the offering price and sold at the closing bid on the first day of trading, during the 15-month period starting January 1980, was 48%. This compared with a mean return for the same trading strategy of 16.3% for the remainder of his six-year trading period. Following Rock's model of investor uncertainty, he initially investigated whether these hot issues related to increased uncertainty about expected returns for the firms during this time period. He concluded that the significant underpricing could be traced to one industry, natural resources, and the uncertainty about results for this industry during the hot issue time frame. During the remainder of the six-year period, the returns for natural resources offerings were not significantly different from other industries. He concluded that there is supporting evidence for Rock's theory of uncertainty leading to greater underpricing. He argued that providing greater information to investors should lead to a lesser need to underprice the IPO.
Several other explanations for rational underpricing of IPOs have received less attention than Rock's model. One is outlined in Houge, Loughran, Suchanek, and Yan (2001). They based their model on a theory proposed by Miller (1977), that in markets with short selling restrictions, such as IPOs, prices are determined by "optimistic" investors. Those pessimistic about the prospects for the firm do not enter the market until they are allowed to short sell the shares, a rational activity given their view of the future. Tinic (1988) developed and tested a theory that underpricing is a form of insurance against law suits and damage to the reputation of the investment banker in situations where the firm does not perform well after the IPO.
Implicit in Rock's model is the assumption that investors who have done their homework know more about the prospects for the firm than either the firm itself or the underwriter. Allen and Faulhaber (1989) presented a slightly different model of uncertainty about the firm. They assumed that the firm itself has the greatest knowledge about its own prospects, but has no credible way to pass this information on to investors. It is rational for the firm to underprice its IPO as a signal to investors. Investors who experience good returns on the firm's IPO will be more likely to purchase shares in a secondary offering. Firms with good prospects are thus willing to underprice their IPOs because they can make up the loss on future share offerings. Firms with poor prospects are reluctant to underprice because they are not likely to have the opportunity to make this up on a second offering of equity. Allen and Faulhaber did not test their model but argued that Ritter's "hot issue" market findings were consistent with their model.
A similar hypothesis was put forward by Ibbotson (1975). He argued that it may be rational for a firm to underprice its IPO if it expects to be going back to the market shortly for a seasoned equity offering, in order to "leave a good taste in the investor's mouth." Welch (1989) developed a formal model to support this. Firms that expect to perform well are willing to underprice their IPOs in order to ensure that future stock offerings are well received. Firms that do not expect to do as well run the risk that this will be found out by investors prior to their secondary offerings and these firms will be unable to recoup what was left on the table by the initial underpricing. It then may become rational for the firms with poor prospects to correctly price their IPOs.
These last two models fit well with the REIT situation. Most REITs have a particular portfolio of properties in mind when they issue an IPO. This was certainly true for the Canadian REIT offerings in the late 1990s. In many cases the REIT management had been operating the properties for a period of time and was, in effect, transferring ownership and refinancing existing assets. Their knowledge of future prospects for these assets was significantly greater than would be the case for a small manufacturing company using its IPO to expand operations and markets, for example.
The REIT legal structure in Canada also makes this model appealing. REITs must pay out 95% of their earnings each year in order to avoid taxation at the trust level. This means that, to expand, the REIT must frequently go back to the market for additional funds by selling more units. There have been numerous returns to the market by Canadian REITs since their IPOs. This behaviour fits with both the Allen/Faulhaber and Welch models.
In the real estate literature, researchers have recently examined REIT IPOs to see if underpricing is the norm for REITs as it is for industrial firms. Wang, Chan, and Gau (1992) (WCG) found significant overpricing for REIT IPOs, in contrast to the underpricing generally found in the finance literature. They documented a first-- day price decline of 2.82% on a sample of 87 REIT IPOs issued between 1971 and 1988. Balogh and Corgel (1992) obtained the same result for a smaller sample of REITs during the latter portion of WCG's study period, although the overpricing was not large. Below, Zaman, and McIntosh (1995) (BZM) found REIT IPOs to be correctly priced. They attributed the findings for WCG to using bid-price to calculate returns, instead of ask-price or a bid-ask average, which they argue is the proper price to use in calculating returns when testing for under or overpricing.
Ling and Ryngaert (1997) examined IPOs for 85 REITs over the period 1991-1994 in the U.S. In contrast to the earlier REIT IPO studies, they found that, during this time period, the IPOs were underpriced; on average, the price appreciated 3.6% on the first day of trading. The size of the return dropped as the reputation of the underwriter increased (this is used as a proxy for uncertainty about the quality of the offering); it increased with the level of institutional ownership. This evidence is consistent with the winner's curse explanation used to account for underpricing in industrial company IPOs.
Ghosh, Nag, and Sirmans (2000a) found similar underpricing also existed for seasoned equity offerings (SEOs) for REITs in the post 1990 market. They attributed this to the same factors that explain underpriced IPOs: uncertainty about the value of the REIT assets and performance post-offering and the significant proportion of institutional investment in post-1990 REITs.
As discussed above, REITs are required to distribute 95% of their net income as dividends to avoid taxation at the trust level. Their tax exempt status makes debt less attractive for them. Thus, to grow through asset acquisition, they need frequent access to the capital market. Ghosh, Nag, and Sirmans (2000b) argued that REITs underprice their IPOs in order to entice investors to purchase secondary offerings. Their analysis of REIT IPO and SEO linked offerings between 1992 and 1996 found that REITs that more substantially underpriced their IPOs were more likely to sell seasoned offerings sooner and to raise more funds through the joint IPO-SEO offerings. However, they found the SEOs also to be underpriced, disputing the expectation that IPO underpricing is overcome through successful subsequent equity offerings. While this latter study might be considered weak support for Welch's model, the existence of underpricing in the SEOs is troubling.
Given the relative small number and newness of REITs in Canada, a full analysis of REIT underpricing was not possible. However, a preliminary study to examine pricing of Canadian REIT IPOs is described in the balance of the paper.
Canadian REIT IPO Pilot Study
IPOs of recently introduced Canadian REITs were examined to determine whether excess returns were available on these investments or whether they experienced the overpricing observed for U.S. REIT IPOs in the 1970s and '80s. IPO data for all 13 REITs trading in Canada as of May 1998 were used in the study.5 Summary data describing these are presented in Table 1. The REITs are now all closed-end funds. However, three of them-RioCan, RealFund, and Canadian REIT-operated for several years as open-ended funds, and were restructured to closed-end funds when the tax legislation governing REITs changed.6 It could be argued that the uncertainty regarding future prospects for these three funds was less than that for the remaining REITs. All IPOs were issued over a relatively short time frame: three in 1993-94 and the remaining 10 between January 1996 and February 1998.
These Canadian REITs concentrate on equity investment; eight of the 13 are diversified, owning a combination of retail, office and industrial properties. The other five, which are all newer REITs, are characterized by specialized investment in a particular type of real estate (hotels, long-term care facilities, or residential buildings.)
The bid, ask, and closing prices7 for each REIT were collected for the first 20 days after start of trading in order to analyze the IPOs of these REITs to look for over- or underpricing. Raw returns were calculated for each day as:
MR, is the market return calculated from the TSE 3009 index between day n and day n-1.
The cumulative raw and market adjusted returns were calculated over the first 20 trading days for three different intervals: days 1-10, 11-20, and 1-20. Mean returns and standard deviations of returns for the individual days and for the cumulative returns were also calculated.
The results of the data analyses appear in Tables 2 and 3. The raw and market adjusted mean daily and cumulative returns are reported in Table 2, as well as the number of REITs with positive returns each day and t-- test results for whether the returns are significantly different from zero. Table 3 reports daily and multi-day returns for individual REITs, separated into diversified and single property type panels.
The results in Table 2 indicate that neither raw nor market-adjusted returns were significantly different from zero on any of the 20 individual days, with the exception of day 17. However, the bottom rows in this table indicate that cumulative market-adjusted returns for days 1-- 10 and 1-20 were significantly positive. Investors, on average, would have done well to buy these REIT units through the IPO, hold them for 10 days and then sell. Table 3 does not indicate that diversification made a significant difference in the returns for the first 20 days. Winners and losers are spread throughout both panels of the table.
The presence of significantly positive cumulative returns led to the question of what might explain this evidence of underpricing. Information was collected on the following features of the REITs in the sample10:
1. The size of the initial public offering - larger REITs might be assumed to have a broader asset mix and lower management fee percentages. They might also be more attractive to institutional investors (generally considered as a proxy for "informed" investors). Ling and Ryngaert (1997) found returns dropped as the size of the offering increased.
2. Whether or not the REIT was one of the three transformed to a closed-end fund (i.e., had a longer history than the new offerings) - the level of uncertainty about performance could be less for these funds since they were already available to the public as investments prior to their re-issuance as REITs. Lower returns for these funds would lend support for the winner's curse theory.
3. Whether the REIT's portfolio was diversified or not -- it could be argued that portfolios of diversified types of properties would be expected to be lower risk than, for example, a REIT which held only hotels.
4. Whether an installment payment option was available to purchase the REIT units - several funds sold units for an initial price of $6.00 with a further $4.00 due within the next several months after the IPO date. This allowed investors more time to assess the REIT before committing all their funds.
5. Maximum allowable debt as specified in the prospectus - higher debt levels imply higher risk and less opportunity to grow through property acquisition without returning to the equity market, given the restrictions on retaining earnings in REITs.
There were insufficient data for satisfactory regression analysis. However, Table 4 reports the correlation coefficients for the variables cited above relative to the market adjusted day 1 (MAI) returns and cumulative returns for days 1-10.
Row 4 indicates that MAI is significantly negatively correlated with "transformed." This implies that if the REIT had been operating as an open-ended fund and thus had a track record prior to the IPO, the underpricing was likely to be smaller. The cumulative returns were also likely to be smaller although the correlation is not significant. This is consistent with lower uncertainty about future results for the fund leading to a smaller requirement for underpricing.
If investors were allowed to buy their units with installment payments, MAI was likely to be higher; the correlation is significant and positive. The correlation for the cumulative returns is also positive but not significant. This result implies that allowing the investor time before he has to commit the full unit price increases the amount of underpricing required. This result seems counterintuitive. It may be that these IPOs were targeted at individual investors-Rock's "uninformed" investors-and that this accounts for the increased underpricing. However, Ling and Ryngaert (1997) found underpricing increased with the amount of institutional ownership.
These are the only significant correlations. The higher the maximum debt specified in the prospectus, the lower the returns; this is the opposite of expectations but the results are not significant. There was not a substantial difference in the maximum debt specified in the prospectuses for the different funds, which may explain this result. Offering size and level of diversification had mixed and non-significant correlations with the two return measures.
Partly, these results are confounded by the fact that the first set of IPOs were the transformed closed-end funds, since they were ready to go to market more quickly (and in fact had been pressing for the tax changes). These funds were all diversified, had an existing history and generally did not use the installment plan for unit selling. Thus, the later specialized REIT IPOs may have had different results because of market timing as much as because of fund characteristics.
These results are preliminary given the immaturity and relatively small size of the market for REITs in Canada. There is already some evidence that the cyclical problems that have plagued REITs historically in the U.S. will be mirrored in Canada. Unit values for most Canadian REITs dropped with the general market decline in the late 1990s. There have been three mergers of pairs of REITs and only three new IPOs have been issued since 1998. Payout ratios have been strong, however, and there have been many successful seasoned equity offerings.
The results of this study, showing some evidence of underpricing of IPOs of REITs in Canada, align with the 1990 results of REIT underpricing in the U.S. Given the similarity of management and ownership structure between the new U.S. and Canadian REITs in the 1990s, this result is not surprising. There is also some weak support for the winner's curse explanation of the REIT underpricing; REITs that were transformed from existing open-ended funds had a smaller degree of underpricing than funds new to the market.
As the REIT market matures in Canada, researchers may want to examine later IPOs to see if underpricing persists. It may be that the cumulative significant returns were a result of the newness of this type of asset class in the market. An examination of seasoned equity offerings by REITs in Canada would provide a richer data set and a method for testing whether the Rock or Welch model seems more applicable. Lastly, an examination of the level of underpricing, if it exists, in other trust IPOs issued in Canada since the tax law changed (for example in the natural resource sector) could be compared to the REIT underpricing found in this study to examine the relative magnitude.
Allen, F. & Faulhaber, G.R. (1989). Signaling by underpricing in the IPO market. Journal of Financial Economics, 15, 303-323.
Balogh, C. & Corgel, J.B. (1992). Initial return behaviour of one class of composite security IPOs: REITs. Working Paper, Bentley College.
Beatty, R.P. & Ritter, J.R. (1986). Investment banking, reputation and the underpricing of initial public offerings. Journal of Financial Economics, 15 (1/2), 213-232.
Below, S., Zaman, M.A., & McIntosh, W. (1995). The pricing of real estate investment trust initial public offerings. Journal of Real Estate Finance and Economics, 11, 5564.
Chalk, A.J. & Peavy, J.W. (1987). Initial public offerings: Daily returns, offering types and the price effect. Financial Analysts Journal, 43 (5), 65-69.
Clayton, J. & MacKinnon, G. (2000). Measuring and explaining changes in REIT liquidity: Moving beyond the bid-ask spread. Real Estate Economics, 28, 89-116.
Corgel, J.B., McIntosh, W., & Ott, S.H. (1995). Real estate investment trusts: A review of the financial economics literature. Journal of Real Estate Literature, 3 (1), 1343.
Ghosh, C., Nag, R., & Sirmans, C.ft (2000a). The pricing of seasoned equity offerings: Evidence from REITs. Real Estate Economics, 28 (3), 363-384.
Ghosh, C., Nag, R., & Sirmans, C.F. (2000b). A test of the signaling value of IPO underpricing with REIT IPO-SEO
pairs. Journal of Real Estate Finance and Economics, 20 (2), 137-154.
Houge, T., Loughran, T, Suchanek, G., & Yan, X. (2001). Divergence of opinion, uncertainty and the quality of initial public offerings. Financial Management, 30 (4), 5-23.
Ibbotson, R. (1975). Price performance of common stock new issues. Journal of Financial Economics, 2, 235-272. Ling, D. & Ryngaert, M. (1997). Valuation uncertainty, institu
tional involvement and the underpricing of IPOs: The case of REITs. Journal of Financial Economics, 43, 433-456. McMahan, J. (1994). The long view: A perspective on the REIT market. Real Estate Issues, August, 1-4.
Miller, E.M. (1977). Risk, uncertainty, and the divergence of opinion. Journal of Finance, 32, 1151-1168.
Ritter, J.R. (1984). The "hot issue" market of 1980. Journal of Business, 57 (2), 215-240.
Rock, K. (1986). Why new issues are underpriced. Journal of Financial Economics, 15, 187-212.
Ross, S. & Klein, R. (1994). Real estate investment trusts for the 1990s. Real Estate Finance Journal, Summer, 37-44. Smith, CW. (1986). Investment banking and the capital acqui
sition process. Journal of Financial Economics, 15, 3-29. Tinic, S.M. (1988). Anatomy of initial public offerings of common stock. Journal of Finance, 43 (4), 789-822.
Wang, K., Chan, S.H., & Gau, G. (1992). Initial public offerings of equity securities: Anomalous evidence using REITs. Journal of Financial Economics, 31 (3), 381-410.
Welch, I. (1989). Seasoned offerings, imitation costs and the underpricing of initial public offerings. Journal of Finance, 44 (2), 421-449.
Jane Londerville* University of Guelph
*Consumer Studies Department, University of Guelph, Guelph, ON, Canada N1G 2W1. E-mail: firstname.lastname@example.org
The author wishes to thank participants at the 1999 AREUEA meetings in New York, especially discussant Jaime Alvayay, for helpful comments on an earlier version of this paper. Dogan Tirtiroglu and two anonymous referees provided very helpful comments on this final draft. Research assistance was provided by Melanie O'Gorman through funding for an Undergraduate Research Assistant position from the University of Guelph.…