Most initial public offerings (IPOs) include a lockup agreement that prohibits pre-IPO shareholders from selling their holdings for a specified period. Industrial IPO lockup periods commonly extend for 180 days, which can vary based on the agreement reached between insiders and underwriters. A lockup agreement assures the market that pre-IPO shareholders do not cash out immediately following the IPO and therefore aligns their interests with those of outside investors.
Lockup agreements can provide a commitment solution to control the moral hazard problem or act as a signal to reduce the level of asymmetric information. Results from previous studies on the relative importance of these two functions of lockup agreements are mixed.
While Brav and Gompers (2003) find evidence to support the commitment device hypothesis, Brau, Lambson and McQueen (2005) support the signaling argument. Examining IPOs in Singapore, Chong and Ho (2007) find a positive relation between lockup length and voluntary earnings forecast disclosure. Their findings show that firms with greater asymmetric information and moral hazard problems feature longer lockup periods.
Another important issue related to IPO lockups is the market reaction to lockup expiration. The terms of lockup agreements are public information; the number of locked shares and the unlock date are reported in the IPO prospectus. The efficient market hypothesis suggests that investors will respond only to unexpected information; therefore, no abnormal returns should be observed around the unlock date. However, the extant evidence does not support this hypothesis. Field and Hanka (2001), Bradley, Jordon, Roten, and Yi (2001), and Ofek and Richardson (2000) find significantly negative abnormal returns around the unlock date for U.S.
industrial IPOs. These negative abnormal returns around the unlock date could possibly be
caused by a downward sloping demand curve, unexpected sales by insiders, or aggressive sales by venture capitalists. Several studies report that negative market responses are more acute for firms with venture capital backing. For example, Bradley et al. (2001) examine a sample of 2,539 IPO firms and find that significant abnormal returns exist only for IPO firms backed by venture capitalists.
Previous studies on IPO lockups focus on industrial IPOs and exclude Real Estate Investment Trusts (REITs). Although Brau, Lambson, and McQueen (2005) incorporate investment funds into their study, they do not separate REITs from other types of investment funds. REITs are different from other investment funds in at least two ways. First, they operate as corporations under certain specific regulatory requirements set up by Internal Revenue Service.
Second, the underlying assets of REITs are real assets instead of financial assets. Consequently, the general conclusions drawn from the studies of IPO lockups for industrial firms or investment funds may not apply to REIT IPOs.
Lockups in REIT IPOs warrant investigation for several reasons. First, the REIT market has grown tremendously in the past 30 years, especially after several regulatory changes in the early 1990s, making REITs a viable investment option for both individual investors and institutional investors. Between 1971 and 2007, the number of REITs in the United States increased from 34 to 152, and their market capitalization increased by more than 200 times, from
$1.5 billion to $312 billion. According to the National Association of Real Estate Investment Trusts (NAREIT), at the end of 2007, the 35-year annual return for the FTSE NAREIT Equity REITs index was 14.15%, compared with 8.6% for the Merrill Lynch Corporate/Government Bond Index, 12.95% for the S&P 500, and 9.67% for the Dow Jones Industrial Average.
Second, studying REIT IPOs allows us to focus on the impact of other factors, rather than
venture capital backing, on the market reaction to lockup expiration. REITs are not normally backed by venture capital, if abnormal returns are mainly caused by aggressive sales by venture capitalists; as reported for industrial IPOs, price movement around the unlock date for REIT IPOs should be relatively stable. However, if abnormal returns occur around the unlock date for REIT IPOs, factors other than aggressive sales by venture capitalists are likely affecting the market response.
Third, examining REIT IPOs would shed new light on the debate on the relevance of the commitment device and signaling hypothesis in explaining IPO lockup length and market reaction around lockup expiration. Asymmetric information and moral hazard problems have been reported as important determinants of lockup length and market responses to lockup expiration for industrial IPOs. REITs have different nature of agency conflict and asymmetric information problems due to several unique characteristics, which in turn, may uniquely impact lockup agreements or result in different market reactions around the unlock date.
Compared to industrial firms, REITs are more liquid and highly regulated, and thus are more transparent. The relative informational symmetric nature of REITs implies that REITs would not rely on lockup length as a signal device, and proxies for asymmetric information would have no significant impacts on the market reaction to lockup expiration.
Whether REITs have more severe agency problems than industrial firms is less straightforward. On one hand, REITs are expected to have a less severe agency problem due to restrictions on income source and payout distribution. REITs are required to distribute at least 90% of their income as dividends to investors, and to earn a minimum of 75% of their incomes from real estate properties or mortgage loans. As fewer free cash flows are left and managers have less discretion over the types of assets they can invest, the moral hazard problem is reduced.
On the other hand, the restrictions on income source and asset type, together with the restriction on ownership structure, weaken the mechanism of corporate control market in the REIT industry (see Ghosh and Sirmans, 2003). In addition to restrictions on income source and asset type, no more than 50% of REIT shares may be held by five or fewer individuals during the second half of any taxable year. These restrictions make inter-industry acquisition extremely difficult and reduce the possibility of hostile takeover, an alternative mechanism that can be used to control the agency problem. For these reasons, Campbell, Ghosh, and Sirmans (2001) find no evidence of any hostile REIT takeovers during the 1990s. Also, Ghosh and Sirmans (2003) find that outside board members of REITs are ineffective in performing their monitoring function.
Lastly, REITs have unique management structure choices. Unlike industrial firms, since 1986 REITs have had the option of being internally or externally managed. The effectiveness of IPO lockup as a commitment device would depend on management types because of different moral hazard problems associated with different structures. If the commitment device hypothesis holds, management structure should have an impact on the choice of lockup length as well as on the market reaction to lockup expiration.
In this study, we use a sample of REIT IPOs for the period from 1980 to 2006 to analyze the motivation behind the lockup agreement, examine the determinants of lockup length, and evaluate the market reaction on lockup expiration. Lockup lengths and market reaction to lockup expiration are two major issues related to IPO lockup. Examining two issues together allows us to have a better understanding on the decision of lockup agreement. Brav and Gompers (2003) also examine both issues and claim that “examine whether cross-sectional differences in abnormal returns around this event can shed light on the competing hypotheses for the existence of lockups” (Brav and Gompers (2003), p.20).
Our findings for REIT IPO lockups are quite different from those reported in previous studies (e.g. Bradley et al. 2001; Brav and Gompers, 2003; Field and Hanka, 2001). Although a smaller percentage of equity REIT IPOs, compared with industrial IPOs, feature a lockup, REIT IPOs have a longer average lockup period. We find a close relation between management structure and lockup length. The Tobit analysis indicates that lockup length increases with offer size and the choice of self-management style. Offer size can be a proxy for asymmetric information and moral hazard problems. Although the positive impact of offer size on lockup length lends support to the commitment solution argument, it contradicts the prediction of the signaling hypothesis. The positive and significant relation between the self-managed dummy and lockup length also illustrates the important role of moral hazard in explaining lockup length.
Contrary to the findings for industrial IPOs in the current literature, our results show no significant abnormal returns for REIT IPOs around lockup expiration for the full sample. Our results imply that there are no unanticipated supply shocks or drastic demand drops around the unlock date, which can be explained by the unique characteristics of REITs. The regression analysis indicate that moral hazard problem, rather than asymmetric information, can partly explain the cross sectional difference in abnormal returns around the unlock date.
This study is organized as follows. We first provide the stylized facts of equity REIT IPOs. We then analyze the determinants of lockup length for REIT IPOs. Next, we examine the market reaction around the unlock date and the determinants of market reaction. Finally, we conclude.