《Moral-hazard-asymmetric-information-and-IPO-lockups_2010_Journal-of-Corporate-Finance.docx》由会员分享,可在线阅读,更多相关《Moral-hazard-asymmetric-information-and-IPO-lockups_2010_Journal-of-Corporate-Finance.docx(14页珍藏版)》请在得力文库 - 分享文档赚钱的网站上搜索。
1、Journal of Corporate Finance 16 (2010) 320332 Moral hazard, asymmetric information and IPO lockups Chris Yung , Jaime F. Zender 1 Leeds School of Business, University of Colorado at Boulder, United States a r t i c l e i n f o a b s t r a c t Article history: Received 3 November 2009 Received in rev
2、ised form 11 December 2009 Accepted 11 December 2009 Available online 7 January 2010 JEL classication: G24 G32 Keywords: Initial public offerings Asymmetric information Moral hazard IPO lockups 1. Introduction Moral hazard and asymmetric information have both been proposed as the motive behind the u
3、se of IPO lockup provisions, with each receiving empirical support in the literature. Rather than consider them to be mutually exclusive motivations, we hypothesize that each is dominant for a different set of rms. We provide novel empirical support for the underwriter certication hypothesis then us
4、e this hypothesis to categorize the rms in our sample. Firms that are certied by a reputable underwriter see a reduction in the severity of asymmetric information relative to other rms and therefore will be more likely to see moral hazard as the friction that motivates the use of the lockup provisio
5、n. For those rms that are unable to obtain high reputation underwriter certication it is relatively more likely that asymmetric information is the motivation for the use of the lockup provision. Based on this separation of rms we introduce and provide empirical support for a novel set of hypotheses
6、concerning the lockup period. 2010 Elsevier B.V. All rights reserved. Recently, considerable attention has been paid to understanding the lockup provision embedded in the contract between an underwriter and a rm engaged in its initial public offering of equity. This literature examines motivations f
7、or the lockup provision, the determinants of the length of the lockup period, and the returns around lockup expiration. Brav and Gompers (2003) and Brau et al. (2005) both examine the motivation for the use of the lockup provision and the determinants of the lockup length and come to opposing conclu
8、sions with the former reporting support for the hypothesis that the lockup provision is used to control moral hazard and the latter reporting support for the hypothesis that lockups are used to control adverse selection. Here, we re-examine this issue and offer a resolution for the apparent conict.
9、We posit that all rms suffer from both moral hazard and adverse selection problems. For some, moral hazard will be the dominant consideration in including the lockup provision in the IPO contract and in determining the lockup length, while asymmetric information will dominate for others. Based on th
10、is hypothesis and the idea that the signal produced by the choice of lockup period is not perfectly revealing we develop a set of testable predictions concerning the length of the lockup period. Underpricing, a central issue in the IPO literature, plays a major role in our analysis. Underpricing is
11、a particularly useful diagnostic in this environment because it is affected by information asymmetry but not by post-issue commitment problems.2 In such an environment, the empirical regularities seen in the full sample (for an “average” rm) do not accurately reect the behavior of * Corresponding au
12、thor. Tel.: +1 303 492 7437. E-mail addresses: chris.yungcolorado.edu (C. Yung), jaime.zendercolorado.edu (J.F. Zender). 1 Tel.: +1 303 492 4689. 2 IPO underpricing is caused by asymmetric information in several canonical models, e.g. Benveniste and Spindt (1989) and Rock (1986). We are not aware of
13、 a model in which underpricing is caused by ex-post commitment problems of insiders. Likewise, surveys such as Jenkinson and Ljungqvist (2001) and Ibbotson et al. (1994) do not entertain the notion that underpricing could be caused by ex-post managerial agency problems. Given that underpricing is re
14、solved extremely rapidly, it is difcult to imagine how it could serve to control moral hazard. 0929-1199/$ see front matter 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jcorpn.2009.12.004 Contents lists available at ScienceDirect Journal of Corporate Finance journal h omepage: www.el sevier
15、. com/l ocate/j corpfi n C. Yung, J.F. Zender / Journal of Corporate Finance 16 (2010) 320332 321 rms for which asymmetric information is the dominant consideration in determining the lockup length nor the behavior of rms for which moral hazard is the dominant friction. We separate our sample based
16、on whether asymmetric information (the asymmetric information subsample) or moral hazard (the moral hazard subsample) is most likely to motivate the use and determine the length of the lockup. This sample bifurcation is the basis of our research design, and it leads to differential predictions on th
17、e co-movements of key variables. For example, consider a rm that suffers from both moral hazard and asymmetric information problems but for which moral hazard is the determinant of the lockup length. Because asymmetric information is present, the IPO is expected to exhibit underpricing. However, con
18、sider a hypothetical shock that increases the level of information asymmetry while leaving the severity of moral hazard unchanged. Such a shock will increase underpricing but not the lockup length for this rm. Conversely, consider a hypothetical shock increasing the severity of moral hazard while le
19、aving information asymmetry unchanged. Such a shock increases the lockup length but not underpricing. In either case, lockup length and underpricing should not co-vary in a cross section of such rms. Suppose instead that information asymmetry is the dominant consideration when choosing the lockup le
20、ngth. The effect of increasing the level of information asymmetry will be to increase both underpricing and the length of the lockup. A shock increasing the severity of the moral hazard problem, however, should not affect underpricing or the lockup length. We therefore predict that lockup length and
21、 underpricing exhibit a positive correlation in the information asymmetry sample. When the lockup is chosen to control a moral hazard problem, it will be chosen recognizing the personal cost to insiders of a poorly diversied position. A given level of control of the managerial moral hazard problem c
22、an be achieved with a combination of post-IPO shares and a given lockup. In order to maintain a constant level of control of the moral hazard problem, at the margin, an increase in the post-IPO ownership by insiders allows a reduction in the length of the lockup. The post-IPO level of ownership will
23、 not, however, have a rst order impact on the nature of the pre-IPO asymmetric information problem. Therefore the level of insider ownership and the length of the lockup should be negatively correlated in the cross section of moral hazard rms and uncorrelated in the sample of asymmetric information
24、rms. Our empirical tests require an identication strategy by which rms may be separated based on which of the frictions, moral hazard or asymmetric information, is likely to be the dominant consideration for the length of the lockup provision. One such candidate proxy is rm size. As a rm grows, its
25、assets are likely to become more transparent and easier to value, which in turn should reduce the severity of information asymmetry (Beatty and Ritter (1986). By contrast, rm size need not diminish agency problems. In particular, managerial misconduct and/or shirking occur at both large and small rm
26、s. Consequently we expect the relative magnitude of these two frictions to vary with rm size. Our second proxy is the reputation of the underwriter. As argued by Beatty and Ritter (1986) (and supported empirically in Carter and Manaster (1990) certication by a high reputation underwriter reduces the
27、 severity of information asymmetry. On the other hand, the analogous argument regarding agency problems is on much thinner ground. In particular, while investment banks can certify the value of an asset, they cannot certify (or prescribe) managers future actions. In fact, without retained (and locke
28、d) equity as a costly signal, insiders themselves have difculty making credible promises regarding their own actions. It is less plausible still that such a commitment could be delegated to a third party not involved in running the rm. Hence, ceteris paribus, underwriter certication should reduce th
29、e severity of information asymmetry relative to that of commitment problems. Therefore it should be more likely that the dominant concern for rms taken public by high reputation underwriters is moral hazard. Conversely, it should be more likely that asymmetric information is the determining friction
30、 for the lockup provision in rms taken public by low reputation underwriters. We empirically verify the relation between both proxies and the level of asymmetric information by examining lockup expiration returns. Expiration returns allow us to consider the severity of asymmetric information at the
31、precise moment when managers trades reveal the quality of their information. More dispersion in post-lockup returns suggests more dispersion in private information.3 Our results are consistent with this view. Larger IPOs and those backed by reputable underwriters are associated with signicantly smal
32、ler dispersion in returns at the expiration of the lockup. Empirically, the bifurcated samples behave as hypothesized. In the asymmetric information subsamples (small rms and those taken public by low reputation underwriters) we nd that lockup length co-varies positively with underpricing. While in
33、the moral hazard subsamples lockup length is uncorrelated with underpricing. We also document the hypothesized different co- movements between the lockup length and post-IPO insider ownership across the two subsamples. The correlation between lockup length and insider ownership is negative in the mo
34、ral hazard subsamples and zero in the asymmetric information subsamples. Two recent papers are closely related to our analysis. Brav and Gompers (2003) examine three motives for the existence of the lockup provision, asymmetric information (signaling), moral hazard (commitment), and rent extraction
35、on the part of issuers. They nd no support for the rent extraction hypothesis and interpret their evidence as suggesting that lockups function primarily as a commitment device to alleviate moral hazard problems between managers and new shareholders. Arguing that the variables used by Brav and Gomper
36、s (2003) to measure moral hazard are more naturally interpreted as indicating the severity of an asymmetric information problem, Brau et al. (2005) (BLM) instead favor the hypothesis that lockups serve as a signal of quality. 3 See also Lowry, Ofcer and Schwert (2006), Wang and Yung (in press) and Y
37、ung et al. (2008) who also identify cross-sectional dispersion in returns with asymmetric information. 322 C. Yung, J.F. Zender / Journal of Corporate Finance 16 (2010) 320332 They develop a signaling model and provide empirical support for its predictions. Their model, however, does not consider mo
38、ral hazard. A common feature of Brav and Gompers (2003) and Brau et al. (2005) is that their analyses consider that the signaling hypothesis and the commitment (or moral hazard) hypothesis are mutually exclusive. An alternative interpretation is that these papers seek to determine which of these fri
39、ctions determines the lockup length for the “average” rm in the universe of rms that have gone public. Brav and Gompers (2003) and BLM (2005) also both use the intuition of a separating signaling equilibrium to generate their empirical predictions concerning the signaling hypothesis. This is a somew
40、hat limiting perspective because if the equilibrium is separating then asymmetric information cannot be a determinant of any underpricing at the time of the IPO. Consequently, their analyses almost entirely ignore underpricing. Reconciling these conicting results, our analysis highlights the importa
41、nce of Brav and Gompers (2003) argument in the moral hazard subsample and BLMs (2005) argument in the asymmetric information subsample. In particular, the aforementioned co-movements emphasized in this paper (lockup length and underpricing; lockup length and insider holdings) differ between the two
42、subsamples. These differences are obscured in the full sample, which emphasizes the importance of considering the motivations for the lockup provision separately. The paper is organized as follows. Section 2 develops our empirical hypotheses. Section 3 describes the data and Section 4 presents our r
43、esults. Section 5 concludes. 2. Hypothesis development 2.1. The role of asymmetric information Firms issuing public equity face a variety of constraints imposed by asymmetric information regarding the value of assets-in- place and growth options. If the market is characterized by heterogeneously inf
44、ormed investors, then offer prices must be discounted on average, either in order to compensate uniformed investors for the winners curse (Rock, 1986) or to persuade informed investors to reveal their information (Benveniste and Spindt, 1989). Compounding the problem, even if issues were fairly pric
45、ed on average, high-quality rms would still suffer mispricing losses because they are pooled with lower quality rms. Given the disclosure and reporting requirements associated with public equity, information concerning rm quality is revealed over time. Strategic trading by speculators in the seconda
46、ry market may hasten this process, however, asymmetric information and the subsequent revelation of information in the aftermarket allow a role for the lockup of insiders shares as a signaling device. Insiders in high-quality rms would prefer to lock shares, committing to sell them only after (it be
47、comes highly probable that) information has been revealed, in order to avoid being pooled with (at least some of) the low-quality rms. Of course, lockups impose personal costs on rm insiders in the form of poor diversication. Insiders in high-quality rms will therefore choose lockups of intermediate
48、 rather than extreme length, balancing the inherent trade-off. Insiders in low-quality rms, while preferring to avoid locking their shares, must choose between mimicking this choice and having their quality truthfully revealed. Heterogeneity within the set of low-quality rms results in a pooling equ
49、ilibrium in which high-quality rms select a lockup period long enough to prevent some but not all low-quality rms from mimicking. An important consequence of this is that there will be underpricing at the IPO despite the signaling efforts of high-quality rms.4 Consider a rm for which the asymmetric information problem is the determinant of the lockup length. When the lockup length is used as a signal of rm quality, an increase in the asymmetry of information magn