48
Author's Accepted Manuscript Large Shareholders and Disclosure Strategies: Evidence from IPO Lockup Expirations Yonca Ertimur, Ewa Sletten, Jayanthi Sunder PII: S0165-4101(14)00029-9 DOI: http://dx.doi.org/10.1016/j.jacceco.2014.06.002 Reference: JAE1019 To appear in: Journal of Accounting and Economics Received date: 11 December 2011 Revised date: 23 May 2014 Accepted date: 4 June 2014 Cite this article as: Yonca Ertimur, Ewa Sletten, Jayanthi Sunder, Large Shareholders and Disclosure Strategies: Evidence from IPO Lockup Expira- tions, Journal of Accounting and Economics, http://dx.doi.org/10.1016/j.jacce- co.2014.06.002 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. www.elsevier.com/locate/jae

Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

Embed Size (px)

Citation preview

Page 1: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

Author's Accepted Manuscript

Large Shareholders and Disclosure Strategies:Evidence from IPO Lockup Expirations

Yonca Ertimur, Ewa Sletten, Jayanthi Sunder

PII: S0165-4101(14)00029-9DOI: http://dx.doi.org/10.1016/j.jacceco.2014.06.002Reference: JAE1019

To appear in: Journal of Accounting and Economics

Received date: 11 December 2011Revised date: 23 May 2014Accepted date: 4 June 2014

Cite this article as: Yonca Ertimur, Ewa Sletten, Jayanthi Sunder, LargeShareholders and Disclosure Strategies: Evidence from IPO Lockup Expira-tions, Journal of Accounting and Economics, http://dx.doi.org/10.1016/j.jacce-co.2014.06.002

This is a PDF file of an unedited manuscript that has been accepted forpublication. As a service to our customers we are providing this early version ofthe manuscript. The manuscript will undergo copyediting, typesetting, andreview of the resulting galley proof before it is published in its final citable form.Please note that during the production process errors may be discovered whichcould affect the content, and all legal disclaimers that apply to the journalpertain.

www.elsevier.com/locate/jae

Page 2: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

1

Large Shareholders and Disclosure Strategies: Evidence from IPO Lockup Expirations*

Yonca Ertimur**

University of Colorado at Boulder

Ewa Sletten§

Boston College

Jayanthi Sunder#

University of Arizona

May 2014

Abstract: We examine the effect of large shareholders’ ex ante selling incentives on firms’ voluntary disclosure choices in the setting of IPO lockup expirations. We find evidence that managers delay disclosures of bad news, not for their own benefit, but to enable influential pre-IPO shareholders to sell their shares at more favorable prices. Delays are more pronounced when aggregate selling incentives are greater, when uncertainty is high, and when venture capitalists, influential investors with strong selling incentives, own more shares. Simultaneously, managers’ disclosure decisions reflect litigation concerns; no significant delays occur when litigation risk is high or when managers trade themselves.

��������� ��������������������������������������������������������������������������������

* We thank Stephen Baginski, Brian Cadman, Fabrizio Ferri, Mei Feng, Alan Jagolinzer, Jeff Ng, Chris Noe (the referee), Jong Chool Park, Sugata Roychowdhury, Katherine Schipper, Thor Sletten, Shyam V. Sunder, Mohan Venkatachalam, Beverly Walther, Ross Watts (the editor), an anonymous referee, and seminar participants at the AAA Annual Meetings, Multinational Finance Society Conference, Arizona State University, Columbia University, INSEAD, MIT Sloan Economics and Finance Seminar, Ohio State University, University of Colorado at Boulder, University of Southern California, University of Texas at Dallas, and University of Utah for their valuable comments. An earlier version of this paper was titled “Voluntary Disclosure Strategy around IPO Lockup Expirations.” ** Leeds School of Business, 419 UCB, 995 Regent Drive, Boulder, CO 80309. E-Mail: [email protected] § Corresponding author. 140 Commonwealth Ave, Fulton 520 D, Chestnut Hill, Ma 02482. E-Mail: [email protected] # 1130 E Helen St, Suite 301G, Tucson, AZ 85712. E-Mail: [email protected]

Page 3: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

2

1. Introduction

Previous studies (Noe 1999, Cheng and Lo 2006, and Rogers 2008) find that managers in

a position to benefit themselves through strategic disclosure do not do so, likely due to the

efficacy of insider trading regulations and the threat of shareholder litigation. However,

managers’ own profit is not the only potential incentive for strategic disclosure. Large

shareholders such as venture capitalists (VCs), private equity and hedge fund investors hold

considerable influence over management (Gompers and Lerner 2004; Brav, Jiang, Partnoy, and

Thomas 2008; Klein and Zur 2009), and it is an open question whether these shareholders exert

their influence to affect (and profit from) firms’ voluntary disclosure strategies. Our findings

suggest large shareholders influence managers of bad-news firms to delay disclosures when these

investors have high selling incentives, especially when litigation risk is low.

We focus on voluntary disclosures around IPO lockup expirations. Lockups are voluntary

agreements between the IPO firm and its underwriter prohibiting pre-IPO shareholders from

selling their stock for a contractually agreed period after the IPO. The lockup expiration provides

a powerful setting in which to identify large shareholders’ selling incentives and study their

effect on disclosure for the following reasons. First, large pre-IPO shareholders influence

managerial decisions through their ownership stakes, board membership, compensation

contracts, and relationships with management (Barry, Muscarella, Peavy, and Vetsuypens 1990;

Lerner 1995). Second, lockup expiration is associated with large-scale selling by pre-IPO

shareholders (Field and Hanka 2001). Third, lockup expiration dates are publicly known and

anticipated. This allows us to identify ex ante selling incentives of influential shareholders

around a specific date. Finally, newly-public firms are characterized by particularly high levels

Page 4: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

3

of information asymmetry between management and the firm’s dispersed shareholders, making

disclosure a powerful tool to influence the stock price at the time of selling.

Voluntary disclosure of private information affects the selling price for pre-IPO

shareholders through two channels. First, disclosure immediately affects the stock price when

news is released. As a result, pre-IPO shareholders favor releasing good news early and delaying

bad news in the lockup expiration quarter to maximize the stock price at the time of sale. Second,

disclosure affects information uncertainty about the firm, which, in turn, impacts how stock price

responds to selling (Kyle 1985). We refer to this phenomenon as “the price impact of selling.”

Recent empirical evidence suggests disclosure of bad news increases firm-specific uncertainty

(Brown, Hillegeist, and Lo 2009; Rogers, Skinner, and Van Buskirk 2009), which intensifies the

price impact of selling. Thus, to avoid increases in uncertainty and mitigate the price impact of

selling, pre-IPO shareholders have an added incentive to favor delaying the disclosure of bad

news until after they sell.

We study voluntary disclosure choices around IPO lockup expirations in the context of

quarterly management earnings forecasts. Management forecasts allow us to conduct precise

tests of the timing of a disclosure relative to an event because managers can preempt news

conveyed by quarterly earnings announcements by issuing forecasts during the quarter. As a

result, we are able to study whether managers accelerate or delay disclosures in the lockup

expiration quarter. Conditioning on whether earnings news is good or bad, we compare forecast

propensity in the lockup expiration quarter with that in benchmark quarters.

In both univariate and multivariate analyses, we find that, conditional on having bad

earnings news, the propensity to issue a forecast is approximately 36% lower in lockup

expiration quarters than in benchmark quarters. In contrast, managers do not accelerate

Page 5: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

4

disclosures of good news in lockup expiration quarters relative to benchmark quarters,

presumably because disclosure of good news is less credible (Hutton, Miller, and Skinner 2003)

and does not reduce firm-specific uncertainty (Rogers et al. 2009). Together, these results

suggest only bad-news firms choose a disclosure strategy that favors pre-IPO shareholders in the

lockup expiration quarter.

Selective disclosure in the lockup expiration quarter should vary with pre-IPO

shareholders’ ex ante selling incentives. We first study aggregate selling incentives, as captured

by predicted abnormal trading volume after the lockup expiration, and find firms are more likely

to delay disclosure in bad-news lockup expiration quarters as predicted trading volume increases

(i.e. as selling incentives increase). Next, because selling incentives and the ability to influence

disclosure strategy likely vary across shareholders, we shift our attention to the selling incentives

of different types of pre-IPO shareholders: VCs, managers, and all other pre-IPO shareholders,

hereafter the residual group. Prior literature suggests VCs both possess considerable influence

over managers and significantly reduce their ownership after the lockup expiration (Field and

Hanka 2001). Accordingly, we predict firms with a higher percentage of VC ownership locked-

up at the time of the IPO (i.e. stronger selling incentives) are more likely to delay disclosure in

bad-news lockup expiration quarters. Our findings are consistent with this prediction. In contrast,

our results suggest managers’ selling incentives do not affect disclosure strategy, consistent with

the evidence in Noe (1999) and Cheng and Lo (2006). Finally, we do not find a significant

relation between ownership of the residual group and disclosure choices, likely because of the

inherent heterogeneity of this group’s selling incentives and ability to influence disclosure

choices.

Page 6: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

5

We further expect the decision to delay disclosures of bad news to be shaped by litigation

risk and firm-specific uncertainty. Litigation risk can deter nondisclosure of bad news (Skinner

1994, 1997; Kasznik and Lev 1995; Field et al. 2005). Consequently, we predict bad-news firms

delay disclosures in response to high selling incentives only when the firm’s litigation risk is

relatively low. Our evidence supports this prediction. Similarly, we find bad-news firms delay

disclosures only when firm-specific uncertainty is high. This result is consistent with bad news

intensifying the price impact of selling through its effect on uncertainty. High uncertainty also

makes it easier for managers to delay disclosure without triggering scrutiny.

The use of ex ante proxies helps us to establish a causal relation between pre-IPO

shareholders’ selling incentives and strategic disclosure. In additional analyses, we test whether

forecast propensity is related to realized sales by readily identifiable pre-IPO shareholders: VCs

and managers. We find bad-news firms are more likely to delay disclosure if VCs sell a

significant fraction of their shares, but only when the managers do not also sell. This result is

likely driven by the asymmetry in SEC Rule 10b-5, which subjects managers but not VCs to

litigation risk associated with insider trading. Consequently, managers select disclosure strategies

that favor the selling incentives of large pre-IPO shareholders, but not when such strategies put

the managers themselves at risk. .

Delaying bad news disclosure benefits pre-IPO shareholders only if the adverse price

reaction to earnings news can be postponed until after these shareholders sell. Comparing stock

returns of bad-news firms with and without forecasts during the lockup expiration quarter, we

find two striking differences. First, non-disclosing firms delay the adverse stock price reaction

until earnings announcement (approximately -7%, on average) and therefore earn higher returns

during the lockup expiration quarter. Second, non-disclosing firms avoid any price impact of

Page 7: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

6

selling after lockup expiration while the price impact for forecasting firms is significantly

negative (approximately -4% at the average trading volume). Collectively, these results suggest

pre-IPO shareholders benefit from delayed disclosure of bad news.

Our study highlights how large investors’ short-horizon incentives affect managerial

disclosure choices. The findings suggest influential shareholders induce managers to undertake

disclosure strategies that do not result in direct benefits for the management, but instead profit

the shareholders. Specifically, we document firms are more likely to delay disclosure of bad

news when these shareholders possess strong selling incentives. Thus, some large and influential

shareholders are able to shape firms’ disclosure choices to their own benefit. This contrasts the

alternative strategy followed by transient institutional investors who invest in firms with higher

disclosure quality (Bushee and Noe 2000).

We also offer insights into how a previously-neglected cost of disclosure—exacerbating

firm-specific information uncertainty—affects disclosure strategies. Increases in uncertainty lead

to a greater price impact of selling after lockup expiration, providing an additional motive for

large shareholders to delay disclosures in bad-news lockup expiration quarters.

Finally, our paper contributes to the IPO literature by documenting that firms adopt

selective disclosure policies that favor pre-IPO shareholders. Gompers and Lerner (1998) suggest

VCs exit by distributing overvalued shares; our paper complements this finding by identifying a

mechanism—the disclosure or non-disclosure of information to the market—through which VCs

increase the value of their shares. Our findings have implications for other settings where large

active shareholders such as hedge funds and private equity investors could similarly influence

disclosure choices.

Page 8: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

7

The remainder of this paper proceeds as follows. Section 2 describes the institutional

background of our research setting and develops our hypotheses. Section 3 outlines our study’s

sample selection procedure. Sections 4 through 6 discuss our results, and Section 7 concludes.

2. Institutional background and hypothesis development

IPO firms typically enter into lockup agreements with their underwriters that restrict pre-

IPO shareholders from selling their shares for a specific period after the IPO.1 These restrictions

appear to be binding on average—trading volume spikes by 85% of previous average volume

when the lockup expires and eventually settles at approximately 40% higher than the lockup

period volume (Field and Hanka 2001; Bradley et al. 2001). The IPO literature attributes this

increase in trading volume to pre-IPO shareholders selling their shares for the first time.

The period following lockup expiration provides researchers with an ex ante proxy for

large shareholder selling incentives for at least two reasons. First, the lockup expiration date is

publicly known and specified in the IPO prospectus. Second, some large pre-IPO shareholders

sell a significant fraction of their shares in the lockup expiration quarter. For example, some VCs

distribute shares to their investors immediately upon lockup expiration (Gompers and Lerner

1998). These investors, in turn, sell their shares in the open market, generating abnormal trading

volume. As for non-VC backed firms, there is sustained abnormal trading volume for several

days following the lockup expiration (Field and Hanka 2001.

Pre-IPO shareholders benefit from executing sales at higher prices. While the lockup

date is publicly known, average abnormal returns following lockup expiration are negative—a

conundrum the finance literature has studied extensively.2 This drop in stock price erodes the

1 Brav and Gompers (2003) report lockup agreements in 99% of the firms in their sample of 2,871 IPOs. Most lockup periods are 180 days long. Field and Hanka (2001) find that the fraction of firms with a 180-day lockup period increased from 43% in 1988 to 91% in 1996. In our final sample of 776 IPO firms from 1995-2005, 92.5% of firms have the standard 180-day lockup period. 2 Prior studies examine several potential explanations for this price drop, including downward sloping demand curves, increase in bid-ask spreads, temporary price pressure from selling, and larger than expected sales by insiders (Bradley et al. 2001; Field and

Page 9: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

8

trading gains to the pre-IPO shareholders and creates particularly strong incentives for disclosure

strategies that allow these shareholders to sell shares at more favorable prices. Disclosing good

news early and delaying bad news until earnings announcements following the lockup expiration

will maximize trading gains for pre-IPO shareholders by altering investors’ assessments of firm

value (Ajinkya and Gift 1984; Waymire 1984; Lang and Lundholm 2000; Hutton et al. 2003;

Sletten 2012). Firms with bad news can delay disclosure without investors inferring the news as

long as the firms can pool with a group of good-news firms for which disclosure is particularly

costly, less credible, or whose managers possess only imprecise information and hence do not

disclose it (Verrecchia 1983; Dye 1985; Acharya, DeMarzo, and Kremer 2011).

Disclosure also affects investors’ uncertainty about firm value. While theory suggests

transparent disclosure reduces market uncertainty (Diamond and Verrecchia 1991), empirical

evidence contradicts this intuition, at least with respect to the disclosure of bad news (Brown,

Hillegeist, and Lo 2009; Rogers, Skinner, and Van Buskirk 2009).3 Rogers et al. (2009), for

example, find that, while good-news management forecasts do not have a significant effect on

uncertainty (measured by the implied volatility of stock options), bad-news forecasts lead to a

sustained increase in uncertainty until the subsequent earnings announcements. When coupled

with intense selling by pre-IPO shareholders, such increases in uncertainty can potentially hurt

stock prices. Specifically, theoretical models (e.g., Kyle 1985) predict a greater price impact of

selling when investors believe they are more informationally disadvantaged.4 Consistent with

Hanka 2001; Cao et al. 2004). While none of these fully explain the negative returns, Field and Hanka (2001) argue that abnormal returns around the unlock day are not large enough to be profitably arbitraged away. 3 Several explanations have been proposed in the finance literature for the increase in volatility after bad news events, a well-documented empirical finding. Black (1976) attributes it to the leverage effect resulting from the drop in price in response to the bad news. Christie (1982) and Schwert (1989) argue that this effect is small. Campbell and Hentschel (1992) propose a model for a volatility feedback explanation. Brown, Harlow, and Tinic (1988) find evidence consistent with this explanation when they document that returns to bad news tend to be larger than returns to good news. 4 In most microstructure models, the price impact of large trades arises from information asymmetry between informed investors and the price protection that less informed investors demand. In our tests, we do not distinguish between information asymmetry and overall uncertainty because of the empirical challenges of isolating the information asymmetry part of market uncertainty.

Page 10: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

9

this prediction, negative returns after lockup expirations are more pronounced among firms

characterized by greater information uncertainty (Bradley et al. 2001; Ofek and Richardson

2000). Therefore, delaying the disclosure of bad news until the next earnings announcement

mitigates the price impact of selling following lockup expiration and increases the trading gains

to pre-IPO shareholders.

These two effects lead to the following hypotheses:

Hypothesis 1A: Forecast propensity of firms with bad quarterly news is significantly lower in

the lockup expiration quarter than in quarters with normal levels of selling.

Hypothesis 1B: Forecast propensity of firms with good quarterly news is significantly higher

in the lockup expiration quarter than in quarters with normal levels of selling.

While, on average, pre-IPO shareholders possess strong selling incentives following the

lockup expiration, there are significant cross-sectional differences in the likelihood and the

magnitude of selling across firms. For example, Field and Hanka (2001) find firms with bigger

run-ups in stock price following an IPO, firms with VC-backing, and firms with more locked-up

shares experience greater selling. We expect the incentives to use selective disclosure to be

stronger among firms that, based on ex ante factors, are likely to experience high levels of selling

after lockup expiration. We hypothesize that:

Hypothesis 2: The strategy of delaying the disclosure of bad news and promptly releasing good

news in the lockup expiration quarter is more pronounced among firms with high aggregate

selling incentives.

Selling incentives vary not only across firms but also across different types of pre-IPO

shareholders. Because pre-IPO shareholders are restricted from selling their shares until after

Information asymmetry effects are likely to be high when there is greater uncertainty over firm value because the potential for private information is greater, and any incremental information potentially provides a significant advantage. Therefore, the overall uncertainty acts as an estimate of the upper bound of information asymmetry.

Page 11: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

10

lockup expiration, the higher the pre-IPO locked-up ownership of a specific shareholder, the

greater the number of shares available for selling or distribution after lockup expiration.

However, for these selling incentives to affect disclosure policy, pre-IPO shareholders must also

possess the ability to influence managers’ disclosure choices.

VCs are an easily identifiable group of pre-IPO shareholders who possess both significant

locked-up ownership at the time of the IPO and an ability to influence firms’ disclosure choices

before lockup expiration. Field and Hanka (2001) find VCs hold a significant fraction of shares

in IPO firms (23% on average) and, among large institutional or corporate shareholders, VCs are

the most likely to sell their shares in the year following the IPO. Further, VCs are typically active

shareholders; they not only finance the start-ups but also play a monitoring and advisory role and

are involved in strategic planning, managerial recruitment and training (Berlin 1998; Gorman

and Sahlman 1989; Hellmann and Puri 2000, 2002; Lerner 1995; Cadman and Sunder 2014).

Thus, VCs develop close ties to and command influence over management that passive

institutional investors are unlikely to enjoy. Consequently, we expect firms with a higher fraction

of locked-up shares held by VCs at the time of the IPO to be more likely to pursue a selective

disclosure strategy.

Managers are another group of pre-IPO shareholders who can benefit from strategic

disclosure choices. However, unlike VCs, whose distributions of shares are exempt from the

SEC’s Section 16 rules, managers face significant litigation risk stemming from SEC Rule 10b-5

and other insider trading regulation. Cheng and Lo (2006) argue that both disclosing good news

and withholding bad news prior to insider selling entail significant litigation risk. Consistent with

this argument, prior literature finds no evidence of firms strategically withholding bad news

before insider sales. In the case of good news, there is an association between disclosure and

Page 12: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

11

subsequent insider selling but this result is driven by managers timing trades following

disclosure of good news rather than strategically disclosing good news prior to their sales (Noe

1999; Cheng and Lo 2006). Thus, we do not expect the disclosure of good news and non-

disclosure of bad-news in the lockup expiration quarter to be related to shares locked held by

managers.

Various groups of shareholders other than VCs and management (e.g., corporate,

institutional, and angel investors, friends and family, as well as rank and file employees) also

likely have significant pre-IPO ownership in the firm. This residual group of shareholders is

heterogeneous with respect to selling incentives and ability to influence disclosure choices. We

therefore do not make a directional prediction about the relation between selective disclosure

behavior and the degree of locked-up ownership among the residual group of shareholders.

In summary, among various pre-IPO shareholders, VCs are a group of investors that is

clearly identifiable and has both strong incentives and the ability to influence the firms’

disclosure strategy. Consequently, we hypothesize that:

Hypothesis 3: The strategy of delaying the disclosure of bad news and promptly releasing good

news in the lockup expiration quarter is more pronounced for firms with higher locked-up VC

ownership at the time of the IPO.

Our hypotheses predict firms engage in strategic forecasting in response to ex ante selling

incentives. In additional analyses in Section 4, we explore the role of litigation risk and

uncertainty. Further, an extension of our hypotheses is, (i) conditional on the nature of the news,

ex post (realized) selling by pre-IPO shareholders is associated with firms’ disclosure choices,

and (ii) disclosure choices influence the timing of related stock returns (i.e., whether returns

Page 13: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

12

occur before or at the earnings announcement) as well as the price impact of trades after lockup

expiration. Sections 5 and 6 present the results associated with ex post selling.

3. Sample selection

Our sample period spans January 1995 through January 2008. We begin with firms in the

SDC database that conducted an IPO from January 1995 through December 2005. The lockup

expiration dates of these IPO firms fall from May 1995 through November 2006. We exclude

ADRs, unit offers, and firms that made secondary offerings during the lockup period. We merge

this dataset with CRSP and COMPUSTAT, which yields a sample of 2,973 firms. For these

firms, we obtain quarterly earnings announcement dates from COMPUSTAT and supplement

them with the dates from I/B/E/S when not available on COMPUSTAT.

For each of our sample firms, we next identify the announcement quarter in which the

lockup expires and, because this is the primary event quarter, we require the availability of

necessary data for this quarter.5 This restriction results in a sample of 2,308 firms. We then

extend this sample to include the quarter immediately preceding the lockup expiration quarter

and four subsequent quarters.6

We require the lockup expiration to be distinct from the earnings announcement and,

therefore, exclude firms for which the lockup expiration coincides with or falls within the 10

days preceding the earnings announcement. This restriction also ensures the pre-IPO

shareholders have sufficient time to trade after the lockup expires and before information is

released through the earnings announcement. We also exclude firms with lockup expirations

5 Announcement quarter is defined as the period starting on the day of the quarterly earnings announcement for quarter t-1 and ending on the day before the quarterly earnings announcement for quarter t. 6 Since the maximum length of a quiet period after the IPO in our sample period is 40 days, and the majority of IPO firms have lockups extending for 180 days, starting the sample only with the quarter immediately preceding the lockup expiration quarter alleviates concerns regarding the potential impact of the quiet period on forecast propensity.

Page 14: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

13

falling within the first ten days of the quarter because this period serves as our estimation

window for quarterly news. These filters result in a sample of 1,705 firms.

Finally, we require all firm-quarters in our sample to have the requisite data to compute

control variables. Among these, an important restriction pertains to our ability to compute

quarterly news: we only use firm-quarters with at least one one-quarter-ahead analyst forecast

issued in the first ten days of the quarter. We do this because we compute the news in the quarter

as the difference between the actual earnings announced at the end of the quarter and the

consensus of analysts’ forecasts of these earnings issued at the beginning of the quarter (see

Appendix A for a detailed definition of all variables). While requiring the availability of

analysts’ forecasts reduces our sample size, it allows us to capture the specific news pertaining to

quarterly earnings and produces a timelier estimate of quarterly news than alternatives (such as

earnings that follow a random walk) would permit. The final sample with required data available

consists of 776 unique IPO firms comprising 3,126 announcement quarters .

Table 1 provides descriptive statistics and univariate tests of differences between firm-

quarters with and without management forecasts. Mean analyst following is 4.74, suggesting a

reasonable level of analyst interest in our sample firms. Mean institutional ownership is 36%,

which is of a similar order of magnitude to descriptive statistics reported by Field and Lowry

(2009). Our sample firms are growth firms—mean Market-to-Book ratio is 4.42—with low levels

of profitability—the mean Return on Assets is -0.03. The magnitude of news seems to matter in

the disclosure decision: the magnitude of bad news is greater and that of good news is smaller in

firm-quarters with a forecast relative to firm-quarters without a forecast. Further, analyst

following is higher in firm-quarter observations where the firm provides a forecast. Finally, firm-

Page 15: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

14

quarter observations where firms provide forecasts are characterized by significantly higher

levels of litigation probability.

4. Firms’ propensity to issue management forecasts in the lockup expiration quarter

4.1 Is there evidence of selective disclosure in the lockup expiration quarter?

To test Hypotheses 1A and 1B, we compare the probability of forecasting during the

lockup expiration quarter to the probability of forecasting during the subsequent four quarters in

sub-samples of firm-quarters with good and bad news. To the extent managers receive news

early, they can voluntarily disclose it through a management forecast, or they can wait until the

end of the quarter to reveal the news at the earnings announcement.7 To measure the news during

the quarter and capture forecasting behavior, we divide each announcement quarter into two

periods. We measure the market’s expectations of earnings as the mean of analysts’ forecasts

issued over the first 10 days of the announcement quarter (news estimation window). We

categorize quarters where the actual earnings are less than (greater than or equal to) the market’s

expectations of earnings as Bad News (Good News) quarters.8,9 We use the remainder of the

quarter, beginning 11 days after the previous earnings announcement (forecast window), to

collect management forecasts, focusing on quantitative management forecasts that pertain to the

current quarter’s earnings per share (EPS) issued during the forecast window as captured by the

First Call Company Issued Guidelines Database (see Figure 1 for a timeline).

7 We focus exclusively on management forecasts as the channel through which managers communicate earnings news. Managers could provide information about components of performance that have implications for the current quarter earnings without providing direct guidance on earnings (e.g., through press releases about contracts won or lost). Analysts and investors also could partially infer earnings news that arises from industry- or economy-wide shocks without any forecasts. However, management forecasts are an important component of a firm’s disclosure strategy, particularly with respect to conveying earnings news (Beyer et al. 2010). Abstracting from other information channels is unlikely to introduce any systematic bias into our analyses. 8 Our Bad News and Good News variables are similar in spirit to the “expectations gap” measure in Kasznik and Lev (1995). Similar to our approach, Kasznik and Lev (1995) use the first 30 days of the announcement quarter to measure the market’s expectations of earnings and focus on the management forecasts issued in the remainder of the announcement quarter. 9 In untabulated tests, we confirm that the news conveyed in the forecast corresponds to the underlying quarterly news. Specifically, in good-news quarters, the median forecast news (measured as earnings forecast less analyst consensus over the first 10 days of the quarter, scaled by price) is 0.002, and the forecast news in bad-news quarters is -0.004.

Page 16: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

15

Table 2 presents the results of our univariate tests. Approximately 31% (=239/(239+537))

of Lockup Expiration and 37% (=723/(723+1,246)) of Post-Lockup Expiration firm-quarters

have bad news during the quarter. Of these bad-news Lockup Expiration firm-quarters 11.7%

have a quantitative EPS forecast compared with 4.5% of good-news Lockup Expiration firm-

quarters, consistent with evidence in the prior literature that bad-news firms are more likely to

disclose management forecasts in general (Kasznik and Lev 1995). More importantly, the mean

forecast propensity for firm-quarters with bad news is lower in the lockup expiration quarter than

in the post-lockup-expiration quarters: 11.7% versus 18.7%, with the difference significant at the

5% level based on two-sided t-tests. In contrast, the mean forecast propensity for good-news

firm-quarters is not significantly different between the lockup expiration and the post-lockup

expiration period: 4.5% versus 6.2%. Our sample of good-news quarters includes firm quarters

with no news (earnings realizations equal to beginning-of-quarter estimates). When we repeat

our tests after excluding no-news firm-quarters, we continue to find no support for Hypothesis

1B.10 Taken together, these findings are consistent with Hypothesis 1A but not consistent with

Hypothesis 1B. The lack of support for Hypothesis 1B is likely driven by the lower credibility of

good news (Hutton et al. 2003). Problems with credibility are likely to be particularly acute for

young IPO firms given just a brief or no forecasting history. Consistent with this explanation, we

find average returns to good-news forecasts in the lockup quarter are 0.3% and are statistically

indistinguishable from zero.

Table 2 also presents mean forecast propensity for pre-lockup-expiration quarters: 16.8%

for bad news and 4.6% for good-news quarters.11 Because forecasts are more frequent in the pre-

10 In further robustness analyses we reclassify quarters with forecasts that convey news inconsistent with the firm’s quarterly earnings news as no-forecast quarters and our inferences are unchanged. 11 We do not use these quarters that are entirely covered by the lockup provisions as a benchmark against which to evaluate disclosure strategies in the lockup quarter for two reasons. First, managers likely adjust their forecasting behavior during these quarters (e.g., by refraining from providing any forecasts) to have flexibility with respect to disclosure strategy in the lockup

Page 17: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

16

lockup quarter than in the lockup expiration quarter, a time-trend in forecasting propensity or an

increase in First Call coverage over a firm’s life are unlikely to explain our results (Chuk,

Matsumoto, and Miller 2013).12 We repeat the analysis in Table 2 for a constant sample of firms

for which we have observations in the pre-lockup, lockup-expiration and all four post-lockup

expiration quarters (not tabulated). We find that forecast propensity is stable across time for good

news quarters, ranging between 7.69% and 8.5%. In contrast, for firms with bad news, forecast

propensity drops from 20% in the pre-lockup period to 12.5% in the lockup-expiration quarter

and then increases to 19% in the post-lockup expiration quarters. Overall, we find strong

univariate evidence that managers delay disclosure of bad news in lockup expiration quarters.13

We also conduct multivariate analyses in which we estimate forecast propensity

separately for good and bad-news samples over event time (results not tabulated). We compare

forecasting behavior in the lockup expiration quarter with the subsequent quarters, after

controlling for other factors that may affect the propensity to issue forecasts (the same set of

control variables that we use in our cross-sectional tests described in Section 4.2.1). The results

of these multivariate analyses are similar to the univariate tests reported in Table 2. In particular,

we find the likelihood of disclosure is significantly lower during the lockup expiration quarter

relative to the post-expiration quarters. The results are economically significant—setting control

variables to their means, we document a 36% reduction in forecast propensity in the lockup

quarter. This is because management forecast practices are sticky—once a firm initiates quarterly earnings forecasts, investors expect the firm to continue forecasting on a quarterly basis (Bhojraj, Libby, and Yang 2011; Chen, Matsumoto, and Rajgopal 2010). Second, selling by pre-IPO shareholders is restricted in the pre-lockup expiration quarters. This alters both the benefits and the costs of selective disclosure—while there are no private gains to withholding bad news (decreasing the benefits of selective disclosure), litigation risk is likely lower in the absence of insider selling (decreasing the costs of selective disclosure). 12 Chuk et al. (2013) show that First Call coverage is not always complete but coverage limitations are not a significant concern for quantitative short-horizon EPS forecasts and for firms with high analyst following (i.e., the types of forecasts and firms that we study). Nevertheless, following their recommendation, we repeat our analysis from Table 2 on a subsample of quarters falling in or after 1998, when the First Call coverage is significantly broader, and find similar results. 13 Firms do not seem to alter the timing of forecasts within the quarter in the lockup expiration quarter relative to other quarters. In particular, firms in bad- (good-) news quarters in our sample issue their forecasts on average on the 57th (59th) day of the quarter which is statistically indistinguishable from the respective averages for the post-lockup quarters.

Page 18: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

17

expiration quarter (statistically significant with p-value less than 0.05) for bad news firms. In

contrast, in the sample of good news firms, the likelihood of disclosure is not different during the

lockup expiration quarter relative to post-expiration quarters.

A potential concern with the tests of Hypotheses 1A and 1B is attrition in sample size

because of the data requirements to calculate the proxy for news during the quarter. We examine

the generalizability of our results to a broader sample using an alternative measure of news based

on market-adjusted returns over the window starting two days after the previous quarter earnings

announcement and ending one day after the current quarter earnings announcement. We classify

quarters where the excess returns, as defined above, are less than (greater than or equal to) zero

as bad (good) news quarters. This returns-based news measure yields results consistent with our

hypotheses (not tabulated). The drawback of the returns-based measure is that it captures both

news pertaining to the quarter’s earnings and to other news about future earnings and growth

prospects. Thus, unlike the analyst-forecast-based measure of news that we use in our reported

analysis, the returns-based measure does not correspond well to the earnings news firms convey

via management forecasts.

4.2 Does selective disclosure behavior vary with ex ante selling incentives in the lockup

expiration quarter?

4.2.1 Main tests

Hypothesis 2 predicts selective disclosure to be more pronounced among firms with

higher aggregate selling incentives. To capture aggregate selling incentives, we estimate a model

for abnormal volume, adapted from Field and Hanka (2001), in which we rely on ex ante factors

to explain the abnormal trading volume after lockup expiration. We find VC-backing, the

percentage of shares locked up, and stock price run-up to be significant determinants of selling

Page 19: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

18

(see Appendix B for results).14 The fitted values from this model, Predicted Abnormal Volume,

serve as our proxy for aggregate selling incentives. Because our focus is the effect of cross-

sectional differences in selling incentives after lockup expiration on disclosure, we restrict the

sample to the lockup expiration quarters and estimate the following model separately for good-

and bad-news firms:

Forecast = �0 + �1 Predicted Abnormal Volume + �2-11 Controls + Year Fixed Effects + �������a�

Per Hypothesis 2, we expect the coefficient on Predicted Abnormal Volume, �1, to be

positive in the subsample of good-news firms and negative in the subsample of bad-news firms.

We further test whether disclosure varies with ex ante selling incentives of three groups

of pre-IPO shareholders: VCs, managers, and the residual group of shareholders. Hypothesis 3

predicts that strategic disclosure behavior will vary with the locked-up ownership of VCs. As

discussed in section 2, we do not expect a significant relation between disclosure and the locked

shares owned by managers because of substantial litigation risk. We remain agnostic with respect

to the residual group. In the sample of lockup expiration quarters, we estimate the following

model separately for good- and bad-news firms:

Forecast = �0 + �1 VC Ownership Locked % + �2 Manager Ownership Locked % +�3 Residual Ownership Locked % +�4-13 Controls + Year Fixed Effects + � (1b)

We measure VC Ownership Locked %, Manager Ownership Locked %, and Residual

Ownership Locked % as the percentage of ownership at the time of the IPO by VCs, managers

and the residual group of pre-IPO shareholders, respectively. Ownership by pre-IPO

shareholders is not available in machine-readable form. We therefore hand collect ownership

figures from the IPO prospectuses. We proxy for percentage of locked-up shares held by each

14 While trading volume is typically not synonymous with selling incentives, in the context of lockup expiration the additional trading volume has been attributed to selling of previously locked up shares. We base our model on factors that Field and Hanka (2001) find to be significant determinants of abnormal trading volume after lockup expirations.

Page 20: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

19

group with the percentage of shares owned by respective groups at the time of the IPO.

Consistent with Hypothesis 3, we expect the coefficient on VC Ownership Locked, �1, to be

positive in the subsample of good-news firms and negative in the subsample of bad-news firms.

We include various control variables in Eqs. (1a) and (1b) motivated by prior literature.

First, to assess whether forecasting propensity increases with the magnitude of the news, we

include News Magnitude, the absolute value of the difference between actual earnings and

analysts’ expectations of earnings scaled by the beginning of the quarter price.

The second set of control variables captures the presence of key market participants that

can influence management forecast propensity. We control for Analyst Following because

disclosure quality is related to analyst following (Lang and Lundholm 1996; Healy, Hutton, and

Palepu 1999) and small, insider-dominated firms use disclosure to attract analysts (Graham,

Harvey, and Rajgopal 2005). We include institutional ownership (% of Institutional Ownership),

because the presence of institutions is usually positively associated with disclosure decisions

(Bushee and Noe 2000; Ajinkya, Bhojraj, and Sengupta 2005).

The third set of control variables captures firm characteristics identified in previous

literature as related to the decision to issue a management forecast. Litigation risk is an important

factor that increases the conditional likelihood of disclosing bad news (Skinner 1994; Field,

Lowry, and Shu 2005). We proxy for litigation risk using Litigation Probability, the fitted value

from the litigation probability model in Rogers and Stocken (2005), who focus on litigation

under SEC Rule 10b-5.15 We also include in the regressions firm size, market-to-book ratio,

15 We confirm the validity of using the Rogers and Stocken (2005) measure for our setting by examining the incidence of securities class-action lawsuits involving our sample firms over the two years following their IPOs. We find that the frequency of lawsuits related to disclosure choices for the above and below median litigation probability samples are 22.4% and 7.47%, respectively. The difference is statistically significant, suggesting that the Rogers and Stocken (2005) model successfully captures litigation probability for our sample of newly public firms. Our results are also robust to using an alternative measure of litigation risk introduced by Francis, Philbrick, and Schipper (1994), which is based on industry classification.

Page 21: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

20

profitability, financial health, and uncertainty (Kasznik and Lev 1995; Bamber and Cheon 1998;

Miller 2002; Ajinkya, Bhojraj, and Sengupta 2005; Hribar and Yang 2010).

Finally, we control for the regulatory environment. Specifically, given the impact of

Regulation Fair Disclosure (hereafter “FD”) on firms’ voluntary disclosures (Heflin,

Subramanyam, and Zhang 2003; Mohanram and Sunder 2006), we control for the shift in

regulatory regime with an indicator variable, After Regulation FD, equal to one if the

announcement quarter ends after the enactment of Regulation FD (October 23, 2000) and zero

otherwise. We also include year fixed effects.

Table 3 reports the results from estimating Eqs. (1a) and (1b). Models 1a and 1b report

the results with Predicted Abnormal Volume as a proxy for aggregate selling incentives of pre-

IPO shareholders. Models 2a and 2b focus on selling incentives of three separate groups of

shareholders captured by VC Ownership Locked %, Manager Ownership Locked %, and

Residual Ownership Locked. When Predicted Abnormal Volume is the variable of interest, we

use all lockup quarter observations. However, models 2a and 2b require ownership data and we

are able to retrieve it only for 571 firms (173 firms with bad news in the lockup expiration

quarter and 398 firms with good news). The attrition is mostly due to the fact that some of our

sample firms had IPOs in 1995 or early 1996 and do not have filings available on the EDGAR

database.

As expected, in the case of firms with bad news in the lockup expiration quarter, the

propensity to delay the news is increasing in both aggregate selling incentives and in the selling

incentives of large, influential shareholders (VCs). The coefficients on Predicted Abnormal

Volume (Model 1a) and VC Ownership Locked % (Model 2a) are negative, and the

corresponding marginal effects are significant, with p-values less than 0.05. The marginal effects

Page 22: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

21

suggest that a one standard deviation increase in Predicted Abnormal Volume (VC Ownership

Locked %) reduces the probability of forecasting by approximately 59% (17%) of the average

forecasting probability in the lockup expiration quarter. Manager Ownership Locked % and

Residual Ownership Locked % are not significantly associated with forecast propensity in the

lockup expiration quarter. In summary, strategic disclosure behavior documented in Table 2 is

more pronounced among bad-news firms with stronger aggregate selling incentives, as well as

firms with greater VC ownership.

As for firms with good news in the lockup expiration quarter, the propensity to issue a

management forecast is not associated with selling incentives as captured by Predicted Abnormal

Volume and VC Ownership Locked %. This finding is consistent with results in Table 2 that, for

good news firms, disclosure behavior is not significantly different in lockup expiration quarters

than in subsequent quarters with normal levels of selling incentives.

4.2.2 Variation with litigation risk and uncertainty

While the incentives to withhold bad news increase in selling incentives, prior literature

suggests firms with bad quarterly news consider additional factors in deciding whether to

disclose the news. These factors include litigation risk (Skinner 1994, 1997; Kasznik and Lev

1995; Field et al. 2005) and firm-specific uncertainty (Rogers et al. 2009). Wary of legal

consequences of nondisclosure, managers in firms subject to high litigation risk may be more

reluctant to delay the disclosure of bad news. Litigation risk matters especially to IPO firms,

which can be sued under Section 11 as well as Section 10b-5 of the Securities Exchange Act of

1934. Section 11 relates to disclosures made in the IPO prospectus (Lowry and Shu 2002), while

Section 10-b5 covers subsequent disclosures and bans any act or omission resulting in fraud or

deceit in connection with the purchase or sale of securities. Consequently, we expect the delays

Page 23: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

22

in disclosure of bad news to occur only at firms with high selling incentives among pre-IPO

shareholders and low firm-specific litigation risk.

We explore the role of litigation risk by estimating variants of Eq. (1a) and (1b). In

particular, we examine the effect of selling incentives separately for firms with high and low

litigation risk during the lockup expiration quarter by splitting Predicted Abnormal Volume and

VC Ownership Locked % into two mutually exclusive variables that capture selling incentives for

either low or high litigation risk groups.16 We expect the negative relation between likelihood of

bad-news disclosure and selling incentives to be driven by firms with low litigation risk

(Predicted Abnormal Volume x Low, VC Ownership Locked % x Low).

The results are reported in Table 4, Models 1a and 2a. They indicate firms with high risk

of litigation do not alter their disclosure patterns even when selling incentives are high; the

coefficients and marginal effects (not reported) on Predicted Abnormal Volume x High and VC

Ownership Locked % x High are not significantly different from zero. In contrast, firms that face

low risk of litigation are less likely to disclose bad news during the lockup-expiration quarter

when selling incentives are high. The coefficient and marginal effects on Predicted Abnormal

Volume x Low and on VC Ownership Locked % x Low are negative and significant, with p-values

less than 0.01 and 0.05, respectively. These results underscore that litigation risk deters firms

from withholding bad news (Skinner 1994; Field et al. 2005).

Firm-specific uncertainty may also affect decisions to delay the disclosure of bad news.

To the extent that high uncertainty translates into greater information asymmetry between

managers and dispersed shareholders, managers’ ability to delay disclosure of bad news until

earnings announcements will be stronger in firms with high uncertainty. More importantly for

16 We tabulate results in which high and low litigation risk and uncertainty are defined as above and below the mean respectively. However, the inferences in this section are unchanged if we define the high versus low groups based on medians.

Page 24: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

23

our setting, greater uncertainty is associated with more pronounced negative returns after lockup

expiration (Bradley et al. 2001; Ofek and Richardson 2000), creating even stronger incentives for

selective disclosure that mitigates these returns. Further, the negative effect of uncertainty on

lockup expiration returns should be particularly pronounced when coupled with intense selling

(Kyle 1985). Considering these combined effects, we expect that selective disclosure strategy

will be concentrated among firms with high selling incentives and high uncertainty.

We estimate expanded versions of Eq. (1a) and (1b), where we split the two proxies for

ex ante selling incentives into two groups based on the level of uncertainty at the start of the

quarter. We measure uncertainty as the standard deviation of daily stock returns in the quarter

preceding the lockup expiration quarter. Model 1b includes Predicted Abnormal Volume x High

and Predicted Abnormal Volume x Low and Model 2b includes VC Ownership Locked % x High

and VC Ownership Locked % x Low. We expect the negative relation between likelihood of

disclosure and selling incentives to be driven by bad-news firms with high uncertainty.

Table 4 Models 1b and 2b present the results. The coefficient and the marginal effect on

Predicted Abnormal Volume x High are negative and significant, with p-values less than 0.01,

while the coefficient and marginal effect on Predicted Abnormal Volume x Low do not differ

significantly from zero. Similarly, the coefficient and the marginal effect on VC Ownership

Locked % x High are negative and significant, with p-values less than 0.05. Surprisingly, the

coefficient of VC Ownership Locked % x Low is also negative, but it is only marginally

significant (at 10% level). Overall, the results in Models 1b and 2b are consistent with firms not

wanting to exacerbate the already high level of uncertainty with bad-news disclosures when they

expect intense selling after lockup expiration.

Page 25: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

24

To summarize, the results in Tables 3 and 4 show that the nondisclosure of bad news is

more pronounced when the aggregate selling incentives and selling incentives of VCs are

greater. This effect is economically significant and is concentrated among firms subject to either

low litigation risk or high uncertainty.

5. Is strategic disclosure behavior associated with higher realized sales?

Unlike the ex ante proxies for selling incentives, realized sales cannot be used to draw

causal inferences about the drivers of the selective disclosure behavior (Noe 1999; Cheng and Lo

2006). Nevertheless, examining the association between realized sales by pre-IPO shareholders

and strategic withholding of bad news is a natural extension of our hypotheses. If selective

disclosure in the lockup expiration quarter enables pre-IPO shareholders to reduce their holdings

at favorable terms, we should observe greater realized sales by large shareholders of the firms

that did not forecast their bad news.

We study the sales of VCs, managers, and the residual shareholders as a function of news

and disclosure behavior in a joint framework using the following model:

Forecast = �0 + �1 Venture Capital High Sell + �2 Manager Sell +�3 Manager Sell x Venture Capital High Sell +�4 Residual Abnormal Volume +�5-14 Controls + Year Fixed Effects + �

(2)

We estimate the above regression in the lockup expiration quarter separately for good-

news and bad-news firm-quarters. Venture Capital High Sell is an indicator variable that captures

above average decrease in VC ownership after lockup expiration. Manager Sell is an indicator

variable that takes the value of one if there are net sales by officers of the firm in the lockup

expiration quarter and zero otherwise.17 We rely on Thompson Reuters for data on trading by

officers and hand-collect VC ownership from IPO prospectuses and post-IPO proxy statements

17 In our sample, managers own shares in the vast majority of IPO firms (83%) but sell shares only in 19% of firms, in line with the 17% reported by Field and Hanka (2001). In bad-news quarters, the fraction of IPO firms with managers selling after lockup expiration is even lower: approximately 11%.

Page 26: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

25

(Appendix C describes the data collection and the issues regarding the nature of the VC

ownership data.) We also construct a variable to approximate selling by the residual group of

pre-IPO shareholders, Residual Abnormal Volume, defined as the aggregate abnormal volume

over the trading window less the change in ownership by VCs and sales by managers.

Table 5 reports the results of the estimation. The likelihood of forecasting is significantly

lower for bad-news firms where VCs sell or distribute a substantial fraction of their shares and

managers of the firm do not sell (i.e., Venture Capital High Sell equals one, and Manager Sell

equals zero). In contrast, when both VCs and managers sell, the likelihood of forecasting bad

news is significantly higher than in quarters with only VC selling. We implement the correction

suggested by Norton, Wang, and Ai (2004) to interpret the significance of the interaction term,

Manager Sell x Venture Capital High Sell. We find the z-statistic on the interaction term is

always positive, and the significance exceeds conventional thresholds for most of the range of

predicted probabilities (0.3 and higher; results not tabulated). These results suggest that

managers delay disclosures of bad news, not for their own benefit, but to enable VCs to sell their

shares at better prices. Even though managers sell relatively infrequently in our setting, when

they do sell, they also disclose bad news to avoid litigation risk stemming from SEC Rule 10b-5.

Thus, managers adjust disclosure policies to benefit large shareholders only when such actions

do not put the managers themselves at risk. We find no association between Residual Abnormal

Volume and the propensity to issue a forecast in bad-news or good-news lockup expiration

quarters. Overall, the evidence in Table 5 strengthens the conclusions from the tests of primary

hypotheses.

Page 27: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

26

6. Disclosure strategy and stock price consequences

We next examine the stock price consequences of disclosure choices to determine

whether pre-IPO shareholders of bad-news firms benefit from delaying disclosure until earnings

announcements. We expect two types of benefits. First, by delaying disclosure, firms with bad

news postpone negative stock returns until the subsequent earnings announcements, allowing

pre-IPO shareholders to secure a higher sale price. Second, delayed disclosure is likely to help

bad-news firms mitigate the negative price impact of selling following the lockup expiration.

The use of non-disclosure to postpone negative returns until the earnings announcement,

when the pre-IPO shareholders have already exited or reduced their stock ownership, has two

implications. Compared with bad-news firms that disclosed the news early, bad-news firms that

delay disclosure should experience (i) less negative returns during the lockup expiration quarter

leading up to the earnings announcement and (ii) more negative returns as the underlying news

becomes public at the time of the earnings announcement. We estimate the following regression

to test these predictions for the subset of bad-news firms:

Abnormal Returns = �0 +�1 No Forecast + �2 News Magnitude + Year and Industry Fixed Effects + �

(3)

We measure Abnormal Returns over (i) the lockup quarter, starting after the news

estimation window (11 days after the previous quarter earnings announcement) and ending two

days before the current quarter earnings announcement (Quarter Excluding Earnings

Announcement), and (ii) the three-day window around the earnings announcement after the

lockup expiration (Earnings Announcement Window) (see Figure 1). The variable of interest is

No Forecast, an indicator variable that equals one if the firm does not issue a forecast during the

lockup expiration quarter and zero otherwise. This variable captures the incremental effect on

Page 28: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

27

stock returns for firms that delayed disclosure relative to firms that issue a forecast during the

quarter. We include News Magnitude because it is likely to affect returns during the quarter.18

Table 6, Model 1, reports the results over the Quarter Excluding Earnings

Announcement. As expected, the coefficient on No Forecast is positive and significant at the

10% level, suggesting bad-news firms that stay silent during the quarter experience higher

returns relative to bad-news firms that forecast earnings. This is largely driven by the strong

negative reaction to forecasts released by bad-news firms; average abnormal returns on the three

days centered on the forecast announcement date are significantly negative (-21.1% with two-

sided t-statistic of -4.36).19

Model 2 reports the results over the Earnings Announcement Window. In this window,

we expect firms with bad news that stayed silent during the quarter to experience a more

negative reaction to the reported earnings news than do firms that already released their news

through a forecast. Consistent with our expectations, the coefficient on No Forecast is negative

and significant at the 10% level. The magnitude of the coefficient indicates that nonforecasting

firms experience on average 7.1% lower earnings announcement returns than do forecasting

firms. Overall, we find support for the notion that, by delaying disclosure until earnings

announcements, bad-news firms shift related negative returns to the earnings announcements and

enable pre-IPO shareholders to secure a higher sale price after lockup expiration.

18 We conduct two untabulated tests to address the concern that characteristics of firms that choose to issue forecasts lead to differences in returns (Kasznik and Lev 1995, Tucker 2007). First, we confirm that the magnitude of news between forecasting and nonforecasting bad-news firms is not significantly different from each other. Second, we estimate an alternative model with all the control variables from Table 3. The results of this additional analysis are qualitatively similar to the results in Table 6. 19 The magnitude of the response to bad news is larger than what has been documented previously. For example, Hutton et al. (2003) report an average three-day return of -9.9% in response to bad-news forecast announcements. However, given our sample of young IPO firms with high growth opportunities, we expect a greater price reaction to news, consistent with the findings in Collins and Kothari (1989) and Lang (1991).

Page 29: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

28

Next, we investigate the second potential benefit of delayed disclosure to pre-IPO

shareholders of bad-news firms—reduced price impact of selling following lockup expiration. To

test for the price impact of selling, we estimate the following model:

Abnormal Returns (Trading Window) = �0 +�1 No Forecast + �2 News Magnitude + �3 Abnormal Volume + �4 Abnormal Volume x No Forecast + Year and Industry Fixed Effects + �

(4)

Abnormal Returns (Trading Window) are the market-adjusted buy-and-hold returns over

the period starting on the lockup expiration date and ending two days before the earnings

announcement. Abnormal Volume is the abnormal trading volume over the same period. The

coefficient on Abnormal Volume, �3, captures the price impact of selling for the group of

forecasters while the interaction term with No Forecast allows us to examine the incremental

price impact for the group of firms that did not forecast. Therefore, �3+�4 captures the overall

price impact for the firms that do not forecast. In the sample of firms with bad news, we expect

�3 to be significant and negative and �3+�4 to be statistically indistinguishable from zero or at

least have a lower magnitude than �3.

The results in Table 6, Model 3 indicate significant negative price impact of selling for

bad-news firms that release their news early via a forecast experience—the coefficient on

Abnormal Volume, �3, is -0.259 and significant at the 5% level. This price impact translates to

more than 4% negative returns at the average abnormal trading volume. This price impact,

however, is mitigated for firms that delayed the disclosure until earnings announcement; the

coefficient on Abnormal Volume x No Forecast, �4, is 0.188. The sum of �3 and �4 is not

statistically different from zero (untabulated �2 test), suggesting that nondisclosing bad-news

firms do not experience any adverse price impact of selling.

Page 30: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

29

Taken together, these results suggest that delaying disclosure of bad news until an

earnings announcement has two effects on returns. First, firms postpone the price drop associated

with the release of bad news. Second, the delay mitigates the adverse price impact of selling after

the lockup expiration date. Comparing the earnings announcement returns and the price impact

of selling for the forecasting and nonforecasting firms, we estimate that by delaying disclosure,

firms help pre-IPO shareholders enjoy a sale price that is approximately 11% higher, on average.

7. Conclusion

We examine whether large and influential pre-IPO shareholders with high selling

incentives impact firms’ voluntary disclosures, leading firms to manage investor perceptions of

firm value and mitigate uncertainty-related stock price movements after lockup expirations. We

find that, for firms with bad quarterly news, the propensity to issue an earnings forecast is

significantly lower in the lockup expiration quarter than in later quarters, a strategy that lets pre-

IPO shareholders sell shares or exit the firm at favorable prices. The likelihood of delaying bad

news in the lockup expiration quarter is higher when the aggregate selling incentives and selling

incentives of VCs are greater, especially when litigation risk is low and firm-specific uncertainty

is high. In contrast, the selling incentives of managers and other pre-IPO shareholders do not

affect disclosure choices. Finally, supplemental tests of realized sales reveal that bad-news firms

are more likely to delay disclosure if VCs sell significant quantities of shares but only when

managers do not also sell. Thus, consistent with prior literature, managers seem wary of

litigation risk and do not personally profit from strategic nondisclosure, but rather engage in

strategic non-disclosure to benefit influential pre-IPO shareholders such as VCs.

Our analysis of abnormal returns in the lockup expiration quarter suggests that, by

delaying disclosure, bad-news firms postpone the negative returns associated with releasing the

Page 31: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

30

news until the earnings announcement and avoid the adverse price impact of selling after lockup

expiration. This enables pre-IPO shareholders to sell their shares at higher prices.

Overall, our findings are consistent with the selling incentives of large influential

shareholders shaping disclosure choices. Our results indicate that while managers do not engage

in strategic disclosures before their own trades, they do delay disclosures of bad news to benefit

other large shareholders. This implies disclosures are not always made in a timely manner,

allowing large, influential shareholders to profit at the expense of other investors. Our evidence

from lockup expirations provides insights into other settings in which large shareholders such as

hedge funds and private equity investors have incentives and ability to influence the flow of

information to dispersed shareholders.

Page 32: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

31

References

Acharya V.V., P. DeMarzo, and I. Kremer, 2011. Endogenous information flows and the clustering of announcements. American Economic Review, 101(7): 2955–2979.

Ajinkya, B., S. Bhojraj, and P. Sengupta, 2005. The association between outside directors, institutional investors and the properties of management earnings forecasts. Journal of Accounting Research 43: 343-374.

Ajinkya, B. and M. J. Gift, 1984. Corporate managers’ earnings forecasts and symmetrical adjustments of market expectations. Journal of Accounting Research 22: 425-444.

Bamber, L.S., and Y.S. Cheon, 1998. Discretionary management earnings forecast disclosures: Antecedents and outcomes associated with forecast venue and forecast specificity choices. Journal of Accounting Research 36: 417-441.

Barry, C.B., C.J. Muscarella, J.W. Peavy, III, and M.R. Vetsuypens, 1990. The role of venture capital in the creation of public companies: evidence from the going public process. Journal of Financial Economics 4: 447-471.

Berlin, M., January/ February 1998. That thing venture capitalists do. Business Review. 15-26.

Bhojraj, S., R. Libby, and H. Yang. 2011. Analyzing guidance at the firm level: The effect of reputation-building and learning-by-doing on guidance frequency and guidance properties. Working Paper Cornell University and University of Pennsylvania.

Black, F., 1976. Studies of stock, price volatility changes. Proceedings of the 1976 meetings of the Business and Economics Statistics Section. American Statistical Association: 177-181.

Bradley, D. J., B. D. Jordan, I.C. Roten, and H. Yi, 2001. Venture capital and IPO lockup expiration: An empirical analysis. Journal of Financial Research 14: 465-492.

Brav, A. and P. A. Gompers, 2003, The role of lockups in initial public offerings. Review of Financial Studies 16: 1-29.

Brav, A., W. Jiang, F. Partnoy, and R. Thomas, 2008. Hedge fund activism, corporate governance, and firm performance. Journal of Finance 63: 1729-1775. Brown, K. C., W.V. Harlow, and S. M. Tinic, 1988. Risk aversion, uncertain information, and

market efficiency. Journal of Financial Economics 22: 355-385.

Brown, S., S.A. Hillegeist, and K. Lo, 2009. The effect of earnings surprises on information asymmetry, Journal of Accounting and Economics 47: 208-225.

Bushee, B., and C. Noe, 2000, Corporate disclosure practices, institutional investors, and stock return volatility. Journal of Accounting Research 38: 171-202.

Page 33: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

32

Cadman, B., and J. Sunder, 2014. Investor myopia and CEO horizon incentives. Forthcoming in The Accounting Review.

Campbell, J. Y. and L. Hentschel. 1992. No news is good news: An asymmetric model of changing volatility in stock returns. Journal of Financial Economics. 31: 281-318.

Cao, C., L. Field, and G. Hanka, 2004. Does insider trading impair market liquidity? Evidence from lockup expirations. Journal of Financial and Quantitative Analysis. 39: 25-46.

Carter, R. and S. Manaster, 1990. Initial public offerings and underwriter reputation. Journal of Finance: 45: 1045-1068.

Chen, S., D. Matsumoto, and S. Rajgopal, 2010. Is silence golden? An empirical analysis of firms that stop giving quarterly earnings guidance. Journal of Accounting and Economics 51: 134-150.

Cheng, Q. and K. Lo, 2006. Insider trading and voluntary disclosures. Journal of Accounting Research 44: 815-848.

Christie, A., 1982. The stochastic behavior of common stock variances: Value, leverage, and interest rate effects. Journal of Financial Economics 10: 407-432.

Chuk, E., D. Matsumoto, and G. Miller, 2013. Assessing methods of identifying management forecasts: CIG vs. researcher collected. Journal of Accounting and Economics 55(1): 23-42.

Collins, D. W., and S.P. Kothari, 1989. An analysis of intertemporal and cross-sectional determinants of earnings response coefficients. Journal of Accounting and Economics 11: 143-181.

Diamond, D. W. and R. Verrecchia, 1991. Disclosure, liquidity and the cost of capital. Journal of Finance 46: 1325-1359.

Dye R.A., 1985. Disclosure of nonproprietary information. Journal of Accounting Research 23: 123-145.

Field, L. C. and G. Hanka, 2001. The expiration of IPO share lockups. Journal of Finance 56: 471 - 500.

Field, L.C., M. Lowry, and S. Shu, 2005. Does disclosure deter or trigger litigation? Journal of Accounting and Economics 39: 487-507.

Field, L.C. and M. Lowry, 2009. Institutional versus individual investment in IPOs: The importance of firm fundamentals. Journal of Financial and Quantitative Analysis. 65: 489-516.

Francis, J., D. Philbrick, and K. Schipper, 1994. Shareholder litigation and corporate disclosures. Journal of Accounting Research 32: 137-164.

Page 34: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

33

Gompers, P. and J. Lerner, 1998. Venture capital distributions: Short-run and long-run reactions. The Journal of Finance 53: 2161-2183.

Gompers, P. and J. Lerner, 2004. The venture capital cycle. MIT Press. Gorman, M., and W. Sahlman, 1989. What do venture capitalists do? Journal of Business

Venturing 4: 231-248. Graham J.R., C. R. Harvey and S. Rajgopal, 2005. The economic implications of corporate

financial reporting. Journal of Accounting and Economics 40: 3-73.

Healy, P., A. Hutton, and K. Palepu, 1999. Stock performance and intermediation changes surrounding sustained increases in disclosure. Contemporary Accounting Research 16: 485-520.

Heflin F., K. R. Subramanyam and Y. Zhang, 2003. Regulation FD and the financial information environment: Early evidence. The Accounting Review 78: 1-37.

Hellmann, T. and M. Puri, 2000. The interaction between product market and financing strategy: The role of venture capital. Review of Financial Studies 13: 959-984.

Hellmann, T. and M. Puri, 2002. Venture capital and the professionalization of start-up firms: empirical evidence. Journal of Finance 57: 169-197.

Hribar, P., and H. Yang, 2010. Does CEO overconfidence affect management forecasting and

subsequent earnings management? Working Paper University of Iowa.

Hutton, A., G. Miller, and D. Skinner, 2003. The role of supplementary statements with management earnings forecasts. Journal of Accounting Research 41: 867-890.

Kasznik R., and B. Lev, 1995. To warn or not to warn: Management disclosures in the face of an earnings surprise. The Accounting Review 70: 113-134.

Klein, A., and E. Zur, 2009. Entrepreneurial shareholder activism: Hedge funds and other private investors. Journal of Finance 64: 187-229.

Kyle, A.S., 1985. Continuous auctions and insider trading. Econometrica 53: 1315–1336.

Lang, M.H., 1991. Time-varying stock price response to earnings induced by uncertainty about the time-series process of earnings. Journal of Accounting Research 29: 229-257.

Lang, M.H., and R. J. Lundholm, 1996. Corporate disclosure policy and analyst behavior. The Accounting Review 71: 467-492.

Lang, M.H., and R. J. Lundholm, 2000. Voluntary disclosure and equity offerings: reducing information asymmetry or hyping the stock? Contemporary Accounting Research 17: 623-662.

Page 35: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

34

Lerner, J., 1995. Venture capitalists and the oversight of private firms. The Journal of Finance 50: 301-318

Loughran, T., and J. R. Ritter, 2004. Why has IPO underpricing changed over time? Financial

Management 33: 5-37.

Lowry, M., and S. Shu, 2002. Litigation risk and IPO underpricing. Journal of Financial Economics 65: 309-335.

Miller, G. S., 2002. Earnings performance and discretionary disclosure. Journal of Accounting Research 40: 173-204.

Mohanram, P.S., and S.V. Sunder, 2006. How has Regulation FD affected the operations of financial analysts? Contemporary Accounting Research 23: 491-525.

Noe, C. F., 1999. Voluntary disclosures and insider transactions. Journal of Accounting and Economics 27: 305 - 326.

Norton, E.C., H. Wang, and C. Ai, 2004. Computing interaction effects and standard errors in logit and probit models. The Stata Journal 4: 154-167.

Ofek, E. and M. Richardson, 2000. The IPO lock-up period: Implications for market efficiency and downward sloping demand curves. Working Paper New York University.

Rogers, J. L., 2008. Disclosure quality and management trading incentives. Journal of Accounting Research 46: 1265-1296.

Rogers, J.L. and P. Stocken, 2005. Credibility of management forecasts. The Accounting Review 80: 1233-1260.

Rogers, J.L., D. J. Skinner, A. Van Buskirk, 2009. Earnings guidance and market uncertainty. Journal of Accounting and Economics 48: 90-109.

Schwert G. W., 1989. Why does stock market volatility change over time? Journal of Finance 44: 1115-1153. Skinner, D.J., 1994. Why firms voluntarily disclose bad news? Journal of Accounting Research

32: 38-60.

Skinner, D.J., 1997. Earnings disclosures and stockholder lawsuits. Journal of Accounting and Economics 23: 249-282.

Sletten, E., 2012, The effect of stock price on discretionary disclosure. Review of Accounting Studies 17: 96-133.

Tucker, J, 2007. Is openness penalized? Stock returns around earnings warnings. The Accounting Review 82: 1055-1087.

Page 36: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

35

Verrecchia, R. E., 1983. Discretionary disclosure. Journal of Accounting and Economics 5: 179-194.

Waymire, G. 1984. Additional evidence on the information content of management earnings forecasts. Journal of Accounting Research 22: 703-718.

Page 37: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

36

Appendix A – Variable Definitions Variable Name Variable Definition Abnormal Returns Market adjusted buy and hold returns over the [-1,+1] window where day (Earnings Announcement) 0 is the quarterly earnings announcement. Source: CRSP

Abnormal Returns Market adjusted buy and hold returns from 11 days after the previous (Quarter Excluding earnings announcement through day -2 relative to the quarterly earnings Earnings Announcement) announcement date. Source: CRSP

Abnormal Returns Market adjusted buy and hold returns from the lockup expiration (Trading Window) day through day -2 relative to the quarterly earnings announcement date.

Source: CRSP

Abnormal Volume The average daily abnormal trading volume multiplied by the number of (Trading Window) days in the trading window (lockup expiration day through day -2

relative to the quarterly earnings announcement date). The average daily abnormal trading volume is the difference between the average volumes over the trading window and days -50 to -6 relative to the lockup expiration, scaled by shares outstanding. Source: CRSP

After Regulation FD An indicator variable that is equal to one if the announcement quarter ends after the enactment of Regulation FD (October 23, 2000) and zero otherwise.

Analyst Following Natural logarithm of one plus the number of analysts that issue at least one (one-quarter- or one-year-ahead) earnings forecast for the firm during the announcement quarter. Source: I/B/E/S.

Bad News An indicator variable that is equal to one when the signed News Magnitude (defined below) is less than zero. Source: I/B/E/S, First Call.

Forecast An indicator variable that is equal to one if the firm issued at least one quantitative forecast pertaining to this quarter’s EPS in a given quarter (excluding those issued in the first 10-days), and zero otherwise. Source: First Call.

Good News An indicator variable that is equal to one when the signed News Magnitude (defined below) is greater than or equal to zero. Source: I/B/E/S, First Call.

High Tech Firm An indicator variable that is equal to one for firms in the following SIC industries: 2833, 2834, 2835, 2836, 3570, 3571, 3572, 3576, 3577, 3661, 3674, 4812, 4813, 5045, 5961, 7370, 7371, 7372, 7373. Source: Compustat.

Litigation Probability The cumulative density function of the fitted value from the litigation probability model in Rogers and Stocken (2005) computed for a given quarter. Source: CRSP, Compustat, IBES, TAQ.

Manager Ownership Locked % Percentage of shares outstanding held by officers at the time of the IPO. Source: SEC Edgar database.

Manager Sell An indicator variable that is equal to one if the number of shares sold exceeds the number of shares purchased by company officers during the lockup expiration quarter. Source: Thomson Financial.

Page 38: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

37

Market-to-Book Ratio Market capitalization scaled by the book value of equity for a given quarter. Source: Compustat.

News Magnitude Absolute value of the difference between actual quarterly earnings and analysts’ expectations of earnings (proxied by the average earnings forecasts issued over the first 10 days of a given quarter) scaled by beginning of quarter stock price. Source: I/B/E/S, First Call and CRSP.

No Forecast An indicator variable that is equal to one if the firm does not issue at least one quantitative forecast pertaining to the current quarter’s EPS during the lockup expiration quarter. Source: First Call.

% of Institutional Ownership Percentage of shares outstanding held by institutional shareholders measured at the latest TFN report date that falls in a given quarter. Source: Thomson Financial.

% of Shares Locked One minus the percentage of shares outstanding sold in the IPO. Source: SDC, SEC Edgar database.

Predicted Abnormal Volume Predicted value from the Abnormal Volume Model in Appendix B.

Residual Ownership Locked % Percentage of shares outstanding held by shareholders other than officers and VCs at the time of the IPO. Source: SEC Edgar database.

Return on Assets Income before extraordinary items for a given quarter scaled by total assets as of the end of the quarter. Source: Compustat.

Run-up Market adjusted buy and hold returns over the window starting from the issue date of the IPO and ending one day before the start of the quarter during which the lockup expires. Source: CRSP.

Size (in millions) The market value of equity at the beginning of a given quarter. Source: Compustat.

Standard Deviation of Returns The standard deviation of daily stock returns in the preceding quarter. Source: CRSP

Venture Capital Backed An indicator variable that is equal to one if the firm is venture capital backed and zero otherwise. Source: SDC, SEC Edgar database.

Venture Capital High Sell An indicator variable that is equal to one if the firm is venture capital backed and experienced an above average decrease in VC ownership following lockup expiration. Source: SDC, SEC Edgar database.

VC Ownership Locked % Percentage of shares outstanding held by VCs at the time of the IPO. Source: SEC Edgar database.

Top-tier Underwriter An indicator variable that is equal to one if the underwriter for the IPO has a modified Carter Manaster Rank of 9.1 (Carter and Manaster 1990, Loughran and Ritter 2004). We thank Jay Ritter for making the data available at http://bear.cba.ufl.edu/ritter/ipolink.htm.

Uncertainty The standard deviation of daily stock returns in the quarter preceding the lockup expiration quarter. Source: CRSP.

Z-Score < 1.81 An indicator variable that is equal to one if the firm has an Altman Z-score of less than 1.81 in a given quarter and zero otherwise. Source: CRSP and Compustat.

Page 39: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

38

Appendix B – Abnormal Trading Volume Prediction Model The following table presents the results from an OLS regression in which Abnormal Volume is the dependent variable. The sample consists of 776 IPO-firm-lockup expiration quarters over the 1995-2007 period. ***, **, and * denote p-values less than 0.01, 0.05 and 0.1, respectively. All variables are defined in Appendix A.

Dependent Variable = Abnormal Volume

Variable Coefficient SE Intercept -0.221 ** 0.10Run-up 0.039 *** 0.01Venture Capital Backed 0.130 *** 0.03Top-tier Underwriter 0.009 0.03% of Shares Locked 0.003 *** 0.00High Tech Firm 0.019 0.04

Observations 776Adjusted R2 0.063

Page 40: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

39

Appendix C – Measurement of VC Sales

VCs, unlike managers and directors, are not treated as insiders and, therefore, are not required to report their trading behavior around the lockup-expiration date (Field and Hanka 2001, Gompers and Lerner 1998). Following the prior literature, we construct a measure of VC selling/distribution based on changes in VC ownership that we hand collect from the IPO prospectus and the post-lockup-expiration proxy statements. We use this as a proxy for trading behavior around the lockup expiration. We identify VCs among the list of beneficial shareholders reported in both the prospectus and the proxy statement. Firms report the percentage of beneficial ownership for (i) each stockholder who is known by the company to beneficially own 5% or more of any class of its capital stock, (ii) each of its present executive officers who are also named in the Summary Compensation Table, and (iii) each of its directors. To collect ownership information for VCs, for each of our 347 VC-backed IPO firms, we: 1) Search for the prospectus (424B) on the SEC Edgar database. 2) Identify the VCs in the “Principal Stockholders” table. Record the sum of their ownership. 3) Search for the first proxy statement filed after lockup expiration on the SEC Edgar database. 4) Note the ownership of the VCs identified in Step 2 above. The above process results in a sample of 260 VC-backed firms for which we are able to compute the change in VC ownership as a percentage of shares outstanding.1 The table below reports the distribution of the percentage point change in VC ownership stake: Mean p1 p25 Median p75 p99 Change in VC Ownership -9.77% -49.90% -14.30% -6.15% -1.20% 3.9% Unfortunately, the data are noisy for a number of reasons: � While firms tend to report VC ownership in the prospectus regardless of the magnitude of the

ownership stake, companies are not required to report beneficial owners with less than 5% ownership in their proxy statements unless a VC partner is sitting on the board of directors. This non-disclosure of smaller ownership stakes in the post-IPO proxies leads us to overstate changes in ownership.

� Because proxy statements are filed annually, we cannot identify the precise timing of the change in ownership and we assume that it occurs when the lockup expires.

� The prospectus estimates the ownership stake of the pre-IPO shareholders as of the completion of the IPO using projected IPO share sales. If the actual number of shares sold in the IPO differs from the proposed amount in the prospectus, the ownership stake reported in the post-IPO proxy will be different even in the absence of any selling or distribution.

The above challenges in data collection lead to two problems. First, identifying VC selling or distribution is not reliable because (a) misidentification of the size of the IPO may be wrongly interpreted as ownership changes and (b) some changes in ownership may be driven by the lack of data on VC ownership below 5%. To illustrate (b), a VC with 6% of ownership at the time of the IPO and 4.5% ownership at the time the first proxy after the lockup expiration is filed will be incorrectly classified as experiencing a 6 percentage point decrease in ownership. To address this issue, we require the change in ownership to be above the mean to be considered significant. Second, the magnitude of the change in ownership may not be entirely accurate because of the lack of data on VC ownership below 5%. Therefore, rather than using a continuous variable measuring change in ownership, we rely on an indicator variable. We define Venture Capital High Sell as equal to one if the percentage change in VC ownership is above the mean, i.e., 9.77%, and zero otherwise. 1 We are unable to obtain the prospectuses for some of the firms that had their IPOs in the early part of the sample period (46 and 21 IPOs in 1995 and 1996, respectively), because the SEC Edgar database was initiated in 1996.

Page 41: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

40

Tab

le 1

– D

escr

iptiv

e St

atis

tics

Tabl

e 1

repo

rts d

escr

iptiv

e st

atis

tics

for

our

final

sam

ple

of 3

,126

firm

-qua

rter

obse

rvat

ions

. **

*, *

*, a

nd *

den

ote

p-va

lues

les

s th

an 0

.01,

0.0

5 an

d 0.

1,

resp

ectiv

ely.

All

varia

bles

are

def

ined

in A

ppen

dix

A.

Al

l Obs

erva

tions

Fo

reca

st =

0

Fore

cast

= 1

Fo

reca

st =

1 v

s. Fo

reca

st =

0

N =

3,1

26

N =

2,8

32

N =

294

Var

iabl

e M

ean

Stan

dard

D

evia

tion

Mea

n M

ean

Diff

eren

ce in

M

eans

t-

stat

istic

Ba

d N

ews

0.34

0 0.

474

0.31

2 0.

612

0.30

0 10

.53

***

Goo

d N

ews

0.66

0 0.

474

0.68

8 0.

388

-0.3

00

-10.

53

***

New

s Mag

nitu

de fo

r firm

-qua

rters

with

Goo

d N

ews

0.00

2 0.

006

0.00

2 0.

001

-0.0

01

-3.1

9 **

* N

ews M

agni

tude

for f

irm-q

uarte

rs w

ith B

ad N

ews

0.00

4 0.

012

0.00

3 0.

008

0.00

4 6.

15

***

Anal

yst F

ollo

wing

4.

739

3.36

6 4.

624

5.84

4 1.

220

5.94

**

* %

of I

nstit

utio

nal O

wner

ship

0.

360

0.24

3 0.

359

0.36

9 0.

011

0.71

Si

ze (i

n m

illio

ns)

1,02

8 3,

201

1,00

2 1,

276

274

1.40

M

arke

t-to-

Book

Rat

io

4.41

5 18

.670

4.

365

4.90

5 0.

541

0.47

Re

turn

on

Asse

ts

-0.0

25

0.20

5 -0

.026

-0

.013

0.

013

1.00

Z-

Scor

e <

1.8

1 0.

397

0.48

9 0.

393

0.43

5 0.

042

1.41

Li

tigat

ion

Prob

abili

ty

0.00

5 0.

013

0.00

5 0.

011

0.00

6 7.

36

***

Stan

dard

Dev

iatio

n of

Ret

urns

0.

046

0.02

5 0.

046

0.04

7 0.

001

0.97

Af

ter R

egul

atio

n FD

0.

379

0.48

5 0.

376

0.40

5 0.

028

0.95

Page 42: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

41

Tab

le 2

– D

istr

ibut

ion

of F

orec

asts

by

Peri

od R

elat

ive

to L

ocku

p E

xpir

atio

n an

d by

Sig

n of

Qua

rter

ly N

ews

Tabl

e 2

repo

rts th

e fr

actio

n of

EPS

man

agem

ent f

orec

asts

by

perio

d re

lativ

e to

lock

up e

xpira

tion

and

by th

e na

ture

of t

he q

uarte

rly n

ews (

Goo

d N

ews v

ersu

s Bad

N

ews)

. Pre

-Loc

kup

Expi

ratio

n Q

uart

ers

are

anno

unce

men

t qu

arte

rs i

mm

edia

tely

pre

cedi

ng t

he q

uarte

r du

ring

whi

ch t

he l

ocku

p ex

pire

s. Lo

ckup

Exp

iratio

n Q

uart

ers a

re th

e an

noun

cem

ent q

uarte

rs d

urin

g w

hich

the

lock

up e

xpire

s. Po

st-L

ocku

p Ex

pira

tion

Qua

rters

are

the

four

ann

ounc

emen

t qua

rters

subs

eque

nt to

the

anno

unce

men

t qua

rter d

urin

g w

hich

the

lock

up e

xpire

s **

*, *

*, a

nd *

den

ote

p-va

lues

less

than

0.0

1, 0

.05

and

0.1,

resp

ectiv

ely.

Goo

d N

ews

and

Bad

New

s ar

e de

fined

in A

ppen

dix

A.

Ba

d N

ews

Goo

d N

ews

Peri

odN

umbe

r of

Fi

rm-Q

uart

ers

Mea

n of

F

orec

ast

Num

ber

of

Firm

-Qua

rter

sM

ean

ofF

orec

ast

(1

) Pre

-Loc

kup

Expi

ratio

n Q

uart

ers

101

0.16

828

00.

046

(2) L

ocku

p Ex

pira

tion

Qua

rter

s 23

90.

117

537

0.04

5(3

) Pos

t-Loc

kup

Expi

ratio

n Q

uart

ers

723

0.18

71,

246

0.06

2

T-t

est f

or d

iffer

ence

s(2

) vs (

3)-2

.49**

(2) v

s (3)

-1.4

3

Page 43: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

42

Tab

le 3

– E

x A

nte

Selli

ng In

cent

ives

and

For

ecas

t Pro

pens

ity in

the

Loc

kup

Exp

irat

ion

Qua

rter

Ta

ble

3 re

ports

res

ults

fro

m e

stim

atin

g Eq

uatio

ns (

1a)

and

(1b)

on

subs

ampl

es o

f fir

m-q

uarte

rs w

ith G

ood

New

s an

d Ba

d N

ews.

The

depe

nden

t va

riabl

e is

Fo

reca

st, w

hich

take

s th

e va

lue

of o

ne if

the

firm

issu

ed a

t lea

st o

ne q

uant

itativ

e EP

S fo

reca

st in

the

fore

cast

win

dow

of t

he lo

ckup

qua

rter,

and

zero

oth

erw

ise.

In

Mod

els

1a a

nd 1

b, w

e lo

se o

bser

vatio

ns b

ecau

se s

ome

dich

otom

ous

inde

pend

ent v

aria

bles

pre

dict

for

ecas

ting

beha

vior

per

fect

ly. I

n M

odel

s 2a

and

2b,

we

furth

er lo

se o

bser

vatio

ns w

ith n

o av

aila

bilit

y of

ow

ners

hip

data

on

EDG

AR

. **

*, *

*, a

nd *

den

ote

p-va

lues

less

than

0.0

1, 0

.05

and

0.1,

resp

ectiv

ely.

Yea

r fix

ed

effe

cts a

re n

ot re

porte

d. A

ll ex

plan

ator

y va

riabl

es a

re d

efin

ed in

App

endi

x A

.

Dep

ende

nt V

aria

ble

= F

orec

ast

D

epen

dent

Var

iabl

e =

For

ecas

t

Bad

New

s G

ood

New

s

Bad

New

s G

ood

New

s

Mod

el 1

aM

odel

1b

M

odel

2a

Mod

el 2

bV

aria

ble

Coe

ffic

ient

SEC

oeff

icie

ntSE

Coe

ffic

ient

SEC

oeff

icie

ntSE

Inte

rcep

t -6

.788

**2.

69-6

.137

***

1.92

-9

.969

***

3.05

-4.6

03**

2.29

Pred

icte

d Ab

norm

al V

olum

e -3

.627

**1.

60-0

.364

1.10

VC

Ow

ners

hip

Lock

ed %

-0

.032

**0.

020.

005

0.01

Offi

cers

Ow

ners

hip

Lock

ed %

-0

.020

0.02

0.00

90.

01Re

sidu

al O

wne

rshi

p Lo

cked

%

-0.0

140.

010.

011

0.01

New

s Mag

nitu

de

17.1

08**

6.75

-24.

186

18.9

2 16

.516

**7.

92-2

6.72

720

.56

Anal

yst F

ollo

win

g 0.

400

0.36

0.04

20.

27

0.42

20.

41-0

.032

0.33

% o

f Ins

titut

iona

l Ow

ners

hip

-0.2

810.

62-0

.114

0.44

-0

.564

0.67

-0.0

040.

44Si

ze (i

n m

illio

ns)

0.25

6*

0.15

0.03

70.

11

0.18

30.

16-0

.052

0.14

Mar

ket-t

o-Bo

ok R

atio

0.

004

0.03

0.00

10.

00

0.00

60.

03-0

.001

0.01

Retu

rn o

n As

sets

5.

945

***

2.02

0.17

30.

69

7.59

9**

3.04

0.47

71.

14Z-

Scor

e <

1.8

1 -0

.224

0.29

0.31

00.

19

-0.3

700.

320.

356

*0.

21Li

tigat

ion

Prob

abili

ty

22.9

28**

*6.

9718

.331

*10

.42

23.7

24**

*7.

7318

.810

*10

.92

Stan

dard

Dev

iatio

n of

Ret

urns

15

.926

**7.

66-4

.571

5.97

14

.113

*8.

21-3

.987

6.30

Afte

r Reg

ulat

ion

FD

5.31

1**

*0.

770.

882

0.66

5.

143

***

0.85

0.85

60.

67

N

235

524

173

398

Pseu

do R

2 0.

262

0.09

50.

303

0.08

7

Page 44: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

43

Tab

le 4

– R

ole

of L

itiga

tion

Ris

k an

d Fi

rm-s

peci

fic U

ncer

tain

ty

Tabl

e 4

repo

rts re

sults

from

Eq.

(1a)

and

(1b)

mod

ified

to s

plit

the

prox

ies

for s

ellin

g in

cent

ives

, Pre

dict

ed A

bnor

mal

Vol

ume

and

VC O

wne

rshi

p Lo

cked

% in

to

two

varia

bles

cap

turin

g se

lling

ince

ntiv

es fo

r hig

h- a

nd lo

w-

Litig

atio

n Ri

sk a

nd U

ncer

tain

ty e

stim

ated

for

Bad

New

s fir

m q

uarte

rs. T

he d

epen

dent

var

iabl

e is

Fo

reca

st, w

hich

take

s th

e va

lue

of o

ne if

the

firm

issu

ed a

t lea

st o

ne q

uant

itativ

e EP

S fo

reca

st in

the

fore

cast

win

dow

of t

he lo

ckup

qua

rter,

and

zero

oth

erw

ise.

In

Mod

els

1a a

nd 1

b, w

e lo

se o

bser

vatio

ns b

ecau

se s

ome

dich

otom

ous

inde

pend

ent v

aria

bles

pre

dict

for

ecas

ting

beha

vior

per

fect

ly. I

n M

odel

s 2a

and

2b,

we

furth

er lo

se o

bser

vatio

ns w

ith n

o av

aila

bilit

y of

ow

ners

hip

data

on

EDG

AR

. **

*, *

*, a

nd *

den

ote

p-va

lues

less

than

0.0

1, 0

.05

and

0.1,

resp

ectiv

ely.

Yea

r fix

ed

effe

cts a

re n

ot re

porte

d. A

ll ex

plan

ator

y va

riabl

es a

re d

efin

ed in

App

endi

x A

.

Dep

ende

nt V

aria

ble

= F

orec

ast

D

epen

dent

Var

iabl

e =

For

ecas

tM

odel

1a

Mod

el 1

bM

odel

2a

Mod

el 2

bG

roup

s: H

igh

or L

ow

Litig

atio

n Ri

sk

Gro

ups:

Hig

h or

Low

U

ncer

tain

ty

Gro

ups:

Hig

h or

Low

Li

tigat

ion

Risk

G

roup

s: H

igh

or L

ow

Unc

erta

inty

V

aria

ble

Coe

ffic

ient

SEC

oeff

icie

ntSE

Coe

ffic

ient

SEC

oeff

icie

ntSE

Inte

rcep

t -6

.269

**2.

89-7

.347

***

2.79

-1

0.29

3**

*3.

26-1

0.72

2**

*3.

44Pr

edic

ted

Abno

rmal

Vol

ume

x H

igh

-1.0

861.

79-4

.904

***

1.73

Pr

edic

ted

Abno

rmal

Vol

ume

x Lo

w

-5.6

63**

*1.

98-2

.816

1.86

VC

Ow

ners

hip

Lock

ed %

x H

igh

-0.0

180.

02-0

.058

***

0.02

VC O

wne

rshi

p Lo

cked

% x

Low

-0

.044

**0.

02-0

.026

*0.

02O

ffice

rs O

wne

rshi

p Lo

cked

%

-0.0

210.

02-0

.021

0.02

Resi

dual

Ow

ners

hip

Lock

ed %

-0

.015

0.01

-0.0

150.

01N

ews M

agni

tude

14

.754

*7.

6416

.953

**6.

64

15.1

37*

8.23

17.0

34**

7.69

Anal

yst F

ollo

win

g 0.

471

0.37

0.38

90.

36

0.45

90.

410.

434

0.41

% o

f Ins

titut

iona

l Ow

ners

hip

-0.0

240.

68-0

.258

0.63

-0

.518

0.69

-0.6

180.

69Si

ze (i

n m

illio

ns)

0.23

60.

150.

271

*0.

15

0.17

90.

160.

222

0.17

Mar

ket-t

o-Bo

ok R

atio

0.

003

0.04

-0.0

010.

03

0.00

60.

030.

004

0.04

Retu

rn o

n As

sets

5.

737

**2.

335.

842

***

1.96

6.

692

**2.

837.

254

**2.

90Z-

Scor

e <

1.8

1 -0

.270

0.29

-0.1

630.

30

-0.3

950.

32-0

.328

0.33

Litig

atio

n Pr

obab

ility

12

.774

*7.

0825

.663

***

7.24

18

.450

**8.

2227

.721

***

8.03

Stan

dard

Dev

iatio

n of

Ret

urns

10

.298

8.02

22.3

96**

9.00

15

.044

*8.

0820

.371

**9.

83Af

ter R

egul

atio

n FD

4.

708

***

0.77

5.47

0**

*0.

85

5.52

5**

*0.

985.

013

***

0.82

N

235

235

173

173

Pseu

do R

2 0.

306

0.26

90.

314

0.30

6

��

��

Page 45: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

44

Table 5 –Management Forecast Propensity and Realized Selling in the Lockup Expiration Quarter Table 5 reports results from estimating Equation (2) on subsamples of firm-quarters with Good News and Bad News. The sample size differs from Table 3 because of data requirements to capture selling by VCs and managers. The sample includes VC-backed firms only when their ownership data are available. The dependent variable is Forecast, which takes the value of one if the firm issued at least one quantitative EPS forecast in the forecast window of the lockup expiration quarter, and zero otherwise. Residual Abnormal Volume is measured as the aggregate abnormal volume over the trading window less the change in ownership by VCs and sales by managers over the same window. ***, **, and * denote p-values less than 0.01, 0.05 and 0.1, respectively. Year fixed effects are suppressed. All explanatory variables are defined in Appendix A.

Dependent Variable = Forecast Model 1a Model 1bBad News Good News

Variable Coefficient SE Coefficient SE

Intercept -4.038 2.77 -5.347 *** 1.96Venture Capital High Sell -4.946 *** 0.82 -0.180 0.46Manager Sell -0.174 0.47 0.304 0.26Manager Sell x Venture Capital High Sell 5.715 *** 1.40 0.099 0.62Residual Abnormal Volume 0.004 0.01 0.005 0.01News Magnitude 12.970 8.22 -18.344 15.84Analyst Following 0.620 * 0.36 -0.066 0.27% of Institutional Ownership -0.405 0.61 -0.225 0.44Size (in millions) 0.106 0.14 0.011 0.10Market-to-Book Ratio -0.007 0.04 0.000 0.00Return on Assets 8.432 *** 2.81 0.399 1.13Z-Score < 1.81 -0.084 0.32 0.437 ** 0.20Litigation Probability 20.742 ** 9.25 16.768 10.96Standard Deviation of Returns 10.564 7.46 -6.381 6.39After Regulation FD 5.156 *** 0.73 1.018 0.64

Year Fixed Effects Yes Yes

N 203 468 Pseudo R2 0.299 0.102

Page 46: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

45

Tab

le 6

– A

bnor

mal

Ret

urns

in th

e L

ocku

p E

xpir

atio

n Q

uart

ers w

ith B

ad N

ews

Tabl

e 6

pres

ents

the

resu

lts fr

om th

e es

timat

ion

of E

quat

ion

(3) (

Mod

els

1 an

d 2)

and

Equ

atio

n (4

) (M

odel

3) f

or fi

rms

with

bad

new

s. Th

e de

pend

ent v

aria

ble

in

Mod

el 1

, Abn

orm

al R

etur

ns (

Qua

rter

Exc

ludi

ng E

arni

ngs

Anno

unce

men

t), a

re m

arke

t-adj

uste

d bu

y an

d ho

ld r

etur

ns f

rom

11

days

afte

r th

e pr

evio

us e

arni

ngs

anno

unce

men

t th

roug

h 2

days

prio

r to

ear

ning

s an

noun

cem

ent;

the

depe

nden

t va

riabl

e in

Mod

el 2

, Abn

orm

al R

etur

ns (

Earn

ings

Ann

ounc

emen

t), a

re m

arke

t-ad

just

ed b

uy a

nd h

old

retu

rns

on d

ays

-1 to

+1

rela

tive

to e

arni

ngs

anno

unce

men

t; an

d th

e de

pend

ent v

aria

ble

in M

odel

3, A

bnor

mal

Ret

urns

(Tra

ding

Win

dow)

, ar

e m

arke

t-adj

uste

d bu

y an

d ho

ld r

etur

ns f

rom

lock

up e

xpira

tion

thro

ugh

2 da

ys p

rior

to e

arni

ngs

anno

unce

men

t. W

e lo

se tw

o ob

serv

atio

ns r

elat

ive

to T

able

2

beca

use

of m

issi

ng r

etur

ns d

ata.

***

, **,

and

* d

enot

e p-

valu

es le

ss th

an 0

.01,

0.0

5 an

d 0.

1, r

espe

ctiv

ely.

Yea

r an

d in

dust

ry f

ixed

eff

ects

are

sup

pres

sed.

All

expl

anat

ory

varia

bles

are

def

ined

in A

ppen

dix

A.

��M

odel

1M

odel

2M

odel

3D

epen

dent

Var

iabl

e =

Abn

orm

al R

etur

ns

�� Q

uart

er E

xclu

ding

Ea

rnin

gs

Anno

unce

men

t

Earn

ings

An

noun

cem

ent W

indo

w

Trad

ing

Win

dow

Var

iabl

eC

oeff

.

SEC

oeff

.

SEC

oeff

.

SEIn

terc

ept

-0.7

02

***

0.19

-0

.117

**

0.05

-0

.031

0.

12

No

Fore

cast

0.

133

* �0.

07

-0.0

71

* �0.

04

0.01

3 0.

06

New

s Mag

nitu

de

-1.0

94

1.39

-0

.356

0.

39

-0.5

32

0.95

Ab

norm

al V

olum

e

-0.2

59

**0.

11

Abno

rmal

Vol

ume

x N

o Fo

reca

st

0.18

8 0.

12

N

237

237

237

R2

0.21

2

0.16

7

0.15

1 ��

��

Page 47: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

46

Figure 1: Timeline for Lockup Expiration Quarter

Earnings announcement t

10-day window for estimating analyst earnings consensus (News estimation

window)

Earnings announcement t-1

Window for measuring management forecasts

(Forecast Window)

Actual earnings reported

Lockup Expiration

Trading Window

Page 48: Large shareholders and disclosure strategies: Evidence from IPO lockup expirations

47

� We�study�how�large�shareholders’�selling�incentives�affect�voluntary�disclosures.��

� Managers�delay�disclosure�of�bad�news�for�the�benefit�of�influential�shareholders.�

� The�effect�is�stronger�when�large�shareholders’�selling�incentives�are�greater.��

� The�effect�is�concentrated�in�firms�with�high�uncertainty�and�low�litigation�risk.�

� Managers�themselves�do�not�benefit�from�the�delay�of�bad�news.�