No news is good news: voluntary disclosure in the face of litigation

• Published date: 01 November 2016
• Author: Iván Marinovic,Felipe Varas
• DOI: http://doi.org/10.1111/1756-2171.12156
• Date: 01 November 2016

RAND Journal of Economics

Vol.47, No. 4, Winter 2016

pp. 822–856

No news is good news: voluntary disclosure

in the face of litigation

Iv´

an Marinovic∗

and

Felipe Varas∗∗

Westudy disclosure dynamics when the firm value evolves stochastically over time. The presence

of litigation risk, arising from the failureto disclose unfavorable information, crowds out positive

disclosures. Litigation risk mitigates firms’ tendency to use inefficient disclosure policies. Froma

policy perspective, we show that a stricter legal environment may be an efficient way to stimulate

information transmission in capital markets, particularly when the nature of information is

proprietary. We model the endogeneity of litigation risk in a dynamic setting and shed light on

the empirical controversy regarding whether disclosure preempts or triggers litigation.

1. Introduction

Firms receive information on an ongoing basis. Productivity shocks originate from many

sources, including innovation breakthroughs, the arrival of business opportunities, frictions in

negotiations with labor unions, and breakdowns of supply-chain relationships. Capital market

perceptions pressure managers to disclose information frequently because the stock price per-

formance is affected by disclosures and lack thereof (see, e.g., Graham, Harvey, and Rajgopal,

2005).

Firm managers may thus feel inclined to disclose good news,but disclosing positive news can

be costly.Concealing bad news likewise can be risky because that information might be revealed

by external sources, perhaps triggering costly litigation. For example, in 2012, the US Justice

Department announced that GlaxoSmithKline (GSK) had agreed to plead guilty and pay a $3

billion fine for withholding information about the cardiovascular risk of Avandia, its antidiabetes

drug. Avandia’s problems began in 2007, when a study published in the New England Journal

of Medicine (Nissen and Wolski, 2007) found that the drug carried a higher risk of heart attacks

than alternative drugs. The 2012 settlement stems from claims made by four GSK employees,

∗Stanford University; imvial@stanford.edu.

∗∗Duke University; felipe.varas@duke.edu.

We thank the Editor, Mark Armstrong, and two anonymous reviewers for their constructive comments, which helped

us to improve the manuscript. We also thank Jeremy Bertomeu, Kyle Bagwell, Ron Dye, Ilan Guttman, Cam Harvey,

Mike Harrison, Mirko Heinle, Jon Levin, Andy Skrzypacz, and workshop participants at Duke University, CMU, EPFL,

Wharton, Stanford University,Universidad de Chile, and the University of Houston for helpful comments.

822 C2016, The RAND Corporation.

MARINOVIC AND VARAS / 823

who tipped off the government about the firm’s concealment of two internal studies that preceded

Nissen and Wolski(2007). Avandia prescriptions and GSK’s stock price dropped sharply after the

publication of Nissen and Wolski (2007). In another example, Toyota was forced to pay a $1.2

billion penalty to settle the criminal probe into its handling of unintended acceleration problems

that led to the recall of 8.1 million vehicles beginning in 2009.1

Westudy managers’ incentives to disclose information when withholding negative informa-

tion is costly due to the presence of litigation risk. How are managers’ disclosure policies affected

by the possibility of litigation? When do managers delay the release of adverse information?

We provide a model that highlights a fundamental asymmetry between positive and negative

information. When the market expects positive information to be disclosed, no newsis bad news.

The market punishes a firm that withholds information. In contrast, when the market expects

negative information to be disclosed, no news is good news and the market rewards the absence

of disclosure. This asymmetry is important because silence, unlike a positive disclosure which

cannot be imitated, is something that firms can imitate. The presence of litigation risk therefore

allows firms to signal their good standing to the market by delaying the release of information.

We show that this trade-off determines both the timing and the content of disclosures. More im-

portant, we show that litigation risk may reduce market skepticism, thus mitigating the tendency

of managers to rely on inefficient disclosure policies.

We present a continuous-time disclosure model. Specifically, we analyze a disclosure game

between the manager of a firm and a mass of buyers (the market) when the firm’s asset value

evolves stochastically over time. The evolution of the asset value is described by a continuous-

time Markov chain that fluctuates between two possible states: low asset value and high asset

value. The manager can disclose private information at any point in time and as often as desired.

Unraveling is not possible in equilibrium because disclosing good news is costly (so the firm

cannot do it continuously). Concealing bad news is risky because there the public news process

has a positive arrival rate when the asset value is low and will trigger costly litigation if the

manager fails to disclose the information before the news arrives. The market is competitive

and continuously adjusts the firm’s stock price, based on all public information. The manager

maximizes the present value of the firm’s future stock prices, perhaps because his compensation,

at each point in time, is proportional to the firm’s stock price.

The presence of litigation risk suggests that managers will preempt the announcement of

bad news by voluntarily disclosing it to avoid litigation costs. This preemption strategy is well

documented empirically (see, e.g., Skinner, 1994; Lev, 1995; Johnson, Kasznik, and Nelson,

2001) but not well understood in theory. In fact, the strategy poses a conceptual difficulty: if the

market expects the manager to reveal bad news at a given time, then the manager’s silence will

be interpreted as a clear sign that the manager’s information is favorable, which will lead to a

rise in the stock price. However, rewarding silence in this way cannot be part of an equilibrium

because rewarding silence, unlike providing verifiable good news, is something that all firms can

do, including those experiencing financial problems. Our analysis reveals that in equilibrium, the

firm can release bad news but only probabilistically. Indeed, the equilibrium predicts that when

falling stock prices reach a certain threshold, the firm will reveal bad news with a probability

that depends on the arrival intensity of the public news, the cost of litigation, and the proprietary

cost of disclosing good news. At that point, the stock price will remain constant until bad news

is finally disclosed.

Litigation risk not only leads to the preemption of bad news but, more important, crowds

out the disclosure of good news because silence is interpreted, per se, as a favorable signal of

the firm’sprospects. The manager can reveal good news in two ways: by explicitlydisclosing the

good news and bearing the disclosure cost or by remaining quiet, thus signalling the good news.

1Former Attorney General Eric Holder called the settlement the largest US criminal penalty ever imposed on a

car company and asserted, “We can say for certain that Toyotaintentionally concealed information and misled the public

about the safety issues behind these recalls.”

C

The RAND Corporation 2016.

824 / THE RAND JOURNAL OF ECONOMICS

There is a pecking order: the manager prefers to use silence when the firm’s undervaluation is

moderate and to use disclosure when the firm’s undervaluation is severe.

The presence of litigation risk may be desirable evenfrom the firm’s perspective. By creating

a new communication mechanism that allows managers to convey good news without disclos-

ing it, the existence of litigation risk allows the firm to save on proprietary disclosure costs.2

From a policy perspective, this means that a harsh legal environment may be a cost-effective

way of improving information transmission, especially in settings where information is highly

proprietary.

Our model is stylized but tractable and allows us to study interesting applications.

In Section 5, we make litigation risk endogenous by considering the incentives to monitor

the firm. For example, the US False Claims Act encourages people with knowledge of suspected

false claims to sue on the government’s behalf. If the Justice Department joins such a lawsuit,

the plaintiff can receive 15% to 25% of recoveries. This law creates significant incentives for

whistle-blowers to monitor firms’ disclosure behavior, thereby creating endogenous litigation

risk.3To capture the endogeneity of litigation risk, we assume the presence of a whistle-blower

who may investigate the firm at a cost. The whistle-blower receives a reward, paid by the firm, if

he can establish that the firm concealed negative information. We demonstrate that the whistle-

blower tends to investigate firms with relative low prior performance for whom the market

uncertainty is relatively large (firms that have remained silent for a lengthy time). The presence

of the whistle-blower means that, in equilibrium, there must always be a positive probability of

concealment, no matter how large the litigation cost. This extension helps clarify the puzzling

relation between litigation risk and disclosure documented empirically by Francis, Philbrick, and

Schipper (1994) suggest that disclosure triggers litigation. In our setting, litigation and disclosure

are simultaneously determined but are not causally linked.

In Section 6, we assume that withholding information for a longer period of time is more

costly to the firm, perhaps because more people traded the stock at an inflated price. We char-

acterize the amount of disclosure delay and its determinants and show that after a certain point

investors will cease to update the stock price even though the manager mayreceive and withhold

negative information—until the manager eventuallydiscloses some information. In Section 7, we

consider the case when the firm’s cash flows havea finite, predeter mined life, perhaps because the

firm’sasset is protected by a patent. In this setting, we show that firms tend to delay disclosures of

adverse events until the expiration approaches, and to cluster the release of negative information

just before its expiration. We also show that the longer the expiration, the more slowly the firm

reveals the adverse information. Tothe best of our knowledge, these predictions are new and have

not yet been tested.

Our results also apply to the problem of product quality certification (see, e.g., Dranove

and Jin, 2010). Indeed, the model can be interpreted as one in which a monopolist sells a product

of unknown quality to a mass of buyers. At each point in time, buyers purchase the good of

unobservable quality at a price that equals the good’s expected quality. The monopolist has the

option to certify the product’s quality at a cost to influence the trajectory of future prices. This

certification-like interpretation, originally adopted by Jovanovic (1982), highlights the parallel

between corporate disclosures and quality certification.

The rest of the article is organized as follows. Section 2 presents the setting. Section 3

analyzes the baseline model without litigation risk. Section 4 introduces litigation risk. Section

5 studies endogenous public news in settings where fact checkers monitor the firm’s disclosure

behavior. Sections 6 and 7 consider two extensions. Section 8 concludes.

2When adverse shocks are permanent, the result is stronger: the manager may prefer a high litigation risk to zero

litigation risk. Numerical computations suggest that this result still holds if negative shocks are sufficientlypersistent.

3An article in the Wall StreetJournal on July 24, 2014 states that Dr. William LaCorte, a “serial whistle-blower”

received a $38 million cut under a federal lawthat encourages fraud repor ting. Much of this sum wasfrom a $250 million

US settlement with Merck in 2008 over allegations that it overchargedMedicaid for the heartburn dr ug Pepcid.

C

The RAND Corporation 2016.