Information transmission during the trial: The role of punitive damages and legal costs

• Date: 01 April 2018
• DOI: http://doi.org/10.1111/jpet.12259
• Author: Felix Munoz‐Garcia,Ana Espínola‐Arredondo,Pitchayaporn Tantihkarnchana
• Published date: 01 April 2018

Received: 16 January 2017 Accepted: 17 May 2017

DOI: 10.1111/jpet.12259

ARTICLE

Information transmission during the trial: The role

of punitive damages and legal costs

Ana Espínola-Arredondo1Felix Munoz-Garcia1

Pitchayaporn Tantihkarnchana2

1WashingtonState University

2ThePennsylvania State University

Wewould like to thank the editor and two anony-

mousreferees for their insightful comments and

suggestions.We are also thankful to James J.

Jozefowicz,Moon Y. Parks, and participants at

the90th Annual Western Economic Association

Internationalconference for providing valuable

commentsand feedback on the paper.

AnaEspínola-Arredondo, 111C Hulbert Hall,

WashingtonState University, Pullman, WA99164

(anaespinola@wsu.edu).

FelixMunoz-Garcia, 101G Hulbert Hall,

WashingtonState University, Pullman, WA99164

(fmunoz@wsu.edu).

PitchayapornTantihkarnchana,317 Kern Gradu-

ateBuilding, The Pennsylvania State University,

UniversityPark, PA 16802 (pxt34@psu.edu).

This paper studies an incomplete information model in which a pre-

ventable accident occurred. The judge determining punitive dam-

ages observes the firm’s (defendant) investment decisions, but is

uninformed about the firm’s experience adopting safety measures.

Our model allows firms to file an appeal if the judge’s verdict is incor-

rect, which the judge may accept or reject. We identify under which

conditions a separating equilibrium exists where the firm’s invest-

ment decisions signal its type to the judge, who responds with a cor-

rect verdict, thus avoiding future appeals. Our paper also finds con-

ditions under which a pooling equilibrium exists whereby the firm’s

investment in precaution conceals its type from the judge, who can

respondwith an incorrect verdict thus giving rise to appeals. Further-

more, we show that the separating equilibrium is more likelyto arise

if the percentage of revenue that defendants are required to pay in

punitive damages decreases, if the punitive-to-compensatory ratio

increases, and if the legal cost of filing an appeal increases.

1INTRODUCTION

The 1989 Exxon Oil spill released 11 million gallons of crude oil into Prince William Sound in Alaska and created vast

economic and environmental damages. In September 1994, a jury awarded $286.8 million for compensatory damages

to commercial fishermen and $5 billion in punitive damages to be paid byExxon (Duffield, 1997). After several appeals,

punitive damages were reduced to $507.5 million. Punitive damage is the monetary compensation awardedto injured

parties. This amount goes beyond what is necessary to compensate individuals for losses (compensatory damage) and

it is intended to punish and encourage the defendant to be more cautious.1A similar pattern of reduced punitive dam-

ages after appeals occurred in the hot coffee lawsuit in Liebeck v.McDonald’s (1994).2We show that unreasonably high

punitive damages can induce firms to file unnecessary appeals and demonstrate that the judge’s choice of punitive

damage affects firms’ decision to invest in precaution.

1SeeKemezy v.Peters, 79 F. 3d 33 (7th Cir.1996). In the Exxon Oil spill, the amount of compensatory damages was evaluated using the contingent evaluation

method;for more details, see Carson et al. (1992).

2In this case, the jury awarded Liebeck, a 79-year-old woman who accidentally spilled hot coffee in her pelvic region and suffered third-degree burns,

$2.7 million in punitive damages, and $160,000 in compensatory damages. This amount was approximately 2 days of McDonald’s coffee sales revenueat

thattime. However, punitive damages were eventually reduced to $480,000 (Shapiro, 1995).

136 c

2017 Wiley Periodicals,Inc. wileyonlinelibrary.com/journal/jpet Journal of Public Economic Theory.2018;20:136–155.

ESPÍNOLA-ARREDONDO ET AL.137

Our model examines a setting of incomplete information in which the judge, upon observing a preventableaccident

(e.g., environmental disaster), cannot observe whether the type of the defendant (firm) is either reckless or cautious.

A reckless type of firm has a poorer managerial ability adopting safety measures and employee training than the cau-

tious type of firm. For instance, in the 2010 Deepwater Horizon oil spill (orBP oil spill), Judge Carl J. Barbier of Federal

District Court in New Orleans recognized that, in order to infer BP’s negligence in the spill, he had to deal with conflict-

ing arguments from the government and the company(e.g., he wrote that the evidence was “voluminous, dense, highly

technical, and conflicting”; see Schwartzjan (2015). The structure of the game is the following: in the first period, the

firm decides its precaution level, either high or low; and the judge, after observing such investment, chooses punitive

damages based on either a share of the firm’s revenue (wealth approach) or as a proportion of compensatory damages

(ratio approach).3In the second period, the firm can appeal the judge’s verdict, as in the case of the Exxon Oil spill,

which the judge can accept or reject.

We identify a separating equilibrium where the cautious firm chooses a high precaution level while the reckless

type invests in low, which emerges when investment costs are intermediate. In this setting, the judge can infer the

firm’s type upon observing its investment decision, leading him to a correct verdict, that is, the ratio approach after

a high precaution but the wealth approach otherwise. As a consequence, in this equilibrium outcome no appeals are

filed, saving time and resources to both parties. We also find a pooling equilibrium where both types of firm invest

in high precaution, which emerges when investment costs are relatively low. In this context, the firm’s investmentin

precaution conceals its type from the judge, who responds with the wealth or ratio approach depending on his prior

beliefs (i.e., frequency of cautious types). Hence, the cautious type needs to appeal in the second period when the judge

responds with the wealth approach; an appeal that is only due to the judge’s lack of information in the first period.4

Our results show that the emergence of these equilibria depends on (1) the burden of the wealth approach (as a

percentage of revenue); (2) the size of the punitive-to-compensatory ratio; and (3) the legal cost that the firm incurs

when appealing the judge’s decision. First, more severe punitive damages based on the wealth approach shrink the

range of parametersunder which the (informative) separating equilibrium arises but expand those for which the (unin-

formative)pooling equilibrium exists. Intuitively, in the separating equilibrium the reckless type has more incentives to

deviate toward a high investmentas the wealth approach it receives in equilibrium becomes more costly. The opposite

argument applies in the pooling equilibrium where both types of firm receive the ratio approach. In this case, a devia-

tion toward low precaution would be followed by more costly punitive damages based on the wealth approach. Hence,

both types of firm have less incentivesto deviate.

Second, if the punitive-to-compensatory ratio increases, the separating (pooling) equilibrium is sustained under

larger (more restrictive) parameter values. In the separatingequilibrium, the reckless type has less incentives to devi-

ate toward high precaution since the gap in punitive damages between the ratio and wealth approach diminishes. By

contrast, in the pooling equilibrium the judge responds with the ratioapproach. Therefore, an increase in the punitive-

to-compensatory ratio provides more incentives for firms to deviatetoward a low investment. Hence, judicial systems

that recommend punitive damages to be a small share of the defendant’s revenue using the wealth approach and/or

large punitive-to-compensatory ratios promote the emergence of the separating equilibrium in which information

flowsfrom the defendant’s actions to the judge, ultimatelyinducing correct verdicts and no further appeals. In contrast,

a legal system requiring defendants to paya large percentage of their revenue in punitive damages and low punitive-to-

compensatory ratios would reduce firms’ incentives to behaveas under the separating equilibrium. As a consequence,

these punitive damages would prevent information transmission,leading to more likely incorrect verdicts followed by

appeals.

3One of the main reasons for courts to consider the defendant’s wealth is to deter and punish highly reprehensible conducts (see Orr,2004). For instance,

in California severalcourts have not allowed punitive damages to exceed 10% of the defendant’s net worth. Regarding the ratio approach, the U.S. Supreme

Courtsuggests that a punitive-to-compensatory ratio below 4:1 is reasonable. Eisenberg, LaFountain, Ostrom, and Rottman (2002) and Segalla (2005) discuss

the factors affecting the selection of punitive damages. However,in cases where a settlement is reached, the punitive-to-compensatory ratio is often below

one.For instance, in the BP oil spill, compensatory damages were above $20 billion (considering both cleanup costs and payments to harmed individuals), while

inthe settlement between BP and the Justice Department, BP agreed to pay $5.5 billion.

4Wealso find the existence of a pooling equilibrium in which both types of firm choose to invest in low precaution. However, we show that such equilibrium

doesnot survive the Cho and Kreps (1987) intuitive criterion because it is based on insensible off-the-equilibrium beliefs.