Moral Judgment and Moral Heuristics in Breach of Contract

University of Pennsylvania Carey Law School University of Pennsylvania Carey Law School

Penn Carey Law: Legal Scholarship Repository Penn Carey Law: Legal Scholarship Repository

All Faculty Scholarship Faculty Works

2009

Moral Judgment and Moral Heuristics in Breach of Contract Moral Judgment and Moral Heuristics in Breach of Contract

Tess Wilkinson-Ryan

University of Pennsylvania Carey Law School

Author ORCID Identi;er:

Follow this and additional works at: https://scholarship.law.upenn.edu/faculty_scholarship

Part of the Commercial Law Commons, Contracts Commons, Law and Economics Commons, and the

Law and Psychology Commons

Repository Citation Repository Citation

Wilkinson-Ryan, Tess, "Moral Judgment and Moral Heuristics in Breach of Contract" (2009).

All Faculty

Scholarship

. 2573.

https://scholarship.law.upenn.edu/faculty_scholarship/2573

This Article is brought to you for free and open access by the Faculty Works at Penn Carey Law: Legal Scholarship

Repository. It has been accepted for inclusion in All Faculty Scholarship by an authorized administrator of Penn

Carey Law: Legal Scholarship Repository. For more information, please contact biddlerepos@law.upenn.edu.

Moral Judgment and Moral Heuristics in Breach of Contract

ABSTRACT

Most people think that breaking a promise is immoral, and that a breach of contract is a

kind of broken promise. However, the law does not explicitly recognize the moral context of

breach of contract. Using a series of web-based questionnaires, we asked subjects to read breach

of contract cases and answer questions about the legal, financial, and moral implications of each

case. Our results suggest that people are quite sensitive to the moral dimensions of a breach of

contract, especially the perceived intentions of the breacher. In the first study, we framed the

motivation for a contractor’s breach as either the chance to make more money or the chance to

avoid losing money. Subjects were more punitive when the motivation appeared to be greed

(breach to gain) than when the motivation appeared to be fear (breach to avoid loss). In the

second study, we manipulated the timing of the negotiation over damages, comparing cases in

which the promisor asks to negotiate damages before definitively breaching (as in a liquidated

damages clause) with cases in which the breach has already occurred. We predicted that once the

contract is breached, the moral violation becomes very salient, and we found that subjects were

less punitive when they setting damages ex ante. Finally, results from the third study suggest that

subjects seemed to believe that intentionally breaking a contractual promise is a punishable

moral harm in itself. When presented with identical losses, one from an intentional breach of

contract and the other from a negligent tort, subjects were more punitive toward the breacher

than the negligent tortfeasor. They treated willful breach as an intentional harm.

Moral Judgment and Moral Heuristics in Breach of Contract

Tess Wilkinson-Ryan, J.D., M.A.

Penn Faculty Fellow, University of Pennsylvania Law School

Jonathan Baron, Ph.D.

Professor of Psychology, University of Pennsylvania

Journal of Empirical Legal Studies, forthcoming

INTRODUCTION

Most people agree that breaking a promise is immoral. Because the legal construct of

contract is tied so closely to the moral notion of a promise, breach of contract would seem to fall

into the same category of moral harm as a broken promise. However, the law does not explicitly

recognize the moral context of breach of contract. The penalty for breach of contract is limited

(with some exceptions) to money damages in the amount necessary to put the non-breaching

party in as good a position as he would have been in had the contract been performed.

contract, a damages award does not depend on the reason for breach, and there are no punitive

damages for intentional breach of contract.

This means that there are no legal distinctions

between cases that elicit very different moral intuitions; morally salient factors like the motives

and intentions of the breacher are legally irrelevant. Though this may seem straightforward to

lawyers and academics, many people find this rule at odds with the idea of the moral obligation

of a contract. In this paper, we argue that commonsense notions of contract and promise find

RESTATEMENT (SECOND) OF CONTRACTS § 264

E. ALLEN FARNSWORTH, Contracts § 12.3 (3

ed. 1999) (“The principal interest protected by

that law is the expectation interest, measured by the amount of money required to put the injured

party in as good a position as that party would have been in had the contract been performed…

‘Willful’ breaches should not be distinguished from other breaches.”)

contract law generally counter-intuitive and efficient breach downright objectionable. Not only

does our legal system appear indifferent to a moral harm (breaking a promise), it permits

breachers to profit from a moral violation. Here, we ask when laypeople consider breach to be

immoral, which moral principles and moral heuristics they employ to make that judgment, and to

what extent their moral reasoning affects their legal and financial decision-making.

Our results suggest that people are quite sensitive to the moral dimensions of a breach of

contract. Subjects evaluated a series of breach of contract cases, and we asked them to set

damages awards and to answer questions about the moral culpability of the actors. In the first

study, we framed the motivation for a contractor’s breach as either the chance to make more

money or the chance to avoid losing money. Subjects were more punitive when the motivation

appeared to be greed (breach to gain) than when the motivation appeared to be fear (breach to

avoid loss). Subjects indicated that punitive damages were appropriate for the intentional, greedy

breacher. We also manipulated the timing of the breach, comparing cases in which the promisor

asks to negotiate damages before definitively breaching with cases in which the breach is a fait

accompli. We predicted that once the contract is breached, the moral violation becomes very

salient. In fact, subjects suggested a lower penalty for breach when they setting damages ex ante.

Finally, subjects appeared to equate willful breach of contract with intentional harm. They set the

damages for willful breach of contract at a higher level than the same loss caused by negligence.

I. NORMATIVE AND DESCRIPTIVE THEORIES OF MORALITY IN CONTRACT

Damages for breach of contract are measured in terms of the promisee’s subjective

expected benefit.

In a breach of contracts case, the " victim"/promisee is awarded the full value

of the benefit of the contract, so her interests are actually advanced to the point that they would

RESTATEMENT (SECOND) OF CONTRACTS § 344; FARNSWORTH, supra note 4, § 12.1.

have been had the contract been completed, not simply returned to their ex ante position, as in a

torts case.

However, punitive damages are not awarded in contracts, though consequential

damages are a permitted in the event that the breach has caused other losses to the promisee.

Specific performance is rarely awarded, and is reserved for the few cases in which money

damages would be unable to redress the promisee's interests.

Under these rules, the breacher's

motivation for the breach is mostly irrelevant. There are no legal differences between a case in

which a job becomes unprofitable (say, due to a rise in the cost of materials) and a case in which

a job becomes less appealing due to a more lucrative offer elsewhere

, though we will offer data

to suggest that most people make a moral distinction between these cases.

Economic theory supports the rule of expectation damages, which arguably provides

optimal incentives for efficient contracts (and efficient breaches).

It is not worth it to the

promisor to breach the contract unless the cost of performance exceeds the benefit of

performance to the promisee. That is, the promisor's own self-interested calculation of costs and

FARNSWORTH, supra note 4, §12.1.

FARNSWORTH, supra note 4, § 12.9.

Id. at § 12.6 (“[E]quitable relief would not be granted if the legal remedy of damages was

adequate to protect the injured party.”) See also, T. Anthony Kronman, Specific Performance. 45

U. CHI. L. REV. 351, 365 (1978) (arguing that the adequacy test for money damages “draws the

line between specific performance and money damages in the way that most contracting parties

would draw it were they free to make their own rules concerning remedies for breach and had

they deliberated about the matter at the time of contracting.”)

See Globe Ref. Co. v. Landa Cotton Oil Co., 190 U.S. 540, 544 (1903) (in which Holmes notes

that “if a contract is broken the measure of damages generally is the same, whatever the cause of

the breach.”) Damages in contract are meant to be compensatory, not punitive. See FARNSWORTH

§ 12.8 (“it is a fundamental tenet of the law of contract remedies that an injured party should not

be put in a better position than had the contract been performed.”) For a judicial argument in

favor of the efficiency of this position, see Patton v. Mid-Continent Sys., 841 F.2d 74f2 (7

Cir.

1988) (in which Judge Posner argues that “even if the breach is deliberate, it is not necessarily

blameworthy. The promisor may simply have discovered that his performance is worth more to

someone else. If so, efficiency is promoted by allowing him to break his promise, provided he

makes good the promisee’s actual losses.”)

See, e.g., RICHARD POSNER, ECONOMIC ANALYSIS OF LAW ch. 4 (3

ed. 1986); STEVEN

SHAVELL, FOUNDATIONS OF THE ECONOMIC ANALYSIS OF LAW, 2004.

benefits of breach integrates the benefit of the contract to the promisee. This forces the promisor

to internalize the costs of breach to both parties, and encourages the promisor only to breach in

cases in which the breach would be overall welfare-maximizing. Were the damages set at a lower

level, the promisor might have an incentive to breach more often and in cases in which the

overall welfare of the parties would not be maximized via a breach. If the damages are set at a

higher level, or if specific performance is required, promisors may be deterred from breaching in

cases in which the cost of performance is higher than the value of the performance to the

promisee, which is inefficient on its face. In addition, a higher level of damages increases the

promisor's risk; some contractors may be unwilling to pay extra to compensate for this risk, so

that contracts that could benefit both parties would not be made.

Of course, there may still be sound moral arguments against efficient breach. The

traditional moral view holds that a contract is a promise, and that breaking a promise is immoral.

The promise theory of contract relies on the underlying autonomy of individuals insofar as they

can choose to obligate themselves in such a way that courts must respect (e.g., enforce) their

self-imposed obligation.

Under this theory, expectation damages are justified based on the

freely-undertaken promise.

The moral view of contracts seems to be in line with common

notions about what it means to make a promise. However, even the promise theory does not

argue with expectation damages as the correct remedy for breach of contract. We will present

evidence that many people find expectation damages inadequate to remedy an intentional breach.

Norms, Schemas, and the Psychological Contract

The sanctity of a promise is an unusually resonant moral principle. Cultural psychologists

See FRIED, supra note 1.

Id. at 18-20.

have identified the rule of contract as one of only three universal moral norms.

In economics

experiments, subjects who are otherwise willing to lie about their intentions in a cooperation

game can be trusted to keep their word if they have explicitly promised, even without threat of

sanctions.

Our prediction is that people hold such strong beliefs about the moral rules of

promise and contract that they will use these rules to inform their legal decision-making.

When a contract is incomplete or informal, people use their own ethical principles to fill

in the spaces. Stuart Macaulay reported in 1963 that many, if not most, businessmen actually had

a preference for relying on “common honesty and decency” or industry and social norms rather

than formal contracts.

“Commitments are to be honored in almost all situations; one does not

welsh on a deal.”

More recently, behavioral researchers have begun to elaborate on the notion

of what it means to honor a commitment. Sandra Robinson and Denise Rousseau introduced the

concept of the “psychological contract,” a construct meant to encompass the beliefs or

perceptions of the parties about the conditions of a reciprocal exchange.

This is particularly apt

in the employment context, their primary focus, because even when a written employment

contract exists, it is unlikely to iterate the various contributions, obligations, and inducements

that characterize an employer/employee relationship over time. Robinson and Rousseau found

that perceived violations of the psychological contract have real effects for employers, including

higher turnover and lower employee satisfaction. Like Macaulay’s study, their research suggests

Paul Robinson, Robert Kurzban & Owen Jones, The Origins of Shared Intuitions of Justice, 16

VAND. L. REV. 1633 (2007).

Robyn Dawes, Jeanne McTavish & Harriet Shaklee, Behavior, Communication, and

Assumptions about Other People’s Behavior in a Common Dilemma Situation, 35 J.

PERSONALITY & SOC. PSYCH. 1 (1977).

Stewart Macaulay, Non-Contractual Relations in Business: A Preliminary Study, 28 AM. SOC.

REV. 55 (1963).

Id. at 63.

Sandra L. Robinson & Denise M. Rousseau, Violating the Psychological Contract: Not the

Exception but the Norm, 15 J. ORG. BEH. 245. (1994).

that parties look to moral norms governing the unwritten contract to define the notion of

performance and breach.

In the case that the contract is formal and adequate to the situation, lay intuitions are still

relevant in how people interpret its terms. Dennis Stolle and Andrew Slain studied the effects of

exculpatory clauses in contracts on consumer behavior.

They found that exculpatory language

in a standard form contract had a deterrent effect on subjects’ likelihood to seek compensation,

even when it is not clear that the exculpatory clause would be legally enforceable. Parties’ beliefs

about the contract were informed by the terms of the contract itself as well as their intuitions or

beliefs about contracts in general, namely, that they are enforceable as written. In this study, we

are extending these questions about contract schemas into the domain of the remedy for breach

of contract, asking how people define the contractual obligation and what they require for

compensation.

Moral Heuristics in Contract

People’s beliefs about contracts are affected by their moral intuitions, even when those

intuitions are inconsistent or unreasonable. Cass Sunstein has recently catalogued a partial list of

moral heuristics—short-cuts, or rules of thumb—that people use in order to make moral

judgments.

A heuristic need not be normatively wrong in an absolute sense; rather, heuristics

are just simple rules, often rules that are useful in most of the cases in which they apply.

Nonetheless, as Sunstein argues, moral judgments based on these kinds of rules of thumb may be

less sound than moral judgments that result from deliberative reasoning.

In this paper, we will

Dennis P. Stolle & Andrew J. Slain, Standard Form Contracts and Contract Schemas: A

Preliminary Investigation of the Effects of Exculpatory Clauses on Consumers’ Propensity to

Sue, 15 BEH. SCIENCES AND THE LAW 83 (1997).

Cass Sunstein, Moral Heuristics, 28 BEH. & BRAIN SCIENCES 531 (2005).

Id. at 534-535.

be considering the kinds of heuristics that might affect moral judgment in the field of contracts.

To the extent that subjects find breach of contract to be morally wrong, prior research

suggests that they will focus on the punishment that the breacher “deserves” rather than the rules

that would create the most efficient incentives. The rule of expectation damages and efficient

breach sets optimal economic incentives for the making and breaking of contracts. However, we

know that people are often insensitive to these kinds of incentives, especially when the incentive

structure fails to adequately punish adequately for moral harms. Jonathan Baron and Ilana Ritov

have studied intuitions about penalties and compensation in tort law.

Though subjects were

highly sensitive to the moral distinctions, they ignored the information about deterrence

altogether, uniformly imposing punishments based on the moral rule that the punishment should

be proportionate to the outrageousness of the act, whether the punishment would be useful,

pointless, or even harmful.

In breach of contracts cases, people must make similar judgments.

Our hypothesis is that they will be highly sensitive to the moral dimensions of efficient breach,

ignoring the incentive effects of various possible policies. In the criminal context, John Darley

and Paul Robinson have repeatedly shown that subjects are more sensitive to morally salient

information than they are to factors associated with deterrence rationales, and that people punish

in line with retributive theories of justice.

This evidence suggested to us that people might also

be more averse to losses coming from someone who has promised to confer a benefit (e.g., a

promisor) than from someone with a neutral status (say, a negligent tortfeasor).

With the legal, moral, and economic arguments of normative theorists in mind, and in

Jonathan

Baron & Ilana Ritov. Intuitions about Penalties and Compensation in the Context of

Tort Law, 7 J. RISK & UNCERTAINTY 33 (1993).

Id.

Paul Robinson & John Darley, The Utility of Desert, 91 NW. U. L. REV. 453 (1997). Kevin

Carlsmith, John Darley & Paul Robinson, Why Do We Punish: Deterrence and Just Deserts as

Motives for Punishment, 83 J. PERSONALITY & SOC. PSYCH. 284 (2002).

light of evidence from psychology and behavioral economics, we undertook empirical research

to examine the nature of moral intuition and moral judgment in breach of contract cases.

II. EMPIRICAL STUDIES

Subjects in all studies were members of a panel recruited over a 10-year period, mostly

through their own efforts at searching for ways to earn money by completing questionnaires.

Approximately 90% of respondents were U.S. residents (with the rest mostly from Canada). The

panel is roughly representative of the adult U.S. population in terms of income, age, and

education

but not in terms of sex, because (for unknown reasons) women predominate in our

respondent pool.

For each study, we sent email to about 500 members of the panel, saying how much the

study paid and where to find it on the World Wide Web. Each study was a series of separate

web pages, programmed in JavaScript. The first page provided brief instructions. Each of the

others presented a case, until the last, which asked for (optional) comments and sometimes

contained additional questions. Each case had a space for optional comments. (We report some

comments in the discussion of the results to suggest possible conclusions and implications of the

quantitative results. Because they were optional, comments were not coded and analyzed

systematically.) Otherwise the subject had to answer all questions in order to proceed. The order

of cases was randomized (with one exception). The study was removed when about 75

responses had been submitted, with a target of 80. The studies are all available at

http://www.psych.upenn.edu/~baron/tess/. In a pilot study we asked subjects to read a series of

breach of contract scenarios and to answer questions about them. Subjects showed some

Linda Babcock, Michelle Gelfand, Deborah Small & Heidi Stayn. The Propensity to Initiate

Negotiations: Toward a Broader Understanding of Negotiation Behavior. Manuscript, Carnegie-

Mellon University (2003).

willingness to demand damages higher than expectation level as well as a preference for specific

performance as a remedy for breach.

B. Experiment 1: Loss Aversion and Motives for Breach

We predicted that subjects would be more sympathetic to a breacher who breaks a

contract in order to avoid a loss than an otherwise identical breacher who breaks a contract in

order to accept a more lucrative deal elsewhere. This is essentially a framing effect: in both

cases, the breacher will make more money by breaking the contract than by honoring it.

However, because people are more sensitive to the prospect of a loss than a foregone gain, we

hypothesized that they would find the notion of breaching in order to gain to be more offensive

than breaching to avoid a loss. We assume that subjects take the status quo to be the expected

benefit of the original contract, such that a lower profit for the promisor is seen as a loss. Thus,

we predicted that subjects would impose higher penalties in a breach to gain case than in a

breach to avoid loss case, and that they would find the former more morally objectionable than

the latter.

I. Method

In Experiment 1, 83 subjects participated, with 78% female. Ages ranged from 22 to 69,

with a median age of 42. In this study, we used a between-subjects manipulation, where half the

subjects read about a breach to avoid loss, and the other half read about a breach to gain. The

scenario described a contract for a kitchen renovation. In the Avoid Loss condition, subjects read

that "the contractor learns that the price of cabinets and countertop has skyrocketed, and the

contract price will barely cover the cost of materials. He decides to break his contract in order to

take other, more profitable work."

In the Gain condition, subjects read that "the contractor learns that there is a shortage of

skilled renovators in a nearby area, and he could charge much more there for a similar project.

He decides to break his contract in order to take other, more profitable work."

Subjects were asked to choose an appropriate amount of damages. We constructed an

eight-point scale based on the available numbers in the scenario. For these cases, subjects could

choose among the following 8 options: no compensation; a number between nothing and

expectation damages (e.g., “more than 0 but less than ...”); expectation damages; a number

between expectation damages and the difference between the buyer's reserve and the seller's

reserve; the difference between the buyer's reserve and the seller's reserve; more than the

difference between buyer's reserve and seller's reserve but less than the contract price; the

contract price; or a number greater than contract price. For the purposes of analysis, we used the

choices as points on a scale from one to eight.

Subjects were also asked to indicate to what extent the breacher should feel guilty should

he choose to pay the damages and break the contract.

II. Results

Subjects who saw the breach to gain scenario set the damages at a mean level of 4.08,

while subjects who saw the breach to avoid loss scenario set the level of damages at 3.14

(difference is significant, t=2.592, df=76.40, p=.0114), where 3 was expectation level. Subjects

did not assign significantly different levels of guilt on a three-point scale in the two conditions,

In this case, expectation damages were $3,000. Subjects chose from the following options:

1) No compensation

2) More than nothing but less than $3,000

3) $3,000

4) More than $3,000 but less than $9,000

5) $9,000

6) More than $9,000 but less than $20,000

7) $20,000

8) More than $20,000

though the small existing difference was in the hypothesized direction (more guilt in the breach

to gain scenario).

Using the same scenario, we saw the same main effect in a within-subjects design

, with

subjects assigning higher damages in the Gain condition than the Avoid Loss condition (mean

difference on 8-point scale=.341, t=2.303, df=87, p=.0237). We also found that subjects assigned

significantly more guilt in the Gain condition than in the Avoid Loss condition (mean difference

on a 3-point scale=.261, t=3.663, df=87, p=.0004).

C. Experiment 2: Timing

In this study, we wanted to know whether the time of the damages negotiation would

matter. We thought that once the breach is a fait accompli, the most salient fact in the scenario is

the broken promise. However, there are usually other opportunities to negotiate damages before

the breach has occurred. Therefore, we hypothesized that subjects would set lower damages if

they were negotiating the cost of a possible future breach rather than assigning blame once the

contract is irreparably broken.

We also wanted to collect initial data that would help us understand the origins of

subjects' punitive attitudes toward breach of contract cases. One possibility was that subjects

were not being prompted to think of the economic consequences of their judgments, that they

were not calculating the costs in a way that would allow them to see that expectation damages

would fully compensate the promisee. Another implication of the misunderstanding hypothesis is

that subjects would not see the economic incentives for expectation damages, both from the point

of view of general economic welfare and from the point of view of the promisee, who would be

Within-subjects data were collected as part of Experiment 2, where the gain/avoid loss

manipulation was simply tacked on to the beginning of the timing experiment so that we could

see within-subjects results of the framing manipulation.

unable to secure a contract at all under a system in which the penalties for breach were so

burdensome as to be a deterrent to contract at all. Another possibility is that subjects would

explicitly choose to impose an additional penalty, reasoning that the breach was a moral harm of

some kind and therefore should be discouraged. In Experiment 2, we asked subjects to indicate

the optimal level of damages overall, then from an economic point of view, and finally from a

moral point of view.

Because we were particularly interested in subjects' explicit attitudes, we used entirely

within-subjects tests for this study.

1. Methods

88 subjects participated in Experiment 2; 67% were female. Subjects ranged in age from

18 to 64, with a median age of 35.5.

This study includes a total of 5 screens, or 5 scenarios. The first two are identical to those

in Experiment 1: the kitchen renovation breach to gain, and the kitchen renovation breach to

avoid loss. The next three scenarios tested the timing hypothesis, and all involve the same

vignette describing a contract for a party rental. A couple, the Bakers, rents out a local restaurant

for their anniversary party, and the owner of the restaurant, Dave, is offered a much bigger fee

for the same night. We showed subjects three versions. In the first version, subjects read that the

couple "discuss with Dave what would happen if he breaks the contract. Then the Bakers set the

date and invite friends and family." In the second version, the owner approached the couple with

a request to negotiate an amount that would release him from the contract: "Dave calls to cancel

the Bakers' reservation...He informs Mrs. Baker that another party loved the space so much that

they offered $25,000 for the same night. Dave and Mrs. Baker discuss the situation and agree on

a fee that Dave will pay to release him from their contract." Finally, in the third version, the

breach is a foregone conclusion and the couple is not invited to negotiate the damages amount.

Subjects are told that an "impartial mediator to the situation must now decide whether and how

much Dave should compensate the Bakers." Each subject saw all three versions of the scenario.

We asked subjects to indicate the fee or penalty that they would set.

We also asked

subjects to indicate the optimal level of damages from two other perspectives. We asked them to

indicate the most efficient damages level: "If you wanted to set a policy that would yield the best

economic outcomes overall, what compensation level would you choose?" We also asked them

to indicate the level of damages most in line with "the best moral rule."

2. Results

Subjects assigned lower damages when the amount was negotiated rather than decided

after the breach. As we described above, we had three conditions: a liquidated-damages clause in

the contract; a request to be released from the contract; and a simple breach. The mean scale

values for compensation (where expectation damages are 3) were, respectively, 4.51, 4.53, and

4.88. The first two conditions did not differ, but the request and simple breach differed

significantly (t=2.079, df=87, p=.0405). In sum, subjects favored higher damages once the

promisor has definitively breached, but are slightly less punitive when they are asked to set a fee

for breach before the contract is irreparably broken.

Recall that in this study we also tested a within-subjects version of the framing study

In this case, expectation damages were $2,000. Subjects chose from the following options:

1) No compensation

2) More than nothing but less than $2,000

3) $2,000

4) More than $2,000 but less than $3,000

5) $3,000

6) More than $3,000 but less than $11,000

7) $11,000

8) More than $11,000

reported in Experiment 1.

Finally, we thought that subjects might indicate that the economically optimal level of

damages would be lower than the level that they set for the breacher. They did not. However,

they did indicate that the damages best reflected by the moral rule would be higher than the level

of damages they chose. Averaging across all the scenarios in Experiment 2, subjects set the

"moral damages" .422 points higher than their own damages assignment (t=5.302, df=87,

p<.001). This effect interacted significantly with the framing manipulation. The difference

between the moral penalty and the level of damages that subjects actually set was bigger in the

breach-to-gain case than in the avoid-loss case (t=2.341, df=87, p=.0216). The effect did not

interact significantly with the timing manipulation.

Experiment 3: Is a contract a promise?

In the first two experiments, our results suggested that people are often willing to impose

damages well above expectation level, and that they find breach morally objectionable (or at

least guilt-inducing) even when damages are paid. In this study, we want to ask what, exactly, is

the moral content of a contract. Our hypothesis is that subjects consider breaking a promise to be

a harm in itself, and that they will punish an intentional harm more severely than a negligent

harm.

In order to test this experimentally, we compared breaking a contract with causing an

identical loss negligently. Each of three scenarios was presented in two different versions. In one

version, subjects read a case of efficient breach, in which the promisor breaks the contract in

order to accept a more lucrative contract elsewhere. In the control condition, the same contract is

rendered impossible to complete when a third party negligently causes a harm that in turn

prevents performance. In this case, we are asking subjects not about the damages that would be

paid by the promisor, who is presumably excused from the contract, but rather the liability

damages paid by the third party. So, for example, either a contractor breaks the contract to take

another job, or a contractor is prevented from completing the contracted-for job because of a gas

leak negligently caused by a third party. In each condition, the financial harm is identical, and

the harm is confined to the harm of not being able to realize the benefit of the contract.

In this study, we made several important changes and additions. First, we asked subjects

to name the dollar amount of damages themselves, rather than asking them to choose among

levels of compensation as in the first two studies. Though this method introduces the problem of

anchoring effects and subject typing errors, it permits subjects to name an exact amount of

damages. Second, we directly asked subjects about the morality of the act of breaching the

contract rather than relying on the subjects' assignment of guilty feelings to the breacher (or on

the appropriate compensation from a moral point of view). Third, we asked subjects what the

promisor should do (breach or perform), and whether the law should force the promisor to honor

the contract.

1. Method

80 subjects answered a questionnaire on the World Wide Web. 76.3% of subjects were

female. Subjects ranged in age from 24 to 79, with a median age of 44. Subjects read six

scenarios and answered questions about each. In this study, there were three core scenarios

describing a contract. They describe contracts for backyard landscaping, floor refinishing, a party

rental, and a sale of used books. Below is one example:

Scenario:

The Millers are getting ready to sell their condo. Their real estate agent tells them that

their condo will be worth $10,000 more if they get the floors refinished (and you should

assume that the agent is correct). They meet with Todd, from Todd's Hardwood Floors,

agree on a price of $6,000 to refinish the floors, and they all sign the contract. Todd is

going to refinish the floors right before the condo goes on the market in early October.

Three of these scenarios were presented in one of two conditions, the promise condition and the

no-promise condition.

Promise Condition:

About three days before he is slated to work on the Millers floors, Todd gets an offer to

refinish all of the floors in another apartment building. If he accepts this offer, he will

make much more money, but he will not be able to refinish the Miller's floors. Todd

decides to take the new job and break his contract with the Millers.

No Promise Condition:

About three days before Todd is slated to start working on the Millers' floors, the owners

of the next apartment down, the Bakers, are trying to move a gas line in their own condo,

even though it is quite dangerous. They do not take proper precautions and the line

breaks, gas leaks, and the Millers' apartment is too toxic for Todd to do his work. The

Millers have already moved into their new home, so they are not personally affected by

the fumes. However, the fumes are quite toxic in the Millers' condo for more than a week,

so Todd is unable to refinish the floors before the condominium goes on the market.

(Note that Todd does not need to be compensated, because he gets a similar, highly

paying job for the same week once he is released from his contract with the Millers.)

In this study, we blocked the conditions. This means that half of the subjects first saw

each scenario in the promise condition, and then saw each scenario in the no promise condition;

the other half of the subjects saw the scenarios in the reverse order (first all no-promise, then all

promise). Within the blocks, the order of scenarios was random. We did this for two reasons. The

first is simply that we thought it would be less confusing for subjects. The second is that it

permitted us to do between-subjects analyses of their responses. By looking only at the first

block of responses for each subject, we can see their responses unaffected by their implicit

comparisons to the other condition.

2. Results

Our between-subjects tests of guilt and morality indicate that subjects thought that a

person who caused harm via breaking a contractual promise was more immoral and should feel

more guilt than a person who caused harm via negligence. On a 7-point scale where 7 is

extremely immoral and 1 is not immoral at all, the average rating of for a negligent wrongdoer

was 3.3, and the average rating for a breacher who caused an identical loss was 5.1 (test of the

difference is significant: t=5.281, df=74.58, p<.0001). Similarly, on a 7-point scale, subjects on

average felt that a negligent wrongdoer’s guilt should rate a 3.3, and a breacher’s guilt should

rate a 5 (Within-subject comparisons were also significant at p<.001. Test of the difference is

significant: t=4.686, df=76.01, p<.0001).

The responses to the question about damages were open-ended, and their distribution was

highly skewed. Taking the logarithm removed the skewness, but the distributions still had long

tails on both sides (i.e., an excess of extreme responses both high and low). Accordingly we

report trimmed means (10% on each side) based on the logarithms, but we use nonparametric

tests for inference. We report both t-tests and Wilcoxon tests for the tests of differences below.

Table 1 shows the means for each case in each condition, shown as a ratio of expectation

damages (Thus, if the promisee’s expectation was $1,000 and the average subject response was

$2,000, the ratio is 2. If the average subject response was $1,000, the ratio is 1. This was based

on the antilog of the mean log.)

Case 1: Refinish

Floors

Case 2: Landscape

Yard

Case 3: Cater Party

Promise: Mean ratio

1.12

3.12

2.33

No Promise: Mean ratio

0.82

0.79

1.58

We analyzed the difference between the promise and no-promise condition in two ways.

First, we looked at our between-subjects measure, comparing the promise responses of the group

who saw the promise cases first to the no-promise responses of the group who saw the no-

promise responses first. Though the pattern of means and log ratios were consistent with our

hypothesis (higher damages in the promise condition), the test of differences was not statistically

significant (t=.4142, df=74.99, p=.6799. The Wilcoxon test also showed non-significance.). We

then looked at the within-subjects comparison. For this analysis, we took the mean of the

difference of each subject’s promise and no-promise damages responses. This was a highly

significant difference (t=4.318, df=79, p<.0001. Wilcoxon test, p=.001), meaning that when they

saw both kinds of case, subjects imposed much higher damages overall in the promise cases than

in the no-promise cases.

Finally, subjects were also inclined to report that the promisors should honor the contract

rather than pay damages and breach (even though it was clear in every case that the promisor

would be better off by breaching). Responses were fairly consistent across cases, with an average

of 75.8% of subjects responding that the promisor should honor the contract in a given case.

Subjects also indicated that they thought that the law should force the promisor, in many cases,

to honor the contract and perform. The average percentage of respondents who thought the law

should require specific performance was 66.7%.

II. DISCUSSION & IMPLICATIONS

A. Interpretation of Findings

Subjects distinguished between cases in which the promisor breaches in order to avoid a

loss of some kind and cases in which the promisor has been given a better offer; they imposed

higher damages on the latter and indicated that he should feel guiltier for his breach. These

contract scenarios seemed to tap into subjects’ moral intuitions about what it means to make and

break a promise. In typical usage, a promise is a promise to perform, not a promise to do

something as valuable as performance. Most non-contractual promises happen between people in

a non-market-based relationship.

In a personal relationship, it is inappropriate to offer money to

remedy a broken promise. Broken promises are harmful on two counts—the loss incurred by the

lack of performance, and the relational rift. This is not (or need not be) the situation, though,

with the homeowner and the flooring contractor or kitchen renovator. Their promises are only

commercial in nature, and the sole purpose of the contract is to provide each with a monetary

benefit. However, as long as subjects believe that breaking a contractual promise is a moral

violation, it is reasonable to think that breachers should not be permitted to profit (above the

expected benefit of the contract) from their intentional bad act. It may be for this reason that

subjects demanded higher damages when the incentive for breach was framed as a potential gain

rather than loss avoidance.

We also used a timing manipulation to compare subjects' responses to the question of what they

would demand, ex ante, to release the promisor from the contract, as opposed to how they would

apportion blame ex post. We predicted that subjects would give more moderate responses when

asked to consider a negotiated breach or a liquidated damages clause and more punitive when

asked to assign a penalty once the harm has been done. We thought that once faced with the fact

of the breach, the moral violation and the personal betrayal become very salient. Subjects were in

fact less punitive when asked to suggest an acceptable fee for releasing the contractor from his

obligation than when they were asked to levy damages after an irreparable breach.

Subjects’ responses in this study are in line with other findings about the effect of a sense

of betrayal on moral reasoning. In this study, we find that subjects imposed higher damages

when the cause of the lost benefit is the promisor than when the harm is caused by someone who

In psychological terms, we might think of the subjects as applying rules from the communal

sharing domain to the market pricing domain. See Alan Page Fiske, The Four Elementary Forms

of Sociality: Framework for a Unified Theory of Social Relations, 99 PSYCH. REV. 689 (1992).

has not made any promises. Sunstein suggests that people may see a betrayal of trust as an

independent harm. This is reasonable in the case of a personal promise. It is not clear, though,

that this reasoning is helpful for thinking through breach of contract cases in the legal context. If

expectation damages are to be fully enforced, the promisee's interests are protected. We can

argue that interpersonal trust is either not required (because external sanctions exist to enforce

the legal duties) or is not violated, because when the breacher pays damages he puts the promisee

in the position she would be in had the breach not occurred. It seems as though people reason

from a premise of trust, but in fact, the point of writing the contract out and making it legally

binding is to obviate the need for trust.

Finally, the most important implication of this study for law is that intuition treats a

contract like an obligation in tort. When a promisor breaches in order to make more money,

subjects want to treat him like an intentional tortfeasor, including levying punitive damages. In

Experiment 3, we explicitly contrasted two cases in which the promisee loses an expected

benefit. In the case of a negligence, subjects wanted the wrong-doer to pay an amount close to

the cost of the loss. However, when the wrong-doer intentionally breached the contract, subjects

consistently imposed higher damages, in some cases damages as high as two or three times the

cost of the loss.

These findings may also have implications for the ability of parties to make and break

contracts efficiently. People’s moral intuitions about contract law may make breach less frequent

than is economically efficient, and may inhibit the parties’ ability to settle out of court if there is

a breach. When parties disagree about appropriate damages in light of a breach of contract, they

may be less likely to settle out of court and more likely to undertake litigation. A possible

implication of our results is that parties will have different ideas about what constitutes a fair

resolution.

If parties think that a contract is a promise to perform, and not simply to confer a benefit

as valuable as performance, they may be less likely to breach at all. For individuals who are not

familiar with the rule of expectation damages (and we take this to be most laypeople), this would

be true because they believe that they are legally (and morally) obligated to perform. Researchers

have found that people believe that exculpatory clauses in contracts mean that they cannot seek

compensation.

A secondary finding in that line of research is that even though subjects

believed that exculpatory clauses could legally relieve a firm’s obligation to compensate its

consumers for harms it had caused, they did not think that contracts with exculpatory language

were less fair than contracts that left open the possibility of consumer compensation.

Our study

offers a kind of extension to these findings. In both cases, subjects thought that the parties were

morally bound by the specific language of the contract, even when contract law says that the

exculpatory clause is unenforceable or that the promisor can pay rather than perform.

More sophisticated parties may also be deterred from efficient breach because they do not

want to offend their customers or get a bad reputation. As we discussed above, empirical

research has demonstrated the real effects of psychological breach.

The loss of consumer trust

may have financial effects that override the potential profits from the breach, no matter that the

consumer’s judgment seems irrational. Parties to a contract will have different interactions with

one another and with the legal system depending on their beliefs about contracts. These beliefs,

in turn, may be partially informed by intuitions imported from the moral domain.

CONCLUSION

Stolle & Slain, supra note 24, at 91.

Id.

Robinson & Rousseau, supra note 23.

“The line between legal and moral guidelines is a very blurry one in my mind,”

commented one subject. “I don’t think a judge can clearly exercise one without the other having

some bearing.” Our results suggest that the connection between law and morality is not a

philosophical abstraction; for most people, it is an entrenched component of their intuitions about

legal decision-making. People’s moral responses to the scenarios in our studies affected their

legal judgments. Subjects imposed higher damages for parties whom they judged to be more

morally culpable. Empirical results like those we have presented here have bearing on practical

legal matters, including bargaining during contract drafting as well as negotiations over the

breach of a contract. These results may also bear on moral theories of breach of contract, as we

identify some discontinuities and tensions between intuition and reason.