PENALTY CLAUSES AND LIQUIDATED DAMAGES
This chapter surveys the economic literature on stipulated damages. In the
literature there seems to be a consensus that liquidated and underliquidated
damages should be respected. Liquidated damages can be a rational option,
especially if parties have more information about the possible losses than
judges. Underliquidated damages may can serve as a technique to let parties
share the risk of increased production costs.
Penalty clauses, on the other hand, have been the subject of a fierce
controversy for a long time. Most authors seem to defend the prohibition of
penalty clauses. Yet it can be argued that they should be allowed under some
JEL classification: K12
Keywords: Stipulated Damages, Underliquidated Damages, Remedies,
Breach of Contract
Sometimes parties to a contract ex ante agree upon how much compensation
will have to be paid should one of them breach the contract. These stipulated
damages are called ‘liquidated damages’ when they are ex ante reasonable
estimations of the true losses. They are called <underliquidated damages’
when they are meant to be undercompensatory and ‘penalty clauses’ when
they are deliberately overcompensatory in order to create an additional
sanction (‘penalty’). In common law penalty clauses are forbidden according
to the ‘penalty doctrine’. Liquidated and underliquidated damages on the
other hand are allowed.
This regulation of penalty clauses has received so much attention in the
law and economics literature that it deserves a separate section. The first
analyses did indeed not make clear why common law restricts the freedom of
contracting parties in this way. In 1979 Posner (1979, p. 290) described the
penalty doctrine as ‘a major unexplained puzzle in the economic theory of
law’. The penalty doctrine seemed to form an important exception to
Posner’s general thesis that common law is efficient.
Before starting to review the literature it might be helpful to pay
attention to the precise definition of the borderlines between penalty clauses
and liquidated damages.
Rea (1984a) has argued that much of the confusion around the penalty
doctrine occurs because of the failure to distinguish ex ante and ex post
valuation of losses. It is not because stipulated damages ex post turn out to
be overcompensatory that they will be considered penalty clauses by
American courts. American courts only check whether the clauses are
reasonable ex ante. When clauses are ex ante reasonable estimations of the
true losses, or to put it differently, when clauses reflect the expected losses,
they will be enforced.
In addition, liquidated damages are meant to compensate fully for losses,
including subjective and other hard to prove losses. It is possible that courts
mistakenly consider liquidated damages to be penalty clauses because they
underestimate subjective costs. Yet, from an analytical point of view these
clauses should be seen as liquidated damages. An early article of Goetz and
Scott (1977) can serve as an illustration of this confusion. Goetz and Scott
argued that a penalty clause is often just a way for a contracting party to
insure itself against idiosyncratic harm or otherwise uncompensated
damages. They give the example of a group of alumni who attach an
enormous subjective value to attending a baseball game of their college
team. Therefore, they stipulate in the contract with the bus driver that the
latter wil have to pay a high amount of damages in case he arrives late. This
amount is high but just reflects their true subjective losses. Moreover, the
promisor will very often be a better insurer than a traditional third party
insurer. This argument, however, is in fact just a plea for respecting
liquidated damage clauses and not necessarily an argument for
supracompensatory penalty clauses.
2. Liquidated Damages, Penalty Clauses and Underliquidated Damages
as Compared to Other Remedies
Many arguments for or against penalty clauses hold for other, less contested
remedies as well. Therefore, it is important to situate these sanctions within
the complete set of contract remedies, including the reliance measure, the
expectation measure and specific performance.
At first sight, liquidated damages seem to be equivalent to the
expectation measure. Liquidated damages are meant to fully compensate the
promisee and ‘fully’ includes the expected profits (only in cases were the
expected profits are compensated, the promisee is really indifferent with
regard to performance or breach). The major difference seems to lie in the
fact that liquidated damages are calculated ex ante (at the time of
contracting) while the expectation damages are calculated or made
operational ex post (after breaching). One major disadvantage of expectation
damages that is that they discourage efficient entering into contracts (see
Chapter 4600), which holds for liquidated damages as well. To put it
differently, if sellers demand liquidated damages in case of breach, fewer
buyers than optimal will sign the contract.
With respect to breaching decisions, however, there is a difference. The
expectation measure leads to efficient breaching (Birmingham, 1970;
Shavell, 1980). Liquidated damages, however, are determined ex ante, so at
the time the promisor decides whether to breach or to perform, he may
already know that the stipulated damages are over- or undercompensatory. If
they are overcompensatory, the promisor might overperform; if they are
undercompensatory, the promisor might overbreach. This difference may be
smaller than it seems. Judges do not compensate perfectly under the
expectation measure either. If the promisor expects the judge to
overcompensate he will overperform. If the promisor expects the judge to
undercompensate he will overbreach.
To summarize, liquidated damages may create incentives to
overperforming as well as to overbreaching. Whether incentives are worse
than under the expectation measure depends on how well judges valuate the
losses. If judges systematically and seriously undercompensate, then
liquidated damages may be the superior remedy when it comes to incentives
Penalty clauses and specific performance have in common that both are
supracompensatory sanctions. One consequence is that both lead to
overperformance, or to put it differently, that both discourage efficient
breaches. Penalty clauses, however, award an amount of money to the
promisee, while specific performance makes the promisor perform in natura.
At first sight this leads to an important difference: only penalty clauses
create an incentive for the promisee to induce breach. But much depends on
how specific performance works in practice. Under American law, nonperforming
promisors have to pay a fine on the basis of ‘contempt of court’.
This fine is paid to the court so it does not alter the incentives of the
promisee But in France, Belgium and the Netherlands, specific performance
is obtained by ‘astreintes’, high amounts of money (usually for each day of
non-performance), determined by a judge but to be paid to the promisee.
Astreintes create incentives to induce the other party’s breach as well.
An important difference between penalty clauses and specific
performance lies in the temporal dimension (De Geest, 1994); for a more
general discussion of the temporal dimensions of liability rules and property
rules see Levmore, 1997).
Specific performance needs some time before it becomes operational. It
works only for the future and not for the past. If by the time the promisee
goes to court performance in natura is no longer possible (or no longer
desirable), then courts will award just damages. Suppose someone had
promised to deliver a wedding cake at the promisee’s wedding, but did not
show up. Specific performance makes no sense afterwards, so awarding
damages is all the courts can do. On the other hand, penalty clauses can be
awarded for breaches in the past. This means that a specific performance
regime might in the real world be compensatory and that parties wanting
extracompensatory sanctions may have no choice but to include penalty
clauses in their contracts.
Finally, underliquidated damages can be the equivalent of anything inbetween
a perfectly applied expectation measure and no damages at all.
They may be the equivalent of reliance damages but also of an
undercompensatory expectation measure. Just as in the case of the reliance
measure they lead to overbreaching. The problem of undercontracting will
be smaller than under the expectation measure or even nonexistent, if they
are set just to compensate reliance damages.
A. Liquidated Damages
3. Liquidated Damages Avoid Ex Post Valuation Difficulties
With liquidated damages losses are estimated ex ante, at the time of
contracting. Such clauses avoid that judges have to compute the damages ex
post. It is well known that judges may have serious difficulties in finding out
the true losses. This holds especially for subjective harm. It is impossible for
a judge to know the promisee’s preferences precisely. Nor can he rely on
what the promisee tells him, because the latter has no incentive to reveal his
preferences in an honest way. This kind of preference revelation problem
does not arise when the loss is determined ex ante. At that time the parties
are still free to enter the contract or not. The higher the amount stipulated,
the higher the price: the debtor will require compensation for his additional
prevention costs and for the additional risk he has to insure. An ex ante
estimation is also useful for other forms of damage which are difficult to
prove. The costs of forgone chances are one example. Lacking clear
evidence, courts will underestimate such losses.
In this respect, liquidated damages reduce the risk of legal error. The
probability that a judge will err in estimating the damage is largely
eliminated. Of course the danger that a judge will erroneously consider the
clause to be a forbidden penalty clause still exists. The risk that a judge
wrongly applies the liquidated damage clause even though the debtor did not
breach the contract, also remains.
A somewhat related advantage of liquidated damage clauses is that they
allow us to avoid adverse selection problems. These problems arise when the
seller does not know how high the losses of the buyer will be in case of
breach. Because of this lack of information, the seller will assume ex ante
that the potential promisee represents an average risk. The doctrine of
foreseeability (Hadley v. Baxendale) serves to mitigate the adverse selection
problems. It forces people whose damage may be higher than usual to
communicate this to the seller. A liquidated damage clause can perform this
function as well. However one crucial condition is that the liquidated
damage clause is tailor-made. If liquidated damages do not take into account
the individual characteristics of the buyer, but consist of standard term
clauses, they no longer have this function.
4. Liquidated Damages Create Incentives for Optimal Reliance and
Cooter (1985) proved that damage clauses create a ‘double responsibility at
the margin’. First, the promisor receives an incentive to make an optimal
amount of precaution costs. Second, the promisee gets a perfect incentive not
to make too many reliance costs, because he will receive a fixed
compensation in case of a breach of contract, regardless of the amount of
reliance costs he has effectively incurred. This is not the case with normal
expectation or reliance damages awarded by a judge, because judges
generally do not examine whether the reliance costs made by the promisee
were excessive or not. To be sure, the same result could be reached with
damages that are judicially determined by compensating only for the
efficient reliance costs. In practice, however, insuperable difficulties of proof
often arise, which hamper the application of this remedy.
5. Liquidated Damages Substantially Increase Ex Ante Transaction
To estimate the losses ex ante is nearly always more expensive than ex post.
Ex ante, damages have be to estimated under all circumstances, even if ex
post no loss is incurred. All states of the world have to be taken into account,
while an ex post determination requires only information about the one
situation that has really occurred. Finally, ex post there is usually more
information on what really happened than ex ante. To put if differently, the
ex ante information costs are higher because what will happen is still
6. Ex Post Undercompensatory Liquidated Damages Lead to Ineffecient
Breaches, Ex Post Overcompensatory Liquidated Damages to
Liquidated damage clauses aim at correctly compensating the promisee’s
loss. However they are set ex ante, when full information may not be
available yet. If ex post the compensation turns out to be clearly lower than
the real damage, incentive problems arise, as briefly explained in Section 2.
If ex post the compensation turns out to exceed the actual loss clearly, the
promisee will have an incentive to head for an inefficient breach of contract.
The promisor will not breach the contract, even if it is efficient to do so. In
case of a breach of contract, the promisee is overinsured; the contract thus
contains a gambling clause. In return for this the creditor has had to pay a
Other aspects may outweigh this disadvantage. A liquidated damage
clause contains information on the possible magnitude of the loss. This
information is useful for the promisor who has to decide whether a breach of
contract is efficient or not. Because this information is produced in advance,
the promisor may make more efficient decisions. The promisor will also be
able to decide more quickly whether the good has to be sold to a third party.
For the same reason, a liquidated damage clause can also lead to a more
optimal level of prevention costs. If the loss or the possible loss is lower than
what the debtor would normally expect, it is also desirable that he makes
fewer prevention costs than usual. The opposite holds when the damage is
higher than usual.
B. Penalty Clauses
7. The Rationality Argument and the Signing Without Reading Problem
One earlier argument against the penalty doctrine was directly derived from
the rationality assumption in economics: we may assume that people are
rational, so if they sign a penalty clause they must have good reasons to do
so. The contract must be efficient, otherwise they would not have signed it.
This argument is implicitly used in the very first economic paper on the
subject (Barton, 1972, p. 286). Barton argued that all liquidated damage
clauses should be enforced without distinction, at least when the provision
was knowledgeably and fairly bargained-for.
The rationality argument is more explicit in Kronman and Posner
(1979). Criticizing Clarkson, Miller and Muris (1978) they argue that if
penalty clauses would really create an incentive for the promisee to induce
the promisor to breach, this danger would be reflected in the parties’
negotiations over the contract price or other terms of the contract. If the
penalty clause survives the negotiation process, this is presumably because
the benefits to the promisee exceed the costs to the promisor. According to
Kronman and Posner (1979) there may be room for intervention if one
thinks that the parties cannot assess these costs correctly, but then the basis
for intervention is paternalism. As a limitation on the freedom to contract,
the penalty doctrine is more paternalistic than the contractual incapacity of
minors or the invalidity of contracts of self-enslavement. A somewhat
similar idea can be found in Farber (1983) who argued that in addition to
contract law (the principle of freedom of contract) seems to embody a kind
of safety belt against individual catastrophic losses (including losses
resulting from penalty clauses).
Ulen (1984, p. 356) used the rationality assumption to defend penalty
clauses very explicitly ‘there is every reason to believe that they will
stipulate the most efficient remedy, considering all the factors’.
This argument, however, is not very convincing. Even though economic
science is based on the assumption of rational behavior, the existence of
information problems is generally accepted. ‘Rational’ does not mean
It cannot be denied that allowing penalty clauses implies an important
danger: signing without reading (Mackaay, 1982; De Geest, 1994).
Contracting parties do not always read or understand what they are signing.
The party drafting the contract may speculate on this and include certain
clauses which would never have been accepted if they had been read. It must
be clear that this problem is not merely theoretical: penalty clauses may ruin
one’s business and make others rich.
The signing without reading problem is elaborated by Mackaay (1982).
Contract terms are in general more costly to verify than most physical
features of commodities. Since most consumers will not inspect the terms,
the market may adapt ‘harsh-term-low-price’ policies. In order to counter
this perversity, some form of government intervention may be desirable.
Of course it would be undesirable to solve the signing-without-reading
problem by declaring all signed contracts void.
A better strategy is therefore to prove that under certain circumstances
some clauses will never be signed by rational, well-informed parties. The
fact that some people in the real world do sign such clauses can then be
considered as evidence that there was a serious procedural deficiency in the
formation of the contract (for a good example of this strategy see Rea,
8. The Costs of Legal Error are Higher for Penalty Clauses than for
Accepting a penalty clause is always a bit risky, even for those promisors
who perform exactly as promised. There is always a chance that a judge will
erroneously enforce a penalty clause, due to either a mistake or difficulties of
proof. A promisor who behaves optimally but cannot prove, for example,
that there was force majeure, could wrongly be sentenced to pay an
enormous amount. The risk of legal error is very hard to insure.
A comparison can be drawn with criminal law. If the judicial system
worked perfectly and if everybody behaved rationally, punishments could be
infinitely harsh. After all, a rational person would never commit a crime and
therefore nobody would be erroneously convicted. But just because the
judicial system works imperfectly, sanctions should not be infinitely harsh.
The same holds for sanctions for breach of contract.
9. Penalty Clauses Discourage Efficient Breaching
In the second edition of his textbook, Posner (1977, p. 93) suggested that the
law may refuse to enforce penalty clauses because they give an incentive to
complete the contract, even when breaching would be efficient. This
argument is correct, but should be put in context, as argued in Section 2. All
extracompensatory sanctions have this disadvantage. This applies to specific
performance, to liquidated damages that ex post turn out to be
overcompensatory, and to the expectation measure if judges systematically
overcompensate the promisees.
10. Penalty Clauses give Promisees an Incentive to Induce Breach of
This was first elaborated by Clarkson, Miller and Muris (1978). Suppose I
sign a contract with a builder. If the builder fails to perform in time, my true
losses are $1,000, but a penalty clause will award me $1,001,000. So I hope
the promisor will breach the contract and I will even do whatever I can to
make the promisor miss the deadline. However in many cases I will not have
any opportunity to mislead or hinder the debtor. Even if I have the
opportunity I must be able to do it covertly, because if the debtor can prove
that I am responsible for the breach, I will get no compensation. Covenants
not to compete, and clauses where the sole relation between the parties is
that of borrower and lender are a few examples given by Clarkson, Miller
and Muris (1978) where inducement of breach is no danger. But where
covertly induced breaching is possible, penalty clauses should be forbidden.
For a similar argument, see X (1978).
In a student note (X, 1978) the thesis that preagreed damage clauses
should be reasonable in the light of actual harm is defended too. This test
enables full compensation of the non-breaching party to occur since the
preagreed damage clause will be enforced as long as the breaching party is
unable to show that actual damages are less than the sum of objective and
subjective damages. In addition, the ex post test is consistent with the
discouragement of wagering and breach-inducing activities.
Theoretically, the argument is correct. Yet it applies to all other
supracompensatory sanctions as well. As argued in Section 2, there is a
similar problem when specific performance is made operational through the
use of ‘astreintes’. The major question is in how many cases the promisee
has an opportunity to covertly induce breach. It is not so easy to make a
debtor fail when he is aware of the dramatic consequences of his breach.
However a technique (not considered by Clarkson, Miller and Muris, 1978)
that has more chances to succeed is to let a promisor sign a contract that he
did not read.
11. Do Penalty Clauses Lead to More Trials?
Landes and Posner (1979) mentioned that historically the penalty doctrine
was created at a time when judges were paid from litigation fees. As argued
by Rea (1984a) this argument does not explain why the penalty doctrine
persisted in an era of subsidized courts.
Rubin (1981) built on the theory of self-enforcing contracts introduced by
Telser (1980) and argued that penalty clauses are not self-enforcing; they
can be enforced only by court intervention. Every breach of contract by the
debtor will therefore lead to a trial. The costs of these trials are not fully
borne by the parties themselves. Courts are largely subsidized. Because some
costs are externalized, parties will stipulate penalty clauses more often than
is optimal. Rea (1984a) remarked, however, that this argument could be
used in many more cases. There are numerous areas in which courts seem to
encourage the use of judicial resources, even though they are supplied at
little cost. Furthermore, the additional uncertainty associated with the
unenforceability of penalty clauses might increase litigation too.
The reasoning of Rubin (1981) is implicitly contradicted by the modern
economic literature on settlement versus trial (see Chapters 7000 and 7400).
In principle rational parties do not go to court. They prefer a settlement
rather than an expensive trial. Why would a rational debtor refuse to pay the
damages stipulated in the contract spontaneously if these clauses are valid in
that legal system? It would cost him a lot more if the promisee went to court.
A trial between two rational parties is only possible if they have different
perceptions of their chances to win. This will happen more often if it is hard
to predict the decision of the judge. With a legal rule allowing all penalty
clauses, the probability of different perceptions will be reduced to nearly
zero. The judge will certainly declare the penalty clause valid. Discussions,
on the other hand, may arise if penalty clauses are forbidden while
liquidated damage clauses are allowed, or when such clauses are forbidden if
they exceed the highest damage possible. In those cases the judge has to
draw a line and where that line lies is not always clear. Whether this is a
valid argument against regulations prohibiting (certain) penalty clauses, is
doubtful, however. The problem of the excessive subsidization of courts can
be solved more easily by increasing the trial costs.
12. Penalty Clauses as a Barrier to Entry?
In the literature it has been argued that penalty clauses entail another serious
potential danger: their use as a barrier to entry.
Diamond and Maskin (1979) are the first to clearly demonstrate the
external effects of damages. In their model, individuals search for partners
and negotiate contracts to produce output. The individuals may decide to
breach their contract and form a new partnership. Since the surplus of this
new partnership depends on the damages that a breaching party has to pay to
its old partner, they also affect its new partner. By stipulating liquidated
damages parties to a contract can therefore exert some monopoly power over
By illustrating how liquidated damage clauses can reduce competition
Aghion and Bolton (1987) again stress the external effects of such clauses.
In their model an incumbent seller and a buyer can sign a contract
specifying a price and liquidated damages in case the buyer breaches the
contract. In line with Diamond and Maskin (1979) it is shown that this
contract gives the parties some joint monopoly power over a potential
competitor. The buyer will not breach the contract and trade with a new
entrant unless the latter compensates him for the liquidated damages that he
owes to the incumbent seller. The liquidated damages thus act as an entry
fee which enables the contracting parties to appropriate part of the surplus of
a more efficient producer who enters the market. Because of this entry fee a
welfare-enhancing entry is blocked and the contract introduces a social cost.
Chung (1992) analyzes the effects of the penalty doctrine in a framework
similar to that of Aghion and Bolton (1987). But in his model the buyer also
has to make a reliance decision and the third party is a new buyer with a
higher valuation for the good. He shows again that by stipulating in the
initial contract a high level of liquidated damages, the parties to a contract
could raise the price that a third party has to pay to induce a breach of
contract. However, when the reliance investment of the buyer is fixed, the
penalty doctrine, by putting an upper limit on the enforceable stipulated
damages, eliminates this inefficiency and allows the implementation of the
first-best outcome. This no longer holds when the reliance investment of the
buyer is variable. Then the promisee will overrely on it, even with the
penalty doctrine in place.
In the literature, the Aghion and Bolton model is often used as an
argument against penalty clauses. However, this barrier-to-entry-argument
applies to the expectation measure and to liquidated damages as well.
Consider a contract for the duration of 10 years that binds all consumers
in a market to a monopolistic seller. The production costs of the current
monopolist are 50 and the price is 100. The sanction for breach is the
expectation measure (the standard remedy under both Anglo-American and
continental law), which means that a breaching buyer has to pay 50. A new
entrant with a production cost of 50 offers the same product at a price of 51.
None of the buyers will breach. If the product is sold at production costs,
that is 50, consumers are indifferent. But if we assume that changing
partners always involves positive transaction costs (a very plausible
assumption), the new entrant will still not attract any of the existing
consumers. So the expectation measure, which is the normal sanction for
breach in Anglo-American law, as well as in continental law, can serve as a
barrier to entry as well.
Aghion and Bolton (1987) merely view their paper as an illustration of
the much broader problem of the endogenous creation of switching costs.
Other examples include advance deposits in rental contracts, frequent flyer
programs, trading stamps, deferred rebates by shipping firms, fixed fees in
franchise contracts and so on. Even if the penalty doctrine is needed to
counter the use of damage clauses as a barrier to entry, it will not necessarily
capture other contract clauses which may have the same effect. Given the
variety and complexity of potential contract clauses, antitrust authorities face
an almost impossible task here.
13. Can Penalty Clauses Have a Signalling Function?
The idea that penalty clauses can be useful as a signal for a promisor’s
reliability was first articulated by Posner (1977, p. 93). According to Posner,
this signalling function is important especially for new entrants in the
market who have not yet built up a reputation. An effective way for a
promisor to convince other parties that he will perform as promised is to
offer a penalty clause against himself. The fact that the promisor is willing
to offer such a heavy sanction on his non-performance signals that he is
convinced that he is willing and able to perform. Thus a penalty clause has a
communicative function: it is a signal of reliability sent by the debtor.
Although this idea has been seriously criticized in the literature, Posner
retains this argument in the third and the fourth editions of his handbook
(Posner, 1986, 1992).
Apparently, this argument was countered in a convincing way by Muris
(1981) and Rea (1984a). According to those authors a liquidated damage
clause can perform this function equally well. A liquidated damage clause
fully compensates the promisee, making him indifferent as to whether the
promisor will breach or perform. So if a new entrant promises to fully
compensate the losses in case of non-performance, potential parties are
satisfied too. But compared to penalty clauses, liquidated damages entail
fewer incentive problems. This may be reflected in a lower price of the
product or service. A similar argument was developed by Schwartz (1990).
De Geest (1994), however, has argued that under some conditions, the
signalling theory becomes defendable. It is cheaper to draft a penalty clause
than a liquidated damage clause. In the case of a liquidated damage clause
one has to form an idea of all possible damage cases that might occur and to
compute an average of them. For a penalty clause it suffices to estimate the
highest possible damage.
A penalty clause may therefore be desirable when a precise liquidated
damage clause is too costly and the courts systematically award a
compensation which is too small or which is exact on average but uncertain.
A debtor may, for instance, want to use a standard form contract with a
penalty clause to his own detriment. It may be too costly to draft a liquidated
damages clause for every person individually. Therefore a standard form
contract is used with a penalty clause compensating for the highest loss
possible. For most creditors the compensation exceeds the real damage and
even the ex ante estimation of it. Therefore, one clearly has to do with a
If the loss of the average creditor is taken as a starting point, an adverse
selection problem would therefore arise in such cases. The creditors with the
highest real damage would not contract. On the other hand no high
premium has to be demanded in return for the penalty clause when the only
risk consists of non-performance by the debtor. For a bona fide debtor this
risk is almost equal to zero, leaving aside the legal error costs.
However, this signalling argument can never justify a penalty clause that
applies to force majeure cases (De Geest, 1994). Such a penalty clause that
applies in case of force majeure is an instance of overinsurance, a gambling
clause. The suggestion of Posner (1973) that courts are taking a stand
against gambling contracts makes sense in that respect.
14. Penalty Clauses as Punitive Sanctions when the Apprehension Rate
is Lower than 1
In the economic literature on tort law it is generally accepted that punitive
damages may be economically desirable when the injurer has a significant
chance of escaping liability for the harm he caused (Polinsky and Shavell,
While apprehension rates are generally higher in contract than in tort
cases, there may be contract cases where it is difficult to discover or prove
the debtor’s breach. It has been argued that punitive damages should be
awarded in those cases (Perlstein, 1992; Polinsky and Shavell, 1998, pp.
936-939). See also Farber (1980) and Schwartz (1990) for a discussion of
the economic function of punitive damages in breach of contract disputes.
Penalty clauses may remedy this problem as well. Their extracompensatory
nature compensates apprehension rates lower than 1. Which
of both remedies is the best - punitive damages or penalty clauses - is still a
relatively unexplored issue. Punitive damages have one comparative
advantage: the apprehension rate can be calculated more realistically,
because it is determined ex post, when more information about what
happened is available. But this ex ante lack-of-information argument is a
more general one that can be used against liquidated damages as well (ex
ante it is more difficult to predict the magnitude of the harm than ex post). It
may therefore be insufficient in itself to justify the penalty doctrine.
15. Penalty Clauses as a Technique to Obtain Backward-Looking
In Section 2 we have already discussed an important difference between
penalty clauses and the specific performance remedy. Specific performance
only works for the future and not for the past. If performance in natura is no
longer desirable, courts will award just damages. On the other hand, penalty
clauses can be awarded for breaches in the past. This means that the
traditional specific performance remedy may sometimes not be enough of a
disadvantage to make the promisor perform specifically. Penalty clauses may
therefore be the only technique to really guarantee specific performance.
16. Penalty Clauses as a Technique to Create Risk Sharing
Polinsky (1983) demonstrated that stipulated damages, which differ from
expectation damages, can be necessary to obtain an optimal risk allocation.
Under some conditions, penalty clauses can be a technique to let parties
share the risk of a higher third party bid.
Consider a contract to sell a Van Gogh painting. The price is 100 but the
buyer’s true valuation is 150. There is a chance that before the date of
delivery a third party will arrive offering 200. Suppose that the parties want
to share that risk equally. This result can be obtained by stipulating in the
contract that the seller should pay damages amounting to 75 in case of a
breach. Should there be a third party offer, then the seller will indeed breach
and make an additional profit of 25 (a price increase of 100 minus 75
damages). The buyer will be 25 better off as well. Yet the damages of 75 are
overcompensatory (at least if the buyer had no sufficient knowledge of the
market to have found the third party himself).
A penalty clause exceeding 100 (in the latter example) allocates the risk
entirely to the buyer. At first sight, penalty clauses are not necessary here,
however, since the specific performance remedy obtains the same result. Yet,
rational parties may prefer penalty clauses in those cases for the reasons
discussed in the former section (specific performance cannot always restore
breaches in the past). Of course, penalty clauses can be a good second-best
solution when specific performance could clear the job as well but is not
available in the legal system.
17. Do Penalty Clauses Increase the Number of Bankruptcies? Penalty
Clauses to Sanction Debtors Financing their Commercial Activities
via Delayed Payments
According to Farber (1983) penalty clauses lead to more bankruptcies. A
debtor will have to pay a much higher amount in case he breaches and will
therefore more easily get into financial distress. If the bankrupted party is
involved in a web of long-term business relations with others, these too will
be disrupted. If the other contracting parties fail to include penalty clauses in
their own contracts, they will be disadvantaged when it comes to receiving a
share of the assets in bankruptcy. If the use of penalty clauses were
widespread, a wave of business failures triggered in a recession by penalty
clauses could have a severe effect on investor confidence.
But Farber (1983) does not take into account another cause of insolvency,
as argued by De Geest (1994). Entrepreneurs who are working with other
people’s capital sometimes take on excessive commercial risks. This danger
is particularly serious when they finance their activities with fixed interest
loans. Then the downward risk of the entrepreneurs is externalized while the
upward risk is internalized. This is why shareholdership is a more common
way of raising money. Yet there is a subtle way to coerce non-assenting
people to lend money at a fixed interest rate: let them work for you or deliver
goods and simply delay paying them. The creditor can then go to court, but
it may take a long time before he is able to enforce his rights. This problem
is most pressing in case of corporations that are about to go bankrupt.
A penalty clause will not really prevent the entrepreneurs from taking on
commercial risks. They can always turn to any owner of capital they may
find. The latter will only lend out his money if he believes that he will
benefit from it. In case of intrinsically loss-making projects this is almost
impossible. A penalty clause in relation to an obligation to pay a sum of
money just makes it more costly to force creditors to lend capital against
their will. Penalty clauses will direct capital seeking entrepreneurs from
non-specialized borrowers (painters, sellers, builders, ...) to specialized
borrowers (bank, shareholders, ...). The result is economically desirable and
may lead to a lower number of bankruptcies, or at least to lower drains in
case of bankruptcy.
By the same token, a penalty clause may be efficient in the case of loans.
A penalty clause that corresponds to the highest possible risk premium
possible will create proper incentives for a debtor to borrow additional
money from parties that explicitly give their consent.
C. Underliquidated Damages
18. Unliquidated Damages as a Technique to Create Risk Sharing
Probably, the most important function of underliquidated damages is to
make risk sharing possible for production cost uncertainties.
In case of underliquidated damages against a producer, the risk of
increased production costs is shared between the debtor and the creditor. The
debtor only bears this risk to the extent of his own profit and a fraction of the
consumer surplus of the creditor. The creditor bears the risk to the extent of
the remaining part of his consumer surplus.
Underliquidated damages against a consumer are possible as well. They
imply that a consumer who cancels a reservation or an order, does not have
to compensate the total profit of the producer or the seller. The risk of
decreased utility is shared between the creditor and the debtor. The creditor
bears this risk to the extent of his own consumer surplus and only a fraction
of the profit of the debtor. The debtor bears the risk to the extent of the
remaining part of his profit. Underliquidated damages thus allow agreement
on, for instance, the reliance measure, the default rule prescribing the
19. Other Advantages and Disadvantages of Underliquidated Damages
Underliquidated damages create a number of inefficient incentives. The
creditor gets an incentive to avoid by all means an efficient breach of
contract, because his compensation given a breach of contract is lower than
his consumer surplus with specific performance. The debtor gets an
incentive to head for an inefficient breach of contract in case of increased
performance costs of the seller or decreased utility for the buyer.
Underliquidated damages for the risk of a higher third party offer boil down
to a gambling clause. With respect to the allocation of that risk, this is
always undesirable unless the creditor is risk seeking.
But underliquidated damages resemble the reliance measure. Also, with
the latter, the compensation is lower than the real (expectation) damage.
Compared with expectation damages, reliance damages have one big
advantage: they lead to an optimal level of contracting.
Because rational parties may have good reasons to stipulate
underliquidated damages, it is not desirable to readjust this clause. That
would be a type of regulation which does not necessarily improve the
As shown by Stole (1992) underliquidated damages may also play a role
as screening device. Standardized underliquidated damages can give highvalue
buyers an incentive to reveal their true valuations. We could compare
this with the foreseeabilility doctrine in ordinary contract law: if parties do
no use stipulated damages, courts will undercompensate the promisor if his
losses were unforeseeable to the promisee (Hadley v. Baxendale in the
common law). The economic function of that rule is to make promisees
reveal in advance that they are more vulnerable.
20. Liquidated and Underliquidated Damages should be Allowed
In the literature there seems to be a consensus that liquidated and
underliquidated damages should be respected. This means that judges
should not test liquidated damages against the real ex post damage. If judges
always corrected liquidated damages that turn out to differ from the real
losses, this would mean that the ex post valuation of the judge is the only
thing that ultimately counts (except for those cases where judges are
uncertain as to what are the true losses and put the burden of proof with the
party that argues that the true losses differ from what is stipulated ex ante).
Liquidated damages can be a rational option, especially if parties have
more information about the possible losses than judges. Underliquidated
damages may among other things be useful to let parties share the risk of
increased production costs.
21. Under what Conditions should Penalty Clauses be Allowed?
Penalty clauses, on the other hand, have been the subject of a fierce
controversy for a long time. Most of authors seems to defend the penalty
doctrine - the common law doctrine that forbids penalty clauses. It cannot be
denied that penalty clauses have many disadvantages. Yet they may have a
number of functions as well.
Though penalty clauses should be forbidden in most cases, we
nevertheless believe that they should be allowed under a few well-defined
(a) A penalty clause against the drafter of a standard term contract
should be allowed when it reflects the highest possible losses. Consumers
may indeed have varying losses in case of breach. Drafting a tailor-made
liquidated damages clause for each consumer individually may simply be too
costly. The signing-without-reading problem is extremely unlikely to occur
if the penalty clause is conceived against the drafter of the contract.
(b) A penalty clause for the obligation to pay an amount of money in
time should be allowed. Here it is quite rational to set the interest rate so that
it reflects the highest possible commercial risk. An interest rate of 200
percent a year may discourage bad-faith promisors to finance risky,
commercial projects by retarding payment to non-consenting promisees. Of
course there is still a signing-without-reading danger here. But this may be
limited if the interest rate is not set at a nearly infinite level but at one that
just reflects a very high commercial risk.
Yet in case of force majeure, these penalty clauses should not be
enforceable even if conditions (a) and (b) are fulfilled. It is impossible for a
penalty clause to have a signalling function here. This would be an instance
of overinsurance, a gambling clause.