For over a century England.s judicial system decided land disputes by ordering
disputants.legal representatives to bludgeon one another before an arena of spectating
citizens. The victor won the property right for his principal. The vanquished lost his
cause and, if he were unlucky, his life. People called these combats trials by battle. This
paper investigates the law and economics of trial by battle. In a feudal world where high
transaction costs confounded the Coase theorem, I argue that trial by battle allocated
disputed property rights e¢ ciently. It did this by allocating contested property to
the higher bidder in an all-pay auction. Trial by battle.s .auctions. permitted rent
seeking. But they encouraged less rent seeking than the obvious alternative: a .rst-
price ascending-bid auction.
_I thank Gary Becker, Omri Ben-Shahar, Peter Boettke, Chris Coyne, Ariella Elema, Lee Fennell, Tom
Ginsburg, Mark Koyama, William Landes, Anup Malani, Jonathan Masur, Eric Posner, George Souri,
participants in the University of Chicago and Northwestern University.s Judicial Behavior Workshop, the
editors, two anonymous reviewers, and especially Richard Posner and Jesse Shapiro for helpful suggestions
and conversation. I also thank the Becker Center on Chicago Price Theory at the University of Chicago,
where I conducted this research, and the Mercatus Center at George Mason University.
yEmail: PLeeson@GMU.edu. Address: George Mason University, Department of Economics, MS 3G4,
Fairfax, VA 22030.
.When man is emerging from barbarism, the struggle between the rising powers
of reason and the waning forces of credulity, prejudice, and custom, is full of
. Henry C. Lea, Superstition and Force (1866: 73).
Modern legal battles are antagonistic and acrimonious. But they aren.t literally battles.
Disputants don.t resolve con.icts with quartersta¤s. Their lawyers don.t .ght to the death.
This wasn.t always so. For over a century England.s judicial system decided land disputes
by ordering disputants. legal representatives to bludgeon one another before an arena of
spectating citizens. The victor won the property right for his principal. The vanquished lost
his cause and, if he were unlucky, his life. People called these combats trials by battle.1
To modern observers trial by battle is an icon of medieval backwardness. Montesquieu
called it .monstrous.(1748 : 563). The institution.s barbarity seems equalled only by
its senselessness. As Richard Posner put it, .trial by battle.is one of those .legal practices
that no one defends any more.(1988: 858).
Almost no one. This paper defends trial by battle. It examines trial by battle in England
as judges used it to decide property disputes from the Norman Conquest to 1179.2 I argue
that judicial combat was sensible and e¤ective. In a feudal world where high transaction costs
confounded the Coase theorem, trial by battle allocated disputed property rights e¢ ciently.
Trials by battle were literal .ghts for property rights. I model these trials as all-pay
auctions. Disputants .bid.for contested property by hiring champions who fought on their
behalf. Better champions were more expensive and more likely to defeat their adversaries in
combat. Since willingness to pay for champions was correlated with how much disputants
1As indicated below, in medieval documents trial by battle is commonly called .duellum..Subsequent
commentators on this institution called it .trial by battle. or .judicial combat,.both of which terms this
2 I don.t consider trial by battle as England.s criminal justice system used it. Nor do I consider trial
by battle in England.s courts of chivalry where judges used it to decide cases involving a¤ronts to honor,
treason, and criminal acts committed abroad. See Russell (1980b) on judicial combat in criminal appeals.
See Russell (2008) on judicial combat in courts of chivalry. For classic treatments of the variety of single
combats and their history in England and elsewhere, see Selden (1610), Gibson (1848), and Nielson (1891).
For examples of judicial duels outside of England in cases unrelated to land disputes, see Howland (1901).
valued contested land, trial by combat tended to allocate such land to the higher-valuing
This .auction. permitted rent seeking. But it encouraged less rent seeking than the
obvious alternative: a .rst-price ascending-bid auction. Further, unlike these auctions, trial
by battle converted part of its social cost into social bene.t: judicial combats entertained
My analysis explains how a seemingly irrational legal institution. trial by battle. is
consistent with rational, maximizing behavior. It illuminates why this apparently ine¢ cient
institution played a central role in England.s legal system for so long. Most important, it
demonstrates how societies can use legal arrangements to substitute for the Coase theorem
where high transaction costs preclude trade.
Economists have said nothing about trial by battle.3 Schwartz, Baxter, and Ryan (1984),
Posner (1996), and Kingston and Wright (2009) discuss duels of honor.4 These are distinct
from and, except for the fact that they involve two combatants, unrelated to judicial duels,
which I consider. Duels of honor were private, unsanctioned, and often legally prohibited
battles waged to redress insults or transgressions of honori.c norms. They weren.t trials
used to decide property rights in legal disputes. Trial by battle is also distinct from and
unrelated to battles between enemy groups fought by a single representative from each side.
The former was a judicial procedure for allocating disputed land. The latter was a diplomatic
procedure for reducing war.s cost.
This paper is most closely connected to two strands of literature. The .rst uses rational
choice theory to understand unusual legal institutions. Friedman (1979) was among the
.rst contributors to this literature. He considers the economics of legal institutions that
stateless people in medieval Iceland used to create social order. Posner (1980) explores the
economics of legal systems in primitive societies. Leeson (2007a, 2009a, 2009b) examines the
economics of 18th-century pirates.legal institutions. He also considers the legal arrangements
3There are a few exceptions to this. Clark (2007) brie.y acknowledges trial by battle. He restates the
conventional wisdom that it was ine¢ cient. Tullock (1980a) mentions trial by battle in passing. He uses it
to mock the modern adversary system of dispute resolution. Zywicki (2008: 44) endorses Tullock.s view that
trial by battle was pure waste (though not Tullock.s analogous criticism of the modern adversary system):
.The trial by battle, of course, is a classic rent-seeking interaction, as there is no social surplus generated by
resolving disputes in that manner..
4Volckart (2004) considers feuding in late medieval Germany.
that warring hostiles created along the 16th-century Anglo-Scottish border (Leeson 2009c).5
Most recently, Leeson (2010) analyzes the law and economics of medieval judicial ordeals.
The second strand of related literature explores the economics of European legal tradi-
tions. Hayek (1960), La Porta et al. (1998), Glaeser and Shleifer (2002), and Djankov et
al. (2003) consider how legal institutions diverged in England and continental Europe in the
Middle Ages and how these institutions in.uenced property rights in those places. I consid-
ers how a key legal institution found throughout Europe before that divergence in.uenced
property rights in England: trial by battle.
The Norman Conquest introduced trial by battle (duellum) to England.6 Until 1179 it was
England.s primary trial procedure for deciding land ownership disputes.7 During this period
one person challenged another person.s claim to a piece of land by initiating an action called
a .writ of right..8 The plainti¤in such an action was called the .demandant..The defendant
was called the .tenant..The demandant initiated his challenge by requesting the crown to
issue an order compelling the tenant to appear before a court to defend his property.9
A colorable claim was necessary for the demandant.s challenge to make it to trial. There
was no guarantee the crown would ful.ll the demandant.s request for a writ. It therefore
behooved him to supply some evidence of his claim.s plausibility. For instance, a land
charter documenting the demandant.s connection to the contested property might be helpful.
Similarly, the court could reject the demandant.s claim if he failed to produce a reliable
witness who would swear to his connection to the disputed land. The court required the
tenant to produce a witness for this purpose too. These .screens.helped prevent some bogus
5Leeson (2007b, 2008) examines private legal arrangements that precolonial Africans used to support
cooperation without government.
6Duellum was also called bellum, for instance in Domesday Book.
7In a small minority of cases, unilateral ordeal, witness investigation, testimony of a hundred court, and
jury-like arrangements were used to resolve property disputes (see, for instance, van Caenegem (1990: 82,
50-51). The law limited judicial combat to cases involving land worth at least 50p., disallowed combat
between disputants of widely di¤erent status, and exempted some towns from battle (Russell 1980a).
8Sometime in the mid-12th century no one could initiate a real property dispute without seeking and
receiving such a writ from the king (Watkin 1979).
9The demandant purchased the writ, which also ordered a particular court to hear the case. Judges heard
land cases in seignorial courts, county courts, and royal courts.
property challenges from making it to trial. But they did so very imperfectly.
Ideally the legal system would.ve liked to assign disputed property rights to those rights.
.true.owners. Unfortunately for 11th- and 12th-century judges, evidence that could help
them do this was in short supply. Judges frequently faced a situation in which the evidence
at their disposal amounted to disputants.and their witnesses.(and perhaps their charters.)
competing, plausible claims. Without helpful evidence, in many cases judges couldn.t iden-
tify disputed land.s true owner. In these cases judges pretended to divine that owner.s
identity instead. Their method of doing so was judicial combat. Trial by battle.s ostensible
justi.cation was as simple as it was absurd: God favored the rightful disputant.s cause. So
he would favor that disputant.s cause in a physical .ght.
Despite its supposedly superstitious underpinnings, trial by battle had secular origins
(Russell 1980a: 112). Further, unlike unilateral ordeals, superstition wasn.t important to
trial by battle.s operation or ability to produce socially desirable results (Bartlett 1986;
Leeson 2010). As I describe below, judicial combat.s productivity rested .rmly in earthly
Trial by battle.s basic form in property cases in the 11th and 12th centuries remained
similar in the 13th century. Our detailed descriptions of some of this form.s aspects are from
still later trials. However, their general features are applicable to trial by battle.s heyday.
The demandant pled before the court by o¤ering to prove his right to the disputed land
on his champion.s body. Consider the demandant.s plea in a case from 1198 (Russell 1959:
Matthew, the son of William, sought against Ralph of Wicherle and Beatrice,
his wife, a wood and other land at Ellenthorpe as the right and dowry of his wife,
Emma, whereof the said Matthew was seised as of right and dowry in the time
of King Henry by taking the issues thereof from wood, timber and pasturing pigs
to the value of 5/4d; and this he o¤ered to prove against him by his freeman
Utling, who o¤ered to prove this against him as the court should adjudge as of
his sight, or by another if any ill should befall him.
The tenant pled by denying the demandant.s claim and o¤ering his own champion as
Ralph and Beatrice came and denied the right and seisin of the said Matthew
by a certain freeman of theirs, Hugh of Floketon, who o¤ered to deny this by his
body, or by another.
If the court couldn.t establish the rightful disputant.s identity, it
adjudged that there should be a battle between [their champions]. The pledges
of Hugh (defending) were Ralph his lord and Robert, the son of Payn. The pledges
of Utling were Matthew, the son of William, and Robert of Cove. A day was
given to them on the coming of the justices into those parts.
In theory the law required the demandant.s champion to be a witness to his right to the
disputed land. The champion had to claim that he observed the demandant.s ancestor.s
seisin. Alternatively he could claim that his deceased father observed it and instructed him
to defend the demandant.s right.
In practice the law permitted demandants to hire champions. A tenant could object to
the demandant.s champion on the grounds that he was hired. But .professional champions
were so frequently used that the courts paid no attention to this particular objection..So
tenants didn.t bother. .There appears to be no recorded case relating to land where one of
the parties objected to the other.s champion solely on the ground that he was hired for the
occasion.(Russell 1959: 257).10 In 1275 judges dropped the charade. The law abandoned
the requirement that demandant champions be witnesses.
The law never even theoretically restricted who tenants could use as champions. Unlike
demandants, tenants could also choose to .ght in person. Though they almost never did.
Later law eliminated this choice. It required tenants to use champions too.
After the disputants pled, the judge asked the champions if they were prepared to wager
battle. To show they were the champions passed him a glove with 1d. in each of its .ngers.
The judge then gave a day when the champions would .ght. Two men from each disputant.s
side pledged to attend.
On the appointed day the champions came to the designated arena and swore oaths a¢ rm-
ing their principal.s rightness in the cause. They also promised they hadn.t concealed charms
on their bodies or resorted to sorcery. Eleventh- and 12th-century arenas were makeshift.
10Russell (1959: 243) suggests that the earlier in the period one goes, the more likely it was probably the
case that demandant champions were genuine witnesses.
Later ones were more elaborate and specially constructed for the purpose. Sixteenth-century
records describe the .lists.as (Russell 1983a: 126):
an even and level piece of ground, set out square, 60 feet on each side due
east, west, north and south, and a place or seat for the justices of the bench was
made without and above the lists, and covered with furniture of the same bench
in Westminster Hall, and a bar made there for serjeants-at-law.
Before battle began the presiding justices made an announcement forbidding spectator
interference. The justices. injunction before a 17th-century combat conjures images of a
deadly tennis match (Russell 1983: 126):
The justices command, in the Queen Majesty.s name, that no person of what
estate, degree, or condition that he be, being present, be so hardy to give any
token or sign, by countenance, speech, or language, either to the prover of the
defender, whereby the one of them may take advantage of the other; and no
person remove, but will keep his place; and that every person or persons keep
their staves and their weapons to themselves; and su¤er neither the said prover
nor defender to take any of their weapons or any other thing, that may stand
either to the said prover or defender any avail, upon pain of forfeiture of lands,
tenements, goods, chattels, and imprisonment of their bodies, and making .ne
and ransom at the Queen.s pleasure.11
The demandant.s champion could win trial by battle in two ways: killing his adversary
or forcing him to submit. A champion submitted to his opponent by uttering .craven..The
tenant.s champion could win in a third way: pushing a stalemate until nightfall. Battle began
before noon. Justices adjudged the tenant.s champion victorious if he remained standing
when the stars appeared.
The victorious champion won the contested property right for his principal. The presiding
judges concluded the trial by ordering the disputed land to his principal.s possession and
announcing his principal.s good title publicly (Russell 1983a: 127):
The King to the sheri¤, greetings. I command you that, without delay, you
give possession to X of [description of land], concerning which there was a suit
between him and Y in my court; because such land is adjudged to him in my
court by battle.12
11This injunction was made at a judicial combat trying a criminal case.
12This announcement concluded a 13th-century trial by battle.
Champions.post-trial fate depended. If both survived, the winner enjoyed the glory of
victory and an improvement in his reputation as a hired thug. The loser was less fortunate.
He paid a £ 3 .ne for perjury and .lost his law:. the judges declared him infamous. He
could never again bear witness in another.s legal dispute (Russell 1980a: 116, 123; Lea 1866:
3 A Theory of Trial by Battle
3.1 Sticky Property Rights
When inadequate evidence prevented medieval judges from assigning disputed property
rights to their true owners, they attempted to do the next-best thing they could do: al-
locate disputed property rights to their higher-valuing users. If transaction costs are zero,
legal systems can rely on private bargaining to allocate disputed property rights e¢ ciently
(Coase 1960). Since transaction costs aren.t zero, how judges allocate disputed property
rights matters. How much it matters varies in proportion to transaction costs.height. If
transaction costs are low, it.s relatively unimportant who judges assign disputed property
rights to: transaction costs typically permit exchange to move rights to persons with more
valuable uses for them.14 If transaction costs are high, it.s very important who judges assign
disputed property rights to: transaction costs typically preclude Coasean exchange.
High transaction costs make property rights .sticky..They prevent markets from reshuf-
.ing rights to higher-valuing users. When rights are sticky, if judges get initial allocations
.wrong,.disputed property rights get stuck in lower-valuing users.hands. Thus the higher
transaction costs are, the more concern a legal system interested in e¢ ciency will show for
getting initial allocations .right..
Land rights in Norman England were near the extreme end of the transaction-cost-of-
13Since the demandant.s champion had to be a witness, in theory a champion who lost a battle, and thus
lost his law, might be prevented from working again as a champion for a demandant (though, it would seem,
not a tenant). In view of courts. unwillingness to uphold the demandant witness rule in the .rst place,
however, it.s questionable whether they would.ve, or in some cases could.ve, enforced this rule.
14Even if transaction costs are zero, there.s still some bene.t of judges assigning disputed property to
higher-valuing users. If disputants know the legal system allocates contested rights this way, they have an
ex ante incentive to use their property in the way that maximizes its social value.
trade spectrum. They were sticky. Anglo-Norman legal institutions therefore showed great
concern for assigning disputed property rights to the higher-valuing user.15 Trial by battle
was that concern.s result.
The feudal system made Anglo-Norman land rights sticky.16 That system created a chain
of lord-tenant relationships extending downward to the lowliest tenant who held his tenement
of some lord, but of whom no lowlier tenant held of him, and extending upward to a baron
or great lord, a tenant-in-chief who held of the king.
The chain of land holders that constituted feudal property arrangements created third
parties with direct interests in tenants.land-related decisions. Those decisions threatened
to impose large externalities on them. Among the most important such decisions were those
relating to land.s alienation.
Alienation had two forms: substitution and subinfeudation. Substitution replaced a link
in the feudal chain. Subinfeudation created a new link it.17
A tenant who substituted his land sold his spot in the feudal chain to someone else.
That buyer purchased the land rights the tenant previously enjoyed. the lord.s protection,
the ability to support himself by the land, and so on. He held of the tenant.s former lord.
The buyer also purchased the obligations of performing the services of holding that land the
tenant previously had. knight.s fees (or service), work, produce, and the duty to pay other
feudal incidents, such as .aids.and .relief..A tenant who subinfeudated his land sold some
portion of his tenement to a buyer but remained a tenant of his lord. This made him the
buyer.s lord and the buyer his lord.s sub-tenant.
The third parties with the strongest interest in land alienations were the alienor.s heirs.
the would-be successors of his holding. and his immediate lord. Subinfeudation threatened
these individuals.interests in alienated property. A tenant might subinfeudate his land for
15O¢ cially, the Anglo-Norman period closed with the end of Stephen I.s reign in 1154. The Angevin period
followed it. Thus trial by battle in the years I.m concerned with (1066-1179) overlapped both periods. Despite
this overlap and the resulting technical inaccuracy, for want of a better term, when I refer to Norman England
or Anglo-Norman legal institutions, I.m referring to England during the period 1066-1179.
16Technically it.s incorrect to speak of land ownership in the context of feudal relations. One should speak
of land tenure and holding or seisin. Tenant ownership doesn.t emerge until the late 12th and early 13th cen-
turies. I discuss this development below. References in my discussion to land ownership, buying/selling land,
and so on should be understand to refer to land tenure/holding and the buying/selling of tenures/holdings.
17For a good summary of substitution and subinfeudation and the problems alienation created, see Baker
an up-front payment and small service from the buyer. When he died, all his heir was entitled
to was the small service his buyer owed.
Further, that service might be the performance of some duty the subinfeudator owed as a
service to his lord. The subinfeudator.s concern was the buyer.s ability to make the up-front
payment rather than his ability to perform the service. However, since the buyer.s failure to
perform for the tenant could a¤ect the tenant.s ability to meet the service he owed his lord,
subinfeudation could injure the lord.s interest.
Subinfeudation could also injure the lord.s interest by precluding his claim to escheat.
If a tenant died and no heir was forthcoming, or if the tenant committed a felony, or failed
to appear in his lord.s court, his property fell to his lord. By inserting a tenant below him
through subinfeudation, the subinfeudator could enjoy this right instead.
Substitution posed similar problems. If a tenant substituted his holding, his heirs.interest
in that land was extinguished. Land he sold was land his heir couldn.t inherit. If a tenant
sold his holding to a less reliable or competent person, his lord su¤ered. The lord became
less likely to receive the service owed him attached to that holding. An old tenant who sold
his property to a young person also damaged his lord who would now have to wait longer to
If a tenant granted his property to a religious house, the injury his lord su¤ered was
still greater. Such grants relieved the new holder, such as a church or monastery, of the
obligation to render the services the former tenant owed his lord. Churches and monasteries
usually held land in alms. The only services they were obligated to provide were spiritual
ones, typically prayers for the granting tenant and perhaps his lord.
To prevent alienors from injuring their heirs and lords, norms developed in Norman
England, bolstered in some areas by formal law, requiring or making it very desirable for
tenants to get their heirs.and lords.consent to alienate land.18 These norms were .exible.
For instance, if the lord.s, tenant.s, and heir.s interests were clearly aligned, receiving explicit
consent to alienate was usually unnecessary. In contrast, if a tenant sought to grant his land
18Feudal property arrangements created another externality problem relating to land alienation: a lord.s
decision to alienate his property could injure his tenant who the alienation would place under a new lord
(.attornment.). Some restrictions also developed to regulate this problem. For instance, a tenant couldn.t
be forced to do homage to a new lord who was his enemy.
to the Church, consent was mandatory: the lord exercised veto power over the tenant.s desire
Feudal property arrangements created a host of externality problems. Thus they required
rules of consent governing land alienations. But these rules had an unfortunate side e¤ect:
they dramatically increased the transaction costs of trading land, sti.ing its reallocation.
.[M]ultiple consents required from people with diverse standards and concerns retarded the
use of land as an economic asset.(Palmer 1985: 387). They made Anglo-Norman property
rights in land sticky.
3.2 Violent Auctions
When property rights are sticky, if judges can.t identify disputed property rights.true owner,
it.s important for them to allocate those rights to the higher-valuing disputant. But judges
have a problem: they don.t know which disputant values the disputed rights more. Trial by
battle was Norman England.s solution to this problem. It was a medieval demand-revelation
mechanism that identi.ed the higher-valuing disputant and allocated disputed land rights
The Anglo-Norman legal system used trial by battle to hold .violent auctions.for con-
tested land. In these .auctions.legal disputants .bid.on contested land by spending on
champions who literally fought for property rights on their employers.behalf. Better cham-
pions were more likely to win these combats.
The best developed reputations for their skill in the arena. Thirteenth-century champion
William of Copeland.s name preceded him. It was known far and wide, from Yorkshire to
Somerset. .The mere sight of him was enough to scare any tenant who might have considered
countering his challenge..Copeland.s contemporary, Robert of Clopton, .was [also] in great
demand as a champion.in the early 13th century (Russell 1959: 259, 246).
Because they were in greater demand, better champions commanded higher prices. The
Abbot of Glastonbury paid 13th-century champion Henry of Fernberg £ 20 to battle on his
behalf in a property dispute. The terms of Fernberg.s contract stipulated partial payment
19The Church. itself a large land owner. had its own rules governing land alienations. To alienate Church
land, a landholder required the prelate.s and chapter.s consent. See, Cheney (1985).
when he wagered battle, another part before he fought, and the rest if he struck his opponent
but once in the arena. An evidently inferior 13th-century champion, John of Smerill, com-
manded less than half this amount for agreeing to battle for William Heynton. His contract
paid him only £ 8 if he defeated his opponent and nothing if he failed to land a blow (Russell
In contrast to the medieval land market, the champion market was .uid. Champions
switched allegiances before battle, reshu ing themselves into the service of the higher bidder.
They were happy .to desert to the other side if the inducement was su¢ ciently great.(Russell
1959: 254-256). The feudal structure of land rights heightened transaction costs in the land
market. But it didn.t a¤ect those costs in the champion market. Unlike alienating his land,
which could require a tenant to secure his lord.s and heirs.consent, the tenant required no
one.s consent to hire a champion.
Hiring a superior champion wasn.t the only way for medieval disputants to .bid. on
disputed land. They could also hire more champions. Only one champion fought. But
purchasing multiple champions. especially the better ones. shrank the other disputant.s
choices, leaving him fewer and inferior options.
In 1220 a demandant named Cliveden contested the right to a parcel of land then under
the tenancy of fellow named Ken. Ken hired four champions, one of them the redoubtable
William of Copeland. Similarly, in a case of contested .shing rights between the Abbot of
Meaux and the Abbot of St. Mary.s of York, Meaux hired seven champions .at great cost..
Meaux was attempting to .monopolise the market.for professional battlers to .compel the
other Abbot to employ a second-rate champion.(Russell 1959: 246, 255).
To see how trial by battle.s violent auctions a¤ected contested property.s allocation,
consider two medieval Englishmen, Eustace and Osbert. Eustace goes before the king.s
court and claims the farmland Osbert occupies is his. Osbert denies Eustace.s claim. Both
o¤er to prove their right on their champion.s body. Property rights in land are perfectly
sticky: the transaction cost of trading them is prohibitive. Whoever the legal system awards
the farmland to will be its permanent holder.
20Hiring champions, even under a contract as advantageous to the employer as the one Heynton negotiated,
probably always required some up-front expense. For example, Heynton had to put up a parcel of his property
to collateralize his promise to pay Smerill if Smerill won.
The court doesn.t know who the farmland truly belongs to. It orders trial by battle.
There are two champions available for hire: Fernberg and Smerill. Fernberg has a reputation
as a great .ghter. Smerill doesn.t. Both champions sell their services to the highest bidder.
Eustace is a more productive farmer than Osbert. So he values the contested land
more. Eustace is therefore willing to pay more for Fernberg.s services than Osbert. He hires
Fernberg, leaving Osbert with Smerill. The combat.s probable outcome is Fernberg.s victory.
Eustace, the higher-valuing user, wins the property right. Trial by battle has used a violent
auction to reveal the higher-valuing user.s identity and allocate the contested land to him.
It has substituted for the Coase theorem where sticky property rights prevented trade from
allocating contested farmland e¢ ciently.
As in any auction, in trial by battle.s violent auction, .bids. and valuations weren.t
perfectly correlated. This imperfection isn.t just true of literal auctions and implicit ones
such as trial by battle. It.s equally true of those .auctions.we call .markets..Auctions, like
markets, only tend to allocate resources to their higher-valuing users. One reason e¢ cient
allocation is a tendency instead of a certainty is that bidders have di¤erent endowments.
Because they have di¤erent endowments, bids and valuations may diverge.
Credit markets, which allow bidders to make bids using others.funds, can help mitigate
this divergence. But credit markets are imperfect. So liquidity constraints may still in.uence
bids. Like modern market participants, medieval citizens could also turn to credit markets
if necessary (see, for instance, Koyama 2010a, 2010b). Medieval credit markets were un-
doubtedly more imperfect than modern ones. And it.s unclear whether those markets made
loans to legal contestants seeking champions. Still, at least in principle, a higher-valuing but
liquidity constrained disputant could borrow for this purpose if he needed to.
A second factor may have also helped mitigate the divergence between bids and valuations
that large endowment disparities might create under trial by battle. According to Russell
(1980: 120), under Norman England.s legal system .Battle seems to have been barred be-
tween people of widely di¤ering status..Presumably disputants with similar .status.had
similar endowments. This bar likely ameliorated the in.uence wealth di¤erences exerted on
trial by battle.s outcomes, improving violent auctions.ability to allocate disputed property
rights e¢ ciently.21
4 Rent Seeking
Trial by battle.s violent auctions not only encouraged the e¢ cient allocation of disputed
property rights when those rights.true owners couldn.t be identi.ed. They encouraged less
rent seeking than one might expect. Wherever there are auctions, there are bids. And
wherever there are bids, there are bid recipients. In the context of a .legal auction.such
as trial by battle, bid recipients pose a problem. The bids themselves are simply transfers.
However, they nefariously in.uence bid recipients.incentives. Since bid recipients.incomes
depend on bids, which in turn depend on land disputes to generate those bids, legal auctions.
bid recipients have an incentive to permit or instigate illegitimate property con.icts.
Illegitimate land disputes, which result from bid recipients.attempts to raise their in-
comes instead of from genuinely felt ownership disagreements, undermine property rights.
They constitute socially costly rent-seeking activity rather than socially productive owner-
ship resolution. Individuals who confront the spectre of rampant rent seeking are insecure in
their property rights. They live in constant fear that fraudulent legal challengers will deprive
them of their property and therefore have weak incentives to invest in their land.
The higher the bids disputants make in a legal auction, the greater is bid recipients.
payo¤ of rent seeking, and the weaker is individuals.incentive to invest in their land. Higher
bids, and thus bid receipts, make it more pro.table for bid recipients to permit and initiate
illegitimate property disputes. Higher bids, and thus bid receipts, also raise bid recipi-
ents.incomes. By increasing the relative payo¤ of being a bid recipient, higher bids attract
resources away from socially productive, wealth-creating industries into the socially unpro-
ductive, bid-recipient industry. Suppressing a legal auction.s social cost therefore requires
21 If he couldn.t access credit markets, a tenant might nevertheless be able to get his lord to chip in to
help him hire an appropriate champion. If in his lord.s eyes the tenant was the higher-valuing user, the lord
would.ve been happy to do this. In this situation the lord would.ve lost revenue if his tenant lost the case
since the challenger was less productive and thus would.ve produced less for him. Instead, if in the lord.s
eyes his tenant wasn.t the higher-valuing user, he wouldn.t have been willing to chip in to help the tenant
hire a good champion. Such a tenant, then, would.ve been more likely to lose his land. But this outcome
would still be e¢ cient: the higher-valuing user would tend to win the contested property. Thus lords may
have had an interest to behave in ways that promoted e¢ cient allocation under trial by battle.
suppressing the height of the bids it generates.
In trial by battle.s violent legal auctions, the bid recipients were champions. Thus cham-
pions (and complicit demandants) had an incentive to rent seek. The better ones could
encourage unscrupulous demandants to initiate fraudulent claims, challenging ownership to
land the demandant knew wasn.t his. Without data on medieval disputants.champion ex-
penditures, it.s impossible to measure the extent of such rent seeking under trial by battle.
But indirect evidence suggests rent seeking wasn.t rampant.
Citizens could and occasionally did hire champions on retainer. At least one English king
hired a champion this way. He paid his champion 3d. per day whether he used the thug.s
services or not. Champion Thomas of Bruges managed to sell his services on retainer as
well. However, champion retainers were uncommon. .[M]ost people.came .to terms with
an available champion only when litigation was imminent.(Russell 1959: 253, 254).
The infrequency of champions on retainer suggests property disputes weren.t ubiquitous.
Most people didn.t feel their land rights were so insecure as to warrant the employment of
a permanent champion to defend them. If rent seeking had been rampant, illegitimate land
disputes would.ve been rampant too. Perpetually threatened by the specter of fraudulent
demandants and eager to perpetrate fraudulent claims of their own, most people would.ve
found it worthwhile to keep their champion of choice at his ready in their permanent em-
ployment. The fact that they didn.t is reassuring.
There are other reasons most medieval Englishmen may have found it unpro.table to keep
champions on retainer. Many people may have been unable to a¤ord retained champions
though they would.ve liked them. Alternatively, the supply of quality champions may have
been very elastic, precluding the need to have a stable of champions on retainer since citizens
could procure them easily on the spot market when the need arose.
Still, it.s reasonable to expect to .nd a large number of retained legal representatives
under a legal system in which people feel that their property rights are constantly threatened
by rent-seeking litigiousness or in which rampant rent-seeking opportunity gives them an
incentive to behave litigiously themselves. The rarity of retained champions in medieval
England therefore suggests that rent seeking under trial by battle wasn.t ubiquitous.
This helps resolve a puzzle that trial by battle.s violent auctions pose: Why didn.t Nor-
man England.s legal system use .regular.auctions. the .rst-price ascending-bid variety. to
auction contested property rights to disputants instead?22
Because regular auctions would.ve encouraged more rent seeking than violent ones. As
in violent auctions, in regular ones, too, there are bid recipients. Those recipients.identity
depends on the auction.s arrangement. There are two obvious bid recipients in a regular
legal auction: the loser and the legal system. Both arrangements encourage rent seeking.
If auction proceeds accrue to losers, individuals have an incentive to initiate baseless legal
disputes to extort current owners. If proceeds accrue to the legal system, say to the king,
or to the judges, o¢ cials have an incentive to permit and create .ctitious property con.icts.
For example, a judge may ignore the absence of basic evidence required to render a claim
colorable, such as the presence of witnesses, and permit the case to go forward to auction. Or
he might ignore clear evidence of the tenant.s ownership and pretend he can.t be sure who
the land belongs to, again allowing the case to go to auction to decide its outcome. The judge
may even instigate illegitimate disputes, encouraging a citizen to fraudulently challenge an
existing landholder.s claim, o¤ering some of the auction proceeds in exchange.23
Trial by battle.s violent auctions encouraged less rent seeking than regular auctions. and
thus were less socially costly. because they generated lower bid receipts, which motivate
rent-seeking behavior. To see why trial by battle.s violent auctions generated lower bid
receipts than regular auctions, consider two risk-neutral legal disputants, a tenant, T, and a
demandant, D. T values the disputed land vT . D values it more: vD > vT > 0. Disputants
know their own and each other.s values. Judges don.t. They require some demand-revelation
mechanism to identify the higher-valuing disputant.
The amount legal disputants spend to in.uence contested land.s legal assignment. their
.bids.. measures the resources .owing to parties with an interest in permitting or instigating
22A sealed high-bid auction produces the same total spending as a .rst-price ascending-bid auction (see
Hirshleifer and Riley 1992: 373). Thus violent auction.s rent-seeking superiority to .regular.auction, which
I derive below, applies equally to the sealed high-bid auction, which one might consider an equally obvious
alternative to trial by battle.s violent auction.
23Friedman.s (1999) excellent paper, which points to the ine¢ ciency of e¢ cient punishments, is closely
related to this. He makes the point that with an e¢ cient punishment, the cost borne by the punished is
captured as a corresponding bene.t by someone else. This is both what makes the punishment e¢ cient, but
also what may make it ine¢ cient in that this situation creates an incentive for the bene.t.s recipient to seek
rents. Using auctions to allocate disputed property rights is analogous.
illegitimate land disputes. the bid recipients. In a regular auction this amount equals the
contested property.s value to the lower-valuing disputant: vT . D knows that if he bids less
than this, T will outbid him. If he bids ” > 0 more, T drops out. D wins the auction. He
spends vT + ” to do so.
Trial by battle is di¤erent. Its violent auction is equivalent to an imperfectly discrimi-
nating all-pay auction with asymmetric valuations.24 In such an auction contestants make
expenditures to improve their probability of winning some prize that has a di¤erent value
to each of them. These expenditures are equivalent to .bids.for the prize. Each contestant
pays his bid. His probability of winning the prize depends on how much he spends to win it
compared to the other contestant.
Ceteris paribus, the more a contestant spends to win the prize, the more likely he is to
win it and vice versa. The auction is imperfectly discriminating because neither contestant
wins the prize with certainty as long as the other contestant spends something to win it too.
The higher .bidder.is more likely to win. But it.s possible for the lower bidder to upset
In trial by battle.s violent auctions the contestants are the legal disputants, T and D,
contesting each other.s right to a piece land. The prize is the land, which they value vT
and vD, respectively, where vD > vT > 0. The disputants bid by spending on champions
who .ght in the arena for their employers. right. T and D spend t > 0 and d > 0 on
champions, respectively, to improve their chance of winning the contested land. They make
their expenditures simultaneously and independently. Expenditures on champions and the
land.s value are in the same units.
Intuitively two factors determine how much a disputant will be willing to spend to win
the contested land, and thus the probability that he wins it, in such an auction: how much he
values the contested land and how much his adversary values it. Ceteris paribus, a disputant
will be willing to spend more to win contested land that he values more and vice versa.
So his optimal spending level depends partly on his valuation of the land. Ceteris paribus,
a disputant will also be willing to spend more to win contested land when his adversary
spends more to win it and vice versa. His optimal spending level is his best response to his
24For example, see Nti (1999) who models a rent-seeking contest with asymmetric valuations.
adversary.s spending level. When his adversary spends more, he must spend more himself
to buy the same chance of winning. So a disputant.s optimal level of spending on champions
also depends partly on his adversary.s level of spending. And that depends partly on his
adversary.s valuation of the land.
To calculate how much disputants spend to in.uence contested land.s legal assignment
under trial by battle, I must .rst determine how much each disputant bids for this land in
a violent auction in equilibrium. Tullock.s (1980b) contest success function describes each
disputant.s probability of winning that auction given his and his adversary.s expenditures
on champions. D.s probability of winning the contested land under trial by battle is
_D(d; t) =
d_ + t_ : (1.1)
T.s probability of winning is
_T (d; t) = 1