Production, Information C os ts , and

Economic Organization
BY ARMEN A. ALC HIAN AND HAROLD DEMSETZ*
The mark of a capitalis tic s ociety is that
res ources are owned and allocated by s uch
nongovernmental organizations as firms ,
hous eholds , and markets . Res ource owners
increas e productivity through cooperative
s pecialization and this leads to the demand
for economic organizations which facili-
tate cooperation. When a lumber mill
employs a cabinetmaker, cooperation be-
tween s pecialis ts is achieved within a firm,
and when a cabinetmaker purchas es wood
from a lumberman, the cooperation takes
place acros s markets (or between firms ).
Two important problems face a theory of
economic organization-to explain the
conditions that determine whether the
gains from s pecialization and cooperative
production can better be obtained within
an organization like the firm, or acros s
markets , and to explain the s tructure of
the organization.
It is common to s ee the firm charac-
terized by the power to s ettle is s ues by
fiat, by authority, or by dis ciplinary action
s uperior to that available in the conven-
tional market. This is delus ion. The firm
does not own all its inputs . It has no
power of fiat, no authority, no dis ciplinary
action any different in the s lightes t degree
from ordinary market contracting be-
tween any two people. I can "punis h" you
only by withholding future bus ines s or by
s eeking redres s in the courts for any failure
to honor our exchange agreement. That is
exactly all that any employer can do. He
can fire or s ue, jus t as I can fire my grocer
by s topping purchas es from him or s ue
him for delivering faulty products . What
then is the content of the pres umed power
to manage and as s ign workers to various
tas ks ? Exactly the s ame as one little con-
s umer's power to manage and as s ign his
grocer to various tas ks . The s ingle con-
s umer can as s ign his grocer to the tas k of
obtaining whatever the cus tomer can in-
duce the grocer to provide at a price ac-
ceptable to both parties . That is precis ely
all that an employer can do to an em-
ployee. To s peak of managing, directing,
or as s igning workers to various tas ks is a
deceptive way of noting that the employer
continually is involved in renegotiation of
contracts on terms that mus t be acceptable
to both parties . Telling an employee to
type this letter rather than to file that
document is like my telling a grocer to
s ell me this brand of tuna rather than that
brand of bread. I have no contract to con-
tinue to purchas e from the grocer and
neither the employer nor the employee is
bound by any contractual obligations to
continue their relations hip. Long-term
contracts between employer and em-
ployee are not the es s ence of the organiza-
tion we call a firm. My grocer can count
on my returning day after day and pur-
chas ing his s ervices and goods even with
the prices not always marked on the goods
-becaus e I know what they are-and he
adapts his activity to conform to my
directions to him as to what I want each
day . .. he is not my employee.
Wherein then is the relations hip be-
tween a grocer and his employee different
from that between a grocer and his cus -
*
Profes s ors of economics at the Univers ity of C ali-
fornia, Los Angeles . Acknowledgment is made for finan-
cial aid from the E. Lilly Endowment, Inc. grant to
UC LA for res earch in the behavioral effects of property
rights .
777
778 THE AMERIC AN EC ONOMIC REVIEW
tomers ? It is in a team us e of inputs and a
centralized pos ition of s ome party in the
contractual arrangements of all other in-
puts . It is the centralized contractual agent
in a team productive proces s not s ome
s uperior authoritarian directive or dis -
ciplinary power. Exactly what is a team
proces s and why does it induce the con-
tractual form, called the firm? Thes e prob-
lems motivate the inquiry of this paper.
I. The Metering Problem
The economic organization through
which input owners cooperate will make
better us e of their comparative advantages
to the extent that it facilitates the pay-
ment of rewards in accord with produc-
tivity. If rewards were random, and with-
out regard to productive effort, no in-
centive to productive effort would be pro-
vided by the organization; and if rewards
were negatively correlated with produc-
tivity the organization would be s ubject
to s abotage. Two key demands are placed
on an economic organization-metering
input productivity and metering rewards .'
Metering problems s ometimes can be
res olved well through the exchange of
products acros s competitive markets , be-
caus e in many s ituations markets yield a
high correlation between rewards and
productivity. If a farmer increas es his out-
put of wheat by 10 percent at the pre-
vailing market price, his receipts als o in-
creas e by 10 percent. This method of or-
ganizing economic activity meters the
output directly, reveals the marginal prod-
uct and apportions the rewards to re-
s ource owners in accord with that direct
meas urement of their outputs . The s ucces s
of this decentralized, market exchange in
promoting productive s pecialization re-
quires that changes in market rewards fall
on thos e res pons ible for changes in output.2
The clas s ic relations hip in economics
that runs from marginal productivity to
the dis tribution of income implicitly as -
s umes the exis tence of an organization, be
it the market or the firm, that allocates
rewards to res ources in accord with their
productivity. The problem of economic
organization, the economical means of
metering productivity and rewards , is not
confronted directly in the clas s ical anal-
ys is of production and dis tribution. In-
s tead, that analys is tends to as s ume s uf-
ficiently economic or zero cos t means ,
as if productivity automatically created
its reward. We conjecture the direction of
caus ation is the revers e the s pecific s ys -
I
Meter means to meas ure and als o to apportion. One
can meter (meas ure) output and one can als o meter
(control) the output. We us e the word to denote both;
the context s hould indicate which.
2
A producer's wealth would be reduced by the pres -
ent capitalized value of the future income los t by los s of
reputation. Reputation, i.e., credibility, is an as s et,
which is another way of s aying that reliable information
about expected performance is both a cos tly and a
valuable good. For acts of God that interfere with con-
tract performance, both parties have incentives to
reach a s ettlement akin to that which would have been
reached if s uch events had been covered by s pecific con-
tingencv claus es . The reas on, again, is that a reputation
for "hones t" dealings -i.e., for actions s imilar to thos e
that would probably have been reached had the con-
tract provided this contingency-is wealth.
Almos t every contract is open-ended in that many
contingencies are uncovered. For example, if a fire
delays production of a promis ed product by A to B,
and if B contends that A has not fulfilled the contract,
how is the dis pute s ettled and what recompens e, if any,
does A grant to B? A pers on uninitiated in s uch ques -
tions may be s urpris ed by the extent to which contracts
permit either party to es cape performance or to nullify
the contract. In fact, it is hard to imagine any contract,
which, when taken s olely in terms of its s tipulations ,
could not be evaded by one of the parties . Yet that is
the ruling, viable type of contract. Why? Undoubtedly
the bes t dis cus s ion that we have s een on this ques tion is
by Stewart Macaulay.
There are means not only of detecting or preventing
cheating, but als o for deciding how to allocate the los s es
or gains of unpredictable events or quality of items
exchanged. Sales contracts contain warranties , guaran-
tees , collateral, return privileges and penalty claus es for
s pecific nonperformance. Thes e are means of as s ignment
of ris ks of los s es of cheating. A lower price without war-
ranty-an "as is " purchas e-places more of the ris k on
the buyer while the s eller buys ins urance agains t los s es
of his "cheating." On the other hand, a warranty or
return privilege or s ervice contract places more ris k on
the s eller with ins urance being bought by the buyer.
ALC HIAN AND DEMSETZ: EC ONOMIC ORGANIZATION 779
tem of rewarding which is relied upon
s timulates a particular productivity re-
s pons e. If the economic organization
meters poorly, with rewards and produc-
tivity only loos ely correlated, then pro-
ductivity will be s maller; but if the eco-
nomic organization meters well produc-
tivity will be greater. What makes meter-
ing difficult and hence induces means of
economizing on metering cos ts ?
II. Team Production
Two men jointly lift heavy cargo into
trucks . Solely by obs erving the total
weight loaded per day, it is impos s ible to
determine each pers on's marginal pro-
ductivity. With team production it is
difficult, s olely by obs erving total output,
to either define or determine each indivi-
dual's contribution to this output of the
cooperating inputs . The output is yielded
by a team, by definition, and it is not a
s um of s eparable outputs of each of its
members . Team production of Z involves
at leas t two inputs , Xi and
Xj,
with
a2Z1aXiaXj3X,0.3
The production func-
tion is not s eparable into two functions
each involving only inputs Xi
or only in-
puts
Xi.
C ons equently there is no s um
of Z of two s eparable functions to treat
as the Z of the team production function.
(An example of a s eparable cas e is Z= aX'
+bXj'
which is s eparable into Zi= aX' and
Zj= bX%,
and Z=
Zi+Zj.
This is not team
production.) There exis t production tech-
niques in which the Z obtained is greater
than if
Xi and Xj
had produced s eparable
Z. Team production will be us ed if it
yields an output enough larger than the
s um of s eparable production of Z to cover
the cos ts of organizing and dis ciplining
team members -the topics of this paper.4
Us ual explanations of the gains from
cooperative behavior rely on exchange
and production in accord with the com-
parative advantage s pecialization prin-
ciple with s eparable additive production.
However, as s ugges ted above there is a
s ource of gain from cooperative activity
involving working as a team, wherein in-
dividual cooperating inputs do not yield
identifiable, s eparate products which can
be s ummed to meas ure the total output.
For this cooperative productive activity,
here called "team" production, meas uring
marginal productivity and making pay-
ments in accord therewith is more expen-
s ive by an order of magnitude than for
s eparable production functions .
Team production, to repeat, is produc-
tion in which 1) s everal types of res ources
are us ed and 2) the product is not a s um
of s eparable outputs of each cooperating
res ource. An additional factor creates a
team organization problem-3) not all re-
s ources us ed in team production belong to
one pers on.
We do not inquire into why all the
jointly us ed res ources are not owned by
one pers on, but ins tead into the types of
organization, contracts , and informational
and payment procedures us ed among
owners of teamed inputs . With res pect to
the one-owner cas e, perhaps it is s ufficient
merely to note that (a) s lavery is pro-
hibited, (b) one might as s ume ris k aver-
s ion as a reas on for one pers on's not bor-
rowing enough to purchas e all the as s ets
or s ources of s ervices rather than renting
them, and (c) the purchas e-res ale s pread
may be s o large that cos ts of s hort-term
owners hip exceed rental cos ts . Our prob-
lem is viewed bas ically as one of organiza-
tion among different people, not of the
phys ical goods or s ervices , however much
there mus t be s election and choice of com-
bination of the latter.
How can the members of a team be re-
warded and induced to work efficiently?
I
The function is s eparable into additive functions if
the cros s partial derivative is zero, i.e., if
a2Z/1XjaX;=
O.
4
With s ufficient generality of notation and concep-
tion this team production function could be formulated
as a cas e of the generalized production function inter-
pretation given by our colleague, E. A. Thomps on.
780 THE AMERIC AN EC ONOMIC REVIEW
In team production, marginal products of
cooperative team members are not s o
directly and s eparably (i.e.,
cheaply) ob-
s ervable. What a team offers to the
market can be taken as the marginal
product of the team but not of the team
members . The cos ts of metering or as cer-
taining the marginal products of the team's
members is what calls forth new organiza-
tions and procedures . C lues to each input's
productivity can be s ecured by obs erving
behavior of individual inputs . When lifting
cargo into the truck, how rapidly does a
man move to the next piece to be loaded,
how many cigarette breaks does he
take, does the item being lifted tilt down-
ward toward his s ide?
If detecting s uch behavior were cos t-
les s , neither party would have an incen-
tive to s hirk, becaus e neither could impos e
the cos t of his s hirking on the other (if
their cooperation was agreed to volun-
tarily). But s ince cos ts mus t be incurred
to monitor each other, each input owner
will have more incentive to s hirk when he
works as part of a team, than if his per-
formance could be monitored eas ily or if
he did not work as a team. If there is a
net increas e in productivity available by
team production, net of the metering cos t
as s ociated with dis ciplining the team, then
team production will be relied upon rather
than a multitude of bilateral exchange of
s eparable individual outputs .
Both leis ure and higher income enter a
pers on's utility function.5 Hence, each
pers on s hould adjus t his work and realized
reward s o as to equate the marginal rate of
s ubs titution between leis ure and produc-
tion of real output to his marginal rate of
s ubs titution in cons umption. That is , he
would adjus t his rate of work to bring his
demand prices of leis ure and output to
equality with their true cos ts . However,
with detection, policing, monitoring,
mea-
s uring or metering cos ts , each pers on will
be induced to take more leis ure, becaus e
the effect of relaxing on his realized (re-
ward) rate of s ubs titution between output
and leis ure will be les s than the effect on
the true rate of s ubs titution. His realized
cos t of leis ure will fall more than the true
cos t of leis ure, s o he "buys " more leis ure
(i.e., more nonpecuniary reward).
If his relaxation cannot be detected per-
fectly at zero cos t, part of its effects will
be borne by others in the team, thus mak-
ing his realized cos t of relaxation les s than
the true total cos t to the team. The dif-
ficulty of detecting s uch actions permits
the private cos ts of his actions to be les s
than their full cos ts . Since each pers on
res ponds to his private realizable rate of
s ubs titution (in production) rather than
the true total (i.e., s ocial) rate, and s o
long as there are cos ts for other people to
detect his s hift toward relaxation, it will
not pay (them) to force him to readjus t
completely by making him realize the
true cos t. Only enough efforts will be
made to equate the marginal gains of de-
tection activity with the marginal cos ts of
detection; and that implies a lower rate of
productive effort and more s hirking than
in a cos tles s monitoring, or meas uring,
world.
In a univers ity, the faculty us e office
telephones , paper, and mail for pers onal
us es beyond s trict univers ity productivity.
The univers ity adminis trators could s top
s uch practices by identifying the res pon-
s ible pers on in each cas e, but they can do s o
only at higher cos ts than adminis trators
are willing to incur. The extra cos ts of
identifying each party (rather than merely
identifying the pres ence of s uch activity)
would exceed the s avings from diminis hed
faculty "turpitudinal peccadilloes ."
So
the faculty is allowed s ome degree of
"privileges , perquis ites , or fringe benefits ."
And the total of the pecuniary wages paid
5
More precis ely:
"if anything other than pecuniary
income enters his ultility function." Leis ure s tands for all
nonpecuniary income for s implicity of expos ition.
ALC HIAN AND DEMSETZ: EC ONOMIC ORGANIZATION 781
is lower becaus e of this irreducible (at
acceptable cos ts ) degree of amenity-s eiz-
ing activity. Pay is lower in pecuniary
terms and higher in leis ure, conveniences ,
and eas e of work. But s till every pers on
would prefer to s ee detection made more
effective (if it were s omehow pos s ible to
monitor cos tles s ly) s o that he, as part of
the now more effectively producing team,
could thereby realize a higher pecuniary
pay and les s leis ure. If everyone could, at
zero cos t, have his reward-realized rate
brought to the true production pos s ibility
real rate, all could achieve a more pre-
ferred pos ition. But detection of the re-
s pons ible parties is cos tly; that cos t acts
like a tax on work rewards .6 Viable s hirk-
ing is the res ult.
What forms of organizing team produc-
tion will lower the cos t of detecting "per-
formance" (i.e., marginal productivity)
and bring pers onally realized rates of
s ubs titution clos er to true rates of s ub-
s titution? Market competition, in prin-
ciple, could monitor s ome team produc-
tion. (It already organizes teams .) Input
owners who are not team members can
offer, in return for a s maller s hare of the
team's rewards , to replace exces s ively (i.e.,
overpaid) s hirking members . Market com-
petition among potential team members
would determine team members hip and
individual rewards . There would be no
team leader, manager, organizer, owner,
or employer. For s uch decentralized or-
ganizational control to work, outs iders ,
pos s ibly after obs erving each team's total
output, can s peculate about their capabili-
ties as team members and, by a market
competitive proces s , revis ed teams with
greater productive ability will be formed
and s us tained. Incumbent members will be
cons trained by threats of replacement by
outs iders offering s ervices for lower reward
s hares or offering greater rewards to the
other members of the team. Any team
member who s hirked in the expectation
that the reduced output effect would not
be attributed to him will be dis placed if
his activity is detected. Teams of produc-
tive inputs , like bus ines s units , would
evolve in apparent s pontaneity in the
market-without any central organizing
agent, team manager, or bos s .
But completely effective control cannot
be expected from individualized market
competition for two reas ons . Firs t, for
this competition to be completely effec-
tive, new challengers for team members hip
mus t know where, and to what extent,
s hirking is a s erious problem, i.e., know
they can increas e net output as compared
with the inputs they replace. To the extent
that this is true it is probably pos s ible for
exis ting fellow team members to recognize
the s hirking. But, by definition, the detec-
tion of s hirking by obs erving team output
is cos tly for team production. Secondly, as -
s ume the pres ence of detection cos ts , and
as s ume that in order to s ecure a place on
the team a new input owner mus t accept
a s maller s hare of rewards (or a promis e to
produce more). Then his incentive to s hirk
would s till be at leas t as great as the in-
centives of the inputs replaced, becaus e he
s till bears les s than the entire reduction in
team output for which he is res pons ible.
III. The C las s ical Firm
One method of reducing s hirking is for
s omeone to s pecialize as a monitor to check
the input performance of team members .7
6
Do not as s ume that the s ole res ult of the cos t of
detecting s hirking is one form of payment (more leis ure
and les s take home money). With s everal members of
the team, each has an incentive to cheat agains t each
other by engaging in more than the average amount of
s uch leis ure if the employer can not tell at zero cos t
which employee is taking more than average. As a
res ult the total productivity of the team is lowered.
Shirking detection cos ts thus change the form of pay-
ment and als o res ult in lower total rewards . Becaus e the
cros s partial derivatives are pos itive, s hirking reduces
other people's marginal products .
I
What is meant by
performance?
Input energy, ini-
tiative, work attitude, pers piration, rate of exhaus tion?
(C ontinuedl)
782 THE AMERIC AN EC ONOMIC REVIEW
But who will monitor the monitor? One
cons traint on the monitor is the afores aid
market competition offered by other
monitors , but for reas ons already given,
that is not perfectly effective. Another
cons traint can be impos ed on the monitor:
give him title to the net earnings of the
team, net of payments to other inputs .
If owners of cooperating inputs agree with
the monitor that he is to receive any
res idual product above pres cribed amounts
(hopefully, the marginal value products of
the other inputs ), the monitor will have
an added incentive not to s hirk as a
monitor. Specialization in monitoring plus
reliance on a res idual claimant s tatus will
reduce s hirking; but additional links are
needed to forge the firm of clas s ical eco-
nomic theory. How will the res idual
claimant monitor the other inputs ?
We us e the term monitor to connote
s everal activities in addition to its dis -
ciplinary connotation. It connotes mea-
s uring output performance, apportioning
rewards , obs erving the input behavior of
inputs as means of detecting or es timating
their marginal productivity and giving as -
s ignments or ins tructions in what to do
and how to do it. (It als o includes , as we
s hall s how later, authority to terminate
or revis e contracts .) Perhaps the contras t
between a football coach and team cap-
tain is helpful. The coach s elects s trategies
and tactics and s ends in ins tructions
about what plays to utilize. The captain
is es s entially an obs erver and reporter of
the performance at clos e hand of the mem-
bers . The latter is an ins pector-s teward
and the former a s upervis or manager.
For the pres ent all thes e activities are in-
cluded in the rubric "monitoring." All
thes e tas ks are, in principle, negotiable
acros s markets , but we are pres uming that
s uch market meas urement of marginal
productivities and job reas s ignments are
not s o cheaply performed for team pro-
duction. And in particular our analys is
s ugges ts that it is not s o much the cos ts
of s pontaneous ly negotiating contracts in
the markets among groups for team pro-
duction as it is the detection of the per-
formance of individual members of the
team that calls for the organization noted
here.
The s pecialis t who receives the res idual
rewards will be the monitor of the mem-
bers of the team (i.e., will manage the us e
of cooperative inputs ). The monitor earns
his res idual through the reduction in
s hirking that he brings about, not only by
the prices that he agrees to pay the owners
of the inputs , but als o by obs erving and
directing the actions or us es of thes e in-
puts . Managing or examining the ways to
which inputs are us ed in team production
is a method of metering the marginal pro-
ductivity of individual inputs to the team's
output.
To dis cipline team members and reduce
s hirking, the res idual claimant mus t have
power to revis e the contract terms and in-
centives of individual members without
having to terminate or alter every other
input's contract. Hence, team members
who s eek to increas e their productivity
will as s ign to the monitor not only the
res idual claimant right but als o the right
to alter individual members hip and per-
formance on the team. Each team mem-
ber, of cours e, can terminate his own
members hip (i.e., quit the team), but
only the monitor may unilaterally ter-
minate the members hip of any of the
Or output? It is the latter that is s ought-the effect or
output. But performance is nicely ambiguous becaus e it
s ugges ts both input and output. It is nicely ambiguous
becaus e as we s hall s ee, s ometimes by ins pecting a team
member's input activitv we can better judge his output
effect, perhaps not with complete accuracy but better
than by watching the output of the team. It is not always
the cas e that watching input activity is the only or bes t
means of detecting, meas uring or monitoring output
effects of each team member, but in s ome cas es it is a
us eful way. For the moment the word performance
glos s es over thes e as pects and facilitates concentration
on other is s ues .
ALC HIAN AND DEMSETZ: EC ONOMIC ORGANIZATION 783
other members without neces s arily ter-
minating the team its elf or his as s ociation
with the team; and he alone can expand or
reduce members hip, alter the mix of
members hip, or s ell the right to be the
res idual claimant-monitor of the team. It
is this entire bundle of rights : 1) to be a
res idual claimant; 2) to obs erve input
behavior; 3) to be the central party com-
mon to all contracts with inputs ; 4) to
alter the members hip of the team; and
5) to s ell thes e rights , that defines the
owners hip (or the employer) of the clas s ical
(capitalis t, free-enterpris e) firm. The
coales cing of thes e rights has aris en, our
analys is as s erts , becaus e it res olves the
s hirking-information problem of team
production better than does the noncen-
tralized contractual arrangement.
The relations hip of each team member
to the owner of the firm (i.e., the party
common to all input contracts and the
res idual claimant) is s imply a "quid pro
quo" contract. Each makes a purchas e
and s ale. The employee "orders " the owner
of the team to pay him money in the s ame
s ens e that the employer directs the team
member to perform certain acts . The
employee can terminate the contract as
readily as can the employer, and long-
term contracts , therefore, are not an es -
s ential attribute of the firm. Nor are
"authoritarian," "dictational," or "fiat"
attributes relevant to the conception of the
firm or its efficiency.
In s ummary, two neces s ary conditions
exis t for the emergence of the firm on the
prior as s umption that more than pecuniary
wealth enter utility functions : 1) It is
pos s ible to increas e productivity through
team-oriented production, a production
technique for which it is cos tly to directly
meas ure the marginal outputs of the co-
operating inputs . This makes it more
difficult to res trict s hirking through s imple
market exchange between cooperating in-
puts . 2) It is economical to es timate mar-
ginal productivity by obs erving or s pecify-
ing input behavior. The s imultaneous oc-
currence of both thes e preconditions leads
to the contractual organization of inputs ,
known as the clas s ical capitalis t firms with
(a) joint input production, (b) s everal in-
put owners , (c) one party who is common
to all the contracts of the joint inputs , (d)
who has rights to renegotiate any input's
contract independently of contracts with
other input owners , (e) who holds the
res idual claim, and (f) who has the right
to s ell his central contractual res idual
s tatus .8
Other Theories of the Firm
At this juncture, as an as ide, we briefly
place this theory of the firm in the contexts
of thos e offered by Ronald C oas e and
Frank Knight.9 Our view of the firm is not
neces s arily incons is tent with C oas e's ; we
attempt to go further and identify refut-
able implications . C oas e's penetrating in-
s ight is to make more of the fact that
markets do not operate cos tles s ly, and he
relies on the cos t of us ing markets to form
contracts as his bas ic explanation for the
exis tence of firms . We do not dis agree with
the propos ition that, ceteris paribus , the
higher is the cos t of trans acting acros s
markets the greater will be the compara-
tive advantage of organizing res ources
within the firm; it is a difficult propos ition
to dis agree with or to refute. We could
with equal eas e s ubs cribe to a theory of
the firm bas ed on the cos t of managing,
for s urely it is true that, ceteris paribus ,
the lower is the cos t of managing the
greater will be the comparative advantage
of organizing res ources within the firm. To
move the theory forward, it is neces s ary
to know what is meant by a firm and to
8
Removal of (b) converts a capitalis t proprietary firm
to a s ocialis t firm.
9
Recognition mus t als o be made to the s eminal in-
quiries by Morris Silver anid Richard Aus ter, and by
H. B. Malmgren.
784 THE AMERIC AN EC ONOMIC REVIEW
explain the circums tances under which
the cos t of "managing" res ources is low
relative to the cos t of allocating res ources
through market trans action. The concep-
tion of and rationale for the clas s ical firm
that we propos e takes a s tep down the
path pointed out by C oas e toward that
goal. C ons ideration of team production,
team organization, difficulty in metering
outputs , and the problem of s hirking are
important to our explanation but, s o far
as we can as certain, not in C oas e's . C oas e's
analys is ins ofar as it had heretofore been
developed would s ugges t open-ended con-
tracts but does not appear to imply any-
thing more-neither the res idual claimant
s tatus nor the dis tinction between em-
ployee and s ubcontractor s tatus (nor any
of the implications indicated below). And
it is not true that employees are generally
employed on the bas is of long-term con-
tractual arrangements any more than on a
s eries of s hort-term or indefinite length
contracts .
The importance of our propos ed addi-
tional elements is revealed, for example,
by the explanation of why the pers on to
whom the control monitor is res pons ible
receives the res idual, and als o by our
later dis cus s ion of the implications about
the corporation, partners hips , and profit
s haring. Thes e alternative forms for or-
ganization of the firm are difficult to re-
s olve on the bas is of market trans action
cos ts only. Our expos ition als o s ugges ts a
definition of the clas s ical firm-s omething
crucial that was heretofore abs ent.
In addition, s ometimes a technological
development will lower the cos t of market
trans actions while, at the s ame time, it
expands the role of the firm. When the
"putting out" s ys tem was us ed for weav-
ing, inputs were organized largely through
market negotiations . With the develop-
ment of efficient central s ources of power,
it became economical to perform weaving
in proximity to the power s ource and to
engage in team production. The bringing
in of weavers s urely mus t have res ulted in
a reduction in the cos t of negotiating
(forming) contracts . Yet, what we ob-
s erve is the beginning of the factory s ys -
tem in which inputs are organized within
a firm. Why? The weavers did not s imply
move to a common s ource of power that
they could tap like an electric line, pur-
chas ing power while they us ed their own
equipment. Now team production in the
joint us e of equipment became more im-
portant. The meas urement of marginal
productivity, which now involved interac-
tions between workers , es pecially through
their joint us e of machines , became more
difficult though contract negotiating cos t
was reduced, while managing the behavior
of inputs became eas ier becaus e of the in-
creas ed centralization of activity. The
firm as an organization expanded even
though the cos t of trans actions was re-
duced by the advent of centralized power.
The s ame could be s aid for modern as -
s embly lines . Hence the emergence of
central power s ources expanded the s cope
of productive activity in which the firm
enjoyed a comparative advantage as an
organizational form.
Some economis ts , following Knight,
have identified the bearing of ris ks of
wealth changes with the director or central
employer without explaining why that is
a viable arrangement. Pres umably, the
more ris k-avers e inputs become employees
rather than owners of the clas s ical firm.
Ris k avers enes s and uncertainty with re-
gard to the firm's fortunes have little, if
anything, to do with our explanation al-
though it helps to explain why all re-
s ources in a team are not owned by one
pers on. That is , the role of ris k taken in
the s ens e of abs orbing the windfalls that
buffet the firm becaus e of unfores een com-
petition, technological change, or fluc-
tuations in demand are not central to our
theory, although it is true that imperfect
knowledge and, therefore, ris k, in this
s ens e of ris k, underlie the problem of
ALC HIAN AND DEMSETZ: EC ONOMIC ORGANIZATION 785
monitoring team behavior. We deduce the
s ys tem of paying the manager with a
res idual claim (the equity) from the des ire
to have efficient means to reduce s hirking
s o as to make team production economical
and not from the s maller avers ion to the
ris ks of enterpris e in a dynamic economy.
We conjecture that "dis tribution-of-ris k"
is not a valid rationale for the exis tence
and organization of the clas s ical firm.
Although we have emphas ized team
production as creating a cos tly metering
tas k and have treated team production as
an es s ential (neces s ary?) condition for the
firm, would not other obs tacles to cheap
metering als o call forth the s ame kind of
contractual arrangement here denoted as
a firm? For example, s uppos e a farmer
produces wheat in an eas ily as certained
quantity but with s ubtle and difficult to
detect quality variations determined by
how the farmer grew the wheat. A vertical
integration could allow a purchas er to
control the farmer's behavior in order to
more economically es timate productivity.
But this is not a cas e of joint or team
production, unles s "information" can be
cons idered part of the product. (While a
good cas e could be made for that broader
conception of production, we s hall ignore
it here.) Ins tead of forming a firm, a buyer
can contract to have his ins pector on the
s ite of production, jus t as home builders
contract with architects to s upervis e build-
ing contracts ; that arrangement is not a
firm. Still, a firm might be organized in
the production of many products wherein
no team production or jointnes s of us e of
s eparately owned res ources is involved.
This pos s ibility rather clearly indicates
a broader, or complementary, approach
to that which we have chos en. 1) As we do
in this paper, it can be argued that the
firm is the particular policing device
utilized when joint team production is
pres ent. If other s ources of high policing
cos ts aris e, as in the wheat cas e jus t in-
dicated, s ome other form of contractual ar-
rangement will be us ed. Thus to each
s ource of informational cos t there may be
a different type of policing and contractual
arrangement. 2) On the other hand, one
can s ay that where policing is difficult
acros s markets , various forms of contrac-
tual arrangements are devis ed, but there is
no reas on for that known as the firm to be
uniquely related or even highly correlated
with team production, as defined here. It
might be us ed equally probably and viably
for other s ources of high policing cos t. We
have not intens ively analyzed other
s ources , and we can only note that our
current and readily revis able conjecture
is that 1) is valid, and has motivated us in
our current endeavor. In any event, the
tes t of the theory advanced here is to s ee
whether the conditions we have identified
are neces s ary for firms to have long-run
viability rather than merely births with
high infant mortality. C onglomerate firms
or collections of s eparate production agen-
cies into one owning organization can be in-
terpreted as an inves tment trus t or in-
ves tment divers ification device-prob-
ably along the lines that motivated
Knight's interpretation. A holding com-
pany can be called a firm, becaus e of the
common as s ociation of the word firm with
any owners hip unit that owns income
s ources . The term firm as commonly us ed
is s o turgid of meaning that we can not
hope to explain every entity to which the
name is attached in common or even tech-
nical literature. Ins tead, we s eek to iden-
tify and explain a particular contractual
arrangement induced by the cos t of in-
formation factors analyzed in this paper.
IV. Types of Firms
A. Profit-Sharing
Firms
Explicit in our explanation of the
capitalis t firm is the as s umption that the
cos t of managing the team's inputs by a
central monitor, who dis ciplines hims elf
becaus e he is a res idual claimant, is low
786 THE AMERIC AN EC ONOMIC REVIEW
relative to the cos t of metering the mar-
ginal outputs of team members .
If we look within a firm to s ee who
monitors -hires , fires , changes , promotes ,
and renegotiates -we s hould find him be-
ing a res idual claimant or, at leas t, one
whos e pay or reward is more than any
others correlated with fluctuations in the
res idual value of the firm. They more
likely will have options or rights or bonus es
than will inputs with other tas ks .
An implicit "auxiliary" as s umption of
our explanation of the firm is that the
cos t of team production is increas ed if the
res idual claim is not held entirely by the
central monitor. That is , we as s ume that
if profit s haring had to be relied upon for
all team members , los s es from the res ult-
ing increas e in central monitor s hirking
would exceed the output gains from the
increas ed incentives of other team mem-
bers not to s hirk. If the optimal team s ize
is only two owners of inputs , then an
equal divis ion of profits and los s es be-
tween them will leave each with s tronger
incentives to reduce s hirking than if the
optimal team s ize is large, for in the latter
cas e only a s maller percentage of the los s es
occas ioned by the s hirker will be borne by
him. Incentives to s hirk are pos itively re-
lated to the optimal s ize of the team under
an equal profit-s haring s cheme.10
The preceding does not imply that profit
s haring is never viable. Profit s haring to
encourage s elf-policing is more appropriate
for s mall teams . And, indeed, where input
owners are free to make whatever con-
tractual arrangements s uit them, as gen-
erally is true in capitalis t economies , profit
s haring s eems largely limited to partner-
s hips with a relatively s mall number of
active"1 partners . Another advantage of
s uch arrangements for s maller teams is
that it permits more effective reciprocal
monitoring among inputs . Monitoring
need not be entirely s pecialized.
Profit s haring is more viable if s mall
team s ize is as s ociated with s ituations
where the cos t of s pecialized management
of inputs is large relative to the increas ed
productivity potential in team effort. We
conjecture that the cos t of managing team
inputs increas es if the productivity of a
team member is difficult to correlate with
his behavior. In "artis tic" or "profes -
s ional" work, watching a man's activities
is not a good clue to what he is actually
thinking or doing with his mind. While it
is relatively eas y to manage or direct the
loading of trucks by a team of dock
workers where input activity is s o highly
related in an obvious way to output, it is
more difficult to manage and direct a
lawyer in the preparation and pres enta-
tion of a cas e. Dock workers can be di-
rected in detail without the monitor him-
s elf loading the truck, and as s embly line
workers can be monitored by varying the
s peed of the as s embly line, but detailed
direction in the preparation of a law cas e
would require in much greater degree that
the monitor prepare the cas e hims elf. As
a res ult, artis tic or profes s ional inputs ,
s uch as lawyers , advertis ing s pecialis ts ,
and doctors , will be given relatively freer
reign with regard to individual behavior.
If the management of inputs is relatively
cos tly, or ineffective, as it would s eem to
be in thes e cas es , but, nonetheles s if team
effort is more productive than s eparable
production with exchange acros s markets ,
then there will develop a tendency to us e
profit-s haring s chemes to provide incen-
tives to avoid s hirking.
12
10
While the degree to which res idual claims are cen-
tralized will affect the s ize of the team, this will be only
one of manv factors that determine team s ize, s o as an
approximation, we can treat team s ize as exogenous ly
determined. Under certain as s umptions about the
s hape of the "typical" utility function, the incentive to
avoid s hirking with unequal profit-s haring
can be mea-
s ured by the Herfindahl index.
11 The us e of the word active will be clarified in our
dis cus s ion of the corporation, which follows below.
12
Some s haring contracts , like crop s haring, or rental
ALC HIAN AND DEMSETZ: EC ONOMIC ORGANIZATION 787
B. Socialis t Firms
We have analyzed the clas s ical propri-
etors hip and the profit-s haring firms in the
context of free as s ociation and choice of
economic organization. Such organizations
need not be the mos t viable when political
cons traints limit the forms of organization
that can be chos en. It is one thing to have
profit s haring when profes s ional or artis tic
talents are us ed by s mall teams . But if
political or tax or s ubs idy cons iderations
induce profit-s haring techniques when
thes e are not otherwis e economically
jus tified, then additional management
techniques will be developed to help re-
duce the degree of s hirking.
For example, mos t, if not all, firms in
Jugos lavia are owned by the employees in
the res tricted s ens e that all s hare in the
res idual. This is true for large firms and
for firms which employ nonartis tic, or
nonprofes s ional, workers as well. With a
decay of political cons traints , mos t of
thes e firms could be expected to rely on
paid wages rather than s hares in the res id-
ual. This res ts on our auxiliary as s ump-
tion that general s haring in the res idual
res ults in los s es from enhanced s hirking
by the monitor that exceed the gains from
reduced s hirking by res idual-s haring em-
ployees . If this were not s o, profit s haring
with employees s hould have occurred more
frequently in Wes tern s ocieties where s uch
organizations are neither banned nor
preferred politically. Where res idual s har-
ing by employees is politically impos ed,
as in Jugos lavia, we are led to expect that
s ome management technique will aris e to
reduce the s hirking by the central monitor,
a technique that will not be found fre-
quently in Wes tern s ocieties s ince the
monitor retains all (or much) of the re-
s idual in the Wes t and profit s haring is
largely confined to s mall, profes s ional-
artis tic team production s ituations . We do
find in the larger s cale res idual-s haring
firms in Jugos lavia that there are em-
ployee committees that can recommend
(to the s tate) the termination of a man-
ager's contract (veto his continuance)
with the enterpris e. We conjecture that
the workers ' committee is given the right
to recommend the termination of the
manager's contract precis ely becaus e the
general s haring of the res idual increas es
''exces s ively" the manager's incentive to
s hirk. 13
C . The C orporation
All firms mus t initially acquire com-
mand over s ome res ources . The corpora-
tion does s o primarily by s elling promis es
of future returns to thos e who (as creditors
or owners ) provide financial capital. In
s ome s ituations res ources can be acquired
in advance from cons umers by promis es
of future delivery (for example, advance
s ale of a propos ed book). Or where the
firm is a few artis tic or profes s ional per-
s ons , each can "chip in" with time and
talent until the s ale of s ervices brings in
revenues . For the mos t part, capital can
be acquired more cheaply if many (ris k-
avers e) inves tors contribute s mall por-
tions to a large inves tment. The economies
of rais ing large s ums of equity capital in
this way s ugges t that modifications in the
relations hip among corporate inputs are
required to cope with the s hirking problem
payments bas ed on gros s s ales in retail s tores , come
clos e to profit s haring. However, it is gros s output s har-
ing rather than profit s haring. We are unable to
s pecify
the implications of the difference. We refer the reader to
S. N. C heung.
13 Incidentally, inves tment activity will be changed.
The inability to capitalize the inves tment value as
"take-home" proviate property wealth of the members of
the firm means that the benefits of the inves tment mus t
be taken as annual income by thos e who are employed
at the time of the income. Inves tment will be confined
more to thos e with s horter life and with higher rates or
pay-offs
if the alternative of
inves ting
is
paying
out the
firm's income to its employees to take home and us e as
private property. For a
development
of this
propos i-
tion, s ee the papeis by Eirik Furobotn and Svetozar
Pejovich, and by Pejovich.
788 THE AMERIC AN EC ONOMIC REVIEW
that aris es with profit s haring among large
numbers of corporate s tockholders . One
modification is limited liability, es pecially
for firms that are large relative to a s tock-
holder's wealth. It s erves to protect s tock-
holders from large los s es no matter how
they are caus ed.
If every s tock owner participated in
each decis ion in a corporation, not only
would large bureaucratic cos ts be in-
curred, but many would s hirk the tas k of
becoming well informed on the is s ue to be
decided, s ince the los s es as s ociated with
unexpectedly bad decis ions will be borne
in large part by the many other corporate
s hareholders . More effective control of
corporate activity is achieved for mos t
purpos es by trans ferring decis ion author-
ity to a s maller group, whos e main function
is to negotiate with and manage (renegotiate
with) the other inputs of the team. The
corporate s tockholders retain the authority
to revis e the members hip of the manage-
ment group and over major decis ions that
affect the s tructure of the corporation or
its dis s olution.
As a res ult a new modification of part-
ners hips is induced-the right to s ale of
corporate s hares without approval of any
other s tockholders . Any s hareholder can
remove his wealth from control by thos e
with whom he has differences of opinion.
Rather than try to control the decis ions
of the management, which is harder to do
with many s tockholders than with only a
few, unres tricted s alability provides a
more acceptable es cape to each s tock-
holder from continued policies with which
he dis agrees .
Indeed, the policing of managerial
s hirking relies on acros s -market competi-
tion from new groups of would-be man-
agers as well as competition from members
within the firm who s eek to dis place exis t-
ing management. In addition to competi-
tion from outs ide and ins ide managers ,
control is facilitated by the temporary
congealing of s hare votes into voting blocs
owned by one or a few contenders . Proxy
battles or s tock-purchas es concentrate the
votes required to dis place the exis ting
management or modify managerial policies .
But it is more than a change in policy that
is s ought by the newly formed financial
interes ts , whether of new s tockholders or
not. It is the capitalization of expected
future benefits into s tock prices that con-
centrates on the innovators the wealth
gains of their actions if they own large
numbers of s hares . Without capitalization
of future benefits , there would be les s in-
centive to incur the cos ts required to exert
informed decis ive influence on the corpo-
ration's policies and managing pers onnel.
Temporarily, the s tructure of owners hip is
reformed, moving away from diffus ed
owners hip into decis ive power blocs , and
this is a trans ient res urgence of the clas -
s ical firm with power again concentrated
in thos e who have title to the res idual.
In as s es s ing the s ignificance of s tock-
holders ' power it is not the us ual diffus ion
of voting power that is s ignificant but in-
s tead the frequency with which voting
congeals into decis ive changes . Even a
one-man owned company may have a
long term with jus t one manager-con-
tinuous ly being approved by the owner.
Similarly a dis pers ed voting power corpo-
ration may be als o characterized by a
long-lived management. The ques tion is
the probability of replacement of the
management if it behaves in ways not ac-
ceptable to a majority of the s tockholders .
The unres tricted s alability of s tock and
the trans fer of proxies enhances the prob-
ability of decis ive action in the event cur-
rent s tockholders or any outs ider believes
that management is not doing a good job
with the corporation. We are not compar-
ing the corporate res pons ivenes s to that
of a s ingle proprietors hip; ins tead, we are
indicating features of the corporate s truc-
ture that are induced by the problem of
ALC HIAN AND DEMSETZ: EC ONOMIC ORGANIZATION 789
delegated authority to manager-moni-
tors . 14
D. Mutual and Nonprofit Firms
The benefits obtained by the new man-
agement are greater if the s tock can be
purchas ed and s old, becaus e this enables
capitalization of anticipated future im-
provements into pres ent wealth of new
managers who bought s tock and created
a larger capital by their management
changes . But in nonprofit corporations ,
colleges , churches , country clubs , mutual
s avings banks , mutual ins urance com-
panies , and "coops ," the future cons e-
quences of improved management are not
14
Ins tead of thinking of s hareholders as joint owners ,
we can think of them as inves tors , like bondholders ,
except that the s tockholders are more optimis tic than
bondholders about the enterpris e pros pects . Ins tead of
buying bonds in the corporation, thus enjoying s maller
ris ks , s hareholders prefer to inves t funds with a greater
realizable return if the firm pros pers as expected, but
with s maller (pos s ibly negative) returns if the firm per-
forms in a manner clos er to that expected by the more
pes s imis tic inves tors . The pes s imis tic inves tors , in
turn, regard only the bonds as likely to pay off.
If the entrepreneur-organizer is to rais e capital on the
bes t terms to him, it is to his advantage, as well as that
of pros pective inves tors , to recognize thes e differences in
expectations . The res idual claim on earnings enjoyed by
s hareholders does not s erve the function of enhancing
their efficiency as monitors in the general s ituation. The
s tockholders are "merely" the les s ris k-avers e or the
more optimis tic member of the group that finances the
firm. Being more optimis tic than the average and s eeing
a higher mean value future return, they are willing to
pay more for a certificate that allows them to realize
gain on their expectations . One method of doing s o is to
buy claims to the dis tribution of returns that "they s ee"
while bondholders , who are more pes s imis tic, purchas e a
claim to the dis tribution that they s ee as more likely to
emerge. Stockholders are then comparable to warrant
holders . They care not about the voting rights (us ually
not attached to warrants ); they are in the s ame pos ition
in s o far as voting rights are concerned as are bond-
holders . The only difference is in the probability dis tri-
bujtion of rewards and the terms on which they can
place their bets .
If we treat bondholders , preferred and convertible
preferred s tockholders , and common s tockholders and
warrant holders as s imply different clas s es of inves tors -
differing not only in their ris k avers enes s but in their
beliefs about the probability dis tribution of the firm's
future earnings , why s hould s tockholders be regarded as
"owners " in any s ens e dis tinct from the other financial
inves tors ? The entrepreneur-organizer, who let us
as s ume is the chief operating officer and s ole repos itory
of control of the corporation, does not find his authority
res iding in common s tockholders (except in the cas e of a
take over). Does this type of control make any differ-
ence in the way the firm is conducted? Would it make
any difference in the kinds of behavior that would be
tolerated by competing managers and inves tors (and we
here deliberately refrain from thinking of them as
owner-s tockholders in the traditional s ens e)?
Inves tment old timers recall a s ignificant incidence of
nonvoting common s tock, now prohibited in corpora-
tions whos e s tock is traded on lis ted exchanges . (Why
prohibited?) The entrepreneur in thos e days could hold
voting s hares while inves tors held nonvoting s hares ,
which in every other res pect were identical. Nonvoting
s hare holders were s imply inves tors devoid of owners hip
connotations . The control and behavior of ins ide owners
in s uch corporations has never, s o far as we have as cer-
tained, been carefully s tudied. For example, at the
s imples t level of interes t, does the evidence indicate that
nonvoting s hareholders fared any wors e becaus e of not
having voting rights ? Did owners permit the nonvoting
holders the normal return available to voting s hare-
holders ? Though evidence is prohibitively expens ive to
obtain, it is remarkable that voting and nonvoting
s hares s old for es s entially identical prices , even during
s ome proxy battles . However, our cas ual evidence de-
s erves no more than interes t-initiating weight.
One more point. The facade is deceptive. Ins tead of
nonvoting s hares , today we have warrants , convertible
preferred s tocks all of which are s olely or partly "equity"
claims without voting rights , though they could be con-
verted into voting s hares .
In s um, is it the cas e that the s tockholder-inves tor
relations hip is one emanating from the divis ion of
owners hlip among s everal people, or is it that the collec-
tion of inves tment funds from people of varying antici-
pations is the underlying factor? If the latter, why
s hould any of them be thought of as the owners in
whom voting rights , whatever they may s ignify or how-
ever exercis able, s hould res ide in order to enhance effi-
ciency? Why voting rights in any of the outs ide, par-
ticipating inves tors ?
Our initial perception of this pos s ibly s ignificant dif-
ference in interpretation was precipitated by Henry
Manne. A reading of his paper makes it clear that it is
hard to unders tand why an inves tor who wis hes to back
and "s hare" in the cons equences of s ome new bus ines s
s hould neces s arily have to acquire voting power (i.e.,
power to change the manager-operator) in order to
inves t in the venture. In fact, we inves t in s ome ven-
tures in the hope that no other s tockholders will be s o
"foolis h" as to trv to tos s out the incumbent manage-
ment. We want him to have the power to s tay in office,
and for the pros pect
of s haring in his fortunes we buy
nonvoting common s tock. Our willingnes s to inves t is
enhanced by the knowledge that we can act legally via
fraud, embezzlement and other laws to help as s ure that
we outs ide inves tors will not be "milked" beyond our
initial dis counted anticipations .
790 THE AMERIC AN EC ONOMIC REVIEW
capitalized into pres ent wealth of s tock-
holders . (As if to make more difficult that
competition by new would-be monitors ,
mutiple s hares of owners hip in thos e en-
terpris es cannot be bought by one pers on.)
One s hould, therefore, find greater s hirk-
ing in nonprofit, mutually owned enter-
pris es . (This s ugges ts that nonprofit en-
terpris es are es pecially appropriate in
realms of endeavor where more s hirking is
des ired and where redirected us es of the
enterpris e in res pons e to market-revealed
values is les s des ired.)
E. Partners hips
Team production in artis tic or profes -
s ional intellectual s kills will more likely
be by partners hips than other types of
team production. This amounts to market-
organized team activity and to a non-
employer s tatus . Self-monitoring partner-
s hips , therefore, will be us ed rather than
employer-employee contracts , and thes e
organizations will be s mall to prevent an
exces s ive dilution of efforts through s hirk-
ing. Als o, partners hips are more likely to
occur among relatives or long-s tanding
acquaintances , not neces s arily becaus e
they s hare a common utility function, but
als o becaus e each knows better the other's
work characteris tics and tendencies to
s hirk.
F. Employee Unions
Employee unions , whatever els e they
do, perform as monitors for employees .
Employers monitor employees and s imi-
larly employees monitor an employer's per-
formance. Are correct wages paid on time
and in good currency? Us ually, this is
extremely eas y to check. But s ome forms
of employer performance are les s eas y to
meter and are more s ubject to employer
s hirking. Fringe benefits often are in non-
pecuniary, contingent form; medical, hos -
pital, and accident ins urance, and retire-
ment pens ions are contingent payments
or performances partly in kind by em-
ployers to employees . Each employee
cannot judge the character of s uch pay-
ments as eas ily as money wages . Ins ur-
ance is a contingent payment-what the
employee will get upon the contingent
event may come as a dis appointment. If
he could eas ily determine what other
employees had gotten upon s uch con-
tingent events he could judge more ac-
curately the performance by the employer.
He could "trus t" the employer not to
s hirk in s uch fringe contingent payments ,
but he would prefer an effective and eco-
nomic monitor of thos e payments . We s ee
a s pecialis t monitor-the union employees '
agent-hired by them and monitoring
thos e as pects of employer payment mos t
difficult for the employees to monitor. Em-
ployees s hould be willing to employ a
s pecialis t monitor to adminis ter s uch
hard-to-detect employer performance,
even though their monitor has incentives
to us e pens ion and retirement funds not
entirely for the benefit of employees .
V. Team Spirit and Loyalty
Every team member would prefer a
team in which no one, not even hims elf,
s hirked. Then the true marginal cos ts and
values could be equated to achieve more
preferred pos itions . If one could enhance
a common interes t in nons hirking in the
guis e of a team loyalty or team s pirit, the
team would be more efficient. In thos e
s ports where team activity is mos t clearly
exemplified, the s ens e of loyalty and team
s pirit is mos t s trongly urged. Obvious ly
the team is better, with team s pirit and
loyalty, becaus e of the reduced s hirking-
not becaus e of s ome other feature in-
herent in loyalty or s pirit as s uch.'5
15
Sports Leagues : Profes s ional s ports contes ts among
teams is typically conducted by a league of teams . We
as s ume that s ports cons umers are interes ted not only in
abs olute s porting s kill but als o in s kills relative to other
teams . Being s lightly better than oppos ing teams en-
ables one to claim a major portion of the receipts ; the
ALC HIAN AND DEMSETZ: EC ONOMIC ORGANIZATION 791
C orporations and bus ines s firms try to
ins till a s pirit of loyalty. This s hould not
be viewed s imply as a device to increas e
profits by over-working or mis leading the
employees , nor as an adoles cent urge for
belonging. It promotes a clos er approxima-
tion to the employees ' potentially avail-
able true rates of s ubs titution between
production and leis ure and enables each
team member to achieve a more preferred
s ituation. The difficulty, of cours e, is to
create economically that team s pirit and
loyalty. It can be preached with an aura
of moral code of conduct a morality with
literally the s ame bas is as the ten com-
mandments to res trict our conduct to-
ward what we would choos e if we bore our
full cos ts .
VI. Kinds of Inputs Owned
by the Firm
To this point the dis cus s ion has ex-
amined why firms , as we have defined
them, exis t? That is , why is there an
owner-employer who is the common
party to contracts with other owners of
inputs in team activity? The ans wer to
that ques tion s hould als o indicate the kind
of the jointly us ed res ources likely to be
owned by the central-owner-monitor and
the kind likely to be hired from people
who are not team-owners . C an we identify
characteris tics or features of various in-
puts that lead to their being hired or to
their being owned by the firm?
How can res idual-claimant, central-
employer-owner demons trate ability to
pay the other hired inputs the promis ed
amount in the event of a los s ? He can pay
in advance or he can commit wealth s uf-
ficient to cover negative res iduals . The
latter will take the form of machines , land,
buildings , or raw materials committed to
the firm. C ommitments of labor-wealth
(i.e., human wealth) given the property
rights in people, is les s feas ible. Thes e con-
s iderations s ugges t that res idual claim-
ants owners of the firm will be inves -
tors of res alable capital equipment in the
firm. The goods or inputs more likely to
be inves ted, than rented, by the owners
of the enterpris e, will have higher res ale
values relative to the initial cos t and will
have longer expected us e in a firm relative
to the economic life of the good.
But beyond thes e factors are thos e de-
veloped above to explain the exis tence of
inferior team does not releas e res ources and reduce
cos ts ,
s ince they were expected in the play of contes t. Hence,
abs olute s kill is developed beyond the equality of margi-
nal inves tment in s porting s kill with its true s ocial
marginal value product. It follows there will be a ten-
dency to overinves t in training athletes and developing
teams . "Revers e s hirking" aris es , as budding players
are induced to overpractice hyperactively relative to the
s ocial marginal value of their enhanced s kills . To pre-
vent overinves tment, the teams s eek an agreement with
each other to res trict practice, s ize of teams , and even
pay of the team members (which reduces incentives of
young people to overinves t in developing s kills ). Ideally,
if all the contes tant teams were owned by one owner,
overinves tment in s ports would be avoided, much as
owners hip of common fis heries or underground oil or
water res erve would prevent overinves tment. This
hyperactivity (to s ugges t the oppos ite of s hirking) is
controlled by the league of teams , wherein the league
adopts a common s et of cons traints on each team's be-
havior. In effect, the teams are no longer really owned
by the team owners but are s upervis ed by them, much
as the franchis ers of s ome product. They are not full-
fledged owners of their bus ines s , including the brand
name, and can not "do what thev wis h" as franchis es .
C omparable to the franchis er, is the league commis -
s ioner or conference pres ident, who s eeks to res train
hyperactivity, as individual team s upervis ors compete
with each other and caus e external dis economies . Such
res traints are us ually regarded as anticompetitive, anti-
s ocial, collus ive-cartel devices to res train free open com-
petition, and reduce players ' s alaries . However, the
interpretation pres ented here is premis ed on an attempt
to avoid hyperinves tment in team s ports production.
Of cours e, the team operators have an incentive, once
the league is formed and res traints are placed on hyper-
inves tment activity, to go further and obtain the private
benefits of monopoly res triction. To what extent over-
inves tment is replaced by monopoly res triction is not
yet determinable; nor have we s een an empirical tes t of
thes e two competing, but mutually cons is tent interpre-
tations . (This interpretation of league-s ports activity
was propos ed by Earl Thomps on and formulated by
Michael C anes .) Again, athletic teams clearly exemplify
the s pecialization of monitoring with captains and
coaches ; a captain detects s hirkers while the coach trains
and s elects s trategies and tactics . Both functions may
be centralized in one pers on.
792 THE AMERIC AN EC ONOMIC REVIEW
the ins titution known as the firm the
cos ts of detecting output performance.
When a durable res ource is us ed it will
have a marginal product and a deprecia-
tion. Its us e requires payment to cover at
leas t us e-induced depreciation; unles s that
us er cos t is s pecifically detectable, pay-
ment for it will be demanded in accord
with expected depreciation. And we can
as certain circums tances for each. An in-
des tructible hammer with a readily de-
tectable marginal product has zero us er
cos t. But s uppos e the hammer were de-
s tructible and that careles s (which is
eas ier than careful) us e is more abus ive
and caus es greater depreciation of the
hammer. Suppos e in addition the abus e
is eas ier to detect by obs erving the way it
is us ed than by obs erving only the ham-
mer after its us e, or by meas uring the
output s cored from a hammer by a
laborer. If the hammer were rented and
us ed in the abs ence of the owner, the de-
preciation would be greater than if the us e
were obs erved by the owner and the us er
charged in accord with the impos ed de-
preciation. (C areles s us e is more likely
than careful us e-if one does not pay for
the greater depreciation.) An abs entee
owner would therefore as k for a higher
rental price becaus e of the higher expected
us er cos t than if the item were us ed by the
owner. The expectation is higher becaus e
of the greater difficulty of obs erving
s pecific us er cos t, by ins pection of the
hammer after us e. Renting is therefore in
this cas e more cos tly than owner us e. This
is the valid content of the mis leading ex-
pres s ions about owners hip being more
economical than renting ignoring all
other factors that may work in the oppo-
s ite direction, like tax provis ion, s hort-
term occupancy and capital ris k avoid-
ance.
Better examples are tools of the trade.
Watch repairers , engineers , and carpenters
tend to own their own tools es pecially if
they are portable. Trucks are more likely
to be employee owned rather than other
equally expens ive team inputs becaus e it
is relatively cheap for the driver to police
the care taken in us ing a truck. Policing
the us e of trucks by a nondriver owner is
more likely to occur for trucks that are
not s pecialized to one driver, like public
trans it bus s es .
The factor with which we are concerned
here is one related to the cos ts of monitor-
ing not only the gros s product performance
of an input but als o the abus e or deprecia-
tion inflicted on the input in the cours e of
its us e. If depreciation or us er cos t is more
cheaply detected when the owner can s ee
its us e than by only s eeing the input be-
fore and after, there is a force toward
owner us e rather than renting. Res ources
whos e us er cos t is harder to detect when
us ed by s omeone els e, tend on this count
to be owner-us ed. Abs entee owners hip, in
the lay language, will be les s likely. As -
s ume momentarily that labor s ervice can-
not be performed in the abs ence of its
owner. The labor owner can more cheaply
monitor any abus e of hims elf than if s ome-
how labor-s ervices could be provided with-
out the labor owner obs erving its mode of
us e or knowing what was happening. Als o
his incentive to abus e hims elf is increas ed
if he does not own hims elf."6
16
Profes s ional athletes in bas eball, football, and bas -
ketball, where athletes having s old their s ource of
s ervice to the team owners upon entering into s ports
activity, are owned by team owners . Here the team
owners mus t monitor the athletes ' phys ical condition
and behavior to protect the team owners ' wealth. The
athlete has les s (not, no) incentive to protect or enhance
his athletic prowes s s ince capital value changes have les s
impact on his own wealth and more on the team owners .
Thus , s ome athletes s ign up for big initial bonus es
(repres enting pres ent capital value of future s ervices ).
Future s alaries are lower by the annuity value of the
prepaid "bonus " and hence the athlete has les s to los e by
s ubs equent abus e of his athletic prowes s . Any decline in
his s ubs equent s ervice value would in part be borne by
the team owner who owns the players ' future s ervice.
This does not s ay thes e los s es of future s alaries have no
effect on pres ervation of athletic talent (we are not mak-
ing a "s unk cos t" error). Ins tead, we as s ert that the
ALC HIAN AND DEMSETZ: EC ONOMIC ORGANIZATION 793
The s imilarity between the preceding
analys is and the ques tion of abs entee
landlordis m and of s harecropping ar-
rangements is no accident. The s ame fac-
tors which explain the contractual ar-
rangements known as a firm help to explain
the incidence of tenancy, labor hiring or
s harecropping. 17
VII. Firms as a Specialized Market
Ins titution for C ollecting, C ollating,
and Selling Input Information
The firm s erves as a highly s pecialized
s urrogate market. Any pers on contem-
plating a joint-input activity mus t s earch
and detect the qualities of available joint
inputs . He could contact an employment
agency, but that agency in a s mall town
would have little advantage over a large
firm with many inputs . The employer, by
virtue of monitoring many inputs , ac-
quires s pecial s uperior information about
their productive talents . This aids his
directive (i.e., market hiring) efficiency. He
"s ells " his information to employee-inputs
as he aids them in as certaining good input
combinations for team activity. Thos e
who work as employees or who rent s er-
vices to him are us ing him to dis cern s u-
perior combinations of inputs . Not only
does the director-employer "decide" what
each input will produce, he als o es timates
which heterogeneous inputs will work to-
gether jointly more efficiently, and he
does this in the context of a privately
owned market for forming teams . The de-
partment s tore is a firm and is a s uperior
private market. People who s hop and
work in one town can as well s hop and
work in a privately owned firm.
This marketing function is obs cured in
the theoretical literature by the as s ump-
tion of homogeneous factors . Or it is
tacitly left for individuals to do thems elves
via pers onal market s earch, much as if a
pers on had to s earch without benefit of
s pecialis t retailers . Whether or not the
firm aros e becaus e of this efficient in-
formation s ervice, it gives the director-
employer more knowledge about the
productive talents of the team's inputs ,
and a bas is for s uperior decis ions about
efficient or profitable combinations of
thos e heterogeneous res ources .
In other words , opportunities for profit-
able team production by inputs already
within the firm may be as certained more
economically and accurately than for re-
s ources outs ide the firm. Superior com-
binations of inputs can be more econom-
ically identified and formed from res ources
already us ed in the organization than by
obtaining new res ources (and knowledge
of them) from the outs ide. Promotion and
revis ioni of employee as s ignments (con-
tracts ) will be preferred by a firm to the
hiring of new inputs . To the extent that
this occurs there is reas on to expect the
firm to be able to operate as a conglom-
erate rather than pers is t in producing
a s ingle product. Efficient production
with heterogeneous res ources is a res ult
not of having better res ources but in know-
ing more accurately the relative productive
performances of thos e res ources . Poorer
res ources can be paid les s in accord with
their inferiority; greater accuracy of
pres ervation is reduced, not eliminated, becaus e the
amount of los s of wealth s uffered is s maller. The athlete
will s pend les s to maintain or enhance his prowes s
thereafter. The effect of this revis ed incentive s ys tem is
evidenced in comparis ons of the kinds of attention and
care impos ed on the athletes at the "expens e of the
team owner" in the cas e where atheletes ' future s ervies
are owned by the team owner with that where future
labor s ervice values are owned by the athlete hims elf.
Why athletes ' future athletic s ervices are owned by the
team owners rather than being hired is a ques tion we
s hould be able to ans wer. One pres umption is carteliza-
tion and monops ony gains to team owners . Another is
exactly the theory being expounded in this paper-cos ts
of monitoring production of athletes ; we know not on
which to rely.
17
The analys is us ed by C heung in explaining the
prevalence of s harecropping and land tenancy arrange-
ments is built s quarely on the s ame factors -the cos ts
of detecting output performance of jointly us ed inputs
in team production and the cos ts of detecting us er cos ts
impos ed on the various inputs if owner us ed or if rented.
794 THE AMERIC AN EC ONOMIC REVIEW
knowledge of the potential and actual pro-
ductive actions of inputs rather than hav-
ing high productivity res ources makes a
firm (or an as s ignment of inputs ) profit-
able.18
VIII. Summary
While ordinary contracts facilitate ef-
ficient s pecialization according to com-
parative advantage, a s pecial clas s of con-
tracts among a group of joint inputs to a
team production proces s is commonly
us ed for team production. Ins tead of
multilateral contracts among all the joint
inputs ' owners , a central common party
to a s et of bilateral contracts facilitates
efficient organization of the joint inguts in
team production. The terms of the con-
tracts form the bas is of the entity called
the firm-es pecially appropriate for or-
ganizing team production proces s es .
Team productive activity is that in
which a union, or joint us e, of inputs yields
a larger output than the s um of the prod-
ucts of the s eparately us ed inputs . This
team production requires -like all other
production proces s es an as s es s ment of
marginal productivities if efficient pro-
duction is to be achieved. Nons eparability
of the products of s everal differently
owned joint inputs rais es the cos t of as -
s es s ing the marginal productivities of
thos e res ources or s ervices of each input
owner. Monitoring or metering the pro-
ductivities to match marginal produc-
tivities to cos ts of inputs and thereby to
reduce s hirking can be achieved more
economically (than by acros s market bi-
lateral negotiations among inputs ) in a
firm.
The es s ence of the clas s ical firm is
identified here as a contractual s tructure
with: 1) joint input production; 2) s everal
input owners ; 3) one party who is com-
mon to all the contracts of the joint in-
puts ; 4) who has rights to renegotiate any
input's contract independently of con-
tracts with other input owners ; 5) who
holds the res idual claim; and 6) who has
the right to s ell his central contractual
res idual s tatus . The central agent is called
the firm's owner and the employer. No
authoritarian control is involved; the ar-
rangement is s imply a contractual s truc-
ture s ubject to continuous renegotiation
with the central agent. The contractual
s tructure aris es as a means of enhancing
efficient organization of team production.
In particular, the ability to detect s hirk-
ing among owners of jointly us ed inputs in
team production is enhanced (detection
cos ts are reduced) by this arrangement and
the dis cipline (by revis ion of contracts ) of
input owners is made more economic.
Tes table implications are s ugges ted by
the analys is of different types of organiza-
tions -nonprofit, proprietary for profit,
unions , cooperatives , partners hips , and by
the kinds of inputs that tend to be owned
by the firm in contras t to thos e employed
by the firm.
We conclude with a highly conjectural
18
According to our interpretation, the firm is a
s pecialized s urrogate for a market for team us e of inputs ;
it provides s uperior (i.e., cheaper) collection and colla-
tion of knowledge about heterogeneous res ources . The
greater the s et of inputs about which knowledge of per-
formance is being collated within a firm the greater are
the pres ent cos ts of the collation activity. Then, the
larger the firm (market) the greater the attenuation of
monitor control. To counter this force, the firm will be
divis ionalized in ways that economize on thos e cos ts -
jus t as will the market be s pecialized. So far as we can
as certain, other theories of the reas ons for firms have no
s uch implications .
In Japan, employees by cus tom work nearly their
entire lives with one firm, and the firm agrees to that
expectation. Firms will tend to be large and conglomer-
ate to enable a broader s cope of input revis ion. Each
firm is , in effect, a s mall economy engaging in "intra-
national and international" trade. Analogous ly, Amer-
icans expect to s pend their whole lives in the United
States , and the bigger the country, in terms of variety
of res ources , the eas ier it is to adjus t to changing tas tes
and circums tances . Japan, with its lifetime employees ,
s hould be characterized more by large, conglomerate
firms . Pres umably, at s ome s ize of the firm, s pecialized
knowledge about inputs becomes as expens ive to trans -
mit acros s divis ions of the firms as it does acros s markets
to other firms .
ALC HIAN AND DEMSETZ: EC ONOMIC ORGANIZATION 795
but pos s ibly s ignificant interpretation. As
a cons equence of the flow of information
to the central party (employer), the firm
takes on the characteris tic of an efficient
market in that information about the pro-
ductive characteris tics of a large s et of
s pecific inputs is now more cheaply avail-
able. Better recombinations or new us es of
res ources can be more efficiently as cer-
tained than by the conventional s earch
through the general market. In this s ens e
inputs compete with each other within and
via a firm rather than s olely acros s markets
as conventionally conceived. Emphas is on
interfirm competition obs cures intrafirm
competition among inputs . C onceiving
competition as the revelation and exchange
of knowledge or information about quali-
ties , potential us es of different inputs in
different potential applications indicates
that the firm is a device for enchancing
competition among s ets of input res ources
as well as a device for more efficiently re-
warding the inputs . In contras t to markets
and cities which can be viewed as publicly
or nonowned market places , the firm can
be cons idered a privately owned market;
if s o, we could cons ider the firm and the
ordinary market as competing types of
markets , competition between private
proprietary markets and public or com-
munal markets . C ould it be that the
market s uffers from the defects of com-
munal property rights in organizing and
influencing us es of valuable res ources ?
REFERENC ES
M. C anes , "A Model of a Sports League," un-
publis hed doctoral dis s ertation, UC LA 1970.
S. N. C heung, The Theory of Share Tenancy,
C hicago 1969.
R. H. C oas e, "The Nature of the Firm," Eco-
nomica, Nov. 1937, 4, 386-405; reprinted in
G. J. Stigler and K. Boulding, eds ., Readings
in Price Theory, Homewood 1952, 331-51.
E. Furobotn and S. Pejovich, "Property Rights
and the Behavior of the Firm in a Socialis t
State," Zeits chrift fiir Nationalokonomie,
1970,30,431-454.
F. H. Knight, Ris k, Uncertainty and Profit,
New York 1965.
S. Macaulay, "Non-C ontractual Relations in
Bus ines s : A Preliminary Study," Amer. So-
ciological Rev., 1968, 28, 55-69.
H. B. Malmgren, "Information, Expectations
and the Theory of the Firm," Quart J. Econ.,
Aug. 1961, 75, 399-421.
H. Manne, "Our Two C orporation Sys tems :
Law and Economics ," Virginia Law Rev.,
Mar. 1967, 53, No. 2, 259-84.
S. Pejovich, "The Firm, Monetary Policy and
Property Rights in a Planned Economy,"
Wes tern Econ. J., Sept. 1969, 7, 193-200.
M. Silver and R. Aus ter, "Entrepreneurs hip,
Profit, and the Limits on Firm Size," J. Bus .
Univ. C hicago, Apr. 1969, 42, 277-81.
E. A. Thomps on, "Nonpecuniary Rewards and
the Aggregate Production Function," Rev.
Econ. Statis t., Nov. 1970, 52, 395-404.