Sei sulla pagina 1di 19

CneprER Foun

Multiple Actors:
Confl,ict and Politics

hapte r 3looked atdecision making when partners do not

have consistent preferences or identities. As was suggested
there, inconsistencies lead to complicatiofls, and one common
instinct of theorists of multiple actor decision making is to see
the decision making problem as that of converting inconsistent
partnerships into teams by aligning preferences and identities.
That leads to concerns with contracts, incentives, selection, so-
cialization, and attention that seek to remove or reduce incon-
sistencies. Studies of collaboration befween two uneasy paft-
ners form the background for examining broader social systems
and enduring social institutions involving mutually inconsistent
decision makers.
This chapter shifts attention to those larger systems and con-
siders a set of ideas about decision making involvittg multiple
inconsistent actors that are less inclined to emph asize elirninat-
ing conflict in preferences or identities. It examines two classic
metaphors of decision making in the face of inconsistent prefer-
cnces and identities. The first metaphor pictures decision mak-
ing as based on a power struggle. It asks: Who gets what,,when,

140 A t,tuMtiR oN DricrstoN MAKTNc
Multiple Actors: Conflict und kilitics l4l
and how? The second metaphor pictures decision making as
porvasive design features of organizations, presumes and im-
coalition formation. It asks: How are partners found, how are
agreements negotiated and enforced? Finally, the chapter con- lx)ses differences in power.
The metaphor of power is a paft of ordinary langu zge, easily
siders some effects of attention and other factors on decision
understood by ordinary actors. ft is used in everyduy conversa-
t ions and in storytelling. It fills the professional literature and
When rnultiple actor decisions are considered in this w&y,
the self-help books of airport bookstores. The familiarity of
they are often called "polit ical" or "conflictual," not because
power and the ease with which the term is used in daily dis-
the process necessarily involves the institutions or practices of
course are great assets. Feople usually have no difficulty an-
government, nor because it is necessarily chara cterized by
vio- swering questions about who has power, and they often agree
lence or outbursts of emotionality, but because decision ,rrukr6
with each other. Discussions of power have a ring of reality to
sustain inconsistent preferences or identities. IVIany familiar
them. They sound right. At the same time, the familiarity of the
systems for collective decision making, including frel markets
iclea also makes it difficult to use in understanding decisions. It
and systems of governan ce, are politi cal in the sense that they
needs to be defined carefully and developed precisely. When
cteate mechanisms for decisions without agreement on either
that is done, power loses some of its charm as a general-pur-
preferences or identities. Market mechaniims use prices
and pose explanation of decision making.
coercively enforced contracts. Systems of demortátir gover-
nance use voting, constitutions, political parties, and ro.irively
enforced legislation.
The resulting decisions are sometimes confusing if consid- 4.7.7 Power as Getting What You Wqnt
ered from the perspective of a single, coherent deciiion maker
The basic idea of decision making in the face of inconsistency is
or a team. They sometimes seem to bring together people who
that different people want to have different things or to fulfill
share nothing beyond indifference toward each other's wishes.
different identities, and not everyone can have everything de-
They sometimes seem to reflect everyone's second choice and
sired. As a resul!, individuals (and groups) struggle, competing
no one's first choice. They sometimes seem not to be imple-
and cooperating-'with eaih other, trying to satisfy their individ-
mented, or to be implemented in ways thatchange the original
ual preferences and identities. Power is the capability to get
decisioll. They involve complicated combinations of trust and
what you want or to fulfill your identity.
distrust and substantial uncertainties that are only partially re-
solved by the process of decision.
rL Although the assumption canbe misleadiilg, most theories of
decision making assume that power is desired, that each indi-
vidual wants to have decisions rnade that are consistent with his
4.1 Decisions and power or her preferences and identities. The distribution of power in a
society, therefore, is a distribution of advantage in the pursuit
One of the ideas most commonly invoked to talk about decision
of a life consistent with personal values. The standard presump-
making when interests and identities are inconsistent is the
idea tion of democracy is equality of power. The standard presump-
of power. The concept of power reflects the intuitive notion of
tion of personal ambition is the pursuit of power. The standard
struggle, with outcomes determined by the relative strengths
of presumption of decision making is the struggle for power and,
contending forces. Sg*e people seem to get more of whát
they through power, for desired outcomes.
want than do others. Individuals and groups consciously pursue
Students of power use two kinds of models in examining de-
power and knowledge about power. Ffierarchy, one of the
most cision makin g. Force models of power portray decisions as being
142 A Pr{tMriR oN DECtsroN MAKTNG Multiple Actors: Conflict and Polüics 143

weighted combinations of the wishes of participants. Exchange COMPLICAIING THE MODEL

models of power see choices as being produced by voluntary There are numerous reasons why estimates of power are not re-
exchanges. They use "powet" as another word for "trading liable across decisions, and each of them defines a possible
advantage." elaboration of the basic model:
L.Power is both positionaland behavioral. Effectiveness due
4.7.2 Force Models of Power
to being in a particular position or havin g a particular role is
In a force vision of power, decisions by a collectivity are por- confounded with effectiveness that stems from a particular sryle
trayed as the result of various kinds of social"averaging" proces- or cleverness of action. The two aspects are unlikely to be well
ses. Those processes consolidate conflicting wishes into a com- correlated and will, in any event, have to be estimated separately.
promise decision. The exact compromise that will be realized 2. Power is domain specifrc. A person powerful in one domain
depends on the relative powers of the parties. There are many is not necessarily powerful in another. There is not a single
variants on the details of such processes, but the central idea is index of power for an individual decision maker, but different
that conflicting desires are pooled to arrive at ajoint decision. powers for different decision arenas. The domain specificity of
power is observed not only in government but also in business
THE SIMPLE FORCE MODEL firms, families, and churches. A force model has to have sub-
scripts to reflect the domain involved.
The simplest force model is one that assumes that decisions ate 3. Potential actors in a deciSion process have to be activated.
the weighted ayerage of the wishes of individual participants, Potential power is not always exercised. Attention is uncertain.
where the weights are the relative power of the various individ,- Energy may be deflected to other concerns. As a result, aforce
uals involved. The vision of the decision process is one in which model of power should decompose power into potential power
each individual participant has a wish that can be represented
and activation, with realized power being their product. But
by a number and some amount of power. As the individuals once this done, the model requires an independent estimate of
apply their powers in support of their wishes, tñe system either power or activation in order to use data on decisions to
records a decision that reflects the net effects.
estimate the other. And since attention is not constant over
This model is a simple and elegant variation on standard no- time, the model requires time subscripts on attention.
tions of force. If a particular decision outcome and the prior 4. The use of power affects power. There is evidence for
wishes of decision makers are known, the power of specific de-
force depletion: Afrequent obsenration is that power is "wast-
cision makers can be estimated. Alternatively, if the powers and
ed," used up. The usual presumed mechanism is tttut the exer-
the wishes are known, the decision outcome can be predicted; cise of power expends good will of past favors that must be re-
or if the powers and outcomes ate known, the wishes can be newed if power is not to decline. Power is a fixed resource,
which the exercise of power decreases. There is also evidence
For the simple force model to be useful as a basis for under- for force conditioning: Pow er acts like a skill that increases with
standitg decision making, the wishes of participants must be exercise. The mechanism may be a gain in competence or it
observable, and po\Mer estimates must be stable over repeated
may be a gain in reputation. fn particular, the exercise of power
observations of outcomes. The m.ain problem with the model is leads others to concede power.
its clear inability to fit dataon multiple person decision making.
If power indices ate estimated from two different decisions2 the All these additions to the basic model are plausible. Each has
indices do not agree. some basis in observations of real decision processes. Unfortu-
l'44 A t)RtMHtr tlN DECrstoN MAKtNc
Multiple Actrtrs: Cortflict and Politic,^r 145

nately, &s the model becomes increasingly realistic, it becomes other actor by observing individual wishes and resulting out-
more difficult to use empirically. In particular, using empirical comes and inferring individual power indices (bV assuming that
data to estim ate the various factors becomes-impoisible. The
those who get what they want have more power). Although it
amount of data required is several orders of magnitude greater
might be imagined that initial reputations for power would be
than the amount of dataimaginable.
¿tugmented by the processes of reputation formation, under
The result is that force models and metaphors arevery gen- fairly general conditions a different result is produced. If per-
eral, but they do not lend themselves to empirirul confirmation
ceived power determines actual outcomes, a process of updat-
or disconfirmation. Power can be conceiu"¿ as a force that ing perceptions by observing outcomes actually converges to a
weights wishes to determine an outcoffie, but such a conception
stable distribution of perceived power that depends not at all
is not particularly useful unless there are independent ways
of on the initial distribution of perceived power but only on the
estimating power.
clistribution of wishes.
These features of force models are of some significance in
thinking about ways of equalizing power in a social system.
Despite those empirical difficulties, force models of power are They suggest, for example, that strategies for reducing dispari-
useful in suggesting a source of power that might easity be over- ties in power probably should include not only redistribution of
looked. A conspicuous feature- of such modeh is
-ifru;;;;, resources but also redistribution of personal preferences and
comes from the relation between one's own wishes utt¿ the identities. That result has implications for demo cratic gover-
wishes of others. There are advantages to having preferences nance. It indicates that a theory of demo cratic governance that
and identities that aÍe consistent with those of poríriful people. emphasizes the role of demo uaticatly elected officials in re-
This advantage is sometimes decried as a ,orrrplication in mea- flecting the wishes of citizens may encounter intractable diffi-
suring power (distinguishing the chameleon fróm the "g.r,.rine- culties in achievirg political equality. Political equality may re-
ly" powerful), but it is a realphenomenon also. ff decision mak- quire not only efforts to make government responsive to wishes
ers want what other people want, they are more likely to get but also efforts to affect the wishes.
what they want.
More generally, considerable advantage comes from having DISAPPOINTMENTS IN FORCE CONCEPTIONS OF POWER
wishes that lie close to the "center of gravify" of the rest of the
The idea of power as force yields some interesting nonintuitive
system. Conversely, individuals who have wishes that lie far
ideas, pafticularly those that consider dynamic changes of
from the "center of grav ity" will experience persistent power- power over time and the role of wishes in reputations for
lessness. Interestingly enough, from tttr perspective of
theories power. Flowever, it is also an idea that has a history of disap-
of demo ctacy, such powerlessness will not ü" particularly ,"- pointment as a concept for interpreting decision making.
duced by givit g citizens who are thus disadv antáged more
trol over resources or access to decisions. Power as Thutology. "Pow er" tends to become a post hoc label
Consider, for examPle, how perceptions of power (and thus fbr the unexplained variance of students of multiple person de*
power) ate influenced by the distribution of preferences or cision makin g. Post hoc labels for unexplained variances have a
identities as translated into wishes. Suppose the po*rr of an in-
long tradition in social and behavioral science. They creep into
dividual is the power he or she is periéive¿ Uy btners to have. the economics literature, using words like "utility" or "risk";
Suppose further that each actor estirnates th; power of
each into the sociology literature, using words like "norms" or "legit-
146 A PRTMUR tlN DHClstoN MAKTNC Multi¡tle Actors; Conflict untl htlilic,y lr47

imacy"; into the anthropology literature, using words like "cul- I)ilficulties in Measuring Powen If differences in power are enor-
ture" or "tradition"; and into the psychology literature, using nlous, as they ate sometimes when the power of dominant
words like "personality" or "habit."
!¡roups is compared with the power of weak groups in large so-
In each case, there is nothing intrinsically wrong with the cial or political systems, measures of power app ear to be fairly
label or the intuitive idea behind it. The labels become prob- useful. The power measures taken from one situation extend to
lematic, however, when they are used as generic after-the- ¿t number of others, and broad predictions can be made. For ex-
fact explanations of events. That is the frequent fate of the co[- irmple, by almost any measure of power, poor people ate sys-
cept of power. ff it is determined that someone has "power" by tomatically weaker politically than other citizens in almost all
observitg that they get what they want, using power as an "ex-
llolitical systems.
planation" of why they get what they want is an exercise in On the other hand, when measures of relative power among
tautology. active participants in a smaller political system, for example in
¿rn organ ization, are used to predict decisions, the results ate

The Symbolic Significance of Talking about Power. Power is also usually disappointing. It does not appear to be possible to pre-
linked to important personal and social beliefs. ft is part of the clict future outcomes reliably by estimating power from previ-
grammar of modern politi cal and social realism that contrasts ous outcomes. Our restricted abilify to measure power com-
itself with grammars of idealism, with its hopes for civic duty bines with the complexity of power to lose whatever "signal"
and the integrity of office. Power is tied to beliefs in the impor- there may be in "pow er" in the "noise" of other factors.
tance of individuals. Ideologies of individualism are couched in One reason the effort is unrewarding lies in the elusive na-
terms of self-esteetn and empowerment. Power is tied to col- ture of fundamental ideas about power. Consider, for example,
cepts about the proper relationships among individuals, for ex* the following three problems:
ample, democratic beliefs in equality. Many societies seek to re- I. Interpersonal comparisons of wish fuffill*ent. If power is to
duce disparities in power. be measured by the ability of individuals to get what they want,
As a result of those connections to beliefs that are deeply how can comparative stateménts about relative power be
held, the term "power" mc)bilizes considerable emotional force. made? Power comparisons require that the extent to which one
Its use protects a person from being considered naive, and person's wishes are realized be compared with the extent to
shows attention to concerns over disparities in power. Thus, the which another's are. This presents a number of problems. A
term is often a signal by which individuals (including students of measure based on wish fulfillment requires us to determine
decision making) proclaim both their sophistication about the whether some people came closer to their desired preferences
world and their concern for the position of the weak. or identities than others. For a variety of reasons, interpersonal
Images of power ate symbolically signifi cant in another comparisons of subjective utilities are viewed with doubt and
sense. They evoke conceptions of life as struggle and conquest, are usually rejected in theories of choice.
domination and subordination. They link "pow ef' to ideas of 2. Strategic expression of wishes. The expression of prefer-
forcefulness and traditional conceptions of masculinity, thus ences and the declaration of identities are strategic weapons of
tend to define decisions (ur well as the study of decision mak- negotiation. Those who want to assess them accurately must
ing) as being domains for establishing "manliness." Those asso- therefore solve the problern of measuring things that are being
ciations of the concept of power have consequences both for consciously and strategically manipulated.
making decisions and for interpreting decision processes. 3. Pliability of wishes. The usual treatrnent of power implicitly
l4tt A PRTMnR oN DricrsroN MAKIN(; Itlultiple Actors: Conflict and ltrilltits l4q
considers wishes to be fixed and measures how f'()urse to the word. Within an exchange model, the ability to
well the world
conforms to them. fn teality, preferences and identities prlrsue one's preferences or fulfill one's identity (i.e., powgr)
adapt at
least as much as the world does, and measures
of their fulfill- tlcpends on three things: control over the rules, control over re*
ment aÍe, atleast to some extent, measures of their $ources, and control over preferences and identities. So any ex-
For these reasofls, as well as others, the search for e hange theory of power considers the ways in which these three
a good em-
pirical measure of power seems to be endless urrá ilrc determined.
('ctntrol over Rules. Exchanges, like all social interactions, take
¡rlace within a social structure of rules. Social rules set the play-
4.1.3 Exchange Modek of power
ing field for decisions and constrain them. Social rules specify
The crucial difficult¡r in using force conceptions proper decision procedures and proper justifications for deci-
of power is the
fact that power weights cannot be independently sions. Some things are taken as given, some questions remain
observed but
must usually be estimated from their consequences. unasked, and some alternatives are unexamined. The ability to
models of power address that problem by fécusing ¿rf'fect those constraints is a fundamental source of power.
on a small
number of factors that provide a trading áOu antage One of the principal forms of contention in the study of deci-
of voluntary exchange.
v '¿)
in a system sion making is the claim that theories of choice systematically
ignore questions of how the rules of decision makin g arc speci-
THE SIMPLE MODEL OFEXCHANGE lied while devoting enormous attention to the ways individuals
()perate within the rules provided to them. Theories of power
The fundamental idea in an exchange model is that consider only the exercise of power within a taken-for-
(individuals, groups, organ izatio*) enter into given set of rules may overlook the effects of these social "con*
volun tary ex-
change relationships regulated by ro*. system
of rules. Each stitutions" on decisions. As generations of legislators have dis-
participant brings resources to the arena. Those coverod, the rules of legislative procedure ate not neutral.
resources in-
clude such things as moiley, prope rW, knowledge,
access to others, rights and authorities, and
information. The
process of choice is one of arranging mutually Control over Resources. Power in an exchange model comes
acceptable trades
within the rules. Each indiviOuát seeks to improve his f'rom control over resources desired by others. Whatever is re-
or her quired to satisfy one's own preferences or to fulfill one's own
own position by trading with other individuals.
When the ex-
changes reach a place where no more legal identity is obtained by exchanging resources. When decision
and mutually ac-
ceptable trades are possible, the process stops. makers have something others want, they can exchange it for
something that they want. Most studies of power in the ex-
EACTORS IN TRADING ADVANTAGE change tradition emph asize the simple point that possession of
desired resources gives power. In order to be powerful, decision
In an exchange model, it is imagined that people improve makers seek control over resources.
positions by trading money, status, affection,^o,
any other re- Such ideas have been used in studies of decision making to
source that they control and another wants. fn"
*oiO ,,po;;;, yield predictions associated with resource dependenry theories:
is superfluous to such theories. The theories could that individuals and organizations will respond to internal and
b* present-
ed, understood, and used as theories of exchange cxternal forces that control vital resources, and that individuals
without re-
150 A pRrMtrR tlN DticrstoN MAKINC
Multiple Actors: Conflict and Politics l5 I
and organizations will seek to limit their dependence on such
resources. Control over resources empowers individuals and
groups. Thus, the standard advice for becoming powerful: Be. H,xchange theories of power suggest that decision making can
come rich; seize a host age; build a better mousetrap. be seen as a grand insurance scheme in which favors are offered
today in return for an option on reciprocal favors in the future.
Actions are taken in the expectation of future favors, but the
Control over Preferences and ldentities. As parents and advertis. ] process of exchange is complicated by the uncertainfSr of the fu-
itg agencies have repeatedly demonstrated, a possible way of ¡ ture. I)ecision makers know that they may need help in the fu-
gaining power within an exchange model is to transform thg
ture but rarely know what kind of help, or when, or from whom.
wants of others so that they demand goods you can provide. fÍI,
Thus, they spread favors broadly, hoping thereby to buy "insur-
stead of seeking to provide the things that othérs want, A
¿rnce" against future needs.
power-seeker tries to induce others to want the things he or she
Favors ate done with some expectation of return, but risk
can provide. In the literature on leadership, this is cátted trang.
i ¿tversion leads most actors in a social system to give much more
formational leadership. In some other literatures it leads to the '
than they demand in retum. This imbalance is accentuated by
"problem of the happy slave," for power comes frorn making
people happy with what is given to them.
the tendency of recipients to discount past favors heavily:
"What have you done for me lately?" This structure of recipro-
Changing others is not the only way to change preference$,
cal insurance in favors, like rules of politeness in social interac-
and identities. It is also possible to imagine changing onr's own ,
tion, strengthens the system and builds a community in which
self. In particular, notice the other side of the piopósition that
the level of unsolicited favor-giving is quite a bit higher than the
trading advantage is well served by having things that otherE,
demands for reciprocity.
want. TLading advantage is also well served by wanting things i

that others do not. With some exceptions, it is deviant turtlg j
and identities that are more easily satisfied through voluntary 1

exchange. A person who craves Swedish licorice *itt generuür 4.2 Decisions and Coalitions
do well in a world of chocolate lovers. - 7 ;


This proposition differs substantially from the comparable A second political metaphor for decision processes in multiple
proposition derived from force models of power. thr de. actor systems is one that highlights bargaining and coalition
cision process produces decisions by "averaging" over partici. formation. The distinction between bargaining and coalitions
pant wishes, participants with deviant wishes will be persistently on the one hand and power on the other is somewhat arbitrary.
disadvantaged. When the decision process arcanges volun ta,ry í
Force models of power, for example, presume some process
trades among participants, participants with de-viant prefer. by which wishes rre combined, unO coalition formatibn is a
ences will be persistently favored. One relatively pure fórm of prime candidate for such a process. Exchange models of pow-
exchange decision making is the "logroll." togrótt processeg
gr presume a process by which exchanges are negotiated,
tend to improve the relative position of deviatrir (r.!. p*opit ¿rnd bargaining is a prime candidate for such a process. Bar-
with strong values or identities unrelated to the valuer or idén- gaining and coalition formation, however, provide foci that
tities of others). This is, perhaps, one of the reasons why such a put greater emphasis on the interactive social aspects of exer-
procedure strikes some people as essential and strikes others as cising control over decision making. Within this view, decision
perverse. making involves horse-trading and logrolling, associations and
alliances. t
152 A PRTMHR clN DucrsroN MAKTNG
Multiple Actors; Conflict and Politics I 53
4.2.1 The ldea of Coalition OI}JECTIVES AI{D RULES
Ideas of bargaining and coalition formation, like ideas ln the resulting bargaining process, there are two key decisions
power' are organized around an assumption of preference Io be made: Who will be in a coalition, and how will the spoils
identity inconsistency. I)ecisions are made un¿ cooperative bc divided among coalition members? Actors attempt to affect
programs are pursued within a context of potential lonflict. t heir personal shares of the spoils, deciding what coalition to
Theories of bargaining and coalition formatión emphasize
two .ioin based on an estimation of the way those allocations serye
important aspects of multiple actor decision making: t heir preferences or help fulfill their identities. Within a partic-

First, they emphasize the structure of the formal decision ular constitutional order , rational a:ctors generally desire to be-
making systeffi, the rules that are (at least implicitly) treated Iong to a winning coalition, to which end they seek allies. How-
as inviolate by participants. Those constitutive rules over, they do not want to have too many allies, for each ally
the resources available to persons occupying particular roles makes a claim on coalition winnings. They like to be members
(e.g. citizen, legislator, diiector, executive) and speciff the of a coalition that is large enough to gain control of the system,
combinations of roles that constitute a winning routition llut no larger. The "minimal winning coalition," if it can be
pable of declaring a legitimate decision. For example, iclentified and achieved, will maximize individual payoffs. In a
dard legislative rules for voting specify that any coalition world of uncertainty, subcoalitiofls, and maneuver, however,
;;; somewhat larger coalitions are likely. Rule-following Actors, on
taining more than half of the votes can tegltimately make
decisions for the entire community. Sometimes such rules the other hand, seek coalitions and distributions that will satisff
are themselves imagined to be the result of decisions negoti- the norms of proper behavior they accept. They have standards
ated within a supergame, but they are considered as given. rrf "fairness" and expectations about legitim ate coalitions and
Second, ideas of bargainitg and coalition formation empha- legitim ate coalition actions.
size the ways in which individual decision makers pursue Conventional theories typically assume the existence of a
objectives by makitg deals. The primary objective is to form monetary prize that goes to the winning coalition and is then
coalition capable (within the rules) of making decisions favor- distributed among the coalition members. In such cases, the
able to its members. Coalitions are formed by entering distribution of winnings has the rather special limitation that
agreements with others that speciff the decisions that one person's gains are another person's losses. The demands of
will be
made by the coalition. Those decisions simultaneously e>aracl one coalition member are strictly conflicting with the demands
resources from the system through coordinated action of another.
coalition members and distribute those resources within the There are winning coalitions that involve fewer than all par-
coalition through competition among coalition members. ticipants and winning coalitions have broad rights of decision
making. Theories of coalition formation within such rule struc-
tures are redistributive in appearance. They allow members of a
4.2.2 Building Coalitions winning coalition to redistribute the resources of the organiza-
Most theories of coalition building presume that individual tion or social system in their own favor. Subject to limitations
ac., imposed by the rules, winners can (ctax.') losers and claim the re-
tors have well-defined preferences or identities and that they
enter into coalitions to satisfy those preferences and fulfill sulting "revenue." There is ample evidence that the redistribu*
those tive appearance of "winner-take-all" rule structures is not mis-
identities. The actors are assumed to do the best they
can,given lcading, since such systems have often produced sigpificant
the rules of the game and the demands of other participaniu.
Multiple Ac:tors: Cottflict und Politic,,y I 55

There ate, however, some restrictions on redistribution. tttade (at least implicitly) in this kind of process. Which of the
most obvious restriction is the ability of the system to irnpossibly large number of possible coalitions will be consid-
that losers will be bound by the decisions of winners. urcd? What is the nature of coalition contracts? How can part-
theories of decision making through coalition formation tters to a coalition enforce their agreements when they are not
sume a constitutional order that enforces the rules cfTected simultaneously? How do variations in mobilization of
and deci,
sions within the rules. The conditions for sustaining
thator¿rt participants affect the operation of the formal rules specifyitrg
-u*olgrational actors require thatparticipants be uble to calr*., winning coalitions? How ate particular identities called to at-
late that, even if those rules currently plaóe them in
a ¿isaOvfl1. tention or forgotten? How ane inconsistent rules applied in
taged positioil, in the long run tfréii interests will
be bettef fbrming coalitions?
seryed by accepting the rules of the game than
by moving to ail.
other game. Similarly, identity-f"Wtns actors must presumg
4.2.3 ComplementariQ in Demands
in the presell d.rcisionsysrem wlr br,
run' be more consistent with their unfotding identities
ilil; i
than The imagery of coalition formation used so far centers on the
undertaking a different system. idea that a winnittg coalition gains a certain amount of resources
Although it typically pays to join the winning coalition, which it then can distribute among its members. The idea is bor-
therÉ j
are exceptions: The more you demand of the qpoils, rowed from theories of games in which there is some collective
the less at. i
tractive you become as a member; and there is atwuír payoff associated with a coalition. It is useful imag er!, but it fails
the possi.
bility that a particular individual may be be ttw off in a ..Iósing,, to capture an important aspect of coalition formation in deci-
than in a winning coalition. As in the case of simple sion making: the role of demand complementarity.
ships, effective strategies depend, among othei things,
whether the process occurs only once or many times, CONGRUENCE AND INDIFFERENCE
whether everyone can be assumed to be "tational"
or whether a
small number act irrationally. When a coalition is built in order to secure and distribute
monetary winningS, it is usually assumed that the demands of
LIMITS TO IDEAS ABOUT COALITION FORMATION coalition members are strictly opposittg. Although different
coalitions may receive different payoffs, the share of the pay-
The central problem with many ideas about coalition off given one member of any coalition reduces the share that
that are found in the literature is that they adopt relative$ can be given to other members. It is this feature of the demands
forms of uncomplicated ration ality or rule followirg. that leads to the theorem about the "minimum winning
theories of coalition formation tend to make heroic coalition."
tion assumptions, often assuming that all participants
il;; p*;. 1
Decision making, however, often involves the crafting of a
policy decision, such as a decision to purchase an affay of
mation' Simple theories of rule foilowing tend to assume equipment, to approve a budget, to pursue a particular set of
rules ate obvious, shared, and always evoked. These pure advertising strategies, or to enact a piece of legislation. In these
seem not to capture coalition formation in multi-person, cases, the demands of potential coalition members are ordinari-
goal, multi-identity, multi-arena situations in
conditions of lim. ly not strictly opposittg. A demand that an organ ízation pur-
ited ration aliry,limited attention, and limited consistency. chase a new piece of equipment and a demand that it exhibit
Behavioral students of coalition formation suggest "modernity" are complementary.A policy that meets one de-
thát there
are a large number of difficult assessments that need
to be mand reduces the marginal cost of meeting the other.

complementarities of demands come

in many forffis, but two
"r^f:": :::-::: ;.':r::': -"::
extreme cases illusttate the importance rhese fearures or
of demand complemen-
tatity to policy formation. The first is l)ropositions about bargaining advantage. In particular, the
the ;;;;' of congruencery likelihood of any particular individual belng on u *inning
coalition members who want policies coali-
tha t aremutually support. titrn depends on the degree of complemenlarity between
ive' The most natural coalition imaginable his or
a coalition of persons altr of who*l
in decision making is hcr own preferences and identitieJ and those 'of otners.
antthe same policy. Each Com-
additional coalition member adds ¡rlementatity may come from either congruence of demands or
strength to the coalition with- tnutual indifference. Thus, being on a wiñning coalition
out exacting any "cost." Sales agents is facili-
whose customers all desire tated by having preferences and identities that either mesh
precisely the product being sold
urc involved in such a coalition. completely with the preferences and identities of others
Each additional customer requires or by
no change in the design of having preferences and identities that do not connect
the product' rn a fike mann er, at all.
acoalition of ádvocates of a par- The advantage produced by mutual indifference is illustrated
ticular tax reform can add additional
advocates without cost. by a sfylized story of traditional politics involving
SimilarlY, individuals with congruent three contest-
identities form coalitions ing groups: capitalists, workers, and farmers. rn this stylized
more easily than individuals with incongruent
identities. par- portrait, it may be plausible to assume that a winning roáition
entlchild, boss/subordinate, or buyer/seli;?ñ,
formed than are combinations of
ur. more easily consists of any two of these groups. From the present poi"t
roles that do not fit together. oi
The second extreme case of complementarify view, the traditional coalition advantage of farmers has
the case of mutual indffirence. Tivo in demands is that their demands were more consistent with the demands
individuats who are indif- of
ferent to each other's dr*unds cither workers or capitalists than either of the others were
satis$r both' In effect, indifference
can f";^Jñiirt;;iiri"" ro cach other. While workers and capitalists often tended
to have
provides another form of oppositg interests, farmers tended to want things that were
congruence' Policies that do not rel-
affert orr* another and id;;: atively unimportant to the other two groups.
ties that do not touch one another :

are natural allies in situa- Since farmers were able to contribute votes to the coalition
tions in which coalition size makes
a difference. of which they were a member and were relatively "low-cost,,
In practice, neithel complete ,orrg*ence
nor complete indif- coalition members, the only two winning coalitions thatwere
ference is commor, but the principíe
extends to demands that likely were between capitalists and farmers and between
overlap without being identical. work-
óustomers who want similar crs and farmers. Thus, farmers should always be found
products are more easily accommodated on the
by a single supplier than winning side. A casual historical observation is that, for many
customers who wanl radically different
ones. Advocates of rel at-, years, farmers tended to wind up in the winning coalition
ed tax policies can form a cóalition with
more readily than advocates one of the other groups (either workers or capitáists).
of conflicting tax policies. Rule followers The his-
with consistent rule sets torical story is excessively simple, but the point is clear: Havirg
can form coalitions more easily
than those with inconsistent sets.
demands that are complementary with others makes one
The propositions are intuiti*ry
obvious, but they Iead to sorne u prrl
features of coalition formation f'erred coalition member.
that are sometimes overlooked.
For example, demand complemen
taritymakes it ltkü;il; win-
ning coalitions will contain more I'OLICY LOGROLLS AS COALITIONS
members than the minimurn
required to win. The policy cost
of additional members is less Complementaúty is a key feature of coalition formation
than the increased tetürityihey provide
against misestimates of through policy "logrolls. " A logroll is a coalition of individuals
strength or random fluctuations in
decision outcomes. or groups who are largely indifferent to each other's demands
158 A PRrMriR trN DHCrsroN MAKrNci
Multiple Actors: Conllict antl Politics I59
but agree jointly to support each other so that each lorming coalitions among mutually indifferent coalition mem*
can have
what he or she wants. one decision maker agrees
io *ufport ,r* hcrs is an important paft both of theories of multi-actor deci-
pet project of another (about which the former
decision maker sicln making and of observed decision making. Logrolls are
is indifferent), receiving in return similarly indiffereni filund not only in the United States Congress but also in busi-
for a favorite project. ness firms, military organizations, and universities.
The usual example of a preference-based logroll
is the annu-
al "Rivers and Harbors Aót" of the united sátes
collection of local projects that collectively commLJ"
vote, although any one of them alone would pr;t;úy-"¡-iry It must be observed, however, that logrolls do not appear to
supported by more than a handful of legislatorr. "ot ex. occur as frequently as might be expected from an analysis of the
tt úsual coalition building advantages they offer. One clear reason for
ample of an identity-based logroll is thi mutual "
tolerance of
pergonal identities traditionally found in liberal the relative infrequency oi logrolls is that they require toler-
¿"-o"r*i.r. ¿rnce-even encouragement-of differences. If the first instinct
Logroll coalitions are particurarry attractive in a worrd
of sin- of decision makers is to try to convert others with tastes or con-
gle-issue participants, participants who
have passionate de-
mands on a few issues and much weaker feelings ceptions of self difterent from their own, they are unlikely to
on -ort. po,
e¡ayfle, contemporary democratic political syjems ,""ro f orm logrolls with them.
pur-. There are also practical problems in organizing coalitions
ticularly attracted Iogrons u-ong single-issue purtiripurrts
committed to protecting based on mutual indifference:
their claim-s oripublic
contemporary business-decision systems seem partir"l*iy;
L. Problems of discovering partners. Most people find it easier
tracted to budget logrons among single-issue puiti.ipu.rÁ to identify people who agree or disagree with them than people
mitted to protecting their claims to UuOgetary who are indifferent to them. Similar or opposing preferences
r.rpport. Á _or"
lpocryphal example would be a marriage in whicñ a wife makes lead people to attend the same meetings and to be engaged
the decisions about where to live, what to eat, how with each other. As a result, it is ordinarily more difficult for de-
to dress, and
how to raise the children; while her husband cision makers to identify people who are indifferent to the
makes deci;io;;
abo.ut the family's policy toward foreign
irrt"rriJtiorrut things they value than to identify people with attitudes or iden*
trade, and military strategy. "utionr, tities that do not touch the same dimensions.
As the example of liberal democracy suggests, there 2. Problems of organization. A coalition of mutually indiffer-
is a
sense in which borh modern democrary ánd
il'o¿"in organaa- ent people is more likely than a coalition of similar thinking
tions are designed as rogroils. Individual freedom, people to require conscious planning and strategic action,
and division of labor can an be seen as ways rather than gradual social commitment. Since participants in
tions in order to make decisions. Each of those "t seems to be such coalitions do not speak the same language or share the
adopted more readily and to function more same expectations, and are not linked in interconnected identi-
smoothry when
there is a substantial area of mutual indifference. political ties, coalition formation does not arise naturally from social in-
tems and organizations contribute to mutual teraction. In throwing together unfamiliar bedfellows, such
uulajp information barriers that promote mutual ignorance. uy ooalitions are confusing not only to observers but also to partic-
Policy logrolls are rüays of making decisions ipants.
trraT are easity
overlooked when decision processes áre framed
u. 3. Problems of trust. Logrolls are difficult to sustain if their
lem solving, rule following, or exercises of force. "itt!, irob-
The ióa of ilgreements require exchanges over time. Since bargaiñs offen
Muttiple Actors: Conflict and Politics l6l
unfold over time, and there are only weak mechanisms for sn-
mediate conflict. At the same time, outcomes depend on which
forcing agreements, logrolls tend to require a certain amount
of individuals are activated. Actors move in and out of the arena
trust. But trust is hard to maintain among mutually indifferent
in response to various claims on their attention. Since demands
on attention are constantly shifting, the climate of decision is
4- Problems of strategic
faktfrcation The logic of logrolling in- unstable in many small ways that cumulatively affect the course
vites falsification of preferences and identities A decision
maker of events. Decision makers are pressed to meet the inconsistent
who agrees with a potential partner on one issue and knows
that demands of a changing group of actors. Consequently, deci-
that individual is indifferent about a second issue may be tempt-
sions are likely to be unstable over time and space. For exam-
ed to act strategically. fnstead of revealing the agreement
on the ple, the interests and identities activated during the adoption of
first issue, the decision maker claims indifference in order to
a policy will often be different from those activated during its
arrange a logroll on the second issue. Since this strategic
oppor- implementation.
tunity is known to everyone, it undermines the alreadJ, tenuous
The instabilities may conceal regularities that should also be
basis for negotiation among the mutually indifferent.
noted. Although any particular decision is subject to idiosyn-
These difficulties do not prevent logrolls, in either political cratic flows of attention from possible participants, the statis-
or business alliances. They do, however, suggest some reasons tical distribution of attention imposes a certain amount of
why some of the coalitions that offer the hlghest payoffs are aggregate consistency on the processes. Some people are dis-
also the hardest to organ ize and sustain. fn u *orld of advantaged by the rules of participation, by other claims on
mutual indifference is a source of power and a basis for roii- their attention, and by inadequate resources to exchange for di-
tion. In a world of cooperative joinf action, mutual indifference rect involvement. fn any particular situation they may be able to
is a source of difficulty and lack of solidarity. Since mutual overcome those disadvantages, but not on a regular basis.
difference and mutual support are difficult io achieve simulta-
neously, winning coalitions tend to be internally ineffective, 4.3.7 Participation in Muttiple Actor Decision Making
whereas internally effective coalitions tend to be unable
to gain
support broad enough to win. Participation patterns affect decision efficieilcy, decision out-
comes, and decision acceptance. To interpret the shifting mix of
participants in multiperson decisions, consider two aspects of
4.3 Participation and Decision rnstabilities
the activation of participants in decision making. First, what arc
As was discussed in Chapter 1 and 2, modern theories of limited the constraints on participation? What are the rules that regu-
late the involvement of potential decision makers in decision
SJionality and of rule following emphasize factors of attention:
Which alternatives are considered? Which consequences? arenas? Rules speciSr the rights and responsibilities of actors:
preferences? Which identities are considered? Which
rules? In
Who can participate? Who must participate? Who cannot par-
a similar fashion, multiple actor theories of decision making ticipate? S"cottd, how do decision makers allo cate attention to
focus on the question of which potential interpersonal decisions within the rules? How do the other claims of life im-
tencies are evoked, thus on who participates, when, and pinge on decision participation?
The demands of time and the constráints of rules assure
different combinations of participants will be activated in dif- RULES OF PARTICIPAIION
ferent places. As a result, not all the potential contradictions
in Insofar as preferences and identities are shared or consistent,
preferences and identities come into pluy, and there
is less i;- the pattern of participation makes little difference. In ancient
l,62 A t,t{tMrirr oN DrictstoN MAKTNC
Mulri¡tle Actors; Cortflict un| kilitics 163

Athens it was possible to imagine choosing leaders by lot, since

I1¡r access under the slogan "no taxation without representa-
citizens could be assumed to be essentially equivaleni. In an or-
tion." On the other hand, in some theories of judicial and bu-
ganization with consistent preferences and identities, individu-
roaucratic decision making, personal consequences are bases
als often enjoy broad authority to act on the behalf of the col-
frrr forbidding participation. Judges are expected to disquali$r
lective because they ate presumed to share common goals and
themselves from participating in decisions that affect their
personal interests. In both cases, the magnitude of personal
If preferences and identities are inconsistent, however, the consequences from a decision is a factor affecting the right to
decisions made will depend on who participates. Left to therr-
selves, some individuals would participate too much (from the
2. Social benefits. Some individuals are seen as having made,
point of view of the system), and some too little. For example,
or as being able to make, greater contributions to a society than
effective and talented participants may exit from the process,
other individuals. Those contributions are generally used to jus-
because they bear the costs of participation themselvei, while
tify greater access to decision making. The arguments are of
gains from their participation flow primarily to others. This col*
two kinds. First, there is an argument of competence. It seems
lective flight of the talented leads to decisions unattractive to
appropriate to arrange that individuals with greater relevant
most. Conversely, some individuals ate, for personal reasoflso
competence be more active in making a decision than individu-
more eager to particip ate than their contributions justify from
als with less competence. Second, there is an argument of com-
the point of view of others. Their disproportionate activation
pensation. It seems appropriate that individuals who are more
also leads to decisions unattractive to most.
valuable to a society-in the sense that they have contributed to
As a result, every social system has rules of participation with
the society-should be compensated by being allowed and en-
respect to every decisiotr, rules that require some people to par-
couraged to participate more than those individuals who are
ticipate, allow others to participate, and forbid th¿ paiticipation
less valuable.
of still others. Justiffing such rules is an important part of poht-
3. Creating community. Participation rules are not just devices
ical and decision philosophy. Who should participate in deci-
for regulating the substantive content of decisiolls. They ate
sions on the allocation of water rights in a watershed? Who
also slrmbols of and instruments for the creation of a communi-
should participate in decisions abouithe closing of an industrial
ty. Rights and obligations to participate are linked to accep-
tance as colleagues or members of a communify. They symbol-
Although a detailed discussion of participation rules and
ize individual significance and the existence of a meaningful
their justifications is beyond the scope of this book, it may be
collective. Participating, or not participating, in decision mak-
helpful to note that participation rules typically reflect three im-
ing is an importantcertification of "citizenship," of being a per-
portant concerns:
son of importance and recognizíng the responsibilities of im-
1-. Personal conseql¿ences. Any particular decision has more portalce. A personal sense of efficacy (ot alternatively a
important consequences for some individuals than for others. personal sense of alienation) depends on participation in deci-
Participation rights are often made contingent on the extent to sion making.
which the individual is personally affected. In democratic theo-
Because these considerations arenot necessarily clear in their
ry, it is generally argued that a proper decision system will pro- jgint implications, and because different people may differ with
vide greatet access to those who are affected by a decision than
respect to how they interp rct them, the contention over partici-
to those who are not. Early colonists in America campaigned
potiott rules in decision systems is a critical constitutiona! strug-
164 A pRTMERoN DrcrsroN MAKTNc
Multiple Actors: Conflict and Politics I 65
gle. It often leads individuals to fight
for participation rights in which participation can be indirect as well as direct, the ways
that, once acquired, they will scarJely
quent inconsistencies between the participation
ii .u,rr.,
"*"; indicatedfrr. in which future participation is affected by past participation,
personal preferences and identities -and and the ways in which patterns of participation affect the legiti-
tñe p*ti.ifuiion de, macy of decisions
manded or permitted by others.

Indirect Participation. Theories of limited attention typically as-

PARTICIPAIION PATTERNS WITHIN THE sume that a fixed amount of time and energy is allocated among
Much of the structure of participation is found contending claims. Attending to one set of problems, prefer-
in the rules. cnces, choice situations, or identities precludes attending to oth-
Mosr decisions carry with lhem rules mand";iñ,
uiroüng, o, ers. Individuals enter a decision arena directly with their voices
prohibiting the involvement of particular individuars
or roles. and physical energies. They complain, protest, organize, and
Nevertheless, there is room for üme behaüoral
variation with- argue. I)ire ct participation in one decision arena makes simul-
in the rules. some-decisions happen because
particurar individ- taneous direct participation in another difficult or impossible.
uals who might wel.haye been piesent
actualiy were not, or be. The assumption is useful, but it clearly is not quite correct. In
cause some individuals who might well
- háve been absent particular, it underestimates the importance of indirect atten-
actuallywere present. tion, or in this case, indirect participation. People often dele-
-Participation decisions can be seen as actions gate their concerns to an appropriate agent (a representative,
taken by eithera rational actor or a rule-fonowid".i;^'iil;;;
relatively conscious factors make participation leader, lawyero or lobbyist). By using representatives, decision
;ñ;;;: participants eircumvent the limits on attention. In effect, by
1' Salience. Decisions are perceived to be important substituting monetary resources (in the case of a hired repre-
to pref- ;

erences or identities. sentative), social capital (in the case of a volunteer representa-
2' Efficaqt. participation is seen to have an effect tive), or threats of future retaliation (itt the case of a partner)
on the out-
qomes of decisions. for time, participants are able to be in several places at the
3' Efficienq' There are no better alternatives same time. The complication, of course, is that agents are im-
for achieüng
preferences or fulfilling identities. perfect representatives.
rt is not surprising, The example of retaliation is a reminder of a second impor-
perhaps, to find that individuals are
more tant form of indirect participation. Absent participants are rep-
l+"ty. to participate in delisions when rr,Jr-"*rllür-"rt,
identities are affected o, resented by the rottsCiousn-ess of others that if itt.y act suffi-
1lT ryhen they are not, more likely to ciently atvariance with the wishes of those who are absent, the-v
participate when they think they can affect ¿"Éiri".
will mobil ize the absent to be present. Customers are "prese nt"
more rit"lá"rr""-",
than when they think they cannor, and
pate when they think they have no arternative ñti.i- by virtue of their threats to switch their allegiance to other
ways ár urüng up- brands. Employees are "present" by virtue of their threats to
propriately or accomplishing what they
want than when they leave for other employment. Citizens are "present" by virtue of
think there are otherways.
their threats to emigrate to other political parties or other lands,
. Tl" principal issues associated with the a[ocation of.atten- Threats of participation need not be overt. They ate implicit
tion have already been considered (subsections
1.3.1 and 2.2.3) in any social relationship. Decision makers consider the possi-
and will not be reviewed herg. It may be
useful, however, to ble effects of their deliberations and their actions on the mobi-
note three special features of participation
behavior: fhe ways lization of potential participants. As they do so, they make
I66 A PRIMER oN DECISIoN MAKING Multiple Actors: Conflict and Politics 167

guesses about the wishes of others and about the likelihood that On the other hand, participation has numerous positive side
those wishes will lead to a change in participation. Those guess- consequences. It provides social certification of position and
es affect actions, but they are clearly subject to substantial un- ()pportunities for interaction. It sometimes occurs in the stimu-
certainty. Exactly what the threat is and whether it will, in fact, Iating circumstances of time pressure and social excitement. It
be executed is likely to be unclear. As a result, the efficiency of offers explicit confirmation of human importance in general
threats is balanced by their imprecision. This, in turn, periodi- ¿rnd of the importance of particular individuals. I)ecision mak-
cally leads threat-givers to execute threats to participate in ors often complain about the pressures of the role, the unrea-
order to confirm the genuineness of the threats-even though sonable demands made of them, and the stresses of being re-
they would otherwise be happy to stay away and allow minor sponsible; but those pressures lend an excitement to decision
deviations from their most desired courses of action. making that enhances its attractiveness.
That dynamic makes mobilization of the weak a two-edged The net effect of these countervailing effects is not easy to
sword. Because participation makes a difference, mobil izing predict in every case, but it seems likely that in most cases frus-
the weak increases their influence. But if mobilization of the tration is slow in onset but increases with time, while excite-
weak stimulates policies that encroach too much on the prerog* ment is immediate but wanes with time. As a result ) a natural
atives of the absent strong, the strong will be activated and act sequence to be expected is one in which participation first in-
to restore their position. The difference between the policy creases the attraction to decision making and then gradually de-
point at which they withdraw from direct participation and the creases it.
policy point at which they reenter defines both the possibilities
for strengthening the weak through mobilization and the dan- Effects of Participation on Decision Legitimacy. Participation
gers of doing so.,
also affects satisfaction with the process and outcome. The
most common assumption in discussions of participation in de-
Effects of Participation on Participation. Participation in decision cision making is that acceptance of a decision depends on the
making has two important effects on subsequent participation. pattern of participation in a decision process. In introducing a
On the one hand, participation is very likely to b; frusüating. new technolog¡r or changing a standard operating procedure,
Although there are some tendencies to develop illusions of de- those who participated in the decision are more likely than non-
cision effectiveness in order to confirm the expectations that participants to believe in the correctness and efficacy of the new
originally stimulated participation, the observable conse- technolog¡r or procedure.
quences of participating in a decision process are likely to be The argument extends to the more general legitimacy of de-
less than anticipated. The social nature of decision making ren- cision processes. At least within the ethos of Western demo cta-
ders personal influence hard to determine, and the complexities cy, decision processes gain legitim acy through a sense of in-
of social causation make the ultim ate effects of a decision hard volvement in them. The involvement need not be formally
to predict. Circulation among decision makers is often driven democratic, indeed it normally is not, but it includes a sense of
by outsiders believing they can do better than current decision being consulted, of having one's opinion's heard, and of confi-
makers, fighting for the right to participate, becomittg dis- clence that the decisions, in some sense, "represent" attention
abused of their ambitions as a result of experience at the frus- Ltl one's concerns.
trations of decision making, and abandoning the field to the The link between participation and decision legitimacy has
next wave of crusaders seeking access to the pior"rr. lccl to a host of managerial tactics for providing the illusion of
168 A pRTMER oN DECrsroN MAKTNc Multiple Actors: Cr¡r{lic.t uncl l\ilitit:l; I ó9

involvement in decision making without genuine influence. over policies and then, using the same personnel, achieve an ef-
Meetings are held to "solicit input" on decisions already made, fective implementation of a policy that earlier was vigorously
or long before any issues are clear enough to frame meaningfui opposed by a substantial minoúty?
alternatives. Personal interactions are laced with comments Implementation difficulties can be explained partly by prob-
confirming the importance of suggestions made. This tactical lems of incomplete and unshared information. Most decision
corruption of participation has, in turn, led to considerable sus. makers are neither omnipresent nor omniscient; they have lim-
picion of participation initiatives. Whatever attraction there ited abilities to attend to all events and limited knowledge by
might be in involvement is compromised by doubt that the in- which to interpret observable actions. Because of these limita-
volvement is genuine. tions, differences might be expected between the decisions of
The general consequence is likely to be a considerable deval- policy makers and the decisions taken by those responsible for
uation of participation. For the reasons outlined above, deci- implementing the policy, even in the absence of inconsistencies
sion impact is difficult to establish in the best of circumstances. in preferences and identities.
One person's effects on a decision ate lost in the effects of oth- Inconsistencies complicate the story by introducing conflict.
ers' and one decision's effects are lost in the general confusions Decision makers, their allies, and their opponents seek to rene-
of history. When those ambiguities are increásed by suspicions gotiate policies and practices after they are "decided." Policies
of the process, experience is likely to teach that paiticipátion ip announced by policy makers are opportunities for others to pur-
a fraud and a waste.
sue their own visions, and those responsible for implementing
poliry will usually have reasons for pursuing preferences or iden-
4.3.2 The Implementation of Decisions tities that ate different from those pursued by policy makers.
The problems are endemic and have been extensively dis-
One of the more persistent problems in multiple-actor decision cussed in the literature on or ganizations, &s well as in treatises
making is the problem of implementation. Foi decisions to have on optimal contracts, incentive schemes, and theories of
effects, they must be implemented. The study of org anizations agency. Decision processes are particularly vulnerable to imple-
has repeatedly examined the way polici"r uná programs adopt-
mentation problems when it is not possible to specify complete
ed by boards of directors, Iegislatures, or top managements are coalition agreements at the time of making a decision. Poliry
subsequently executed, modified, and elaborated Uy ttrose who makers cannot anticipate all of the contingencies that wilt occur
implement them. Decision histories abound with ,ur", of unim- in the course of implementation. Without discretion on the part
plemented, partially implemented, or exotically implemented of administrators, implementation is likely to sacrifice intelli-
actions. gence to stand ardízation. Decision makers know this and ac:
cept that implementation of decisions calls for local informa-
tion and locál expertise. However, most of the procedures that
Constructing appropriate mechanisms to induce administrators facilit ate intelligent administrative elaboration of policy also in-
to execute policy decisions is one of the central concerns of or- vite creative administrative deviations.
ganization theory. Ffow can administrative identities be con-
structed so that the personal preferences of an administrator do INSTABILITIES OF COALITIONS
not twist a policy decision? How can a decision process first In order to implement policies, coalitions require stability
stimulate disagreement and factionalisrn through open debate across time. Temporal stability is continually underrnined by
170 A pr{rMrin oN l)ricrsroN MAKTNc
Multiple Ac,lrtr,r. Cortflict und l\ililit.,y 17 I

problems of attention-it is not easy to sustain the attention

of irchieve consensus or mutual indifference leads to ambiguous
all coalition members against alternative claims on their time
and energy. Decision making often elicits attention that ¡rolicies (e.g . party platforms in national elections). Ambiguous
is not policies allow for a healthy level of selective interpretation on
sustained once a choice is made. Implementation also
requires the part of not-quite-natural allies. 'When ambiguities are clari-
stability across institutions. Since th; pofitical forces active
in fied in the course of implernentation, the coalition tends to fall
decision making ate not always the same as the political
forces itpart.
active in implementation, institutional stabiliry ñ problematic.
The politi calforces active in one part of an otgun ization 2" Outcome optimism. Forming a winning coalition alrnost al-
are dif- ways leads coalition members to overestimate the positive
ferent from the political forces active in another part. The polit.
ical reality facing a police officer on the street áuring a consequences to be expected. There are both motivational and
riot is structural reasons for this. Structurally, since programs ate
different from the politi calreality facing a city councit debating
police policy. These complications of coalitions across rarely adopted when expectations are erroneously pessimistic,
time and the sample of programs actually adopted is more likely to ex-
space are not simply pathologies of social systems. They
reflect hibit errors of overoptimism than of overpessimism. Winning
some important aspects of how those systems manage
to survive coalitions will suffer from the winner's curse. Policies ate
and thrive.
Forming a coalition in order to suppo rt apolicy, whether adopted because their probable return has been exaggerated.
in a This is not because decisions are made foolishlyi but precisely
legislature or in a board room, involu.r standurd techniques
of . because decision makers estim ate expected retürn ,rrrd., con-
horsetrading, persuasion, bribes, threats, and management
of ditions of uncertainty, and expectations for the alternative
information. Those are the conventional procedures of
discus- with the highest expected return will (otr average) be unduly
sion, politics, and policy formation. They ut" well conceived
to optimistic.
help participants form coalitions, exploie support for alterna-
tive policies, and develop a viable policy. Much of the genius The structural exaggeration of expectations is exacerbated by
of the social dynamics of coalition formation. In order to build a
modern org anizational leadership lies in skills at producing
pol- coalition, coalition members will system atically exaggerate the
icy from the conflicting and incño ate ideas, demands, préóorr-
ceptiors, and prejudices of the groups to which organ .g¡ational probable good effects of the policy and will rystem atically un-
leadership must attend. derestimate the probable negative effects. Policies will, on aver-
At the heart of several of these techniques for achieving poli- &ge, be oversold.
cies, however, are three features of decision making Both structural and social sources of overestimation lead, on
thaimake averQge, to post decision disappointment on the part of at least
coalitions unstable and thereby make implementatiott particu*
larly problematic: some coalition members. Thus, great hopes lead to action, but
great hopes are invitations to disappointment. This in turn
I. Decision ambiguity. Ambiguity is frequently an advantage leads both to an erosion of support and to an awareness of "fail-
in the development of a coalition to support a decisio¡. It is ures of implementation." As the policy is revealed to provide
easier to conceal or ignore disagreements if policies fewer payoffs to coalition members than they anticipated, the
are*tiit* coalition tends to fall apart.
with provisions or terminology that can be interpreted differ-
ently by different people. rn assembling a coalition to 3. Support exaggeration Few major policies could be adopted
support a without some supporters for whom the policy is relatively unim-
policy, it is often necessary to make thé terms of the
ugtéément portant except as a politicalbargain. They may be persuaded to
unclear in order to hide or suppress conflicts. The effort
to ioin a coalition by claims of loy alty or friendship, or by a logroll
Muhiple Actors: Conflict and politics 173
in which their support is traded for support on other issues. In,.,
addition, a prominent feature of decision making is that indi,.u lcrt portrait, most students of decision
making in economic,
viduals political, military educational,
d groups enter a decision atena for a variety of rea, religious, and other sociar insti_
?t Ir¡rions see invorving
sons, only some of which are concerned with the content of muttipte actors with inconsistent
decisioll. A typical coalition includes members who support thg ¡rrcf'crences and identities. .itrr"ow designed
decision primarily ^^ ¡l- ' t tl
.rr r ¡ v'lving consistent,indiüdual pr"tÉr"n."s for situations in_
^-i*^-:I-- so they will be recorded as a supporter-i and identities has
sccmed less than adequate
For them, the decision has syrnbolic significance, but iti lmple. for deafing with such situation-sl
mentation does not. As history moves from the adoption óf students of decisionmaking t uu"
* ,""og ¡izedthe difficulties,
decision to its implernentation, the coalition tends to fail apart, lhough many of them have pi"i"tt"o
to-maintai'u ii"ory or
,i rrrrltiple actors thatconsiso ürt"tt
*ideas about howmurtiple
...e i
¡rctors could be induced to
4.4 single Actors and Multiple Actors
fo-rm consistent pr"¡";;;s

t, . irlontities, to be a "team-" wirttL.r*sing and

awareness of the ex-
Students of decision making usually insist on a distinction b0., tent to which multiple-actor
decision making proceeds without
et rn sistency-througll. -
tween single-actor, or individual, decision making, on the ong' more,, political,, proi"rr", f", _uLing
hand, and multiple-actor, or organ izational, decision *ut in;,, tlccisions without aehieving ugr""Á".r,
on preferences or con_
sistency in identities-has
on the other. The distinctions are reflected in the differ*ncJi,i ó-"" in.i"ur"¿ reluctance to use sin_
between Chapters 3 and 4, on the one hand, and the two chap.*: ¡¡lr:-actor theories to compreh"no
What has become somewhat less -utiiple-actor situations.
ters preceding them, on the other. They stem from some unddr-_i clóar is *fr"ifr*
dcsigned for multip_I" *rorS
standable resistance on the part of students of org anizations to,,
hcnd individual deóisio" -igii, i' fuct, come ,o "'rt "ory
characterizing organizations as having the kinds áf pteferenri'i -ut ing"b;ter than theories based on
individual consistency. It is a
and identity consistency assumed in most theories of individual soletim
decision making.
j' ind ividuat s;-;*r consistenr o**:il. JJiIT#J,?[::,JXI
There is a history. Economic theories of markets and politi. i clcarly suc! u porlruyur is oftln I substantiar fiction.
processes of individuar The
cal theories of international relations, for example, were origi., choice often seem to be as fifled
with in_
nally theories of individuals. Individual entrepreneurs met l, ll,_l:-:ry:
as the processes of collective choice. Insofar as
ls true, it is possible (though
individualworkersandcustomerSinmarkets.mbividualSoVof. by no means assured) that individ_
,irl processes of choice án ú" uná"trtoo¿
eigns, or their agents, met other sovereigns in international in. ,

r.rms better than they can be

in organizational
teractions. It seemed natural to treat such decision makers ut understood in the classic individ-
u¿rlistic terms.
individuals. As those decision makers grew into large, compli.
cated corporations, in one case, and large,complicatéd political such an argument, however, is
somewhat beyond predomi-
rr¿tnt current thinking. It
institutions, in the other, thatframework was tátuitred, a fiction is clear that*urry il;;; thathave
fi'trm observations oi org anizational come
reinforced by a parallel legal fiction transforming such corpo. decision making have be-
rate bodies into imagi nary beings with many of tñe legal rights eome stand ard' in treatments
of individual decision making, but
I hq converse is
and properties of individuals. also clear. As this book perhaps
f irrn' many of
serves to con-
IJltim ately, it was inevitable that the internal coherenóe of the ideas about decision making
sitlns of individual choice found in discus-
corporate actors would be challenged, as it has been in several are also found in dilcussions
of orga-
domains of thought, including theóries of decision making. Al- rlizational choice. Ideas are borrowed
back and forth with,
though popular journalism sometimes presents a more consis- ct)nsiderable ease and, on the
whole, effectiveness. Many of the
concepts and processes discussed
in chapters-i and z,f tlris

174 A pRTMER 0N DrcrsroN MAKTNc

book can be slid into Chapters 3 and 4, andvice versa. fnco

tency of prefere-nces and identities has turned out to be,
importaút ractor in decision making, b:, it not tp
feli-1bly connectád to tn" ái¡ri*ri¿¡'L¿**op single acrors g