The New Institutionalism: Organizational Factors in Political Life

Author(s): James G. March and Johan P. Olsen

Source: The American Political Science Review , Sep., 1984, Vol. 78, No. 3 (Sep., 1984), pp.
734-749
Published by: American Political Science Association

Stable URL: http://www.jstor.com/stable/1961840

REFERENCES
Linked references are available on JSTOR for this article:
http://www.jstor.com/stable/1961840?seq=1&cid=pdf-
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

American Political Science Association and Cambridge University Press are collaborating with
JSTOR to digitize, preserve and extend access to The American Political Science Review

This content downloaded from

132.174.250.76 on Tue, 18 Aug 2020 13:17:37 UTC
All use subject to https://about.jstor.org/terms
 The New Institutionalism:
Organizational Factors in Political Life

JAMES G. MARCH
Stanford University

JOHAN P. OLSEN
University of Bergen

Contemporary theories of politics tend to portray politics as a reflection of society, political

phenomena as the aggregate consequences of individual behavior, action as the result of choices based
on calculated self-interest, history as efficient in reaching unique and appropriate outcomes, and
decision making and the allocation of resources as the central foci of political life. Some recent
theoretical thought in political science, however, blends elements of these theoretical styles into an
older concern with institutions. This new institutionalism emphasizes the relative autonomy of
political institutions, possibilities for inefficiency in history, and the importance of symbolic action to
an understanding of politics. Such ideas have a reasonable empirical basis, but they are not charac-
terized by powerful theoreticalforms. Some directions for theoretical research may, however, be iden-
tified in institutionalist conceptions of political order.

In most contemporary theories of politics, coherent or consistent; it is not completely

traditional political institutions, such as the legis- legitimate; but neither can it be entirely ignored.
lature, the legal system, and the state, as well as This resurgence of concern with institutions is a
traditional economic institutions, such as the cumulative consequence of the modern transfor-
firm, have receded in importance from the posi- mation of social institutions and persistent com-
tion they held in the earlier theories of political mentary from observers of them. Social, political,
scientists such as J.W. Burgess or W.W. and economic institutions have become larger,
Willoughby, economists such as Thorstein Veblen considerably more complex and resourceful, and
or John R. Commons, and sociologists such as prima facie more important to collective life.
Max Weber. From a behavioral point of view, Most of the major actors in modern economic and
formally organized social institutions have come political systems are formal organizations, and
to be portrayed simply as arenas within which the institutions of law and bureaucracy occupy a
political behavior, driven by more fundamental dominant role in contemporary life.
factors, occurs. From a normative point of view, Attention to political institutions has increased
ideas that embedded morality in institutions, such in the literature on legislatures (Shepsle & We-
as law or bureaucracy, and that emphasized ingast, 1983), budgets (Padgett, 1981), public
citizenship as a foundation for personal identity, policymaking (Ashford, 197-; Scharpf, 1977),
have given way to ideas of moral individualism local government (Kjellberg, 1975), and political
and an emphasis on conflicting interests. elites (Robins, 1976). It is manifest in studies of
In recent years, however, a new institutionalism the origin of the state (Wright, 1977) and the
has appeared in political science. It is far from development of national administrative capacity
(Skowronek, 1982), in analyses of the breakdown
of democratic regimes (Potter, 1979), and in
discussions of corporatism (Berger, 1981; Olsen,
Received: October 6, 1983 1981; Schmitter & Lehmbruch, 1979). It is
Accepted for publication: December 6, 1983 reflected in the Marxist rediscovery of the state as
This research has been supported by grants from the a problem in political economy (Jessop, 1977) and
Norwegian Research Council for Science and of the importance of organizational factors for
Humanities, the Norwegian Ministry of Consumer Af- understanding that role (Therborn, 1980). It is
fairs and Government Administration, the Mellon present in studies of formal organizations and
Foundation, the Spencer Foundation, the Stanford
particularly in studies of the place of such
Graduate School of Business, and the Hoover Institu-
organizations in the implementation of public
tion. We are grateful for comments by Julia W. Ball,
Michael D. Cohen, Stephen D. Krasner, Martin
policy (Hanf & Scharpf, 1978). It is visible in at-
Landau, Todd LaPorte, W. Richard Scott, and William tempts to link the study of the state to natural
Siffin. science (Masters, 1983) and to the humanities

734

This content downloaded from

132.174.250.76 on Tue, 18 Aug 2020 13:17:37 UTC
All use subject to https://about.jstor.org/terms
1984 Organizational Factors in Political Life 735

(Geertz, 1980), as well as in a renewed interest in social stratification of a modern society with its
making historical-comparative studies of the state associated distribution of wealth and income has
(Evans, Rueschemeyer, & Skocpol, 1983; obvious major effects on political events. Class
Hayward & Berki, 1979; Krasner, in press). differences translate into political differences with
In this article we examine some aspects of these great reliability across time and across cultures;
developments and their implications for develop- differences in the organization and ideology of
ing a theoretical understanding of how political social class seem to lead to predictable differences
life is organized. We approach the task from the in political organization and institutions (Tilly,
perspective of students of formal organizations. 1978). Other analyses at the same level of aggrega-
The argument, however, extends beyond organi- tion make the structure and process of politics a
zation theory to a more general view of the place function of physical environment, geography, and
of institutions in politics and the possibilities for a climate; of ethnicity, language, and culture; of
political theory that is attentive to them. economic conditions and development; or of
demography, technology, ideology, or religion.
Theoretical Styles of Contemporary Plausible arguments which make political life a
Political Science derivative of one or more of these broad con-
textual forces have been developed, and it is not
Although the concept of institution has never hard to find empirical data to support the
disappeared from theoretical political science, the arguments. Although there are a number of
basic vision that has characterized theories of relatively precise contextual theories, the major
politics since about 1950 is (a) contextual, inclined theoretical significance of these ideas from the
to see politics as an integral part of society, less in- present point of view is less the specific forms of
clined to differentiate the polity from the rest of the theories than their general inclination to see
society; (b) reductionist, inclined to see political the causal links between society and polity as run-
phenomena as the aggregate consequences of in- ning from the former to the latter, rather than the
dividual behavior, less inclined to ascribe the out- other way around. It is assumed that class, geog-
comes of politics to organizational structures and raphy, climate, ethnicity, language, culture,
rules of appropriate behavior; (c) utilitarian, in- economic conditions, demography, technology,
clined to see action as the product of calculated ideology, and religion all affect politics but are
self-interest, less inclined to see political actors as not significantly affected by politics.
responding to obligations and duties; (d) func-
tionalist, inclined to see history as an efficient The Macro Consequences of Micro Behavior:
mechanism for reaching uniquely appropriate Reductionism
equilibria, less concerned with the possibilities for
maladaptation and non-uniqueness in historical Historically, political theory has treated politi-
development; and (e) instrumentalist, inclined to cal institutions as determining, ordering, or modi-
define decision making and the allocation of fying individual motives, and as acting autono-
resources as the central concerns of political life, mously in terms of institutional needs. In con-
less attentive to the ways in which political life is trast, substantial elements of modern theoretical
organized around the development of meaning work in political science assume that political phe-
through symbols, rituals, and ceremonies. nomena are best understood as the aggregate con-
sequences of behavior comprehensible at the in-
Politics as Subordinate to Exogenous Forces: dividual or group level.
Contextualism Such theories depend on two presumptions.
The first presumption is that a political system
Historically, political scientists and political consists of a number (often a large number) of
philosophers have tended to treat political institu-elementary actors. Human behavior at the level of
tions, particularly the state, as independent fac- these elementary actors may be seen as conscious,
tors, important to the ordering and understanding calculated, and flexible, or as unconscious,
of collective life (Heller, 1933). Modern political habitual, and rigid. In either case, the preferences
scientists, with few exceptions, have not. The and powers of the actors are exogenous to the
state has lost its position of centrality in the political system, depending on their positions in
discipline; interest in comprehensive forms of the social and economic system. The second
political organization has declined; political presumption is that collective behavior is best
events are defined more as epiphenomena thanunderstoodas as stemming from the (possibly in-
actions necessary to an understanding of society; tricate) interweaving of behavior understandable
politics mirrors its context (Easton, 1968). at a lower level of aggregation. Discovering, or
The most conspicuous contextual factor cited in deducing, the collective consequences may be dif-
recent writing is the social class structure. The ficult, even impossible; but the central faith is that

This content downloaded from

132.174.250.76 on Tue, 18 Aug 2020 13:17:37 UTC
All use subject to https://about.jstor.org/terms
736 The American Political Science Review Vol. 78

outcomes at the collective level depend only on competition. Within this theory we find individual
the intricacies of the interactions among the in- species, each adapting to an environment through
dividual actors, that concepts suggesting autono- survival, mutation, and reproduction. Selection
mous behavior at the aggregate level are certainly and changes in population distributions within the
superfluous and probably deleterious. environment are assumed to be understandable as
Within such a perspective, for example, the consequences of the actions of individual actors
behavior of an organization is the consequence of that, in combination with the actions of others
the interlocking choices by individuals and sub- and the potential capacity of the environment,
units, each acting in terms of expectations and produce a distribution of types.
preferences manifested at those levels (Niskanen,
1971). The behavior of a market is the conse- Action as the Making of Calculated Decisions:
quence of the interlocking choices by individuals Utilitarianism
and firms, each acting in terms of a set of expecta-
tions and preferences manifested at those levels Historically, political science has emphasized
(Stigler, 1952). It is not necessary that the micro the ways in which political behavior was embed-
processes involve choice, of course. Aggregate ded in an institutional structure of rules, norms,
behavior in a group can be defined as the conse- expectations, and traditions that severely limited
quence of the interlocking of trial-and-error learn- the free play of individual will and calculation
ing occurring at the individual level (Lave & (Wolin, 1960). In contrast, modern political sci-
March, 1975). Or the aggregate behavior of an in- ence has, for the most part, described political
dustry can be defined as the consequence of the events as the consequence of calculated decisions.
interlocking of standard operating procedures and Not just in political science, but throughout mod-
accounting rules followed at the level of the in- ern theoretical work in the social sciences, the pre-
dividual firm (Nelson & Winter, 1982). eminent vision of human behavior is a vision of
There is nothing intrinsic to a perspective that choice. Life is characterized as deliberate decision
emphasizes the macro consequences of micro ac- making.
tions which requires that the elementary units be The details of the choice metaphor vary from
individuals. All that is required is that the one treatment to another, but the characteristic
behavior of a more comprehensive system be form is one that assumes choices stem from two
decomposable to elementary behaviors explicable guesses about the future. The first is a guess about
at a less comprehensive level. In practice, the uncertain future consequences of possible cur-
however, in most of the social sciences, the ac- rent action. Decision theorists recognize that
tions of individual human beings are considered human limitations may restrict the precision of
to determine the flow of events in a larger social the estimates, that the estimates may be biased,
system. Outcomes at the system level are thought and that the information on which the estimates
to be determined by the interactions of individuals are based may be costly; but information about
acting consistently in terms of the axioms of in- probable consequences is assumed to be impor-
dividual behavior, whatever they may be. Thus, tant to a choice. From this assumption comes an
we make assumptions about individual consumers emphasis on the power of information and exper-
to understand markets, about voters to under- tise (Crozier, 1964) and the importance of reliable
stand politics, and about bureaucrats to under- and unbiased information sources (Nisbet & Ross,
stand bureaucracies. 1980). Although numerous psychological experi-
The two best-specified theories of aggregate ments have indicated that the guesses of human
behavior in the social sciences, the economic subjects are biased (Kahneman, Slovic, & Tver-
theory of markets and the ecological theory of en- sky, 1982), it has not been easy to formulate alter-
vironmental competition, exemplify the modern natives to the simple notion that the guesses of ex-
style. Consider the theory of markets. Within this perienced humans are, on average, accurate. As a
theory we find individual consumers, each at- result, most theories of choice present decisions as
tempting to make purchases at the best possible being, on average, sensible. In their political ver-
prices considering his or her own preferences and sions, choice theories assume that, on average,
alternatives, and individual producers, each at- voters vote intelligently with respect to their in-
tempting to make production and pricing deci- terests; legislators organize sensible coalitions,
sions that result in the best possible return con- given their interests; and nation states voluntarily
sidering his or her own preferences and alter- enter alliances that, on average, improve theii
natives. The behavior of the market is assumed to positions.
be understandable as a consequence of these in- The second guess on which intentional, antic
dividual actors making choices that, in aggregate, ipatory choice is based is a guess about a decision
fit together into market phenomena. Consider maker's uncertain future preferences for possible
similarly the ecological theory of environmental future outcomes. In any theory of deliberate

This content downloaded from

132.174.250.76 on Tue, 18 Aug 2020 13:17:37 UTC
All use subject to https://about.jstor.org/terms
1984 Organizational Factors in Political Life 737

choice, action depends on the decision maker's ciency is a standard, although usually not explicit,
values. Since the consequences of interest are to assumption of much of modern social science.
be realized in the future, it is necessary to an- Economic theories of markets and ecological
ticipate not only what will happen but how the theories of competition, for example, are con-
decision maker will feel about those outcomes cerned with the characteristics of an equilibrium,
when they are experienced (March, 1978). The if one exists. They are used to predict differences
compexities of the second guess are largely ig- (e.g., in markets, organizational structures,
nored by theories of choice. In their standard population, technologies) that will be observed, at
forms, the theories assume that preferences are equilibrium, in different environments. Similarly,
stable, thus that current preferences are good some postwar theories of political parties see
predictors of future preferences; that preferences party orientation and organization as equilibrium
are unambiguous and consistent, thus that a solutions to problems of survival in a competitive
choice will be clearly indicated, given the first political environment (Downs, 1957). The as-
guess; and that preferences are exogenous, thus sumption of historical efficiency makes such
that whatever process generates preferences, it theories largely indifferent to the behavioral real-
precedes choice and is independent of the choice ity of the micro processes that are assumed. For
process. In one of the best-developed forms of example, competition can be assumed to eliminate
choice theories, these assumptions about prefer- action that is inconsistent with the logic of sur-
ences are taken as axioms, and preferences are vival. Examples include theories of market
discovered not by asking decision makers to equilibria, such as those found in recent ideas of
report them but by defining a "revealed prefer- efficient capital markets (Sharpe, 1970); theories
ence" function that satisfies the axioms and is of organizational structure, such as those found in
consistent with choices made by a decision maker recent ideas of industrial organization (William-
(Luce & Raiffa, 1957). Although the empirical ex- son, 1978); and theories of political parties, such
istence of consistent revealed preferences has been as those found in ideas of political economy
the subject of considerable debate (Becker & (Olson, 1965).
Stigler, 1977; Kahneman, Slovic, & Tversky, History cannot be guaranteed to be efficient.
1982), the theoretical idea forms the basis of ex- An equilibrium may not exist. Even if there is an
tensive analytical development and empirical ex- equilibrium, historical processes can easily be
ploration. slow enough relative to the rate of change in the
environment that the equilibrium of the process is
The Efficiency of History: Functionalism unlikely to be achieved before the environment,
and thus the equilibrium, changes. By assuming
Historically, political theory has been am- quickness, theories of political behavior avoid
bivalent about the efficiency of history. Like focusing on transient phenomena that might be
other social scientists, students of political less predictable and more subject to effects from
development have been inclined to accept an idea the details of the processes involved. For example,
of progress, the more or less inexorable historical when it is predicted that political parties will come
movement toward some more "advanced" level. to identical positions in an environment of single-
At the same time, political histories have often peaked voter preferences, it is assumed that party
emphasized the unique significance of a particular adjustment will be much more rapid than will be
sequence of events or choices, the impact of a par- changes in voter preferences. Efficiency also re-
ticular campaign strategy or speech, or the par- quires that the equilibrium be unique and achiev-
ticular tactics of international negotiation. In able. Processes with multiple equilibria are, of
modern usage, the terminology of progress has course, easily specified and frequently observed.
been largely replaced by a terminology of sur- What makes them unattractive is not their rarity,
vival, but for the most part, in contemporary but their intractability and the indeterminacy of
theoretical political science, institutions and their outcomes. It is no accident that the most
behavior are thought to evolve through some common principle of theories in the social sciences
form of efficient historical process. is the optimization principle, and that one of the
An efficient historical process, in these terms, is
greatest concerns in such theories is showing that
one that moves rapidly to a unique solution, con- a process has a unique optimum that is guaranteed
ditional on current environmental conditions, to be achieved.
thus independent of the historical path. The
equilibrium may involve a stochastically stable The Primacy of Outcomes: Instrumentalism
distribution or a fixed point, but we require a
solution that is achieved relatively rapidly and is Historically, theories of political institutions
independent of the details of historical events portrayed political decision making primarily as a
leading to it. The assumption of historical effi- process for developing a sense of purpose, direc-

This content downloaded from

132.174.250.76 on Tue, 18 Aug 2020 13:17:37 UTC
All use subject to https://about.jstor.org/terms
738 The American Political Science Review Vol. 78

tion, identity, and belonging. Politics was a ve- back to considerations that typified earlier forms
hicle for educating citizens and improving cultural of theory in political science. We do not mean to
values. Although there are exceptions, the suggest, however, that the new and the old are
modern perspective in political science has identical. It would probably be more accurate to
generally given primacy to outcomes and either ig- describe recent thinking as blending elements of
nored symbolic actions or seen symbols as part of an old institutionalism into the non-institution-
manipulative efforts to control outcomes, rather alist styles of recent theories of politics.
than the other way around. This new institutionalism can be presented and
Modern polities are as replete with symbols, discussed as an epistemological perspective of
ritual, ceremony, and myth as the societies profound importance to understanding social
more familiar to anthropological tradition. Politi- science, but for our purposes, it is more useful to
cians announce public support for positions they define it in terms of a narrow collection of
fail to defend in private (Edelman, 1964). Legis- challenges to contemporary theoretical thinking in
lators vote for legislation while remaining indif- political science, a small set of relatively technical
ferent to its implementation (Pressman & Wildav- ideas of primary interest to professional students
sky, 1973). Administrators solicit public par- of political life. The ideas deemphasize the
ticipation in decision making in order to secure dependence of the polity on society in favor of an
public support for policies to which they are interdependence between relatively autonomous
already committed. Chief executives advocate social and political institutions; they deemphasize
reorganization of the public bureaucracy, an- the simple primacy of micro processes and effi-
nounce plans for making reorganizations, and cient histories in favor of relatively complex pro-
regularly abandon the plans (March & Olsen, cesses and historical inefficiency; they deempha-
1983). Information is gathered, policy alternatives size metaphors of choice and allocative outcomes
are defined, and cost-benefit analyses are pur- in favor of other logics of action and the centrality
sued, but they seem more intended to reassure of meaning and symbolic action. The ideas are not
observers of the appropriateness of actions being all mutually consistent. Indeed, some of them
taken than to influence the actions (Feldman & seem mutually inconsistent. For example, ideas
March, 1981). based on the assumption that large institutional
In modern discussions of politics, these sym- structures (e.g., organizations, legislatures,
bolic actions are characteristically portrayed as states) can be portrayed as rationally coherent
strategic moves by self-conscious political actors. autonomous actors are uneasy companions for
Rituals and ceremonies are defined as window- ideas suggesting that political action is inade-
dressing for the real political processes, or as in- quately described in terms of rationality and
struments by which the clever and the powerful choice.
exploit the naive and the weak. The hiring of ex-
perts lends legitimacy to policies (Meyer & The Causal Position of Political Institutions
Rowan, 1977); associating unpopular moves with
popular symbols is reassuring (Edelman, 1964). Without denying the importance of both the
Control over symbols is a basis of power, like con- social context of politics and the motives of in-
trol over other resources (Pfeffer, 1981a); and the dividual actors, the new institutionalism insists on
use of symbols is part of a struggle over political a more autonomous role for political institutions.
outcomes (Cohen, 1974). The state is not only affected by society but also
affects it (Katzenstein, 1978; Krasner, 1978;
Institutionalist Perspectives Nordlinger, 1981; Skocpol, 1979; Stephan, 1978).
Political democracy depends not only on
The new institutionalism is not peculiar to economic and social conditions but also on the
political science. Renewed interest in institutions design of political institutions. The bureaucratic
is characteristic of recent trends in economics, agency, the legislative committee, and the ap-
which has discovered law, contracts, hierarchies, pellate court are arenas for contending social
standard operating procedures, professional forces, but they are also collections of standard
codes, and social norms (Akerlof, 1980). It is also operating procedures and structures that define
seen in anthropology and sociology, although and defend interests. They are political actors in
non-institutionalist visions never succeeded in their own right.
those fields to the extent that they did in political The argument that institutions can be treated as
science and economics. Nor are the institutionalist political actors is a claim of institutional
ideas entirely new. By labeling the collection of coherence and autonomy. The claim of coherence
ideas "the new institutionalism," we mean to is necessary in order to treat institutions as deci-
note the fact that there was indeed an "old institu- sion makers. From such a point of view, the issue
tionalism," that cycles in ideas have brought us is whether we wish to picture the state (or some

This content downloaded from

132.174.250.76 on Tue, 18 Aug 2020 13:17:37 UTC
All use subject to https://about.jstor.org/terms
1984 Organizational Factors in Political Life 739

other political institution) as making choice on the fecting tastes. Similarly, conventional theories of
basis of some collective interest or intention (e.g., politics assume that a voter's exposure to and
preferences, goals, purposes), alternatives, and choice of a candidate do not change that voter's
expectations (Levi, 1981). There is no necessary preferences for various attributes that a candidate
answer to the question unless we impose one. might possess, although they may change a voter's
Whether it makes pragmatic theoretical sense to beliefs about which candidates possess which at-
impute interests, expectations, and the other tributes. The new institutionalism, in company
paraphernalia of coherent intelligence to an in- with most research on preferences, argues that
stitution is neither more nor less problematic, a preferences and meanings develop in politics, as in
priori, than whether it makes sense to impute the rest of life, through a combination of educa-
them to an individual (Kahneman, 1982; March & tion, indoctrination, and experience. They are
Shapira, 1982). The pragmatic answer appears to neither stable nor exogenous (Cohen & Axelrod,
be that the coherence of institutions varies but is in press). If political preferences are molded
sometimes substantial enough to justify viewing a through political experiences, or by political in-
collectivity as acting coherently. stitutions, it is awkward to have a theory that
The claim of autonomy is necessary to establish presumes preferences are exogenous to the
that political institutions are more than simple political process. And if preferences are not ex-
mirrors of social forces. Empirical observations ogenous to the political process, it is awkward to
seem to indicate that processes internal to political picture the political system as strictly dependent
institutions, although possibly triggered by exter- on the society associated with it.
nal events, affect the flow of history. Programs The contrast between the two kinds of notions
adopted as a simple political compromise by a is found most starkly in theories of political
legislature become endowed with separate mean- leadership. One classic idea of political leadership
ing and force by having an agency established to emphasizes the creation of winning political coali-
deal with them (Skocpol & Finegold, 1982). The tions among participants with given demands
establishment of public policies, or competition (March, 1970). The leadership role is that of a
among bureaucrats or legislators, activates and broker: providing information, identifying pos-
organizes otherwise quiescent identities and social sible coalitions, and facilitating side-payments
cleavages (Olsen & Saetren, 1980; Tilly, 1978). and the development of logrolls. Such a view of
Policy experts within the political system develop leadership is implicit in the theory of the political
and shape the understanding of policy issues and process that has been developed in political
alternatives (Heclo, 1974). science in recent decades. A second conception of
Such phenomena are not routinely accom- leadership emphasizes the transformation of
modated by modern political theory, which makes preferences, both those of the leader and those of
political outcomes a function of three primary the followers (Burns, 1978; Selznick, 1957).
factors: the distribution of preferences (interests) Leaders interact with other leaders and are co-
among political actors, the distribution of opted into new beliefs and commitments. The
resources (powers), and the constraints imposed leadership role is that of an educator, stimulating
by the rules of the game (constitutions). Each of and accepting changing worldviews, redefining
these is treated as exogenous to the political meanings, stimulating commitments. Such a view
system. That is, preferences are developed within is more conspicuous in the ideas of the new in-
a society and transmitted through socialization, stitutionalism.
resources are distributed among political actors by The distribution of political resources is also
some broad social processes, and rules of the partly determined endogenously. Political institu-
game are either stable or change by a revolu- tions affect the distribution of resources, which in
tionary intervention exogenous to ordinary turn affects the power of political actors, and
political activities. thereby affects political institutions. Wealth,
The idea that preferences are produced and social standing, reputation for power, knowledge
changed by a process that is exogenous to the pro- of alternatives, and attention are not easily
cesses of choice is fundamental to modern deci- described as exogenous to the political process
sion theory. In the "revealed preference" version and political institutions. Holding office provides
of the theory, preferences must be stable in order participation rights and alters the distribution of
for the theory to be testable. In other versions, power and access (Egeberg, 1981; Laegreid &
preferences can change, but choice itself does not Olsen, 1978). The policy alternatives of leaders
produce a change in preferences. Conventional are not defined completely by exogenous forces,
theories of markets, for example, picture advertis- but are shaped by existing administrative agencies
ing and experience as providing information (Skocpol, 1980; Skocpol & Finegold, 1982;
about alternatives and their properties, not as af- Skowronek, 1982). The outcomes of the political

This content downloaded from

132.174.250.76 on Tue, 18 Aug 2020 13:17:37 UTC
All use subject to https://about.jstor.org/terms
740 The American Political Science Review Vol. 78

process modify reputations be rash to

for assume that
power, errors in in
which expectations have
turn modify political outcomes (Enderud, 1976; a normal distribution with a mean of zero. The
March, 1966). allocation of attention may be critical to the flow
Finally, the third exogenous factor in conven- of events. The responsiveness of the political
tional theories of politics, the rules of the game, is system to environmental pressures may, at least in
not really exogenous either. Constitutions, laws, the short run, depend on the amount of slack in
contracts, and customary rules of politics make the system, and on the ways in which accounting
many potential actions or considerations il- numbers are produced and fudged. The system
legitimate or unnoticed; some alternatives are ex- may not come close to trying to resolve conflict
cluded from the agenda before politics begins but simply attend sequentially to the demands
(Bachrach & Baratz, 1962), but these constraints placed on it (Cyert & March, 1963). Learning may
are not imposed full-blown by an external social be superstitious, and fallacious rules of inference
system; they develop within the context of may persist for long periods (Nisbet & Ross,
political institutions. Public agencies create rules 1980). At the limit, the connections between prob-
and have them sanctioned by politicians (Eckhofflems and solutions may be less dominated by a
& Jacobsen, 1960), and revolutionary changes are logic of causal linkages between means and ends
initiated and pursued by military bureaucrats than by the less problematic temporal linkages of
(Trimberger, 1978). simultaneity (Cohen, March, & Olsen, 1972).
Theories of collective behavior most commonly
The Causal Complexity of Political History simplify the potential morass of collective com-
plexity by one of two classic routes. The first is
Theories of politics tend to assume a relatively statistical aggregation. In its usual guise, aggrega-
uncomplicated intermeshing among the elemen- tion assumes that the factors affecting outcomes
tary units of a political system. There may be can be divided into two groups, one systematic
many individuals, groups, or classes involved, but and the other random. Thus, for example, we
they are relatively undifferentiated and their inter- might assume that in a population of voters there
actions are relatively simple. Empirical observa- are many factors affecting electoral choice. Some
tions of political systems, on the other hand, of those factors (e.g., income) have impacts on
often stress the institutional complexity of the vote that are strong and consistent across in-
modern states (Ashford, 1977; Scharpf, 1977) and dividuals. Other factors (e.g., specific policy
identify a rather complicated intertwining of in- issues) have impacts that are weaker or less consis-
stitutions, individuals, and events. Alternatives tent or less well understood. If we assume the lat-
are not automatically provided to a decision ter factors can be treated as noise, that is, that
maker; they have to be found. Search for alter- they are independent, randomly distributed
natives occurs in an organized context in which variables, the systematic factors will be clear in
problems are not only looking for solutions, but the aggregate results. In this way, conventional
solutions are looking for problems. Information assumptions of aggregation impose a statistical
about the consequences of alternatives is order on the results.
generated and communicated through organized The second classical simplification is the
institutions, so expectations depend on the struc- assumption of historical efficiency. Although the
ture of linkages within the system, as well as the argument is usually associated with theories of
ways in which biases and counter-biases cumulate natural selection and best specified in modern
(Simon, 1957a, b). Guesses about future pref- theories of population biology, the basic idea of
erences are developed within institutions historical efficiency is implicit in many modern
dedicated to defining and modifying values and theories. Regardless of the complexity or apparent
the meanings of actions (Cyert & March, 1963; anomalies of human behavior, historical pro-
March & Olsen, 1976). There are many such in- cesses are assumed to eliminate rules for behavior
stitutions, some nested within others, with multi-that are not solutions to an appropriate joint op-
ple, overlapping connections (Long, 1958). Na- timization problem. Thus, a prediction based on
tional political systems fit into international solving the optimization problem will correctly
political systems and are composed of numerous predict behavior, regardless of whether the actors
subsystems, some of which extend beyond na- involved formulate or solve that problem expli-
tional boundaries. citly (Friedman, 1953). For example, we might
If this complexity is not decomposable an- predict the outcome of a complicated political
alytically into smaller systems or susceptible to negotiation by assuming that the actors are each
some relatively simple aggregation techniques, the acting rationally on the basis of complete infor-
theoretical problems of understanding social mation about each other and the world, even
history are not easily accommodated within con- though we recognize that such assumptions are
temporary theoretical styles. For example, it may quite false as a description of individual behavior.

This content downloaded from

132.174.250.76 on Tue, 18 Aug 2020 13:17:37 UTC
All use subject to https://about.jstor.org/terms
1984 Organizational Factors in Political Life 741
Students of institutions suggest alternative be specified or solved by the observer, or if it is
theoretical simplifications for understanding com- impossible to identify the precise mechanisms by
plex political systems, most commonly the which historical experience is transformed into
assumption of a political structure. By a political current action. Unless the process is specified, it is
structure we mean a collection of institutions, impossible to examine either the likelihood that a
rules of behavior, norms, roles, physical arrange- particular equilibrium will be achieved or how
ments, buildings, and archives that are relatively long it will take.
invariant in the face of turnover of individuals
and relatively resilient to the idiosyncratic pref- Politics as an Interpretation of Life
erences and expectations of individuals. In con-
trast to theories that assume action is choice based A conception of politics as decision making is at
on individual values and expectations, theories of least as old as Plato and Aristotle. It is reflected in
political structure assume action is the fulfillment the language and concerns of political thought,
of duties and obligations. The difference is impor- from the earliest political philosophers through
tant. In a choice metaphor, we assume that Bentham to Merriam and Lasswell. Who getswhat
political actors consult personal preferences and and how? For the most part, contemporary
subjective expectations, then select actions that theory in political science considers politics and
are as consistent as possible with those preferences political behavior in such instrumental terms. The
and expectations. In a duty metaphor, we assume intent of actions is found in their outcomes, and
that political actors associate certain actions with the organizing principle of a political system is the
certain situations by rules of appropriateness. allocation of scarce resources in the face of con-
What is appropriate for a particular person in a flict of interest. Thus, action is choice, choice is
particular situation is defined by the political and made in terms of expectations about its conse-
social system and transmitted through socializa- quences, meanings are organized to affect
tion. choices, and symbols are curtains that obscure the
Political structure simplifies a complex world real politics, or artifacts of an effort to make deci-
for the individuals in it. It does not necessarily, sions.
however, simplify the problems of the political Parts of the new institutionalism are challenges
theorist. The complex intermeshing of rule-driven to this primacy of outcomes. These challenges
behavior may be just as difficult to unravel as the echo another ancient theme of political thought,
complex intermeshing of preference-driven the idea that politics creates and confirms inter-
behavior. As a result, there has long been a pretations of life. Through politics, individuals
tendency to combine ideas of political structure develop themselves, their communities, and the
with ideas of historical efficiency. If individual public good. In this view, participation in civic life
behavior is driven by rules within a political struc- is the highest form of activity for a civilized per-
ture, then it is possible to imagine that historical son. The ideas find post-Hellenistic voices in J.S.
experience accumulates over generations of in- Mill, Pateman (1970), and Lafferty (1981).
dividual experience. The information about that Politics is regarded as education, as a place for
experience is encoded in institutional rules discovering, elaborating, and expressing mean-
(Nelson & Winter, 1982). This argument is a ings, establishing shared (or opposing) concep-
familiar one to political discourse. It has been a tions of experience, values, and the nature of ex-
part of conservative doctrine for hundreds of istence. It is symbolic, not in the recent sense of
years, forming a basis for defending both tradi- symbols as devices of the powerful for confusing
tional rules of behavior and the existing political the weak, but more in the sense of symbols as the
order. instruments of interpretive order.
The advantage to treating behavior as rule The primary source of the institutionalist
driven, in addition to its apparent consistency challenge is empirical. Observers of processes of
with numerous observations, is not that it is pos- decision making regularly discern features that are
sible thereby to "save" a belief in historical effi- hard to relate to an outcome-oriented conception
ciency; rather, it is that it leads more naturally of collective choice. The pleasures are often in the
than does treating behavior as optimization to an process. Potential participants seem to care as
examination of the specific ways in which history much for the right to participate as for the fact of
is encoded into rules, and thus to making the idea participation; participants recall features of the
of historical efficiency more attentive to the pos- process more easily and vividly than they do its
sible limiting conditions for efficiency, and more outcomes; heated argument leads to decisions
likely to generate interesting predictions about without concern about their implementation; in-
multiple equilibria or long time paths. In fact, the formation relevant to a decision is requested but
assumption of efficiency becomes mostly a matter not considered; authority is demanded but not ex-
of faith if the joint optimization problem cannot ercised (Feldman & March, 1981; March & Olsen,

This content downloaded from

132.174.250.76 on Tue, 18 Aug 2020 13:17:37 UTC
All use subject to https://about.jstor.org/terms
742 The American Political Science Review Vol. 78

1976). These observations are often reported as not necessarily efficient, but it would be of greater
anomalies, as symptoms of some kind of perver- help if we were able to show the specific ways by
sity in the systems that were observed, paradox- which specific history-dependent processes lead to
ical. The appearance of paradox, however, is a outcomes that are either non-unique or long
product of our theoretical presumption that the delayed under some conditions. It is plausible to
main point of a decision-making process is a deci- argue that politics is filled with behavior that is
sion. For many purposes, that presumption may difficult to fit into a utilitarian model, but the
be misleading. The processes of politics may be plausibility would be augmented if we could
more central than their outcomes. describe an alternative model. And it is pro-
Politics and governance are important social vocative to note the importance of symbols,
rituals. In older worlds in which the major causal ritual, ceremony, and myth in political life, but we
force producing historical experience was the will cannot sustain the provocation without a clearer
of the gods, social rituals were organized around specification of how theories of politics are af-
ceremonies by which that will was discovered and fected by such a vision.
influenced. Most contemporary developed Moving from the subtle judgments of empirical
societies, being somewhat more secular in their knowledge to an appropriate theoretical formula-
conceptions of causality, believe that experience is tion is no easier in the analysis of politics than it is
produced by a combination of natural laws and elsewhere. It requires not only further empirical
intentional human action. In these societies, studies but also theoretical research. By theoret-
therefore, social and political rituals are organized ical research we mean primarily the development
around the consultation of expertise and the mak- of ideas, concepts, and models based on empirical
ing of decisions (Olsen, 1970). The procedures of observations and relevant to a behavioral under-
decision that we observe are reflections and standing and prescriptive ordering of political life.
reminders of this modern, secular conception of The objective is not impossible. Thirty years ago,
the social order. They are signals and symbols of empirical students of organizations made two ma-
the appropriateness of events, not in the sense jor criticisms of the existing theory of organiza-
that what happened needs to be viewed as tional decision making. The first criticism was
desirable or pleasant, but in the sense that what that the theory made extraordinary time and in-
happened can be viewed as having occurred in the formation demands on organizations (March &
way things happen (Feldman & March, 1981). The Simon, 1958; Simon, 1957a, b). Information and
usual term is "legitimate"; but legitimacy may time were treated as freely available resources. To
denote something narrower than is intended, for ask that all consequences of all alternatives be
what rituals seek to establish is not only the moral known precisely seemed unreasonable in the face
virtue of events but also their necessity. of empirical evidence that organizations con-
sidered only a small number of alternatives, ex-
Theoretical Research and Political Institutions amined only a small number of consequences
related to only a subset of organizational goals,
Human actions, social contexts, and institu- and made relatively imprecise estimates.
tions work upon each other in complicated ways, The second criticism was that the theory as-
and these complex, interactive processes of action sumed that all participants in an organization
and the formation of meaning are important to shared the same goals, or if they did not, that con-
political life. Institutions seem to be neither flict among them could be readily managed
neutral reflections of exogenous environmental through the terms of some prior agreement (Cyert
forces nor neutral arenas for the performances of & March, 1963; March, 1962). In the case of a
individuals driven by exogenous preferences and political organization, the agreement was a coali-
expectations. As a result, contemporary political tion contract, or constitution, by which all
theory is probably overly sanguine about the members of a coalition or polity agreed to be
possibilities for a theory of politics that ignores bound to the policies specified through bargaining
political institutions. or legislation. Thus, the familiar distinction be-
For the most part, however, the relevant tween "politics" and "administration." In the
theoretical work remains to be done. It is in- case of an economic organization, the agreement
teresting to suggest that political institutions and was an employment contract by which employees,
the society are interdependent, but that statement in return for the payment of wages, agreed to act
needs to find a richer theoretical expression. It is as though they had the same goals as the owner or
appropriate to observe that political institutions other legitimate policy maker. Empirical studies
can be treated as actors in much the same way we seemed to indicate that conflict was endemic in
treat individuals as actors, but we need more organizations and that it tended to be inter-
detailed demonstrations of the usefulness of doing minable rather than settled by prior agreements.
so. There is good sense in noting that history is These criticisms began to have serious impact

This content downloaded from

132.174.250.76 on Tue, 18 Aug 2020 13:17:37 UTC
All use subject to https://about.jstor.org/terms
1984 Organizational Factors in Political Life 743

on formal theories of organized action when they conceptions on which a modest amount of theo-
were translated into useful theoretical statements retical work might yield rewards.
through the development of information Historical Order. The concept of historical
economics and theories of agency. Such theories order implicit in contemporary theory emphasizes
consider information as a scarce resource subject the efficiency of historical processes, the ways in
to strategic action in a world populated by self- which history moves quickly and inexorably to a
interested rational actors. Ideas drawn from unique outcome, normally in some sense an op-
organizational studies of bounded rationality and timum. An institutional theory would specify how
internal conflict permeate modern economic historical processes are affected by specific
theory in the form of discussions of moral hazard, characteristics of political institutions, and it
asymmetric information, agency, signalling, and would provide greater theoretical understanding
optimal information strategies (Hirshleifer & of the inefficiencies of history, i.e., historical pro-
Riley, 1979). Most students of organizations cesses that do not have equilibria, take extended
would argue that these theories are also in- periods of time, lead to non-unique equilibria, or
complete, but it is clear that the earlier empirical result in unique but suboptimal outcomes. Theo-
criticisms have reformed theoretical thinking. retical attention to the inefficiencies of history in-
The new institutionalism would benefit from volves a greater concern for the ways in which in-
similar theoretical development if it could be ac- stitutions learn from their experience (Etheredge,
complished. Like the early observations about 1976) and the possibilities that learning will pro-
bounded rationality and internal conflict, obser- duce adjustments that are slower or faster than
vations about the importance of institutions have are appropriate or are misguided. It involves try-
generally taken the form of criticism of existing ing to specify the conditions under which the se-
theoretical ideas rather than the delineation of an quential branches of history turn back upon each
alternative set of precise theoretical concepts. other and the conditions under which they
Developing a comprehensive theoretical structure diverge. It involves characterizing the role of stan-
for institutional thinking is, of course, a pro- dard operating procedures, professions, and ex-
digious and pretentious task, not one that will be pertise in storing and recalling history.
undertaken here. We can, however, identify a few Temporal Order. In most theories of action, we
ideas associated with the new institutionalism that assume things are ordered by their consequential
might warrant theoretical attention. connections. Means are linked to appropriate
ends; causes are linked to effects they produce;
consequences are linked to actions that lead to
Institutional Conceptions of Order them and to preferences they affect; solutions are
linked to problems they solve. Such concepts of
Institutional thinking emphasizes the part order underlie theories of choice. Deviations from
played by institutional structures in imposing consequential order are viewed as interesting aber-
elements of order on a potentially inchoate world. rations, disturbances of a system otherwise held
Traditional political theory involved considerable together by the way wanting something leads to
attention to the order produced by political con- doing something connected to the want, and do-
tracts and reflected in constitutions, laws, and ing something leads to consequences related to the
other stable rules, or by a community of moral intention. Temporal order provides an alternative
obligation, often inspired and buttressed by religi-in which linkages are less consequential than tem-
ous dogma (Berki, 1979; Waterstone, 1966). For poral. Things are connected by virtue of their
the most part, modern political theory eschews simultaneous presence or arrival. In a culture with
such concerns and focuses on aggregation and a strong sense of monthly or yearly cycles or of
historical efficiency superimposed on two other birth cohorts, we should not be overly surprised
kinds of order: the order imposed by reason and by temporal order. In many human situations the
the order imposed by competition and coercion. most easily identified property of objects or
Reason is recognized in ideas of rationality and in-
events is the time subscripts associated with them.
tentional action; it finds institutional expression Thus, students of time allocation in organizations
in the hierarchical organization of means and ends have observed the ways in which attention to
(and thus in formally planned institutions). Com- problems seems to be determined as much by the
petition and coercion are recognized in ideas of time of their arrival as by assessments of their im-
conflict of interest, power, bargaining, survival, portance. A classic form of temporal order is
and war; they find institutional expression in elec-found in queuing theory, although most discus-
tions and policymaking. Theoretical research rele- sions of queuing are embedded in a consequential
vant to the new institutionalism would involve structure in which queues are either indistinguish-
elaborating additional notions of political order. able or distinguishable only by their processing
We believe it is possible to identify at least six suchtimes.

This content downloaded from

132.174.250.76 on Tue, 18 Aug 2020 13:17:37 UTC
All use subject to https://about.jstor.org/terms
744 The American Political Science Review Vol. 78

Endogenous Order. Much of contemporary students of legislatures imagine that a legislature is

theory emphasizes the way order is imposed on best understood in terms of lawmaking, and
political institutions by an external environment. students of courts imagine that a court is best
From this perspective, for example, power within understood in terms of adjudicating. Alterna-
a political system is determined by possession of tively, a human institution can be studied and in-
resources in the environment, interests are deter- terpreted as the cross-section of the lives of the
mined by position in the external world, and people involved. The idea that collective behavior
coherence within an institution is assured by the can be understood as a mosaic of private lives
exigencies of existence. Thus, order is effectively links contemporary theoretical thought to similar
exogenous to the institution and does not depend ideas among qualitative students of human
on properties of the institution or processes within behavior and novelists (Krieger, 1983). A focus on
it. Students of institutions have suggested a institutional demography combines such a vision
number of ways in which internal institutional of organized life with attention to a property of
processes affect things like the power distribution,individual lives that is itself a product of the in-
the distribution of preferences, or the manage- stitutional structure-the individual career
ment of control. As a result, they invite theoret- (March & March, 1978; Pfeffer, 1981b). The the-
ical development of models appropriate for oretical requirements include useful concepts of
understanding the ways in which interests and the ways in which organizations adapt through
preferences develop within the context of institu- turnover, institutions are driven by their cohort
tional action, the ways reputations for power structures, and the pursuit of careers and profes-
evolve as a result of the outcomes of politics, the sional standards dictates the flow of events.
ways in which the process of controlling purposive Symbolic Order. Students of formal organiza-
organizations produces unanticipated conse- tions have called attention to the ordering force of
quences, and the ways in which the course of deci- symbols, rituals, ceremonies, stories, and drama
sion making within political systems systematic- in political life (March, 1981; March & Olsen,
ally, and endogenously, results in illusions of suc- 1976, 1983; Meyer & Rowan, 1977; Pfeffer,
cess and failure. 1981a; Pondy, 1978). Symbols permeate politics
Normative Order. It is a commonplace observa-in a subtle and diffuse way, providing interpretive
tion in empirical social science that behavior is coherence to political life. Many of the activities
constrained and dictated by cultural dicta and and experiences of politics are defined by their
social norms. Although self-interest undoubtedly relation to myths and symbols that antedate them
permeates politics, action is often based more on and that are widely shared. At the same time,
discovering the normatively appropriate behavior symbolic behavior is also a strategic element in
than on calculating the return expected from political competition. Individuals and groups are
alternative choices. As a result, political behavior, frequently hypocritical, reciting sacred myths
like other behavior, can be described in terms of without believing them and while violating their
duties, obligations, roles, and rules. Such a implications. The traditional problem with such-
description has not, however, been translated into observations is not doubt about their veracity but
any very compelling theoretical form. Some ef- about our ability to translate them into useful the-
forts have been made to rationalize normative oretical statements without excessive damage to
rules, such as altruism (Kurz, 1978) and reciproc- their meaning. Theoretical development reflective
ity (Axelrod, 1980), or to specify the conditions of an institutional perspective would include an
for their evolution (Axelrod & Hamilton, 1981; examination of the ways in which the tendencies
Trivers, 1971). From an institutionalist perspec- toward consistency and inconsistency in beliefs af-
tive, such efforts are exemplary, but they tend to fect the organization of political meaning, the
limit attention to the comparative statics of in- ways in which "exemplary centers" (Geertz, 1980)
dividual norms. A broader theoretical examina- create social order through ceremony, and the
tion of normative order would consider the rela- ways in which symbolic behavior transforms more
tions among norms, the significance of ambiguityinstrumental behavior and is transformed by it. In
and inconsistency in norms, and the time path of particular, a serious theoretical understanding of
the transformation of normative structures. A myths, symbols, and rituals must include some at-
theoretical understanding of such conventional tention to the dynamics of symbols, to the pro-
norms as those surrounding trust and legitimacy cesses by which symbols shape the behavior not
seems likely to be particularly germane to political only of the innocent but of the society as a whole.
analysis.
Demographic Order. It is tempting for students Examples of Possible Theoretical Research
of politics, as for students of other human
endeavor, to find order defined in terms of the Within these six conceptions of order, there are
logic of their particular domain of interest. Thus, possibilities for theoretical research attentive to

This content downloaded from

132.174.250.76 on Tue, 18 Aug 2020 13:17:37 UTC
All use subject to https://about.jstor.org/terms
1984 Organizational Factors in Political Life 745
the insights of students of institutions. Such ing current historical events. In effect, the chance
research is institutional in two respects: First, it is
fluctuations of history change the baselines of the
oriented to one or more of the institutionalist con- next step of the historical process. Common
ceptions of order; second, it tries to illuminate descriptions of incremental policy processes make
how institutional and organizational factors af- them appear to be in the nature of martingales.
fect political events. As examples, consider the The distribution of possible outcomes from a
following: policy process is pictured as resulting from com-
Example 1: Policy Martingales. Many models petition and bargaining over incremental adjust-
of history recognize that specific historical events ments in the current policy; the policy actually
involve elements of chance. The unique historical adopted is a draw from that distribution. This
happening may be a draw from some probability martingale property of policymaking is not in-
distribution of possible events. Even in cases dependent of institutional factors. Indeed, it
where chance, strictly considered, is not viewed as seems a prototypic institutional characteristic.
vital, any specific event is seen as the consequence Policies, once adopted, are embedded into institu-
of a complicated interweaving of factors impossi- tions. They are associated with rules, expecta-
ble to predict with precision in a single case. In the tions, and commitments. By affecting attention
independent trial version of such models, any and aspirations, they affect the future search
specific historical event is subject to various kinds behavior of political participants.
of random fluctuations but, in the long run, Martingales diverge more rapidly than do in-
unlikely events at a particular time are balanced dependent trial processes; that is, for a given
by different unlikely events at a subsequent time. amount of chance variation in each time period,
The specific realizations of the historical process the variance across possible outcomes after some
that comprise the events of today are independent number of periods will be substantially greater in
of the specific realizations that comprise the a martingale. As a result, the precision with which
events of yesterday. Each specific event of an un- specific realizations of the process can be an-
folding history is relatively difficult to predict, but ticipated is considerably less. Thus, policy mar-
prediction is not improved by knowledge of the tingales are related to, but not identical to,
history of past realizations of that process. various less precise ideas of forks in history, of
It is possible to see political policymaking as an critical events that made a difference. There is a
independent trial process. Suppose we think of sense in which the first step is more important
policy as the result of bargaining among political than any subsequent one, but it is a limited sense.
actors with prior preferences and resources, but In a martingale process all events are forks; the
subject to trial-by-trial variation attributable to policy paths of two political systems with identical
specific, unpredictable and uncontrollable fac- underlying political conditions will be radically
tors. Then understanding the short-run outcomes different simply because of the way in which
of a policy process would depend on considerable (possibly small) perturbations shift the focus of
detail of the specific situation. A student of in- political pressure.
stitutions might well observe that the details of the Example 2: Experiential Learning. It is a fre-
way attention is organized, how alternatives are quent observation of institutionalism that institu-
presented, what information is available, which tions accumulate historical experience through
participants are free from other demands, how in- learning. The results and inferences of past ex-
stitutional memory is consulted, and a host of perience are stored in standard operating pro-
other factors would affect the specific political cedures, professional rules, and the elementary
policy adopted at a specific time. At the same rules of thumb of a practical person. These
time, however, such factors are irrelevant (or elements of historical knowledge have been por-
redundant) to understanding the long-term mix of trayed both as forms of irrational retrogression
policies. Such an understanding is possible simply and as carriers of wisdom, and it is not hard to
from a knowledge of the underlying political pro- specify environmental situations in which either
cess and any systematic institutional biases. characterization would be appropriate. What is
Not all policymaking processes are independent less clear is whether we can model the processes of
trial processes. Many of them seem to be more in institutional learning. Although there have been
the nature of martingales (Feller, 1950). Like an some loose arguments that experiential learning
independent trial process, a martingale process is will, in the long run, lead to the discovery and
subject to chance variation, but the variations ac- adoption of optimal strategies, little theoretical
cumulate. What distinguishes a martingale is the effort has been devoted to specifying precisely the
property that the expected value of the process at conditions under which learning from experience
one time is equal to the realization of the process leads to optimal behavior, or to relating those
at the preceding time. This property makes the conditions to features of institutional structure or
specific path of history important to understand- life.

This content downloaded from

132.174.250.76 on Tue, 18 Aug 2020 13:17:37 UTC
All use subject to https://about.jstor.org/terms
746 The American Political Science Review Vol. 78

Consider the following simple model of learn- taneity, relatively few problems are solved, and
ing (Levinthal & March, 1982). A decision- choices are made for the most part either before
making institution simultaneously learns along any problems are connected to them (oversight),
three dimensions. First it modifies its strategy; or after the problems have abandoned one choice
that is, it changes the likelihood of making one to associate themselves with another (flight). This
choice rather than another among the alternative situation of extreme loose coupling, called an
activities available to it. Subjective success leads open structure in the original discussions of the
to increasing the chance of repeating a choice; garbage can, has attracted most of the attention in
subjective failure leads to decreasing the chance of the literature, and empirical studies have revealed
repeating a choice (March & Olsen, 1976). Sec- decision processes that appear to approximate
ond, an institution modifies its competences; that such an open structure (March & Olsen, 1976).
is, it changes the skill it has at the various ac- Not all decision situations are quite so unstruc-
tivities in which it might engage. Competence at tured, however. We can characterize a choice
an activity increases with experience at it; it situation in terms of two structures. The first is
decreases with time (Preston & Keachie, 1964). the access structure, a relation between problems
Third, an institution modifies its aspirations; that (or solutions) and choice opportunities. The ac-
is, it changes its definition of subjective success. cess structure may require, allow, or not allow a
Aspirations move in the direction of past perfor- particular problem, if activated, to be attached to
mance (Cyert & March, 1963). It is clear that in- a particular choice. The second structure is the
stitutional factors affect several of the key decision structure, a relation between decision
features of such learning. The learning rates makers and choice opportunities. This structure
associated with the three kinds of learning are may require, allow, or not allow that a particular
partly a function of features of the institution. decision maker participate in the making of a par-
The degree of loose coupling in an organization ticular choice. Access and decision structures can
affects the precision with which choices are made, be imagined in any kind of arbitrary configura-
outcomes observed, aspirations expressed, and tion, but two special forms have been considered
competences realized. Thus, it can be expressed as formally. A specialized structure is one that is
various forms of noise in the process. Organiza- decomposed into substructures that are open.
tional slack affects the degree of centralization inThus, a specialized decision structure is one in
the organization, and thus the linkage among which it is possible to divide choice opportunities
subunits. and decision makers into subgroups and match
The three dimensions of learning obviously in- the two sets of subgroups so that every decision
teract. For example, learning of aspirations af- maker in a particular subgroup of decision makers
fects the definition of subjective success, and has access to every choice opportunity in the
thereby affects the learning of strategies. Learning matched set of choice opportunities, but to no
of competences affects performance outcomes, other. A hierarchical structure is one that expands
and thereby affects the learning of both strategiesaccess rights as a function of hierarchical rank.
and aspirations. Learning of strategies affects For example, in a hierarchical access structure,
choices, and thereby affects the learning of com- problems and choices are ordered, and each prob-
petences. The model can be explored to discover lem has access to choices of the same or lower
the circumstances under which it reaches an rank. The differences made by these structures
equilibrium, and, among those the circumstances, have been noted both formally (Cohen, March, &
those under which it reaches an optimum. It can Olsen, 1972) and empirically (Egeberg, 1981;
also be combined into more complicated struc- Olsen, 1983), but the empirical and theoretical ex-
tures of learning where the choices of one institu- amination of garbage-can processes within access
tion affect the outcomes of another (e.g., com- and decision structures that are not completely
petition and cooperation), and where the learning open is barely begun.
institution is itself composed of learning subunits.
Example 3: Garbage Cans. Garbage-can models Subtle Phenomena and Simple Theories
of organizational choice have been suggested as a
representation of a particular temporal order. In These examples hardly exhaust the list. Em-
the form most commonly discussed in the litera- pirical observations of reputations for power in
ture, the garbage-can model assumes that prob- politics suggest that such reputations depend
lems, solutions, decision makers, and choice op- heavily on the place of an individual in a political
portunities are independent, exogenous streams structure and on inferences about the relation be-
flowing through a system (Cohen, March, & tween preferences and outcomes. Some simple
Olsen, 1972). They come together in a manner models of the dynamic relations among reputa-
determined by their arrival times. Thus, solutions tions for power, institutional position, pref-
are linked to problems primarily by their simul- erences, and social outcomes would provide a

This content downloaded from

132.174.250.76 on Tue, 18 Aug 2020 13:17:37 UTC
All use subject to https://about.jstor.org/terms
1984 Organizational Factors in Political Life 747

richer understanding of the ways in which power ter, which of course does not exclude the possibil-
reputations affect politics. Empirical observations ity that it is also the former, we have tried to draw
of post-decision surprises (i.e., deviations of some possible implications for theoretical
realized outcomes from expected outcomes) sug- research in political science. They are, at best,
gest that there are systematic differences between theoretical directions suggested by a sympathetic
the ways in which individuals experience the con- appreciation of a tradition of institutionalist
sequences of their actions and the ways in which thought. Such an effort is a little like trying to
institutions do. Some simple models of institu- write a useful commentary on Heidegger in the
tional expectations, choices, and post-decision form of a Shakespearean sonnet. If it has virtue, it
assessments would clarify the occasions for ex- is in attempting to encourage talking about a
pecting positive or negative surprises from subtle body of thought in a way sufficiently naive
deliberate action. to entice the technically proficient.
What characterizes all of the examples, as well
as the others that might be added, is a relatively
simple approach to institutional phenomena. The References
new institutionalism is often couched in terms of a
contrast between the complexity of reality and the
simplifications provided by existing theories, but Akerlof, G. A. The economics of social customs, of
which unemployment may be one consequence.
theoretical research from an institutional perspec-
Quarterly Journal of Economics, 1980, 95, 749-775.
tive cannot involve the pursuit of enormous con-
Ashford, D. E. Political science and policy studies:
textual detail. It is constrained by the capacity of towards a structural solution. Policy Studies Jour-
human (and artificial) intelligence to cope with nal, 1977, 5, 570-583.
complexity, and although that capacity seems to Axelrod, R. More effective choice in prisoners' di-
expand with time, the rate of expansion continues lemma. Journal of Conflict Resolution, 1980, 24,
to be modest relative to the demands of a fully 379-403.
contextual and institutional theory. From the Axelrod, R., & Hamilton, W. D. The evolution of co-
point of view of theoretical research, conse- operation. Science, 1981, 211, 1390-1396.
Bachrach, P., & Baratz, M. The two faces of power.
quently, the new institutionalism is probably bet-
American Political Science Review, 1962, 56,
ter viewed as a search for alternative ideas that
947-952.
simplify the subtleties of empirical wisdom in a Becker, G. S., & Stigler, G. J. De gustibus non est
theoretically useful way. disputandum. American Economic Review, 1977,
67, 76-90.
Conclusion Berger, S. (Ed.). Organizing interests in Europe:
pluralism, corporatism, and the transformation of
The institutionalism we have considered is politics. Cambridge: Cambridge University Press,
1981.
neither a theory nor a coherent critique of one. It
Berki, R. N. State and society: an antithesis of modern
is simply an argument that the organization of
political thought. In J. E. S. Hayward & R. N.
political life makes a difference. Some of the Berki (Eds.), State and society in contemporary
things we have noted are fragments of ideas; Europe. Oxford: Martin Robertson, 1979.
others are somewhat more systematic in develop- Burns, J. M. Leadership. New York: Harper and Row,
ing a theme or reporting a series of observations. 1978.
They are held together by an awareness of a set of Cohen, A. P. Two-dimensional man: an essay on the
phenomena that are more easily observed than ex- anthropology of power and symbolism in complex
plicated. Insofar as the ideas are consistent, the societies. London: Routledge and Kegan Paul, 1974.
Cohen, M. D., & Axelrod, R. Coping with complexity:
consistency is sustained partly by ambiguity.
The adaptive value of changing utility. American
Many of the core ideas seem plausible and have
Economic Review, 1984, 74, 30-42.
been durable, but plausible durability (as Cohen, M. D., March, J. G., & Olsen, J. P. A garbage
numerous students of the history of knowledge can model of organizational choice. Administrative
have observed) is neither necessary nor sufficient Science Quarterly, 1972, 17, 1-25.
for good sense. Crozier, M. The bureaucratic phenomenon. Chicago:
The new institutionalism is an empirically based University of Chicago Press, 1964.
prejudice, an assertion that what we observe in Cyert, R. M., & March, J. G. A behavioral theory of
the world is inconsistent with the ways in which the firm. Englewood Cliffs, N.J.: Prentice-Hall,
Inc., 1963.
contemporary theories ask us to talk. Like other
Downs, A. An economic theory of democracy. New
prejudices in knowledge, it may be wrong-headed
York: Harper and Row, 1957.
or muddle-headed, but it may also be a useful Easton, D. Political science. In D. L. Sills (Ed.), Inter-
continuation of that gentle confrontation between national encyclopedia of the social sciences. Vol. 12.
the wise and the smart that describes much of in- New York and London: Macmillan and Free Press,
tellectual history. On the chance that it is the lat- 1968.

This content downloaded from

132.174.250.76 on Tue, 18 Aug 2020 13:17:37 UTC
All use subject to https://about.jstor.org/terms
748 The American Political Science Review Vol. 78

Eckhoff, T., & Jacobsen, K. D. Rationality and respon- American Economic Review, 1978, 68, 216-222.
sibility in administrative and judicial decision- Laegreid, P., & Olsen, J. P. Byraakrati og beslutninger.
making. Copenhagen: Munksgaard, 1960. Bergen: Universitetsforlaget, 1978.
Edelman, M. The symbolic uses of politics. Urbana: Lafferty, W. M. Participation and democracy in Nor-
University of Illinois Press, 1964. way. Oslo: Universitetsforlaget, 1981.
Egeberg, M. Stat og organisasjoner. Bergen: Univer- Lave, C. A., & March, J. G. An introduction to models
sitetsforlaget, 1981. in the social sciences. New York: Harper and Row,
Enderud, H. The perception of power. In J. G. March 1975.
& J. P. Olsen (Eds.), Ambiguity and choice in Levi, M. The predatory theory of rule. Politics and
organizations. Bergen: Universitetsforlaget, 1976. Society, 1981, 10, 431-465.
Etheredge, L. S. The case of the unreturned cafeteria Levinthal, D., & March, J. G. A model of adaptive
trays. Washington, D.C.: American Political organizational search. Journal of Economic
Science Association, 1976. Behavior and Organization, 1982, 2, 307-333.
Evans, P., Rueschemeyer, D., & Skocpol. T. (Eds.). Long, N. The local community as an ecology of games.
Bringing the state back in. Report of a conference on American Journal of Sociology, 1958, 44, 251-261.
research implications of current theories of the state. Luce, R. D., & Raiffa, H. Games and decisions. New
Unpublished manuscript, 1983. York: Wiley, 1957.
Feldman, M. S., & March, J. G. Information as signal March, J. C., & March, J. G. Performance sampling in
and symbol. Administrative Science Quarterly, social matches. Administrative Science Quarterly,
1981, 26, 171-186. 1978, 23, 434453.
Feller, W. An introduction to probability theory and its March, J. G. Bounded rationality, ambiguity, and the
applications, Vol. I. New York: Wiley, 1950. engineering of choice. Bell Journal of Economics,
Friedman, M. Essays in positive economics. Chicago: 1978, 9, 587-608.
University of Chicago Press, 1953. March, J. G. The business firm as a political coalition.
Geertz, C. Negara: the theater state in nineteenth- Journal of Politics, 1962, 24, 662-678.
century Bali. Princeton, N.J.: Princeton University March, J. G. Decisions in organizations and theories of
Press, 1980. choice. In A. H. van de Ven & W. F. Joyce (Eds.),
Hanf, K., & Scharpf, F. (Eds.). Interorganizational Perspectives on organizational design and behavior.
policy making: limits to coordination and central New York: Wiley, 1981.
control. London: Sage, 1978. March, J. G. Politics and the city. In K. Arrow, J. S.
Hayward, J. E. S., & Berki, R. N. (Eds.). State and Coleman, A. Downs, & J. G. March (Eds.), Urban
society in contemporary Europe. Oxford: Martin processes as viewed by the social sciences.
Robertson, 1979. Washington, D.C.: The Urban Institute, 1970.
Heclo, H. Modern socialpolicies in Britain and Sweden. March, J. G. The power of power. In D. Easton (Ed.),
New Haven, Conn.: Yale University Press, 1974. Varieties of political theory. Englewood Cliffs,
Heller, H. Political science. In E. R. A. Seligman & N.J.: Prentice-Hall, 1966.
A. Johnson (Eds.), Encyclopedia of the social March, J. G., & Olsen, J. P. Ambiguity and choice in
sciences. New York: Macmillan, 1933 (1957). organizations. Bergen: Universitetsforlaget, 1976.
March, J. G., & Olsen, J. P. Organizing political life:
Hirschleifer, J., & Riley, J. G. The analytics of uncer-
tainty and information--an expository survey. Jour- What administrative reorganization tells us about
nal of Economic Literature, 1979, 17, 1375-1421. government. American Political Science Review,
Jessop, B. Recent theories of the capitalist state. In 1983, 77, 281-296.
Cambridge Journal of Economics, 1977, 1, 353-373. March, J. G., & Shapiro, Z. Behavioral decision theory
Kahneman, D. Bureaucracies, minds, and the human and organizational decision theory. In G. R.
engineering of decisions. In G. R. Ungson & D. N. Ungson & D. N. Braunstein (Eds.), Decision mak-
Braunstein (Eds.), Decision making: an inter- ing: an interdisciplinary inquiry. Boston: Kent
disciplinary inquiry. Boston: Kent, 1982. 1982.
Kahneman, D., Slovic, P., & Tversky, A. (Eds.). March, J. G., & Simon, H. A. Organizations. New
Judgment under uncertainty: heuristics and biases. York: Wiley, 1958.
Cambridge: Cambridge University Press, 1982. Masters, R. D. The biological nature of the state.
Katzenstein, P. J. (Ed.). Between power and plenty: World Politics, 1983, 35, 161-193.
foreign economic policies of advanced industrial Meyer, J. W., & Rowan, B. Institutionalized organiza-
states. Madison: University of Wisconsin Press, tions: formal structure as myth and ceremony.
1978. American Journal of Sociology, 1977, 83, 340-363.
Kjellberg, F. Political institutionalization. London: Nelson, R. R., & Winter, S. G. An evolutionary theory
Wiley, 1975. of economic change. Cambridge, Mass.: Harvard
Krasner, S. D. Defending the national interest: raw University Press, 1982.
materials, investments and the U.S. foreign policy.Nisbet, R., & Ross, L. Human inference: strategies and
Princeton, N.J.: Princeton University Press, 1978. shortcomings of social judgement. Englewood
Krasner, S. D. Approaches to the state: alternative Cliffs, N.J.: Prentice-Hall, 1980.
conceptions and historical dynamics. Comparative Niskanen, W. A. Bureaucracy and representative
Politics, in press. government. Chicago: Rand McNally, 1971.
Krieger, S. Mirror dance. Philadelphia: Temple Univer- Nordlinger, E. On the autonomy of the democratic
sity Press, 1983. state. Cambridge, Mass.: Harvard University Press,
Kurz, M. Altruism as an outcome of social interaction. 1981.

This content downloaded from

132.174.250.76 on Tue, 18 Aug 2020 13:17:37 UTC
All use subject to https://about.jstor.org/terms
 1984 Organizational Factors in Political Life 749

Olsen, J. P. Local budgeting: decision making or a corporatist intermediation. Beverly Hills, Calif.:
ritual act? Scandinavian Political Studies, 1970, 5, Sage, 1979.
85-118. Selznick, P. Leadership in administration. Evanston,
Olsen, J. P. Organizational integration in government. Ill.: Northwestern University Press, 1957.
In P. Nystrom & W. Starbuck (Eds.), Handbook of Sharpe, W. F. Portfolio theory and capital markets.
organizational design. Vol. 2. New York: Oxford New York: McGraw-Hill, 1970.
University Press, 1981. Shepsle, K., & Weingast, B. Institutionalizing majority
Olsen, J. P. Organized democracy. Bergen: Univer- rule: A social choice theory with policy implications.
sitetsforlaget, 1983. American Economic Review, 1983, 73, 357-372.
Olsen, J. P., & Saetren, H. Aksjoner og demokrati. Simon, H. A. Administrative behavior. (2nd ed.) New
Bergen: Universitetsforlaget, 1980. York: Macmillan, 1957a.
Olson, M. The logic of collective action. Cambridge, Simon, H. A. Models of man. New York: Wiley,
Mass.: Harvard University Press, 1965. 1957b.
Padgett, J. F. Hierarchy and ecological control in Skocpol, T. Political response to capitalist crises: neo-
federal budgetary decision making. American Jour- Marxist theories of the state and the case of the New
nal of Sociology, 1981, 87, 75-129. Deal. Politics and Society, 1980, 10(2), 155-201.
Pateman, C. Participation and democratic theory. Skocpol, T. States and social revolutions: a comparative
Cambridge: Cambridge University Press, 1970. analysis of France, Russia, and China. Cambridge
Pfeffer, J. Management as symbolic action: The Crea- and New York: Cambridge University Press, 1979.
tion and maintenance of organizational paradigms. Skocpol, T., & Finegold, K. State capacity and eco-
In L. Cummings & B. M. Staw (Eds.), Research in nomic intervention in the early New Deal. Political
organizational behavior. Vol. 3. Greenwich, Conn.: Science Quarterly, 1982, 97, 255-278.
JAI Press, 1981(a). Skowronek, S. Building a new American state. Cam-
Pfeffer, J. Some consequences of organizational de- bridge: Cambridge University Press, 1982.
mography: Potential impacts of an aging work force Stephan, A. C. The state and society: Peru in compara-
on formal organizations. In S. B. Kiesler, J. N. tive perspective. Princeton, N.J.: Princeton Univer-
Morgan, & V. K. Oppenheimer (Eds.), Aging: social sity Press, 1978.
change. New York: Academic Press, 1981(b). Stigler, G. J. The theory of price (Rev. ed.) New York:
Pondy, L. R. Leadership as a language game. In M. W. Macmillan, 1952.
McCall, Jr. & M. M. Lombardo (Eds.), Leadership. Therborn, G. What does the ruling class do when it
Durham, N.C.: Duke University Press, 1978. rules? London: Verso, 1980.
Potter, A. L. Political institutions, political decay and Tilly, C. (Ed.). From mobilization to revolution. Read-
the Argentine crises of 1930. Unpublished doctoral ing, Mass.: Addison-Wesley, 1978.
dissertation, Stanford University, 1979. Trimberger, E. K. Revolution from above: Military
Pressman, J. L., & Wildavsky, A. B. Implementation. bureaucrats and development in Japan, Turkey,
Berkeley: University of California Press, 1973. Egypt, and Peru. New Brunswick, N.J.: Transaction
Preston, L. D., & Keachie, E. C. Cost functions and Books, 1978.
progress functions: an integration. American Trivers, R. The evolution of reciprocal altruism. Quar-
Economic Review, 1964, 54, 100-108. terly Review of Biology, 1971, 46, 35-57.
Robins, R. S. Political institutionalization and the Waterstone, G. C. Order and counterorder. Dualism
integration of elites. Beverly Hills, Calif.: Sage, in western culture. New York: Philosophical
1976. Library, 1966.
Scharpf, F. W. Does organization matter? Task struc- Williamson, 0. E. Markets and hierarchies. New York:
ture and interaction in the ministerial bureaucracy. Free Press, 1978.
In E. H. Burack & A. R. Negandhi (Eds.), Wolin, S. Politics and vision. Boston: Little, Brown,
Organizational design: theoretical perspectives and 1960.
empirical findings. Kent, Ohio: Kent State Universi- Wright, H. T. Recent research on the origin of the
ty Press, 1977. state. Annual Review of Anthropology, 1977, 6,
Schmitter, P., & Lehmbruch, G. (Eds.). Trends toward 379-397.

This content downloaded from

132.174.250.76 on Tue, 18 Aug 2020 13:17:37 UTC
All use subject to https://about.jstor.org/terms