The study of political institutions was dominant within political science in Britain and the US in the early twentieth century. Until the 1950s, institutionalism enjoyed a privileged status within the discipline—its assumptions and practices as well as its methodological and theoretical premises were rarely questioned. let alone subject to the behavioralist critique (Lowndes 2002). 

The institutional approach can be understood as a subject matter, as a method, and as a theory. As a subject matter, the study of political institutions is central to the identity of the discipline of political science (Rhodes 1995).

To quote Rhodes (1995, 43), “If there is any subject matter at all that political scientists can claim exclusively for their own, a subject mater that does not require acquisition of the analytical tools of sister fields and that sustains their claim to autonomous existence, it is, of course, formal-legal political structure.”

Public administration, a sub-discipline within political science, has the study of institutions as its key characteristic. Public administration is the study of “the institutional arrangements for the provision of public services” (Hood 1987, 504 in Rhodes 1995, 52) or”the study of public bureaucracies” (Rhodes 1979, 7 in Rhodes 1995, 52). William Robson (1975, 195 in Rhodes 1995, 52) describes the dominant approach in public administration as institutional: “It concentrated attention on the authorities engaged in public administration, analyzed their history, structure, functions, powers, and relationships. It enquired how they worked and the degree of effectiveness they achieved.”

As a method, the traditional or classic institutional approach is “descriptive-inductive, formal-legal, and historical-comparative.” It is descriptive because it employs the techniques of the historian and explores specific events, eras, people, and institutions and inductive because inferences are drawn from repeated observations (Rhodes 1995, 43).

As such, the classic institutional approach systematically describes and analyzes phenomena that have occurred in the past and explain contemporary political phenomena with reference to past events. The goal is to explain and understand but not to formulate laws (Kavanagh 1991, 482 in Rhodes 1995, 42).

The institutional approach also applies the formal-legal inquiry. Formal because it involves the study of formai governmental organizations, and legal because it includes the study of public law (Eckstein 1979, 2 in Rhodes 1995, 44). An example of formal-legal methods in the study of political institutions is constitutional studies (Rhodes 1995).

The classic or traditional institutional approach is also comparative. Woodrow Wilson (1989, xxxiv in Rhodes 1995, 45) argued that one’s “institutions can be understood and appreciated only by those who know other systems of government . . . By the use of a thorough comparative and historical method . . . a general clarification of views may be obtained.”

As a theory, the classic or traditional institutional approach does not only make statements about the causes and consequences of political institutions. It also espouses the political value of democracy (Rhodes 1995).

Proponents of the approach treat the functioning and fate of democracies (dependent variable) as a function of, or influenced by legal rules and procedures (independent variable). Moreover, the approach offers an opportunity for infusing into the empirical study of politics the analysis clf political values (Rhodes 1995). Influenced by the political philosophy of Michael Oakeshott, Johnson (1975, 276-7 in Rhodes 1995, 47) describes the rationale for the study of political institutions in the following manner:

political institutions express particular choices about how political relationships ought to be shaped; they are in the nature of continuing injunctions to members of a society that they should try to conduct themselves in specific ways when engaged in the pursuit of political ends. This is to define political institutions as necessarily containing a normative element.

Critics to the classic or traditional approach attack the approach’s limitation both in terms of scope and method. Peters (1999,6-11 in Lowndes 2002, 92) describes the ‘proto-theory’ of the traditional approach as “normative (concerned with ‘good government’), structuralist (structures determine political behavior), historicist (the central influence of history), legalist (law plays a major role in governing) and holistic (concerned with describing and comparing whole systems of government) (Lowndes 2002,
92)

Similarly, Roy Macridis, a comparativist in political science, critiques the approach’s subject matter and method while focusing on the study of comparative government. He claims that, comparative government was `excessively formalistic in its approach to political institutions’, did not have ‘a sophisticated awareness of the informal arrangements of society and of their role in the formation of decisions and the exercise of power’; was ‘insensitive to the nonpolitical determinants of political behavior’; was ‘descriptive rather than problem-solving, or analytic in its method’; was insensitive to hypotheses and their verification; and therefore, was unable to formulate a comparative `political theory of dynamics’ (Macridis 1963, 47-8 in Rhodes 1995, 48).

The historical methods and legal analysis of the classic institutional approach are inadequate. Historical methods cannot explain systematically the structure and behaviour of governments due to its focus on the unique. The gap between the formal statements of the law and the practice of government renders legal analysis ineffective (Rhodes 1995).

David Easton, the most influential critic of the traditional study of politics, found the classic institutional approach wanting on two grounds.

First, the analysis of law and institutions could not explain policy or power because it did not cover all the relevant variables (Easton, 1971, ch. 6) Second, ‘hyperfactualism’, or ‘reverence for the fact’ (p. 75), meant that political scientists suffered from ‘theoretical malnutrition’ (p. 77), neglecting the general framework within which these facts could acquire meaning (p. 89) (Rhodes 1995, 49).

Other critics noted that the approach was concerned with the institutions of government, and yet operated with a restricted understanding of its subject matter. Its focus was on formal rules and organizations rattler than informal conventions and on official structures of government rather than broader institutional constraints on governance (within and outside of the state (Lowndes 2002).

By the 1980s, the traditional or classic institutional approach has declined in its importance in political science. March and Olsen (1984, 734 in Lowndes 2002, 94 and Rhodes 1995, 53) coined the term “new institutionalism” critiquing the traditional or classic institutional approach as having “‘receded in importance from the position they held in the earlier theories of political scientists’.” Asserting that political institutions played a more autonomous role in shaping political outcome, they make claims that,

The bureaucratic agency, the legislative committee, and the appellate court are arenas for contending social forces, but they are also collections of standard operating procedures and structures that define and defend interests. They are political actors in their own right (March and Olsen 1984, 738 in Lowndes 2002, 94 and Rhodes 1995, 53).

In contrast to the traditional or classic institutional approach, now referred to as the “old institutionalism,” new institutionalism has a much broader, yet sophisticated definition of its subject matter.

Political institutions are no longer equated with political organizations: `institution’ is understood more broadly to refer to a ‘stable, recurring pattern of behavior (Goodin 1996:22). The new institutionalists are concerned with the informal conventio
ns of political life as well as with formal constitutional and organizational structures. New attention is paid to the way in which institutions embody values and power relationships, and to the obstacles as well as the opportunities that confront institutional design. Crucially, new institutionalists concern themselves not just with the impact of institutions upon individuals, but with the interaction between institutions and individuals (Lowndes 2002, 91).

New institutionalists argue that institutions do matter. In their seminal article on new institutionalism, March and Olsen (1984) emphasized the central value of institutions vis-a-vis individual choices in explaining political phenomena. They argue that political behavior is “embedded in an institutional structure of rules, norms, expectations and traditions that several limited the free play of individual will and calculation” (March and Olsen 1984, 736). Burnham et al., (2004, 18) captures it succinctly: “political phenomena could not be simply reduced to the aggregate consequences of individual behavior; rather, “the choices that people make are to a significant extent shaped by the institutions within which they operate.” 

There are several variants of new institutionalism reflecting the divide between ‘normative’ approaches and a new, more sophisticated version of rational choice.

Normative institutionalism argues that political institutions influence actors’ behavior by shaping their ‘values, norms, interests, identities and beliefs’ (March and Olsen 1989:17). Hence ‘normative’ refers to a concern with norms and values as explanatory variables (owing much to the traditions of sociological institutionalism), and not to ‘normative theory’ in the sense of promoting particular norms.

Rational choice institutionalism denies that institutional factors `produce behavior’ or shape individuals’ preferences, which they see as endogenously determined and relatively stable (favoring utility maximization). Political institutions influence behavior by affecting ‘the structure of a situation’ in which individuals select strategies for the pursuit of their preferences (Ostrom 1982: 5-7). Institutions provide information about others’ likely future behavior, and about incentives (and disincentives) attached to different courses of action (Lowndes 2002, 95).