SciELO - Scientific Electronic Library Online

vol.4 edição especial índice de autoresíndice de assuntospesquisa de artigos
Home Pagelista alfabética de periódicos  

Brazilian Political Science Review (Online)

versão On-line ISSN 1981-3821

Braz. political sci. rev. (Online) vol.4 Rio de Janeiro  2009


Preference formation and institutional change*



Sérgio Praça

University of São Paulo (USP), Brazil




This essay critically analyses how historical institutionalists and rational choice scholars study institutional stability and change. Special attention is paid to the thorny issue of how political actors' preferences are formed, with historical institutionalists considering preferences as endogenously formed, and rational choice analysts postulating that preferences are fixed and exogenous. An argument is made in favour of the perspective that considers preferences as being formed within the functioning of the political system over time, endogenously. The essay also proposes the incorporation of ideas and non-decisions as tools to elucidate processes of change.

Keywords: Neo-institutionalism; Rational choice; Historical institutionalism; Political preferences; Institutional change.



It is no exaggeration to say that Brazilian political science has undergone an institutionalist revolution over the last two decades. Analysts have progressively considered institutional variables to the detriment of cultural variables to analyse elections, the behaviour of political parties in the Legislative, presidential choices, social policies and a plethora of other substantive matters linked to the functioning of the Brazilian political system.

However, although the incorporation of institutional variables has dominated many of the themes in Brazilian political science in recent times, leading to divergent diagnoses of the functioning of our coalition presidentialism (Abranches 1988; Figueiredo and Limongi 1999; Ames 2001; Santos 2001; for a review, see Palermo, 2000), less attention has been paid to theoretical issues within institutionalism.1 There is little we know about how institutions are sustained over time and the mechanisms that lead to change within them, though there is recurring use of the concept of path dependence in some works (Couto 1998; Fernandes 2002). Not just in Brazil is this a feature of political science: there is reasonable consensus among rational choice institutionalists and historical institutionalists2 that mechanisms for institutional stability and change are far from being well understood.

This essay aims to critically analyse how historical institutionalists and rational choice scholars have studied institutional stability and change and to propose the more effective incorporation of ideas in circulation and non-decisions as tools to elucidate processes of change. Special attention will be paid to the thorny issue of how political actors' preferences are formed, a point of discord between these two lines of analysis, with historical institutionalists considering preferences as endogenously formed and rational choice analysts postulating that preferences are fixed and exogenous.

I begin by analysing how the old institutionalists used the perspective of punctuated equilibrium, at times allied to the concept of path dependence, to explain institutional change. After discussing preference formation mechanisms, I argue that historical institutionalism displays some advantages when analysing institutional change , though two challenges still remain.


Punctuated Equilibrium: The Old Institutionalist Perspective

Punctuated equilibrium, the first analytic perspective utilized by institutionalists to study institutional change, can certainly find various substantive examples that justify its use. As Krasner (1984, 234) puts it, this perspective may be thus described: institutional structures do not respond in an immediate, fluid and incremental fashion to alterations in the domestic and/or international environment. Change is seen as rare, difficult, episodic and dramatic, and not as continuous and incremental. Political crises are catalysts and, once they are over, the institutional arrangements that result from them respond more to incentives generated endogenously than to external phenomena. The adoption of an electoral system that organizes congressional elections in a given country can be a good example. Majoritarian electoral systems encourage a type of coordination between political parties and other actors that can differ enormously from the type of coordination (or lack thereof) under proportional representation systems. Maurice Duverger's (1954) classic "laws" clearly associate the type of interaction we can expect among political actors under a certain electoral system.

It is not difficult to imagine that parties, interest groups, potential candidates, deputies, mayors, campaign financers and other actors organize their resources and focus on aspects X and Y to the detriment of others, foreseeing which will be most important, under a given electoral system, to ensure success in elections, success in the adoption of certain public policies and success in terms of participation in a governing coalition if possible, to cite the three main political objectives considered by Müller and Strom (1999). Therefore, political actors act in pursuance of certain goals considering the institutional constraints imposed by the system. Adapting to these constraints and obtaining political success is not an easy task. Therefore, once actors learn how to act and obtain electoral success under a certain type of system, changing it is rare. Katz (2005, 58) states that only 14 countries that have been democratic since 1950 have adopted a new electoral system. Discreet reforms, undertaken within the existing electoral system, are more numerous.

Electoral systems are institutions whose wholesale replacement implies high costs, for they operate as points around which political actors exchange information about their behaviours when interacting, thus strengthening certain expectations. Therefore, the cost of maintaining as unsatisfactory institution may be lower than that of creating and reproducing another institutional design. From this rationale flows the statement, recurrent in the historical institutionalist literature, that initial choices about institutions constrain the number of institutional alternatives in the future. In Krasner's (1984, 240) words, "once a critical choice has been made, it cannot be taken back. There may be a wide range of possible resolutions of a particular state-building crisis. But once a path is taken it canalizes future developments. (...) past choices preclude certain strategies or make them very costly".

Reinforcing this point, Pierson (2004, 142-153) points out three factors that encourage the stability of institutional designs: a) problems of coordination:3 there is great intrinsic difficulty in coordinating actors around an alternative to the existing institutional solution;4 b) number of veto points: political actors design institutions that are hard to change in order to tie not only their own hands5 but also their successors' hands. They give veto powers to actors that could be harmed by institutional change. It is important to identify the actors that have veto powers and resources at their disposal, for it is possible that an actor is deeply opposed to a certain change but only has a voice and no vote; c) positive feedback: actors adhere more strongly over time to existing institutional arrangements. Commitments and interaction with other actors become more credible and are strengthened over time. Hence, the incentives to alter institutions diminish.6 Furthermore, actors' preferences about institutional arrangements are dynamic, not static. Actors develop skills, behaviours and interactions that work very well under the present institutions and not under others, or it is unknown whether they work well under other institutions and it would be rather costly to find out. Inter-linkages between diverse institutional arrangements also affect the desire for reforms, acting as another positive feedback mechanism.

There is no doubt as to the fact that the punctuated equilibrium line of analysis explains certain institutional choices well, above all those with highly important implications for political actors' coordination and expectations about the behaviours of others. But this perspective has at least three serious limitations: it says little about ongoing reproduction mechanisms that ultimately allow a certain institution to sustain itself over time; it often stresses the centrality of "critical junctures" whilst not making explicit when we can expect them to occur and not stating whether the critical juncture takes place due to a single political event (e.g., the unveiling of a corruption scandal) or to a conjunction of several processes (e.g., a scandal plus a change in macroeconomic policy); and, lastly, it hugely underestimates institutions' capacity to change little and/or in gradual fashion.

Pierson (2004) points out four lines of argument that seek to explain institutional stability and change. Beyond the punctuated equilibrium perspective, he considers: a) the line that considers political "losers" as catalysts of change; b) the line that emphasises the importance of various simultaneous processes for change to occur; and c) the line that focuses the role of entrepreneurs. It is interesting to point out what one line clarifies and another blurs, whether the lines are complementary or contradictory, their advantages and limitations. The punctuated equilibrium perspective has had its main flaws presented. Its two main strengths are: considering exogenous factors as impacting for purposes of institutional change and considering institutional change over time, albeit ignoring the importance of small, gradual and cumulative changes.

The line of argument that considers the importance of political losers as agents of change has the merit of showing how political conflicts internal to institutions affect their changes, but displays two disadvantages: it minimizes changes that are of interest to the political "winners", i.e., to the best-placed political actors in a certain institution, and disregards the fact that certain institutional changes can be of interest both to the "winners" and to the "losers".7

The perspective that emphasises the importance of several simultaneous processes for change to occur, in turn, has the enormous advantages that come with considering internal political conflicts as well as exogenous causes, typical of punctuated equilibrium explanations. However, it has two limitations: there are perhaps few institutions that can plausibly be affected by various processes simultaneously and it is extremely difficult to forecast when change tends to occur.

Lastly, there is the analytic perspective that takes into account the role of entrepreneurs, with the advantage of showing that certain political actors — well-prepared and well-positioned — can guide processes of institutional change. It has the flaw of exaggerating the power of certain entrepreneurs, excessively minimizing the importance of factors both exogenous and endogenous to political conflict as catalysts of change, as well as the possibility that certain institutions are the object of so little dispute that it would be unlikely for a political actor to use up resources in order to change it.

All the lines of argument spelled out can complement one another. Schickler (2001), for example, places great importance on entrepreneurs of changes in the standing orders of the US Congress, while at the same time analysing in detail the structure of the internal political conflict that made their emergence possible.

Pierson (2004) sets off from this review of lines of analysis to identify the five main problems with the institutionalist literature on change: 1) little notion about when one can expect change to happen: the typical answer of public policy literature is given by Kingdon (1995) when he states that change occurs when various processes ("streams") converge to make a certain subject salient;8 2) little notion about when one can expect a certain pattern of institutional change rather than others; 3) selection bias: the cases selected by institutional change analysts may be completely atypical (Pierson, 2004, 140), incurring in what Geddes (2003) calls selection by the dependent variable — in other words, by choosing cases where change took place, scholars fail to notice cases where change did not occur;9 4) tendency to focus the analysis on catalysts of change, observing only the last moment of what might be a very long process; 5) excessive focus on actors' pressure for institutional reforms, whilst disregarding those who advocate institutional stability.10

All these problems will be tackled, directly or indirectly, over the course of this essay. Returning to the punctuated equilibrium perspective, it is not rare to find authors who associate the notion of rare and drastic change to the concept of path dependence, one of the concepts used in the social sciences most often and with the least rigour, frequently limited to the vague notion that "the past matters" to explain a certain institutional design.11 This association occurs because analysts who use the punctuated equilibrium perspective somehow need to answer the following question: if an institution rarely changes, what explains so much stability? One of the flaws in the institutionalist literature pointed out by Pierson (2004), and mentioned above, is precisely that there is a very strong tendency to focus the analyses only on the catalysts of change, observing just the tip of what may be a very long process. Therefore, one must deepen the analysis of the mechanisms of institutional reproduction that may be going on. Stating that the trajectory of a certain institution depends on the initial path chosen is an answer that dodges the issue. Other extremely pertinent questions follow from this one: how can one state that a certain institutional process is path dependent? Can every observation of institutional stability be explained by this concept? Which mechanisms of institutional reproduction are most common, and in which contexts? In the next section, I analyse these points based on Mahoney's (2000) essay, seeking to evaluate the heuristic reach of the concept of path dependence as well.


Path Dependence and Institutional Reproduction Mechanisms

As has already been mentioned, in the current literature there are various uses of the concept of path dependence. The most frequent is the one I call 'soft use', which states that the "key claim of path-dependent arguments is that these earlier choices create legacies or institutions that last a long time and are very difficult to reverse. Thus earlier choices change the costs and benefits associated with later choices, and may even determine the existence of later choices" (Geddes 2003, 139). Although not mistaken, this definition is incomplete when compared to Mahoney's (2000, 507), according to whom a path dependent process can be identified as a historical sequence in which contingent events cause institutional patterns or chains of events with deterministic properties. Highlighting the importance of contingent events as initial triggers and arguing that the historical sequence can be both of a causal nature (reactive sequence) and of a self-reproducing nature (self-reinforcing sequence) are two highly important points in better characterizing path dependent processes and being able to separate them from institutional continuities better explained by other arguments.

According to Mahoney (2000, 510), there are three characteristics that define path dependence: 1) it is a causal process extremely sensitive to events that occurred in the initial stages of a certain historical trajectory; 2) the initial historical events are contingent,12 and cannot be explained by previous events or "initial conditions"; 3) the path dependent trajectory is also characterized by relatively deterministic causal patterns or what may be called "inertia": once the process that will determine a certain political result begins, this process tends to remain constant until the result is achieved. The nature of this inertial process varies: in a self-reinforcing sequence, the inertia is characterized by mechanisms that reproduce certain institutional patterns over time; in a reactive sequence, the inertia is characterized by mechanisms of action and reaction that confer upon the chain of historical events an "inherent logic" according to which an event leads naturally to another.

The sequence of the self-reinforcing type is the one of most interest to this essay. It can be described as follows: at point 1 in time, there are multiple alternatives — A, B and C — available to the political actors, the theory being unable to forecast or explain which option will be adopted; at point 2 in time, during a critical juncture, option B is favoured over others, and this event is contingent; lastly, in the later stages (at points 3, 4 etc in time), option B profits from its initial advantage and is reproduced in a stable manner (Mahoney 2000, 514).

The main characteristic that differentiates path dependent processes from other institutional reproduction processes is the fact that the initial event must be contingent, and cannot be explained by previous events or "initial conditions". This is the probable motive why path dependent processes are rare: it is hard to find an event that cannot be plausibly linked to a previous one. This may also be the reason for much of the confusion surrounding the concept, for institutional reproduction processes that were not initiated by contingent events are often defined by analysts as path dependent. But what is a contingent event after all? According to Mahoney (2000, 513),

[...] contingency refers to the inability of theory to predict or explain, either deterministically or probabilistically, the occurrence of a specific outcome. A contingent event is therefore an occurrence that was not expected to take place, given certain theoretical understandings of how causal processes work.

It is interesting to note that the concept of critical juncture, previously associated in this essay with explanations of the punctuated equilibrium type, reappears in this discussion on path dependence in a restricted and analytically rigorous fashion. Whilst in the eyes of scholars who espouse the former perspective critical junctures only seem to be the various periods when some kind of significant institutional change/choice takes place, in path dependent processes critical junctures are moments in which contingent events occur. They can be defined, according to Mahoney (2000, 513), through counterfactual analysis, with the analyst imagining how the institutional trajectory would have taken place had a different option been selected. In other words, the critical juncture is characterized as the moment at which a contingent choice is made, leading to rather different results from those that would be plausible after other, theoretically imaginable choices.

After the contingent event results in a certain institution, what are the mechanisms responsible for its reproduction? Much of the institutionalist literature, as Pierson (2004, 142-153) has argued, habitually attributes the stability of a certain institutional design to problems of coordination between actors, to the large number of veto points and to positive feedback that encourages actors to adhere to the institution already in existence, thus making the adoption of new arrangements less likely. These three factors are an excellent starting point for one to think about institutional reproduction, but they do not detail the mechanisms responsible for it to happen. Does the positive feedback of a certain institution affect all actors in the same manner? Could a certain institutional arrangement be considered legitimate by some and absolutely misguided by others?

By identifying four useful theoretical perspectives to explain how institutional reproduction occurs in path dependent processes — utilitarist, functionalist, power-related and legitimacy-related — Mahoney (2000, 516) helps one disaggregate the reproduction mechanisms. According to the utilitarist perspective, a certain institution is reproduced because the potential benefits of change are seen as smaller than the costs of change. This argument about the costs of change is used both by "first generation" institutionalists like Krasner (1984, 235) and by rational choice scholars.

The functionalist perspective in turn establishes — in a manner similar to what Pierson terms positive feedback — that institutions are reproduced because they have consequences thought of as positive for the political system. These include integration among actors, incentives to adaptation and incentives to survival. The institution serves a certain function; the institution expands because this function is seen as positive; the institution is ever more capable of performing this function and thus becomes consolidated over time. The "strong" functionalist version considers these positive consequences also to be the causes of institutional reproduction.

According to the legitimacy-related perspective, once the institution is selected in contingent fashion, it will be reinforced by means of processes of progressive legitimation, even if other previously available institutional alternatives had been more legitimate (Mahoney 2000, 523). With regard to changes in the Brazilian budgetary process from 1988 to 2006, it is interesting to note that budget amendments proposed individually by congressmen were progressively de-legitimized after two scandals in 1993 and 2006, and this did not result in the end of this institution. Rather, it resulted in the creation and prioritization, over time, of collective amendments (proposed by state caucuses and parliamentary committees). There are two interesting lessons to be drawn from this episode: the process of de-legitimization of individual amendments was one of the causes of the institutional layering (Streeck and Thelen, 2005; more on this in the fourth section of the essay) and the four perspectives about institutional reproduction set out by Mahoney (2000) are not necessarily exclusive — in this case, the power-related mechanism of institutional reproduction acted simultaneously with the legitimacy-related mechanism.

The power-related perspective (Mahoney 2000, 521), like the utilitarist perspective, assumes that actors make decisions considering the relative costs and benefits generated by the institution. But the power-related explanation emphasises the fact that institutions distribute costs and benefits unequally throughout the political system, and also stresses that actors with different resource allocations will have conflicting interests with regard to institutional reproduction. In other words, it is likely that losers in the political arena will try to change institutions. According to this perspective, an institution can persist even when most individuals or political groups would like to change it, as long as a certain elite benefits from the institutional arrangement in force and has enough power to ensure its reproduction. This perspective offers the advantage of considering the process of institutional stability and reproduction something dynamic, in which the same actors can have different preferences regarding a certain institution over time. This point will be further developed in the next section of this essay.

A possible example of path dependent processes to think about is that of certain economic institutions established in Andean countries in the early 20th century with the help of US economist Edwin Kemmerer, a subject studied by Drake (1989). Kemmerer's trips to Bolivia, Chile, Colombia, Ecuador and Peru between 1923 and 1931 may be considered contingent events, for no theory can forecast the adoption of institutions simply because a specialist suggests so. The path dependence perspective, as we have seen, has as its differential precisely this type of contingent, unexpected occurrence as the trigger for institutional choices.13 In this case, Kemmerer acted as the entrepreneur of certain institutional ideas (the creation of central banks, for example), making the most of the fact that several actors of the political and economic fields in each country had already advocated similar ideas previously (Drake 1989, 3).

In the next section I argue that historical institutionalism, owing to the fact that it explicitly considers preference formation as something endogenous to the political system — that is, something that alters over time due to factors internal to the system —, presents a certain advantage in relation to rational choice institutionalism for the study of institutional change, since the latter analytic line tends to consider preferences as exogenously fixed.


Preference Formation: Differences between Rational Choice and Historical Institutionalism

This session tackles the issue of which institutionalist analytic perspective is best equipped to study institutional change. Out of the three "neo-institutionalisms" (Hall and Taylor 2003) in vogue — namely, rational choice institutionalism, sociological institutionalism14 and historical institutionalism — the latter has the advantage of considering political actors' preferences as to institutional designs as something formed endogenously over time, rather that taking the view that these preferences are exogenously fixed, as do rational choice analysts.

The rational choice analytic line is guided by four main aspects as to the study of institutions. The first, referent to the definition of the problem, is that institutions are considered preference aggregation mechanisms, with the basic functions of coordinating political actors and solving collective action problems. In the words of Peters (1999, 45), this analytic line considers institutions as capable of conferring a certain regularity of results that benefit all the participants. These results would necessarily be irrational if obtained on the basis of the aggregation of isolated individual actions, without the coordination of the institutional arrangement. In Thelen's (1999) view, this line argues that institutions are considered coordination mechanisms that generate or sustain certain equilibrium.

In the second aspect, regarding the relationship between preferences and institutions, rational choice institutionalists consider political actors to have preferences that are fixed and determined exogenously to the institutions to which they belong, and that are basically oriented for these actors to maximize their utility through votes. Rational choice scholars do not usually consider other mechanisms for the formation of these preferences (Peters 1999, 15).

With reference to the third aspect, the origin of institutions, rational choice scholars consider that institutions are formed by a group of political actors aware of the functions that a certain institution should have, in a sort of tabula rasa (blank slate). The actors have the prerogative of creating or destroying institutions without having to worry about the historical processes and structural context within which they find themselves (Peters 1999, 54-55).

Lastly, according to Peters (1999, 56), most analysts from this perspective consider institutional change as exogenous and it is generally ignored, except as a new problem to model once it occurs.15 And even if the change is considered in this sense, practically the only cause suggested for it to occur is that the institution ceased performing a certain function for which it was designed.

On the four points considered above, historical institutionalists steer the following course. Starting off from a certain understanding, originating in the social sciences and in economics, of institutions as "formal rules, compliance procedures, and standard operating practices that structure the relationship between individuals in various units of the polity and the economy" (Hall 1986, 7), this line ascribes great importance to the role of historical processes for the operation of political institutions. This line does not necessary deny all the assumptions of rational choice institutionalism — the role of institutions as generators of equilibrium, for example (Immergut 1998, 7) — but offers tools of analysis to contextualize the history and ideas that allow institutional equilibrium X rather than Y to be achieved.

As for the relationship between preferences and institutions, also to be dealt with further on, historical institutionalists consider that political actors do not possess fixed preferences. Rather, preferences are determined in a fashion endogenous to institutions, and may vary over time. On the matter of the origin of institutions, historical institutionalists view them as being formed by historical processes, not depending much on the intentions of the actors that organized them initially (Thelen 1999). Lastly, this analytic perspective, as seen in the first section of this essay, used to consider institutional change as resulting from exogenous shocks, configuring situations of punctuated equilibrium. More recently, analysts from this line have turned their attention to establishing tools of analysis to better understand how institutional change can be generated endogenously (Mahoney and Thelen 2009).

The point to discuss now regards how political actors' preferences are formed according to these two analytic perspectives. In the eyes of Katznelson and Weingast (2005, 3), political preferences have been described and understood in three ways by the institutionalist literature: a) the theory indicates the preferences of the relevant actors in exogenous fashion; b) preferences are thought of as caused by historical processes; c) preferences are believed to be induced by strategic circumstances and the interaction between actors within certain institutions and/or social processes. This interaction encourages or persuades a certain actor to have a specific kind of preference.

Rational choice literature, as already mentioned, is the one most associated with the exogenous indication of actors' preferences, but Katznelson and Weingast (2005) warn that in the case of some preferences — like US congressmen's preference for re-election —, this exogenous indication was sufficiently preceded by careful analyses that consider institutional contexts, such as Mayhew's (1974). If all the literature on the US Congress takes representatives' re-election as an axiom, it is because this has a strong empirical and historical basis.16 This point notwithstanding, it is not unfair to consider that historical institutionalists pay far more attention to the endogenous mechanisms of actors' preference formation than do rational choice scholars. Katznelson and Weingast (2005, 2) seem to point to a certain convergence around historical concerns to understand how preferences are formed, stating that

We know too little about preferences, where they come from or how they are generated. (...) Historical and rational choice institutionalist scholars have been converging on the idea that because institutions often generate sufficiently strong incentives for actors, whether medieval kings or members of the modern US Congress, we can derive a form of preference based on the compelling logic of institutions embedded in particular historical situations; or at least come to understand how a given institutional milieu both constrains and shapes the repertoire of available preferences.

In spite of this relatively recent convergence, Thelen (1999, 375-377), while praising the fact that rational choice analysts have progressively incorporated cultural and/or contextual factors to determine actors' preferences, indicates that one of the points that differentiates historical institutionalism is that it considers that individuals' interests have structural, endogenous origins. Therefore, individuals' strategic interaction need not be analysed. She contrasts this with the notion that rational choice analysts emphasise that political results need to be understood considering the actions and behaviours of individuals that behave strategically.

For Geddes (2003, 183), one is dealing with a question of semantics.

The everyday meaning of 'preferences' includes both the kinds of underlying goals that are referred to as preferences in the rational choice idiom and also attitudes toward (preferences about) choices or activities that would help achieve the goals. These attitudes are not referred to as preferences in the rational choice idiom; they are called strategies, strategic choices, or, occasionally, second-order preferences. They include policy preferences, institutional preferences, and most other preferences about real-life choices. Second-order preferences are choices of strategies for achieving first-order preferences. (...) The politician's first-order preference for remaining in office remains unchanged, but he rationally picks the policy or institutional strategy he considers most likely, in the circumstances he faces, to help him achieve that goal.

Hence, second-order preferences are almost always endogenous in rational choice literature, but they are called strategies rather than preferences. For their part, first-order preferences are considerably stable over time (Geddes 2003, 182), and even if they change, this can be easily incorporated into the analysis.

This link between second-order preferences (about policies, about institutions) and first-order preferences (about re-election, about remaining in political office) is often obscure or irrelevant to the actor's decision about the second-order preference. It is an exaggeration to state that a certain actor prefers institution X rather than Y simply because he/she wants to be re-elected, and not because, after the re-election, he/she simultaneously wishes to pursue a certain macroeconomic result and/or fight corruption, for example. Not every institutional choice is strategic; choices may be made on the basis of other criteria. Therefore, reducing institutional choice to what Geddes calls first-order preferences means limiting excessively the plausible motives why political actors choose what they choose.

After all, why is it important to understand how political actors' preferences are formed? Mainly for two reasons. The first is: in case there is strong evidence that actors' preferences for a certain institution are influenced by exogenous factors, or that they may be considered fixed in advance and practically immutable over time, the tools of historical institutionalism will be of little help to the analysis. The second is that the analysis of preferences at moments of critical juncture can reveal practically the full range of institutional options at actors' disposal (Katznelson 2003).17

In order to study institutional change, the notion that preferences are formed endogenously can be analytically fruitful and is slowly being shared by scholars who follow both historical institutionalism and rational choice institutionalism. The main challenge is to evaluate on a case-by-case basis what each analytic perspective elucidates and conceals. Some of the advantages of historical institutionalism for the study of institutional change will be listed and contrasted with the rational choice perspective below.


Advantages and Limitations of Historical Institutionalism

Although Peters (1999, 70) states that historical institutionalism is not a fertile perspective with which to reflect upon institutional change due to its strong emphasis on the stability of institutions, it is undeniable that various scholars have recently concerned themselves with explaining causes and processes of institutional change (Mahoney 2000; Thelen 2003; 2006; Pierson 2004; Mahoney and Thelen 2009). Beyond the theoretical and analytic gain of considering actors' preferences as endogenously formed, what other advantages does the historical institutionalist perspective offer to the study of institutional change? In this section, I offer two interconnected arguments. The first is that historical institutionalism offers some especially interesting analytic tools to study institutional change. The second is that it supplies useful analytic tools for understanding the dynamics of change. This argument is grounded in the work of Streeck and Thelen (2005), who, in turn, find support in authors like Hacker (2004) and Schickler (2001) to construct a fertile typology of institutional change processes.

The advantages are made particularly clear when placed side-by-side with rational choice analysts such as Shepsle (2006), for whom political institutions are "fixed exogenous constraints", rarely alterable over time; there is no chance of political actors behaving according to informal institutions, for behaviours that deviate from the formal rules would imply rethinking of the very definition of institution.18

At this point it is worth pausing for a word on informal institutions and the very definition of institution. Shepsle (2006) refers to Russian parliamentarians as political actors who "constantly test how much they can get away with", concluding that the Duma "hardly seems 'institutionalized' at all". In other words, he discards informality as an empirical regularity to be analysed. To be fair to rational choice analysts, though, authors like Douglass North (1990), do incorporate informal institutions, whilst icons of historical institutionalism such as Streeck and Thelen (2005, 10) state:

Informal institutions by no means exist only in premodern societies; in fact informal norms enforced by community disapproval are universally present in social life. They are, however, not the subject of our study. This is because to the extent that modern economies are political economies — that is, governed by politics — they are mainly controlled by norms and sanctions that are formalized.

By saying this, however, these authors fail to consider the fact that various formal changes are preceded by informal arrangements. They also disregard the fact that informal institutions may constitute equilibria, as Knight (1992) convincingly argues.

Historical institutonalists such as Tsai (2006) and Lindner (2003) incorporate informal institutions into explanations of institutional change. Also, the recent theoretical contribution by Mahoney and Thelen (2009) strongly incorporates informality into historical institutionalism understanding how the compliance of certain actors with an institution can vary, thus establishing an elegant dialogue between informality and the theoretical corpus of historical institutionalism.

According to Mahoney and Villegas (2007, 74-81), historical institutionalism also offers specific tools that can be useful in studying change, such as techniques to analyse necessary and/or sufficient causes, and procedures to test hypotheses in the specific analysis of a case; a central concern with the temporal dimensions of political explanation; the possibility for the analyst to greatly deepen his/her case study with empirical material from primary and/or secondary sources; and, lastly, the proposal of hypotheses based on the intersection of various causal processes.19

When two or more causal processes come together at a certain moment, this can have a major impact on institutional dynamics. Again, Brazilian budgeting institutions offer an interesting case study. It may be considered that the cash limit to individual budgeting amendments, which began in the 1990s, happened because of two parallel processes: the response of the Brazilian political system to corruption in the Joint Budget Committee, unveiled in 1993 (Krieger, Rodrigues and Bonassa 1994), and the macroeconomic policies initiated by President Fernando Henrique Cardoso, in an effort to strengthen the Executive's control over the budgetary process. In this case, considering the limitation in amendments as a response to scandal overestimates the responsiveness of the Brazilian political system. But solely considering this limitation as a macroeconomic choice makes the analyst ignore the parallel — and relevant — political process started by the 1993-1994 Parliamentary Inquiry Committee (CPI). The most adequate logic is to consider both processes.

Following the differentiation proposed by Taylor (2009) between causes and processes of institutional change, I now turn to Streeck and Thelen's (2005) analytic framework of processes of change. They express concern over the current state of studies on institutional change when they state that

[...] we must avoid being caught in a conceptual schema that provides only for either incremental change supporting institutional continuity through reproductive adaptation, or disruptive change causing institutional breakdown and innovation and thereby resulting in discontinuity. (...) we argue that equating incremental with adaptive and reproductive minor change, and major change with, mostly exogenous, disruption of continuity, makes excessively high demands on 'real' change to be recognized as such and tends to reduce most or all observable changes to adjustment for the purpose of stability (Streeck and Thelen 2005, 8).

They propose five possible processes of institutional change that necessarily have an endogenous origin, which does not prevent exogenous shocks from serving as triggers to initiate change as well. The processes are: 1) displacement: traditional arrangements are gradually replaced by new "logics of action"; 2) exhaustion: institutional arrangements contain devices that, if not updated, lead to the disappearance of these arrangements; 3) drift: institutional designs with different impacts at times A and B are maintained, resulting in different consequences;20 4) layering: new institutional arrangements are gradually superimposed upon previous arrangements; and 5) conversion: institutional arrangements have objectives, competencies and/or prerogatives redefined over time in order to attend to new political interests21 (Streeck and Thelen 2005, 19).

Their proposal is interesting for two reasons. The first is the fact that it encompasses several points already covered by other analytic perspectives and/or other authors linked to historical institutionalism.22 In a process similar to what Streeck and Thelen (2005) characterize as drift, sociologists Clemens and Cook (1999, 451) state that institutional learning can lead old rules to obtain new results — in other words, different effects with the same institutions. Shepsle (2006, 16), in turn, in a contribution of the rational choice line to understand mechanisms and processes of institutional change, proposes seven possible causes, among them the "incorporation of new political actors", something that Thelen (2003, 232) associates with conversion.

Streeck and Thelen's (2005) framework also allows one to think how political conflicts can lead to partial changes in institutions, in the form of what they call layering, following Schickler (2001). One is dealing here with the partial renegotiation of some elements belonging to a certain institutional design, while other elements remain unaffected (Thelen, 2003, 225). The layering process does not necessarily result in incongruent institutions like most of the ones Schickler (2001) studies. In Brazil, a case of congruent layering took place in the budgetary process from 1993 onwards, with respect to the coexistence of individual and collective amendments to the budget. In relation to when we can expect layering to occur, Thelen (2003, 232) points out that this process of institutional change can be associated to empirical cases in which the political context changes and new challenges come to the fore, but the vast majority of political actors are the same.

In spite of the pertinence of the Streeck-Thelen proposal to explain several empirical cases of institutional change, there is a great challenge ahead, partially considered by Boas (2007). It is integrating this analytic perspective with path dependent processes. In cases of drift, for instance, is there institutional change or reproduction? After all, the institution retains its design but obtains different results. As we have seen, some central banks studied by Drake (1989) did not necessarily change their design to change function. It seems to be possible to integrate at least the process of drift laid out by Streeck and Thelen (2005) with a more general model of path dependence, contrary to Boas' (2007, 34) opinion.

It is worth remembering that the theoretical gain in considering the distribution of power and resources as being determinants in the stability and change of institutions is shared by only a few scholars of the rational choice analytic line, such as Levi (1990) and Knight (1992). Particularly the concept of contingent consent proposed by Levi (1990) speaks to historical institutionalism. Her main theoretical point is that certain actors' lack of compliance with an institution can be a driver of institutional change. There are two factors, according to Levi (1990, 410-411), that can compel some actors not to comply with a certain institutional design. The first is that these actors may obtain new information or new power resources so as to cultivate an interest in changing the institution in order to achieve more favourable results. The second is that trust in the compliance of the other actors with the institution can diminish drastically, above all when those who most benefit from the existing institutional design are caught abusing the power they have. And this would lead the other actors also not to comply with the institution.

However, Levi's (1990) argument contains a hint of tautology. Some actors may want to change the institution so as to obtain from it more favourable results, or because they feel that the other actors do not comply satisfactorily. The resources available to these actors are obviously important, but how would the internal distribution of these resources change for the formerly disadvantaged actors then to be able to carry out institutional change? Levi (1990) does not provide a clear answer. In a recent text (Levi, 2009), she gloats that she is one of the few rational choice scholars to takes account of the internal distribution of power to understand the functioning of institutions, but does not elaborate on how this distribution of resources can be altered is such a way that brings about institutional change.

A good theoretical complement can be the also recent argument by Mahoney and Thelen (2009) about institutional dynamics. While they strongly consider the role of actors' compliance — which brings them rather close to the argument by Levi (1990) about contingent consent — they set out clear hypotheses on the types of actors that can be led not to comply with a certain institution and under which circumstances this is most likely to occuur. Therefore, the rational choice and historical institutionalist analytic lines complement one another interestingly in this discussion.

In the next section I deal with two more general challenges to the study of institutional change, namely, incorporating ideas in circulation in more systematic fashion and paying more attention to non-decisions.


Challenges: Incorporating Ideas and Non-Decisions

For some time now, historical institutionalists have ascribed an important role to the analysis of ideas in institutional change (Hall 1993). Recently, some authors have proposed a return to ideas as the central focus of studies on change, arguing that the "conflict of ideas" is fundamental for one to understand why certain institutional alterations occur and others do not (Peters, Pierre and King 2005). But the use of ideas strictu sensu to study institutional change is pertinent only for institutions that clearly have strong associations with certain political results. An independent central bank, for example, can easily be identified with low inflation rates if compared with a politically-controlled central bank. Matters such as inflation and the extent of state intervention in the economy are extremely polemical. They arouse debate and a broad spectrum of ideas, solutions and political proposals. However, the vast majority of political institutions are not clearly associated with a certain political result about which ideas circulate. This is so not just because the design of certain institutions (budgetary institutions, for instance) does not tend to be politically salient on the agenda, but also because the association between a type of institutional design and a certain political result is rarely clear. What kind of budgetary process leads to fiscal deficits?23 Do centralized legislatures generate less clientelistic public policies than parliaments organized in decentralized fashion?

The debate on institutional design is often restricted to the institutions themselves, with very little circulation of ideas about them in society or even in academia. It is interesting to note that while Pierson (2006) emphasises the fact that certain public policies have the characteristics of institutions, the opposite is also true, as certain institutions are characterized, like public policies, by the fact that they encourage the formation of ideas and proposals about their design. Other institutions are characterized by the comfortable inertia of discretion.

Pierson (2004) states that one of the factors present in economic markets and that might lead institutions to perfect themselves, i.e., competition, is either rare or inexistent in the field of politics. He claims this happens because "in most cases, institutions are not really subject to direct competition at all. Instead, single institutional arrangements, or sets of rules, typically have a monopoly over a particular part of the political terrain" (Pierson 2004, 128). The author is certainly correct with respect to direct competition, but there is an "imaginary" competition that can be very strong in relation to institutional designs adopted previously24 (Elster et al. 1998, 60) and with regard to institutions adopted in other countries.

One of the advantages of analysing critical junctures such as the most recent National Constituent Assembly in Brazil is that it allows us to consider around which ideas change emerges. Other analytic perspectives place excessive emphasis on the interests at play and have little to say about ideas on institutional change in circulation. Of course, for certain kinds of institutional change, the "ideas in circulation" can be no more than one or two "focal points" shared by an epistemic community (e.g., an independent Central Bank). Other kinds of institutional change may theoretically involve many different ideas and, worse still, ideas shared by divergent and/or contradictory interests (e.g., the fact that both "progressives" and "conservatives" defended the creation of the Joint Budget Committee during the Constituent Assembly). In this case, it is necessary to make explicit why each interest, for different motives, espouses a certain "institutional idea". In this sense, the analysis of ideas in circulation as proposed both by Kingdon (1995) and Katznelson (2003) can be extremely fruitful in understanding institutional change.

Another challenge before institutionalist scholars is the incorporation of non-decisions into the analysis, in order to avoid selection bias (Capoccia and Kelemen 2007). While analysts look only at cases of change, much would certainly be revealed by also analysing changes that were aired but did not occur due to the configuration of the political struggle internal to the institution or to other factors. The initial comment on this matter was made by Bachrach and Baratz (1963, 632):

Many investigators have also mistakenly assumed that power and its correlatives are activated and can be observed only in decision-making situations. They have overlooked the equally, if not more important area of what might be called 'nondecision-making', i.e., the practice of limiting the scope of actual decision-making to 'safe' issues by manipulating the dominant community values, myths, and political institutions and procedures.25

Although this definition has more to do with the manipulation of the decision-making agenda, it is extremely interesting for one to think about institutional change and also public policy change.

Some studies of public policies, such as Diniz (2005), about labour legislation since the Sarney presidency, and Mancuso (2007), about the legislative agenda of the Brazilian business segment and the extent to which it was successful during the 1990s, take non-decisions into account. The analysis by Diniz (2005) is undertaken more with the perspective of evaluating the Executive's agenda power and how sincerely it reveals its political preferences by means of formal bill proposals. She finds that not every defeat of a bill proposed by the Executive should be seen as a failure, since the president sometimes proposes something just to stanch complaints and to please certain interest groups. It thus becomes clear how certain non-decisions reveal the manipulation of the political agenda by the Executive and contribute to explaining power conflicts that would be ignored in case the analyst took into account only bills explicitly approved or rejected in committees or the plenary. For his part, Mancuso (2007, 149-189) shows that over the course of the 1990s, Brazilian industrialists benefited from 56 (60.9%) out of the 92 non-decisions regarding bills it had an interest in. The author comments that there are "successes that consist in the rejection, withdrawal or filing of bills that industry repudiates [and] mean relief for the segment, for they prevent the cost of doing business in Brazil from going up (...)" (Mancuso, 2007, 155). It is another example of how the analysis of non-decisions can elucidate the dispute over a certain public policy, for if the scholar in this case had limited himself to the decisions actually made, the success of the business lobbt would have been underestimated.

It would be fruitful if in their studies on institutional change, analysts started also considering attempts at change that were unsuccessful. The argument by Taylor (2009) about the role of gradual policymaking as the engine of the autonomy of the Brazilian Central Bank would probably be even stronger if the author considered attempts at formal change in the bank's regime — in other words, "non-decisions". Not only would the epistemic community of economists and members of the government in favour of autonomy become clearer. This would also be the case for the political, economic and bureaucratic forces opposed to this idea and to the inclusion of this proposal for institutional change on the governmental agenda. In other words, the reasons why the autonomy of the Brazilian Central Bank has been won slowly, endogenously and, in a way, in a manner less formalized than it might, would be clearer.

Although this essay has displayed a certain predilection for the perspective of historical institutionalism, the superiority of this line for the study of institutional change remains to be sought on a case-by-case basis. It would not be surprising if studies that sought to make the most of the advantages of rational choice and of some of the strategies of historical institutionalism — as advocated by the "analytic narratives" approach proposed by Bates et al. (1998; see also Stryker 1996 and Pedriana 2005), for instance — were more successful than monolithic attempts. Perhaps, in the end, the possible synthesis of these two institutionalisms is the theoretical response best equipped to answer empirical regularities that puzzle us political scientists



1 Recent exceptions are Arretche (2007) and Peres (2008).

2 I adopt here the differentiation between institutionalisms proposed by Hall and Taylor (2003), among many other authors.

3 See also Carey (2000).

4 Clearly not all political actors need to agree for an institutional change to take place. Kreppel (2003) analyses changes in the standing orders of the European Parliament and shows that several significant alterations have been passed with the support of just two large parliamentary groups that, naturally, join coalitions in that parliament. In other words, the support level in question is far from unanimity.

5 The original argument is Elster's (1979).

6 But perhaps this perspective exaggerates how much current institutions please actors.

7 In his study on internal changes in the US Congress from 1890 to 1989, Schickler (2001) is quite emphatic on this point. In the case of the Brazilian budgetary process, Praça (2008) shows that the National Constituent Assembly witnessed alliances of conservative and progressive interests around the design of several institutions, notably the maintenance of the Joint Budget Committee created in 1969.

8 The rationale is instigating, but this is not a good answer to explain institutional change because many institutions will never stand out on the public agenda, and this does not necessarily mean that they are stable.

9 One way of minimizing this problem is to study, within cases of change, moments when there were non-decisions — in other words, change was raised but did not occur. I tackle this point in the last section of the article.

10 Again, paying attention to non-decisions is fundamental, for battles over changes that were not implemented reveal a great deal about ongoing mechanisms of institutional reproduction.

11 See Pierson (2000) for more criticism of the mistaken use of path dependence and Sartori (1970) for the original argument about the conceptual stretching with which this concept currently suffers.

12 The criteria used to define whether a certain event is contingent will be examined below.

13 Mahoney (2000, 532) states that contingent events can sometimes be explained by the emergence of a political entrepreneur.

14 I will not consider sociological institutionalism in this essay because although analysts of this line share with rational choice colleagues a rather broad definition of institutions, the main variable they consider to explain the genesis and persistence of a certain institution is its legitimacy (Hall and Taylor 2003, 211-212). Though certainly a relevant variable in understanding institutional dynamics, legitimacy is only one out of four institutional reproduction mechanisms pointed out by Mahoney (2000). Therefore, I will focus the essay on the other two institutionalist approaches Hall and Taylor draw attention to.

15 Weingast (2002) states that the rational choice perspective confers extremely little importance to the explanations of institutional change; see Greif and Laitin (2004) for an attempt, based on rational choice, to integrate some aspects of historical institutionalism with the aim of understanding the process of institutional reforms over time. The vast majority of institutional change analyses start off from the historical institutionalist side.

16 One of the mistakes most commonly made in analysing the behaviour of parties and parliamentarians in Brazil is the adoption of the re-election axiom without considering other preferences and institutional incentives available to the members of our Congress. Articles like Samuels (2003) are useful to clear up this problem.

17 According to this author, paying "systematic attention to preferences inside the elephantine moments of change identified by comparative historical social science can advance understanding of how the multiple possibilities inside unsettled moments of uncommon choice were resolved. The very character of critical junctures as relatively open times produced by concatenations of structural processes invite elucidations of the preferences and choices of the actors — grand to ordinary — placed inside such situations when the potentiality of alternatives explodes as previous constraints on belief and action erode" (Katznelson 2003, 277).

18 This point was extensively dealt with in the previous section.

19 This is a challenge yet to be considered by historical institutionalists.

20 An example of this process may be extracted from the analysis by Taylor (2009) on the prerogatives of the Brazilian Central Bank: the control over monetary policy always was in the remit of the Bank (or of the Office of the Superintendent of Money and Credit, SUMOC), but this prerogative in practice was irrelevant during time A (1945-1994), and has become extremely relevant during time B (1995- ).

21 Taylor (2009) cites five public policies that indirectly affected certain prerogatives and arrangements regarding the Central Bank, thus contributing, in the case of three of them (Collor Plan, renegotiation of the foreign debt and implementation of the fixed exchange rate), to the conversion of the organization.

22 Pierson (2004) considers three processes of institutional change: layering, functional conversion and diffusion. All of them echo the Streeck-Thelen proposal.

23 Although Heller (1997) argues that decentralized budgetary processes result in fiscal deficits, the Executive usually has so much autonomy to solve macroeconomic questions that this association becomes less plausible.

24 There were plenty of allusions to this type of "competition" between institutions during the debates of 1987 and 1988 on Brazil's new constitution. Economist and former minister Mario Henrique Simonsen, when called upon to attend a joint meeting of the Finances Committee and of the Budget and Financial Oversight Subcommittee, held on 30 April, 1987, had the following to say about parliamentarians' prerogative to amend the budget: "With regard to Congress' capacity to transfer funds and create new expenditures during its assessment of the budget bill, my impression is that a balance should be struck between the 1946 and the 1967 Federal Constitution. Doubtless, the 1967 Constitution tied the Legislative's hands too much. On the other hand, the 1946 Constitution permitted something on the whole undesirable, i.e., creating expenditure without indicating the source of funding. In my view, if one were to establish something similar to what exists in the German Constitution, according to which any member of Congress can propose expenditure as long as the corresponding revenue source is indicated, that would already be an improvement in relation to the 1967 Constitution and a brake on an excessive public-sector deficit."

25 Moe (2005) proposes a similar and updated version of this argument.


Bibliographical References

Abranches, Sérgio. 1988. Presidencialismo de coalizão: o dilema institucional brasileiro. Dados 31 (1): 5-34.

Ames, Barry. 2001. The deadlock of democracy in Brazil. Ann Arbor: Michigan University Press.

Arretche, Marta. 2007. A agenda institucional. Revista Brasileira de Ciências Sociais 22 (64): 147-151.

Bachrach, Peter, and Morton Baratz. 1963. Decisions and nondecisions: An analytical framework. American Political Science Review 57 (3): 632-642.

Bates, Robert H., Avner Greif, Margaret Levi, Jean-Laurent Rosenthal, and Barry R. Weingast. 1998. Analytic narratives. Princeton: Princeton University Press.

Boas, Taylor. 2007. Conceptualizing continuity and change: The composite-standard model of path dependence. Journal of Theoretical Politics 19 (1): 33-54.

Capoccia, Giovanni, and Daniel R. Kelemen. 2007. The study of critical junctures: Theory, narrative, and counterfactuals in historical institutionalism. World Politics 59:341-369.

Carey, John M. 2000. Parchment, equilibria, and institutions. Comparative Political Studies 33 (6/7): 735-761.

Clemens, Elizabeth, and James Cook. 1999. Politics and institutionalism: Explaining durability and change. Annual Review of Sociology 25:441-466.

Couto, Cláudio Gonçalves. 1998. A longa constituinte: Reforma do Estado e fluidez institucional no Brasil. Dados 41 (1): 51-86.

Diniz, Simone. 2005. Interações entre os poderes executivo e legislativo no processo decisório: Avaliando sucesso e fracasso presidencial. Dados 48 ( 2): 333-369.

Drake, Paul. 1989. The money doctor in the Andes: The kemmerer missions, 1923-1933. Durham: Duke University Press.

Duverger, Maurice. 1954. Political parties. New York: Wiley.

Elster, Jon. 1979. Ulysses and the sirens. Cambridge: Cambridge University Press.

Elster, Jon. Claus Offe, and Ulrick K. Preuss (with Frank Boenker, Ulrike Goetting, and Friedbert W. Rueb). 1998. Institutional design in post-communist societies: Rebuilding the ship at sea. Cambridge: Cambridge University Press.

Fernandes, Antônio Sérgio Araújo Fernandes. 2002. Path dependency e os estudos históricos comparados. BIB - Revista Brasileira de Informação Bibliográfica em Ciências Sociais, no. 53:79-102.

Figueiredo, Argelina, and Fernando Limongi. 1999. Executivo e Legislativo na Nova Ordem Constitucional. Rio de Janeiro: Ed. FGV.

Geddes, Barbara. 2003. Paradigms and sand castles: Theory building and research design in comparative politics. Ann Arbor: University of Michigan Press.

Greif, Avner, and David Laitin. 2004. A theory of endogenous institutional change. American Political Science Review 98 (4): 633-652.

Hacker, Jacob. 2004. Privatizing risk without privatizing the Welfare State: The hidden politics of social policy retrenchment in the United States. American Political Science Review 98 (2): 243-260.

Hall, Peter. 1986. Governing the economy: The politics of state intervention in Britain and France. New York: Oxford University Press.

____. 1993. Policy paradigms, social learning, and the State: The case of economic policymaking in Britain. Comparative Politics 25 (3): 275-296.

Hall, Peter A., and Rosemary C. R. Taylor. 2003. As três versões do neo- institucionalismo. Lua Nova, no. 58:193-223.

Heller, William. 1997. Bicameralism and budget deficits: The effect of parliamentary structure on government spending. Legislative Studies Quarterly 22.

Immergut, Ellen. 1998. The theoretical core of the new institutionalism. Politics & Society 26 (1):5-34.

Katz, Richard S. 2005 "Why are there so many (or so few) electoral reforms?". In The politics of electoral systems, ed. Michael Gallagher and Paul Mitchell, 57-76. Oxford: Oxford University Press.

Katznelson, Ira. 2003. "Periodization and preferences: Reflections on purposive action in comparative historical social science". In Comparative Historical Analysis in the Social Sciences, ed. James Mahoney and Dietrich Rueschemeyer, 270-301. Cambridge: Cambridge University Press.

Katznelson, Ira, and Barry R. Weingast. 2005. "Intersections between historical and rational choice institutionalism". In Preferences and situations: Points of intersection between historical and rational choice institutionalism, ed. Ira Katznelson and Barry Weingast, 1-24. New York: Russell Sage Foundation.

Kingdon, John W. 1995. Agendas, alternatives and public policies. New York: Addison-Wesley.

Knight, Jack. 1992. Institutions and social conflict. Cambridge: Cambridge University Press.

Krasner, Stephen D. 1984. Approaches to the state: Alternative conceptions and historical dynamics. Comparative Politics 16 (2): 223-246.

Kreppel, Amie. 2003. Necessary but not sufficient: Understanding the impact of treaty reform on the internal development of the European Parliament. Journal of European Public Policy 10 (6): 884-911.

Krieger, Gustavo, Fernando Rodrigues, and Elvis Bonassa. 1994. Os donos do Congresso. São Paulo: Ática.

Levi, Margaret. 1990. "A logic of institutional change". In The limits of rationality, ed. Karen Schweers Cook and Margaret Levi, 402-418. Chicago, University of Chicago Press.

____. 2009. "Reconsiderations of rational choice in comparative and historical analysis". In Comparative Politics: rationality, culture, and structure, ed. Mark I. Lichbach and Alan Zuckerman, 117-133. Cambridge: Cambridge University Press.

Lindner, Johannes. 2003. Institutional stability and change: Two sides of the same coin. Journal of European Public Policy 10 (6): 912-935.

Mahoney, James. 2000. Path dependence in historical sociology. Theory and Society 29 (4): 507-548.

Mahoney, James, and Celso M. Villegas. 2007. "Historical enquiry and comparative politics". In The Oxford Handbook of Comparative Politics, ed. Carles Boix and Susan Stokes, 73-89. Oxford: Oxford University Press.

Mahoney, James, and Kathleen Thelen. 2009. "A theory of gradual institutional change". In Explaining institutional change: Ambiguity, agency, and power, ed. James Mahoney and Kathleen Thelen. Cambridge: Cambridge University Press.

Mancuso, Wagner Pralon. 2007. O lobby da indústria no Congresso Nacional: Empresariado e política no Brasil contemporâneo. São Paulo: Edusp/Humanitas.

Mayhew, David. 1974. Congress: The electoral connection. New Haven: Yale University Press.

Moe, Terry. 2005. Power and political institutions. Perspectives on Politics 3 (2):215-233.

Müller, Wolfgang, and Kaare Strom. 1999. Policy, office, or votes?: How political parties in western europe make hard decisions. Cambridge: Cambridge University Press.

North, Douglass C. 1990. Institutions, institutional change and economic performance. New York: Cambridge University Press.

Palermo, Vicente. 2000. Como se governa o Brasil? O debate sobre instituições políticas e gestão de governo. Dados 43 (3): 521-527.

Pedriana, Nicholas. 2005. Rational choice, structural context, and increasing returns: A strategy for analytic narrative in Historical Sociology. Sociological Methods & Research 33 (3): 349-382.

Peres, Paulo Sérgio. 2008. Comportamento ou instituições? A evolução histórica do neo-institucionalismo da ciência política. Revista Brasileira de Ciências Sociais 23 (68): 53-71.

Peters, Guy. 1999. Institutional theory in Political Science: The "new institutionalism". London: Continuum.

Peters, Guy, Jon Pierre, and Desmond S. King. 2005. The politics of path dependency: Political conflict in historical institutionalism. Journal of Politics 67 (4): 1275-1300.

Pierson, Paul. 2000. Increasing returns, path dependence and the study of politics. American Political Science Review 94 (2): 251-267.

____. 2004. Politics in time: History, institutions and social analysis. Princeton: Princeton University Press.

____. 2006. "Public Policies as Institutions". In Rethinking political institutions: The art of the State, ed. Ian Shapiro, Stephen Skowronek and Daniel Gavin. New York: New York University Press.

Praça, Sérgio. 2008. Corrupção e processo orçamentário no Brasil: 1986-1993. Text presented at the Sixth Meeting of the Brazilian Political Science Association, State University of Campinas (Unicamp), August 30th, in Campinas, São Paulo, Brazil.

Samuels, David. 2003. Ambition, federalism, and legislative politics in Brazil. Cambridge: Cambridge University Press.

Santos, Fabiano. 2001. O poder legislativo no presidencialismo de coalizão. Belo Horizonte: Ed. UFMG.

Sartori, Giovanni. 1970. Concept misformation in comparative politics. American Political Science Review 64:1033-1053.

Schickler, Eric. 2001. Disjointed pluralism: Institutional innovation and the development of the U.S. Congress. Princeton: Princeton University Press.

Shepsle, Kenneth. 2006. "Old questions and new answers about institutions": The Riker objection revisited. In The Oxford Handbook of Political Economy, ed. Barry Weingast and Donald Wittman. Oxford, Oxford University Press.

Streeck, Wolfgang, and Kathleen Thelen, ed. 2005. Beyond continuity: Institutional change in Advanced Political Economies. Oxford: Oxford University Press.

Stryker, Robin. 1996. Beyond History vs. theory: Strategic narrative and sociological explanation. Sociological Methods & Research 24:304-352.

Taylor, Matthew M. 2009. Institutional development through policymaking: A case study of the Brazilian Central Bank. World Politics 61 (3):487-515.

Thelen, Kathleen. 1999. Historical institutionalism in comparative politics. Annual Review of Political Science 2: 369-404.

____. 2003. "How institutions evolve": Insights from comparative-historical analysis. In Comparative historical analysis in the Social Sciences, ed. James Mahoney and Dietrich Rueschemeyer. New York: Cambridge University Press.

____. 2006. "Institutions and social change": The evolution of vocational training in Germany. In Rethinking political institutions: The art of the state, ed. Ian Shapiro, Stephen Skowronek and Daniel Gavin. New York: New York University Press.

Tsai, Kellee S. 2006. Adaptive informal institutions and endogenous institutional change in China. World Politics 59:116-141.

Weingast, Barry. 2002. "Rational-choice institutionalism". In Political Science: The state of the discipline, ed. Ira Katznelson and H. Milner. New York: Norton.



Submitted in December, 2008
Accepted in June, 2009



* I am deeply grateful to Matthew Taylor and Lincoln Noronha for their comments on an earlier version of the essay, as well as to Julio Faundez for his bibliographical recommendations and to the reviewers for their suggestions. Financial support provided by Capes and CNPq were fundamental.