Explanations of Institutional Change: Reflecting on a “Missing Diagonal”

Previous Focus in Explaining Institutional Change

The most widespread approach in explaining institutional change is the idea of a “punctuated equilibrium” (Eldredge and Gould 1972)—that is, a dual model of path dependency and critical junctures. Longer phases of institutional stability are interrupted by short moments of institutional flux in which sudden change is made possible. These critical junctures are routinely defined as brief and consequential moments in time. Major change happens in condensed periods of time. The choice at the critical juncture canalizes future choices and leads to self-reinforcing reproduction mechanisms (Pierson 2000a). Among others, prominent macrohistorical comparisons (Collier and Collier 1991; Mahoney 2001b; Moore [1966] 1993), policy analyses (Baumgartner and Jones 1993), and more recently studies in international relations and global governance (Fioretos 2017; Zürn 2018) have set the explanatory agenda. However, how these brief and consequential moments in which change is made possible emerge in the first place remains unclear. This lacuna is the weak spot of the neo-institutionalist theorem of critical juncture and path dependency (Dunning 2017). Change comes in the form of a sudden and exogenous shock, but it remains by and large an unexplained explainer.

Against this backdrop, a powerful line of critique has developed. Instead of thinking about exogenous shocks as the explainer for institutional change, it argues to focus more on endogenous and gradual forms of change (Greif and Laitin 2004; Mahoney and Thelen 2010; Streeck and Thelen 2005; Weyland 2008). Institutions have been thought to exhibit so much internal inertia and stickiness that, once created, institutional reproduction and stabilization was the more natural research focus, whereas theoretical leverage for thinking about change has been relatively small (Immergut 2006). However, Knight (1992) has already reminded us that we should see institutions as power-laden vehicles and that there is constant struggle over the proper meaning of an institution.

In this light, our attention in institutionalist explanations should be focused on deficiencies in both rule making and rule taking. Institutions “often remain ambiguous and always are subject to interpretation, debate, and contestation” (Mahoney and Thelen 2010, 11). Enactment and enforcement problems, regulatory loopholes, control gaps, and circumventions as well as semantic struggles, imprecisions, implicit assumptions, and cognitive limits of rule-takers are some of the endogenous sources that endow actors pushing for change. These types of endogenous change are seen as incremental. Streeck and Thelen (2005, 31) helpfully characterize “five types of gradual transformation,” demonstrating its internal variety: exhaustion, drift, layering, displacement, and conversion. In all of these five processes, endogenous causes work over longer periods to alter existing institutional substance.

Two innovations are particularly important. First, previous approaches in neo-institutionalism tended to overlook that rule compliance should not be taken for granted (see also Levitsky and Murillo 2009). Instead, compliance with the institution is a key variable and not a fixed parameter. This insight opens up space for thinking more systematically about change that is caused within the institution itself. Second, different types of change agents are identified that can push gradually within the institutional boundaries for alternatives. Institutions define only what is “legitimately expected behavior” (Knight 1992, 15). That means that institutions shape actions and “exert patterned higher-order effects” (Clemens and Cook 1999, 444). Yet, they are not to be misunderstood as straightjackets. They leave interpretational gaps and room for maneuver, which change agents can gradually make use of. These change agents constitute the movers that are needed for explaining incremental change. Whereas in the previous dualism of critical juncture and path dependence the exogenous shock is the cause that carries the explanatory weight, here it is the endogenous enabling and endowing of actors that use their operative leeway within an institution. Institutions become plastic (Hall 2016). The agents are able to gradually replace and modify the institutional arrangement in which they operate. Table 1 shows the two dominant poles in explaining institutional change, disentangling already the source of the cause from its time horizon.

Table 1. Dominant Approaches in Explaining Institutional Change

Note: Source: Own table.

The “Missing Diagonal”: Moving Towards a Full Typology of Institutional Change

Reviewing these theoretical perspectives, I argue that the dominant arguments about institutional change conflate two dimensions. In current accounts, the source and the time horizon of the cause for change seem to be too tightly coupled. They do not need to correlate as much as the state of art suggests. On one hand, punctuated equilibrium relies on exogenous shocks that operate rapidly. On the other hand, the recent critique closely links endogenously caused change with gradual long-term processes. I propose to explicitly disentangle them from each other. I organize the fuzzy field of institutional change along two axes: (1) the sources of the cause and (2) the time horizon of how long the cause operates. By opening up a two-dimensional space, the typology highlights a “missing diagonal” in our explanations of institutional change.

The first axis is the source for institutional change. The driver for change can be exogenous or endogenous. Please note that an exogenous cause is not only external to an institution but also generated outside this very institution, thus working from the outside in. In contrast, a cause can also emerge within the institution itself, being so a product of inherent institutional properties and working from the inside out. An illustrative way of understanding the difference between exogenous and endogenous is to delineate system from its environment as natural scientists routinely do. For example, in a chemical reaction the system is the mixture of elements that react with each other, separated by the glass wall of a test tube. Everything happening outside the test tube is environment.

The work of social scientists is usually less straightforward than this natural science analogy suggests. The complex interplay and interconnectedness of political and social phenomena complicates the distinction between not only internal and external (location) but also endogenous and exogenous factors (location plus generation). The glass wall of the test tube needs often to be explicitly established, putting the researcher’s perspective on center stage. Compared with natural scientists, social scientists engage more actively in defining what lies inside or outside the test tube and in determining whether a cause is actually generated within or outside the test tube. Additionally, as Greif and Laitin (2004) argue, what counts as an exogenous and an endogenous factor can also vary over time. A factor can be exogenous to an institution in the short run, but it can be endogenized in the institutionalist explanation in the long run.

Yet, the guiding idea from the test tube analogy should be transferred to the social sciences. It is rightly noted that institutions are an “ill-defined concept” (Scharpf 1997, 38) and that there is some Babylonian confusion about what an institution is and what it entails. Thinking of the social science equivalent to the glass wall of the test tube can be understood as a plea for greater analytical precision in drawing the borders of an institution, clarifying what is generated inside and outside of it. In making the amorphous concept of institutions more manageable, the work of Crawford and Ostrom (1995) on a common “grammar of institutions” provides a helpful guide here. Answers to the following five key questions force us to make explicit the often implicit understanding of the institution under study: To whom does the institution actually apply (A)? What is the deontic operator (D)? What is the aim (I)? What are the conditions (C)? And, what sanctions are to be imposed for deviant behavior (“or else,” O)? This ADICO scheme helps scholars to be more precise about what the institution actually entails (and to decide what is endogenous and exogenous to it).

We should also be aware that a social science institution is never as sterile as a test tube. Therefore, an endogenous institutional explanation refers not only to the chemical elements—the explanatory factors—that a social scientist deliberately selects and puts into the test tube but also to the bulges, dents, and cracks of the tube’s wall as well as potential internal impurities, pollutions, and contaminations. It is these institutional imperfections that a social scientist needs to consider in an endogenous explanation. Inherent institutional properties empower or even produce actors so that institutions contain within themselves the seeds for change. A guiding rule of thumb for concrete research projects is to take the literal meaning of endogeneity seriously. An endogenous explanation means that institutional properties generate within this very institution change agents and structural dynamics that cause the institution to change.

Besides the source of the cause, the second axis of the typology refers to the time horizon of the cause. The cause can either work suddenly and create brief moments of rupture or operate incrementally over long stretches of time (Abbott 2001; Pierson 2000b; 2004). The general idea is to take the temporal structure of variables and explanations more seriously (Büthe 2002; Griffin 1992; Grzymala-Busse 2011; Sewell 1996). Here, I account for one crucial temporal distinction: duration as the temporal length in which a cause works. ¹ A cause can hit the target institution suddenly like a karate chop. A financial crisis might be an adequate social science equivalent here, often creating turbulence in which institutional actors have insufficient time to adapt to changing circumstances, instead needing to react hastily. Or, the causal effect of one variable on another may be stretched out to gradually unfold over extended periods, working like the “termites in the basement” (Mahoney and Thelen 2010, 31; Streeck and Thelen 2005). In social sciences, adequate equivalents are “‘sociological’ variables like demography, literacy, or technology” (Pierson 2004, 14). In contrast to the sudden fluidity of a crisis, these causes operate with a longer time horizon.

It is difficult to give generalizable operational advice regarding how to distinguish between long- and short-term causes. Social scientists are trained to temporally zoom in and out, readjusting their aperture depending on the specific research project. Years or even decades of protest may be defined as short-term causes in order to explain, for example, the changing political culture in a postcolonial country. Meanwhile, years would be a rather long time horizon for explaining foreign policy changes in the Cuban Missile Crisis. The distinction between sudden and gradually operating causes therefore depends on the time horizon of the institution under study as well as the (spatio)temporal scope conditions of the particular research project.

The proposed typology draws major inspiration from the work of Pierson (2003). In his typology, Pierson also emphasizes the temporal structure of variables. Yet, he disentangles the time horizons of causes and outcomes from each other. In contrast to him, I am only interested in the explanatory side and focus explicitly on the causes, not on the outcomes. Yet, combining these dimensions (source of cause, time horizon of cause, and time horizon of outcome) would result in a three-dimensional typology in which the eight corners of a cube represent the different types of change. Beyond the time horizon of the outcome, it might also be worthwhile to connect the typology with the scope of the outcome. I share the criticism of Grzymala-Busse (2011, 1286) that there is no need to assume an inherent connection between sudden causes leading automatically to large-scale change, whereas gradually working causes are assumed to produce “only” small changes. Similarly, the proposed typology can also be complemented with insightful approaches that account for the rigidity of the institutional environment to explain what type of change agent is likely to emerge (Mahoney and Thelen 2010) and what forms of change are expected to occur (Riedl and Roberts 2017; Streeck and Thelen 2005). Also, the types of explanation might be productively connected to a more nuanced discussion of concrete causal mechanisms, decomposing the involved entities and their activities (Machamer, Darden, and Craver 2000), specifying potential microfoundations (Hedström and Swedberg 1996; Tilly 2001), and accounting for further important aspects of temporality (Grzymala-Busse 2011). Here, I limit the typological axes to two major constituents of a cause: its origin and its duration.

By cross-tabulating these two major causal constituents, we arrive at a full typology of institutional change. While the upper left and the lower right cells are dominant in current accounts of institutional change, less attention is placed on the missing diagonal of type II and type III change. Table 2 provides an overview, already indicating the illustrative natural science metaphors and the social science examples that I discuss in the following section.

Table 2. A Typology of Institutional Change

Note: Source: Own table.

Exogenously Driven Ruptures (Type I Change)

Exogenously driven rupture—that is, the type I change in the upper left quadrant—follows the dual model of long phases of stability that are suddenly interrupted by surprising events in which unmodeled forces intervene abruptly, making continuation of the existing rules of the game unlikely. However, the idea of “punctuated equilibrium” (Eldredge and Gould 1972) that is borrowed from paleobiology is subject to some misunderstandings. ² In the social sciences, two distinct understandings of a punctuated equilibrium coexist in parallel, causing considerable confusion: phyletic gradualism that is best illustrated with a branching tree metaphor on one hand and the original Gould and Eldredge idea of punctuated equilibrium that follows a step function on the other hand. Clarifying this confusion helps us to better understand the different explanatory routes. Figure 1 situates the reader by showing graphically where the major differences lie.

Note: Source: Own figure.

Figure 1. Phyletic Gradualism versus Punctuated Equilibrium

The two understandings differ in one important manner. The exogenous shock in the branching tree metaphor makes paths diverge and represents a Weberian railroad switch, whereas the step function can be understood as putting an institution on a new track. An important example for the former is the classic study by Collier and Collier (1991, 27). By using Robert Frost’s poem “The Road Not Taken” as a key metaphor, they invoke the image of a “‘branching point’ from which historical development moves onto a new path” (Hall and Taylor 1996, 942). It represents a forking path. Other scholars have refined this template (Capoccia and Kelemen 2007; Collier and Munck 2017; Mahoney 2001b; Pierson 2000a; Rixen and Viola 2015), but they agree that once a critical choice is made, it forecloses other viable options and can hardly be taken back, which reflects the idea of a branching tree.

One might speculate that Stephan Krasner’s (1984, 240–44) influential article led to the conceptual confusion in the discipline about what punctuated equilibrium is. He juxtaposed Gould and Eldredge’s (1972; 1977) idea of “punctuated equilibrium” and the branching tree metaphor, implying that the one can be equated with the other. In the aftermath, branching tree metaphors and punctuated equilibria are almost always used interchangeably in the social sciences, particularly in historical neo-institutionalism. Yet, Gould and Eldredge argue exactly against such a branching tree metaphor. Branching trees, they argue, represent Darwinian evolutionary thinking. In contrast, they propose a step function as the appropriate illustration for punctuated equilibria.

Instead of gradual transformations, punctuated equilibrium argues that “rapid events of speciation in isolated subpopulations” (Eldredge and Gould 1972, 110) lead to sudden and brief interruptions of long phases of morphological stability. A homeostatic equilibrium is abruptly disturbed. In contrast to the branching tree metaphor, in which “splitting, when it occurs, produces a slow and very gradual divergence of forms” (Eldredge and Gould 1972, 109), punctuated equilibrium is characterized by two mechanisms: a rapid isolating mechanism at the beginning that is then followed by local adaptation.

What has largely been overlooked in the social sciences is this isolating mechanism. Gould and Eldredge emphasize that a rapid geographical dislocation (e.g., the break-up of an island) isolates a subpopulation that needs to adapt to new circumstances. It is allopatric—that is, the descendants stem from other ancestors. Yet, and this is important to note, it is not a saltational theory. It does not argue with jumps in evolution but instead with dislocation that makes the identification of the concrete isolating mechanism so important (Gould and Eldredge 1993).

Arguments that follow the branching tree metaphor are more common in the social sciences than the step function. A good illustration is Roberts’s (2014) study on the effects of the economic crisis on the stability of party systems in Latin America. Advancing a classic critical juncture framework (Capoccia 2015; Collier and Collier 1991, 27–39; Mahoney 2001a; 2001b; Lipset and Rokkan 1967), Roberts (2014, 10) argues that an “exogenous shock—the 1982 debt crisis—bankrupted developmental states and forced economic adjustment.” In an “activating critical juncture” (Riedl and Roberts 2017) antecedent conditions like the depth of the previous state-led development policy and the nature of the former party system (elitist vs. labor-mobilizing) predisposed differential disruptive effects on post-critical juncture party systems. In this “neoliberal critical juncture” (Roberts 2014, 89–93), the change of economic models after the debt crisis was the “major mover of party system change” (Mainwaring 2016, 707). I interpret the exogenous shock of the debt crisis as an important choice moment, resulting in a forking path. In my reading, its argumentative structure follows the branching tree model.

Arguments that employ the Gould/Eldredge idea of punctuated equilibrium—that is, they incorporate not a railroad switch but setting the institution on a new railroad track—are rarer. As depicted above, the exogenous shock must be so strong that it puts an institutional order in a new environment, creating an aliud. War is an appropriate candidate for such a far-reaching and comprehensive shock. A landmark study in this regard is, for example, Ikenberry’s (2001) masterful analysis, After Victory. Ikenberry explains the set-up of institutional orders after great power wars such as the Napoleonic Wars 1814/15, the First and Second World Wars 1918 and 1945, and the end of the Cold War 1989/1990. These chaotic major postwar junctures are perceived by Ikenberry (2001, 3) as such extraordinary situations that state leaders found themselves in “unusually advantageous positions” to “remake international order.” He argues that those who won the war neither exploited their advantageous position nor abandoned the other states. Instead, they used their dominant position during these watershed moments to develop mechanisms of strategic restraint in order to build and lock in a new political order that bound together other states in a set of favorable and institutionalized postwar power relations (Ikenberry 2001, 50–79). I read Ikenberry’s study as being closer to the allopatric speciation model than to phyletic gradualism. The international orders were disrupted during these great power wars, leading Ikenberry (2001, 7) to employ Katzenstein’s notion of history as a “sequence of irregular big bangs.” A step function illustrates this change trajectory best. When drawing the trajectory on paper, one cannot draw its progression in one fluid motion; rather, one must lift the pencil to start anew at a different point.

The distinction between the two understandings of “punctuated equilibrium” has direct methodological repercussions. The starting point for Gould and Eldredge was actually a methodological concern. The standard Darwinian argument is that all gaps in the fossil record, sooner or later, need to be closed, and if we did not yet manage to do so, then we just need to dig deeper to bring the dots closer together. In other words, if gaps cannot be closed and graded sequences be completed, then we have simply not found enough empirical evidence to prove the theory right, effectively making the theory “virtually unfalsifiable” (Eldredge and Gould 1972, 90). What we can learn from Gould and Eldredge is their fundamental insight that our theoretical expectations sometimes color our empirical perceptions and that gaps in the sequences should not automatically be interpreted as imperfections in our data. Instead, we need to acknowledge that “many breaks in the fossil record are real” (Eldredge and Gould 1972, 96).

This means that, methodologically speaking, in some instances we should not overestimate continuity and gear our data collection effort towards a steady, linear, and continuous path; instead, we should take seriously the existence of breaks and discontinuities. Breaks are then turned into a legitimate and fruitful object of study. If we follow this route, then the burden of proof lies in demonstrating a social science equivalent to an isolating mechanism that explains why we observe a leap and why an institution has been put on a new track. The research perspective therefore changes, shifting focus from developmental pathways—typically entailing the analysis of critical antecedents, continuities before and after crises, and institutional legacies—to explanations of paradigmatic change—that is, institutional change so fundamental in nature that it breaks with preexisting traditions. The comprehensiveness and magnitude of wars are prime candidates for punctuated equilibria, but it can also be smaller events depending on the institution of study. Yet, the key question remains: Are we interested in continuous legacy arguments, or do we “allow” for discontinuity in our explanations?

The distinction between branching trees and step functions also speaks to the long-standing discussion among historical institutionalists regarding to what extent change is necessary for making arguments about critical junctures. For example, Collier and Munck (2017) see it as a necessary component, whereas Capoccia and Kelemen (2007, 352) state expressis verbis that “change is not a necessary element of a critical juncture.” If arguments follow the step function architecture, institutional change is almost inevitable. A sharp and sudden change is to be expected after the isolation sets in. However, if arguments follow the branching tree structure, this condition can be relaxed. In a critical juncture, the original path may be continued without changing direction (see also Dunning 2017).

To take stock, there are two different explanatory modi of type I change that are often both misleadingly subsumed under the umbrella of “punctuated equilibria.” While the original Gould/Eldredge model of allopatric speciation resembles a step function, the social science “adaptation” of it more frequently follows the branching tree model—ironically, exactly the one that Gould and Eldredge (1972) argued against in their classic study. Branching tree models mark the divergence of paths due to an exogenous shock, highlighting continuity. In contrast, the step function model sees exogenous shocks as causing isolation from each other and offering a new beginning, so emphasizing discontinuity.

Endogenous Gradual Change (Type IV Change)

While it is safe to say that the critical juncture literature is dominant in explaining rapid change, type IV change dominates explanations of gradual institutional change. It emerged as a direct reaction to the explanatory void of critical juncture arguments. Instead of emphasizing punctuated equilibria, it asks how critical junctures emerge in the first place and how an institution can be changed from within.

Radioactive decay is a fitting analogy from the natural sciences. Here, the driver for change lies within the chemical element itself. Compared with stable chemical elements, radioactive elements are characterized by an inherently unstable nucleus. While the number of protons per nucleus is fixed, the number of neutrons that stabilize the atomic core can vary. Neutrons offset the repulsion of positively charged protons. If there are too many or too few neutrons in relation to the optimal number of protons per nucleus, the atomic core becomes unstable and subject to radioactive decay. The period that it takes for half of the unstable atoms to undergo such radioactive decay is measured in the half-lives of the isotopes. The crucial idea that we can adopt here for the social sciences is that the inner “architecture” of the atomic cores is responsible for the radioactive decay. In other words, the potential for change is already implanted in the composition of protons and neutrons within the nuclei, serving, therefore, as an apt analogy for endogenously caused change. Moreover, as naturally existing radioactive elements have long half-lives, they are also well suited for illustrating long time horizons and gradual forms of change.

Translated to the social sciences, the inner architecture in which the electric repulsion of protons is offset by neutrons can be compared to binding mechanisms within institutions. Following the recent work of Capoccia (2016), those institutions in particular develop “bite” in which two mechanisms are at play. On one hand, Capoccia convincingly demonstrates that the institutionalization of cultural categories makes deviant behavior costlier. On the other hand, the allocation of power over the timing of reforms is critical because it makes it more difficult for contesting actors to build and shift coalitions. If institutional incumbents can bind change agents and offset their repulsive energies, they remain resilient.

Going beyond the crude dichotomy of winners and losers within an institution, Mahoney and Thelen (2010) suggest different types of change agents. In their stimulating work, they distinguish between subversives, insurrectionaries ³ , symbionts, and opportunists. Whereas subversives might disguise their true intentions and go undetected, gradually pushing for change, insurrectionaries openly mobilize dissent. Symbionts, in turn, rely on the preservation of the institution and might either parasitically contradict the spirit of the institution or mutualistically “derive benefit from rules they did not write” (Mahoney and Thelen 2010, 24). Finally, opportunists are characterized by ambiguous preferences, exploiting the institution for their own sake. If these change agents can be “neutralized,” an institution is able to defy change. Otherwise, they become the actors behind decay processes.

Illustrative social science examples for type iv change are abundant. As discussed above, in critiquing the dualism of stasis and compressed moments of change, a new research agenda has emerged in recent years. One of the cornerstones of this research strand is the comparative historical analysis of Kathleen Thelen (2004) on labor skill formation regimes in Germany, Britain, the United States, and Japan. Particularly in her longitudinal study of skill formation in Germany, Thelen demonstrates that the development of vocational training in Germany can be traced back to the 1870s and evolved incrementally via feedback loops in which both labor and capital recognized the beneficial role of handicraft chambers (Handwerkskammer) that gained “parapublic authority” (Thelen 2004, 40) to organize skill formation. In her later terms, the handicraft chamber turned out to be a mutualistic symbiont. The development of German skill formation was therefore not the result of “big bangs” or critical junctures (which might have been probable given Germany’s rich history of turning points); instead, it was the more “subtle and incremental changes occurring in relatively “settled” rather than “unsettled” times” (Thelen 2004, xiii) that account for its gradual evolution.

A further empirical example in which the logic of endogenous gradual change can be observed in admirable clarity is the work by Valerie Bunce (1999) on the dissolution of the Soviet Union. The Soviet state institutions that disintegrated gradually have been characterized by Bunce as “subversive.” Although—like in the reference to radioactive decay—exogenous forces have served as catalyst, the seeds of their own demise have been already implanted in the original institutional design. The causes were to be found inside rather than outside the institutions under study (Kalyvas 1999).

Exogenous Gradual Change (Type II Change)

Erosion is the most common metaphor employed to describe all types of gradual processes. Not only in the social sciences but also beyond, erosion is often equated with any sort of gradual weakening. Yet, having a closer look at its geological roots, this broad understanding is not only too vague but also misleading. In geology, erosion is a physical process by which solid particles are moved, transported, and later sedimented. In general, two sets of factors need to be distinguished. On one hand, erosivity factors refer to the strength of wind and water—the exogenous factors. On the other hand, erodibility factors refer to the characteristics of the soil—that is, the vulnerability to winds or water (Julien 2010, 4–27). While the structure and condition of the soil influences to what extent the sediment can be set in motion so that substance is eroded, the actual driver for change is always exogenous: wind or water that “attacks” the soil. An intuitive example from the social sciences is the effect of economic sanctions on authoritarian regimes (Hufbauer et al. 2007). Like different types of soils, authoritarian regimes differ in their vulnerability to economic sanctions. Certain erodibility factors define the robustness or porousness of the targeted authoritarian regime and influence the speed and scope of change. Yet, the cause of this form of gradual change (international sanctions) is exogenous.

Despite its ubiquitous usage, clear-cut erosion processes are difficult to find in the classic social science canon. Although not being marked as such, an illustrating example is the classic book by Jack Goldstone (1991). He advances a complex argument on why states broke down in early modern Europe and Asia during the seventeenth and nineteenth centuries. I interpret Goldstone’s study as an erosion argument in which political institutions are gradually weakened over extended periods. This weakening is caused by a factor that is exogenous to the political institutions: demographic change. Goldstone’s work can be best understood if we break up his argument into erosivity and erodibility factors. Similar to the sanction argument made above, states differed in their vulnerability to demographic pressure. However, “revealing answers lie in the broad-based impact that sustained population growth (or decline) had on the economic, social, and political institutions in agrarian-bureaucratic states” (Goldstone 1991, xxiv). Demographic pressure is like the continuous wind that weakens institutional substance and gradually leads to a decline of nationwide rule-making and rule-enforcing capacity, first in terms of declining financial endowment and then subsequently in terms of waning military capability. Demographic pressure can be so understood as the exogenous mover that gradually unfolds and weakens state institutions.

Endogenously Driven Ruptures (Type III Change)

In contrast to the punctuated equilibrium models, endogenous ruptures mark type III changes. This type constitutes the second end of the “missing diagonal.” Type III changes are caused by shocks that do not stem from outside the explanatory test tube but have their cause inside this tube. Endogenous ruptures might be the most surprising and least intuitive type of change. Yet, I argue that we have neglected this way of explaining institutional change. The proposed typology is able to detect and point us to this highly interesting type that needs to be theorized in greater depth. Institutional change can occur due to shocks, but these shocks do not need to come from the outside; they can also be caused endogenously.

If we take a look at natural science and medicine, we get a better idea of what endogenous shocks can be. Think of a lung embolism in which a travelling blood clot that circulated in the body gets caught in a too-narrow artery. The thrombus itself may evolve slowly or quickly. But, what is more important, once a part of this thrombus separates from the vascular wall and becomes an embolus, it can turn rapidly into an embolism in which the blood flow is abruptly blocked and that might even be lethal. A second intuitive example is an autoimmune reaction. Although still being extensively researched, it is safe to say that it is an unexpected and rapid reaction in which the body does not recognize a cell as being benign any longer but instead sees it as harmful and erroneously fights against it. What is worth noting in these examples is that the process is driven endogenously—that is, within the system—and that these events can occur suddenly and without prior warnings in even previously undamaged blood vessels and healthy parts of the body.

Identifying social science examples for the lower-right quadrant is challenging. Endogenous ruptures often demand a powerful actor, a Schumpeterian norm entrepreneur who is produced, empowered, or suddenly motivated by (perverse) incentive structures within the institution and who is able to break sharply with long-held institutional inertia. Military coups can serve as adequate social science equivalents to endogenous ruptures. The pioneering work of Janowitz (1964) already highlighted endogenous explanations for military interventions into politics. In a similar vein, Nordlinger (1977) and Perlmutter (1977, 89–114) describe praetorian institutions in which the political sociology of the officer corps is emphasized for explaining military coups. The norm entrepreneurs that sharply break with institutional inertia by staging a military coup are bred within the military itself, constituting a kind of “young Turks” argument within the highest military ranks.

While these explanations point to the sudden empowerment of actors, the classic study on party systems by Sartori (1976) refers to an institution-inherent force that causes these sudden ruptures. Being informed by historical case studies of the Weimar Republic, the French Fourth Republic, and Chile until 1973, Sartori argues that a party system with five to six parties that is characterized by the existence of an anti-system party and a bilateral opposition (i.e., an opposition left and right from the center) “suffices to identify” what he calls a “polarized pluralism.” In such a polarized pluralism, centrifugal forces dominate party competition. Creating a situation of “center-fleeing hemorrhage” (Sartori 1976, 136), he is convinced that “the thing we know for sure” is that a centrifugal polity “is doomed: it can only, and quickly, end in deflagration” (Sartori 1976, 145). It is this quick deflagration due to centrifugal forces that represents the idea of an endogenously caused rupture best. The cause for change works like a fast-acting poison that is innate to the institution of the party system itself.

Eichengreen’s (1996) work on the end of the Bretton Woods international monetary system can be interpreted in a similar way. While not being a classic neo-institutionalist argument, his insights can be made fruitful to this field of research. In general, the stability of the Bretton Woods system depended on the conviction that the US dollar was “as good as gold.” Yet, what “made it dynamically unstable” (Eichengreen 1996, 116) was a construction flaw, an inbuilt breaking point. The so-called “Triffin dilemma” highlights an inherent tension between long-term exchange rate stability by guaranteeing that accumulating dollar reserves was attractive due to the unquestioned convertibility on one hand and the willingness of the US to offer additional dollars to meet global demand and guarantee liquidity on the other hand. Despite critical voices from its very inception, the Triffin dilemma lay dormant until the US foreign monetary liabilities exceeded its gold reserves in the 1960s. A significant dollar overhang resulted. When the US government was not willing to reduce the global liquidity problem by reversing its perennial balance of payment deficits, thus running risk of economic contraction and unemployment, France’s president de Gaulle was among the most vocal critics of the United States’ “exorbitant privilege” of running deficits, threatening to liquidate its reserves, and the German Bundesbank, though initially backing the dollar, also abruptly withdrew its support and floated the mark in spring 1971, fearing domestic inflation. In retrospect, it is fair to say that the endogenous source for the breakdown of the Bretton Woods system was an inherent fracture point that, once visible, incentivized actors to suddenly revise their cost-benefit calculations, not willing to pay a domestic prize for stabilizing an international monetary system. They no longer accepted the institutional rules of the game but instead broke sharply with them (see also Frieden 2006, 339–60).

To sum up, although they are still underresearched, endogenous ruptures can happen in at least three ways. First, from the study of military coups we can learn about the sudden empowerment of military officers, sometimes even acting out of the second row. Second, with Sartori we can better appreciate inherent deflagrating factors (centrifugal competition) within institutions. Third, from Eichengreen’s explanation of the end of Bretton Woods, we can better understand an abrupt change in institutional incentive structure. All three ways sharply deviate from the perspective of endogenous decay processes in that they do not assume weak actors under an institutional spell that gradually push for change but instead place emphasis on the abrupt creation, empowerment, or changing motivation of powerful actors and their disruptive behavior.

Disentangling Multilayered and Sequential Explanations of Institutional Change

In the methodological process tracing literature, heavy emphasis is put on disentangling the intermediate steps that ultimately lead to an outcome. The crux of process tracing lies in collecting (necessary and/or sufficient) empirical evidence to test a hypothesized causal mechanism that usually consists of several intermediate steps. The challenge of process tracing is to demonstrate why and how causal chain links operate with each other (Beach and Pedersen 2016; Bennett and Checkel 2015; Rohlfing 2013). Even if these intermediate steps are “tightly coupled” (Pierson 2004, 88), they need to be distinguished for analytical clarity. Applying Kant’s dictum that even if two things cannot be separated, they can still be distinguished, it is essential to identify the different parts of a process and to adequately account for the “fractal-like nature” (Grzymala-Busse 2011, 1282) of political phenomena.

Explaining institutional change is no exception. While some institutional explanations are straightforward and can be located in one of the cells of the proposed typology, others are fuzzier. They involve processes that coincide, overlap, and affect each other in multiple sublayers and may or may not include diachronic, causal chain, or sequence arguments (Abbott 1995). Yet, I argue that the proposed typology helps to identify the individual building blocks of how complex change arguments are actually constructed, constituting an important stepping-stone for more analytical clarity.

Take, for example, the classic study Social Origins of Dictatorship and Democracy by Barrington Moore ([1966] 1993). Moore’s work is almost unparalleled in its empirical depth and has inspired generations of scholars to pursue what has later been called comparative historical analysis. Yet, his informal theory building has also caused intensive debates and confusion regarding his core argument (Mahoney 2003; Skocpol 1973). I will demonstrate the fruitfulness of the proposed typology by carving out the most important building blocks of his complex argument and locating them in one of the four cells of the typology. By so doing, I attempt to narrow down his argument to its basic explanatory skeleton. I focus on his third path towards modern states, the communist route, and will use the Chinese case as an illustration to show that the proposed typology is able to act as a helpful compass, guiding scholars through Moore’s sometimes distracting and overwhelming historical detail. In other words, the typology introduced here effectively “disciplines” his theory building.

Two caveats are in place. First, within this article, I focus only on the reconstruction of Moore’s argument. The modern process-tracing literature would additionally demand that concrete empirically observable implications for Moore’s theory be spelled out. Yet, Moore often remains on a rather abstract macro level in his argumentation. A more fine-grained microfoundation of his argument needs to be delegated to future research. Second, a full account of Moore’s causal mechanism cannot be fulfilled here. Doing justice to the complex and nuanced study of causal mechanisms demands not only the identification of involved entities and their associated activities but also that the “productive continuity” (Machamer, Darden, and Craver 2000, 3) between the subcomponents of the mechanism (Beach and Pedersen 2016, 23–44; Rohlfing 2012, 23–60) be addressed. This is beyond the scope of this endeavor. Instead, I concentrate on the different elements of Moore’s argument. Figure 2 illustrates this argument graphically, highlighting the four types of explaining change.

Note: Source: Own figure.

Figure 2. Graphic Illustration of Moore’s Argument on China’s Imperial Rule

Moore’s goal is to explain the “varied political roles played by the landed upper elites and the peasantry in the transformation from agrarian societies … to modern industrial ones” (Moore [1966] 1993, xvii). In my reading of Moore’s work, the institution under analysis is the system of Imperial rule in China. This rule is described as a tenancy arrangement between landlords who aspired to join the bureaucratic ranks through educational attainment (for himself or his family) on one hand and the peasants that were controlled by this Imperial bureaucracy on the other hand.

Moore’s argument about the end of this ruling system is predominantly about endogenous decay. Yet, he also incorporates erosion arguments and adds endogenous and exogenous ruptures at a later stage. I identify four parallel decay arguments. First, the exploitative nature of the tenancy arrangement demanded cheap labor and (in Marxist parlance) a reserve army of unemployed peasants, leading to overpopulation. Second, the Imperial examination system bred an oversupply of bureaucrats that resulted in a large number of degree candidates. Third, the institution incentivized corruption. The Imperial bureaucracy relied on a system of local corruption that undermined the effectiveness of tax collection but was important in maintaining a system of privileges for the local gentry and landlord-scholars (Moore [1966] 1993, 170–73). Finally, Moore advances an understanding of an agricultural sector that lacked any motivation for rationalization due to an abundance of labor and the absence of an urban market (179–80).

These four endogenous decay processes are complemented with an exogenously driven erosion process. Moore argues that the West eroded Chinese Imperial substance by the constant influx of foreign merchants, particularly after the Opium Wars. These foreign merchants gradually weakened the traditional role of the scholar-official. A “new hybrid society” emerged “in which power and social position no longer rested securely in the hands of those with classic education” (Moore [1966] 1993, 176). The point of attack for this exogenous erosion was therefore the numerically small commercial, trading, and finance class that usually resided in the big cities in coastal areas. The Imperial bureaucracy was no longer able to prevent or, at the least, absorb and control bourgeoning commercial elements.

This amalgamation of old landed elites and a newly emerging trading and commercial sector was the social underpinning for a political order in which the “link between rulers and ruled was weak and largely artificial, liable to snap under any severe strain” (Moore [1966] 1993, 205). In this situation, Moore voices a very strong opinion that the peasantry had no capacity for self-organization and describes a “relatively atomistic peasant society” (211) that had too little experience in organizing and mobilizing protest beyond the clan. An exogenous shock, the Japanese occupation, changed the situation. This occupation “performed two essential revolutionary tasks for the Communists” (223)—namely, the elimination of the old elites and the forging of new solidarity among the oppressed peasants. This exogenous shock substantially contributed to the bonds between the rulers and ruled ultimately snapping.

Finally, Moore emphasizes in his work a strategic ideological shift after 1927 of the Communist Party. Until 1926, the Communists did not begin to “display any serious interest in using the peasants as the base for a revolutionary movement” (Moore [1966] 1993, 223). A year later, they suddenly switched from fruitless attempts to gain power via the Marxist orthodoxy of winning proletarians in the cities to the Maoist interpretation of relying on the peasantry. By concentrating on the peasantry, the Communist Party could leverage on an inbuilt weak point of traditional Imperial rule. The Chinese landlord–tenant relationship was a “device for squeezing an economic surplus out of the peasant” (179). The peasants were not only exploited by the landlords; the heavy focus on scholarly education to join the Imperial bureaucracy also led to growing alienation. Moore describes these landlords as completely detached, they “seemed to play no part whatever in the actual work of cultivation, not even a supervisory one” (167). This exploitation-cum-alienation relationship between peasants and landlords was a major Achilles heel for the traditional Imperial rule arrangement that Mao’s new ideological strategy could capitalize on. The sudden ideological conversion from a proletarian to a peasant revolution represents an endogenous punctuation to the old Chinese ruling system that significantly contributed to its downfall.

To sum up, Moore’s argument about China’s path to communism is one of four overlapping streams of endogenous decay, with an additional erosion process by the West that caused a hybridization of the traditional scholar-based hierarchical society. Moreover, the ideological shift of the Chinese Communist Party and the Japanese occupation were disruptive, encouraging the exploited peasantry to mobilize against the ancient regime.

Erosion versus Decay: The Democratic Backsliding Debate

On a global scale, we observe that populist voices are becoming stronger, that nationalism is experiencing an upsurge, that democratic qualities are deteriorating, and that some countries are even embarking on an autocratization path. Against this backdrop, a number of important research projects and publications have tackled the problem of setbacks, interestingly often under the banner of “democratic erosion.” Drawing attention to only some prominent publications in the last years, McFaul (2018) discusses, for example, the “erosion of Russian democracy,” Gamboa (2017) the “erosion of democracy in Colombia and Venezuela,” Bermeo (2016, 14) puts forward that “troubled democracies today are more likely to erode rather than to shatter,” and Diamond (2015, 147) observes the “significant erosion in electoral fairness, political pluralism, and civic space for opposition and dissent” in Turkey.

All of these authors employ a language of “erosion,” but they do not take full advantage of the concept. If we agree that erosion is an exogenously driven process, the cause for the democratic decline in these countries must rest outside the democratic institutions. The aforementioned authors, however, do not make a distinction between endogenous and exogenous drivers in their analysis of democratic backsliding. Instead, it seems that “erosion” has degenerated into a catch-all term that covers all forms of gradual change. The distinction is not only of analytical value; it is also of profound relevance for political praxis. For policy advice, scholars need to know where they locate the cause, whether inside or outside the institution whose development they seek to explain.

In an insightful review, Waldner and Lust (2018) have identified different explanatory strands for democratic backsliding. The clearest candidate for an exogenous factor in explaining democratic decline stems from the international arena. Levitsky and Way (2010) proposed the compelling idea that, depending on the Western linkages of an authoritarian country, leverage towards this country increases. Although they are more concerned with democratization processes, this insight can be applied to the reverse process of autocratization as well. The more a backsliding country A is linked in economic, social, or communicative ways to an authoritarian country B, the higher the potential influence of country B on country A.

Beyond linkage and leverage, unintentional autocratic diffusion processes, as well as an attractiveness of authoritarian values, are also probable candidates for explaining democratic erosion. Yet, empirical research has been rather cautious here (Bank 2017). If taking a historical perspective, exogeneity can also stem from a different angle. If we follow the line of argument that today’s democratic backsliding can be traced back to a series of previous low-quality democratizations in highly unequal and also economically weaker countries (Waldner and Lust 2018, 101–2), then this past history of low-quality democratizations could constitute an exogenous cause for today’s decline trend.

More convincing arguments are of an endogenous nature. As Neumann (1933) already showed in his classic account about the rise of Nazi Germany and the inherent dangers of the Weimar Constitution, backsliding often happens within democratic core institutions that have been initially built to guarantee horizontal and vertical accountability (Slater and Arugay 2018). Within weakly institutionalized settings, such as in super-presidential political systems, relatively unconstrained actors within these institutions constitute endogenous causes for backsliding. But even within strongly institutionalized settings, it is not automatically the robustness of electoral institutions or its parliamentary or presidential system that explains democratic stability but rather underlying endogenous causes. Cheibub (2007), for example, revisits the famous hypothesis that parliamentarism is more conducive for democratic stability because it mitigates previously existing societal conflict lines (Linz 1990). In his explanation, Cheibub endogenizes historical legacies. It is not the seemingly divisive nature of presidentialism that has caused more democratic backslidings and breakdowns but the fact that presidential systems have been established in unhospitable contexts for democratic survival, particularly in former military regimes. Arguments about the legacies of autocratic communist regimes on citizens’ attitudes take a similar explanatory route of endogenizing the past (Pop-Eleches and Tucker 2017). An explicit decay argument takes this historical legacy seriously and traces its influence on democratic decline today. Not all endangered democracies erode; some also decay.

Taking stock, beyond vaguely referring to interplays between endogenous and exogenous factors, the democratic backsliding debate provides rich empirical material to demonstrate the importance of being accurate in our arguments. Locating the source of the malaise, either inside or outside the institution, is essential not only for the sake of analytical clarity but also for competently advising political praxis.

Endogenous Ruptures as an Underexplored Research Area

Almost two decades ago, Weingast (2002, 692) prognosticated that “the endogenous emergence, choice, and survival of institutions are likely to be the major topic of the next decade.” Weingast was correct in his assessment, except for one important qualification. Thanks particularly to the research agenda of Thelen and her collaborators (Mahoney and Thelen 2010; Streeck and Thelen 2005; Thelen 2004), we know much more about how endogenous gradual change unfolds.

Recently, Capoccia (2016, 1100) provided an important stimulus when he called for a more “robust theorization of the role of institutions in causing political outcomes.” While focusing on gradually evolving institutions, he criticizes that the institutionalist research agenda remains too often a mere “epiphenomenal intermediary” (Capoccia 2016, 1100) between the strategies of actors and the aggregation of their preferences to broader outcomes, without playing an independent causal role. I concur with his crucial intervention—and would add that we need to expand our explanatory toolkit. It is rapid endogenous developments that are largely overlooked. The goal of the typology proposed here is, therefore, to point us to a type of change that has not received the same intellectual emphasis as its gradualist counterpart.

A major concern is to call for more attention to power and enterprising actors in neo-institutionalist theory. So far, the entering wedge for endogenous gradual explanations has been to take agency within institutions more seriously. Change agents of various sorts push internally for more behavioral leeway (Knight 1992; Mahoney and Thelen 2010; Streeck and Thelen 2005). Behavioral compliance of rule takers with institutional constraints decreed by rule makers should not be assumed but put into question. Due to the constraining effect of institutions, endogenous actors have routinely been conceptualized as weak and as gradually pushing for change. What has been overlooked is to broaden the perspective. Actors should not only be understood as mere rule takers but rather as also actively shaping the rules. Institutions consist not only of rule makers and rule takers but also of rule shapers.

Three locales can be identified in which change agents are particularly likely to emerge. First, Hall (2016) has recently reminded us that we should think more in terms of actor coalitions within institutional settings. Shifting power asymmetries between group members open space for endogenous change. Second, an institution comes usually not in the singular but as part of a larger institutional assemblage. Institutions are ensembles that unite subinstitutions under one umbrella. Yet, these subinstitutions are not seamlessly continuous but often only patchily linked with each other. The interspace between those loosely linked mosaic elements can be a fertile ground for the emergence of endogenous actors pushing for change. Third, Orren and Skowronek (1996) have focused on the development of institutions over time, marking the “intercurrence” of institutions. As institutions outlive each other and do not follow a strict sequence, an array of overlapping institutional settings result. Like for the spatial argument within institutional assortments, it is the nonsimultaneous temporal development of multiple institutions created at various times and operating within their own time logic that leaves elbowroom for endogenous actors. While these locales have been productively applied for gradual change, they should be transplanted to rupture arguments as well. Power asymmetries within actor coalitions, the interspace, and the intercurrence of institutional assemblages might also produce powerful actors that punctuate institutional inertia from within.

A focus on powerful actors points to a further research strand that social science explanations of ruptures should embrace. Cognitive heuristics that stem from social psychology are still underexplored in political science, but they can represent an interesting explanation for why (even risk-averse) actors break sharply with institutional demands (Kahnemann, Tversky, and Slovic 1982). Actor rationality is often bounded so that actors overreact in crisis situations. Instead of thinking about full information and a sufficient ability to rationally process this amount of information, the modal type of decision making is one of informational scarcity, informational asymmetry, time pressure, and exaggerated sense of urgency.

These mental shortcuts do not necessarily need to be inferior and error prone (Gigerenzer and Gaissmaier 2010), but what can be learned is that these shortcuts and unexpected actions often serve as an entry point for endogenous ruptures. While heuristics have found a remarkable entry into the study of electoral behavior and social movements, they are less applied to elite decisions (Fortunato and Stevenson 2019; Lau and Redlawsk 2001; Weyland 2019). It is these (lopsided) decisions by powerful, but often myopic, actors within institutional settings that can yield sudden and dramatic consequences, constituting an endogenous rupture and resulting in turbulent times.

The typology of explaining institutional change highlights an overlooked research field. Although actors have been characterized previously as being weak and pushing gradually for more room to maneuver, endogenous ruptures are both undertheorized and empirically underexplored in political science. Yet, powerful actors can serve as Schumpeterian entrepreneurs and break institutional inertia, guided by cognitive heuristics and producing (unintended) turbulences. Institutions still structure politics, but powerful actors can appear in all the situations in which powerless actors have been conceptualized: as independent change agents, as a result of shifting power asymmetries within coalitions, and in the interspaces and in the intercurrences over time between overlapping mosaic elements of larger institutional assemblages. These are the locales where we should look for sources of endogenous ruptures as much as we already did when studying endogenous decay.