1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Theories of institutions and institutional change
  5. 3. Empirical cases of institutional evolution: contemporary changes in German capitalism
  6. 4. Theoretical approaches to institutional change
  7. Acknowledgements
  8. References

The political-economic institutions that have traditionally reconciled economic efficiency with social solidarity in the advanced industrial countries, and specifically in the so-called ‘coordinated market economies’, are indisputably under pressure today. However, scholars disagree on the trajectory and significance of the institutional changes we can observe in many of these countries, and they generally lack the conceptual tools that would be necessary to resolve these disagreements. This article attempts to break through this theoretical impasse by providing a framework for determining the direction, identifying the mode, and assessing the meaning of the changes we can observe in levels of both economic coordination and social solidarity.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Theories of institutions and institutional change
  5. 3. Empirical cases of institutional evolution: contemporary changes in German capitalism
  6. 4. Theoretical approaches to institutional change
  7. Acknowledgements
  8. References

This article is a composite contribution that weaves together elements from several different projects on which I have been working over the past few years (Mahoney and Thelen 2010b; Martin and Thelen 2007; Palier and Thelen 2008; Streeck and Thelen 2005a; Swank et al. 2008; Thelen 2004; Thelen and Kume 2006). While grounded primarily in an empirical analysis of political-economic change in the advanced political economies, the paper draws as well on the conceptual and theoretical framework that I have been developing to understand institutional evolution more generally. I attempt to illustrate the utility of these concepts and this framework by putting them to work in the analysis of contemporary institutional change in advanced industrial economies.

The empirical focus of a great deal of my work relates to a central debate in the literature on the political economies of the most developed democracies concerning distinct ‘varieties of capitalism’ (VofC), a debate that has been very much shaped by the theory advanced by Peter Hall and David Soskice (Hall and Soskice 2001). As is well known, the Hall/Soskice framework is constructed around a fundamental distinction between so-called ‘liberal’ market economies (hereafter: LMEs) and ‘co-ordinated’ market economies (CMEs). The key difference goes back to the question of how employers co-ordinate their activities — whether mostly through the market (as in the liberal economies), or through various arrangements that allow firms to achieve joint gains through co-operation (as in the co-ordinated economies). This literature has focused special attention on the rather different institutional arrangements one finds in the CMEs that distinguish them from LMEs, including but not limited to: co-ordinated collective bargaining, arrangements for worker participation and voice at the plant level, well-developed systems of vocational education and training, durable ties between firms and their suppliers, and financial arrangements that can supply firms with ‘patient’ capital (Hall and Soskice 2001; Kitschelt et al. 1999a; Thelen 2001). From the beginning, this body of work has sought to refute the idea that contemporary market pressures (broadly captured under the headings of ‘globalization’ and/or ‘deindustrialization’) will drive a convergence on a single ‘best’ or ‘most efficient’ model of capitalism. In contrast to such arguments, the central claim at the heart of the VofC framework has been to insist that these systems — while in no way impervious to change — are nonetheless very durable; they operate on different logics and each exhibits strong self-enforcing tendencies (Hall and Soskice 2001: ch. 1; Hall and Soskice 2003: esp. 245–49).

Beyond its extraordinary influence in scholarly debates, this framework has provided some reassurance for those who worry about the breakdown of the institutions characteristic of the CMEs, which are widely seen as supporting a more egalitarian form of capitalism. Many studies have documented the resilience of key political-economic arrangements in the CMEs (and their continued distance from LMEs) despite new strains (Culpepper 2005; Goyer 2007; Hall 2007; Hall and Gingerich 2004; Pontusson 2005; Wallerstein et al. 1997). Other analyses point in a less sanguine direction, and in general, debates on the resilience of egalitarian capitalism tend to revolve around competing interpretations of how robust or fragile are the institutions that have traditionally defined these co-ordinated systems (see, e.g. the exchange on wage bargaining institutions in Germany in this journal: Hassel 1999, 2002, and Klikauer 2002).

My own earlier work focused primarily on the consequences of different institutional arrangements that characterize these different types of political economies, particularly in the co-ordinated countries. Over the past several years, however, I have turned my attention more to the origins of these institutions, as well as the related questions of what sustains them but also how they evolve and change through time. It seems clear to me that, if we want to know whether these institutions will survive or not, we will have to make some headway on how to conceptualize and explain institutional change itself.

This article is an attempt to advance that project. It proceeds in three steps. A first section lays out the theoretical terrain, including some of the puzzles and questions that have motivated my recent research. The second section presents empirical cases to illustrate some of the specific arguments I have been trying to advance within the context of the debates on varieties of capitalism. A third and final section returns to the theoretical level to discuss my general approach to studying institutional change, highlighting some of its main features and suggesting that this approach enjoys some distinct advantages over alternative perspectives in helping us understand how institutions evolve over time.

2. Theories of institutions and institutional change

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Theories of institutions and institutional change
  5. 3. Empirical cases of institutional evolution: contemporary changes in German capitalism
  6. 4. Theoretical approaches to institutional change
  7. Acknowledgements
  8. References

One of the most striking and surprising features of the literature on institutional change is its continued relative underdevelopment. The traditional comfort zone for institutional analysis in comparative politics and comparative political economy has involved the analysis of comparative statics, in which institutions are invoked as an independent or intervening variable to explain some other outcome — for example, policy differences or divergent patterns of social or political stratification. To this day, scholars are still generally more apt to ask what institutions do than how they evolve and change through time.1 A small though growing body of work has begun to explore the origins of some of the key institutions that are now seen as defining distinct models of capitalism (Iversen and Soskice 2009; Mares 2000; Martin and Swank 2008; Swenson 2002; Thelen 2004). However, the question of origins is not exactly the same as the question of institutional change, and on the latter much work remains to be done. Within the VofC literature, for example, there is still a strong tendency to emphasize continuity through time in the basic structure and logic of the two broad models of political economy. The origins of these systems lie in the distant past, but once they are up and running they are seen as rather stable and most of the feedback in each system is seen to be positive and operating to support arrangements that shore up the existing logic.2

The varieties of capitalism framework is often criticized as ‘static’ (Deeg and Jackson 2007; Schmidt 2002: 111), but in fact it has given us a very powerful and highly dynamic model of institutional reproduction.3 In line with recent work highlighting the important role played by employers in forging the institutions of co-ordination in the first place, VofC scholars point especially to employers’ continued support for these institutions in explaining institutional stability. A previous literature (on democratic corporatism) saw many of the institutions of co-ordination (e.g. centralized wage bargaining) as a product of labour strength. As such, these institutions were seen as vulnerable to employer rollback in a context — such as that prevailing in many of the advanced democracies since the 1970s — of high unemployment and declining union membership. By contrast, VofC scholars argue that within the CMEs employers have organized their production strategies around the existence of these arrangements, and therefore will not necessarily abandon these institutions in a period of market turbulence. This line of reasoning, well documented in many cases, takes us well beyond previous vague arguments about institutional ‘stickiness’ or ‘inertia’ as explanations of institutional stability.

However, as illuminating as this framework has been on the question of institutional reproduction, scholars working in this tradition have generally had much less to say about institutional change over time.4 This is a feature that the varieties of capitalism scholarship in fact shares with most institutional analysis (see also Hay 2005). Since the idea of persistence is virtually built into the definition of an institution, it should perhaps not be a surprise that the question of change is a weak spot in the literature as a whole and indeed across all varieties of institutionalism (Hall and Taylor 1996; Mahoney and Thelen 2010b).

If one looks to the broader literature in political science, a good deal of work falls back on a punctuated equilibrium model of change. The core idea is that key political or political-economic institutions are constructed in the context of some historical choice point in the past, and then once they are in place, they lay out an enduring logic of political development. In the historical institutionalist literature, one sees this in the language of ‘critical junctures’ (or choice points) that occur in the past, and the historical trajectories that flow from the legacies they produce. Much of this work emphasizes long stretches of institutional stability, periodically interrupted by episodes of relatively rapid innovation. These moments of innovation are usually associated with some kind of exogenous shock — for example, revolution, defeat in war or regime change — that disrupts the previous arrangements and opens the door for significant institutional innovation. The dominant model, then, is one of discontinuous change, and it is a model, therefore, that encourages us to draw a rather sharp line between the logic and analysis of institutional reproduction on one hand and that of institutional change on the other.

This general view is often also linked to claims about the relative weight of ‘agency’ versus ‘structure’ in various phases. One thinks, for example, of the seminal article by Ira Katznelson, which argues that institutions are mostly constraining in these long trajectory periods, and posits that critical juncture moments open up more space for agency (Katznelson 2003). Katznelson stresses high contingency in these episodes, such that actor choice becomes especially consequential in establishing new historic pathways. In so doing, Katznelson is invoking and embracing a familiar punctuated equilibrium view that emphasizes alternating periods of stability and structure punctuated by moments of agency and choice.

Finally, in the last several years as well, some of these ideas have come to be attached to arguments about path dependence in politics (e.g. Mahoney 2000; Pierson 2000). Many such arguments take on this same general structure concerning choices (or ‘paths taken’) at particular junctures whose effects get ‘locked in’ through positive feedback or increasing returns. Such models again encourage us to think about change in a particular way, namely as discontinuous, and characterized by the ‘breakdown’ of one set of institutions or relationships and its replacement with another.

Clearly, sometimes institutional change is abrupt and discontinuous: the collapse of political institutions in the former Soviet Union and Eastern Europe is an example. However, it is not at all clear that this model captures the most common way institutions change over time. My own work has focused on political-economic institutions in Germany, a country that has experienced some rather large shocks over the course of the twentieth century, including several regime changes, defeat in two world wars, and foreign occupation. If the model of punctuated equilibrium in the context of major historical break points works anywhere, it should have traction in this country. However, observing this case carefully — certainly for the political-economic institutions that I study — what is actually much more impressive is how resilient many of these institutions have proved even through these massive disjunctures (e.g. Thelen 2004). It seems that big breaks do not necessarily mean big openings for radical institutional reconfiguration.

But conversely, it is also not at all clear that all significant change has to have its source in a major exogenous shock that upends old arrangements and somehow clears the way for new ones. On the contrary, many of the cases I have been studying over the past several years (both historical and contemporary) suggest that significant change often takes place gradually and through a cumulation of seemingly small adjustments even in what Ann Swidler would call relatively ‘settled’ times, or at least in the absence of some obvious historic rupture (Swidler 1986). Table 1, which is drawn from joint work with Wolfgang Streeck, locates the focus of much of my work recently (Streeck and Thelen 2005a: 9). This table breaks out ‘process of change’ on one dimension — which can be incremental or abrupt — and the results of change on the other — which can produce either continuity or discontinuity.

Table 1.  Types of Institutional Change: Processes and Results
 Result of change
Process of changeIncrementalReproduction by adaptationGradual transformation
AbruptSurvival and returnBreakdown and replacement

A great deal of the work within the varieties of capitalism tradition emphasizes institutional reproduction through adaptation even in the face of new pressures (upper-left quadrant). As noted above, the framework provides a compelling account of the forces that help stabilize existing institutions even in the face of new pressures. The problem is that this literature also has a tendency to over-predict stability, or in some cases to code whatever changes we might observe as minor adjustments in a pattern of overall reproduction of the model (Streeck and Thelen 2005a: 8). And, as noted above, when it comes to models of change in the literature, the punctuated equilibrium model of breakdown and replacement (lower right quadrant) seems to be the only game in town.

Much of my recent work, therefore, has been motivated by the fact that we do not have adequate conceptual tools to capture and analyse forms of change that unfold incrementally but with transformative effects (upper right quadrant in Table 1, in bold). The significant changes that are indisputably transpiring in the political economies of the advanced industrial countries (particularly in the CMEs) are mostly of this variety and therefore do not conform to the dominant punctuated equilibrium model of institutional breakdown/replacement. My work therefore has tried to advance our understanding of institutional change by focusing on two fronts. First, my empirical work has attempted to articulate a political coalitional theory of institutional change that underscores the way in which shifts in the political coalitions on which institutions rest drive changes in the form and functions those institutions take over time (see especially Thelen 2004; also Palier and Thelen 2008; Thelen and Kume 2006). Second, and more generally, together with Wolfgang Streeck and with James Mahoney, I have tried to develop some conceptual tools that can both capture and help us analyse these forms of institutional change that are gradual but cumulatively transformative, rather than abrupt and discontinuous (Mahoney and Thelen 2010b; Streeck and Thelen 2005a, see also Djelic and Quack 2003).

Before turning to contemporary empirical examples, a word is in order about the case through which I came to the issue of institutional change. This was the case of the German system of vocational education and training, whose origins and evolution I tracked over about a century of development in the context of a broader book project (Thelen 2004). The vocational education and training system recommended itself for intensive study because this is a set of institutions that is widely seen as a core foundation of Germany's version of a ‘co-ordinated market economy’. The system has been widely celebrated in the literature as a model solution to many of the collective action problems that often plague private sector training regimes (Acemoglu and Pischke 1999; Finegold and Soskice 1988). It is a system that has generated a rather large pool of portable skills, and one that therefore has correctly been seen as a key institutional support for Germany's famous high-wage, high-skill, high-value-added manufacturing model. For all these reasons, this is a system that is seen as having helped to reconcile Germany's comparatively strong unions with strong performance in world markets (Streeck 1997).

Against the backdrop of its current form and functions, an examination of the origins of the German training system revealed some surprises. Looking back, one finds that the core institutional innovation around which this contemporary pillar of social partnership came to be built was a piece of legislation passed in 1897 by an authoritarian government.5 The original system did not envision a role for industry, let alone of course conceive of a place at the table for organized labour. Indeed, if anything, the system was designed to weaken unions by shoring up and supporting a class of independent artisans as a political bulwark against Germany's then quite revolutionary working class movement. In short, this is a system with roots in the distant past, but one that by the late twentieth century had been very significantly reconfigured, even if there are some core elements that survived in a recognizable form — despite the transition to democracy, the decline of the artisanal economy, and all of the political ruptures Germany experienced in the intervening period. The most striking aspect of this whole process, however, is that the most important revisions to this system took place outside of the historic break points. They did not take place through massive institutional re-engineering, but rather through incremental changes that over time cumulated into a very significant institutional transformation.

The argument that I develop to account for this transformation clashes strongly with perspectives that separate out periods of stability from those of change, and contradicts as well the Katznelson idea of periods characterized by the primacy of agency or structure. Instead, the argument that I construct is a political-coalitional theory of institutional reproduction and transformation that traces significant changes in the form these institutions take and the functions they perform in society to shifts in the coalitional base on which these institutions rest (Thelen 2004). Like the path dependence literature, my explanation features positive feedback, but these feedback effects are not associated with the faithful reproduction of the system as originally constituted (as in traditional path-dependency accounts). Rather, the role that positive feedback plays in my analysis is to illuminate processes through which institutions created for one set of purposes and resting on the shoulders of one set of actors could be carried forward on the shoulders of another coalition altogether. These features explain why the case of German vocational training is one that combines elements of increasing returns and positive feedback (and therefore some striking elements of stability even through these massive historic breakpoints) with incremental changes that add up to a real transformation.

3. Empirical cases of institutional evolution: contemporary changes in German capitalism

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Theories of institutions and institutional change
  5. 3. Empirical cases of institutional evolution: contemporary changes in German capitalism
  6. 4. Theoretical approaches to institutional change
  7. Acknowledgements
  8. References

The insights derived from this historical study inform my approach to understanding institutional developments in the present. Clearly, many of the institutional arrangements that have traditionally distinguished CMEs such as Germany from the liberal model are under tremendous strain. New pressures emanate above all from globalization and the decline of manufacturing in these mature industrial economies.

To set the scene for a discussion of contemporary changes, we recall that there are two broad arguments in the literature concerning the future of co-ordinated capitalism. On one hand, there is the classic varieties-of-capitalism perspective, which sees the divergent institutional arrangements characteristic of LMEs and CMEs as fundamentally robust and resilient (Hall 2007; Hall and Soskice 2003). As noted above, these two models constitute quite distinct equilibria, each having strong self-enforcing characteristics. Specifically for the CMEs, the argument is that employers will not abandon or dismantle institutions such as co-ordinated wage bargaining or skill formation because they have organized their production strategies around these institutions and thus have come to rely on them for their success in the market (Thelen 2000). Clearly, scholars in this camp acknowledge some of the changes that have occurred but they do not see these changes as pulling CMEs inexorably in the direction of LMEs, a conclusion that many studies appear to support (Deeg and Lütz 2000; Goyer 2007; Hall 2007; Kenworthy 2001). VofC scholars thus tend to see the changes that are occurring as essentially adaptive rather than transformative — not undermining existing arrangements and perhaps even serving to shore them up (Hall 2007; Hall and Gingerich 2004).

On the other hand, recent years have seen the resurgence of a convergence thesis — albeit one that is considerably more sophisticated than earlier versions. Authors in this camp perceive in contemporary developments a genuine breakdown, or decay through ‘self-exhaustion’, of the arrangements that have distinguished the co-ordinated political economies in the past (Glyn 2006; Howell 2003; Streeck 2009; Streeck and Hassel 2004). They are likely to code many of the very same changes that VofC scholars see as adaptive as instead transformative, and involving a relatively straightforward trend toward liberalization. Recent developments have therefore led some to wonder if there is just one (liberal) variety of capitalism, after all (Howell 2003; see also Goodin 2003).

My own perspective comes out of several related projects, including joint work with Ikuo Kume, which examines developments in Germany, Sweden, and Japan (Thelen and Kume 2006), with Bruno Palier, which compares France and Germany (Palier and Thelen 2008), and with Duane Swank and Cathie Jo Martin, which tracks broad trends across 18 OECD countries and looks in more detail at select countries including Germany and Denmark (Martin and Thelen 2007; Swank et al. 2008). In this article, however, I will focus on the German case, touching briefly on developments in three different institutional arenas — industrial relations, vocational training, and labour market and social policy. Determining the direction of change in Germany should be instructive because this country has always been widely perceived as a crucial (even paradigmatic) case in the VofC debates. But looking across several different institutional arenas in one country allows us also to draw some conclusions about the conditions that produce different modes of change, which is something to which I want to return in the conclusion.

To preview, I find that the overall direction of change in this case is not well captured by the new (or resurgent) convergence theories or even the language of liberalization. Stepping back from the current period, the German political economy over the past 100+ years exhibits features that distinguish it fundamentally from the liberal model. These features have survived massive political and economic shifts and shocks, and they are unlikely to evaporate in the face of current trends. How, then to capture the changes that are indisputably unfolding in this country? Rather than a breakdown of co-ordination, developments across all three of the realms I have examined point instead to a significant reconfiguration of co-ordination on the basis of a very different (much narrower) political coalition, one that in Germany centres especially though not exclusively on large firms in manufacturing and their workers (Thelen and Kume 2006; see also Carlin and Soskice 2009).

The core framework that I use to make sense of these developments is presented in Figure 1, for which the central insight comes from work by Martin Höpner, though I have adapted his framework to fit my slightly different categories (Höpner 2007). This framework retains the core distinction between LMEs and CMEs — which I believe continues to capture something very important about the way that different political economies are organized. However, there has always been considerable variation not just between LMEs and CMEs but also among CMEs themselves. These differences are captured on the second, vertical, dimension, which ranges from very solidaristic forms of co-ordination to what (following Swenson) I have elsewhere called segmentalist forms of co-ordination.6


Figure 1. Two Dimensions of Comparative Capitalism. Adapted, with permission, from Höpner (2007).

Download figure to PowerPoint

To clarify the difference between solidarism and segmentalism we can think about the different forms that co-ordination takes across CMEs. Take for example the realm of vocational education and training, a defining feature of such systems. In some countries training is organized — either in schools or in firms themselves — to reach a large number of youths, and to provide portable skills through standardization in the content of training. These systems are solidaristic in the sense that they are relatively encompassing in scope and relatively uniform in content. However, as elaborated elsewhere (Thelen 2004: ch. 1), the Japanese system of training is also part of a co-ordinated system but operates on a very different, more segmentalist logic — a system which therefore has rather different implications for labour strength and distributional outcomes of concern to many political economists. In the Japanese case, training is concentrated in large firms, and the skills that youth acquire are company specific (i.e. not organized around national occupational labour markets, but instead around internal labour markets within firms). Or take the example of industrial relations. In Scandinavia, co-ordinated wage bargaining has been very encompassing, producing contracts or framing agreements that have traditionally generated a high degree of uniformity across firms of different sizes as well as across sectors. Japan also achieves a high score on key measures of wage co-ordination (e.g. Kenworthy 2001 assigns a 5 to Japan on a five-point scale), but it is clear that co-ordination is structured and achieved very differently in this case. In Japan, large manufacturing firms co-ordinate among themselves, and while the settlements they reach establish benchmarks for other firms and sectors, this system by no means generates similarly egalitarian wages and working conditions across industries nor, especially, across firms of different sizes.

This difference in levels of solidarism/segmentalism among CMEs is captured in Figure 1, which provides a highly stylized depiction of where Sweden and Japan would be located in this two-dimensional space — with Sweden exhibiting high levels of both co-ordination and solidarism, while Japan is co-ordinated but organized according to a more segmentalist logic.

More importantly, this framework brings out features of the current transition in some countries that are lost or obscured in debates that are organized around tracking movements on a one-dimensional continuum that ranges from ‘liberal’ to ‘co-ordinated’. In previous work, for example, I have argued that developments in key areas in Germany do not conform to a model of liberalization so much as they represent a trend toward segmentation and dualization (Thelen and Kume 2006). Whether or not formal institutions for co-ordination break down, a narrowing of their coalitional base may have very important implications for outcomes of interest, most obviously for labour strength and distributional outcomes. How such narrowing might occur is not so much a question of institutions as it is of political dynamics and especially of the political-coalitional foundations on which institutions rest. High levels of solidarism can give way to greater segmentalism gradually, for example, as employment shifts from manufacturing to services. Thus, if — as in some realms in Germany — solidaristic institutions and practices that were developed for industry in the 1950s and 1960s fail to take hold in services in the 1990s and 2000s, the coalitional base on which these institutions rest will narrow perforce and almost imperceptibly. The important conceptual point, however, is that such developments do not ‘register’ as change on the varieties of capitalism dimension (the x-axis), since they do not represent a breakdown of co-ordination so much as they do a reconfiguration of co-ordination on less solidaristic terms (as indicated by the arrow in Figure 1) (see, especially, Höpner 2007: 16–17; also Thelen and Kume 2006).7 This explains why there seems sometimes to be a ‘dialogue of the deaf’ between VofC and convergence theories even if the proponents of each are looking at the very same institutions. I turn now to three arenas in which significant changes in the German political economy appear to conform to such a pattern of dualization.

Collective Bargaining

In the 1990s a significant literature predicted the breakdown of centralized bargaining through competitive deregulation (Flecker and Schulten 1999; Kapstein 1996; Katz and Darbishire 1999; Mahnkopf and Altvater 1995; Martin and Ross 1999). The core logic behind these predictions ran roughly as follows. Because employers had originally agreed to centralize bargaining to secure wage moderation in the inflationary context of the 1970s, the shift towards persistently high unemployment in the 1980s removed the rationale for their continued participation. Since they could now count on the discipline of the market to control wages, employers were expected to launch a neo-liberal attack against the ‘rigidities’ of centralized bargaining (Siebert 1997).

As we now know, however, centralized bargaining arrangements did not collapse as predicted (Wallerstein and Golden 2000; Wallerstein et al. 1997). In the case of Germany, nothing much has changed in terms of the formal structure of negotiations, which are still conducted at the multi-industrial level and by unions that in the meantime — through mergers — have actually become more encompassing than they were before. The level of formal stability we observe in Germany is striking when one considers the inflammatory rhetoric levelled against centralized bargaining in recent years by some quite powerful actors, notably Hans-Olaf Henkel, the former head of the German Industry Association (BDI). Formal stability in Germany also stands out against the backdrop of other cases such as Britain and Denmark which did in fact witness real formal decentralization in the 1980s and 1990s (Wallerstein and Golden 2000). While it has clearly become extremely controversial in some circles, co-ordinated wage bargaining in Germany has not broken down. These are the signs of stability that VofC scholars emphasize and invoke to document the relative resilience of CME institutions even in the face of new pressures (Hall and Gingerich 2004; Thelen 2000).

However, as Jacob Hacker has put it in a very different context, the formal stability of this system only really captures half the truth (Hacker 2005). The fact is that there have indeed been some rather consequential changes in bargaining, even if these changes have not taken the form of a formal decentralization, let alone a complete breakdown of traditional institutions. The most important of these trends in Germany — much commented upon in the literature — has been a noticeable shrinkage in the coverage of collective bargaining (e.g. Hassel 1999). So while it is certainly noteworthy that there has been no free fall into deregulation, it is also true that significantly fewer workers are now covered by the deals that get negotiated within the context of these still rather centralized negotiations. Declining coverage has resulted both from individual firms (especially small firms) opting out of these agreements, but also from a creeping expansion in the number of workers in various non-standard employment relations, especially in services (Thelen and van Wijnbergen 2003; on the growth of atypical work, see especially Eichhorst and Kaiser 2006). In short, the formal institutions are stable but at the same time, they cover a shrinking core of workers, concentrated especially though of course not exclusively in large manufacturing firms.

As the political coalitional foundations of these bargaining institutions narrow, however, one can observe as well a subtle shift in the functions that these institutions are performing. Where collective bargaining outcomes used to establish the parameters or a baseline for plant bargaining, settlements are now in many cases organized around ratifying some of the deals that have already been struck by strong works councils in large firms and their employers (Rehder and Hassel 2001). In other words, what we see is something that John Goldthorpe wrote about 25 years ago, namely a trend toward ‘dualism’ (or what I have been calling segmentalism) that involves growing divide between a system that still covers (and covers well) a significant proportion of ‘core’ workers, but with larger numbers of workers outside that core (Goldthorpe 1984). If one focuses excessively on the formal structure of bargaining, one risks missing very significant signs of drift in the system of collective bargaining, as it covers a shrinking core of workers (on ‘drift’, see especially Hacker 2005; also Mahoney and Thelen 2010b; Streeck and Thelen 2005b).

Vocational Education and Training

A second arena is vocational education and training, where the mode and direction of change are similar. This is an especially interesting case because, unlike the heavily criticized institutions of centralized wage bargaining, the vocational training system is extremely popular in Germany. It enjoys wide support not just from academics but from the full range of major organized interests. Organized business and labour, Social Democrats and Christian Democrats: all are committed to preserving the German system of training and adapting it to changes in production technology and sectoral employment patterns. Thus, the past decade has seen strong co-operation across party lines, as well as across the class divide, in implementing reforms designed to update skill profiles and training to adjust to new production requirements (Thelen and Busemeyer 2008). Because skills are more complex and technically demanding than before, arguably the training has gotten better.

However, the problems currently plaguing the German training system have less to do with the quality of training then they do with the quantity of training slots available to youth — above all, shortfalls in the supply of apprenticeships offered by firms. This problem is in some significant measure a function of the failure of the service sector (the only real source of employment growth in Germany) to embrace the traditional ‘dual system’ of training on anywhere near the same scale as in manufacturing (Culpepper and Thelen 2007; Thelen 2007). Noteworthy progress has been made to develop new professions in key emerging services (e.g. information and communications technologies — ICT), and some branches of the service sector (banking, notably) have well developed apprenticeship training. But overall, service sector firms have been far less likely to train, and so as employment structures shift, overall training opportunities have declined.8

In addition, however, there is also evidence that the updating operations mentioned above may themselves have had a somewhat corrosive side effect on the system even in manufacturing. Germany's smallest firms — traditionally very important players in the German vocational training system — in many cases can no longer afford to sponsor the more demanding training that is now called for (Wagner 1997, 1999). Thus, the share of apprentices trained in the country's smallest firms has fallen very significantly over the past two decades (Thelen 2007: 253).

Similar to collective bargaining, if one looks at this system from one angle — for example, observing the continued strong demand on the part of youth for in-firm training and noting recent successes in developing new occupations as well as ongoing progress in updating older ones — one sees continued support and many signs of positive adaptation. Such developments seem to confirm the institution's resilience and adaptability in the face of new technological and market challenges. However, an equally important development — which is not necessarily ‘picked up’ on a one-dimensional LME-CME continuum — is again a significant shrinkage in the system's coverage — as fewer youth are able to secure training spots and as this set of institutions too is increasingly carried forward on the shoulders of a narrower base of business participation — as manufacturing declines and as the smallest firms drop out. With this shrinkage comes as well the threat (and to some extent the reality at least in some large companies) of new pressures that push away from the portable skills that traditionally characterized the German model toward training geared more to companies’ own specific skill needs — in other words pressures here too towards more segmentalism in training to go along with increasing dualism in collective bargaining (Thelen 2007; Thelen and Busemeyer 2008).

Labour Market and Welfare Institutions

Developments in the areas of labour market and social welfare institutions exhibit a different mode of change, even if the direction of change is quite similar to the other examples given above. As is well known, Paul Pierson's work on path dependence grew out of observations about the resilience of key social programmes that despite all sorts of pressures seemed to be politically ‘locked in’ by the power of the constituencies these programmes themselves generated over time (Pierson 1994). This kind of lock-in model has been especially applied to welfare regimes such as the German one that are built around a social insurance logic. In such systems, the contributions a person makes during his or her working life take on the character of deferred wages, which makes it extremely difficult to retrench (Palier and Martin 2007). Pierson is certainly correct on one level: many of the core protections traditionally afforded to those in standard employment relationships in Germany — job security provisions, some aspects of status-based social insurance — have proved to be politically more or less unassailable.

However, even if outright dismantling of the traditional protections has been politically impossible, the system as a whole is not as locked in as that literature sometimes implies. In fact, very significant changes have been accomplished in a different way, through what Eric Schickler has called institutional layering (Schickler 2001). This mode of change depicts a situation in which the original institutions are left in place, but new elements are added alongside the old system, elements that have transformative potential in the longer run (Mahoney and Thelen 2010b; Streeck and Thelen 2005b). This is the pattern of change one observes in German labour market institutions, through reforms that promote the growth of a low-wage sector and the accompanying changes this has prompted with respect to the two-tier system of unemployment support that emerged under the infamous Hartz reforms.

Beginning with labour market policy: recent reforms in Germany have mostly not touched traditional employment protections for core workers. In fact, the safeguards that core workers in the German labour market enjoy are as strong as ever; there are multiple sources of protection rooted in a combination of legislative rules and collectively bargained regulations, but augmented as well in large firms by powerful works councils and in a growing number of cases company-level employment pacts that provide explicit job guarantees (Rehder 2003). However, alongside this stable core, recent reforms have also encouraged the growth of low-wage work, particularly in the service sector with the expansion of so-called mini-jobs and other forms of atypical employment. In other words, labour market reforms in the late 1990s and early 2000s have essentially created a new layer, separate and indeed ‘encapsulated’ from those in standard employment relationships (Eichhorst and Kaiser 2006: 16, 21).

Social policy is an interesting case, as Hartz IV has mostly (with justification) been coded as an assault on some of the protections traditionally enjoyed by labour market insiders (Trampusch 2005). The dramatic reduction in unemployment insurance, from three years to one year, certainly sparked noisy demonstrations that gave voice to outrage and alarm at the thought that a regular worker who fell on hard times might be dumped into the ranks of the poor and supported at a minimum level regardless of his previous contribution record (Dribbusch 2004). It is debatable, however, whether or not labour market insiders are the ones who are most affected by the reform. Clearly, skilled workers in core manufacturing industries are unlikely to feel the bite of Hartz IV. As noted above, workers in large firms in particular enjoy very strong employment protection, and skilled workers (especially in the West) are unlikely to remain unemployed for over a year. Instead, the most obvious losers under the new legislation are unskilled workers and, especially, the long-term unemployed, over half of whom are easterners. The uneven impact of Hartz IV explains why, after an initial nationwide outcry, demonstrations were increasingly concentrated in the East once the law was actually being implemented (Dribbusch 2004).

Moreover, whatever the balance sheet in terms of winners and losers on the benefit-cut side of the reform, what is in any event certainly true is that Hartz IV also stabilized and shored up the traditional social insurance system on the financing side of the reform. It did so by relieving the social insurance funds of the obligation to provide benefits for people whose contribution records were never going to be able to cover their entitlements under the old system (particularly easterners, but also some in unsteady or atypical employment). In other words, focusing on the financing side of the reform, Hartz IV was very much part of the same kind of quarantine operation that serves (in classic dualist fashion) to shore up the traditional social insurance model by sharpening the distinction between workers who are still linked to the core labour market (therefore still covered by that system) — and those who are outside this system and fall into an alternative, parallel world of means-tested social assistance that operates on a completely different logic: not status/income maintenance but poverty alleviation (Eichhorst and Kaiser 2006; Palier and Thelen 2008). Viewed in this light we again see a pattern of change that stabilizes the old system (albeit for a smaller core) by adding, alongside it, a parallel system of state-financed means-tested assistance (i.e. a classic ‘layering’ operation) — which also conforms to a segmentalist direction of change.

Résumé: Direction of Change

All in all, the direction of change in Germany in these three realms seems more to conform to a pattern of dualization than liberalization. The coalitional dynamics that have produced this are complex, and ‘insiders’ have not won all the key reform battles (see, e.g. Häusermann forthcoming). However, simplifying somewhat, it also appears to be the case that the very same coalition that has allowed the country to avoid succumbing to liberalization has also actually prevented it from sustaining high levels of solidarism.

In collaborative work with Duane Swank and Cathie Jo Martin, we track changes over time in a broader range of 18 OECD countries in the two dimensional space introduced in Figure 1. While a full recounting of the results of that analysis lies outside the parameters of this article (but see Swank et al. 2008), a couple of points are worth mentioning and these are depicted schematically in Figure 2. A first is that it appears to be unusual in the extreme for a country to move directly from the NE to the NW quadrant, that is, from strongly co-ordinated and solidaristic to a liberal model (the horizontal arrow in Figure 2).9 This sort of straightforward neo-liberal shift (closest to a strong convergence argument) seems to be ruled out politically by the coalitional alignments prevailing in most co-ordinated economies (Palier and Thelen 2008). More common are moves from the NE to the SE quadrant (the diagonal arrow in Figure 2), that is, towards increasing dualism — which is occurring in some, but not all, CMEs.10


Figure 2. Likely and Unlikely Trajectories of Change in Advanced Political Economies. Source: Palier and Thelen (2008); also Swank et al. (2008).

Download figure to PowerPoint

A second point from the cross-national analysis concerns the clustering of countries overall, and here what we find is that the LMEs mostly occupy the SW quadrant while CMEs occupy various positions in the NE and SE quadrants. The NW quadrant — that is, high solidarism and low co-ordination — by contrast, is more or less empty, suggesting that LMEs cannot sustain high levels of solidarism.11 The relative vacuum here is significant, and it is completely consistent with what we know from cross-national work that suggests that a high degree of employer co-ordination is necessary (though clearly not sufficient) to achieve high levels of social solidarity (see especially Kitschelt et al. 1999b). This pattern is also consistent with historical work on individual countries that clearly demonstrates that employer organization and power are not antithetical to high levels of labour organization and power — and may well be necessary to achieve the latter (Iversen and Soskice 2009; Swenson 2002; Thelen 2004).

Résumé: Modes and Mechanisms of Change

Beyond the direction of change, some observations are in order about the modes and mechanisms of change that are occurring in Germany and other CMEs. Here what is especially striking is that the kinds of changes that are taking place mostly do not involve a frontal attack on traditional institutions. Indeed, as mentioned, in cases such as training, some of these developments may even be partly a side effect of adapting and modernizing the system. Likewise in collective bargaining, demands to revise outright the formal rules of the game (labour law reforms) have mostly not yielded much. The bigger changes have been accomplished more gradually and incrementally, through defections and drift as mentioned above. But, in addition, subtle but significant changes have been sought through efforts to reinterpret existing rules (and with conflicts in some cases being played out more in the courts than in the legislature).

An example of such strategies of change — though institutional conversion— can be found in the area of collective bargaining (on conversion see Streeck and Thelen 2005a, and Mahoney and Thelen 2010b). German employers have sought to challenge the traditional relationship between industry-level bargaining and local (plant-based) negotiations, by bringing court cases centring on the so-called ‘favourability clause’ (Günstigkeitsprinzip) (see especially Rehder 2006, on which I draw here). This rule is central to the German industrial relations system, and has clearly served in the past to shore up centralized bargaining. The law specifies that certain issues (including wages) that are regulated by collective bargaining are in fact reserved for the unions, and it therefore prohibits plant-level works councils from negotiating on these items unless, as the law stipulates, the result of these negotiations would redound ‘to the benefit of the worker’ (zu Gunsten des Arbeitnehmers, therefore: Günstigkeitsprinzip). It is permissible under the law, for example, for works councils to agree to wage increases above the contractual rate, but concession bargaining is ruled out.

There has been an interesting struggle in recent years, played out in the courts, in which firms (sometimes actually in alliance with their works councils) have argued that, in the context of high unemployment, a local deal that trades off lower wages against employment protection does redound ‘to the benefit of the worker’ and therefore is consistent with this law. Thus far, industrial unions have prevailed in the courts on this issue. However, should the courts shift position on this, then — without any kind of major legislative fight or rewriting of the formal law — what will have happened is that a rule that was originally designed to shore up centralization would be converted into a potentially very powerful vehicle for decentralization. Once again, the point is that sometimes the conflicts are played out explicitly over the formal institutions, but some of the most significant changes that one observes in many political-economic institutions today can actually be accomplished without traditional structures or rules breaking down in an absolute sense.

In sum, and just as in the historical example of vocational education and training cited at the outset, the changes that are taking place across all these realms are such that they combine elements of stability and change. Where this is the case, it is not helpful to ask whether the old institutions have broken down or not. The point, rather, is to look at the way in which a shift in the coalition on which these institutions rest has produced a renegotiation of some aspects of the old system, while leaving other features in place.

4. Theoretical approaches to institutional change

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Theories of institutions and institutional change
  5. 3. Empirical cases of institutional evolution: contemporary changes in German capitalism
  6. 4. Theoretical approaches to institutional change
  7. Acknowledgements
  8. References

These observations about the different modes of change we can observe in Germany across these various arenas bring us back to the theoretical level and to the question of how to capture subtle shifts that go undetected if one is operating with a punctuated equilibrium view of change. The upper right quadrant in Table 1 locates the analytic space occupied by various modes of change that are ‘incremental but cumulatively transformative’ (Streeck and Thelen 2005a). The foregoing analysis has touched upon a number of these that are in evidence in the German case: institutional drift (collective bargaining and vocational training), conversion through reinterpretation (Günstigkeitsprinzip), and layering (labour market and social policy).

In joint work, James Mahoney and I have offered some general propositions about the conditions under which one strategy or mode of change is more likely to emerge than another. Our framework links particular modes of change to specific features of the prevailing political context and of existing institutional rules, by asking two broad questions: (i) does the political context afford defenders of the status quo strong or weak veto possibilities? And (ii) does the targeted institution afford actors opportunities for exerting discretion in its implementation and enforcement? (Mahoney and Thelen 2010b). The answers to these questions produce the analytic space depicted in Table 2.

Table 2.  Modes of Change in Various Arenas in the Contemporary German Political Economy
 Characteristics of the targeted institution
Low level of discretion in interpretation/enforcementHigh level of discretion in interpretation/enforcement
Characteristics of the political contextStronger veto possibilitiesLayering (pensions, social insurance)Drift (collective bargaining, vocational training)
Weaker veto possibilitiesDisplacement (relatively rare)Conversion (favorability principle- Günstigkeitsprinzip)

The foregoing discussion provides empirical examples of all of these modes of change except for outright displacement (lower left quadrant). As I have argued, the wholesale replacement of traditional rules and institutions with new ones is rare in the politics of reform in contemporary advanced capitalist economies.

Developments in labour market and social policy in Germany provide an example of institutional layering. This is an area of policy characterized by relatively low levels of discretion in the implementation and enforcement of rules, the most important conflicts being played out instead in the legislature where (for example) the level and duration of unemployment benefits are set. While there may be some room for discretion in enforcement/implementation in individual cases (e.g. as bureaucrats assess a client's qualification for specific programmes), it is fair to say that the more consequential politics are those that are fought directly over the rules themselves — making it less likely that drift or conversion emerge as the main mode of significant change.

Whether the dominant mode of change will be layering or displacement then turns on the number of veto points and the strength of veto players. In the German case, unions and representatives of core workers have been capable of blocking all initiatives on the Right (associated especially with the Free Democratic Party, FDP) to dismantle entirely the traditional social insurance model in favour of a more liberal regime, but (alongside their employers) they have also fended off initiatives from the Left (floated occasionally by the Green Party) to replace the existing system with a more universalistic tax-based model. However, as we have seen, what they have not been able to do is prevent the layering of alternative arrangements alongside the traditional social insurance model — e.g. low-wage alternatives to permanent full-time employment and the associated means-tested programmes described above.

The case of collective bargaining, by contrast, represents a case of drift. Here, too, powerful veto players have either not sought (large manufacturing employers) or been able to block (unions) formal institutional changes that would challenge the basic structure of central bargaining. However, as pointed out above, the system has experienced significant drift as the rules and regulations negotiated in the context of such bargaining now cover a shrinking group of workers. Such change is possible because even if formal institutional change has been blocked, the character of the institution itself allows significant discretion on the part of individual actors in the implementation and enforcement of these deals. In the extreme case, companies can opt out of the agreements altogether (Koalitionsfreiheit) but in addition, individual firms and works councils (and/or unions) can opt to apply the regulations contained in a given agreement strictly or more leniently. This latitude for discretion offers precisely the kind of institutional environment in which drift as a mode of change is most likely to emerge.

Finally, the example of the favourability principle (Günstigkeitsprinzip) represents at least the attempt at a strategy of conversion on the part of employers. In this case, the formal rule specifies what is and is not permissible, but much depends on how this rule is interpreted. Does the rule allow works councils to trade off concessions on one dimension (wages) for benefits on another dimension (employment security)? Or must the local deal improve on the industrial contract on each dimension separately? Such is the question that employers put to the courts, in the hope they would rule in favour of the former. The reality is that works councils already often agree to concessions in the ‘grey zone’ of the law, which therefore transpire under the radar (drift). By bringing the issue to the courts, employers hoped to anchor these practices formally, through an authoritative reinterpretation of the law. Clearly employers could have attempted to revise the law itself (in the legislature), but they correctly perceived that this strategy would run up against the opposition of major veto players (social democrats, unions). Thus, the attempt here was to shift the arena of conflict, that is, to move it to an arena where unions and their political allies have fewer possibilities to veto or block the reform — namely the courts. The idea was in effect to move the conflict from the upper right to the lower right quadrant (conversion).

More examples could be given and more could be said about the particular modes and mechanisms at work — across diverse institutional arenas in Germany but also cross-nationally. I close, however, with a more general point about how we can conceptualize change in the advanced industrial political economies generally.

Even a cursory glance at the institutionalist literature reveals that there are many ways to conceptualize institutions, and different scholars lean in different directions on this issue. The varieties of capitalism framework, for example, draws insight from a Williamsonian perspective that emphasizes the co-ordinating functions institutions perform. Historical institutionalists, by contrast, tend to embrace a slightly different, more political, view, seeing institutions as the legacy of concrete historical processes and battles. These two perspectives are by no means necessarily at odds with one another, and in many ways they usefully shed light on different aspects of institutions. However, approaches that put the power-distributional features of institutions (as opposed to their co-ordinating functions) at the centre of the analysis may enjoy some advantages when it comes to thinking about institutional change.

If one conceives of institutions as sets of rules that are enforced or complied with by actors and organizations of various sorts, this brings into strong relief the fact that rules are never ‘simply’ applied, they are always interpreted, enforced and enacted, and, of course, by actors who have divergent and conflicting interests (Streeck and Thelen 2005a). This feature of institutions is important to bear in mind because analytically it opens up a way of analysing something to which Paul Pierson has drawn attention, namely the ‘gaps’ (as he puts it) that exist from the start, or that emerge over time, between the design of an institution and its actual on-the-ground implementation and effects (Pierson 2004: ch. 4). Where do these gaps come from? The previous discussion alluded to several potential sources.

The first has to do with the cognitive limits of rule makers or institutional designers. This is the simple fact that institutional designers never fully control the uses to which their creations are put (Pierson 2004: esp. 115–22; Elster 2006; Thelen 2004). To return to the first historical example cited above: clearly when Germany's authoritarian leaders created a system for training under the control of a group of reactionary artisans, they did not dream that this would eventually become a core institutional support for ‘diversified quality production’ and social partnership between labour unions and business interests. But this is precisely how these institutions came to function under the somewhat different coalition that inherited them and also adapted them to their own, very different, ends (Thelen 2004: chs 2, 5).

Second, gaps emerge because institution-building is often a matter of political compromise (Palier 2005; Schickler 2001). This means that institutions and rules are often ambiguous from the beginning, almost by design, as a consequence of the particular (often conflicting) coalition of interests that presides over their founding. Examples abound, but one that is touched on in the discussion above is centralized collective bargaining itself. This is a set of institutions which by its very nature is jointly forged and sustained by (at least segments of) labour and capital (Swenson 1991). As such, these arrangements embody and institutionalize different, often conflicting, goals — which can of course also get balanced out rather differently over time.

Third, institutions are not neutral. Because they instantiate power, they are contested. However, since the ‘losers’ in these contests do not always go away, it sometimes happens that actors who are not part of the ‘design coalition’ may nonetheless find ways to occupy and redeploy institutions not of their own making. The example of the German vocational training model comes again to mind — its transformation from a weapon against organized labour into a core foundation for union strength. It is also perhaps worth noting in this context that this is a point that is almost completely lost in the path dependence literature, which invariably stresses how actors adapt their goals and strategies to the prevailing institutions (i.e. with the causal arrow running from institutions to strategies). My point here is something like the opposite, namely, that actors are always trying to bend the institutions and reinterpret the rules to fit their interests and goals (see also Mahoney and Thelen 2010b).

A fourth source of gaps, perhaps better thought of as an umbrella for the others, is time (Pierson 2004: ch. 4, passim). Many of the institutions of interest to scholars of political economy have long outlived their designers. If one thinks about institutions as rules, it is clear that establishing what a rule ‘says’ or requires is something that is always done in a context and with reference to the particular circumstances to which the rule is getting applied (Streeck and Thelen 2005a). Thus, changes in the context can open up tremendous space for reinterpretations that are very far from the intent of the designers, who may be long gone. Conflicts over these issues in fact often begin as soon as a rule is laid down, and provide a very important (though frequently overlooked) mechanism by which a rule is both clarified (‘worked out’) but also modified in practice and over time. This is the point of the example of the Günstigkeitsprinzip, a classic example of a broader class of struggles over what goals institutions serve and how rules should be interpreted — struggles that therefore become important avenues through which these institutions and rules themselves evolve over time.

The general point, however, is that whatever the source of these gaps (cognitive limits, compromise or political contestation), it seems clear that these gaps are where much of the action is when it comes to understanding how institutions can change incrementally and over time. They are the site of political contestation over what institutions are and especially what they do.

This is worth emphasizing because some of the most prominent treatments of institutions in both political science and sociology not only do not problematize these relationships, they seem in fact to rule out these kinds of gaps by definition. This is true for some versions of institutionalism that cleave closely to the classic view from the new institutional economics (NIE) that sees institutions as solutions to various kinds of collective action problems. Scholars embracing this view see institutions as having been designed by actors to achieve joint gains through co-operation (Weingast 2002). In the context of these analyses, that is exactly what these institutions do, and — perhaps even more importantly — that is all they do; there are no unintended consequences or if there are they are not typically problematized. In other words, in these kinds of arguments there is no gap between the design of the institution and the behaviour and outcomes under it.

If we then move to the other end of the spectrum, namely to the new institutionalism in sociology, we find an interesting parallel in the work of some of these scholars, despite the fact that they embrace a radically different view of institutions. Far from a voluntarist view of institutions held together by rational actors acting in their own self-interest, ‘new’ institutional sociologists such as Meyer and Rowan conceive of institutions as scripts that individuals enact almost unconsciously (Meyer and Rowan 1991). Here too, however, there seems to be no gap between the institution and the behaviour under it; the enactment of the rule both reflects and reinforces its existence. Or to take a much cited example: the handshake is the institution.

My claim here is that scholars working within frameworks that conflate conceptually the institution and the behaviour under that institution will find it hard to develop an account of change that leaves room for conflict and agency. They will find it difficult to see how change could ever be generated endogenously; instead, change will seem always to require an exogenous shock or shift that causes the old institution to break down and that creates an opening in which something new can take its place. By contrast, by focusing on the political coalitions and ongoing political manoeuvring among what we might think of as institutional ‘rule takers’ (and between them and ‘rule makers’) (Streeck and Thelen 2005b: 13), we open up room for talking about strategy, conflict, and agency as important all the time, and not just in those rare moments when structures break down entirely.


  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Theories of institutions and institutional change
  5. 3. Empirical cases of institutional evolution: contemporary changes in German capitalism
  6. 4. Theoretical approaches to institutional change
  7. Acknowledgements
  8. References

This article is based on the text of the first annual lecture of the British Journal of Industrial Relations, delivered at the London School of Economics in November 2008. I thank the editors of the journal for the privilege of delivering the inaugural lecture. In terms of arguments presented here, I would like to acknowledge, above all, the members of the working group on ‘Institutional Change and Institutional Complementarities in Advanced Capitalism’ that I co-chaired with Wolfgang Streeck at the Max-Planck-Institut für Gesellschaftsforschung from 2004 to 2007. The results of these ongoing conversations with Martin Höpner, Britta Rehder, and Christine Trampusch shaped my thinking in very profound ways and I am indebted to them and to Wolfgang for this.

  • 1

    There are exceptions and in fact their number is growing. See, for example, Culpepper (2005); Greif and Laitin (2004); Zorn et al. (2008); and the essays in Mahoney and Thelen (2010a), to name just a few from a variety of perspectives.

  • 2

    Although see the recent work by Iversen and Soskice, which analyses the origins but also tracks the co-evolution of various arrangements associated with co-ordinated and liberal economies over a long stretch of time (e.g. Iversen and Soskice 2009; also Iversen and Soskice unpublished manuscript).

  • 3

    For a review and discussion of the various criticisms levelled at the VofC framework, see Hanckéet al. (2007).

  • 4

    For a notable exception, however, see Hall (2007).

  • 5

    For a full account, see Thelen (2004: chs 2 and 5).

  • 6

    Collaborative work with Duane Swank and Cathie Jo Martin develops an index to measure levels of solidarism across countries and over time (Swank et al. 2008).

  • 7

    I owe this insight, and especially the graphic representation of it that I adopt in Figure 1, to comments made by Martin Höpner at a working-group meeting discussing my article with Ikuo Kume at the Max Planck Institute in 2006. See also Höpner's own very illuminating development of this idea in Höpner (2007).

  • 8

    While the content and quality of training are heavily regulated (and firms must be deemed competent to take apprentices), the decision on the part of the qualified firms to train or not is entirely voluntary.

  • 9

    New Zealand may be the most interesting exception (see, e.g., Aberbach and Christensen 2001; Goldfinch 1998).

  • 10

    A major focus of the project with Swank and Martin is precisely to understand the conditions under which high levels of solidarity are sustainable.

  • 11

    With a few possible exceptions, for example, Britain in the 1960s perhaps.


  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Theories of institutions and institutional change
  5. 3. Empirical cases of institutional evolution: contemporary changes in German capitalism
  6. 4. Theoretical approaches to institutional change
  7. Acknowledgements
  8. References
  • Aberbach, J. D. and Christensen, T. (2001). ‘Radical reform in New Zealand: crisis, windows of opportunity, and rational actors. Public Administration, 79 (2): 40322.
  • Acemoglu, D. and Pischke, J.-S. (1999). ‘Beyond Becker: training in imperfect labour markets. Economic Journal, 109: F11242.
  • Carlin, W. and Soskice, D. (2009). ‘German economic performance: disentangling the role of supply-side reforms, macroeconomic policy and coordinated economy institutions. Socio-economic Review, 7 (1): 67100.
  • Culpepper, P. D. (2005). ‘Institutional change in contemporary capitalism: coordinated financial systems since 1990. World Politics, 57: 17399.
  • Culpepper, P. D. and Thelen, K. (2007). ‘Institutions and collective actors in the provision of training: historical and cross-national comparisons’. In K.-U.Mayer and H.Solga (eds.), Skill Formation: Interdisciplinary and Cross-National Perspectives. New York: Cambridge University Press, pp. 2149.
  • Deeg, R. and Jackson, G. (2007). ‘Toward a more dynamic theory of capitalist variety. Socio-economic Review, 5: 14979.
  • Deeg, R. and Lütz, S. (2000). ‘Internationalization and financial federalism: the United States and Germany at the crossroads? Comparative Political Studies, 33 (3): 374405.
  • Djelic, M.-L. and Quack, S. (2003). ‘Conclusion: globalization as a double process of institutional change and institution building’. In M.-L.Djelic and S.Quack (eds.), Globalization and Institutions. Cheltenham: Edward Elgar, pp. 30233.
  • Dribbusch, H. (2004). ‘Major protests against cuts in unemployment assistance’. EIROnline 2004/09. (accessed 31 May 2009).
  • Eichhorst, W. and Kaiser, L. C. (2006). ‘The German labor market: still adjusting badly?’. Discussion Paper No. 2215. Bonn: Forschungsinstitut zur Zukunft der Arbeit.
  • Elster, J. (2006). ‘Beyond rational self-interest: authors and actors in French constitution making’. In I.Shapiro, S.Skowronek and D.Galvin (eds.), Rethinking Political Institutions: The Art of the State. New York: New York University Press, pp. 26082.
  • Finegold, D. and Soskice, D. (1988). ‘The failure of training in Britain: analysis and prescription. Oxford Review of Economic Policy, 4 (3): 2153.
  • Flecker, J. and Schulten, T. (1999). ‘The end of institutional stability: what future for the “German model”?. Economic and Industrial Democracy, 20: 81115.
  • Glyn, A. (2006). Capitalism Unleashed. Oxford: Oxford University Press.
  • Goldfinch, S. (1998). ‘Remaking New Zealand's economic policy: institutional elites as radical innovators 1984–1993. Governance, 11 (2): 177207.
  • Goldthorpe, J. H. (1984). ‘The end of convergence: corporatist and dualist tendencies in modern western societies’. In J. H.Goldthorpe (ed.), Order and Conflict in Contemporary Capitalism. Oxford: Clarendon Press, pp. 31543.
  • Goodin, R. (2003). ‘Choose your capitalism? Comparative European Politics, 1 (2): 20313.
  • Goyer, M. (2007). ‘Capital mobility, varieties of institutional investors and the transforming stability of corporate governance in France and Germany’. In B.Hancké, M.Rhodes and M.Thatcher (eds.), Beyond Varieties of Capitalism: Conflict, Contradictions and Complementarities in the European Economy. Oxford: Oxford University Press, pp. 195220.
  • Greif, A. and Laitin, D. (2004). ‘A theory of endogenous institutional change. American Political Science Review, 98 (4): 63352.
  • Hacker, J. (2005). ‘Policy drift: the hidden politics of US welfare state retrenchment’. In W.Streeck and K.Thelen (eds.), Beyond Continuity: Institutional Change in Advanced Political Economies. Oxford: Oxford University Press, pp. 4082.
  • Hall, P. A. (2007). ‘The evolution of varieties of capitalism in Europe’. In B.Hancké, M.Rhodes and M.Thatcher (eds.), Beyond Varieties of Capitalism. Oxford: Oxford University Press, pp. 3985.
  • Hall, P. A. and Gingerich, D. (2004). Spielarten des Kapitalismus und institutionelle Komplementaritäten in der Makroökonomie: eine empirische Analyse. Berliner Journal für Soziologie, Winter: 532.
  • Hall, P. A. and Soskice, D., eds. (2001). Varieties of Capitalism: The Institutional Foundations of Comparative Advantage. New York: Oxford University Press.
  • Hall, P. A. and Soskice, D. (2003). ‘Varieties of capitalism and institutional change: a response to three critics. Comparative European Politics, 2 (1): 24150.
  • Hall, P. A. and Taylor, R. C. R. (1996). ‘Political science and the three institutionalisms. Political Studies, 44: 93657.
  • Hancké, B., Rhodes, M. and Thatcher, M. (2007). ‘Introduction: beyond varieties of capitalism’. In B.Hancké, M.Rhodes and M.Thatcher (eds.), Beyond Varieties of Capitalism. Oxford: Oxford University Press, pp. 338.
  • Hassel, A. (1999). ‘The erosion of the German system of industrial relations. British Journal of Industrial Relations, 37 (3): 484505.
  • Hassel, A. (2002). ‘The erosion continues: reply. British Journal of Industrial Relations, 40 (2): 30917.
  • Häusermann, S. (forthcoming). Modernization in Hard Times: The Politics of Welfare State Reform in Continental Europe. New York: Cambridge University Press.
  • Hay, C. (2005). ‘Two can play at that game . . . or can they?’ In D.Coates (ed.), Varieties of Capitalism, Varieties of Approaches. New York: Palgrave, pp. 10621.
  • Höpner, M. (2007). ‘Coordination and Organization: The Two Dimensions of Nonliberal Capitalism’. MPIfG Discussion Paper. Cologne: Max Planck Institut für Gesellschaftsforschung.
  • Howell, C. (2003). ‘Varieties of capitalism: and then there was one? Comparative Politics, 36 (1): 10324.
  • Iversen, T. and Soskice, D. (2009). ‘Distribution and redistribution: the shadow from the nineteenth century. World Politics, 61 (3): 43886.
  • Kapstein, E. (1996). ‘Workers and the world economy. Foreign Affairs, 75 (3): 1637.
  • Katz, H. and Darbishire, O. (1999). Converging Divergences. Ithaca, NY: Cornell University Press.
  • Katznelson, I. (2003). ‘Periodization and preferences: reflections on purposive action in comparative historical social science’. In J.Mahoney and D.Rueschemeyer (eds.), Comparative Historical Analysis in the Social Sciences. New York: Cambridge University Press, pp. 270302.
  • Kenworthy, L. (2001). ‘Wage setting measures: a survey and assessment. World Politics, 54: 5798.
  • Kitschelt, H., Lange, P., Marks, G. and Stephens J. D. (eds.) (1999a). Continuity and Change in Contemporary Capitalism. New York: Cambridge University Press.
  • Kitschelt, H., Lange, P., Marks, G. and Stephens J. D. (eds.) (1999b). ‘Convergence and divergence in advanced capitalist democracies’. In H.Kitschelt et al. (eds.), Continuity and Change in Contemporary Capitalism. New York: Cambridge University Press, pp. 42760.
  • Klikauer, T. (2002). ‘Stability in Germany's industrial relations: a critique of Hassel's erosion thesis. British Journal of Industrial Relations, 40 (2): 295308.
  • Mahnkopf, B. and Altvater, E. (1995). ‘Transmission belts of transnational relations? Unions and collective bargaining in the context of European integration. European Journal of Industrial Relations, 1 (1): 10117.
  • Mahoney, J. (2000). ‘Path dependence in historical sociology. Theory and Society, 29: 50748.
  • Mahoney, J. and Thelen, K. (eds.) (2010a). Explaining Institutional Change: Ambiguity, Agency, and Power. New York: Cambridge University Press.
  • Mahoney, J. and Thelen, K. (eds.) (2010b). ‘A theory of gradual institutional change’. In J.Mahoney and Thelen, K. (eds.), Explaining Institutional Change: Ambiguity, Agency, and Power. New York: Cambridge University Press.
  • Mares, I. (2000). ‘Strategic alliances and social policy reform: unemployment insurance in comparative perspective. Politics & Society, 28 (2): 22344.
  • Martin, A. and Ross, G. (eds.) (1999). The Brave New World of European Labour: Comparing Trade Unions Responses to the New European Economy. Oxford: Berghahn.
  • Martin, C. J. and Swank, D. (2008). ‘The political origins of coordinated capitalism. American Political Science Review, 102 (2): 18198.
  • Martin, C. J. and Thelen, K. (2007). ‘The state and coordinated capitalism: contributions of the public sector to social solidarity in postindustrial societies. World Politics, 60: 136.
  • Meyer, J. W. and Rowan, B. (1991). ‘Institutionalized organizations: formal structure as myth and ceremony’. In W. W.Powell and P.DiMaggio (eds.), The New Institutionalism in Organizational Analysis. Chicago: University of Chicago Press, pp. 4162.
  • Palier, B. (2005). ‘Tracing the political processes of path-breaking changes in French social policy’. In W.Streeck and K.Thelen (eds.), Continuity and Discontinuity in Institutional Analysis, Oxford: Oxford University Press, pp. 12744.
  • Palier, B. and Martin, C. (2007). ‘Reforming the Bismarckian welfare systems. Social Policy & Administration, 41 (6): 53554.
  • Palier, B. and Thelen, K. (2008). ‘Dualizing CMEs: Flexibility and Change in Coordinated Market Economies’. Paper read at Council for European Studies, 6–8 March 2008, at Chicago, Illinois.
  • Pierson, P. (1994). Dismantling the Welfare State? Reagan, Thatcher, and the Politics of Retrenchment. Cambridge: Cambridge University Press.
  • Pierson, P. (2000). ‘Increasing returns, path dependence, and the study of politics. American Political Science Review, 94 (2): 25168.
  • Pierson, P. (2004). Politics in Time: History, Institutions, and Political Analysis. Princeton, NJ: Princeton University Press.
  • Pontusson, J. (2005). Inequality and Prosperity: Social Europe vs. Liberal America. New York: New Century Foundation.
  • Rehder, B. (2003). Betriebliche Bündnisse für Arbeit in Deutschland: Mitbestimmung und Flächentarif im Wandel. Frankfurt/Main: Campus.
  • Rehder, B. (2006). ‘Recht und Politik beim Wandel des Flächentarifs. Politische Vierteljahresschrift, 47 (2): 16992.
  • Rehder, B. and Hassel, A. (2001). ‘Institutional Change in the Wage Bargaining System: The Role of Big Companies’. MPIfG Working Paper Series 01/9. Cologne.
  • Schickler, E. (2001). Disjointed Pluralism: Institutional Innovation and the Development of the U.S. Congress. Princeton NJ: Princeton University Press.
  • Schmidt, V. (2002). The Futures of European Capitalism. Oxford: Oxford University Press.
  • Siebert, H. (1997). ‘Labor market rigidities: at the root of unemployment in Europe. Journal of Economic Perspectives, 11: 3754.
  • Streeck, W. (1997). ‘Beneficial constraints: on the economic limits of rational voluntarism’. In J. R.Hollingsworth and R.Boyer (eds.), Contemporary Capitalism: The Embeddedness of Institutions. New York: Cambridge University Press, pp. 197219.
  • Streeck, W. (2009). Re-forming Capitalism: Institutional Change in the German Political Economy. Oxford: Oxford University Press.
  • Streeck, W. and Hassel, A. (2004). ‘The crumbling pillars of social partnership. West European Politics, 26 (4): 10324.
  • Streeck, W. and Thelen, K. (eds.) (2005a). Beyond Continuity: Institutional Change in Advanced Political Economies. Oxford: Oxford University Press.
  • Streeck, W. and Thelen, K. (eds.) (2005b). ‘Introduction: institutional change in advanced political economies’. In W.Streeck and K.Thelen (eds.), Beyond Continuity: Institutional Change in Advanced Political Economies. Oxford: Oxford University Press, pp. 139.
  • Swank, D., Martin, C. J. and Thelen, K. (2008). ‘Institutional Change and the Politics of Social Solidarity in Advanced Industrial Societies’. Paper read at APSA, 2008, at Boston, MA.
  • Swenson, P. (1991). ‘Bringing capital back in, or social democracy reconsidered: employer power, cross-class alliances, and centralization of industrial relations in Denmark and Sweden. World Politics, 43 (4): 51344.
  • Swenson, P. (2002). Capitalists against Markets. New York: Oxford University Press.
  • Swidler, A. (1986). ‘Culture in action: symbols and strategies. American Sociological Review, 51: 27386.
  • Thelen, K. (2000). ‘Why German employers cannot bring themselves to dismantle the German model’. In T.Iversen, J.Pontusson and D.Soskice (eds.), Unions, Employers and Central Banks: Macroeconomic Coordination and Institutional Change in Social Market Economies. New York: Cambridge University Press, pp. 13871.
  • Thelen, K. (2001). ‘Varieties of labor politics in the developed democracies’. In P. A.Hall and D.Soskice (eds.), Varieties of Capitalism. New York: Oxford University Press, pp. 71103.
  • Thelen, K. (2004). How Institutions Evolve: The Political Economy of Skills in Germany, Britain, the United States and Japan. New York: Cambridge University Press.
  • Thelen, K. (2007). ‘Contemporary challenges to the German vocational training system. Regulation and Governance, 1 (3): 24760.
  • Thelen, K. and Busemeyer, M. R. (2008). ‘From Collectivism towards Segmentalism: Institutional Change in German Vocational Training’. MPIfG Discussion Paper 08/13, Cologne.
  • Thelen, K. and Kume, I. (2006). ‘Coordination as a political problem in coordinated market economies. Governance, 19 (1): 1142.
  • Thelen, K. and Van Wijnbergen, C. (2003). ‘The paradox of globalization: labor relations in Germany and beyond. Comparative Political Studies, 36 (8): 85980.
  • Trampusch, C. (2005). ‘Sozialpolitik in Post-Hartz Germany’. In MPIfG Working Paper. Cologne.
  • Wagner, K. (1997). ‘Costs and Other Challenges on the German Apprenticeship System’. Paper read at Workshop on ‘Skills for the 21st Century’, January 10, at Institute for Contemporary German Studies; Johns Hopkins University, Washington DC.
  • Wagner, K. (1999). ‘The German apprenticeship system under strain’. In P. D.Culpepper and D.Finegold (eds.), The German Skills Machine: Sustaining Comparative Advantage in a Global Economy. New York: Berghahn Books, pp. 3775.
  • Wallerstein, M. and Golden, M. (2000). ‘Postwar wage setting in the Nordic countries’. In T.Iversen, J.Pontusson and D.Soskice (eds.), Unions, Employers, and Central Banks: Macroeconomic Coordination and Institutional Change in Social Market Economies. New York: Cambridge University Press, pp. 10737.
  • Wallerstein, M., Golden, M. and Lange, P. (1997). ‘Unions, employers’ associations, and wage-setting institutions in northern and central Europe, 1950–1992. Industrial and Labor Relations Review, 50 (3): 379402.
  • Weingast, B. (2002). ‘Rational-choice institutionalism’. In I.Katznelson and H.Milner (eds.), Political Science: The State of the Discipline. New York: Norton, pp. 66092.
  • Zorn, D., Dobbin, F., Dierkes, J. and Kwok, M. (2008). ‘The new firm: power and sense-making in the construction of shareholder value. Nordiske Organisationsstudier, 3: 4168.