Models of Capitalism COLIN CROUCH. New Political Economy, Vol. 10, No. 4, December 2005

Transcription

1 New Political Economy, Vol. 10, No. 4, December 2005 Models of Capitalism COLIN CROUCH That capitalist economies might take diverse forms has been long recognised by some scholars. Sometimes this diversity has been seen as a matter of evolutionary development. This was true of Max Weber s ideal-type approach, that of the advocates of postwar modernisation theory, and of those who followed Antonio Gramsci s identification of a Fordist phase of capitalism that was deemed to succeed the classic free-market form. This last idea flourished particularly in the French régulationiste school. 1 These approaches, different from each other though they are, all see some forms of capitalism superseding, and as therefore in some sense superior to, earlier modes. Hence these are not theories of a true diversity in the sense of a continuing multiplicity of forms, the historical superiority of any of which might never come to an issue. Analysts willing to adopt a less historicist approach have been rarer. The modern locus classicus was Andrew Shonfield s work, 2 which examined the role of various institutions surrounding the economy various branches of the state, banks, stock exchanges in a number of Western European countries, the United States and Japan. Although he thought some were more efficient than others in particular, he was impressed by those that inserted some elements of planning into otherwise free markets he did not talk in terms of historical transcendence. When more theoretically inclined political scientists and sociologists returned to considering economic questions in the 1980s, they resumed Shonfield s concern with national politico-economic systems and hence national varieties of capitalism. Occasionally sub-types would be recognised within a national economy (mainly with regard to Italy and Spain), but these sub-types have nearly always been geographically subdivided, so the concept of territorially based economies has been retained. This does not mean that each nation-state has been seen as embodying its own unique form of capitalism; rather, national cases are grouped together under a small number of contrasted types. This literature has many achievements. It has provided an intellectual counterweight to easy arguments about globalisation, which predict an inevitable trend towards similarity among the world s economies. Neoinstitutionalist accounts of diversity have provided both theoretical arguments and some empirical demonstrations to suggest that these may be great oversimplifications. However, if we are to model the diversity of economic institutions more scientifically, and particularly if we are to study institutional change and innovation, we need to deconstruct Colin Crouch, Institute of Governance and Public Management, Warwick Business School, University of Warwick, Coventry CV4 7AL, UK. ISSN print; ISSN online=05= # 2005 Taylor & Francis DOI: =

2 Colin Crouch the wholes that contemporary institutionalism takes for granted and discover their constituent elements elements which are able to survive in combinations other than those thus identified. Acceptance of the value of taking this approach would have considerable implications for the future study of capitalist diversity. It has in particular a major methodological consequence: empirical cases should be studied, not to determine to which (singular) of a number of theoretical types they should each be allocated, but to determine which (plural) of these forms are to be found within them, in roughly what proportions, and with what change over time. This alternative is less ambitious than the current fashion, in that it does not enable us to map the economic world with a few parsimonious categories. But it is also more ambitious, partly because it corresponds more closely to the requirements of scientific analysis, but also because it is able to accommodate and account for change taking place within empirical cases. This is something which most of the neoinstitutionalist literature on capitalist diversity finds difficult to do, leading to the functionalism and determinism of much of its analysis. The aim of the present article is to develop this critique of the existing literature, to highlight some promising recent trends and to point towards the new approach indicated above. This last, which involves first deconstructing into constituent elements and then being ready to recombine into new shapes the aggregated forms of currently dominant analyses, is developed more fully elsewhere. 3 Pitfalls in the formulation of types The smallest number of theoretical types consistent with the idea of diversity is two. For almost all writers on models of capitalism, one is always the freemarket model of neoclassical economics. This constitutes the principal intellectual antagonist for neoinstitutionalists, even when they argue that it accounts for only a highly specific form of capitalism. 4 There must be at least one other form to make a theory of diversity: hence dichotomies. At the other extreme there is no theoretical limit to the number of forms that might be identified, but theories rarely propose more than five or six. Given the relatively small number of empirical cases of advanced capitalism for those tied to a national case approach (currently around 25), it is difficult to sustain more than a handful of types without lapsing into empiricism. The work of Michel Albert, who made the original contribution to dualistic analysis, is typical. 5 He modelled two types of capitalism, which were seen in an antagonistic relationship. They are labelled in geocultural terms as Anglo-Saxon and rhénan (Rhenish). The former defines free-market capitalism, considered to be embodied in the Anglophone countries. 6 The second takes its name from certain characteristics considered to be common to the riparian countries of the Rhine: Germany, the Netherlands, Switzerland, more problematically France. However, not only is the author uncertain whether France s institutions fully belong to this type (an anxiety which was one of his main motives in writing the book), but Japan and Scandinavia are considered to be part of it. The broad institutional range gathered together to form this second type is disconcerting. The essential idea is a capacity to make long-term decisions 440

3 Models of Capitalism that maximise certain collective rather than individual goods. But this means ignoring differences among the very diverse forms of collectivism found. It is important to note that this dualism in the identification of types of economy parallels the debate between political philosophies neoliberalism and social democracy which lies behind the analysis and behind most contemporary political debate. 7 This has created some confusion over whether neoinstitutionalism s confrontation is with neoclassical economics, and therefore at the analytical level only; or with neoliberal politics, implying an ideological confrontation; or with all political practices associated with the anti-keynesian and pro-capitalist forces which came to prominence during the period. One form taken by both the scientific and the ideological debate has been dispute over which kind of capitalism delivers the best economic performance. As David Coates showed in his study of models of capitalism, this has been an extraordinarily difficult issue to resolve. 8 He unravelled the complexities of the components of economic growth and other indicators of performance, in particular pointing out the importance for comparative studies of where national cases have stood at particular moments in relation to the overall evolution of the world capitalist system. He showed how it had been a mistake for institutionalists to seize at various times on particular national examples as proving the superiority of economies not based on pure markets: the models selected had a tendency to start to underperform. Analysts have been on stronger grounds when making either a weaker or a different claim. The former is that various kinds of institutional economy can do just as well as (not necessarily better than) a pure market one; the latter is the argument that institutional economies enabled the coexistence of high levels of economic performance alongside the pursuit of certain other social goals (for example, a relatively egalitarian incomes distribution) not readily available to purer market economies. Neoclassical analysis considers how economic actors would behave if a world of perfect markets existed. It usually but not necessarily incorporates the normative assumption that both economy and society would be improved were institutions to take this form, but neoclassical economists are at liberty to consider that this may not always constitute a practical proposition; they are not bound by their analytical approach to any particular policy conclusions, or to consider that the world in reality takes a certain form. It is neoliberalism which, as a political creed rather than a form of analysis, not only definitely adopts a positive normative evaluation of markets, but also believes that they could always be introduced in practice. But in practice not even neoliberals do this. A by-product of the ideological dominance of neoliberalism since the 1980s, and in particular its association with the most powerful nation-state on earth the United States has produced a tendency among even serious analysts to assume that certain practices and institutions constitute part of the neoliberal paradigm just because they are found in the US. The characteristics of the neoliberal model are derived from empirical observation of what is thought to be its main empirical example. But it is logically impossible to derive the characteristics of a theoretical category from the characteristics of an example of it, as the theoretical characteristics have to be known before a case can be considered to be such an example. For example, an extremely 441

4 Colin Crouch powerful, scientifically oriented military sector, tying a number of contracting firms into close and necessarily secretive relations with central government departments, is a fundamental attribute of the US economy, and central to much of its innovative capacity in such sectors as aerospace and computing. The operation of such a military sector has nothing to do with the principles of either neoclassical economics or neoliberal politics. Analysts respond to this in two ways. Some just ignore the existence of this sector and its special characteristics in their account of the US economy. For example, the Organisation for Economic Cooperation and Development (OECD) felt able to describe the US as a country lacking any close support from government for industry. 9 Alternatively, it argues that the defence sector is somehow part of the US liberal model, without noting the difficulties of such an assumption. 10 Indeed, as Campbell and Pedersen argue, at the practical level neoliberalism has not been the monolith that both its advocates and opponents set it up to be. 11 Within it have been contained a diversity of practices, some not particularly coherent with others. Kjær and Pedersen point, for example, to clear differences from the normally presented model in the form taken by so-called neoliberalism in Denmark, 12 whilst King and Wood have even demonstrated significant distinctions between the neoliberalisms of the United Kingdom and the United States in the 1980s, two cases normally seen as joint paradigms. 13 The collection of studies edited by Peter Hall and David Soskice under the name Varieties of Capitalism represents the most ambitious and significant contribution to date of the dualist approach. 14 It draws much from Albert, though it barely acknowledges his contribution. Their book has become the emblematic citation for all studies of diversity in capitalist economies. It is also an example of the preoccupation of many neoinstitutionalists with coming to terms with and, in this case, eventually becoming absorbed by, an idealised version of neoliberalism. It seeks not only to allocate every developed capitalist economy to one or other of two categories, but derives from this account a theory of comparative advantage and a list of the kind of products in which the country will specialise. 15 This is achieved with the aid of certain assumptions concerning what constitutes radical and what incremental innovation a characteristic which is considered to differentiate whole classes of goods and services. It is this factor, combined with its use of this sectoral analysis to account for certain important developments in different national economies during the 1990s, which has made the account so appealing. Despite some ambiguity about a possible third model, these authors work with an essentially dualist approach along the rationale outlined above. They specify, first, a liberal market economy (LME) identified with neoliberal policies, radical innovation, new sectors of the economy and the Anglophone countries (Australia, Canada, Ireland, New Zealand, the United Kingdom, but primarily the US). Germany is at the centre of a second type, called a coordinated market economy (CME), where social and political institutions engage directly in shaping economic action. This form is linked to social democracy, incremental innovation, declining economic sectors and non-anglophone countries. It is odd that the core linguistic uniting characteristic of the LMEs, the only generalisation that really works, is never actually discussed as such. More aware of 442

5 Models of Capitalism Irish sensitivities than most authors, Hall and Soskice always talk of Anglo-Saxon and Irish economies. But, perhaps because like others they resist the far simpler and more accurate Anglophone, they miss some serious potential implications of this. For example, one of the most impressive pieces of evidence cited by them to support their contention that radical innovation is concentrated in LME countries and only incremental innovation in CMEs is work carried out for them on patent citations. 16 This reveals a strong statistical tendency for patents taken out in Anglophone countries to cite scientific sources, while those taken out in continental Europe and Japan tend to cite previous patents or non-scientific sources. The six leading countries out of 18 studied are all Anglophone (headed by Ireland). Prime facie distinction between radical and incremental innovation does seem to be well proxied by that between academic and product citations, and one can see this being related to the character of research in firms, research centres and universities. But it is also possible that firms in Anglophone countries are more likely to cite articles in the overwhelmingly Anglophone literature of global science than those in other countries. Further, liberal market economies are largely defined by their having characteristics determined by common law traditions; these also encourage the use of patenting of innovations to a greater extent than civil law systems. Therefore higher levels of patenting as a legal device, not necessary a reflection of actual innovation will be most widespread in common-law, and hence liberal market, systems. This distortion may help explain why, according to Estevez- Abe, Iversen and Soskice, New Zealand has more radical technological innovative capacity than Germany, Sweden or Switzerland. 17 The LME type of economy depends on labour markets that set wages through pure competition and permit very little regulation to protect employees from insecurity, and on a primary role for stock markets and the maximisation of shareholder value in achieving economic goals. Such an economy is considered by the authors to be poor at making minor adaptive innovations, because employers make inadequate investment in employee skills which might produce such innovations; but it excels at radical innovations, because the combination of free labour markets and external shareholders makes it relatively easy to switch resources rapidly to new and profitable firms and areas of activity. A CME, featuring corporatist wage-setting, strongly regulated labour markets and corporate financing through long-term commitments by banks, follows exactly the reverse logic. Hall and Soskice stress strongly that they are depicting two enduring forms of capitalism, because each has different comparative advantages. However, those of the CME form are located solely in minor adaptations within traditional and declining industries, while LMEs have assigned to them all future-oriented industries and services sectors. In the end, therefore, this is a neoinstitutionalism that fully accepts the logic of neoclassicism set out above: in the long run, all institutions other than the pure market fail to cope with the future. Since these different forms of capitalism are considered to have been the products of historical longues durées, it also means that the German economy never was radically innovative in the past, which requires explaining away many past events in the economic history of such German industries as chemicals, machinery, steel and motor vehicles when these sectors were at the forefront of technological advance. 443

6 Colin Crouch This brings us to a further fundamental point: typologies of this kind are fixed over time; they make no provision for changes in characteristics. As Zeitlin puts it, approaches like that of Hall and Soskice render learning almost impossible. Or, as Bertoldi says, they ignore any impact of change in the world economy and make no allowance for evolutionary development. As Hay has it, this literature tends to take either a spatialising approach (the elaboration of models, as in the cases we are discussing here) or a temporalising one (identifying historical phases, and therefore probably giving more scope to actors capacity to change, but ignoring synchronic diversity). It is not necessary for neoinstitutionalist analysis to be as rigid as this. 18 Hall and Soskice also assume automatically that all innovation within new industries represents radical innovation, while all within old ones can represent only incremental innovation. This is because they use different sectors as proxies for different types of innovation. According to such an approach, when Microsoft launches another mildly changed version of Windows it still represents radical innovation, because information technology is seen as a radical innovation industry; but, when some firms eventually launch the hydrogen-fuelled motor engine, this will only be an incremental innovation, because the motor industry is an old industry. Further, the authors do not confront the leading position of two clearly CMEs (Finland and Sweden) in new telecommunications technologies and the Nordic countries generally in medical technologies. 19 Robert Boyer has shown that the institutional pattern found in the Nordic countries can favour high-technology growth in information and communication technologies as much as the Anglo-American one. This is completely lost in accounts that insist on dualism and an a-priori allocation of institutional patterns. 20 Instead of the a-priori paradigm case methodology, Boyer used Charles Ragin s Booleian techniques to derive institutional patterns empirically. 21 Booleian algebra assigns category values (not interval ones) to the mass of characteristics that constitute a whole. Individual characteristics are identified as either present or absent. This enables a search for shared characteristics in a number of complex empirical cases, assisting the researcher to determine which characteristics tend to be found together, and which are rarely or never associated. A further serious flaw in the Varieties of Capitalism approach is that it misunderstands the work of individual innovative companies. While engaging in radical innovation, firms usually also need to bring out products with minor improvements in order to sustain their position in markets while they wait for a radical innovation to bear fruit; but, according to the Hall-Soskice model, it is not possible for firms within an LME to succeed at incremental innovation. It is a major advance of the approach that they focus on the firm as an actor, rather than take a macroeconomic approach to the study of economic success. However, many of the advantages of this are vitiated by the fact that their model allows the firm virtually no autonomy outside its national macroeconomic context. These authors further follow conventional wisdom in arguing that the superiority of American (or Anglophone) firms over German ones results from the fact that in the Anglophone countries all managerial power is concentrated in the hands of a chief executive officer (CEO) who is required to maximise shareholder value, with employees engaged on a hire-and-fire basis with no representative channels 444

7 Models of Capitalism available to them. Here they are failing to distinguish between the firm as an organisation and as a marketplace. By seeing the CEO s power as being solely to maximise share values by the use of a hire-and-fire approach to management, they are able to present the firm in an LME as solely the latter and not as an organisation. They can therefore dispense with the knowledge accumulated in the theory of the firm, which distinguishes between market and organisation, and presents at least the large firm as an organisation with personnel policies, and with management having a wider range of discretion and possibilities than just maximising share values. This is significant. In reality firms differ considerably in the extent to which they construct organisational systems, internal labour markets and distinctive ways of working, even developing specific corporate cultures, rather than simply establishing themselves as spaces where a number of markets intersect. For example, a firm that develops a distinctive approach to work among its workforce as part of its competitive strategy cannot depend on a hire-and-fire personnel policy. Employees need to be inducted into the firm s approach and are likely to demand some understandings about security if they are to commit themselves in the way that management wants. Rapid hire-and-fire meets neither of these needs. This fundamental difference in corporate strategy has nothing at all to do with differences between LMEs and CMEs; both can exist within either, particularly the former. Neglect of the firm as an organisation is thus a weakness of much neoinstitutionalist analysis. It is caused by the obsession already noted with a dichotomy between two mutually incompatible politico-economic ideologies, a dichotomy in which the distinction between firm and market is not at issue. At times Hall and Soskice seem to regard the organisational structure of the firm (or corporate hierarchy) as a characteristic of both LMEs and CMEs, and therefore an irrelevant variable - though it should be conceded that the relevant passage is worded ambiguously, as follows: All capitalist economies also contain the hierarchies that firms construct to resolve problems that markets do not address adequately... In liberal market economies, these are the institutions on which firms rely to develop the relations on which their core competences depend. 22 They seem here to be building into their model a functionalist balancing item, implying that hierarchy will exist to the extent that it can resolve problems. In that case, why does their theory not build into the features of both LMEs and CMEs those that they would respectively need in order to have them cope with the kinds of innovation that their theory says is impossible for them? At the level of type-building one should not pick and choose which institutional features automatically receive compensation and which do not. As Weber originally formulated the concept, ideal types are one-sided accentuations, pressing home the logical implications of a particular kind of structure. The aim is not to provide an accurate empirical description, but a theoretical category, to be used in the construction of hypotheses. Again, the authors are not building their theory deductively, but are reading back empirical detail from what they want 445

8 Colin Crouch to be their paradigm case of an LME the US into their formulation of the type. It is simply not possible within their methodological approach to ask the question: is everything important that occurs in the US economy the embodiment of free markets? Hall and Soskice do briefly consider diversity within the CME form. Apart from Germany, they also see Japan, Switzerland, the Netherlands, Belgium, Sweden, Norway, Denmark, Finland and Austria as unproblematic though differences between what they call industry-based coordination of the German type and group-based coordination found in Japan and Korea are recognised. 23 In an earlier work Soskice fully recognised these two distinct forms of CME: a northern European model, and the group-co-ordinated East Asian economies. 24 ( Northern Europe is here defined by Soskice to include Italy but not France.) But not much is made of the distinction in the full development of theory or cases. A Mediterranean group (France, Italy, Spain, Portugal, Greece and Turkey) is also given some recognition. Like Albert before them, Hall and Soskice accept that France is somehow different, and consider that a so-called southern European group (including France) probably constitute a third, state-led, post-agrarian model. 25 This at least makes matters more differentiated, although it produces a type curiously unable to distinguish between the French state and the Italian or Greek ones. Sometimes this Mediterranean group is seen as being empirically poised somewhere between the LME and CME model, which enables the authors to insist that LME and CME remain the only points which require theoretical definition. But elsewhere the Mediterranean countries are treated as examples of CMEs; Thelen, for example, treats Italy as almost unambiguously a German-type co-ordinated economy. 26 One of the starting points of the model was an earlier paper by Soskice criticising the Calmfors and Driffill model of wage bargaining. 27 This model had contrasted economies with centralised and decentralised collective bargaining arrangements, classing the French, Italian and Japanese among the latter. Soskice pointed out that, although these three countries were not as coordinated as Germany or Sweden, one could identify within them various mechanisms that ensured some coordination of wage bargaining. He found (within the sample of countries being considered) that only the UK and the US lacked such mechanisms; therefore all other cases were classified as CMEs. Both here and in Hall and Soskice, the basic drive of the dichotomy is to confront the neoclassical model with a single rival type. Beyond dichotomies Some contributors to the study of capitalist diversity have gone beyond dichotomies. Vivien Schmidt has three models of European capitalism: market (very similar to the LME model), managed (with an enabling state that encourages economic actors to cooperate, more or less the CME model) and state (an interventionist state of the French kind). 28 The last is designed to remedy the neglect of this form by Hall and Soskice. Acknowledging that the role of the state has declined considerably in France in recent years, she points out that its background role and historical legacy remain of considerable importance in enhancing national economic capacity. But, indeed, much the same could be said of the US state, 446

9 Models of Capitalism whose role in the vast defence-related sector could well be defined as state enhancement of economic capacity. Schmidt also manages to be sensitive both to change and to its timing. 29 She studies how countries embodying each of these types respond to the challenges of globalisation and Europeanisation. A central hypothesis is that these challenges do not lead to simple convergence. Governments of the various countries have responded in complex ways, producing new forms of diversity. If there is any overall convergence, it is mainly towards a loss of extreme characteristics and thus some sharing of attributes from the various models. And these diversities are full of interesting paradoxes: the UK, having had in many respects the weakest economy of the three, was thus the earliest to be forced to come to terms with the pressures of globalisation. As a result, it now appears better prepared to face that challenge than Germany which, being initially the strongest economically, could delay adjustment. A second hypothesis fundamental to her study is that political discourse has been particularly important in shaping national responses to the challenge. By this Schmidt means, not just that different substantive discourses were adopted, but that these took different forms. She distinguishes between communicative and coordinating discourse forms. The former, more suited to centralised systems like the British and French, inform the public of what needs to be done; the latter, more typical of Germany, is used to develop consensus among powerful actors who cannot be controlled from the centre. This work therefore marks a refreshing shift towards an actor-centred and non-determinist account. Schmidt by no means discounts the existence of very strong structures, within which her actors need to operate. But these are malleable by innovative actors, in particular by politics. She criticises particularly effectively the oversimplified accounts that characterise much rational choice work in international political economy. This, she argues, is a curiously depoliticised form of study of politics, assuming as it does that the interests of nation-states can be modelled in a straightforward way, with fixed, consciously held preferences. She demonstrates effectively how governments in the three countries of concern to her study developed very varied positions in relation to Europeanisation: for example, the UK was quickest to respond to many of the single market initiatives, but slowest to the single currency. This can all be explained, and she provides good explanations, but these require tactical and historically contingent political actors. But Schmidt still follows the practice of identifying empirical cases as standing for ideal types. This is unfortunate, because her own actual practice is well able to cope with the implications of seeing cases as amalgams of types: her actors are creative political schemers, looking for chances to change and innovate, not automata acting out the parts the theorist has set for them. And, as noted, she succeeds in showing how over time individual countries have moved around the triangular space which her particular model of types of capitalism allows them. Several other authors present three or more forms of capitalism, or of elements of capitalism, nearly always retaining a geocultural approach. Gøsta Esping- Andersen s analysis of different types of welfare state embodies variables relating to the outcomes of political struggle, or dominant political traditions, which avoids some of the functionalist implications of the Varieties of Capitalism model. 30 Again, one starting point is free-market or liberal capitalism associated with the 447

10 Colin Crouch Anglophone group of countries, and another is Germany, producing a conservative continental European model. There is, however, a third, social-democratic pole, geographically associated with Scandinavia. Critics of Esping-Andersen s model have concentrated: on identifying mixed cases (Castles and Mitchell); on stressing how the treatment of women in different systems does not seem to correspond to the simple typology (Daly); or on breaking up the over-extended conservative continental category. A fourth type has now been clearly established, separating southern European welfare states from this one on the basis of their particularly large role for the family (Naldini) and other informal institutions (Ferrera). Ebbinghaus, concentrating on policies for combatting early exit from the labour market, which he sees as deeply related to the form of the overall welfare regime, adds a fifth type based on Japan. All these works continue to depend on the characteristics of paradigm cases, which can be highly misleading. For example, Viebrock, in a study of different forms of unemployment benefit systems, has shown how Sweden usually the absolute paradigm case of social democracy has for reasons of political history retained a role for voluntary associations alongside the state in the organisation of its unemployment insurance system. 31 A strong move away from dualism, which neither starts from nor privileges the free-market model, is the scheme of Richard Whitley. 32 He builds up a set of fully sociological models of capitalism based on six types of business system (fragmented, coordinated industrial district, compartmentalised, state-organised, collaborative, and highly coordinated), related to a number of different behavioural characteristics. 33 He also presents five different ideal types of firms (opportunist, artisan, isolated hierarchy, collaborative hierarchy, allied hierarchy) 34 and a diversity of links between these types and certain fundamental institutional contexts (the state financial system, skill development and control, trust and authority relations). 35 Significantly, Whitley s main fields of study are Japan, Korea, Taiwan and other Far Eastern economies, rather than either the American or the German cases, and he is therefore further removed from the obsession with neoliberalism and a contrast between it and a model of organised capitalism that sometimes distorts the analysis of those who concentrate on Western Europe and North America. By far the best and most sophisticated approach to a post-dualist typology of capitalism to date is that established by Bruno Amable. 36 He collected quantitative data on a vast range of characteristics of the national economies of most OECD countries: product markets, labour markets, financial systems and social protection. He uses literally dozens of individual indicators to assess each. He then allows a typology of groups of countries to be formed empirically by these data; he does not start from paradigm cases. This procedure gives him five groups, which, as with other authors, fall into familiar geocultural patterns: market-based (primarily Anglophone), social democratic (Nordic), Asian (Japan and Korea), Mediterranean (southern European) and Continental European (continental Western European less the Nordic and Mediterranean countries). He further finds (as have others 37 ) that this last group does not show much internal coherence, and for some purposes splits it further into two sub-groups: one comprising the Netherlands and Switzerland, the other Austria, Belgium, France and 448

11 Models of Capitalism Germany. Moreover, Amable is not afraid to draw attention to further diversity for some of the characteristics, with the result that countries do not always figure within their normal group. At times Amable lapses from his finely nuanced stance. For example, the book ends with a future- and policy-oriented dialectic between the market-based and a simplified and generalised Continental European model. 38 It seems that engaging in the rhetoric of debate about the future course of capitalism leads always to dualism, even when, as in Amable s case, the best strength of the author s position lies precisely in the demonstration of a far more differentiated world. He also depends necessarily for his data on sources like the OECD which are often constructed with in-built biases. For example, although at one point Amable acknowledges the importance of military-related research and production in many of the high-tech sectors of the US economy, he follows the OECD in excluding all consideration of this from the indicators of the role of the state in the economy and of the regulation of external trade. 39 These minor criticisms apart, Amable has demonstrated that a genuinely scientific approach, using very extensive and diverse kinds of data, produces a useful and coherent typology comprising five or six types, at the same time enabling clear recognition of exceptions within types. Dichotomisers will argue that they are applying the principle of parsimony and Occam s razor to complex schemes of Amable s or Whitley s kind. They will claim that, while there is clearly a loss of information if one collapses Whitley s co-ordinated industrial district, compartmentalised, state-organised, collaborative, and highly co-ordinated mechanisms into the single idea of a CME, that idea seizes on the essential point that divides all these forms from the pure market one: coordination. But, as Scott and Hay have separately argued, parsimony must not become an excuse for inaccuracy and ignoring important diversity. 40 Is coordination the fundamental attribute of all the types in Whitley s list? On what grounds could this quality be regarded as more fundamental than the other characteristics which divide then, especially since the coordination takes place at very different levels? Recent developments in the governance approach draw attention to the role of collective competition goods provided by various governance modes in local economies, without demonstrating anything remotely strong enough to be called national coordination. 41 This suggests the possibility of analyses more moderate than those addressed at the whole macroeconomy. Meanwhile, Hage and Alter have convincingly demonstrated analytical distinctions among several institutional forms. 42 In that case, to apply Occam s razor to reduce them all to one idea of coordination is to cut into serious theoretical and empirical flesh. An explanation becomes more parsimonious than another when it uses a smaller number of explanatory variables while explaining at least as much as its opponent. For example, it is more parsimonious to model the solar system as heliocentric than terracentric, because the former uses far simpler mathematics to account for at least as many planetary movements as the latter. We should be far less impressed with the heliocentrist if she had to say: Forget about the outer planets; this theory is more parsimonious because it just looks at the inner ones. But contemporary social science often makes use of precisely 449

12 Colin Crouch this kind of argument, using the idea of parsimony as meaning a kind of rough, tough macho-theory that concentrates on the big picture and ignores detail. As Whitley s formulations demonstrate, the relationships between different forms and different behavioural characteristics present a varied patchwork of similarities and differences, not a set of polar contrasts. This suggests in turn the fundamental point: that individual empirical cases might well comprise more complex amalgams still of elements from two or more theoretical types. Whitley himself treats a fragmented market model of economic organisation separately from one dominated by large firms, and is therefore able to see the US itself as a hybrid of two different forms of capitalism rather than a pure case. This question has considerable practical implications, which are discussed more fully elsewhere. 43 It is often recognised by authors who speak of hybrid forms. For example, Schmidt suggests strongly that some changes in French institutions are making that case increasingly a hybrid, with borrowing from Germany as well as from neoliberal sources. 44 Jackson suggests that hybridisation, as opposed to simple imitation of the exogenous, is the usual outcome of attempts at borrowing institutions, even under extreme periods of transition, such as Germany or Japan under postwar occupation. 45 Other researchers have shown the power of hybrid cases in achieving important reforms in welfare state organisation. 46 Zeitlin discusses various national cases that have become exceptions to their types as the result of mixing institutional forms at the initiative of what I would call institutional entrepreneurs. Considering an earlier period, Windolf discusses how French family capitalism played an important part in the country s postwar modernisation, merging with advanced financial means of control and the strong state to produce a dynamic new model. Hybridisation deals with only one way in which cases may deviate from types, and it is still very close to the idea of clear, macro-level types, because it sees these as the source of the hybridisation. However, it does constitute an important challenge to simple equations of cases and types. 47 Questioning the centrality of the nation-state The centrality of the nation-state in most typologies of capitalist diversity also needs to be questioned. This centrality is found in most neoinstitutionalist studies, including those on social [that is, national] systems of innovation and production. 48 It is also central to work from the parallel but distinct literature on national systems of innovation. 49 At one level the case is well made. Very extensive elements of governance in the industrial and post-industrial societies of which we have knowledge do operate at the level of the nation-state: states have been the main sources of law, and most associations and organisations target themselves at the state. 50 Given that markets are framed by law, this means that, of the modes of governance usually discussed in governance theory, the state itself (obviously), markets and various levels of associations are all heavily defined at national level, while community and informal associations exist at a lower geographical level. Even research that explicitly works at comparisons between regional or other substate geographical levels often has to acknowledge the importance of 450

13 Models of Capitalism the nation-state as a major instance for the determination of socioeconomic variables. 51 But many macro-level neoinstitutionalists go further than this and postulate virtually hermetically sealed national institutions often because they are concerned to address debates about economic and social policy, and these are mainly conducted at national levels. Radice argues that this has perhaps been particularly the case for left-of-centre analysts desiring to bring the state back in, leading to an exaggeration of the importance of national policy. 52 More generally, neoinstitutionalists are led to stress the nation-state by their functionalist assumptions, which model discrete, autonomous systems, each equipped with their sets of institutions, like a body with its organs. There are also methodological advantages in being able to treat nation-states as discrete units of analysis, as many economic data are produced at national levels. Theorists of the diversity of capitalism are therefore eager to play down the implications of globalisation, and argue intelligently and forcefully against the naive assumptions of much other literature that globalisation somehow abolishes the significance of national differences. 53 However, the position of the nation-state as the definer of the boundaries of cases is not so fixed that it should be taken for granted per definitionem. This is particularly obvious with respect to multinational corporations. As Beyer shows, 54 large firms draw on resources from a range of different national bases; it is very difficult to identify them with particular national types and to see their institutional possibilities as being constrained by their country or countries of location. As Jackson puts it, national models of capitalism are becoming institutionally incomplete. 55 This seems particularly true where international corporations are concerned, but even firms that are nationally owned and operate primarily within one nation-state have access to knowledge, links and practices existing outside the national borders. Radice similarly criticises the national innovation system literature for a kind of mercantilism, arguing that it does not take adequate account of the fact that technology is always a public/private collaboration, and that the private actors are usually global firms. Something always leaks abroad from national programmes; innovation is at once global and national. 56 He also points out the falsity of the dichotomy between so-called globalising and national forces, as though one could identify them and then establish their relative importance. 57 The phenomena associated with globalisation are brought about at the behest of domestic actors working to influence national governments. As Helleiner earlier made the point: internationalisation is not an independent variable, because it is an outcome of state policy. 58 Radice demonstrates a different weakness of nation-state-based analysis by pointing out that all states are not equal as units. 59 The US is able to borrow to fund its deficits in a way not available to others, which means that comparing the performance of that economy with others is not a true comparison of institutional capacities. One can move from that observation to point out that nation-states cannot always be treated as a series of unit instances of the same phenomenon; they are also linked together in a hierarchical way to form an overall system, as Wallerstein and other world-system analysts have showed. 60 For example, the units Portugal and France cannot be treated as equal units within which the effects of various independent variables can be independently 451

14 Colin Crouch and comparatively assessed, because they are partly defined by their relationship to each other. Scott stresses the need to consider a range of levels: world system, society (nation-state), organisational field, organisational population, organisation and organisational sub-system. 61 As he points out, different disciplines tend to look at different components of this. Hollingsworth and Boyer are helpfully explicit that their scheme can be used at subnational and transnational, as well as national levels. 62 We need always to be able to ask: are arguments about the characteristics of national economies limited to specific economic sectors and industrial branches, or do they claim to apply to all? And how far beyond the heartland of the economy does the theory claim to range? If the nation-state is at the heart of the analysis, are political institutions also to be covered by the characterisation? Or does the theory apply even further, to structures like the welfare state, family or religion, for example? As we develop thinking of this kind, we soon come to see that the clear division between endogenous and exogenous that is so fundamental to nation-state-based theories becomes replaced by a continuum of accessibility. Towards a new analytical approach: anticipating recombinant capitalism As noted at the outset, most contributions to the literature on the diversity of capitalism conflate theoretical models and empirical cases through a research strategy that seeks the unique theoretical box to which an individual case must be assigned. For example, Goodin et al., while arguing that the US constitutes a pure type of market welfare regime, acknowledge that 80 per cent of US social protection expenditure goes to social insurance schemes of a corporatist nature and not to the means-tested schemes associated with the market model. 63 However, because they consider that this is a smaller proportion than goes to such schemes in other cases, they claim that they are justified in regarding the US as a paradigm of the pure market model. They do not consider the possibility that the corporatist elements of the welfare system might act complementarily to the market process in the US case, and that the US system might operate differently if it really was a pure market one. In fact, the differences that have been identified among neoinstitutionalist theories have major implications for how they relate theoretical models and empirical cases. There are broadly two ways of doing this: the labelling method and the analytical method. The two approaches are analogous to the two different forms of categorisation found on bottles of mineral water: first, the water is labelled as either still or sparkling the water is, unambiguously, one or other of these types; second, there is set out a detailed chemical analysis of elements and compounds, traces of which can be found in the water the water contains these chemicals. The neoinstitutionalist researcher following the labelling strategy inspects the characteristics of an empirical case and decides which of a limited number of theoretical models (ideally two) it most closely resembles. The case is then considered to be an example of that model and labelled accordingly, all features of it which do not fit the model being considered as noise and disregarded. A clear example is again the study by Goodin et al., which takes three national cases as examples of three models, then reads back empirical features of these cases into the models. In defence of such procedures Hollingsworth claims that, 452

15 Models of Capitalism even if an individual society has more than one social system of production, one will dominate. 64 This is possibly true, but not only should it remain an hypothesis worth testing rather than an a-priori methodological assumption, but the role of minor or hidden institutional forms can have major importance. By contrast, the researcher following the analytical approach considers to what extent traces of each of a series of models can be found within the case; there may be no conclusion as to which form it most closely resembles. Even if there is, that information remains framed in the context of the wider knowledge of its attributes. But it is also necessary to recognise weaknesses of the analytical approach. We rarely have in macrosociology or political economy measuring instruments of the kind at the disposal of the chemist analysing the mineral water. If we could say: the Californian economy comprises x per cent pure market governance, y per cent basic state support and z per cent immigrant community dynamic effects we would be saying something very significant. But we cannot; we can only say: the impact of immigrant communities may be important as catalysts for innovation. The analytical approach thus runs the risk of being wrong-footed as less scientific by an alternative that presents a false scientific precision. Labelling works best when there is only a limited number of models to which cases can be assigned, but these models embrace a wide range of institutions without worrying about excessive complexity. Conversely, the analytical method is most likely to be found among theories that accept a larger number of types but are less ambitious in their institutional coverage. These theories can best demonstrate their richness when showing how complex an individual case can be, and for that require a large number of models. They are therefore only really feasible when a limited number of institutions is being considered. The strongest point of the labelling approach is its clarity. The designation of still or sparkling is always far more prominent on the water bottle than the detailed chemical analysis, and it is the only information in which most consumers are interested. Likewise, policy makers, investors and other users of social research into forms of capitalism probably want to know simply: is this economy like the US or like Germany? The labelling model is also of particular value when measuring instruments are crude. We do not have finely tuned ways of measuring elements within a national economy; but we might be able to say what an economy is more or less like in other words, which simple model does it most resemble? An analytical approach, in contrast, is able to depict the actors within its cases as confronting an empirical complexity made up of elements of a number of models. A number of recent studies suggest that authors are becoming more willing to accept the degree of complication and apparent incoherence that this implies. 65 If these actors are institutional entrepreneurs, then, unlike the actors within a game theory, they can be presented as having the capacity to try to combine these elements in new ways, making use of serendipitous redundancies embedded in the empirical incongruences of their situation. As theorist and real-world actors interact, the former may be able to develop new theoretical cases out of the recombinant institutions produced by the more successful of these attempts. The two approaches present opposed logics of research. What is noise for the labelling approach becomes grist for the mill of explaining what actors can do for the analytical approach. A high degree of diversity within a 453

16 Colin Crouch case, a problem for labelling theory, becomes for an analytical theory a crucial independent variable for explaining capacity for change. Under the conditions of early twenty-first century capitalisms there is not a question of whether an economy will change, but how it is doing so. The accurate study of this situation surely requires a shift from the labelling to the analytical strategy. Notes 1. Robert Boyer & Yves Saillard (eds), Théorie de la régulation: L état des savoirs (La Découverte, 1995). 2. Andrew Shonfield, Modern Capitalism (Oxford University Press, 1964). 3. Colin Crouch, Capitalist Diversity and Change: Recombinant Governance and Institutional Entrepreneurs (Oxford University Press, 2005). 4. Robert Boyer, The variety and unequal performances of really existing markets: farewell to Doctor Pangloss? in: J. Rogers Hollingsworth & Robert Boyer (eds), Contemporary Capitalism: The Embeddedness of Institutions (Cambridge University Press, 1997), pp Michel Albert, Capitalisme contre capitalisme (Seuil, 1991). 6. It has become routine to use the term Anglo-Saxon here, but it is problematic. It was developed originally to group together the collection of peoples of English, German and Scandinavian origin who inhabited England on the eve of the Norman invasion of 1066; it served to contrast them with both the French-speaking (though originally Scandinavian) invaders and the Celtic inhabitants of other parts of the British Isles. Such a term became useful over 800 years later to distinguish the British and other northern European (and by now primarily Protestant) inhabitants of the late 19th century US from more recent Latin-language-speaking and Irish immigrants, and by extension also from Poles and other Catholics as well as from black people and Jews. Its contemporary use by academic social scientists, as well as international organisations like the OECD, seems not only blithely ignorant of these connotations, but also commits the solipsism of using it mainly as a contrast with Germany which includes the Saxon half of the mythical Anglo-Saxon identity. Its contemporary use also normally includes Ireland a people explicitly excluded from both the original and subsequent US terms. It is, in fact, used entirely consistently to identify that group of countries where English is the dominant language and the majority population is white-skinned: the UK, Ireland, the US, Canada, Australia and New Zealand. The correct, unambiguous term, which precisely identifies this group of countries is Anglophone, and one wonders why this clear and accurate term is not used instead of the more popular, exotic but highly dubious alternative. To insist on this point is not pedantry, but draws our attention to certain possible implications of the fact that the economics literature which finds it far easier to make sense of the Anglophone economies than others is itself almost solely Anglophone. 7. John L. Campbell & Ove K. Pedersen, The second movement in institutional analysis, in: John L. Campbell & Ove K. Pedersen (eds), The Rise of Neoliberalism and Institutional Analysis (Princeton University Press, 2001), pp ; and David Coates, Models of Capitalism: Growth and Stagnation in the Modern Era (Polity, 2000). 8. Coates, Models of Capitalism. 9. Organisation for Economic Cooperation and Development, The Jobs Study (OECD, 1994). 10. Bruno Amable, Institutional Complementarity and Diversity of Social Systems of Innovation and Production, Review of International Political Economy, Vol. 7, No. 4 (2000), pp Campbell & Pedersen, The second movement in institutional analysis. 12. Peter Kjær & Ove Pedersen, Translating liberalization: neoliberalism in the Danish negotiated economy, in: Campbell & Pedersen, The Rise of Neoliberalism and Institutional Analysis, pp Desmond King & Stuart Wood, The political economy of neoliberalism: Britain and the United States in the 1980s in: Herbert Kitschelt, Peter Lange, Gary Marks & John Stephens (eds), Continuity and Change in Contemporary Capitalism (Cambridge University Press, 1999), pp Peter A. Hall & David Soskice (eds), Varieties of Capitalism: The Institutional Foundations of Comparative Advantage (Oxford University Press, 2001). 15. Peter Hall & David Soskice, Introduction, in: Hall & Soskice, Varieties of Capitalism, pp Maria Estevez-Abe, Torben Iversen & David Soskice, Social protection and the formation of skills: a reinterpretation of the welfare state, in: Hall & Soskice, Varieties of Capitalism, pp Ibid., p

17 Models of Capitalism 18. Jonathan Zeitlin, Introduction: governing work and welfare in a new economy: European and American experiments, in: Jonathan Zeitlin & David Trubek (eds), Governing Work and Welfare in a New Economy: European and American Experiments (Oxford University Press), 2003, pp. 1 30; M. Bertoldi, Varietà e dinamiche del capitalismo, Stato e Mercato, No. 69 (2003), pp ; and Colin Hay, Common trajectories, variable paces, divergent outcomes? Models of European capitalism under conditions of complex economic interdependence, unpublished paper delivered to Conference of Europeanists, Chicago, March Bruno Amable, The Diversity of Modern Capitalism (Oxford University Press, 2003), ch. 5; and C. Berggren & S. Laestadius, The Embeddedness of Industrial Clusters: The Strength of the Path in the Nordic Telecom System (Kungl. Tekniska Högskolan, 2000). 20. Robert Boyer, New Growth Regimes, but still Institutional Diversity, Socio-Economic Review, Vol. 2, No. 1 (2004), pp. 1 32; and Robert Boyer, The Future of Economic Growth: As New Becomes Old (Edward Elgar, 2004). 21. Charles Ragin, Fuzzy-set Social Science (University of Chicago Press, 2000). 22. Hall & Soskice, Introduction, p Ibid., p David Soskice, Divergent production regimes: coordinated and uncoordinated market economies in the 1980s and 1990s, in: Kitschelt et al., Continuity and Change in Contemporary Capitalism, pp Hall & Soskice, Introduction, p Kathleen Thelen, Varieties of labor politics in the developed democracies, in: Hall & Soskice, Varieties of Capitalism, pp David Soskice, Wage Determination: The Changing Role of Institutions in Advanced Industrialized Countries, Oxford Review of Economic Policy, Vol. 6, No. 4 (1990), pp ; and Lars Calmfors & John Driffill, Bargaining Structure, Corporatism and Macroeconomic Performance, Economic Policy, Vol. 3, No. 1 (1988), pp Vivien Schmidt, The Futures of European Capitalism (Oxford University Press, 2002). 29. Ibid., ch Gøsta Esping-Andersen, The Three Worlds of Welfare Capitalism (Polity, 1990). 31. Frank G. Castles & D. Mitchell, Three Worlds of Welfare Capitalism or Four?, Working Paper No. 63, Luxembourg Income Study, Luxembourg, 1991; Mary Daly, A fine balance: women s labour market participation in international comparison, in: Fritz Scharpf & Vivien Schmidt (eds), Welfare and Work in the Open Economy, Volume 11: Diverse Responses to Common Challenges (Oxford University Press, 2002), pp ; Manuela Naldini, Evolution of Social Policy and the Institutional Definition of Family Models: The Italian and Spanish Cases in Historical and Comparative Perspective, unpublished PhD thesis, European University Institute. Florence, 1999; Maurizio Ferrera, Le trappole del welfare (II Mulino, 1997); Bernhard Ebbinghaus, When labour and capital collude: the political economy of early retirement in Europe, Japan and the USA, in: Berchard Ebbinghaus & Philip Manow (eds), Comparing Welfare Capitalism: Social Policy and Political Economy in Europe, Japan and the USA (Routledge, 2001), pp ; and Elke Viebrock, The Role of Trade Unions as Intermediary Institutions in Unemployment Insurance: A European Comparison, unpublished PhD thesis, European University Institute, Florence, Richard Whitley, Divergent Capitalisms (Oxford University Press, 1999). 33. Ibid., p Ibid., p Ibid., p Amable, The Diversity of Modern Capitalism. 37. Colin Crouch, Social Change in Western Europe (Oxford University Press, 1999). 38. Amable, The Diversity of Modern Capitalism, ch Ibid., p W. R. Scott, Institutions and Organizations (Sage, 2001), ch. 9; and Hay, Common trajectories, variable paces, divergent outcomes?. 41. Colin Crouch, Patrick Le Galès, Carlo Trigilia & Helmut Voelzkow, Local Production Systems in Europe: Rise or Demise? (Oxford University Press, 2001). 42. Jerry Hage & Catherine Alter, A typology of interorganizational relationships and networks, in: Hollingsworth & Boyer, Contemporary Capitalism, pp Crouch, Capitalist Diversity and Change, ch

18 Colin Crouch 44. Schmidt, The Futures of European Capitalism, ch Gregory Jackson, Varieties of capitalism: a review, unpublished manuscript. 46. Maurizio Ferrera, Anton Hemerijck & Martin Rhodes, The Future of Social Europe: Recasting Work and Welfare in the New Economy, Report for the Portuguese Presidency of the European Union, Lisbon, 2000; Anton Hemerijck & Martin Schludi, Sequences of policy failures and effective policy responses, in: Fritz Scharpf & Vivien Schmidt (eds), Welfare and Work in the Open Economy. Volume I: From Vulnerability to Competitiveness (Oxford University Press, 2001), pp Zeitlin, Introduction ; and Paul Windolf, Corporate Networks in Europe and the United States (Oxford University Press, 2002), p Amable, Institutional Complementarity and Diversity of Social Systems of Innovation and Production ; and Robert Boyer & Maurice Didier, Innovation et Croissance (La Documentation Française, 1998). 49. Christopher Freeman, The National System of Innovation in Historical Perspective, Cambridge Journal of Economics, Vol. 19, No. 1 (1995), pp. 5 24; Bengt-Åke Lundvall (ed.), National Systems of Innovation: Towards a Theory of Innovation and Interactive Learning (Pinter, 1992); and Robert R. Nelson, National Innovation Systems: A Comparative Analysis (Oxford University Press, 1993). 50. Colin Crouch, Breaking open black boxes: the implications for sociological theory of European integration, in: Anand Menon & Vincent Wright (eds), From the Nation State to Europe? Essays in Honour of Jack Hayward (Oxford University Press, 2001), pp Susan Cotts Watkins, From Provinces into Nations: Demographic Integration in Western Europe (Princeton University Press, 1991); Andres Rodríguez-Pose, Dynamics of Regional Growth in Europe: Social and Political Factors (Oxford University Press, 1998); Andres Rodríguez-Pose, Convergence or Divergence? Types of Regional Responses to Socio-Economic Change in Western Europe, Tijdschrift voor Economische en Sociale Geografie, Vol. 90, No. 4 (1999), pp Hugh Radice, Globalization and National Capitalisms: Theorizing Convergence and Differentiation, Review of International Political Economy, Vol. 7, No. 4 (2000), pp Hall & Soskice, Introduction, pp ; Whitley, Divergent Capitalisms, ch. 5; and Paul Hirst & Grahame Thompson, Globalization in question: international economic relations and forms of public governance, in: Hollingsworth & Boyer, Contemporary Capitalism, pp J. Beyer, One best way oder Varietät? Strategischer und organisatorischer Wandel von Großunternehmen im Prozess der Internationalisierung, MPIfG Discussion Paper 01/2, Max-Planck-Institut für Gesellschaftsforschung, Cologne, Jackson, Varieties of Capitalism. 56. Hugh Radice, Globalization and National Differences, Competition and Change, Vol. 3, No. 3 (1998), p Ibid., pp Eric Helleiner, States and the Reemergence of Global Finance (Cornell University Press, 1994). 59. Radice, Globalization and National Differences, pp T. K. Hopkins & Immanuel Wallerstein, World-systems Analysis: Theory and Methodology (Sage, 1982). 61. Scott, Institutions and Organizations, p J. Rogers Hollingsworth & Robert Boyer, Introduction, in: Hollingsworth & Boyer, Contemporary Capitalism, p Robert Goodin, B. Headey, R. Muffels & H.-J. Dirven, The Real Worlds of Welfare Capitalism (Cambridge University Press, 1999). 64. J. Rogers Hollingsworth, Continuities and changes in social systems of production: the cases of Japan, Germany, and the United States, in: Hollingsworth & Boyer, Contemporary Capitalism, p Amable, The Diversity of Modern Capitalism; Glenn Morgan, Richard Whitley & Eli Moen (eds), Changing Capitalisms? Complementarities, Contradictions and Capability Development in an International Context (Oxford University Press, 2005); Wolfgang Streeck & Kathleen Thelen (eds), Change and Discontinuity in Institutional Analysis: Explorations in the Dynamics of Advanced Political Economies (Oxford University Press, 2005); Wolfgang Streeck & K. Yamamura (eds), The Origins of Nonliberal Capitalism: Germany and Japan in Comparison (Cornell University Press, 200l); and K. Yamamura & Wolfgang Streeck (eds), The End of Diversity? Prospects for German and Japanese Capitalism (Cornell University Press, 2003). 456

19

20 Economy and Society Volume 38 Number 4 November 2009: Regional and sectoral varieties of capitalism Colin Crouch, Martin Schröder and Helmut Voelzkow Downloaded by [Koc University] at 05:03 26 January 2015 Abstract This study seeks to go beneath the generalizations that constitute characterizations of national economies in order to examine local and sectoral diversity in particular, forms of capitalist organization at the level of localized sectors. It reports on the findings of research based on detailed case histories of local economies in four different types of production: modernized craft manufacturing (furniture), mass production (motor vehicles), high-technology production (biopharmaceuticals) and high-tech services (television film-making). In each case a local economy in Germany (usually seen counter-factually as an example of a particularly national system) was compared with one elsewhere in Europe: respectively, southern Sweden, Hungary (compared with eastern Germany) and the UK (for two studies). In the analysis, companies act rationally in response to sector-specific challenges, being partly bound by the existing institutional framework that they encounter, but partly acting to alter it. Two possibilities are distinguished and found in the cases. In the first (structurally conservative) case, arrangements of governance in the national innovation and production system prove to be beneficial for the companies and their aim to stand up to international competition. Insofar as national institutions help companies to deal with competition on their markets, they will probably try to preserve these arrangements. In the second (innovative) case, companies turn away from the national context and develop their own local governance structure. If the national institutional structure is seen as not adequate or non-fitting to deal with sectorally specific terms of competition, then the internal and external coordination of companies in reaction to challenges posed by the market is likely to deviate from the national structure. In some instances evidence of creative incoherence, where local deviation from the national model provides a creative impulse, is found. Colin Crouch, Warwick Business School, Warwick CV4 7AL, UK. Colin.crouch@wbs.ac.uk Copyright # 2009 Taylor & Francis ISSN print/ online DOI: /

21 Colin Crouch et al.: Regional and sectoral varieties of capitalism 655 Keywords: local production systems; varieties of capitalism; creative incoherence; furniture-making; motor industry; biopharmaceuticals; TV film-making; Germany; Sweden; Hungary; UK. Introduction Downloaded by [Koc University] at 05:03 26 January 2015 Even when national economies are strongly characterized by relative success in certain kinds of economic activity rather than others, there will usually be exceptions: firms, or districts, or whole sectors that thrive in a way that seems incongruous from the perspective of the overall pattern. These exceptions should not be treated as anomalies, or as irritating unexplained residuals that analysts hurry past, with eyes averted and fixed on the big picture of the major features of an economy that are consistent with their theories. Exceptions and minority features are as much a part of reality as predominant characteristics, and worthy of study as telling us something about the full complexity of how economic institutions operate. There are several possible reasons for exceptions, all consistent with accounts that identify overall national patterns: 1. There may be characteristics of a local institutions and infrastructure that support forms of economic organization that differ from and may even defy the overall national architecture. 2. In a stronger version of 1, there may actually be creative incoherences at work, rather than confusion and handicap, in the difference between national and local, whereby local institutional entrepreneurs are able to produce innovative outcomes, working between the contradictory incentives of national and local institutions. By creative incoherence we understand a truly Schumpeterian form of innovation, in which entrepreneurs produce something new by putting together previously untried combinations of elements. The most original examples of all will be those where in the past the separate elements had been actually incompatible with each other, creating therefore a creative incoherence. 3. Firms may at times be less bounded by national institutions than theory about the national base of types of capitalism usually assumes. In a recent comparative research project (Crouch & Voelzkow, 2009; Voelzkow et al., 2006) we studied four different localized sectoral economies that differed in various ways from the predominant characteristics of some of the national economies in which they are found. While examples and case studies do not enable us to prove anything, they do enable us to explore how some of the reasons for exceptions work out in practice. The central focus of the research was Germany, a country often depicted as particularly lacking in institutional diversity. 1 The sectors were chosen to represent highly diverse forms of production: furniture (an instance of modernized craft production); motor vehicles (classic mass production);

22 656 Economy and Society Downloaded by [Koc University] at 05:03 26 January 2015 biopharmaceuticals (a leading high-tech sector); and advanced services (television film-making). From the perspective of recent research on varieties of capitalism theory (Hall & Soskice, 2001b) the first two would be seen as mature sectors characterized by incremental innovation and therefore likely to succeed in Germany; the latter two would be seen as sectors of radical innovation, for which Germany is considered poorly equipped. Cases in the first two should therefore show the German model operating in its classic way; those in latter two should demonstrate one or more of the above forms of support for diversity if their ability to survive and thrive is to be explained. In each case, we compared the German localities with those in other European countries specializing in the same sector, but where a different overall national logic should be expected. For furniture we compared some districts in Sweden: furniture is typically a sector for small to medium-sized firms, while Sweden is regarded as providing an environment favourable only to large corporations. Special factors should therefore be needed to explain successful survival there. For motor vehicles we took the opportunity to carry out a different kind of test: of the transferability of West German institutions to the German neue Länder, therefore a plant in Eastern Germany, and compared what we found there with another central European location outside the frame of German institutions. We therefore compared plants of the Volkswagen/Audi group in East Germany and Hungary. In both the high-tech and advanced services cases we compared west German localities with those in the UK, regarded by Hall and Soskice (ibid.) as a country where such sectors, considered to be characterized by radical innovation, are likely to thrive better than in Germany. In these two studies it is German success that requires special explanation, British success being taken for granted by the varieties of capitalism model. A common starting point for studies of this kind has been the analysis of institutions seen as important in determining different types of economic performance, mainly:. systems of corporate financing and corporate governance, sometimes patterns of inter-firm cooperation;. industrial relations and vocational training;. research and development (R&D). It is assumed that firms behave according to the rules provided for them by the specific institutional arrangements, which thus coordinate and govern them. There is a tendency in studies that are concerned with the varieties of capitalism to reduce more or less drastically the empirical variety of institutional arrangements in different institutional sectors. The variety of institutional configurations that produce innovation and apply it to industrial production is again reduced to a number of models or country families (Rhodes & van Apeldoorn, 1997; Schmidt, 2002). In order to achieve this, the single results of the research undertaken are put together into a holistic image,

23 Colin Crouch et al.: Regional and sectoral varieties of capitalism 657 Downloaded by [Koc University] at 05:03 26 January 2015 which gives the impression of extreme coherence. The basic element that allows fusing the individual results of national particularities into a small number of models lies in the concept of a complementary order. Theoretically speaking, the variety of combined institutional particularities of sectors this follows from the complementary order assumption must be very limited, because institutional particularities have to fit together mutually. Certain combinations reinforce each other; others exclude each other due to their mutual dysfunctionality. A system of corporate finance that rests on incremental innovation in existing and successful products, and thus oriented towards long-term projects, fits into a system of vocational and further training, which invests large efforts into the qualification of employees. Vice versa, a system of corporate financing that is oriented towards the maximizing of short-term profits fits better into a system of vocational and further training that directs investments into the qualification of employees accordingly. However, if these employees have been only poorly trained by the firm, due to the low investments which are generally activated for human resources, then it does not make any sense to keep these exchangeable employees in the firm whenever crisis is ahead. Hire and fire politics at the level of the firm fit a system of short-term entrepreneurial strategies as much as relatively high stability of employment fits production systems with long-term orientations. According to this thesis, the German model does not only result from a number of characteristic institutional particularities, but it also seems to constitute its general structure through the complementary order of these particularities, which fit together in their functional logic and, because of this, mutually reinforce each other, guaranteeing a comparative advantage in diversified quality production (DQP) (Streeck, 1992), but suffering from weaknesses in the branches of the new economy. Some recent research critical of this approach has shown that institutional arrangements may be more loosely coupled within a national innovation and production system, allowing not only for (perhaps temporarily) incoherences at the national level (for instance, between corporate financing and labour participation rights), but also for the creation of autonomous subsystems at the sectoral or local level. (In fact attention had been drawn to sector-specific governance in the early 1990s, before the main surge of interest in national models itself (Hollingsworth, Schmitter & Streeck, 1994).) These subsystems may establish their own governance structures, diverging from the national model. As Trigilia (2004) asks, if the national level is the main determinant of economic diversity, why are the industries seen as nationally typical not distributed evenly across the national territory? In other words, why do clusters exist? That they do exist, and that they cannot all be explained in terms of physical geography, is well established, and not only for traditional sectors (Crouch, Le Galès, Trigilia & Voelzkow, 2001) but also in the so-called high-tech ones (Crouch, Le Galès, Trigilia & Voelzkow, 2004, pt 3; Kenney, 2000; Saxenian, 1999; Swann, Prevezer & Stout, 1998).

24 658 Economy and Society Downloaded by [Koc University] at 05:03 26 January 2015 This diversity does not necessarily contradict theories of the variety of capitalism, or of national innovation systems (Lundvall, 1992; Nelson, 1993; Nelson & Winter, 1982), or of national competitive advantage (Porter, 1990), which ought all to be able to accommodate the idea of further subnational concentrations of governance, provided adequate special factors can be identified. They are merely more interested in delineating the dominant model, while we are interested in understanding the factors that support some deviations from it. The difference between our work and that of these other schools is simply that of focus. Therefore, for example, neither these theories nor that of the varieties of capitalism should have any difficulty with the fact that some areas of Germany specialize in high-quality motor vehicles, others in machine tools and others again in chemicals. 2 Local specialisms that depart from the logic of a national system in this way suggest that the nation-state is not necessarily always the most important level for determining the institutional environment of business. It is important that research pay attention to these instances. The fashion for dealing only in stylized facts encouraged by many economists in the name of parsimony and elegance can easily become an invitation to deal in stereotypes and over-generalizations. On closer inspection, therefore, the idea of a national system starts to seem less solid. It can be at least variegated and at most directly challenged by local institutions; and it may be in competition with both supranational regulatory regimes and internationally available institutions, services and persons (Brose &Voelzkow, 1999; Deeg & Jackson, 2007). These questions become even more acute when we consider parts of the world where national systems are weak. This is, for example, the case with the countries of Central and Eastern Europe. Most of the institutions of the state socialist period seem to have been swept away very rapidly during the early 1990s; that regime had in its time swept away most pre-existing institutions; and even before that, in much of central Europe the national level had in any case had a short-lived period following centuries of foreign domination. Further, during fifteen or so years of post-communist reconstruction, governments and other institutional actors in this part of the world have been presented with various powerful exogenous models. So the idea of national systems remains elusive. For those countries that have entered or are working towards entering the EU, there is the acquis communautaire; for all there is the World Bank, the International Monetary Fund and the OECD, and individual multinational firms considering major investments in production. These exogenous forces do not simply offer ideas for consideration; they offer important inducements to compliance. How then are these national systems to be described? Various contributors to the neo-institutionalist tradition have shown how it is possible to cope with diversity while retaining the concept of institutional constraint (Crouch, 2005; Crouch & Farrell, 2004; Schmidt, 2002; Streeck & Thelen, 2005; Thelen, 2004). There is no claim to predict the kinds of innovation that will be undertaken: this literature is an actor-centred institutionalism (Scharpf, 1997) that provides scope for agents, but they are

25 Colin Crouch et al.: Regional and sectoral varieties of capitalism 659 not completely free agents, their range of initiative being bounded somehow by their institutional starting point, as in a random walk. By looking at specific local cases, we are here better able to explore the potentiality for variation around a strong institutional core that lies in the hands of key actors than is the case with the national accounts that dominate the literature (for other local and sectoral approaches, see Crouch et al., 2001, 2004; Hollingsworth, Schmitter & Streeck, 1994; Swann, Prevezer & Stout, 1998). We can then examine to what extent that localized sector corresponds to the account usually given in the literature of the national institutional structure. Four paired case studies Downloaded by [Koc University] at 05:03 26 January 2015 Our research was conducted through interviews with firms and other institutions involved in the sectors concerned, together with analysis of relevant documents. Our teams, which were in most cases locally-based, also worked closely with experts in the localities concerned. In an article of this length we can clearly not give full accounts of the cases, but can only draw out the lessons that they convey for our general task of exploring how exceptional industries survive within national systems that should not be expected to support them. For more detailed accounts we refer readers to Voelzkow (2007) and Crouch and Voelzkow (2008). Furniture-making in north-west Germany and southern Sweden The production processes around furniture-making were transformed by mechanization, but many of the products do not depend on mechanization for either their design or existence. It has historically been a craft industry, dependent on high levels of manual skill and small firms, and this usually remains the case at the expensive and high added-value end of the range. But fully mechanized production on Fordist principles in large corporations has also developed strongly in the sector. Given that this is a light industry, with portable products, with pre-industrial roots but easily subject to mass production, it is easily susceptible to globalization for low-priced mass-market products. But some production continues in industrialized countries. This gives rise to obvious questions: under what circumstances is this survival possible? In what parts of the market are furniture firms in rich countries operating? Are they thriving? And, most important, what institutional arrangements are associated with their success and in which locations is it possible to find these arrangements? We compared the traditional district Ostwestfalen-Lippe in north-eastern Nordrhein-Westfalen that has about 20,000 people working in furniture production with a similar traditional case in Småland and Västergötland in southern Sweden, where we studied three contrasting clusters. These were: Tibro, a town with about 9000 inhabitants, which has about eighty furniture

26 660 Economy and Society Downloaded by [Koc University] at 05:03 26 January 2015 companies with 1500 employees; Virserum with only 2000 inhabitants, whose furniture industry has disappeared in the last fifty years; and Älmhult, where the success story of Ikea began. The last is a case apart, as Ikea is a single large firm, not an SME cluster. There is some difference between the national situations, as firms in Ostwestfalen-Lippe specialize in the production of kitchen furniture, while the Swedish data are based on furniture except kitchen fixings (but including kitchen tables and chairs, for example). Germany is known as a country in which smaller (mainly medium-sized) enterprises can thrive; they are well organized, and law and the fiscal regime are favourable to them and to family-based business in general. A structure of associations allows for coordination between smaller companies, which is often a functional equivalent to vertical integration. Sweden, in contrast, is an economy of large firms consolidated around an even smaller group of holding companies. The economy has therefore been historically concentrated in massproduction sectors dominated by large corporations: mining, motor vehicles, traditional pharmaceuticals and telecommunications. On the basis of national system analysis, we should predict a very poor outlook for an SME-based furniture industry in Sweden, certainly much poorer than in Germany (though varieties of capitalism theory as such would be unable to distinguish between the two countries, as both are coordinated economies). After some considerable reduction in its size and in particular its labour force, the Ostwestfalen-Lippe cluster is finding continuing success at the very top end of the market. The three southern Swedish clusters have gone in different directions. That in Tibro is flourishing; that in Virserum has more or less collapsed; Ikea enjoys considerable success, even though most of the firm s products are made elsewhere in the world. Our research therefore found a complex pattern of evidence (Sjöberg & Rafiqui, 2006; Voelzkow & Schröder, 2006a). To some extent the difference in level of aggregation between the two studies added to the complexity of outcomes. The Swedish study looks at three local clusters, while the German one looks at the regional level. Also, the Swedish researchers deliberately picked three clusters with different characteristics, so the Swedish case adds complexity by its very design. There are important bank-based modes of corporate finance in the German furniture cluster, but stock markets are increasingly becoming an alternative; two of the biggest companies in the cluster attempted to change into stock companies (Voelzkow & Schröder 2006a). In addition, restrictive credit-issuing criteria, imposed on German Hausbanken by the international regulation of Basel II (the agreement reached within the Basel Committee on Banking Supervision, the self-regulating institution of the international banking sector), have changed the traditionally close relationship between companies and banks in the German cluster. Many of the companies now have to disclose information that the (family) owners consider private. In the three Swedish clusters, corporate finance was more nuanced (Sjöberg & Rafiqui, 2006). Whereas the clusters of Tibro and Virserum have largely used the Swedish model of

27 Colin Crouch et al.: Regional and sectoral varieties of capitalism 661 Downloaded by [Koc University] at 05:03 26 January 2015 debt-based and thus bank-centred financing, as well as customer credit (a rather forgotten but often important mode of financing), the case of Ikea merits further attention. Part of Ikea s success seems to stem from the fact that it tried to support its expansion with its own resources, relying on neither bank-based nor stock-based financing modes, and thereby achieving greater freedom from external agencies. The typical German Mittelstand model of corporate governance can be found in the furniture industry of Ostwestfalen-Lippe (Voelzkow & Schröder, 2006a, 2006b). Of all companies in the cluster, 70 per cent are family-owned and 85 per cent employ fewer than 200 employees. However, family ownership is increasingly combined with a professional management team and some companies have been taken over by foreign competitors, which now expect to see high yields from their new objects of investment. At the same time, corporate takeovers have a tendency to replace traditional company networks. Even though bank-based corporate financing is still prevalent and thus ensures the stakeholder model to a large degree, this hints at a stronger accentuation of shareholder-value in the companies, which undermines stakeholder-based governance. Again, the cluster is conforming to a national model, but that model is itself undergoing change. Although SMEs are not well supported by the Swedish system of economic governance, the Tibro firms had developed institutions more similar to those found in Denmark or central Italy, companies sharing production sites that no single company could afford to acquire itself (Sjöberg & Rafiqui, 2006). Also, the local government supported the sector through place marketing and the provision of infrastructural institutions. The capacity of SMEs to act collectively was thus a functional equivalent to the hierarchical structure of a vertically and horizontally integrated company. Companies in Virserum in contrast did not find ways of cooperating, and one after the other went bankrupt. Ikea, being a big integrated company itself, was in this respect alone more typical of the pattern to be expected in the Swedish model. As we should expect for the German cluster, collective pay agreements (Flächentarifverträge) provide the basis for the remuneration of workers and for other industrial relations issues (Voelzkow & Schröder, 2006a, 2006b). But, once again, general changes entering the German model can be seen within the cluster: an increasing degree of flexibility is integrated into the agreements, turning them into framework agreements (Manteltarifverträge) within which further negotiation can be conducted at company level. Also, increasing price pressure leads supplier companies to outsource production to subcompanies that have not joined the collective pay agreements. A relatively cooperative approach towards industrial relations was cited as one factor in the success of the Tibro cluster over that of Virserum (Sjöberg & Rafiqui, 2006), here conforming to the familiar Swedish model. In Ikea there is a different, internal company effort to create a team spirit, encouraging employees to identify with the company and facilitating managementemployee interaction.

28 662 Economy and Society Downloaded by [Koc University] at 05:03 26 January 2015 In vocational training the German cluster yet again showed arrangements characteristic of the German model as a whole, but in terms of that model as it is changing. The bulk of education is done by the dual system of vocational training, combining on-the-job training and studies in professional schools (Berufsschulen) (Voelzkow & Schröder, 2006a, 2006b). However, the overall German trend towards an upgrading of qualifications can also be found in the cluster and is illustrated by the increasing importance of Berufsakademien that are trying to combine a system which is close to that of a university for applied sciences (Fachhochschule) with that of the traditional system for vocational training. Also typical of the traditional German model is the fact that the vocational training system is managed by the social partners. For the Swedish cases, the now familiar view again holds true. Whereas Virserum engaged in no more than the absolute minimum training, which was delivered in house and on the job, firms in Tibro offered inter-firm collaboration (Sjöberg & Rafiqui, 2006). In so doing, the cluster was able to attract employees from all over Sweden by its well-designed training schemes. In Ikea, as in many large Swedish firms, vocational training is largely internalized within the company, which maintains its own vocational schools. It thus tries somewhat to insulate itself from the Swedish system, where initial vocational education is normally provided within normal schools, not as part of an apprenticeship model. Conclusions SMEs can survive in furniture production only if they are backed up by collective competition goods, i.e. an infrastructure that increases their competitiveness and gives them access to possibilities that would usually be barred due to limited resources of individual companies (Crouch et al., 2001). For the cluster in Ostwestfalen-Lippe, the typical German model of diversified quality production continues to provide this kind of environment. However, it is essential to note that general institutional changes taking place in the country are also facilitating flexibility in this sector: evidence both that change has been necessary for the model and that it is being achieved. (For an account of the German national system as one in a state of considerable change rather than as in equilibrium, see Glassmann [2006a].) This demonstration of the capacity of the German model to change would refute any versions of capitalist diversity theory that take a strong stand on the inevitability of path dependence a characteristic that is not endemic to such theories, but which is occasionally found. In Sweden, the important local variations are highly instructive. Ikea in some respects demonstrates the Swedish system of favouring large corporations (which is very different from the German one) at work, though in practice Ikea deviates from most typical Swedish institutions and constitutes a case in itself. It is a case of exception type 1 (a local in this case corporate context deviant from and even defiant of the national model). To some extent it is also a case of 2, creative incoherence, in that the firm built its distinctive style through a mixture of strongly Swedish characteristics and some that are quite outside the national frame. It is also a case of 3, the lack of reach of a national model.

29 Colin Crouch et al.: Regional and sectoral varieties of capitalism 663 While maintaining important links with Sweden, Ikea today has its headquarters in the Netherlands and has production plants in many countries throughout the world. The crisis of the sector in Virserum also demonstrates what we would expect to be the case for a sector based on SMEs. Tibro, however, demonstrates the capacity of local institutions to develop their own internal coherence, standing out against the prevailing national model. It is then a case of 1, the comparison with Virserum demonstrating the extent to which the firms and local authorities in Tibro had to forge some distinctive institutions to support their industry. Motor-vehicle manufacture in eastern Germany and western Hungary Downloaded by [Koc University] at 05:03 26 January 2015 Motor-vehicle manufacture was the classic industry of the mass-production age, and a major contributor to manufacturing output in many industrial countries in the second half of the twentieth century. Although famously subjected to Taylorist techniques during the early twentieth century, which enabled it to employ low-skilled labour, motor manufacturing was transformed in the last third of that century to become far more demanding of its labour force, suppliers and institutional infrastructure (Womack, Jones & Roos, 1991). In Germany and Japan in particular there were major developments both in advanced machinery and in labour skill that made possible more efficient production and greater sophistication as well as variety of products. These combined with developments originating in Japan, which improved the efficiency and quality of relations between the core manufacturing firms and their suppliers, that depended on both high skills and high levels of organization. It also became an industry in which constant improvements in design and product performance and variety became major instruments of competition. Therefore the automobile industry is seen as one of the core sectors for incremental innovation. A key question arises concerning the institutional requirements of the shift in production away from the core countries. But what does a new production area require if it is to compete at these sophisticated levels of production? The accidents of history here make possible an exceptional experiment (Keune & Tóth, 2006; Piotti, 2006). One among the countries of Central Europe, the German Democratic Republic, had occasion to unite itself with the western Federal Republic of Germany almost directly after the fall of communism, and in theory immediately inherited all West German institutions, an entire acquis nationale. The other developing economies, including Hungary, had both the obligation and the opportunity to construct, or to import, new institutional structures, possibly in continuity with some survivals from the past. We here therefore compare the plant established by the Volkswagen-Audi group in Zwickau, Saxony, in Eastern Germany, with that in Győr in north-western Hungary. Varieties of capitalism theory would predict that, provided the West German institutions had been appropriately transferred, an East German region would

30 664 Economy and Society Downloaded by [Koc University] at 05:03 26 January 2015 soon start to operate as part of a coordinated economy and would therefore succeed at the motor industry. It is not clear that the theory has anything to say about Hungary, as its protagonists have not yet classified the former communist countries according to their typology. However, it might be expected to conclude that, given that post-communist Hungary has not had time to establish deeply embedded institutions of the kind required for a coordinated market economy, this industry should not thrive there. It would be reasonable for varieties of capitalism theory also to conclude that Hungary has probably not had time to develop the preconditions for a sophisticated liberal market economy either. It might therefore predict that neither up-market DQP nor radical innovation would be likely to take hold there, or in the rest of Central and Eastern Europe, expecting instead low-quality mass production. In general, that is the case with these economies. National innovation systems theory would similarly see eastern Germany now following the general (western) German path, and would have to remain silent about Hungary until it had analysed its institutions. From our perspective an important issue is whether eastern German Länder simply inherited West German institutions or whether deviations in the transfer have imported certain novelties. Hungary we see as a meeting-point of many institutional fragments: some national historical residues; some path-independent new approaches; some regional specificities; institutional borrowings; and not least the organizational preferences of VW- Audi itself which may or may not be simple transfers of their practices in Germany and other Western European economies. As we are dealing with a German multinational group, corporate finance and governance are matters for corporate headquarters and do not enter as variables into the comparison between the East German and Hungarian plants. Instead, we focus on relations between VW and Audi, on the one hand, and their local suppliers, on the other, in addition to the other variables. At Zwickau these relations follow a dual pattern (Piotti, 20045). VW has very stable relationships with those suppliers that the company brought with it from Western Germany, though some local suppliers are slowly joining the circle. Inter-company relations tend to be cooperative and supported by the German model of public and semi-public research institutions with which companies collectively cooperate, and which has been imported alongside unification. Even though Audi is delegating more and more production and some research to Győr, the local content of production remains low (Keune & Tóth, 2006; Keune, Kiss & Tóth, 2004). Rather, as production in Győr increases, more and more companies from the West are following Audi and increasing their capacities there. Whereas relations between Audi and its established supplier companies are cooperative, relations with local companies are less intense and less stable. The underlying structure of formerly West German institutions is not transferred to local Hungarian firms, though it is possible that examples will spread to the region from the suppliers who arrived from West Germany with Audi. In Zwickau, the dual system of vocational training works rather well, even though there are no strong associations to support it (Piotti, 20045). The

31 Colin Crouch et al.: Regional and sectoral varieties of capitalism 665 Downloaded by [Koc University] at 05:03 26 January 2015 basic system that existed in the former East Germany was very similar to that in the West. In addition, VW and some other companies seem to be investing in training efforts to ensure a supply of adequately qualified employees, and possibly even to help the region. In reviving the system that existed in Hungary before the end of the communist era, Audi favoured a system that was not very far from the German one, introducing vocational training schools with much the same content as in Germany (Keune & Tóth, 2006). However, the involvement of the social partners in the setting up of training schemes was rejected in favour of direct control of the training schemes through Audi. VW was eager to apply West German collective pay agreements in Eastern Germany (Piotti, 20045, 2006). However, local suppliers strongly refused to adhere to these. VW and the first and second-tier suppliers who followed it from West Germany also brought with them the second pillar of the German industrial relations system, works councils, as an automatic part of the package of institutions now being extended to the neue Länder. Local East German supplier companies have again, however, been more reluctant to do this. They did not consider that works councils served their interests; it is also important to note than in their case VW applied no normative or other pressures to persuade them to accept these Western institutions. In these local firms, works councils tend to be accepted only insofar as they do not oppose management decisions and refrain from representing workers interests against the management. However, it should be noted that this change in the role of the works councils is also taking place in the German model as a whole. Audi is encouraging cooperative industrial relations in Győr (Keune & Tóth, 2006; Keune, Kiss & Tóth, 2004). However, the firm s reluctance to accept strong and unified trade unions indicates that it is a rather limited version of coordinated industrial relations that Audi is trying to establish. Trade unions and works councils are viewed favourably if they facilitate coordination between management and labour, but not if they seek to become powerful stakeholders. Conclusions VW and Audi have brought their own institutional model along with their production sites to Zwickau and Győr respectively. However, this does not mean that they impose the typical West German model of intercompany relations, industrial relations and vocational training on local suppliers. Even though they partially export aspects of the German model to the degree that they deem them supportive to the management s aims, the companies also try to maintain control of the institutions they use. This means especially that they are not willing to accept a strong role for the social partners or for the workforce to the detriment of management power. In our two clusters, VW and Audi are acting as institutional entrepreneurs, supplying collective competition goods for their local economy by actually building up a distinctive institutional structure. In doing so, particular elements of the German model are selectively exported, and there is as one would expect a further difference between the East German and Hungarian contexts, with

32 666 Economy and Society considerably more transfer in the former case. This selectivity enables us to challenge institutional theories that insist on the integrity and necessity of a particular institutional Gestalt. The Zwickau case largely confirms what we should expect from most theories of capitalist diversity: transfer of the German model, but in weakened form though we should recall from our discussion of the furniture case that the German model is itself changing in similar directions in former West Germany. The Győr case and to some extent the Zwickau one demonstrates an exception of type 3: national models are not necessarily determinant. In this case, it is the internal managerial decision-making of a powerful corporation that is able to shape institutions of production. Downloaded by [Koc University] at 05:03 26 January 2015 Biopharmaceuticals around Munich and Cambridge Biotechnology is becoming as emblematic of the early twenty-first century as motor vehicles were of the late twentieth. The sector applies very advanced science and strong links between universities and the rest of the public research base, on the one hand, with flexible and sophisticated firms on the other. It has distinctive and high requirements for skills, financial arrangements, corporate governance and legislative framework. Since it is concerned with human health, it is subjected to intensive regulation, and, although science is moving rapidly, it can often take several years from a scientific discovery to the development of a truly new product and its arrival in the market. The question that arises is: given the precision of these requirements, is this sector one that can thrive only in certain institutional contexts? For varieties of capitalism theory the biochemicals sector is as paradigmatic to the liberal market form of capitalism as the motor industry is to the coordinated market form (Hall & Soskice, 2001a), as the liberal market is considered to favour high-risk entrepreneurial activities in new sectors. The main biotechnology district in Germany is that around Munich, the capital of Bavaria. If the sector has strengths, they need to be explained by one or more of our four sources of exception. Within Europe the UK is usually seen as the closest to a liberal market economy and the first successful biotechnology and high-tech region is usually considered to be that around the city of Cambridge (Casper & Kettler, 2001). We therefore compare Munich with this district, and also examine the relationship of the Cambridge area to the national British institutional structure. Companies in the two clusters are roughly comparable in size, turnover and markets. The Munich cluster is linked to the University of Munich as well as to a Max Planck Institute, whereas the Cambridge cluster has established links to its university. (The following account of both cases depends heavily on Jong [2006].) Due to sectoral specificities, both clusters are highly concentrated, linked to international networks and compete on the same international markets. Both need access to research institutes, venture capital and a management that is capable of linking research to market use of the products.

33 Colin Crouch et al.: Regional and sectoral varieties of capitalism 667 Downloaded by [Koc University] at 05:03 26 January 2015 The UK venture capital industry seems to cater well to the corporate finance needs of the biotech cluster in Cambridge, as would be predicted by most theories. However, it is the cluster in Germany that shows some interesting results. As one might suppose, bank-based modes of financing were not adapted to the needs of Munich-based biotech firms. These shortcomings were however compensated by two functionally equivalent institutions. Firms relied on public funding and used foreign venture capital markets, thereby circumventing the rigidities of the German corporate financing model and its underdeveloped market for venture capital. This raises an important point about the idea of national institutional systems in a global economy with liberalized capital movements: firms can escape certain elements of their national systems. Similarly, Munich firms avoid any difficulties in recruiting top managers that might be created by the primarily internally-oriented German labour market by relying on international labour markets: half the firms in Munich had acquired their CEO from a foreign company, whereas this was the case for only one of nineteen biotech firms in Cambridge. In neither country did national institutions of industrial relations play a part. As largely SMEs employing mainly highly-skilled labour, firms in this sector are rarely unionized; their staffs salaries are not collectively bargained; and in the German cases very few firms have works councils. Neither are national vocational training institutions significant in either case, as the core staff are educated to higher levels than normally associated with this form of training, while national first-level university education systems are also not important, as these firms recruit globally. More relevant are differences at the level of doctoral training, where neither Germany nor the UK has yet rivalled the great US universities in establishing highly innovative graduate schools linked to high-tech firm start-ups (Jong, 2007). This last point leads to consideration of research and development. The varieties of capitalism approach would expect a high degree of cooperation with public research infrastructure for the German cluster (Hall & Soskice, 2001a, p. 26). This was indeed the case, as the share of scientific publications produced in collaboration with public research institutes was twice as high in Munich as in Cambridge. However, the outcome of this was rather surprising: the stronger academic embeddedness of the Munich cluster seems to have been leading to better performance in developing radically innovative products there than in Cambridge. Conclusions The example of Munich-Martinsried illustrates how a sciencebased cluster can bypass the rigidities of the German political economy by docking on to international structures. Apart from the close link to the German innovation system (in the form of contacts with universities and research institutes), biotechnology in Munich constitutes something of a liberal island, which confirms some of the prior studies on biotechnology in Germany (cf. Casper, 2002). It is therefore a case of exception 3 the limitations of

34 668 Economy and Society national structures. Whereas in the motor industry case these structures could become partly subordinate to the managerial strategy of a large firm, here it is more a case of the Europeanization and globalization of economic governance, and a demonstration of how national regimes are today not the same kind of iron cage for entrepreneurs as they possibly were in an earlier period. Governance of the biotechnology cluster in Cambridge, however, largely corresponds to the national stereotype of the liberal market economy of the UK, but with the paradoxical result that incremental innovations are more prevalent in the British cluster. This confirms earlier findings suggesting that the German biotechnology industry is not being outperformed by the British one (Lange, 2005, 2006). As should not surprise us, having an appropriate context does not guarantee success. Downloaded by [Koc University] at 05:03 26 January 2015 Television film-making in Cologne and London Television film-making has been transformed in recent years by both technological and organizational changes. First, developments in camera and lighting technology have transformed the skills needed by camera teams and reduced the numbers of personnel needed in a team. Second, the break-up of public monopoly television programme providers produced a major growth of very small enterprises. The question arises whether one can identify different ways in which actors in different institutional contexts responded (Elbing, 2006; Elbing & Voelzkow, 2006a, 2006b; Glassmann, 2006a, forthcoming). We here again use a comparison between Germany and the UK. In both countries national broadcasting monopolies were required during the 1980s to downsize their in-house production activities and purchase programmes from independent producers. In both cases a large number of very small, projectbased firms developed, and in both anxieties were expressed that some important collective goods provided for the sector by the public producer in particular workforce training would be at risk. The British industry is focused on the national capital, London, though in reality it is concentrated on a very small part of that city, the area known as Soho that has long been a centre for the film and entertainment industry. About 70 per cent of all UK TV and film production companies are situated here. In Germany it is concentrated on Cologne in Nordrhein-Westfalen, close to the previous capital of the Western Republic, Bonn. Almost a third of outsourced German TV production is concentrated in Cologne and the immediately adjacent areas. Differences are to be found in the ways pressure was produced: in the British case, regulations created forced outsourcing, beginning with the launch of Channel 4 as a channel that would commission programmes, primarily from SMEs, rather than make them itself, followed by the implementation of the independent production quota, a statutory quota requiring 25 per cent of terrestrial broadcasters production content to be produced by independent companies. In the German case Standortpolitik was the driving force in pressing broadcasters, namely the

35 Colin Crouch et al.: Regional and sectoral varieties of capitalism 669 Downloaded by [Koc University] at 05:03 26 January 2015 WDR and the newly-established RTL, to outsource and, in particular in the case of RTL, to buy more content produced by companies located in Cologne and the surrounding area instead of from abroad. Both varieties of capitalism and neo-classical theory would expect the British liberal-market economy to be better able to respond to the changes, because of its flexibility. Today both clusters consist of large firms networked to a vast and changing number of smaller companies. Freelancers have an important role in both, and parts of the industry, like film production, rely on freelance work. Thus, project ecologies (Grabher, 2002) have been created based on shorter projectbased working periods and prolonged personal contacts. Companies cooperate in networks, which evolve around these projects and are supported by a cultural milieu, the evening economy, consisting of a lively bar scene, numerous restaurants, cafés, etc., that are in geographic proximity. Due to the rapid establishment and dissolving of companies based on projects, neither long-term-oriented credit financing by banks nor venture capital financing satisfy corporate finance needs. In film production, state subsidies and tax schemes, like sale and leaseback in Great Britain, have become important ways of financing. Contrary to expectations, local saving banks play an important role in the German cluster by providing short-term capital of the kind normally associated with venture capital (Glassmann, 2006, forthcoming). Banks were able to take up risky projects because they target their investment objects by building up intensive links to companies. With this expert knowledge, banks also assume a role close to that of a development agency, linking successful firms and thus providing services for the cluster as a whole; they hereby provide a collective competition good. On one hand, bank financing and the linking of companies are seen as typical for the German model of corporate financing. On the other, by providing a sort of venture capital, banks serve ends that are contrary to their supposed capacities in stereotypes of the German model. Also contrary to expectations, banks in Soho have built up an expert position in the cluster, specializing in knowledge about its needs, providing capital to companies and linking their financing needs to services offered (Elbing, 2006). This contrasts with the expectations of theories of liberal capitalism, in which we are led to expect financing directly via financial markets, not using the interlocutor of a bank, which has long-term relationships to companies and regards itself as a part of the cluster or even as a corporate finance house. Vocational training schemes are available in both clusters. However, institutional arrangements again defy the framework that is set out by the varieties of capitalism approach. The formal German dual system is considered inadequate for the qualification needs of the Cologne media industry, which seeks not formal knowledge but flexibility and creativity intangible qualifications (Glassmann, 2006, forthcoming). Although the state is trying to establish qualification structures and schemes, the system of vocational qualification in Cologne is not mediated by associations, as would be the case for the normal functioning within the German model. The

36 670 Economy and Society Downloaded by [Koc University] at 05:03 26 January 2015 German corporatism is absent. Instead, we find a market for a large number of public or private organizations offering vocational training or further training. In London, the situation is somewhat inverted (Elbing, 2006). Although on-the-job training is common in the industry, state-initiated projects as well as the social partners have played important roles in the establishment of qualification schemes. The Producers Alliance for Cinema and Television (PACT, representing employers) and the Broadcasting, Entertainment, Cinematograph & Theatre Union (BECTU, a trade union), have cooperated with the Sector Skills Council for the Audio Visual Industries (Skillset, a public body) to produce skill modules. PACT and Skillset have even set up funds to which production companies and broadcasters are supposed to contribute in order to pay for vocational and freelance training. Associations thus take up an important role. Not being able to rely on a formally trained workforce, in both cities companies and projects rely on freelancers who are hired on a project base and whose skills and qualifications are ensured via interpersonal networks where reputation is more important than certified knowledge (Elbing, 2006; Glassmann, 2006, forthcoming). Wages and working conditions are predominantly negotiated on an individual basis. This, not surprisingly, is framed by weak formal industrial relations, with a low union density, particularly in the German case, where ver.di (the German media industry workers union) has a density of less than 10 per cent. Following the imperatives of flexibility, the Cologne media cluster thus has a labour market that much resembles that of a liberal market economy. In London, the above-mentioned associations are developing a framework of social regulations in which questions of intellectual property rights, the employment status of freelancers and social security schemes are regulated by social partners. Conclusions Both these cases demonstrate exception 1 deviation from national model expectations in very strongly networked local contexts though neither really constitutes defiance, as they have found support from national institutions (the broadcasting corporations, banks, training organizations and trade unions). They are both cases of 2, in that both create a new and apparently vibrant organizational form by making some surprising new combinations of expected national and unexpected local institutions. The governance of the successful TV and film production sector in London s Soho area, with its considerable informality and flexibility is consistent with much of the expectations of the British model. On the other hand, its regulations and institutional solutions in the areas of corporate financing, vocational training and industrial relations, conform more to those of a coordinated model. Interestingly, it is not so much despite as because of its extreme fragmentation that this cluster needed to generate a degree of coordination if it was to thrive: its firm-level organization structures were too fragile to thrive without external institutional support. This holds true at least from the perspective of SMEs,

37 Colin Crouch et al.: Regional and sectoral varieties of capitalism 671 which profit from these alternative forms of organization, using them to the detriment of institutional alternatives, such as vertical integration of broadcasters and larger production companies. Film and TV production in Cologne, with its equally creative solutions in the areas of corporate financing, vocational training and industrial relations, similarly fails to conform to the stereotype of the embedding German model. Instead, this cluster shows many traits of a liberal market economy, though in the background familiar German institutions are providing the support role that the British case shows such firms need. Conclusions Downloaded by [Koc University] at 05:03 26 January 2015 The main results from our research are summarized in Table 1. Our main interest is concentrated on the expected exceptions, the cases expected to conform to national models being to a large extent a control group. However, there are findings of interest within these control cases. First, in both instances where German firms were expected to do well (furniture and motor vehicles) we had to come to terms with the fact that the German national model itself is not static. Second, being in a liberal market economy has not been enough to guarantee very strong success in British biopharmaceuticals. Third, a remarkable hybrid model has been developing in the British new TV industry, indicating both the relative institutional autonomy of strongly networked local Table 1 Summary of findings Cases Furniture Motor vehicles Biopharmaceuticals TV programmemaking Conforming to expected national model Germany, Ostwestfalen-Lippe: but conforms to a changing German model Germany, Zwickau: but conforms to a changing German model UK, Cambridge: but some weaknesses UK, Soho (London): but also uses distinctive local context to gain creative incoherence (elements of exceptions 1, 2) Conforming to alternative models Sweden, Tibro: divergent, even defiant local context (exception 1) Sweden, Ikea: divergent, even defiant corporate context; creative incoherence; capacity to reach outside national context (exceptions 1, 2, 3) Hungary, Győr: capacity of large firm to reach outside a weak national context (exception 3) Germany, Munich: capacity of firms to access non-national resources through market Germany, Cologne: distinctive local context facilitates creative incoherence (exceptions 1, 2)

38 672 Economy and Society Downloaded by [Koc University] at 05:03 26 January 2015 economies and the weaknesses of a liberal environment for an exceptionally flexible industry. Among the expected exceptions, we find some examples of all the anticipated sources of variation. Local environments can provide institutional support for sectors that have requirements not normally associated with national models: this is the case of furniture in Tibro and new TV programme-making in Cologne (and of course London). However, it is only in Tibro that we can really talk of defiance of national institutions. Ikea deviates in all respects. It has a distinctive environment within Sweden, which although locally rooted is really a corporate rather than a geographical environment, and in several respects defies Swedish national institutions. It does this in a way that creates a highly distinctive corporation, and is hence a creative incoherence. And it is a global firm, using characteristics and resources from across the world. VW-Audi in Hungary provides a further example of how a major corporation can create its own institutions admittedly here in a country that has not yet acquired a distinctive form of national capitalism. The inability of national institutions to form a prison for firms (assuming that they should want to do so in the first place) is also important in explaining the success of the Munich biopharmaceuticals sector. Here the resources of the locality help firms to network and learn how to access resources elsewhere in Europe and the world that are not provided on the German national scene. Finally, the Cologne new TV industry, like that in London, makes use of dense local networks to produce a creative incoherence. In our analysis, companies act rationally in response to sector-specific challenges, being partly bound by the existing institutional framework that they encounter, but partly acting to alter it. Two cases can be distinguished. In the first (structurally conservative) case, arrangements of governance in the national innovation and production system prove to be beneficial for the companies and their aim to stand up to international competition. Insofar as national institutions help companies to deal with competition on their markets, they will probably try to preserve these arrangements. The key example here is the furniture industry in Ostwestfalen-Lippe and German diversified quality production albeit in a nationally changing form. This is so far in conformity with established wisdom: If firms decide to support the regulatory regimes that sustain the comparative institutional advantages of the nation, it is because they also underpin the competitive advantages of the firm (Streeck & Thelen, 2005, p. 25). In the second (innovative) case, companies turn away from the national context and develop their own local governance structure. If the national institutional structure is seen as not adequate or non-fitting to deal with sectorally specific terms of competition, then the internal and external coordination of companies in reaction to challenges posed by the market is likely to deviate from the national structure. This is where the potential of local innovation and production systems can be found. The decoupling from the national production system takes place by a differentiation of the local (or

39 Colin Crouch et al.: Regional and sectoral varieties of capitalism 673 Downloaded by [Koc University] at 05:03 26 January 2015 corporate) level, which is thereby developing its own governance mode, or by reaching beyond the national context to access resources in the global system, either through individual corporate power (Ikea, Audi-VW) or through local networking (Munich). One aim of this article was to identify creative incoherences in the governance of local economic clusters. Competition can also be the driving force here: it dawns on companies that competing enterprises within the cluster are having greater success on markets using different institutional arrangements, which then entails imitation of their arrangements. Another possibility is the setting up of collective action capacities in the local innovation and production systems: rearranging governance becomes a cooperation project. In both cases, companies are not only rule-takers from the national institutions, but also rule-makers of the local system. Since governance of the local economy decouples itself from national institutions and develops its own institutional dynamics, local innovation and production regimes are likely to be the entry gates for institutional change. In the light of these conclusions, we propose a counter-stream model to discussion of production and innovation regimes. The starting point (first stage) is represented by the analysis of Hall and Soskice (2001a): national institutions are the independent variable, steering the actions of companies towards the path prescribed by the encompassing national mode of capitalism (liberal or coordinated). This generates specific strengths (competitive advantages) and weaknesses and predicts a sectoral specialization of entire national economies (see Figure 1). An initial counter-stream model or second stage is documented by some of our case studies: sectoral challenges are perceived as representing the independent variables, which have an influence on economic action in local clusters. Sectoral peculiarities (the terms of competition) provoke a certain strategy by companies. Exigencies of the markets lead to local institution building or, in a very different move, outreach to resources beyond the (National) institutions Firm behaviour Specialization Figure 1 First stage: varieties of capitalism national institutions determine national performance Local or global institutions Firm behaviour Changes in market niche Figure 2 Second stage: varieties of capitalism under international pressure

40 674 Economy and Society National institutions Expected specialization Creative incoherences Actions of companies Local and extranational institutions Unexpected specialization Downloaded by [Koc University] at 05:03 26 January 2015 Figure 3 Third stage: the interplay of national, local and extranational institutions and actions of companies national setting. If cultivated, the ensuing governance mode becomes a local arrangement, which provides for companies a viable alternative to the national corset (see Figure 2). If these two chains of cause and effect are put together, it becomes evident that companies act in two different environments: on the one hand, the institutional context has an impact on their actions. On the other hand, competition in their sector is continuously questioning the usefulness of this institutional context, opening up a search for alternatives. The institutional context itself thereby becomes a matter of entrepreneurial action ( institutional entrepreneurship ). This gives rise to a third model, in which national institutions, local institution-building and extranational outreach are all seen to have an effect on company behaviour (see Figure 3). The continuity of institutional arrangements is thus determined by economic interests that react on specific terms of competition, which vary according to sectors: the durability of an institution can rest substantially, if rarely wholly, on how well it serves the interests of the relevant actors. Where an institution fails to serve those interests, it becomes more fragile and susceptible to defection (Hall & Thelen, 2005, p. 7). If the national institutional context is seen as nonfitting because it ceases to grant a competitive advantage, the affected companies can make use of local or even global institutional arrangements that open up the possibility of escaping the constraints of the national context. By introducing the concept of creative incoherences to grasp the internal divergence in national varieties of capitalism, we argue that it is not only by conforming to national institutions that companies can prosper, but sometimes by diverging from them that companies, localities and even whole sectors may gain competitive advantage by establishing an institution system that is better suited to company needs.

41 Colin Crouch et al.: Regional and sectoral varieties of capitalism 675 Notes Downloaded by [Koc University] at 05:03 26 January The project, Local Production Systems in Europe, was funded by the Volkswagen Stiftung and supervised by Helmut Voelzkow with the assistance of Colin Crouch and Theo Leuenberger. Reference will be made at relevant points in this article to the work carried out by various members of the research team, who carried out extensive on-site research for the respective case studies. Their reports and some contextual chapters were collated in the report, Local production systems in Europe, presented to the Stiftung in 2006 and published as Voelzkow (2007). A reduced version of the research report is published as Innovation in local economies: Germany in comparative context (Crouch & Voelzkow, 2008). 2 In fact, regional innovation theory and Porter s theory would have no difficulty explaining how Swedish institutions can support advanced biotechnology and telecommunications as part of its dominant national model; it is only the varieties of capitalism approach that sees primarily only two forms of economy, liberal market and coordinated market, Sweden always being ranked alongside Germany in the latter category. These other theories would have to recognize as exceptional the small firms and consumer orientation of the southern Swedish furniture industry (Kjær, 1996). Indeed, the fact that this sector is partly characterized by small and medium-sized enterprises (SMEs) provides problems for nearly all accounts of the Swedish economy, which stress its domination by large enterprises (cf. Pontusson, 1997). References Brose, H.-G. & Voelzkow, H. (Eds.) (1999). Institutioneller Kontext wirtschaftlichen Handelns und Globalisierung. Marburg: Metropolis. Casper, S. (2002). National institutional frameworks and high-technology in Germany: The case of biotechnology. In J. R. Hollingsworth, K. Müller & E. Hollingsworth (Eds.), Advancing socioeconomics: An institutionalist perspective (pp ). Lanham, MD: Rowman & Littlefield. Casper, S. & Kettler, H. (2001). National institutional frameworks and the hybridization of entrepreneurial business models: The German and UK biotechnology sectors. Industry and Innovation, 8(1), 530. Crouch, C. (2005). Capitalist diversity and change: Recombinant governance and institutional entrepreneurs. Oxford: Oxford University Press. Crouch, C. & Farrell, H. (2004). Breaking the path of institutional development? Alternatives to the new determinism. Rationality and Society, 16(1), 543. Crouch, C. & Voelzkow, H. (2009). Innovation in local economies: Germany in comparative context. Oxford: Oxford University Press. Crouch, C., Le Galès, P., Trigilia, C., & Voelzkow, H. (Eds.) (2001). Local production systems in Europe: Rise or demise? Oxford: Oxford University Press. Crouch, C., Le Galès, P., Trigilia, C., & Voelzkow, H. (2004). Changing governance of local economies: Response of European local production systems. Oxford: Oxford University Press. Deeg, R. & Jackson, G. (2007). The state of the art: Towards a more dynamic theory of capitalist variety. Socio-Economic Review, 5, Elbing, S. (2006). Film- and TV-production in London: A coordinated local economy in a liberal environment? In H. Voelzkow, C. Crouch & Leuenberger (Eds.), Local production systems in Europe (Unpublished report to Volkswagen Stiftung) (pp ). Frankfurt am Main: VW Stiftung. Elbing, S. & Voelzkow, H. (2006a). Good Regional Governance der

42 676 Economy and Society Downloaded by [Koc University] at 05:03 26 January 2015 Film- und Fernsehproduktion in London und Köln: Überraschende Gemeinsamkeiten in unterschiedlichen nationalen Umwelten. In R. Kleinfeld, H. Plamper & A. Huber (Eds.), Regional governance, Vol. 2, Steuerung, Koordination und Kommunikation in regionalen Netzwerken als neue Form des Regierens (pp. 8397). Osnabrück: Universitätsverlag Osnabrück. Elbing, S. & Voelzkow, H. (2006b). Marktkonstitution und Regulierung der unabhängigen Film- und Fernsehproduktion: Staat, Verbände und Gewerkschaften im deutsch-britischen Vergleich. Industrielle Beziehungen, 13(4), Glassmann, U. (2006a). Rule-breaking and freedom of rules in national production models: How German capitalism departs from the Rhenish equilibrium. In H. Voelzkow, C. Crouch & T. Leuenberger (Eds.), Local production systems in Europe (Unpublished report to Volkswagen Stiftung) (pp. 3069). Frankfurt am Main: VW Stiftung. Glassmann, U. (2006b). Beyond the German model of capitalism: Unorthodox local business development in the Cologne media industry. In H. Voelzkow, C. Crouch & T. Leuenberger (Eds.), Local production systems in Europe (Unpublished report to Volkswagen Stiftung) (pp ). Frankfurt am Main: VW Stiftung. Glassmann, U. (forthcoming). Beyond the German model of capitalism: Unorthodox local business development in the Cologne media industry. European Planning Studies. Grabher, G. (2002). The project ecology of advertising: Tasks, talents and teams. Regional Studies, 36(3), Hall, P. & Soskice, D. (2001a). Introduction. In P. Hall & D. Soskice (Eds.), Varieties of capitalism: The institutional foundations of comparative advantage (pp. 168). Oxford: Oxford University Press. Hall, P. & Soskice, D. (Eds.) (2001b). Varieties of capitalism: The institutional foundations of comparative advantage. Oxford: Oxford University Press. Hall, P. & Thelen, K. (2005). Institutional change in varieties of capitalism. Paper prepared for presentation to the International Sociological Association, Research Committee 19 Annual Conference, Chicago, IL, 8 September. Hollingsworth, J. R., Schmitter, P. C., & Streeck, W. (1994). Governing capitalist economies: Performance and control of economic sectors. Oxford: Oxford University Press. Jong, S. (2006). The development of Munich and Cambridge therapeutic biotech firms: A case study of institutional adaptation. In H. Voelzkow, C. Crouch & T. Leuenberger (Eds.), Local production systems in Europe (Unpublished report to Volkswagen Stiftung) (pp ). Frankfurt am Main: VW Stiftung. Jong, S. (2007). The birth, dispersion, and adaptation of new institutions across institutional systems: A study of the emergence and performance of biotech firms and technology transfer officers in and around publicly funded research centres in three regions. PhD thesis. Florence: European University Institute. Kenney, M. (Ed.) (2000). Understanding Silicon Valley: The anatomy of an entrepreneurial region. Stanford, CA: Stanford University Press. Keune, M. & Tóth, A. (2006). From cathedral in the desert to institutional entrepreneur: The case of Audi in Győr, Hungary. In H. Voelzkow, C. Crouch & T. Leuenberger (Eds.), Local production systems in Europe (Unpublished report to Volkswagen Stiftung) (pp ). Frankfurt am Main: VW Stiftung. Keune, M., Kiss, J., &Tóth, A. (2004). Innovation, actors and path dependency: Change and continuity in local development policy in two Hungarian regions. International Journal for Urban and Regional Research, 28(3), Kjær, P. (1996). The constitution of enterprise: An institutional history of interfirm relations in Swedish furniture manufacturing. Stockholm: Stockholm University. Lange, K. (2005). Deutsche Biotech- Unternehmen und ihre Innovationsfähigkeit im internationalen Vergleich. In Max-Planck-Institut für Gesellschaftsforschung. Tätigkeitsbericht, 2005,

43 Colin Crouch et al.: Regional and sectoral varieties of capitalism 677 Downloaded by [Koc University] at 05:03 26 January 2015 Lange, K. (2006). Deutsche Biotech-Unternehmen und ihre Innovationsfähigkeit im internationalen Vergleich: Eine institutionentheoretische Analyse. PhD thesis. Reichsuniversität Groningen. Lundvall, B.-A. (1992). National systems of innovation: Towards a theory of innovation and interactive learning. London: Pinter. Nelson, R. R. (1993). National innovation systems: A comparative analysis. Oxford: Oxford University Press. Nelson, R. R. & Winter, S. G. (1982). An evolutionary theory of economic change. Cambridge, MA: Harvard University Press. Piotti, G. (20045). Industria automobilistica e processi si cooperazione locale: Il caso di Zwickau. Sviluppo Locale, 11(25), Piotti, G. (2006). Testing the German model in eastern Germany: The case of the automobile industry in Zwickau. In H. Voelzkow, C. Crouch, & T. Leuenberger (Eds.), Local production systems in Europe (Unpublished report to Volkswagen Stiftung) (pp ). Frankfurt am Main: VW Stiftung. Pontusson, J. (1997). Between neoliberalism and the German model: Swedish capitalism in transition. In C. Crouch & W. Streeck (Eds.), Political economy of modern capitalism: Mapping convergence and diversity (pp. 5570). London, Sage. Porter, M. (1990). The competitive advantage of nations. New York: The Free Press. Rhodes, M. & Van Apeldoorn, B. (1997). The transformation of West European capitalism? (Working paper, Robert Schuman Centre). Florence: European University Institute. Saxenian, A. (1999). Silicon Valley s new immigrant entrepreneurs. San Francisco, CA: Public Policy Institute of California. Scharpf, F. (1997). Games real actors play: Actor-centred institutionalism in policy research. Boulder, CO: Westview. Schmidt, V. (2002). The futures of European capitalism. Oxford: Oxford University Press. Sjöberg, Ö. & Rafiqui, P. (2006). Spatial differentiation in national production systems: The case of the furniture industry in Sweden. In H. Voelzkow, C. Crouch, & T. Leuenberger (Eds.), Local production systems in Europe (Unpublished report to Volkswagen Stiftung) (pp ). Frankfurt am Main: VW Stiftung. Streeck, W. (1992). Productive constraints: On the institutional conditions of diversified quality production. In W. Streeck (Ed.), Social institutions and economic performance: Studies on industrial relations in advanced capitalist economies (pp. 140). London: Sage. Streeck, W. & Thelen, K. (Eds.) (2005). Change and discontinuity in institutional analysis: Explorations in the dynamics of advanced political economies. Oxford: Oxford University Press. Swann, G. M. P., Prevezer, M. & Stout, D. (Eds.) (1998). The dynamics of industrial clustering: International comparisons in computing and biotechnology. Oxford: Oxford University Press. Thelen, K. (2004). How institutions evolve. New York: Cambridge University Press. Trigilia, C. (2004). The governance of high-tech districts. In C. Crouch, P. Le Galès, C. Trigilia & H. Voelzkow, Changing governance of local economies: Response of European local production systems (pp ). Oxford: Oxford University Press. Voelzkow, H., Crouch, C., & Leuenberger, T. (Eds.) (2006). Local production systems in Europe (Unpublished report to Volkswagen Stiftung). Frankfurt am Main: VW Stiftung. Voelzkow, H., with Elbing, S., & Schröder, M. (2007). Jenseits nationaler Produktionsmodelle? Die Governance regionaler Wirtschaftscluster: Eine international vergleichende Analyse. Marburg: Metropolis. Voelzkow, H. & Schröder, M. (2006a). Governance of the furniture industry in Ostwestfalen-Lippe and its national context: A local economy conforming to the German model? In H. Voelzkow, C. Crouch, & T. Leuenberger (Eds.), Local production systems in Europe (Unpublished report to Volkswagen Stiftung)

44 678 Economy and Society (pp ). Frankfurt am Main: VW Stiftung. Voelzkow, H. & Schröder, M. (2006b). Die Governance der Möbelindustrie in Ostwestfalen-Lippe und ihr nationaler Kontext: Ein typisch deutsches Wirtschaftscluster? In R. Kleinfeld (Ed.), Muster regionaler Politikgestaltung in Europa (pp. 951). Osnabrück: Verlag Dirk Koentopp. Womack, J., Jones, D., &Roos, D. (1991). The machine that changed the world: The story of lean production. New York: Harper Perennial. Downloaded by [Koc University] at 05:03 26 January 2015 Colin Crouch is Professor of Governance and Public Management at the Business School of the University of Warwick. He is also an external scientific member of the Max Planck Institute for the Study of Societies at Cologne and a Fellow of the British Academy. His most recent books include Social change in Western Europe (1999); (with others) Local production systems in Europe: Rise or demise? (2001); Postdemocrazia (2003) (in English as Post-democracy, 2004); (with others) Changing governance of local economies: Response of European local production systems (2004); and Capitalist diversity and change: Recombinant governance and institutional entrepreneurs (2005). Martin Schröder is a researcher at the Max Planck Institute for the Study of Societies, Cologne. His interests include political economy and economic sociology, welfare states, varieties of capitalism and research on firms. Helmut Voelzkow is Professor of Comparative Social Sciences at the University of Osnabrück. He has published within the fields of economic sociology and comparative European sociology. His recent works include Jenseits nationaler Produktionsmodelle? Die Governance regionaler Wirtschaftscluster: International vergleichende Analysen (with Sabine Elbing und Martin Schröder, 2007); and Die Europäische Union Marionnette oder Regisseur? (edited, with Patricia Bauer, 2004).

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77