Varieties Of Capitalism
For example, if firms make investments in their workers’ skills, unions gain hold-up power that can be levied against the firms. This necessitates an industrial relations systemwhere such hold-up power can be managed. Conversely, workers will be reluctant to acquire firm-specific skills unless firms can make credible long-term commitments, which require a financial system that provide access to “patient” capital, and a corporate governance system where workers are given influence, and so on. Because of these institutional complementarities, one is not likely to find every logically conceivable combination of institutions in the real world. In fact, Soskice (1999) makes the claim that there are only two dominant types: one called liberal market economies (LMEs) and another called coordinated market economies (CMEs). Each is characterized by the extent to which institutions protect and encourage investment in assets that assist firms in pursuing particular product market strategies. In CMEs where firms and workers have invested heavily in assets that are specific to particular companies, industries, or jobs, institutions are designed to protect those investments.14 In LMEs where such institutional protection is missing or weak, market competition encourages economic agents to make investments in general assets since, in the absence of protection, mobility is the best insurance against risks. This does not eliminate specific assets, but it will reduce their relative importance. The VoC argument suggests a very different explanation for the welfare state than power resources theory.Mares (2003), for example, argues that companies and industries that are highly exposed to risk will favor a social insurance systemwhere cost and risk are shared, leading employers to push universalistic unemployment and accident insurance. Although low-risk firms will oppose such spending, it is remarkable that universalism has been promoted by groups of employers since the literature associates it so closely with policies imposed on employers by unions and left governments. Manow (forthcoming) argues that social insurance systems shape the structure of production systems, and Estevez-Abe, Iversen, and Soskice (2001) and Iversen (2005) suggest that social protection (including job protection, unemployment benefits, income protection, and a host of related policies such as public retraining programs and industry subsidies) encourages workers to acquire specific skills, which in turn enhances the ability of firms to compete in certain international market segments.15 The welfare state is thus linked to the economy in a manner that creates beneficial complementaries.
This may help explain the lack of evidence for the deleterious effects of social spending on growth, and why globalization has not spelled the end to the welfare state. A mostly unexplored topic in the VoC literature is the relationship between economic and political institutions. It is striking, for example, that the distinction between LMEs and CMEs is almost perfectly collinear with the distinction between PR and majoritarian electoral systems. One possible explanation, which goes back to Katzenstein’s (1985) work on corporatism, is that PR promotes the representation of specialized interests in the legislature and its committees. At least this would be true if parties have incentives to accommodate each other’s specific interests.16 Majoritarian systems, by contrast, encourage parties to elect strong leaders in order to convince the median voter that they are not beholden to special interests. Such “leadership parties” are consequently not conducive to the protection of specialized interests and therefore encourage economic agents to make investments in more portable assets (say, college degrees as opposed to extensive vocational training). Another reason for the coupling of electoral and economic systems may be that PR serves as a credible commitment to social protection because of its effect on class coalitions and redistribution. A high level of insurance will encourage investment in risky assets and hence support a particular type of firm.17 This is a conjecture that still awaits careful empirical corroboration.
In particular, it will need to be shown that the correlation between electoral systems and production regimes is not a historical accident, but the result of a deliberate design of political institutions by representatives of particular economic interests. The question of institutional origins, of course, is a matter that concerns the entire institutionalist approach to political economy. The more successful political economy is in explaining economic policies and outcomes with reference to the institutional design, the more pressing it is to explain why one design was chosen rather than another (Thelen 1999; Pierson 2000). But the question then is how we can approach this task without being overwhelmed by the complexity of institution-free politics. In the concluding section I ask whether the answer may lie in a structuralist approach.
