Institutions, Complementarities, and Comparative Legal Development

We begin our analysis by considering the relevance for the legal origin hypothesis of the growing consensus within the social sciences that institutions matter for economic performance. In the context of comparative political economy, institutions are defined as “rules of the game” whose purpose or function is to minimize transaction costs associated with market activity (Reference NorthNorth 1990). Put slightly differently, they are “shared beliefs” that impart stability and order to economic interactions (Reference AokiAoki 2001). A first issue in this line of work is to identify the processes by which such institutions have emerged and evolved alongside the development of markets in capitalist economies. Because institutions have the character of a public good, it cannot be assumed that markets will generate them spontaneously; but since the state is composed, in its turn, of agents and coalitions pursuing their own interests and operating under conditions of uncertainty, nor can it be assumed that the public power will necessarily act in the general interest when setting and enforcing rules. Aoki has framed the issue as follows: how do the rules of the game become enforceable, when the enforcer is part of the game? The answer, he suggests, lies in an evolutionary analysis capable of showing “how the rules of the game are endogenously generated, and thus become self-enforcing through the strategic interactions of agents, including the enforcer” (Reference AokiAoki 2001:2).

This implies that institutions are the result, to some degree, of design, but also that ex post adjustment, trial and error, and experimentation play key roles in the emergence of stable forms (Reference StreeckStreeck 2005). It further follows that an understanding of more formal institutions, including the legal framework, must be complemented by an appreciation of how they interact with informal norms, social conventions, and tacit beliefs in shaping behavior. The phenomenon of “order without law” (Reference EllicksonEllickson 1991) may well be confined to a few, exceptional contexts, but, even so, the operation of the legal system in the context of market relationships is likely to be indirect at best, and to depend upon the mobilization of a range of extralegal forces in order to be effective. Enduring institutions have a “self-enforcing” and “self-sustaining” character that cannot be reduced to the formal sanctions and procedures of the law (Reference AokiAoki 2001:5).

A second focus of the new institutional research agenda is on the explanation of diversity. The persistence of different institutional forms across national boundaries, in the face of common technological and market pressures, implies the possibility of multiple equilibria or pathways to economic development. The concept of “institutional complementarity” has been suggested as an important mechanism in the persistence of diversity (Reference AokiAoki 2001:225), but the idea requires some unpacking.

Complementarities arise from the “coevolution” of institutional forms. Coevolution is the process by which adaptations in one institutional context or domain become adjusted or fitted, over time, to those in another.Footnote ² While institutions contain elements of design, the way in which they relate to one another at a systemic level is the result of an evolutionary process whose outcome, to a significant degree, cannot be planned or predicted. Thus complementarities often arise from unexpected contingencies or conjunctions. Once increasing returns set in, arbitrary or contingent initial conditions may become fixed in the manner of what geneticists, referring to certain structural features of DNA, call “frozen accidents” (Reference CrickCrick 1968), or what social scientists, using the much-cited example of the typewriter keyboard, know as “QWERTY phenomena” (Reference DavidDavid 1985; Reference DennettDennett 1995). These are sources of “path dependence,” or the tendency of systems to become locked into specific historical trajectories. By virtue of the existence of multiple pathways to economic development, cross-national diversity follows.

Complementarities persist because they possess a degree of functionality: “the performance of a configuration increases when its elements assume specific properties” (Reference HöpnerHöpner 2005:333). However, functionality does not imply optimality. There is a selective process, but because of path dependence, the rules or practices that emerge over a period of time are at best only “qualifiedly” efficient for their environments. They are not ideal, and their defects may be well known, but the costs of unraveling them and starting again from scratch either outweigh, or are perceived to outweigh, the potential gains, or at least to involve an excessive downside risk of regulatory failure (Reference RoeRoe 1995). Thus imperfect institutions can survive. But at the same time, institutions that appear fixed, because of their persistence over time, may well be destabilized, in their turn, by unpredictable external shocks, or through internally generated tensions or “incoherencies” (Reference BoyerBoyer 2005). Institutional change therefore tends to be characterized by “punctuated equilibrium,” another idea borrowed from evolutionary biology (Reference Eldredge, Gould and EldredgeEldredge & Gould 1985): periods of relative stasis give way at “critical junctures” to phases of accelerated development (Reference AokiAoki 2001:223–4).

The idea of complementarities that has just been outlined can be used to generate a theory of comparative legal evolution. This begins with the observation that market economies share many of the same legal institutions: these range from the basic forms of protection of individual and collective property rights and recognition of the capacity to contract, to more specific types of support for the business enterprise (limited liability and corporate personality), regulation of the employment relationship (labor law), provision of a welfare state (taxation and social security law), and so on. However, the particular forms taken by these institutions differ across national systems, in ways that reflect variations in the evolutionary path. More specifically, the timing of industrialization, the structure of firms and of labor unions, the degree of liquidity of capital markets, and more generally the role of the state in regulating economic life, are among the many factors that might be expected to influence the evolution of distinctive legal “varieties of capitalism” (Reference Casper, Hall and SoskiceCasper 2001; Reference Teubner, Hall and SoskiceTeubner 2001). Thus within the limits set by the market paradigm that all capitalist systems to some degree share, diversity of legal form is to be expected, just as it is for the institutions of governance more generally (Reference Hall, Soskice, Hall and SoskiceHall & Soskice 2001).

Formal divergence may, however, be compatible with a degree of functional continuity across systems (Reference GilsonGilson 2001). This has long been recognized by theories of comparative law. Functional equivalents for a given legal mechanism may be found in completely different areas of law, or outside the legal system altogether, at the level of social norms or commercial practices. It is, as the classic restatement of Reference Zweigert and KötzZweigert and Kötz (1998:39) asserts, a “basic rule of comparative law” that “different legal systems give the same or very similar solutions, even as to detail, to the same problems of life, despite the great differences in their historical development, conceptual structure, and style of operation”; in particular, “we find that as a general rule developed nations answer the needs of legal business in the same or in a very similar way” (Reference Zweigert and KötzZweigert & Kötz 1998:40). In a similar vein, Mattei, who more explicitly incorporates a law-and-economics perspective into comparative law, suggests that there is a “common core of efficient principles hidden in the different technicalities of the legal systems” (1997:144).

By virtue of functional continuity, in addition to observing complementarities within national systems, we are likely to observe functional substitutes or equivalents—institutions that substitute for one another, in the sense of performing a similar function in different ways—across such systems. An example of this is the tendency for the market for corporate control to be underpinned by takeover codes in Britain and Australia, and via a combination of shareholder litigation and securities law in the United States, which we examine in the next subsection. Substitutes may also be observed across different periods within the same national system (we provide an example below concerning industrial action and unfair dismissal law in the United Kingdom). It follows from this and other examples that the hypothesis of functional continuity is not inconsistent with the claim made above for diversity of form. It requires those who wish to argue that national systems are tending to diverge, rather than to converge, to show that legal diversity correlates to real and enduring differences in social and economic phenomena. But in the final analysis, this is an empirical question, and not one that can be settled by a priori reasoning: “[f]unctional accounts of the origins of institutional complementarity must be made compatible with historical-genetic accounts featuring ‘real’ as opposed to efficiency-theoretical causal relations” (Reference StreeckStreeck 2005:365). We therefore need to look more closely at the detail of cross-national diversity in labor law and corporate governance.

Legal Institutions as a Source of Divergence Between National Systems of Production

A comparative institutional analysis of the kind just set out can be seen to underpin the distinction widely drawn in the corporate governance field, following Reference Berglöf, Hopt and WymeerschBerglöf (1997), between systems based on arms-length or outsider control and those based on direct or insider control, a categorization that broadly corresponds to that between “liberal market” and “coordinated market” systems in the literature on “varieties of capitalism” (Reference Hall, Soskice, Hall and SoskiceHall & Soskice 2001). Corporate governance theory views the mechanisms of governance as devices for minimizing the agency costs that arise from (among other things) the separation of ownership and control in large, listed companies. The focus is on structures that are to a large degree incentive-compatible and functional; but it also is recognized that different systems can arrive at distinctive solutions to the agency problem.

In “outsider-oriented” or “liberal market” systems, dispersed ownership and market liquidity enable outside investors to diversify their holdings, thereby spreading the risk of being subject to managerial opportunism while at the same time using the capital market to hold management to account, via the mechanism of the hostile takeover bid. In different systems, different institutions have evolved that facilitate these processes. In the United States, a range of mechanisms, including shareholder litigation and an intensively regulatory regime of securities law, serves to protect minority shareholder interests (Reference CoffeeCoffee 1999). In Britain and other common law countries such as Australia, the model of the takeover code, originating in the City of London, plays a key role, and shareholder litigation is rare. This reflects, to a large degree, the collective voice exercised by institutional investors in the British context, which is not matched to the same degree, historically, in the United States (Reference Black and CoffeeBlack & Coffee 1994; Reference Armour and SkeelArmour & Skeel 2007). Shareholder litigation and takeover codes therefore appear to be substitutes in providing a mechanism for protecting minority shareholders; the presence of one means that there is less need for the other.

By contrast, in the case of “insider-oriented” or “coordinated market” systems, the concentration of ownership allows for direct monitoring and observation of managerial performance, thereby overcoming some of the agency problems inherent in the separation of ownership and control in outsider-based régimes (although this need not imply the absence of laws protecting shareholder interests, which are often quite strong in civil law countries; see Reference SiemsSiems 2005b; Reference Lele and SiemsLele & Siems 2007). Concentration or “block-holding” takes different forms, depending on context; in varying degrees, corporate cross-shareholdings, bank-led governance, and the residue of family-based control and state control can be observed (see the contributions in Reference Hopt and WymeerschHopt & Wymeersch 1997). Again, specific legal institutions have developed to complement the presence of mechanisms of direct control. One of these is the institutionalization of employee voice within the firm; this reflects a sense in which employees are regarded as a core stakeholder group contributing to the sustainability of the enterprise (Reference Rogers, Streeck and FreemanRogers & Streeck 1994). In a few systems, Germany in particular, there is a role for employee-nominated directors on a supervisory board as part of a two-tier board structure. Employee representation within company organs is by no means the general rule, however. In France, most companies have not taken up the option, provided in legislation, of having a dual board, and employee voice, while significant, mostly operates outside corporate structures (Reference Goyer, Hancké, Gospel and PendletonGoyer & Hancké 2005). In Japan, a highly integrative approach to the participation of employees in the firm almost entirely takes the form of social norms rather than legal prescription (Reference LearmountLearmount 2002: Ch. 7).

Thus while there may be a common impetus across systems for incentive-compatible solutions to the separation of ownership and control, as agency theory suggests, this cannot account for the diversity of adaptations found in different national systems (Reference AokiAoki 2001:18; Reference Aguilera and JacksonAguilera & Jackson 2003). The solutions arrived at can be understood as stable states or equilibria that are the consequence of the strategic choices made by agents—that is to say, decisions that are “rational” in the sense of being made with the aim of enhancing their interests; but the institutional forms that characterize any given system are also sufficiently durable to shape those strategic actions. Because of this “feedback loop” between structure and agency, once systems are set on their separate pathways, there is reason to expect them to continue to diverge.

When this type of analysis is extended to include labor law, further complementarities are observable. The institutions of labor law and employment relations intersect with the mechanisms of corporate governance at two levels. One, the level of the firm, is concerned with the constitution and governance of the enterprise, and with the influence of employees within managerial processes. The other, the level of the market, is concerned with how far basic employment conditions are regulated across the labor market as a whole or, alternatively, left up to individual or collective agreement in particular sectors or enterprises.

Thus in so-called liberal market systems, the predominant form of employee representation is collective bargaining between employers and trade unions. Collective bargaining operates in a manner akin to setting up a contractual mechanism for negotiation. This can be done by the employer voluntarily recognizing a particular union or unions, which is the norm in Britain, or through various regulatory mechanisms, which, as in the United States since the 1930s, have required the employer to negotiate with a certified bargaining agent that can demonstrate that it has majority support in the relevant bargaining unit. Since 2001, Britain also has had a system of compulsory recognition, based to a large extent on aspects of the U.S. and Canadian models, but as yet it plays a minor role in shaping the structure of industrial relations. Whatever the degree of state compulsion used to bring about recognition or certification, there are strict limits to how far collective bargaining can go in relation to the core areas of managerial “prerogative,” so that it stops short of co—decisionmaking or codetermination (for the United States, see Reference WeilerWeiler 1990; for Britain, Reference WedderburnWedderburn 1986: Ch. 4). Outside those areas where employers concede collective bargaining or have it forced on them by public regulation, there is no legal obligation to deal with employee representatives. In their emphasis on collective bargaining, these systems may be characterized as voluntarist.

Voluntarism at the level of the enterprise tends to go hand in hand with a partial approach to regulation at market level. Thus although both Britain and the United States have national minimum wage laws and some legislation governing basic terms and conditions such as working hours, the tendency has been for statutory regulation to impose only minimal constraints on the employment contract outside those sectors governed by collective bargaining. As collective bargaining has shrunk, since the 1950s in America and the late 1970s in Britain, so the uneven and partial character of labor market regulation has been accentuated within these systems (for the United States, see Reference WeilerWeiler 1990; for the United Kingdom, see Reference Deakin and WilkinsonDeakin & Wilkinson 1991).

Coordinated market systems, on the other hand, tend to combine an integrative approach to the role of employees in the enterprise with universalism in labor market regulation. Integration implies the incorporation of employee voice directly into the decisionmaking structures of the firm. The form this takes varies considerably across systems. In Germany and the Netherlands, employee representatives sit on supervisory boards; as we have already noted, however, this is exceptional even in mainland Europe. It is more normal for employers to be required to inform and consult employee representatives on personnel matters through enterprise-level structures of various kinds. There are variations; in France, for example, enterprise committees have only a limited say on personnel matters by comparison to works councils in Germany. However, the concept of a general obligation of consultation, which arises independently of whether the employer recognizes a particular trade union for the purposes of collective bargaining, is well established in most continental European systems, in part because of the influence of European Union (EU) standards (on information and consultation laws at the national level and their relationship to corporate governance, see the overviews of Reference Rogers, Streeck and FreemanRogers & Streeck 1994 and Reference Gospel and PendletonGospel & Pendleton 2003; on the relevant EU law, see Reference BarnardBarnard 2006).

Universalism in labor market regulation takes various forms. In Germany, sector-level collective bargaining between trade unions and associations of employers sets basic minima. The effects of agreements can be extended to non-federated employers by statutory order. In France, where sectoral bargaining is also observed, a statutory minimum wage and legislation on working time also underpin terms and conditions of employment. Italy does not have a minimum wage, but it does possess legislative and constitutional mechanisms that, by judicial interpretation, can be deployed to extend the effects of collective agreements. In each case, legal devices ensure that all sectors of the economy, more or less, are subject to labor regulation that guarantees workers certain basic protections in the form of “social rights” (Reference DeakinDeakin 1990).

Thus notwithstanding the considerable differences of legal form that exist across the systems of developed market economies, there is a case for identifying certain underlying complementarities between the mechanisms of labor law and corporate governance (see Table 1 for a summary). The prevailing form of labor regulation at enterprise level has implications for corporate governance because the degree to which employees have rights of consultation and co-decisionmaking affects mechanisms of accountability. In particular, co-decisionmaking involving employees is not easily reconcilable with the notion of outsider, arms-length control in liberal market systems. The direct line of accountability between managers and shareholders would be blurred, and the disciplinary impact of capital markets, in particular the effects of the market for corporate control, would be blunted, it is argued, by the imposition of a legal obligation upon management to consult employee representatives in advance of major corporate restructurings. It is on this basis that critics of codetermination or co-decision procedures claim that they increase agency costs (Reference Hansmann, Kraakman and KraakmanHansmann & Kraakman 2004).

Table 1. Complementarities in Corporate Governance and Labor Law

In the context of coordinated market economies, on the other hand, this more direct form of employee involvement may fit well with concentrated share ownership. Employee representatives may aid investors in the process of monitoring managers and may also bring valuable information on organizational processes to bear on the decisionmaking process, notwithstanding possible costs arising from more extended or protracted decisionmaking processes (Reference Pistor, Blair and RoePistor 1999). Employee representation may also provide a more broadly based mechanism for building trust between workers and investors and in particular for encouraging mutual investments in firm-specific assets (Reference Rogers, Streeck and FreemanRogers & Streeck 1994). Either way, institutionalized employee involvement in the firm may be said to be complementary to block-holding as a particular form of corporate ownership and control (Reference RoeRoe 1999).

At the level of the market, a universal floor of rights, taking wages and conditions out of competition, also has potential implications for corporate governance. In particular, it affects the relative flexibility or rigidity of labor costs across sectors, and therefore the scope for gains to be made through the outsourcing or fragmentation of production; it also has implications for the regulation of small and medium-sized enterprises. Thus it is possible that statutory labor standards set an implicit barrier to entry to entrepreneurial start-ups. This is a controversial and empirically unresolved question; there may be countervailing effects arising from the tendency of labor standards to encourage high levels of investment in training and skills (see Reference Armour and CummingArmour & Cumming 2004; and on the role of regulation with regard to venture capital more generally, Reference Black and GilsonBlack & Gilson 1998; Reference GilsonGilson 2003). However, the existence of functional links between universalism in labor standards, venture capital flows, and entry and exit rates for small and medium-sized enterprises is clearly an issue that merits further empirical investigation.

While the identification of complementarities in this sense appears to be a useful way of explaining varieties of capitalism, a number of questions arise. First, to what extent does the institutional environment determine the strategies adopted by firms? We should expect the strength of “isomorphic” tendencies to differ from one context to another, and for there to be a role for strategic choice by managers at enterprise level. In this vein, Reference Gospel and PendletonGospel and Pendleton (2003, Reference Gospel, Pendleton, Gospel and Pendleton2005) develop a model that seeks to identify the different influences that corporate governance may be expected to have on the approaches of firms to the management of labor. The model is summarized in Table 2.

Table 2. The Impact of Institutional Complementarities on Firm-Level Decisionmaking (Adapted from Reference Gospel and PendletonGospel & Pendleton 2003)

Even allowing for the broad degree of correspondence of institutional influences with managerial strategy that this model implies, there is abundant evidence of enterprises and sectors that go against the trend in all varieties of systems. British and American pharmaceutical firms behave very much along the lines predicted for stakeholder-oriented systems (Reference Gospel and PendletonGospel & Pendleton 2003), as do many utilities and service providers in regulated sectors (see Reference DeakinDeakin et al. 2002, Reference Deakin2006). Conversely, some German and Japanese companies have begun to adopt shareholder value metrics and the business strategies associated with them (Reference LaneLane 2003; Reference LearmountLearmount 2002). Thus it has to be emphasized that legal institutions do not rigidly dictate firm-level practices.

A possible resolution of this issue, which seeks to preserve the explanatory power of the comparative institutionalist approach, is to recognize that the principal “varieties” identified in the literature are ideal types or abstractions, which must not be confused with individual cases, whether at national or enterprise level (Reference CrouchCrouch 2005). The latter have to be studied empirically, in a way which recognizes that they consist of a number of different elements or components, the configuration of which cannot be predicted in advance; inductive approaches, such as case studies, are required to bring out their relevant features (Reference JacksonJackson 2005).

The second question posed by the comparative institutionalist approach concerns the strength of the complementarities that it purports to identify. Just how enduring are these linkages; how much of an obstacle do they present to change at national, industry, or enterprise level; and how far do they form a barrier to transnational convergence? In Germany and Japan, internal labor markets, constructed around implicit promises of job security and high levels of investment in firm-specific training, remained in place during the 1990s and early 2000s, while they have become a rarity in the private sector in United States and Britain. There is also evidence that Japanese and German companies have adjusted to the growing role of external investors and to increased capital market pressures in a way that has left intact (so far at least) the social compromises embodied in those systems (Reference JacobyJacoby 2005; Reference HöpnerHöpner 2005). Thus it is far from clear that a tendency to convergence of either form or function is being observed. It could be argued that even during a period when national systems are increasingly exposed to the effects of transnational capital flows, regulatory competition, and the growing acceptance among policy makers and business elites of a “shareholder value” norm (see Reference Hansmann and KraakmanHansmann & Kraakman 2001), governance mechanisms remain matched to local conditions and reflect particular trajectories of economic development. But, again, this is not a question that can be resolved a priori; it can only be effectively addressed through empirical analysis that is sensitive to the context and history of the cases being studied.

A third problem relates to the specific role of legal institutions. The comparative institutional approach identifies a number of mechanisms that are legal in nature—in particular, aspects of corporate and securities law that affect the rights and interests of shareholders on the one hand, and laws governing employee representation and providing workers with social rights on the other—as having a functional relationship to patterns of ownership and control at the level of the firm. However, this is far from establishing that the legal system has an independent causal influence on the emergence of particular configurations within corporate governance and labor relations. It is possible that the legal system simply reflects underlying movements within the economy, in the sense of expressing the effects of norms and conventions whose origins lie elsewhere. Even a legal system that can be thought of as lending stability to social norms, and thereby imparting to them a greater sense of permanence, would have a limited role to play as a causal factor. It is precisely here that the legal origin hypothesis seeks to break new ground, and we now take a closer look at its claims.

The Labor Index: Does Legal Origin Matter for Labor Law?

Legal origin theorists hold that that the common law or civil law origin of a given country's legal system is a major cause of its approach toward the regulation of business. According to these authors and their network of collaborators, the common law, originating in England, is associated with “decision making by juries, independent judges, and the emphasis on judicial discretion as opposed to codes” (Reference BoteroBotero et al. 2004:1344). By contrast, the civil law, as expressed in the rediscovery of Roman law by mainland European systems in the Middle Ages and the French and German codes of the nineteenth century, “is characterized by less independent judiciaries, the relative unimportance of juries, and a greater role of both substantive and procedural codes as opposed to judicial discretion” (Reference BoteroBotero et al. 2004:1344–5). It follows that

The influence of legal origin on labor regulation is placed in the context of a wider set of claims made by these authors about the effects of the common law/civil law divide. Thus they argue that common law systems provide superior protection to shareholders and creditors than civil law countries, impose lighter entry regulation, have less “formalized” dispute resolution procedures, and rely on private contracting and litigation, rather than regulation, for the enforcement of securities laws (see Reference La PortaLa Porta et al. 1997, Reference La Porta1998, Reference La Porta2000; Reference Djankov and GlaeserDjankov, Glaeser, et al. 2003; Reference Djankov and La PortaDjankov, La Porta, et al. 2005). As in other cases, the influence of legal origin on the labor market is indirect; it is mediated through the practice of regulation, or “regulatory style.” If a system has adopted a particular regulatory approach in one area, it is more likely to do so in another. In addition, the marginal cost of adopting the laws of the parent system are lower than attempting to begin anew with new methods and procedures. Thus “path dependence in the legal and regulatory styles emerges as an efficient adaptation to the previously transplanted legal infrastructure” (Reference BoteroBotero et al. 2004: 1346). Although no reference is made here to the concept of institutional complementarities, the same basic idea seems to be at work.

The method used by proponents of the legal origin approach to test their hypotheses is to construct, in each case, an index intended to measure the strength or weakness of laws on a given area in different systems. Individual country scores are then regressed against a number of possible causal factors or independent variables, including legal origin, to see how far they are correlated with them. Possible correlations with a number of economic outcomes are also measured. The coverage of the labor index, extending to both developed and developing nations, is much broader than the range of comparative country studies in the “varieties of capitalism” literature (see Reference Hall, Soskice, Hall and SoskiceHall & Soskice 2001); whereas exponents of the latter generally confine their analysis to the developed world, one of the principal rationales for the methodological approach of the legal origin school is the possibility it offers of understanding the effects of legal transplantation on the developing world.

In the labor index constructed by Botero et al., each country score is built up from a series of sub-indexes. These cover, respectively, civil rights, employment law, collective labor relations law, and social security law. The sub-indexes in turn consist of various sub-sub-indexes that are broken down further into variables that measure the intensity of legal regulation in particular sub-areas. Thus the employment law sub-index is broken down into the following subcategories: “alternative employment contracts,”“cost of increasing hours worked,”“cost of firing workers,” and “dismissal procedures.” A description is then given of the relevant variables; thus “alternative employment contracts” measures

In each case the strength of regulation is measured on a scale of 0 to 1, with 1 indicating a more protective (that is, employee-favoring) law. The various sub-indexes are aggregated to give a composite score for each country, normalized again to a 0–1 scale.

The results reported in Reference BoteroBotero et al. (2004) essentially extend the legal origin hypothesis to labor regulation. Civil law countries are found to regulate the employment contract more intensively than common law countries; the effect is less, but in the same direction, for industrial relations laws. In relation to social security, French- and Scandinavian-origin countries are more generous than common law–origin systems, but German-origin countries are not. The regressions also indicate that political power matters: countries with a long period of left-wing government during the twentieth century regulate labor more heavily, as do systems with higher union density (with union density being taken as a proxy for the strength of organized labor as a pressure group). However, the effect is not as strong as for legal origin, leading Botero et al. to argue that “the effects of legal origin on the regulation of labor are larger and different from those of politics” (2004:1371) and thereby to reject the argument of Reference RoeRoe (1999, Reference Roe2006) to the effect that political forces are the main driver of cross-national diversity. When the analysis is extended to outcomes, evidence is presented to show that higher scores on the labor index are associated with lower levels of male labor force participation, higher youth unemployment, and a larger unofficial economy. On this basis the authors conclude that labor regulations do not, on the whole, operate to cure market failures; instead, “[t]he results are consistent with the view that legal origins shape regulatory styles, and that such dependence has adverse consequences for at least some measures of efficiency” (Reference BoteroBotero et al. 2004:1378). On this basis they reject the claim that institutions tend, on the whole, to be efficiently matched to underlying economic conditions within national systems.

To evaluate these findings, it is necessary to consider the methodological status of the labor index and what it is able to tell us about complementarity, efficiency, and diversity.

What Is the Labor Index Really Measuring?

The sources of the values attached to the variables in the labor index are simply stated as “the laws of each country” in the dataset, as they stood in the mid-1990s (Reference BoteroBotero et al. 2004: Table I). These are supplemented by a number of cross-country secondary sources, including the International Encyclopaedia of Labour Law and Industrial Relations, the International Labour Organization (ILO)'s Conditions of Work Digests, and the U.S. Social Security Administration's Social Security Programs Throughout the World. On the face of it, then, the labor index measures “law on the books” rather than the impact of legal regulation in practice. Indeed, Botero et al. accept that what they are essentially measuring is “formal legal rules” (2004:1347).

If the gap between formal law and law in practice affected all countries equally, this would not necessarily pose a problem for the indexing methodology; if, on the other hand, civil law systems “wrote down” more of their regulations, with the common law relying on judicial decisions that were not reflected in statutes, there would be a problem. However, Botero et al. reject this critique, for the following reasons:

The claim that labor law is mostly statutory is open to question. Labor legislation is not self-enforcing; courts, from specialist employment tribunals up to appellate and constitutional bodies, play a pivotal role in the development of the substantive rules of labor law, even when these rules have a statutory origin. However, in defense of the methods used and argument deployed by Botero et al., it can be argued, as we see below, that judge-made law is an important source of labor law in both the common law and civil law systems. If their approach suffers from a methodological limitation in focusing on the written law, it is not obvious that this biases their findings for or against one of the principal families of legal systems.

What of their second claim that the index reflects real burdens and real costs? What this appears to mean is that in the construction of particular variables, Botero et al. have attempted to infer the effects of laws upon employer flexibility (or, conversely, worker protection). Thus if the law on dismissal, for example, restricts the grounds upon which termination may take place, the employer's freedom of maneuver is that much less than if it permits termination merely upon the giving of notice.

This is a more serious difficulty. The form taken by a particular law may not tell us much at all about the particular degree of flexibility available to employers. That may depend on factors outside the scope of the labor index, including social and cultural norms beyond the law that govern dismissal.Footnote ³ The social or economic effect of a given legal rule can only be understood by seeing legal rules as part of a system of interlinked norms, some of which are extralegal in nature. As we have seen, this is a basic aspect of the method of comparative law developed by, among others, Reference Zweigert and KötzZweigert and Kötz (1998).

This point can be illustrated with an historical example. Dismissals without notice were often extremely costly to British employers during the 1950s and 1960s, when union strength in the workplace was at its height and sackings were met by industrial action, which was often spontaneous or “unofficial” in character. This was before the introduction of unfair dismissal law (see Reference Deakin and MorrisDeakin & Morris 2005:386–95 for an overview of the emergence of dismissal law in the United Kingdom). At that point, therefore, the United Kingdom would have scored a very low mark (a zero, presumably) for “costs of dismissal” when in practice the situation was quite otherwise. Conversely, the introduction of dismissal legislation in the 1970s and 1980s would have made the law on dismissal look more stringent; but in practice, as union strength was declining, social norms against dismissal were weakening, in part because of the very same unfair dismissal law, which was designed to take disputes out of the workplace and subject to them to a legal procedure that some commentators argued “legitimized” dismissal (Reference CollinsCollins 1995).

This example illustrates the enormous complexity involved in preparing any labor law index. Knowledge of national conditions should qualify the scores attached to particular country variables. However, this magnifies the scale of what is already a huge task.

In this respect, the sheer breadth of the labor index could be said to work in its favor. Because it includes such a wide range of variables, it is possible to argue, in its defense, that the offsetting effects of social and cultural norms are going to be captured at some level. Social sanctions such as the effects of an unofficial strike do not exist in a vacuum; they are affected by the framework of laws governing strikes. Thus if there were a time-series index for the United Kingdom, the decline in the effectiveness of the unofficial strike as a mechanism of job protection in Britain in the 1970s and 1980s might be caught by the falling value of the variable measuring the scope of the right to strike during this period (for an index measuring the fall in protection for the right to strike in Britain, illustrating the potential value of an indexing methodology in this context, see Reference Freeman and PelletierFreeman & Pelletier 1990).

Nevertheless, serious difficulties remain. The indexes represent unweighted measures; we do not know how significant each one is in its contribution to the overall labor market environment. Crucially, for a comparative study, within different countries the significance of particular measures may also differ. This follows from the existence of functional equivalence across systems (Reference Zweigert and KötzZweigert & Kötz 1998). For example, the mandatory requirement that companies bargain with unions or work councils shapes a great deal about the labor environment in Germany, while the lack of this particular written law may be less significant to shaping the labor environment in Japan, where a different style of consensual company politics prevails, making it less necessary for the formal law to intervene (Reference LearmountLearmount 2002).

One possible answer to this objection is that weightings for cross-national indexes are extremely difficult to determine, and that an unweighted index might be less biased than one based on subjective attempts at weighting. The sub-indexes constructed in the labor index are arguably most effective without an attempt at “even factoring” to reduce the values ascribed to laws in particular countries, as this might mean sacrificing some of the carefully constructed detail within the subcategories.

Some of the criticisms just made would affect any index of this kind. Thus another possible approach is to ask how the index of Botero et al. compares to its main rivals. One such index is the Global Competitiveness Report, an annual publication of the World Economic Forum detailing the perceptions of experts and executives on the economic and political environment for business across 50 countries, which surveys more than 3,600 experts globally to assemble more than 200 indicators (World Economic Forum various years). For example, the 2002–3 survey asked respondents, among other things, how flexible or inflexible labor regulations were, and how productive or hostile labor relations were, on a scale of 1 to 10, for the country in which they were resident. Countries were then ranked, based on their average score.

Then there is the work of the Organisation for Economic Co-operation and Development (OECD), including its 1994 report on Labor Standards and Economic Integration. Here the OECD attempted to find a cross-national system for classifying information and measuring stringency surrounding labor and working regulations in OECD countries (OECD 1994). To this end, it considered country ratification of ILO conventions on the existence of minimum wages, wage protection, child employment, employment protection, working hours, weekly rest, paid holidays, and occupational health and safety in 20 countries. These indicators are useful because international conventions are necessarily comparable by country (unlike some nationally originating regulations, which may be incomparable across countries). In addition, they cover a wide and significant area of labor regulation and protection. Finally, these data have been collected for a significant number of countries.Footnote ⁴

In addition, there are various secondary indicators of the impact of legal institutions. For example, we could regard average wage levels as providing some information about the stringency of the legal environment for labor; or consider that union membership offers some information about the protections afforded in the legal environment surrounding collective action and industrial relations. The difficulty is in knowing to what extent the variable concerned is a valid proxy for the effects of the law. Many factors apart from legislation influence wage levels; and wage levels are determined to a greater extent by government legislation in some countries than in others. Similarly, union membership may have determinants other than industrial relations law.

To test the validity of the employment law index, we correlate it with other measures of employee protection, discussed above, for which sufficient cross-national data are available. In particular, we correlate the employment law index separately with the perception of the stringency of labor regulation as reported in the Global Competitiveness Report, and the number of ILO conventions ratified by country (as reported by the OECD). The results are displayed in Figures 1 and 2. The employment law index is highly correlated with managers' and experts' perceptions of the legal environment in countries for which data are available (in other words, the higher the score on the employment law index, the less “flexible” the rating given to labor regulation by managers and experts by country).Footnote ⁵ In addition, the employment law index is correlated, although less strongly, with international indicators of the stringency of labor regulation, as expressed in the number of ILO conventions ratified by various countries.

Figure 1. Employment Laws Index and Ratification of ILO Labor Conventions

Figure 2. Employment Laws Index and Perceptions of Labor Standards

This suggests that the Botero et al. index does capture, to some extent, the way in which labor law is perceived to operate. But quite apart from the question of whether the individual country scores attributed by Botero et al. to particular variables enjoy universal assent (and there is considerable scope for argument here), there are important aspects of labor regulation that are not well captured by their index, in particular the role played by functional equivalents to regulation beyond the formal law. At best, the labor index constructed by Botero et al. is a proxy for the real impact of laws. The issue is how good a proxy it is. It corresponds quite well to a number of alternative indicators that are based on perceptions and opinions of lawyers and managers, but these too are only proxies for the law's impact. Legal and social norms affect one another in complex and subtle ways that cannot be captured by an index that measures formal law, no matter how carefully constructed. Qualitative and case-study approaches remain essential if the results of the indexing approach are to be put into the context of particular national settings. However, institutional detail of the kind that would be needed to validate the scores provided in the index is entirely lacking from the Botero et al. analysis.

Complementarity, Efficiency, and Diversity

As we have seen, Reference BoteroBotero et al. (2004) make a wide claim for complementarity between legal origin and modes of economic regulation, with implications for economic outcomes. Legal origin, they argue, shapes institutions independently of both efficiency and politics. They reject an efficiency-based explanation for the form of institutions because, as we have seen, after controlling for differences in wealth between countries (for which average years of schooling are taken as a proxy), they find that the intensity of labor regulation is, in certain respects, positively correlated with a larger unofficial economy, lower male participation in employment, higher youth unemployment, and higher relative wages for “insiders.” If institutions were efficiently matched with economic conditions, these “inefficiencies,” they suggest, could not persist. They also reject the hypothesis, associated with Reference RoeRoe (1999, Reference Roe2006), that politics is the main driver of regulation, on the grounds that the political composition of governments and variations in union density (a proxy for worker influence) can only explain a limited extent of the divergence between systems that they observe.

The claim that civil law systems are inherently less supportive of market institutions than common law ones was made initially by Reference HayekHayek (1960) and has more recently been revived by Mahoney: “[t]here are structural differences between common and civil law, most notably the greater degree of judicial independence in the former and the lower level of scrutiny of executive action in the latter, that provide governments with more scope for alteration of property and contract rights in civil law countries” (2001:505). If legal origin had this effect, we would expect there to be differential growth rates for common law and civil law countries. However, this claim simply cannot be maintained, even if, for the moment, the methodological limitations of the indexing method are set to one side.

Reference BoteroBotero et al. (2004) are careful not to argue that legal origin causes differences in GDP growth; in their analysis, GDP is not a dependent variable but an exogenous factor for which they attempt to control, as we have just seen. Reference MahoneyMahoney (2001), on the other hand, claims to find evidence of a relationship between common law origins and faster rates of real growth in GDP per head in a sample of developed and developing economies between the 1960s and the 1990s. However, if developed nations alone are considered, this relationship disappears. Reference Hall, Soskice, Hall and SoskiceHall and Soskice (2001:21) show that coordinated market systems, all of which have civil law origins, enjoyed faster economic growth than liberal market regimes, all of which have common law origins, in the 1960s and 1970s, and that the two groups had roughly the same growth levels from the mid-1970s to the mid-1980s. The position was reversed between then and the late 1990s, but at the end of this period (which is the relevant point in time for the construction of the LLSV indexes) GDP per head was still slightly higher, on average, in the coordinated market systems.

In other words, systems with civil law origins seem to have been just as successful as common law systems, among developed economies, in delivering economic growth to their citizens for most of the post–World War II period. This strongly suggests that institutions approximately reflect particular national conditions and trajectories, and that there is no uniquely successful or predominant route to development.

If there are negative effects of civil law origin, they seem to be confined to developing systems. Here, proponents of the legal origin hypothesis offer an argument based on historical contingency. What they identify as the civil law orientation toward centralized state control of the economy may, they suggest, have been efficient in the mainland European (or, to be even more specific, French) context in which it originated, but it gives rise to inefficiencies in the context of the transplantation: “[w]hen a civil law system is transplanted into a country with a ‘bad’ government, it will lead to less secure property rights, heavier intervention and regulation, and more corruption and red tape than does a common law system transplanted into a similar environment” (Reference Glaeser and ShleiferGlaeser & Shleifer 2002:1221).

However, the methodological foundation for this claim is weak, in particular in the context of labor regulation. The dataset developed by Reference BoteroBotero et al. (2004), since it only measures “formal law,” is likely to give an incomplete picture, at best, of the impact of labor law in developing nations, where there is likely to be a significant gap between the legal text and its enforcement. Moreover, the dataset measures the law as it applies to a “normal” employment contract, which is full-time and of indeterminate duration (Reference BoteroBotero et al. 2004:1353); it therefore has little relevance in systems with large informal economies, where very few workers meet this description. This is perhaps why Reference BoteroBotero et al. (2004:1378) find that the link between legal origins and inefficiencies is not borne out when the sample is confined to countries with per capita income below the median. They take this as a further indication of the weakness of the “efficiency” theory, on the grounds that regulation has the most deleterious effects when it is enforced, but it may also reflect the inherent and unavoidable limitations of their dataset in the context of developing systems.

We offer an additional test of the claim that divergence between the common law and civil law locks in inefficiency by considering an alternative dependent variable to those used by Reference MahoneyMahoney (2001) and Reference BoteroBotero et al. (2004), namely, the Human Development Index (HDI). The HDI has been developed as a comprehensive measure of human capabilities by country. It is a composite of indicators on life expectancy at birth, adult literacy, and adjusted per capita income in purchasing power parity (UNDP various years). According to the United Nations Development Programme (UNDP), “[t]he HDI measures the average achievements in a country in three basic dimensions of human development—a long and healthy life, knowledge and a decent standard of living” (UNDP 2000:17). Largely given impetus by the work of Sen on human capabilities (Reference SenSen 1999), the HDI is increasingly widely cited and used. Information for the index has been gathered since 1990 for an increasing number of countries (almost all are covered in the more recent reports) and is published annually in the UNDP's Human Development Report.

We conduct a simple correlation analysis considering a number of legal variables first introduced by Reference La PortaLa Porta et al. (1997) and developed further in subsequent papers, correlated against the HDI for 1999 (which is roughly contemporaneous with most of the LLSV indexes, which use legal and other data from the mid- to late-1990s). Table 3 displays the results. First, when we consider the legal origin clusters (English, German, French, and Scandinavian), the only cluster with a significant positive relationship to human development is the Scandinavian one. Second, employment and industrial relations laws bear no relationship to HDI, but social security law has a highly positive relationship. Third, most of the corporate governance measures developed by LLSV bear no relationship to HDI (but some, such as protection of creditor rights, are strongly negatively correlated). Again, this analysis tells us nothing about causation, but when set against the regressions run by the legal origin authors, it offers a telling comparison. For an increasingly accepted and particularly comprehensive measure of human well-being, legal origin might matter, but not in the way that legal origins scholars predict. There may be a story to be told, as far as human development is concerned, about social security protection. Scandinavian countries, well known for their comprehensive social security systems, produce significantly higher HDI outcomes than the other legal clusters. Countries with high scores on the social security law index in general score highly on the HDI.

Table 3. Correlations Between the HDI and Various Legal Origin Indicators

^* p<0.05 level (2-tailed)

^** p<0.01 level (2-tailed)

Sources: Human Development Index: United Nations Development Programme, Human Development Report (1999).

Legal origin indexes: Reference La PortaLa Porta et al. 1997, Reference La Porta1998; Reference BoteroBotero et al. 2004.

To sum up: the methodological limitations of the legal origin approach make it illegitimate to use that mode of analysis to draw conclusions about the impact of legal systems on economic development. There is no reliable basis on which it can be said that common law institutions are more conducive to economic growth than their civil law counterparts, and the more careful analyses of the legal origin school do not in fact claim that there is such a link. To the extent that it serves as a proxy for the social and economic effects of laws, the labor index could be used to throw light on the welfare and efficiency effects of regulation and to provide some indications of the likely strength of national-level complementarities between labor law and corporate governance. However, the pointers provided by the legal origin approach would need to be supplemented by more detailed institutional accounts of individual country systems for the findings to be regarded as reliable.

What of the wider claim of the legal origin school, to have identified a causal link running from legal origin to regulatory style and economic outcomes? Here the central problem is the mechanism by which legal origin plays the role attributed to it. As we have already suggested, strong path dependencies must be at work if the adoption of one or another of the main legal families at some point in the nineteenth century was still having a preponderant influence on the substance of law at the end of the twentieth. However, the method relied on by the legal origin school is not well equipped to throw light on this question. Legal origin dummy variables, particularly when used in cross-sectional analyses, are crude instruments that may pick up any number of social, political, and historical factors; they can only give us part of the picture. It is time to shift focus, toward an historical explanation for the origins of complementarities.

Historical Origins of Complementarities: The British Experience of Industrialization

Although it is possible to go much farther back in the search for the origins of modern legal diversity (see Reference Glaeser and ShleiferGlaeser & Shleifer 2002), there is general agreement that the period of industrialization in the eighteenth and nineteenth centuries was a formative one in the emergence of modern legal and economic institutions (Reference PistorPistor 2005:8). Today's legal forms are sometimes thought to be the outcome of a selective process that accompanied industrialization, based upon the gradual discarding of institutions without functional value. Some law and economics scholars claim, for example, that modern institutions of corporate governance can be understood as a functional response to the problems of asymmetric information that arise in relations between managers and investors:

But it is difficult to square this essentially “teleological” argument with historical analyses pointing to the contingent circumstances that accompanied the rise of the legal form of the joint stock company in England around the time of the Industrial Revolution. In the eighteenth century, at the point when economic growth was beginning to accelerate, several different legal forms were available for business organization: these included incorporated trading companies established by state charter, along the lines of the East India Company; trusts providing for the management by specialized agents of property owned beneficially by others; and so-called unincorporated companies that were essentially partnerships linking together merchants, managers, and investors (Reference HarrisHarris 2000). The approval of the state was needed to set up a corporate entity with separate personality. In other cases, trading as a corporation could attract criminal liability either at common law or under the 1720 Bubble Act (6 Geo. 1, c. 18), which although not originally enacted with the aim of suppressing corporations acquired this effect for a while through judicial interpretation in the 1800s and 1810s. Joint stock was in use in certain sectors as a means of spreading risk, but it did not normally confer limited liability on investors. The partnership form of most of the early textile and mining enterprises meant that there was a limited degree of integrated management across different industrial sites (Reference PollardPollard 1965).

It was as late as 1844, after repeated lobbying efforts, that the United Kingdom Parliament made incorporation through a registration procedure generally available, and in 1855 and (more completely) in 1856 that it attached limited liability to this new corporate form. Far from being in the vanguard, English law lagged behind a number of other jurisdictions, including France and the United States. One of the major factors behind the adoption of the 1855–56 Acts was the perception that increasing numbers of English-based enterprises were being incorporated overseas in order to take advantage of more amenable legal environments—an early instance of regulatory competition in corporate law (Reference SavilleSaville 1956). But by the middle of the nineteenth century, “the tide had turned, and English company law became the model for Europe” (Reference HarrisHarris 2000:289), with French and German legislation of the 1860s and 1870s copying key features of the English system; it also had some influence on U.S. state-level legislation. On the face of it, then, the dissemination of the basic corporate law model owed much to interest group pressure, regulatory competition, and transplantation through imitation, the very forces identified today as powerful mechanisms for convergence (Reference Hansmann, Kraakman and KraakmanHansmann & Kraakman 2004:5); but in the country of its origin, its relationship to the emergence of an industrial economy was anything but straightforward.

Does any of this matter for today's company law? If company law is the product of selective pressures ensuring, through various means, that rules that do not fit with their environment are selected out, the answer must be no; a smooth progression ensures that function and form are matched. If, on the other hand, the joint stock company was only one of a number of legal forms that could have met the requirements of business, some of which might have done just as well or better, its contemporary preeminence is yet another illustration of the effects of institutional lock-in. An historical perspective on this issue, suggested by Harris's detailed analysis, is that

Thus from a “genealogical” perspective, it is the original conditions surrounding the emergence of the corporation that are of importance for the development of British capitalism, and not the supposedly functional adaptation of legal forms to economic development. Because the first phase of the Industrial Revolution in Britain preceded the general availability of the corporate form by several decades, it is not possible to ascribe Britain's early lead in industrialization to its system of company law. Even following the enactment of limited liability in 1855, whether through inertia, suspicion of the new legal regime, or otherwise, the majority of mining, engineering, and textiles companies were slow to incorporate (Reference Holbrook-JonesHolbrook-Jones 1982). It was only at end of the nineteenth century that incorporation became the norm in these sectors, at the same time as a wave of consolidations and mergers occurred (Reference HannahHannah 1983).

The comparatively late emergence of the legal infrastructure for enterprise in Britain is also evident in the law governing the employment relationship. The British Industrial Revolution was not marked by a straightforward move from household production to factory labor. Workers resisted the factory in large part because many of the early factories were centered around compulsory workhouse labor for those receiving poor relief, or they copied the same model (Reference PollardPollard 1965). Not only did “putting out” survive well into the nineteenth century in numerous industries, but even when production was brought in-house, an “internal contracting” system was maintained under which employers normally dealt with labor intermediaries who, in turn, hired other family members or “underhands” to work for them (Reference LittlerLittler 1982). Partly in response to worker resistance, harsh disciplinary laws were passed to control labor. Breach of the service contract was increasingly criminalized in the eighteenth and nineteenth centuries. In one sense, these laws were indeed “functional” to the “needs” of employers at the time; they were deployed most frequently against craft-based workers who retained a high degree of independence and could command a premium based on the scarcity of their skills. At a point when few manufacturing firms had integrated managerial structures, the criminal law was used as an alternative means of underpinning employer prerogative. Up to the 1870s, there were thousands of prosecutions a year in the industrial heartlands of the north and midlands of England; prosecution rates were driven by the business cycle, increasing in the upturn when labor supply was at its most restricted (Reference Simon and SavilleSimon 1954). The disciplinary emphasis of “master and servant” was to have a formative influence on employment law (Reference Deakin and WilkinsonDeakin & Wilkinson 2005), on the development of collective bargaining (Reference Holbrook-JonesHolbrook-Jones 1982), and on managerial practice:

Like limited liability, the master-servant model was also exported; as a by-product of colonization, it spread to most of the common law world in the course of the nineteenth century (Reference Hay and CravenHay & Craven 2004) and also had an influence upon American law during the early phases of industrialization (Reference Tomlins, Hay and CravenTomlins 2004). The effects of this particular exercise in transplantation have been slow to wear off. Long after the development of integrated organizational forms and modern management techniques reduced the need for the penal enforcement of employment contracts, the legacy of the master-servant model is still discernible within the contractual form of the modern employment relationship (for the United States, see Reference AtlesonAtleson 1983; for Australia, see Reference MerrittMerritt 1982; Reference Howe and MitchellHowe & Mitchell 2001; for England, see Reference FoxFox 1974; Reference Deakin and WilkinsonDeakin & Wilkinson 2005). The reluctance of organized labor to support the use of legal means to underpin labor regulation, and the resulting preference for voluntarist solutions outside the law, principally in the form of collective bargaining, owed much to a deep-rooted belief that the law was an instrument of the employer “class” (see Reference WedderburnWedderburn 1986:16–47).

Thus the case of British industrialization illustrates the limits of a functional analysis of the evolution of the law. There were far-reaching changes in corporate and labor law in the nineteenth century that accompanied the emergence of an industrial society. However, legal institutions were in many respects not particularly responsive to economic needs and, conversely, even when they were modified in response to perceptions of what those needs might be, as in the case of the limited liability reforms of the 1850s, industry, in its turn, was slow to respond to the new possibilities created by the law. Thus the notion that the law incrementally evolved by way of adaptation to the needs of the economy is impossible to maintain. It would be more accurate to say that legal and economic development, while moving in a broadly coevolutionary dynamic in which each one influenced the other, were out of synch for most of the time, and that the process of mutual adjustment was uneven and sporadic, and often had unexpected consequences: an example of punctuated equilibrium.Footnote ⁶

Status and Contract in the Origins of Continental European Labor Law

Thus the characteristic common law notion of the enterprise as the unencumbered property of the employer, with the workers relegated to limited claims, at best, on the surplus from production, owes more to the initial conditions of British industrialization than it does to the supposedly universalizing economic logic of agency theory. There is evidence that the nature of enterprise was understood differently in continental European systems from the very early stages of industrialization (see Reference BiernackiBiernacki 1995). The continental European entrepreneur of this period “had a different conception of his role from the British” one; in societies with “a strong feudal and manorial tradition,” factory owners saw themselves as having “duties as well as the privileges that such a position entails” (Reference LandesLandes 1969:191). This type of industrial paternalism was not simply a legacy of preindustrial practices, however; it owed much to the scarcity of the labor supply in systems where the rural population continued to have access to the land, and hence to alternatives to waged employment, long after this had ceased to be the case in Britain (Reference O'BrienO'Brien 1996). In addition, attitudes on the Continent were reinforced by “public and official opinion”: “[i]n [nineteenth-century] France, the government was sensitive to factory unemployment, keeping watch on hiring and firing and utilizing political pressure when necessary to limit the number of jobless, even in—or rather especially in—severe crises” (Reference LandesLandes 1969:192).

There is also evidence that the legal transition toward a market order played a role in shaping continental European attitudes toward the business enterprise. The various “integrative” conceptions of the enterprise that continue to influence civil law systems today have deep, historical roots; but contrary to the claim that the civilian systems operate by reference to “regulation” in preference to “contract,” a contractual logic played a major role in shaping the law from the beginning.Footnote ⁷ The starting point is the process by which the emerging forms of wage labor were grafted onto the traditional Roman law concept of the locatio conductio in the post-revolutionary codes. In relying on the model of the locatio, the drafters of the codes were grouping work relationships with other types of contracts, the effect being to stress that, in common with them, they were based on exchange. Thus labor, or in some versions labor power—as expressed, for example, in the German term Arbeitskraft—became a commodity that was linked to price (not necessarily the “wage”), through the contract. The further consequence was to align the work relationship with the law of things rather than the law of persons: the notion of the personal “subordination” of the worker was absent from the formulae used by the early codes (Reference Veneziani and HeppleVeneziani 1986; Reference Simitis and SteinmetzSimitis 2000). Thus the codes helped propagate a strongly contractualist notion of the work relationship, at a point when English law and practice were still dominated by the almost “pre-industrial” notion of service: “[t]he codes inspired by the revolutionary ideology of 1789 put contractual analysis at the centre of the juridical conceptualization of the work relationship” (Reference SupiotSupiot 1994:14).

This basic contractual model was accommodated to the growth of industrial and labor legislation in the course of the nineteenth century in two distinct ways. In the French-origin systems, the power of the state to regulate conditions of work was instantiated within the legal system through the concept of ordre public social, that is, a set of minimum, binding conditions that applied as a matter of general law to the employment relationship. The implicit logic of this idea was that in recognizing the formal contractual equality of the parties to the employment relationship, the state also assumed, by way of symmetry, a responsibility for establishing a form of protection for the individual worker, who was thereby placed in a position of “juridical subordination.” In German systems, by contrast, a “communitarian” conception of the enterprise qualified the role of the individual contract. This approach was summed up at the end of the nineteenth century by the (conservative) jurist Gierke's argument that the “eternal juridical truths” of the modernized Romanist tradition simply served to conceal “formulas expressing individualistic and capitalistic assumptions” (Reference GierkeGierke 1895:32). Under Gierke's influence, the employment relationship was oriented away from the law of obligations and toward the law of persons; thus in contrast to the French approach, German law came to recognize the “personal subordination” of the worker in the form of “factual adhesion to the enterprise” or Tatbestand, a process that conferred “a status equivalent to membership of a community” (Reference SupiotSupiot 1994:18).

It has been argued by Supiot that “there is no European country in which the conception of the employment relationship has not been influenced to some degree by each of these two legal cultures, the Romanist and the Germanic” (Reference SupiotSupiot 1994:19). The influence of communitarian thinking was particularly strong in all continental systems, French-origin included, in the first part of the twentieth century, when it overlapped to some degree with fascist ideologies. But it would be excessively reductive to identify communitarianism exclusively with authoritarian notions of the corporative state. The concept of the “interests of the enterprise,” not just as a reference point for defining the mutual obligations of employer and employee but also as a basic principle of company law, is based on an integrative model of the firm that predated the rise of authoritarian regimes and retained an influence after their fall. The result was a “synthesis” of contractual and communitarian elements that has ever since been a source of “structural ambivalence” in the conceptual framework of continental labor law (Reference SupiotSupiot 1994:32).

From this necessarily brief overview, it is clear that the origins of continental labor law were complex. Contrary to the argument of Reference BoteroBotero et al. (2004), the civil law approach cannot be characterized as more regulatory than that of the common law. Thanks to the liberalizing influence of the French Code civil, legal innovation preceded economic development: nearly all the continental systems, including those influenced by the German tradition, acquired a liberal-contractual model of the employment relationship not just in advance of the English common law but, in marked contrast to the British experience, before industrialization affected more than a tiny proportion of the labor force. As industrialization advanced, legal systems responded to the growing wave of labor regulation in ways that incorporated the principle of worker protection while at the same time recognizing the primacy and legitimacy of capitalist modes of economic organization. Thus the civilian approach was different from that of that English common law, but not in the way that the authors of the legal origin hypothesis suggest: regulation was not preferred to contract, it was conjoined with it.

Legal Cultures as “Carriers of History”

What we have just been describing is not simply the development of different systems of substantive rules, but the evolution of distinctive legal cultures. Comparative lawyers have used the terms legal culture and legal style in a number of different ways (Reference Zweigert and KötzZweigert & Kötz 1998; Reference MarkesinisMarkesinis 1994; Reference LegrandLegrand 1999; Reference BellBell 2001; Reference SamuelSamuel 2003; Reference KaganKagan 2003; Reference PistorPistor 2005). However, there is general agreement that legal systems generate commonly understood “ground rules” or shared assumptions, as aids to the interpretation of the law. Thus the meaning attributed to superficially similar legal texts may differ from one jurisdiction to another according to the prevailing juridical style or underlying assumptions guiding legal interpretation. Relevant here are the conceptual modes of thought through which legal discourse creates “epistemological maps” (Reference SamuelSamuel 2003) that describe and categorize economic and social relations. In the terms used by comparative institutional analysis, these cultural reference points are a form of “information compression embodied in an institution [which makes] it possible for boundedly rational agents to efficiently collect and utilize the information necessary for their actions to be consistent with changing internal and external environments” (Reference AokiAoki 2001:14). As such, they are “carriers of history” through which the cultural information of earlier periods is passed on (Reference DavidDavid 1994). At the same time, the conceptual discourse of the legal system, although separate from that of politics or commerce, is also (indirectly) shaped by the social and economic forms alongside which the law has coevolved (Reference FögenFögen 2002). Thus diversity in the experience of industrialization is embedded in the legal forms that have developed to describe the business enterprise and the related economic institutions of market economies. The influence of legal origin on the present-day substance of labor regulation is the result.

If this is the case, we have a rather different explanation for the significance of the law for economic development than that provided by the legal origin school. As we have seen, they have argued that common law systems have an inherent tendency to produce rules that are market-compatible, by virtue of their greater reliance on judicial adjudications and other decentralized forms of law making, in contrast to the civil law, which relies to a greater extent upon centralized regulation and thereby gives rise to greater potential for governmental interference with property and contract rights. The main problem with this line of argument is that it is not an accurate description of the common law/civil law divide, either in general terms or in the specific context of labor regulation. The idea that common law judges have discretion to shape rules to changing economic circumstances, while civilian judges are bound to apply, through rigid deductive logic, the strict legal text of the code is, as Mattei has shown, “dramatically misleading, being based on a superficial and outdated image of the differences between the common law and the civil law” (1997:79). Arguments about whether judicial decisions are a formal “source” of law in civilian systems aside, the prominent role of judicial decisionmaking in the civil law is now clearly understood (see Reference MarkesinisMarkesinis 2003). Notwithstanding the efforts of the drafters of the French civil code to limit judicial influence and curb the doctrine of judicial precedent, “neither before nor after the French codification could any of the civil law systems be fairly characterized as the one described by the French post-revolutionary scholars” (Reference MatteiMattei 1997:83). Many of the doctrines thought to be most characteristic of a distinctive civilian approach to economic regulation, such as the application of the concept of good faith to commercial contracts, were judicial innovations (see Reference Teubner, Hall and SoskiceTeubner 2001; Reference PistorPistor 2005).

When the sources of labor regulation, specifically, are considered, the systems are closer together than Reference BoteroBotero et al. (2004) suppose. The civil law codes of the nineteenth century were flanked, almost from their inception, by statutory provisions dealing with specific areas of labor regulation, including labor mobility and factory conditions (see Reference Veneziani and HeppleVeneziani 1986). Legislation played a very similar role in supplementing judge-made law in the common law systems. If anything, it was the civil law codes that played the major role in disseminating freedom of contract in employment relations in the nineteenth century; the main export of the English legal system during this period was not freedom of contract, but the statutory, and status-based, law of master and servant (Reference Hay and CravenHay & Craven 2004). There was certainly no lack of governmental “intervention” in the labor market during this period (Reference Deakin and WilkinsonDeakin & Wilkinson 2005). The interplay of judge-made law and legislation is, as we have seen, a feature that has been common to all the European systems of company law and labor law since the Industrial Revolution.

But for all that, there were, and are, differences between the common law and the civil law, at the level of core concepts and guiding assumptions. What Reference PistorPistor (2005) calls “legal ground rules” allocate responsibility for the control of economic relationships differently in the common law and civil law. Civil law judges have considerable power to shape the terms of contractual relationships through the application of open-ended general clauses, such as the principle of good faith, in ways which have no equivalent in the common law. Thus in the civil law, the tendency is for freedom of contract to be socially conditioned when, in common law systems, it is, formally, unconstrained (Reference PistorPistor 2005:9). This fundamental difference in approach permeates many contemporary features of labor and corporate law. However, it is not so much the differences in substantive rules that matter; many of these are transient in nature. The enduring difference between common law and civil law systems operates at the level of the ingrained assumptions and understandings that are deployed in legal analysis, and in the values they serve to perpetuate. Above all, the common law and civil law systems operate on different assumptions about the nature of the enterprise and the role of the legal system in regulating it.

We are still only just beginning to understand the historical processes that gave rise to the emergence of distinctive legal cultures, and their economic consequences. However, if we focus on the period surrounding industrialization, and the legal innovations that accompanied it, the outlines of the argument start to become clear. The common law approach to labor regulation was to a large extent the institutional legacy of a certain mode of industrialization, one in which, thanks to the timing of economic change, the law came to instantiate a “contractualist” notion of the enterprise, based on a strong conception of the employer's property rights as the basis for managerial prerogative. Conversely, “integrationist” forms of labor law regulation and corporate governance on the continent of Europe reflect the more gradual experience of industrial development in those systems and the different trajectory from that of Britain, which, as a result, most continental economies assumed.

The differential experience of industrialization in England and its main continental European rivals in the course of the nineteenth century provides no reason to think that there was anything inherently superior about the common law route. The pace of industrialization may have been impeded in France and Germany by the slow movement of labor from the land and by the retention of preindustrial notions of employer responsibility. But few would now suggest that “Britain's earlier industrial revolution can be represented by even the most panglossian of historians as the best of possible paths to the twentieth century” (Reference O'BrienO'Brien 1996:242).

The predominant response of the civilian systems to industrialization was not a propensity to constrain or interfere with economic development. Rather, the view emerged that property and responsibility were two sides of the same coin. In the French tradition, the exercise of public power for the protection of workers, and in the Germanic tradition, the communitarian conception of the enterprise, counterbalanced the support provided by the legal system to freedom of contract and property rights. This was a particular legal conception of a market economy and society, a different one from that which came to predominate in the English common law, but not one that was inherently less compatible with market-based forms of governance.