The advantage to codetermination is the sharing of power throughout all levels of the organization.

7.6.2 Employees

The literature on employee involvement has focused on two questions: does employee involvement come at the expense of shareholders (reduce shareholder wealth), and if contracts are incomplete, is employee involvement efficient? There is little empirical evidence in support of the first question and, to our knowledge, no empirical evidence that would allow us to formulate an answer to the second question.

The incidence of employee involvement is often thought to be limited to Germany’s mandatory codetermination and two-tier boards. In fact, employee involvement is also mandatory in Austria and the Netherlands245 (two-tier boards), Denmark, Sweden, Luxembourg and France246 (one-tier board). Companies operating in two or more member states of the European Union must have a “European Works Council”.247Voluntary codetermination can be found in Finland and Switzerland [Wymeersch (1998)]. In contrast, employees in Japan are not formally represented on the board [Hoshi (1998)], although Japanese corporations are run, supposedly, in the employees’ and not the shareholders’ interest [Allen and Gale (2000)]. Compared to the wealth of opinions on employee involvement, the empirical literature is small, even for countries where such institutions are known to exist, such as Germany.

German codetermination provides for mandatory representation of employees on the supervisory board of corporations248 with three levels of intensity: full parity for coal, iron and steel companies (since 1951),249 quasi-parity for other companies with more than 2000 employees (since 1976)250 and 13 parity for those with 500–2000 employees (since 1994).251 Media companies are exempt.

Does the degree of codetermination adversely affect shareholder wealth or company performance? If codetermination reduces shareholder wealth, shareholders will resent codetermination and they will try to bypass252 or shift board rights to the general assembly. There is some evidence of the former but none for the latter. In 1976 most supervisory boards of corporations subject to the quasi-parity regime did not have to be consulted on important management decisions253 [Gerum et al. (1988)], a clear violation of the recommendations in most corporate governance codes (see Section 6.2).254

If there are losses in shareholder wealth from codetermination, how large are they? Econometric studies of codetermination compare company or sector performance “before and after” the 1951, 1952, 1972 and 1976 reforms or their enforcement by the courts. These studies find no or small effects of codetermination [Svejnar (1981, 1982), Benelli et al. (1987), Baums and Frick (1999)] and/or their samples and methodology are controversial [Gurdon and Rai (1990), FitzRoy and Kraft (1993)]. 255 A recent study relies on the cross-section variation of codetermination intensity, controlling for different types of equity control and company size. It finds codetermination reducing market-to-book-value and return on equity [Gorton and Schmid (2000a)]. Codetermination intensity and its incidence correlate with other factors that are known to matter for stock price and accounting measures of performance, in particular sector and company size, and it is doubtful that one can ever fully control for these factors.

5.3 The ‘Scandinavian School’

In the 1950s the (powerful) trade unions in Scandinavia helped to pass the ‘codetermination legislation.’ This law stated that unions must be involved in technological design—originally motivated by concerns about deskilling and job loss through automation. A form of sociotechnical systems analysis evolved into a set of techniques for studying work places and processes. As computers arrived in the workplace, this came to include progressive computer scientists and social scientists, many of whom now participate in STS publishing and conferences, as well as CSCW and PD (Greenbaum and Kyng 1991, Bjerknes et al. 1987, Bødker 1991, Neumann and Star 1996).

The ‘Scandinavian School’

In the 1950s the (powerful) trade unions across Scandinavia helped to pass the ‘co-determination legislation.’ These laws stated that unions must be involved in technological design – originally motivated by concerns about deskilling and job loss through automation. A form of sociotechnical systems analysis evolved into a set of techniques for studying work places and processes. As computers arrived in the workplace, this came to include progressive computer scientists and social scientists, many of whom now participate in STS publishing and conferences, as well as Computer Supported Cooperative Work, Human Computer Interaction, and Participatory Design conferences (Greenbaum and Kyng, 1991; Bjerknes et al., 1987; Bødker, 1991; Neumann and Star, 1996). Gregory (2003) discusses the democratic bases of the multilevel Scandinavian participatory design approach – in particular their comfort with working through conflict and contradiction.

Voting

The power of shareholders to direct corporate affairs through voting is limited. Shareholders elect directors (exclusively, outside of the codetermination context), and as noted above they vote on certain fundamental matters. But at least in the United States, shareholders vote on matters such as charter amendments, mergers, and dissolution only after the board has first voted in favor of a proposal. Thus, shareholders cannot initiate changes (although shareholders can act on some of these items in some countries without board approval). Moreover, traditionally the power to elect directors has been of limited use in public corporations, since most shareholders do not attend shareholder meetings but instead vote by proxy, and generally only the corporation itself finds it financially worthwhile to distribute proxy material. Thus, shareholders have traditionally only been able to vote on the slate presented to them by the board. The main context in which the shareholder's power to vote for the board has had value (in public corporations) is when there is a hostile takeover, so that bidders can acquire a controlling share and thereby take control. However, boards have been able to find highly effective defenses against takeovers, and for the most part American law has allowed those defenses (British rules have been less accommodating of takeover defenses).

More recently, certain institutional shareholders with an interest in active corporate governance have attempted to expand their ability to influence corporate behavior through voting. Their efforts are discussed below in the Recent Developments section.

1 Introduction

Organizational participation (OP) can be treated from the micro perspective of the individual, the meso level of the organization, and the macro level of national policy. There are substantial differences between countries. The United States and Britain have voluntary schemes based on companies and sometimes on sections or divisions of companies. Most continental European countries have a legal underpinning for OP. For instance, the German codetermination scheme gives employees up to 50% representation at the supervisory board of their two-tier board structure. There are also elected representatives on works councils with legal rights to be consulted in defined circumstances. Similar legislation exists in most continental and Scandinavian countries.

However, formal representative participation does not require legislation. In Britain and the United States, such schemes are set up voluntarily by many companies sometimes related to collective bargaining. Formal schemes, sometimes called joint consultation or employee involvement, become part of company policy and frequently have a written constitution and elected representatives. Informal participation is quite different and can take many forms. It may be encourages by top management, but it is not usually accepted boardroom policy. It is normally left to individual managers, departments, or divisions and can vary from ad hoc meetings to regular forums. Informal participation is an aspect of managerial leadership and changes as managers leave or retire. A particular department may acquire a tradition of OP, which then lasts beyond a particular manager’s style of operating. In voluntary formal and informal OP, the content of deliberation and the amount of influence available to different groups vary considerably.

In Japan, tradition and culture, rather than legislation, have produced formally sanctioned schemes of bottom-up management (ringi seido), the extensive use of quality circles, and the custom of lifelong employment in large companies. Some of these Japanese OP and human resources schemes have begun to weaken.

Forms of Business Entities

A fourth body of law fundamental to the conduct of business prescribes the types of entities by which business can be conducted. A business may be conducted as a sole proprietorship owned by one person, a general partnership, a limited partnership, a limited liability company, or a corporation.

2.1 Unit of Analysis

The job-analysis unit is the job itself and the tasks that comprise it. For this reason, many job-analysis systems relate directly to the tasks and the immediate work conditions. More comprehensive job-analysis systems (see Dunckel 1999, Gael 1988) take into account the fact that essential features (e.g., the potential of the individual for defining goals self-reliantly) and results of the tasks become comprehensible only if the overall organizational conditions and the knowledge, skills and attitudes of the workers are taken into consideration. They therefore extend the analysis to include both the organizational conditions (e.g., degree of responsibility, codetermination potential, values and norms, leadership climate) and the working individuals with their respective knowledge, skills and attitudes.

Methods can be classified, based on widely accepted criteria, according to whether they are more job/task-oriented or worker/person-oriented. Job-oriented analysis is concerned with analyzing tasks, job conditions or job features, without regard for the concrete individuals involved and their different knowledge, skills, abilities and attitudes. Person-oriented analysis, on the other hand, is centred on the person and specifically concerned with differences between individual workers in the perception, interpretation and performance of the job.

Typical job-oriented approaches are: Functional Job Analysis, Task Inventories, Health Services Mobility Approach; typical person-oriented approaches are: Position Analysis Questionnaire, Critical Incident Technique, Ability Requirement Scales (see Gael 1988, Ghorpade 1988).

The distinction between job-oriented and person-oriented analysis is of conceptual and practical significance. In stress research, for example, it is conceptually important to begin by determining stress factors independently of the person involved (and their individual perceptions and coping behavior) in a job-oriented manner, then going on to examine how objectively identical stress factors are perceived and dealt with differently by different individuals and how they affect different people. However, the (person-oriented) analysis of these inter-individual differences means first determining objective stress factors, because we are concerned here with inter-individual differences in relation to objectively identical stress factors. This distinction is of practical significance when planning future workplaces for which there are not yet any workers. Here, the job-oriented approach is the only way of obtaining job analysis information.

Unit of Analysis

The job analysis unit is the job itself and the tasks that comprise it. For this reason, many job analysis systems relate directly to the tasks and the immediate work conditions. More comprehensive job analysis systems (Dunckel, 1999; Gael, 1988) take into account the fact that essential features (e.g., potential for defining goals self-reliantly) and results of the tasks only become comprehensible if the overall organizational conditions and the knowledge, skills, and attitudes of the workers are taken into consideration. They therefore extend the analysis to include both the organizational conditions (e.g., degree of responsibility, codetermination potential, values and norms, leadership climate) and the working individuals with their respective knowledge, skills, and attitudes.

Methods can be classified, based on widely accepted criteria, according to whether they are more job/task-oriented or worker/person-oriented (Brannick et al., 2007). Job-oriented analysis is concerned with analyzing tasks, job conditions, or job features, without regard for the concrete individuals involved and their different knowledge, skills, abilities, and attitudes. Person-oriented analysis, on the other hand, is centered on the person and specifically concerned with differences among individual workers in the perception, interpretation, and performance of the job. Typical job-oriented approaches are: functional job analysis, task inventories, health services mobility approach (Gael, 1988; Ghorpade, 1988); typical person-oriented approaches are: position analysis questionnaire, critical incident technique, ability requirement scales, cognitive task analysis (Gael, 1988; Ghorpade, 1988).

The distinction between job-oriented and person-oriented analysis is of conceptual and practical significance. In stress research, for example, it is conceptually important to begin by determining stress factors independently of the person involved (and their individual perceptions and coping behavior) in a job-oriented manner, then going on to examine how objectively identical stress factors are differently perceived and dealt with by different individuals and how they affect different people. However, the (person-oriented) analysis of these interindividual differences means first determining objective stress factors because we are concerned here with interindividual differences in relation to objectively identical stress factors. This distinction is of practical significance when planning future workplaces for which there are not yet any workers. Here, the job-oriented approach is the only way of obtaining job analysis information.

The Containment of Liberal Statism

Whereas statist elements of New Deal reform met with powerful opposition by the late 1930s, the rudimentary modernization theory embodied in Hoover's associative state was further elaborated during the Cold War, as U.S. manufacturing and credit extension underpinned the non-Soviet world economy. The need for uninterrupted production to reap the profits from rising demand formed the basis for a kind of truce between capital and strong industrial unions based on supervised collective bargaining, negotiated wage adjustments to meet inflation, a union crackdown on wildcat job actions, and elite theories that cast this system as the start of a ‘post-capitalist’ era of widespread abundance. Although President Dwight Eisenhower cautioned against the dangers of a ‘military–industrial complex,’ U.S. foreign policy dictated massive defense production alongside consumer goods manufacture, leading to low unemployment rates and good wages. During the immediate post-World War II era, a mildly redistributive welfare state expansion including the GI Bill, social security extension, and minimum wage protection expressed an ongoing progressive commitment to some measure of social and economic democracy – though for whites only. Failed proposals for universal health care and full employment traced its limits.

The architects of U.S. corporate liberalism never intended it to become a European-style corporatism that involved mandatory codetermination of industrial and labor policy; in the diluted native form, it provided an alternative vision to liberal statism, supplying a degree of constraint on both income-supporting policies and union participation in economic planning that was present in anything like comparable strength nowhere else. More than other distinctive U.S. features, the fuller elaboration of the associative vision even during the deceptively radical 1960s goes far to explain both the special character of modern United States liberalism and the persistence of powerful linkages between ruling business elites and pro-corporate antistatists in political roles.

In light of recent scholarship, we now know much more about the mobilization against the New Deal, including how vehemently anti-New Dealers worked to defeat government planning for economic performance. Opponents of planning clustered around nodes represented by the resurgent free market liberalism of Friedrich von Hayek, who forecast a long slide down the slippery slope to communism for welfare states, by the romanticized heroics of libertarian novelist Ayn Rand's fictional entrepreneurs, and by concerns among business conservatives and Southern industrialists that federal government ‘interference’ might invade local prerogatives regarding labor organization and race. Aiding their project, a ‘fiscal revolution’ in economic policy thinking, underway by 1937 and dominant by the 1960s, offered modified, ‘commercial Keynesian’ policy knowledge suggesting that the measured use of fiscal and monetary levers would enable macroeconomic stabilization without requiring intrusion into capitalist investment practices or extensive interference with labor markets. Those who took this view in the early postwar decades typically saw some benefit in the collective bargaining structures mandated by the National Labor Relations Board (NLRB), whereas they preferred that all decisions regarding production be left to private ordering in the context of a mildly redistributive compensatory state.

Even more important foes of liberal statism were organizations where conservative theorists and high-level business executives engaged in counter-planning against all that the New Deal had inaugurated for the role of government. These efforts began quite early as business conservatives organized in the Liberty League, the National Association of Manufacturers, and the Chamber of Commerce to promote a pro-business agenda against labor unions and Social Security, and to finance economists such as Hayek, who organized the laissez-faire oriented Mount Pelerin Society in 1947. Against his own preference, this organization abandoned reservations about the moral exactions of modern capitalism that remained urgent for conservatives such as Hayek and early Chicago-school economist Frank H. Knight. The economists' laissez-faire movement, linked with anti-communism through the John Birch Society and informed by center-right economic theory through the Brookings Institution and the American Enterprise Institute, found a champion in staunchly anti-New Deal and pro-states rights department store magnate Barry Goldwater in his Phoenix-centered mobilization of business conservatives. Although his 1960 election campaign demonstrated the perils of attacking Social Security, it also set the stage for pro-market conservatives to turn the Republican Party into a vehicle for anti-statist ideology.

To sum up, whereas the first generation of American liberal-statist social scientists had reconstituted government as a knowledgeable, capable, morally responsible regulator and redistributor, and the second generation had modeled the state as social insurer and mediator of class antagonisms, the third, late New Deal generation of statist liberals conceived a state role that captured the planning role that originated in the conservation movement, the social provision role embodied in the Beveridge Plan, and the codetermination methods that took hold in postwar European social capitalism. It was strongly nationalist, and thus reminiscent of older Hamiltonian and Whig models of a development state responsible for energizing and supporting through provision of infrastructure voluntary community efforts leading to balanced economic, moral, and civic development of the commonwealth. Policy discourses of the 1950s, developed around Cold War internal security issues, an increasingly consensual commercial Keynesianism, modernization theories exported from the United States as technologies for liberal internationalism, dreams of ‘post-modernity’ in social relations, and domestic discourses of racial and gender identity, challenged the normative postures that liberal statists and left progressives had formulated between the 1880s and the Fair Deal and inaugurated a new ideological era. The legacy remains, embodied in resilient structures of disciplinary knowledge formation, though once durable discursive norms that demanded grounding political argument in social empiricism have been dramatically compromised in the post-1970s era of ‘take no prisoners’ partisanship that moved the U.S. political spectrum dramatically to the right. Still, absent the polarizing tendencies of a world divided between capitalism and state communism, and even given now the resurgence over recent decades of libertarianism and free market ideology, it becomes ever more obvious that the critical decisions in democratic governance in the West, and beyond, will henceforth be – as they were in the new liberal era, 1880s to 1940s – hard choices between variants of modern liberalism.

2.3.2 Market structure and innovation

Moving on to our consideration of the relationship between market structure and R&D, the empirical patterns are mixed, and not terribly informative. Even before one controls for industry effects, the variance in R&D intensity explained by market concentration is small. Moreover, whatever relationship that exists in cross sections becomes imperceptible with the inclusion of controls for industry characteristics, whether expressed as industry fixed effects or in the form of survey-based and other measures of industry characteristics such as technological opportunity, appropriability conditions, and demand. In parallel to a decades-long accumulation of mixed results, theorists have also spawned an almost equally voluminous and equivocal literature on the link between market structure and innovation.

How should we think about the sensitivity of the link between market structure and innovation to other industry-level variables, no less the modest explanatory power of market structure observed in cross-sectional data? And how should we consider this relationship in light of the lack of theoretical consensus on the question? As noted above, in response to the indeterminacy of the theory and the inconclusiveness of the empirical results, Sutton (1998) develops a class of game-theoretic models yielding a range of possible Nash equilibria that bound a set of possible measurable outcomes. The substantive conclusion of his analysis is consistent with the earlier conclusion of Cohen and Levin (1989), namely that, though market concentration and R&D intensity may be correlated, market concentration is not an independent, important driver of innovation. Although a question of time horizon, the argument that market structure is not a fundamental determinant of innovation turns on the empirical case that market structure may be at least partly a function of other, more plausibly exogenous variables, and possibly innovation itself. Thus, the correlation observed between market structure and R&D intensity reflects either their codetermination or the impact of innovation on market structure.28 Such a conclusion calls into question the “Schumpeterian tradeoff”—at least that between the allocative efficiency tied to ex ante market structure and the dynamic efficiency associated with the pace of technical advance.29

In his review of the “Schumpeterian” (neoclassical) theoretical and empirical literatures, Gilbert (2006) suggests that we should not conclude from the mixed empirical record that market structure does not matter for innovation. Rather, he argues that one should consider that there is no one theory of the relationship, but many. And his response to the sensitivity of the empirical relationship between market structure and R&D to industry effects and the inconclusiveness of the findings is (1) that industry effects “mask” the relationship; (2) that empirical scholars have not controlled for the contingencies highlighted by theorists; and (3) that empirical scholarship suffers from limited data, measures, and methods. Regarding the need to consider the contingencies, empiricists should work harder at understanding the key features of industries and technologies that may condition the relationship, and at least a couple of these contingencies are highlighted by theorists (as well as empiricists), including firms’ abilities to protect their innovations, and the type of innovation in question—particularly process versus product innovation. The challenge for testing the game-theoretic models of R&D rivalry, however, is that only in their most stripped down and simplified version do they provide clear, testable empirical implications (see below), partly because they analyze behavior in highly stylized and counterfactual settings. For example, many models focus on the interaction of a single incumbent and a single prospective entrant, or, alternatively, symmetric competitors. Moreover, many of the results obtained in this literature depend upon typically unverifiable assumptions concerning the distribution of information, the identity of the decision variables, and the sequence of moves.

Regarding measures, there can be little disagreement with Gilbert’s contention that the commonly employed measure of market structure, market concentration, does not accurately reflect the nature or intensity of competition. Progress has been made here, however, particularly with the use of measures of market power, such as the modified Lerner index employed by Nickell (1996) and Aghion et al. (2005), although even those measures are limited. Also promising is the strategy of using multiple measures of competition, as employed by Geroski (1990), and subsequently Artes (2009), with the idea that one wants to look for robust results across them.

For the study of the relationship between competition and innovation, just as fundamental as the absence of appropriate measures of the intensity of competition, is our limited understanding of exactly how firms compete with respect to innovation. Grabowski and Baxter (1973) conducted the first empirical study of strategic interaction, offering weak evidence that firms in the chemical industry engage in competitive matching of R&D investment. Little empirical attention was devoted to the subject until the 1990s (e.g., Cockburn and Henderson, 1994; Khanna, 1995; Lerner, 1997; Meron and Caves, 1991), when a number of scholars looked for evidence of strategic interaction in R&D spending patterns. Due, however, to the contingent quality of the theoretical literature’s results, these studies focus on the simplest form thereof, namely competitive matching in either R&D investment or new product introductions. Henderson and Cockburn (1994) look for evidence of positively correlated project-level research investments within therapeutic drug classes in the ethical pharmaceutical industry. Meron and Caves (1991) try to discern matching in firms’ overall R&D expenditures in 28 US manufacturing industries. Khanna’s (1995) study of market segments within the high-end computer industry and Lerner’s (1997) study of Winchester disk drives search for evidence of matching behavior with respect to product introductions. The difficulty facing all these analyses is that there are numerous explanations for positively correlated R&D or product introduction behaviors other than strategic behavior. These include common changes in industry-level technological opportunity and demand conditions, spillovers from leading firms which increase the marginal productivities of rival R&D, or simply a catchup phenomenon where equally capable firms involved in similar activities all move in the same direction at roughly similar rates with it being a matter of chance that any one firm moves before the others.

Considering the difficulties in controlling for these alternative factors, it is not surprising that the studies, considered together, are inconclusive. Cockburn and Henderson (1994) find that purposive matching does not appear to characterize “the bulk of research investment” in ethical pharmaceuticals. Despite qualitative evidence suggesting some purposive matching, Lerner (1997) cannot reject a simpler probabilistic “catchup” explanation. In contrast, on the basis of quantitative and as well as qualitative evidence, Khanna (1995) concludes that there is competitive matching within segments of the computer industry. For most of the industries in their sample, Meron and Caves (1991) found some evidence of strategic matching within groups of what they identified as core firms within each industry.30

Thus, some of the studies suggest that strategic interaction affects innovative activity and some suggest that it does not. Also, it is hard to know from these results whether the absence of a common result across the different industries examined signifies that strategic interaction matters more or differently in some industries than in others, but, if that is true, it would be interesting to know what conditions its character and importance. Even in the studies that suggest that competitive interaction matters, however, we have little sense of how important it is relative to other factors. Indeed, Geroski (1991c) has speculated that strategic rivalry may be of second-order importance when compared to the influence of factors such as technological opportunity, and Cockburn and Henderson (1994) suggest that, in addition to opportunity, heterogeneous firm capabilities also appear to be much more important. These comments do not, however, suggest that we should dismiss the impact of strategic interaction on innovation, but, rather, suggest that we need to devote more study to the issue.