Comparative Institutional Analysis of Corporate Governance
Comparative Institutional Analysis of Corporate Governance
Abstract and Keywords
This chapter applies a ‘comparative institutional analysis’ approach to CG. It criticizes the property right approach for being a special case, and develops a more general model of governance with shifting allocations of property rights to different internal and external actors, contingent to the overall economic performance of the firm. This contingent relational governance solution is deemed to be complementary with an uncontingently horizontal organization, at least in innovative activities. The Silicon Valley model is discussed as supportive evidence of this mode of CG.
A Modicum of Historical Background
I myself started to get interested in the subject of corporate governance (CG) more than 20 years ago. I became interested in it as a natural outcome of having worked on the theory of the firm as an organization of managers, workers, and investors. When I had an occasion to visit Harvard in 1979–1980, however, very few people in the Department of Economics, except for the late Professor Harvey Leibenstein, had an interest in CG. Most of them had not even heard the terminology. CG was then strictly a subject matter of lawyers and my first systemic exposure to it was through auditing a small seminar at Harvard Law School. My first book in English, The Cooperative Game Theory of the Firm, published in 1984, was an attempt to integrate an economist's approach to the theory of the firm with the corporate governance theory in law that evolved from that personal experience. The references in the book contained a number of writings by legal scholars and legal documents on CG, but the text itself did not have any reference to CG. While writing the book, I debated whether or not I should use the term or not. Regrettably in retrospect, I did not because I felt the term might alienate my potential audience, economists. Instead I used the term ‘legal models of decision-making structure of the firm’ in a primitive attempt to capture the nexus between management and governance. Ironically, the book was then received more warmly by legal scholars than economists.
The term ‘corporate governance’ all of a sudden began to capture the attention of economists in the late 1980s and early 1990s. It happened in a particular atmosphere in the US at that time, when the competitive threat of German and Japanese companies alarmed the American business community. It was said then (for example, at a conference on CG organized by the American Academy of Sciences in 1991) that the erosion of competitiveness of American industries was due to the ‘short-termism’ of the top corporate executives subjected to the constant pressure of stock market signals, whereas their German and Japanese counterparts were insulated from such pressure through the bank-oriented corporate finance and governance. With the perceived reversal of competitiveness in the mid-1990s, however, (p.32) references to the ‘short-termism’ were replaced by references to ‘flexibility’ without much change in the substance of the argument, whereas the bank-oriented system became perceived as out of date because of a series of banking crises all over the world throughout the 1990s. As the force of global capitalism emanating from Wall Street exerted such a stormy influence in the second half of the 1990s, market-oriented corporate financing and governance was declared victorious. All of us know well, however, that CG may mean more than just stockholders' sovereignty, that other stakeholders also matter in the mechanisms of CG. Indeed, some of us have been earnestly doing research on this premise. However, an influential survey article on CG by Shleifer and Vishny (1997: 737) bluntly limited their scope thus: ‘CG deals with the ways in which suppliers of finance to corporations assure themselves of getting a return on their investment.’
Then, all of a sudden, another shock hit the scene: the Enron and WorldCom debacles. These events themselves do not necessarily invalidate the Shleifer-Vishny concept of CG. I suppose they would say that the ways securities and auditing regulations were formulated and practiced were wrong and could be remedied by a more rational design. For reforming the American CG system, they may be right. But in my view what the events teach us is, among other things, that any CG arrangement cannot be absolutely superior to others over time and over space. By saying this, I do not intend to advocate the return to a romantic era of bank-oriented CG system. Given changing environments, it will never be a viable universal solution. However, I would argue that it is a good time to reconsider such issues in broader perspectives like the following: Why is there a variety of CG arrangements? What are the nexuses between the past and possible subsequent trajectories of a CG system? Besides financial markets, do the polity and social and labor relations matter in the formation of a CG arrangement? How can each CG arrangement be improved by legal design? Can and should diverse arrangements eventually be harmonized?
What is CG? The CIA Approach
The issues just mentioned are comparative by nature. But for comparative studies to be productive, it is desirable to go beyond being just descriptive and to work within a unified conceptual and analytical framework for an understanding of diverse CG arrangements. It is also desirable for such a framework to be useful for accommodating insights from different disciplines. In The Cooperative Game theory of the Firm (Aoki 1984), I made a primitive attempt to present such a framework. I tried to view the firm as a coalition of the stockholders (financial resource suppliers) and the workers (firm-specific human resource suppliers) and specify conditions under which a cooperative Nash solution may, or may not, become a self-binding agreement. Law, institutional set-ups for labor and financial markets, social norms, and so forth (p.33) constitute such conditions for evolving in diverse forms (US, UK, Japanese and German). One thing remained unsatisfactory, however. The manager was considered to be performing a mediating role to achieve such a stable solution under surrounding institutional constraints, but he was not explicitly treated as a strategic player. This was certainly a point to be remedied.
In a recent book of mine, Toward a Comparative Institutional Analysis (Aoki 2001), I was able to present a more satisfactory—at least for me—framework for comparative and trans-disciplinary CG. It is based on a generic conceptualization of institutions based on a game-theoretic framework. A game may be described by a tripartite arrangement composed of a set of players, sets of their activated action choices, and the consequence function, which maps each profile of chosen actions by all the players to some consequences. I argue that statutory law constitutes an element determining a form of the consequence function, together with technology, historical legacies, and so on. I then conceptualized an institution in one particular domain of repeated social game (economic, political, or social) as a salient characteristic of its equilibrium reflected in ‘shared beliefs among players regarding the ways how the game is being played’. An institution as such is a summary representation of a stable equilibrium outcome of strategic plays among the players.
Based on this generic conceptualization, I formulated the concept of CG as follows. Let us consider the domain of CG composed of three types of strategic players: the manager, the investors, and the workers. CG is a set of self-enforcing rules (formal or informal) that regulates action choices of those players contingent on evolving states. In particular, managers' beliefs (expectations) regarding possible actions of other players in critical contingencies—such as in the state of corporate financial difficulties—that may constrain his or her action ex ante are essential elements of a CG arrangement.
There are some merits in considering an institution in general, and a CG arrangement in particular, as an equilibrium phenomenon (Aoki 2001: 14–20). Among other things, it provides a rigorous framework for understanding a particular CG arrangement as ‘nexuses (linkages)’ of various elements (such as rights to property, control, rewards, and information, as well as a mode of social relationships). It also provides a rigorous framework for understanding various modes of nexuses (complementarities) between a CG arrangement on the one hand and an organizational architecture and political-economy arrangement on the other. Under the game-theoretic framework, the nexuses of the first type may be often analyzed in terms of linked games. The idea is that a particular profile of strategies in one particular domain of a game may not be sustained as an equilibrium (or there may be many other equilibria) when that game is played separately, but it may become a (unique or distinct) equilibrium when the game is simultaneously played with another game and players belonging to both domains coordinate their own strategies in both games to achieve higher payoffs. That is, (p.34) linking games may create externalities to some or all players. For example, Granovetter's celebrated notion of social-embeddedness may be related to this idea (Granovetter 1985). A certain mode of transaction (for example, relational contracting) may not be feasible among strangers in the bazaar but, if buyers and sellers are continuously in social contracts, it may become a viable alternative because they may be afraid of being ostracized from the domain of social exchange if they default on such contracting. Social embeddedness may be relevant to an understanding of certain types of corporate governance arrangements (for example, corporate grouping found in India, Chile, and elsewhere) (see Granovetter 1992, 1994; Khanna and Palepu 1999, 2000). I will shortly provide another example of linked games for understanding a CG arrangement based on property rights.
Nexuses of one particular arrangement in the corporate governance domain with particular institutional arrangements in other domains (such as the polity and financial and labor markets) may be conceptualized as institutional complementarities and their viability and implications can be analyzed in terms of strategic complementarities within the game-theoretic framework: that is, in terms of how an equilibrium profile of strategic choices of agents in one domain can become strategically complementary to, or conditional on, the equilibrium choices of other agents in other domains. In this way, we can understand the conditional robustness of an overall institutional arrangement of the economy in general, and that of a CG arrangement in particular, as well as the multiplicity of such arrangements. It is often difficult to change a CG arrangement by a mere modification of corporate law unless it is accompanied with complementary changes in other institutions such as in labor and financial markets. I will shortly provide a few examples of institutional complementarities surrounding CG arrangements.
There is another type of nexus surrounding CG, namely, the nexuses between the arrangements of a system and the possible subsequent trajectories feasible for that system. Although I will not elaborate on this here, there is one thing I would like to note. In a game-theoretic analysis, there is usually a multiplicity of equilibria. Usually, game theorists regard a multiplicity of equilibria as troublesome, and they have spent a great deal of research effort, without decisive success, on the so-called refinement of the equilibrium to enable them to identify only one equilibrium out of many possible Nash equilibria. However, I consider that the multiplicity of equilibria of games should not be regarded as bothersome in comparative institutional analysis. On the one hand, by making institutions susceptible to equilibrium analysis, it can be made clear that institutions are humanly devised, yet can be neither arbitrarily designed nor implemented in a discretionary manner. On the other hand, once an institutional bifurcation occurs, even if two economies are exposed to the same technological and market environments afterwards, the subsequent overall institutional arrangements of the two economies may well differ, depending on their respective interim (p.35) institutional trajectories—the phenomenon known as path dependence. Thus, equilibrium and historical analyses are mutually complementary and are both indispensable to comparative institutional analysis.
Having laid out the basics of the game-theoretic framework for comparative institutional analysis, let me move on to examine some interesting examples of linked games and institutional complementarities regarding generic CG arrangements. My purpose here is to illustrate the potential explanatory power of the game-theoretic approach so that arguments remain abstract and generic.
The Hartian Property Rights Approach as a Special Case
The first example is drawn from the celebrated contributions of Oliver Hart and his associates (Grossman and Hart 1985; Hart 1995; Hart and Moore 1990). We reinterpret their major insights in the context of our framework. First, consider the simplest domain of organization constituted of one employer and one worker who invest respectively in relation-specific (firm-specific) human investments. Suppose that the organizational architecture is designed in such a way that the skill of the employer is ‘essential’ in the Hart sense while the worker's is not. This corresponds to a situation in which the manager's task is indispensable to the productive use of physical assets used by herself as well as by the worker so that the worker cannot enhance his productivity without her direction (information processing), even if he owns the entire set of physical assets. On the other hand, the manager can (at least partially) realize her value even without the trained worker if she owns the physical assets. In this case, the second-best solution is for the manager to acquire the ownership of the entire set of physical assets. Only in this way is the manager motivated to accumulate her own essential human assets. The indispensability of the manager's skill may be thought of as consisting of her ability to dictate the use of physical assets to the workers in a productive manner in unforeseeable events, when contract is incomplete. Thus, the ownership of physical assets and the hierarchical coordination of production by the manager and the physical labor of the propertyless worker are ‘linked’. In other words, proprietary firms can win out in competition with other ownership arrangements, in which the hierarchical mode is conventional in the organizational architecture in the economy. However, as we will see later, there can be other modes of organizational architecture (production coordination), so that the nexus of this type is to be regarded as a rather specific case, albeit an important one, even if we limit our attention to generic CG arrangements.
When the manager-cum-owner becomes cash-constrained, she must then raise funds from outside investors through debt contracts (when the cash constraint is moderate) or from stockholders through equity contracts (when it is severe). This situation can be analyzed as a three-person repeated game (p.36) between the investor, the manager, and the worker. In this game, the investor controls the supply of funds, the manager invests/shirks in relation-specific human capital and is engaged in hierarchical coordination, and the worker invests/shirks in relation-specific human capital and is engaged in production using physical assets. One kind of such model can be obtained simply by augmenting the Tirole's model (2001) with the explicit addition of the worker. One can derive the following implication from such model: a value-enhancing takeover by a new stockholder may not necessarily be efficiency-enhancing when the worker retaliates against the ‘breach of trust’ by a new manager with non-cooperation (Shleifer and Summers 1988; Aoki 2001: ch. 11:1).
Institutional Complementarities between Co-determination and Corporatism
In the previous model, it is assumed that the worker may be subjected to the efficiency wage discipline: that is, the worker invests in relation-specific human capital and uses it in the second-best manner in the anticipation of the employer's sharing the surplus with him, as far as the manager has kept her promise to do so. Imagine, however, that a wage rate is fixed by a corporatist agreement between the trade union and the employers' association at the national level and each management is obliged to comply with it. Suppose that, in order to elicit a worker's cooperation under this institutional environment, the employer (suppose too for a while she is a manager-cumowner just like the Hartian proprietor) allows the worker to participate in the ‘residual rights of control’ (Grossman and Hart 1986)—the rights to decide on the use of human and physical assets in contractually unspecified events—provided that the worker has always cooperated (made an effort) in past periods, and otherwise keeps the residual rights of control to herself and does not make any payment beyond what is determined in the corporatist agreement. In a symmetric way, the worker makes a reciprocating effort provided that the employer has always partially relinquished residual rights of control to the worker in the past, and otherwise shirks. Let us assume that the worker can reduce the cost of his effort by participating in the residual rights of control, possibly because of improvements in working conditions, participation in workplace design, pleasure derived from more autonomous control of his work, and so on. This implies that the participation of the worker in the residual rights of control transforms the organizational architecture from a functional hierarchy to a participatory hierarchy. On the other hand, there may be some reduction in the employer's utility in the event of partial relinquishment of residual rights of control, for she may not implement the work plan that she likes best. Still, it can be one equilibrium over periods in which the reciprocating cooperative strategies are sustained by both parties. We cannot make a definite Pareto-ranking between this equilibrium and the Hartian equilibrium.
(p.37) When the equity of the original owner of the firm is still too small relative to the required capital, she inevitably needs to abandon her ownership rights. However, in this case the governance structure cannot be the same as the shareholder governance discussed in the previous section because the worker participates in the residual rights of control. Suppose that both the worker and investors (shareholders and creditors) are able to veto a management action respectively that they prefer less than the status quo, or the reappointment of the manager for the next round of the game, depriving her of an opportunity to obtain employment continuation value. Thus, the worker and investors can exercise separate control rights over the management. Let us call this governance arrangement ‘co-determination’. Then, any unilateral new action that would hurt the worker can be blocked by a worker's veto and by the manager's career concerns. On the other hand, assume that, although the investors supply full financing, they have little useful inside information for facilitating the smooth operation of the participatory hierarchy within the firm, and thus are passive in formulating a business plan. The possibility of restructuring after initial financing can be perceived only by the manager, who has invested in firm-specific human assets. However, the investors can threaten to withdraw finance and the workers can be uncooperative if they so choose. In this setting, it can be proved that participatory hierarchy and co-determination are institutionally complementary to corporatist wage-setting. There may be a stock value-enhancing management plan that can be chosen under stockholder governance but not under co-determination if it is expected to have a welfare-reducing impact on the worker and incite a retaliatory uncooperative choice of efforts by the workers. Co-determination and stockholder sovereignty governance are thus not necessarily Pareto-rankable. Also, it is interesting to note that under co-determination more external financing takes the form of long-term debt contracts, as the interests of debt-holders and those of the worker are more congruent than under functional hierarchy (Aoki 2001: ch. 11:2).
Knowledge Sharing and Relational Contingent Governance: Benefits and Costs
As already mentioned, there has been a persistent stream of thought in the CG literature that the corporation actually is, or at least ought to be, run in the interests of various stakeholders, including the workers, but not in the sole interests of the stockholders. Even Adolph Berle, who was engaged in a harsh debate against such view in the early 1930s, was converted to it in his later career. Recently Jean Tirole, an analytical economist, made this comment in his Presidential Address to the Econometric Society: ‘I will, perhaps unconventionally for an economist, define corporate governance as the design of institutions that induce or force management to internalize the welfare of stakeholders.… There is unfortunately little formal analysis (p.38) of the economics of the stakeholder society’ (2001: 4). Relying on recent developments in contract and control theory, he pointed out that it is difficult to theoretically design multi-task incentives for the manager to satisfy diverse interests of the stakeholders or an effective arrangement for the division of control rights among stakeholders. Even if that is so, it is possible to design a CG arrangement in which control rights shift (as opposed to being divided) between stakeholders contingent on events in the corporate organizational domain: more specifically, between the insiders (the managers and workers) on one hand and a designated agent of the investors on the other, contingent on the (corporate financial) outcome of the stage game in a repeated game context. Thus, I call this governance arrangement ‘relational-contingent governance’. I first derive this mechanism theoretically as a second-best solution to a free-riding problem inherent in the organizational architecture of the knowledge-sharing type.
Suppose that both the manager's skill and the worker's skill become essential in the Hartian sense. This corresponds to the situation in which both the manager and the worker cannot generate surplus value without mutual cooperation, even if they own the entire (or relevant) set of physical assets. In this situation, an ownership arrangement cannot resolve the governance problem. I interpret this situation as one in which the information-processing activities of both manager and worker are crucial inputs for each to be productive. Catching its essential aspect in the simplest form, let us assume that they are symmetrical in their contribution to the organizational output but that each of them cannot precisely observe the other's level of effort. This type of production organization is referred to as ‘team’ (Alchian-DemsetzHolmstrom) or as horizontal hierarchy in contrast to the functional hierarchy in which the indispensability of agents' skills is asymmetric. In this type of organization, free-riding on other members' efforts becomes an inherent moral-hazard problem that cannot be resolved by sharing the outcome among the members alone. There must be an external discipline.
Suppose that this organization (let us refer to it as the ‘H-firm’ and its manager and worker as ‘insiders’) needs some outside financing for productive activity. It is provided by numerous investors who expect a certain level of financial returns. They cannot, however, observe even the aggregate output value of the H-firm ex post, but can observe only the court-verifiable event of its termination. They entrust the enforcement of financial contracts to a particular relational monitor (‘R-monitor’) who can observe the aggregate output value of the H-firm at the end of each period and then exercise control rights contingent on it according to a contract agreed with the H-firm at the beginning of the period. The R-monitor requires a certain expected level of income per period for this service payable from the current output of the H-firm.
In this setting, it can be proved that the following nexus of contingent contracts is the second-best CG arrangement for the free-riding problem (p.39) (Aoki 2001: ch. 11:3). It divides the entire range of the H-firm's possible output value at the end of each period into the following four regions in the order of highest to lowest, and specifies control rights to be exercised by either the insiders or the R-monitor on each of them. In the highest region, insider-control region, both investors and R-monitor get a fixed amount of returns and the residual output value is equally shared exclusively among the insiders. In the next highest R-monitor-control region, control rights to output shift to the R-monitor. The R-monitor pays the same rate of return to the investors as in the insider-control region, pays the agreed minimum income to the insiders, and acquires the non-negative residual. The H-firm continues to the next period. In the next lower bailing-out region, the payment schedules are the same as the previous region except that the output value level is so low that the residual borne by the R-monitor becomes negative. However, the H-firm is still sustained to the next period. This corresponds to the case in which the R-monitor bails out the H-firm comprised of the wealth-constrained insiders. In the lowest termination region, the R-monitor terminates the H-firm after making contractual payments of the minimum income to the insiders and a fixed rate of return to the investors lower than the expected investors' rate. Deficits after the termination are to be borne by the R-monitor.
The nexus of contracts just described defines a basic mechanism of governance regarding both the disposition of the H-firm's output and its continuation at the end of each period. Since control rights shift between the insiders and the R-monitor in a punctuated manner contingent on the value of the H-firm's output, we may call this arrangement ‘relational-contingent governance’. In the insider-control region, the insiders become the residual claimants, as in the case of an insider-controlled firm. However, if such a status were to extend over the entire range of output value, the moral hazard inherent in H-firms would become unavoidable. Further, if the value of output is very low, it may not be sufficient to guarantee the minimum required income of the insiders. For these two reasons, if the value of output falls below a certain level, the residual claimant status shifts to the R-monitor. If the value of output falls even further to below the termination point, the H-firm is terminated and its members have to accept inferior outside options. This efficiency-wage-like discipline can provide incentives for the insiders not to shirk. The outside option value may be taken as a parameter by the insiders of an individual H-firm, but its lowering can be regarded as a (general equilibrium) outcome of the convention of horizontal hierarchies prevailing in organizational architecture: that is, if all firms are structured as H-firms relying on the context-oriented skills of their members, and individuals' skills are geared toward a particular firm, they cannot freely move between the firms without suffering a loss in their employment continuation value. Thus, the effectiveness of relational-contingent governance is enhanced when horizontal hierarchies are established as a convention in the (p.40) organizational field. Conversely, as we have discussed, horizontal hierarchies are run more efficiently when relational contingent governance are institutionalized. Thus, the convention of horizontal hierarchies and contingent relational governance are mutually reinforcing and institutionally complementary.
Since some costs of termination may be born by investors, in practice there may be incentives for the R-monitor to terminate a financially troubled H-firm even when the H-firm should be bailed out. To counteract these incentives, some positive values—rents—needs to be expected for the R-monitor over time and across H-firms for credibly committing to a bailing-out operation whenever it is appropriate to do so. We can then discern one important dilemma inherent in the mechanism of relational-contingent governance. On the one hand, if rents are not sufficiently high, the R-monitor may be motivated to terminate firms that should be bailed out: that is, valuable organization-specific assets may be destroyed even when mildly poor performance occurs due to uncontrollable stochastic events but not to the actions of insiders. On the other hand, if the rents made possible by bailing-out are too high, the monitoring agent may be motivated to bail out a firm that should not be. If such expectation prevails, the mechanism of relational-contingent governance fails to provide proper incentives ex ante for the insiders of horizontal hierarchies to make sufficient effort. The tendency is known to economists as the ‘soft-budget constraint’ syndrome (Kornai 1980). Which syndrome prevails in a particular economy depends on the relative magnitudes of those costs and rents facing relational monitors. Explicit contracts of relational-contingent governance are hard to write in practice because of the complexity of the contractual environments. Further, it may not be possible to determine the rents for each firm, but they may be specified and generated only in a broader institutional context in which they are embedded. In actuality, one cannot assume that costs and rents are arranged in such a way that the second-best solution can be implemented with precision in each organization domain.
It is reasonable to expect that one or another of the syndromes may prevail. Yet, in environments in which rents and costs remain fairly stable, if not balanced exactly in a second-best way, expectations regarding the possible behavior of R-monitors, whoever they may be, may become predictable, and firms of the horizontal hierarchy type may accordingly be disciplined while being able to accumulate and preserve organization-specific assets in a more or less steady fashion. However, when there is an environmental change that drastically transforms the parameter values defining the costs and rents of bailing-out, so that expectations regarding the monitoring agent's possible actions become uncertain, the provision of effective relational-contingent governance will become problematical.
The above discussion has been conducted at a highly abstract level. In particular, I have been silent about who relational monitors can be and what their incentives to bail out financially depressed firms are. There (p.41) are several institutional possibilities of contingent governance relationships: (a) between firms and their main bank; (b) between subsidiary corporations and their holding/management company; (c) between an entrepreneurial start-up firm and a venture capital company; (d) between state-owned enterprises and the government; or (e) between banks and the government regulatory agency. These possibilities and their inherent syndromes are discussed in Aoki (2001: 300–5).
The Silicon Valley Model as a New Mode of CG
We discussed in the previous section CG issues for the production system in which information is symmetrically shared among its members. In contrast, we may conceive of a system whose members jointly, but competitively, pursue some systemic objectives with information differentiation among them as its inherent characteristic. For example, consider a cluster of entrepreneurial firms as observed in Silicon Valley. Each firm is engaged in the development of a module product that is expected to constitute a part of a complex innovative product system. In each modular design multiple entrepreneurial firms compete so that their information processing is hidden, or encapsulated, from one another. This information encapsulation allows another important systemic characteristic: each module can be developed independently of the design of other modules, as far as the interfaces and performance requirements among modules are standardized ex ante or ad interim and known to each entrepreneur. Then, an innovative system may be developed in an evolutionary manner by combining the best-developed product of each module ex post. When system development is extremely complex, this process may have a superior innovative capacity to cases in which system design is done in a hierarchical manner once for all, or design improvements may be done through intense information exchanges and sharing among a fixed set of design subunits. This is so because the process can create option values (Baldwin and Clark 2000) by allowing each module to experiment on diverse designs in the presence of high uncertainty. However, the option value cannot be obtained without costs. The costs are the duplication of development costs within each module. Further, if the cost of development by an entrepreneur has to be financed by outside investors so that possible returns are to be shared with them, entrepreneurial incentives may be compromised without a proper governance arrangement. How can these costs of development be controlled?
Let us consider a game played by the venture capitalist (VC) and two groups of entrepreneurs, each competing for the development of a modular product that may constitute an innovation system. These two modular products may be combined through standardized interfaces. The VC finances the initial development funds to multiple entrepreneurs in each module design and then monitors their design development without necessarily (p.42) observing their effort levels directly. It mediates a modicum of information sharing among entrepreneurs if necessary for the ad interim modification of interface. Eventually the VC selects only one entrepreneur for each module for the completion of its project and realizes its values by bringing it to public offering or arranging an acquisition by an existing company. The realized values can be shared between the VC and the selected entrepreneurs according to ex ante share contracts, but other entrepreneurs do not get anything. It is essentially a tournament game played among entrepreneurs refereed by the VC; and we may call this arrangement ‘VC governance by tournament’. The VC is linked to other financial markets for raising funds, but I do not consider this aspect here.
We can now take stock. The arrangement can create option value at the cost of duplicated development efforts and financing (Baldwin and Clark 2000). The tournament provides additional incentives for the entrepreneur, in contrast to the stand-alone development effort, because marginal benefits of additional effort are composed of marginal expected benefit obtained in the event of winning plus marginal gains obtained from the enhanced probability of winning (Aoki 2001: ch. 14). However, as the number of entrepreneurs competing in each modular design increases, these incentive effects are diluted so that there is an optimal number of entrepreneurs to compete in each module development, depending on the degree of uncertainty involved in development and the expected value of final products (Aoki and Takizawa 2002). Particularly interesting is the following proposition. If the total value of an innovative system is expected to be high, and if the VC's selection of winning entrepreneurs is believed by entrepreneurs to be precise, then it is possible that, even for the same share allocation between entrepreneurs and financiers, VC governance by tournament can elicit higher development efforts from entrepreneurs than arm's-length financing, and that its effect, together with the creation of option value, can compensate the social costs of duplicated development efforts.
Concluding Remarks on the Role of Law
Using a simple generic model, I have shown that there may exist diverse CG arrangements associated with different organizational architectures. Also, I have argued that those arrangements are supported by complementary institutional arrangements in other domains (see Aoki 2001 for a more comprehensive treatment on this subject). This may indicate that a CG arrangement may have a robust property that may be hard to change unless complementary changes occur in other domains. Also, a mode of organizational architecture tends to evolve as a convention, although conscious design elements are also involved. Thus, a particular CG arrangement and a corresponding organizational architecture may co-evolve. Does all this indicate that an attempt to improve on a CG arrangement through the design of statutory law is bound to be futile? Obviously, this is not the case.
(p.43) As I indicated before, statutory laws constitute an element of the consequence function of the game structure. In other words, they may provide information to the players about the possible outcomes of their actions if laws are enforced, although whether or not they are actually enforced is a matter determined through the strategic interplays between the enforcer and other players. Thus, laws affect the outcome of the game through the expectations of the players as well as their incentives. So statutory laws are not institutions per se in my conceptualization, but can induce the evolution of an institution. In particular, codified rules of corporate governance—that is, the legal rights afforded to various agents (particularly shareholders and employees) and the associated legal procedures—define the exogenous rules of the game in the corporate organization domain, and as such they may affect the beliefs and incentives of the agents and thereby corporate performance (La Porta et al. 1998). However, legal rules that are inconsistent with equilibria in complementary domains, particularly with a prevailing convention of organizational architecture, may not yield the outcome intended by the legislature in the corporate organization domain. For example, the Japanese Commercial Code provides minority shareholders with one of the strongest rights at stockholders' meeting. However, its governance arrangement is normally not considered to be stockholder-controlled.
On the other hand, sustainable legal rules for corporate governance may be understood as the codification of an equilibrium arrangement that evolved through a long history of complementary institutions (for example, co-determination in Germany; see Aoki 2001: ch. 6). A careful and systematic study is called for regarding the questions of how the initial institutional conditions, such as the legacies of old institutions and the prevailing informal rules (norms, social ethics, and so on), the kinds and level of the existing stock of human competence, can affect policy impact on subsequent legal evolution and, conversely, how formal rule-setting in the polity interacts with the evolution of the endogenous rules of the games (that is, institutions) in CG and other domains.
Alchian, A. and Demsetz, H. (1972). ‘Production, Information Costs, and Economic Organization’, American Economic Review, 62: 777–95.
Aoki, M. (1984). The Cooperative Game Theory of the Firm. Oxford: Oxford University Press.
Aoki, M. (2001). Towards a Comparative Institutional Analysis. Cambridge, MA: MIT Press.
—— and Takizawa, H. (2002). ‘Information, Incentives and Option Value: The Silicon-Valley Model’, The Journal of Comparative Economics, 30: 759–86.
Granovetter, M. (1985). ‘Economic Action and Social Structure: The Problem of Embeddedness’, American Journal of Sociology, 91: 480–510.
Granovetter, M. (1992). ‘Economic Institutions as Social Constructions: A Framework for Analysis’, Acta Sociologica, 35: 3–11.
Granovetter, M. (1994). ‘Business Groups’, in N.J. Smelser and R. Swedberg (eds.), Handbook of Economic Sociology. Princeton: Princeton University Press.
Grossman, S. and Hart, O. (1986). ‘The Costs and Benefits of Ownership: A Theory of Vertical and Lateral Integration’, Journal of Political Economy, 94: 691–719.
Hart, O. (1995). Firms, Contracts, and Financial Structure. Oxford: Clarendon Press.
Hart, O. and Moore, J. (1990). ‘Property Rights and the Nature of the Firm’, Journal of Political Economy, 98: 1119–58.
Khanna, T. and Palepu, K. (1999). ‘Policy Shocks, Market Intermediaries, and Corporate Strategy: The Evolution of Business Groups in Chile and India’, Journal of Economics and Management Strategy, 8: 271–310.
Khanna, T. (2000). ‘Is Group Affiliation Profitable in Emerging Markets?: An Analysis of Diversified Indian Business Groups’, Journal of Finance, 55: 867–91.
Kornai, J. (1980). Economics of Shortage. Amsterdam: North-Holland.
La Porta, R., Lopez-de-Silanes, F., Shleifer, A., and Vishny, R. (1998). ‘Law and Finance’, Journal of Political Economy, 106: 1113–55.
Shleifer, A. and Summers, L. (1988). ‘Breach of Trust in Hostile Takeovers’, in A. J. Auerbach (ed.), Corporate Takeovers: Causes and Consequence. Chicago: University of Chicago Press.
Shleifer, A. and Vishny, R. (1997). ‘A Survey of Corporate Governance’, Journal of Finance, 52: 737–87.
Tirole, J. (2001). ‘Corporate Governance’, Econometrica, 69: 1–35.