Organizations exist because they cost-effectively coordinate and provide incentives relative to alternative modes of transacting. The article exposes some of the main economic explanations for why organization emerges, what it does and how its scale and scope are determined, highlighting the role of a hierarchical command structure in aligning the members’ objectives with those of the organization and successfully overcoming problems of asymmetric information and collusion.
KeywordsIncentive Scheme Information Problem Incomplete Contract Hide Action Command Structure
Organization is an entity designed to influence behavior and reach collective objectives for its members.
Numerous classifications of organization can be made to emphasize its different functions or characteristics, according to objectives, output type, ownership structure, hierarchical network, degree of delegation of decisions and divisionalization, financial structure, reliance on technology and innovation, measurability of outputs and inputs, power of incentives, size in labor force, output or asset value, and scope measured vertically or horizontally in terms of diversification of output and tasks. In the modern state, almost all economic and social activities involve some form of integration and organization. The raison d’être of organization is cost-effective coordination and provision of incentives in transactions where private negotiations by independent units fail to generate satisfactory outcomes.
The new institutional approach to organization of activity, common in economics and legal studies, puts the transaction to the forefront as the unit of analysis. The key to understanding the organization and its boundaries is thus the nature of transaction costs, viewing the organization as a solution by the actors to executing transactions at minimal cost (Coase 1937).
The informational requirements for value-maximizing coordination of a wide range of activities are huge, by far exceeding the signals contained in market prices. The organization coordinates along multiple dimensions. It allocates productive and consumptive resources across tasks and members. It coordinates over time as well: to “fit the pieces” properly, avoid effort duplication, and produce value in rather unpredictable environments, the organization solves assignment and synchronization problems. It induces mutual adjustments and achieves standardization in processes, outputs, knowledge, and even norms and values. By creating a moral order and identity in the minds of its members, the organization copes better with the rise in communication and coordination costs that follow specialization of its labor motivated by productivity.
Transaction costs have two main sources. The first, bounded rationality, is the limited capacity of humans to foresee the relevant contingencies and describe with precision those foreseen, including the corresponding set of actions that should be taken in each contingency. Bounded rationality implies that the parties to a transaction will draft an incomplete contract, with gaps and missing provisions. Incompleteness is manifest in the contractual terms of idiosyncratic transactions in which parameters such as “quality” and “know-how” are important, yet difficult to define unambiguously. The resulting cost for the parties can be huge because contract incompleteness invites opportunistic behavior. Organization may cope better with contract incompleteness because it replaces the costly renegotiations in unforeseen contingencies by a command structure where people are told what to do. Integration of a transaction is supplemented by implementing an appropriate incentive scheme, instituting a monitoring and command system, and bringing to the forefront career concerns of the parties, inside and outside the organization.
The second source of transaction costs is information problems. Members at all levels of the organization have private information, but their goals are not necessarily congruent with those of the organization. The simplest nonmarket paradigm to study the information problems and their incentive implications is the principal-agent framework. In its canonical form, the principal makes a take-it-or-leave-it contract offer to the agent who, if the parties agree, performs a task which the principal cannot for lack of time or expertise. The principal faces three types of informational obstacles in motivating the agent to take the appropriate action: hidden action (moral hazard), which refers to unobservability of the agent’s actions; hidden knowledge (adverse selection), which refers to the agent’s task-relevant private information; and verifiability of relevant parameters and contract terms by third parties, such as a court of law.
The challenge in controlling the hidden action problem is instituting objective and verifiable performance measures that are somehow correlated with the agent’s unobservable actions, then, using these measures to tailor an incentive package which indirectly and cost-effectively aligns the agent’s goals with those of the principal. However, because the relation between the agent’s actions and performance measures is typically non-deterministic, raising the power of incentives inevitably exposes the agent to risk. This generates a utility loss which the principal bears in the form of a “risk premium” if the agent is risk averse. The solution to the hidden action problem thus poses a dilemma for the organization, namely, the need to balance incentive provision costs against the costs of risk bearing (Grossman and Hart 1983). In practice, firms may choose to offer stock options to their CEOs to motivate them so that they in turn motivate the employees, but the considerable indeterminacy in stock prices coupled with the obstacles in measuring performance typically reduces the efficacy of such performance-based payment schemes (Chiappori and Salanié 2003). These negative consequences amplify in complex and bureaucratized, multilayer organizations (Williamson 1985). Smaller, owner-controlled business organizations have an advantage in coping with incentive problems that are rooted in unobservability of its members’ actions.
The second informational obstacle, hidden knowledge, usually takes the form of private information about individual characteristics, some task specific and others general, such as effort cost or endurance, creativity, and productivity in teams. These are crucial pieces of information for efficient matching of tasks with employees, physical with human resources. Screening methods before and during employment such as interviews, contract design, and probation terms can be used to extract information from the agents, yet in practice the power of these instruments is limited and agents enjoy information rents. To protect their rents, low-cost agents have no incentive to reveal their productivity if the principal is not able to commit to not using the information to the detriment of the agent. The expectation that revelation of productivity will raise future performance standards is known as the ratchet effect (Freixas et al. 1985).
The inability to measure and verify relevant performance parameters such as quality, effort, output, and payments is the third informational obstacle to motivation. The distinction between observability and verifiability is important here. The scope for opportunistic behavior remains strong even if the parties can perfectly observe the relevant parameters, provided courts are not able to verify them. The theory of repeated games suggests that third-party verification is not an issue because cooperative behavior can be sustained as a (subgame-perfect) Nash equilibrium as long as the parties perfectly observe the outcome, but this conclusion collapses when the parties heavily discount future payoffs or if the frequency of transaction is not sufficiently high.
The information problems and their inefficiency consequences amplify in teams, where two or more agents are expected to contribute to a set of common objectives. Each agent has an incentive to free ride, withhold effort, and rely on others. Peer monitoring and pressure are often not strong enough to fully overcome this problem. Restoring optimal incentives in a team context can be very costly too, as the principal may have to pay huge individual rewards for a successful team outcome to fully eliminate the possibility of free riding. Handling these costs leads to differences in the organization’s size, incentive structure, and architecture.
Organization outperforms alternative modes of governance for idiosyncratic transactions that are projected to occur frequently, based on specialized needs (Williamson 1985). Task specialization and relationship-specific investments on assets and knowledge generate dependencies. Because these investments are worth more for the transacting parties in their relation than outside, ideally the relationship should be governed by long-term contracts that fully protect relationship-specific investments. But contracts are incomplete and open to opportunism, which suggests the best environment to protect relationship-specific investments is the vertically integrated command structure of the organization (Grossman and Hart 1986).
The organization can be viewed as a nexus of incomplete contracts between principals and agents in a cascade of relationships (Alchian and Demsetz 1972). The interactions are reflected in a formal architecture, usually a hierarchical structure that shows the degree of centralization, how the constituent units are arranged, authority allocated, and tasks assigned. The channels of this architecture produce decisions, performance supervision at multiple levels, and manage information flows. Aspects of information processing include data generation and establishing a network through which information, processed or raw, is transmitted with maximal processing capability and minimal delay. The lesser the informational disadvantage of a unified governance and the greater the need for coordination, the greater the advantage of organization relative to its alternatives. The organization also has an advantage in performing pressing tasks with minimal delay (Bolton and Farrell 1990).
An organization is informationally efficient if its network minimizes the need for extra information in checking whether a new project or plan of action is worthwhile. Though a reversed tree form of hierarchy can minimize processing delays, usually this is not a fully symmetric, balanced hierarchy. Rather, it is optimal to involve decision nodes at upper levels of the hierarchy in various stages of the data processing exercise (Radner 1992). Placing a group of equally effective people in parallel as a polyarchy ends up selecting a larger number of available projects than a linear hierarchy order where final approval requires sequential approval by all. Therefore, the incidence of type-I statistical error is higher in hierarchy, whereas polyarchy displays higher errors of type II. The relative performance of the two modes, polyarchy or hierarchy, depends on the importance of the two error types (Sah and Stiglitz 1986). Because projects or new ideas vary in this respect, the organization’s flexibility to form polyarchic or hierarchic screens in decision-making is an asset. Organizational form also matters, in particular, regarding coordination of changes and the scale of experimentation with new projects. The M-form has an advantage in coordinating attribute matching and flexibility in choosing the scale of experimentation, whereas the U-form organization achieves better coordination in attribute compatibility (Qian et al. 2006).
Although the formal chart may look hierarchical, many activities and interactions within the organization do not necessarily follow hierarchy. Delegation and decentralization is the norm. The level of decentralization is often defined in terms of the span of control. So, of two organizations with six people, the one in which five specialists support a middle-level manager is more centralized than one in which two pairs of specialists support two middle-level managers. But the real allocation of authority may be quite different, as the organization may selectively and informally authorize the subordinates while formally assigning authority to an overloaded superior. This structure is shown to serve indirectly the motivation of subordinates who know that their ideas and preferences will sometimes be implemented (Aghion and Tirole 1997). The optimal overlap between real and formal authority balances these motivational benefits against the risks resulting from dissipation of control, a trade-off that depends on the degree of congruence between the goals of people working at different formal levels of the hierarchy.
Large organizations in particular are vulnerable to influence activities. Members spend time and resources to affect the decisions made at upper levels, typically in the form of producing and manipulating evidence to develop credentials for promotion. The resulting loss in the organization’s output is termed influence costs (Milgrom and Roberts 1988). In response, the organization should seek an optimal balance between incentives in promotion and incentives in actual positions, as well as credibly modifying the criteria to reduce the attractiveness of positions awaiting promotion. Though improved transparency by adopting clearer rules and criteria is generally considered an organizational asset, it also reduces the cost of identifying whom to influence. If the latter effect dominates, marginal improvements in transparency may well lead to an increase in influence activities (Bac 2001).
Public organizations, including government and most nonprofits, differ from profit-seeking organizations in both fundamental and qualitative respects. The differences are reflected in their incentive structures and hierarchy. It is generally observed that members of organizations with social or political missions face weaker direct incentives than the employees of a firm. This is partly explained by the higher degree of congruence in the goals of the social or political organization and its members, which reduces the need for powerful incentives. However, the prime explanation is in the feasibility of direct incentive schemes. Public organizations pursue multiple goals which typically are hard to measure and even define. Some of their outputs and services are laden with amenity potential, contributing directly to the owners’ and top managers’ utility (Demsetz and Lehn 1985). Multitask models of hidden action have shown that when some goals are measurable and others not, an incentive scheme based on the measurable goals distorts actions to the detriment of the non-measurable goals and, hence, may well worsen the overall performance (Holmstrom and Milgrom 1991). If it does, the organization should reduce the power of direct incentives and possibly rely more on alternative instruments of motivation such as monitoring and emphasizing promotion and career concerns (Holmstrom 1999). A more radical response is to create independent units with clear and distinct goals or, if possible, isolate and confine the measurable goals to new specific divisions where high-powered incentive schemes become feasible and effective. Indeed, units and agencies of government are often characterized by independent missions (Wilson 1989; Tirole 1994). Multi-divisional frameworks may generate a loss of coherence in the overall decision-making process, but the benefits from subjecting the units to different masters under tight systems of checks and balances along financial, operational, and implementation dimensions can be notable. In essence, if incentivized to properly defend the cases for alternative causes, competition among advocates of specific goals and interests in the organization may give rise to good policy setting, similar in function to the adversarial adjudication system in courts of law.
Another threat against which all organizations must carefully safeguard is the possibility of collusion, a formation of an informal coalition to promote the common interests of its members (Tirole 1986). These organizations inside the organization are products of the information problems, combination of hidden action, hidden knowledge, and non-verifiability. Internal collusion comprises members only, whereas external collusion, named “capture” in the regulation context, involves outsiders as well. A supervisor who conceals an employee’s corruption for a benefit in return is internally colluding. Favoritism in transactions with outsiders and the capture of a procurement officer’s or an environmental agency’s decision by a private firm are incidences of external collusion. Internal and external collusion can coexist, as in the case of a coalition of officials which organizes the collection and sharing of corrupt proceeds from the public. Collusion can be temporary, a spontaneous deal ex post at the occasion, or it can take a sustained form whereby the parties involved coordinate ex ante their actions toward the common collusive objective (Bac 1996).
Sustaining collusion is subject to the same qualitative problems the organization faces in sustaining itself. But the threat of collusion imposes new constraints on incentive schemes, the rule-discretion balance, and the span of control. Stricter decision rules reduce the need to rely on the private information of members, thereby, the stakes of outsiders whom they are transacting with. The cost of a shift to rules is the foregone benefit from occasional and selective use of discretion. The rule to auction all contracts, for example, works extremely well on the price dimension and minimizes the likelihood of capture at the cost of other contract dimensions such as quality, speed, reliability, and reputation of the contractor, which typically are non-measurable but observable by an expert, manager, or official. Alternative instruments to reduce the payoffs to collusion include raising the rewards to supervision and evidence of good performance, introducing rotation, and intensifying monitoring. The cures carry along their risks: monitors are subject to the same threat of collusion, unless replaced by incorruptible electronic accountability systems or cameras. Large rewards to human monitors can prevent collusion but are obviously costly and risk inducing extortion with fabricated evidence of wrongdoing. As for rotation policies, they prevent development of trust-based relations that help in sustaining collusive behavior, but trust is also an input for cooperation and productivity in teamwork.
Governance of internal and external transactions and acts is subject to labor, administrative, tort, and criminal laws. Internal administrative procedures are activated upon reports of embezzlement and misdemeanor such as harassment, rules on hiring and firing decisions as well as ensuring job safety standards are imposed by labor laws, and strict (vicarious) liability is instituted for torts and crimes committed by employees. Just as the internal rules of the organization serve to align the objectives of its members with those of the organization, these laws can be viewed as a response to the dissonance between the goals of the organization and the society. So, by holding firms strictly liable for the actions of their employees rather than imposing individual liability, the state de facto delegates monitoring and partially also enforcement of offenses to the firms. Because the firm is in close relation to its employees and better informed about its environment, it can, less expensively than the state, monitor, prevent, and sanction. The firm is then expected to align incentives and partially shift the liability to its employees, passing the sanctions along to those in violation of the law using its own instruments such as wage policy, firing, and other measures. The lower the ability of the employees to pay for the social damage of their acts and the more effective the firm is in sharing its liability with its employees, the stronger the case for strict liability laws (Posner 1999).
The dominant new institutional economic model of the large corporate organization views shareholders as principals and the board of directors as their agents who monitor the managers on behalf of the shareholders. As an intermediate layer between shareholders and managers, the board is not immune to collusion with management against the shareholders. The risk of collusion is in part remedied by the market for corporate control, in part by equity holdings of the directors. This model is criticized for neglecting the fact that organization is a distinct legal entity with rights on its own, protected by the state. It is challenged by new theories proposed to understand corporate governance with directors as the principals and the firm as a productive team where relationship-specific investments are emphasized (Blair and Stout 1999). From a legal point of view, shareholders’ primacy is questionable, and ownership of the shares does not entail ownership of the corporation, which is an autonomous legal entity. The board of directors is entrusted to act on behalf of and for the benefit of the shareholders, with a fiduciary duty to review the decisions and ensure that they serve the corporation’s interests, and a legal right to control the managers’ decisions. Directors are motivated not only by their extrinsic compensation package but reputational concerns as well. They act as a mediating principal balancing the competing interests of the team, which justifies their highly independent legal status to carry out this role effectively.
Theories of organization lack a unified approach in explaining the emergence and function of organization as well as its internal practices comparatively, in contrast with a comprehensive set of alternative modes of transacting. A key feature of these alternatives is that they, too, are “organized” to some extent. Markets and the spectrum of hybrid contractual arrangements such as alliances, partnerships, networks, franchising, and subcontracting practices are not spontaneous, improvised systems. They combine aspects of impersonal exchange and formal organizations, just as organizations combine internal markets, decentralization, and command. But the most important obstacle to the research program on a unified theory of organization is the lack of consensus on modeling-bounded rationality of actors whose decisions and choices shape the form and function of the organization.
- Alchian A, Demsetz H (1972) Production, information costs and economic organization. Am Econ Rev 62:777–795Google Scholar
- Blair M, Stout L (1999) A team production theory of corporate law. V Law Rev 85:247–328Google Scholar
- Chiappori PA, Salanié B (2003) Testing contract theory: a survey of some recent work. In: Dewatripont M, Hansen L, Turnovsky S (eds) Advances in economics and econometrics, vol 1. Cambridge University Press, New YorkGoogle Scholar
- Radner R (1992) Hierarchy: the economics of managing. J Econ Lit 30:1382–1415Google Scholar
- Sah RK, Stiglitz J (1986) The architecture of economic systems: hierarchies and polyarchies. Am Econ Rev 74:716–727Google Scholar
- Tirole J (1986) Hierarchies and bureaucracies: on the role of collusion in organizations. J Law Econ Organ 2:181–214Google Scholar
- Tirole J (1994) The internal organization of government. Oxf Econ Pap 46:1–29Google Scholar
- Williamson O (1985) The economic institutions of capitalism. Free Press, New YorkGoogle Scholar
- Wilson JQ (1989) Bureaucracy: what government agencies do and why they do it. Basic Books, New YorkGoogle Scholar