Principal-Agent Theory

Key Concepts and Terms

• Principal-agent problem. The central dilemma investigated by principal agent theorists is how to get the employee or contractor (agent) to act in the best interests of the principal (the employer) when the employee or contractor has an informational advantage over the principal and has different interests from the principal. Sappington (1991) provides a discussion of principal-agent incentive problems.

  • Privatization as a solution to the principal-agent problem. Principals must balance agency costs against costs of debt financing and other costs associated with not separating ownership (sovereignty in the public sphere) from control. The issue of privatization of government services hinges not only on the relative production costs in the public vs. private sector, but also on agency/transaction costs. Specifically, the issue hinges on finding the optimum point of privatization where the marginal total cost (agency costs plus production costs) equal their marginal benefits. This is made more complicated by the fact that there are many forms of separation of ownership/sovereignty from control, including many forms of contracting out, which can carry high agency costs. Indeed, contractors may be in-house governmental agents, not representing privatization at all. See Jensen and Meckling (1976). The contract state is a reference to one model for solution of the principal-agent problem, namely radical contracting-out of otherwise governmental services. Public administration in New Zealand is often taken as a model.

• Agency costs are a type of transaction cost, reflecting the fact that without incurring costs, it is impossible for principals to ensure agents will act in the principals' interest. Agency costs include the costs of investigating and selecting appropriate agents, gaining information to set performance standards, monitoring agents, bonding payments by the agents, and residual losses. Agency costs have policy implications.
  1. Information costs in contract management. In addition to recognizing that contract management involves agency costs, one may also observe that the informational advantage of the contractor regarding performance means that the contractor may be able to impose high agency costs by resisting the principal's effort to gain information. That is, there is information asymmetry in favor of the agent, who regards information as a source of power and therefore hoards and guards it (Waterman & Meier, 1998). The more difficult for the principal to gain information on performance outcomes, the more likely that contracts will be framed instead in terms of contractor behavior. The more uncertain the outcomes, the more the agent will have an incentive to resist the principal's information-gathering efforts so as to encourage behavioral rather than outcome performance standards. This is a form of the effectiveness (outcome) vs. efficiency (behavior) distinction common in administrative literature.
     • Example: Mahler and Regan's (2005) study of the effect of strong agency Internet presence on congressional oversight was related by the authors to principal-agent theory. By reducing the costs of information-gathering to the principal because of the Internet, Mahler and Regan found, control of the agent (the agencies and agency outcomes) became easier and more effective.

     • Example. In 2006 the Lincoln Group, which contracted with the Department of Defense to place pro-American articles in the Iraqi Press, was revealed to have had a policy of revealing to the DoD as little as possible about their procedures and capabilities (Marx, 2006). The policy was based on LG's perception that the more DoD knew about uncertainties and weaknesses in the company's implementation, the more present and future contracts would be endangered.

  2. Goal incongruity between principal and agent increases the incentive of the agent to withhold information from the principal (Simonsen and Hill, 1998).

  3. Agent risk aversion. Some agents are more risk-aversive than others, perhaps due to organizational culture. Risk aversive agents are more prone to withhold information from principals, increasing agency costs.

  4. Interdependence can also make processes more complex and uncertain, in turn increasing agency costs of obtaining performance information.

  5. Communication costs. The agent may not follow the intent of the principal when there has been insufficient investment of time and personnel, and investment in communications channels, by the principal, resulting in lack of clarity and/or consistency of messages from the principal (Goggin, Bowman, Lester, and O'Toole, 1990).

• Simon on agency costs. Though not considered a "principal-agent theorist," Herbert Simon, a founder of modern management theory, had much to say about the role of agents and agency costs. Simon emphasized that agents faced information costs in the present and uncertainty about the future, limiting their decision-making ability and forcing them to make decisions by seeking the first satisfactory solution rather than optimizing (which would require full information). Where classical "economic man" knew all alternative courses of action and their consequences, had a stable preference function for all consequences, and made optimal decisions by maximization, "administrative man" "satisficed" under "bounded rationality." See Herbert Simon's classic Models of Man (1957).

• Management control theory deals in part with the principal-agent problem viewed as a control problem. Control may be after-the-fact or before-the-fact -- that is, governmental actions after the consequences of the agent's actions are known, or prior to the agent taking action. The latter is, of course, traditional budgeting and regulation. Principal-agent problems are common in budgeting (Smith & Bertozzi, 1998). Control theory, however, suggests that where consequences are homogenous and therefore are easily monitored, after-the-fact controls are more effective. Where consequences are unique and hard to monitor, before-the-fact control is appropriate. By implication, public administrators should employ traditional before-the-fact regulation only in a relatively small subset of activities, but should implement after-the-fact controls in many areas, typically through contractual arrangements which specify consequences desired by the principal and attach related after-the-fact rewards and penalties to the agent.
  • Demand-revealing control mechanisms. As Thompson (1998) notes, before-the-fact and after-the-fact controls both ultimately rely on incentives and sanctions for their effectiveness. However, after-the-fact controls are demand-revealing mechanisms in that they make incentives clear to the agent and allow the agent to determine how best to satisfy the demand schedule that is presented openly. In contrast, before-the-fact controls often are demand-concealing, whether appropriation mandates to governmental organizations to perform services or flexible pricing (cost-plus contracts) of the military sector type. Often before-the-fact controls do not allow the agent -- governmental or private -- to gain a full view of the demand schedule, without which information efficient mobilization of resources is impossible, and even more so when before-the-fact controls impose constraints on the type, manner of use, time of use, etc., of resources.

• Theory of corruption. Governmental corruption can be analyzed in terms of principal-agent theory (Aidt, 2003). Bribery can be seen as a market transaction cost which occurs when public officials have discretionary power (often due to weak institutional constraints) to extract rents from those seeking governmental services, such as licenses and permits. Corruption can be efficient by providing a (black) market mechanism by which the agent (the official) creates a market of escalating bribes such that the service-seekers (the agents) most in need bid the most in bribes and get the services first. Moreover, bribery may substitute for other less efficient means by which the agent may expropriate public wealth, such as direct personal confiscation.

  Efficient corruption also speeds service delivery when the "normal" situation is one of broken bureaucratic processes. Thus in a study by Leff (1964), corruption served to increase food production under price control in Brazil, whereas production stagnated under price control in Chile, where officials were not corrupt and enforced rules. However, corruption may also be inefficient insofar as the objectives of the agents do not reflect social welfare. Corrupt agents may supply government services to maximize bribes, not maximize service delivery. Unlike open auctions, inefficiencies in service allocation under corruption may also arise from the expenditure of resources to keep corrupt deals secret. Moreover, the collection of rents accrues to the agent, not to the government (the principal), and these revenues are not available for reinvestment in more efficient means of service delivery. In the case of corrupt tax officials, government revenues may be reduced considerably. Also, corrupt officials may take steps to assure the continued inefficiency of "normal" service delivery precisely so that citizens must use bribes to obtain needed services. Then from the citizen point of view, corrupt deals lack security because contracts cannot be enforced in courts, and lack of security impedes efficient planning by entrepreneurs. For the principal (government) to regain control over the agent (the corrupt official), monitoring/enforcement must be increased and salaries of officials must be raised to the point where fear of dismissal outweighs hope for profit through bribery. However, costs of monitoring and costs of higher salaries may prove to be great enough to actually reduce overall government revenues compared to relying on low-paid, corrupt agents.

  • Non-benevolent principals. The theory of corruption changes when one assumes both the principal (the government) and the agent (the official) and non-benevolent and corrupt. In such a context, there is no incentive for efficient allocation of government services. Shleifer and Vishny (1998) in their book, The Grabbing Hand, take this viewpoint, which assumes all government officials will be corrupt given the opportunity. Governments adopt economic policies with a view to creating more opportunities for corruption, not to remedy inefficiencies. Increasing the salaries of officials is not a remedy for corruption because there is no risk of being caught since higher officials are corrupt also, not seeking to enforce rules. Higher officials establish monitoring mechanisms only for symbolic reasons, not substantive control.

  • The effect of democratic reforms. The choice is not only between the two models of benevolent principal/corrupt agent and corrupt principal/corrupt agent. In systems marked by separation of powers and/or decentralization of powers, it is possible some principals will be benevolent and others corrupt. Separation of powers, decentralization, and transparency in goverment measures can mitigate corruption. However, such measures only work to the extent that power is shifted from corrupt to benevolent principals. For instance, decentralization to local levels may shift power to more democratically accountable and benevolent principals, or to more corrupt principals more under the control of local interests. Moreover, decentralization of spending decisions while continuing central control of revenue collection is likely to increase corruption, since the interests of central principals and local agents will diverge more sharply.

Assumptions

• Economic motivation. Principal-agent theory rests on a basis in economics, with the relationship between the principal and the agent defined by the contract. Relationship problems are approached through a contractual perspective; solutions lie in changing the contract. In evaluating writings from this perspective, consider if adequate emphasis has been given to non-economic considerations: cultural, social, psychological, political. (For instance, Wood & Waterman [1994] note that reponsiveness of the bureaucrat to political control is the norm, not the exception, even though political controls are not part of the employee contract).
  • Negative view of agents. By and large, principal-agent theory tends to assume agents have an innate tendency to seek autonomy from organizational rules, to minimize the burden of discharging responsibilities, and to hoard rather than pass through information (since information is power) (Frederickson & Smith, 2003).

• Acultural dynamics. Principal-agent theory can ascribe an over-simplistic Skinnerian psychology to agents, emphasizing their responses to stimuli from the agent and overlooking factors associated with organizational culture. Jones (2003: 404) calls the Skinnerian underpinnings of principal-agent theory "antiquated" and leading to the neglect of important non-contractual aspects of bureaucratic relationships. That is, control may rest as much or more on organizational culture as on the principal-agent contract and related incentives and disincentives to the agent. Brehm & Gates (1993), for instance, noted "the overwhelming importance of attributes of the organizational culture" in their study of supervision/control over police behaviors.

• Descriptive vs. normative theory. Principal-agent theory can be descriptive theory (as in small group experimental studies) or it can be normative theory (prescribing contracting-out). These two poles of principal-agent theory are not necessarily compatible. As descriptive theory, one must approach empirically the question of whether after-the-fact controls are more effective in all settings (cf., Schick, 1998). As normative theory, principal-agent theory is an umbrella for dissimilar bedfellows ranging from radical Milton Friedman-type privatization to quite moderate reforms involving performance contracting within traditional governmental organizations.
  

Illustrative Hypotheses

• The more ambiguous the relation of means to ends, the greater the information advantage of the agent. The greater the agent's information advantage, the higher the information costs to the principal.

• The more ambiguous the relation of means to ends, the less effective a priori regulation and rules stipulated by the principal.

• The more complex the process (ex. the more autonomous and semi-autonomous units involved), the higher the information costs.

• The lower the information costs to the principal, the better the oversight of the agent. The more the oversight of the agent, the better the performance.

• The higher the information costs, the more likely the contract will be specified in terms of agent efficiency rather than agent effectiveness.

• The more risk-aversive the agent, the more the agent will increase information costs to the principal.

• The more transparency provisions are written into the contract, the lower the information costs.

• The more performance outcomes are written into the contract, the more successful the outsourcing.

• The more the agent puts up in performance bond, the better the performance.

• The more the goals of the selected agent are similar to those of the principal, the more outcomes will be those sought by the principal.

• The better the information provided by past agents, the more effective the performance standards framed by the principal.

• The less the investment in principal-agent communication channels, the more the goal incongruity between principal and agent.

• The more the goal incongruity, the greater the information costs.

• The higher the cost of information, the more the principal will satisfice rather than optimize.

• Raising salaries of officials reduces corruption in states characterized by benevolent principals and corrupt agents, but not in states characterized by corrupt principals and corrupt agents.

• Decentralization reduces corruption only when central officials are less benevolent and democratically accountable than are lower officials.

Frequently Asked Questions

• Where can I find materials on principal-agent theory?
• Is principal agent theory the same as intelligent agent theory (social agent simulation)?
  
  
  
  
• Where can I find materials on principal-agent theory?
  Principal-agent theory has been covered frequently in The RAND Journal of Economics, among others. For a critique of principal agent theory as applied to the U.S. budget process, see A Critique of Principal-agent Delegation Theory: Us Budget Process Reforms 1985-1996, by Jasmine L. Farrier.

  
  

• Is principal agent theory the same as intelligent agent theory (social agent simulation)?
  No. Intelligent agent theory or social agent simulation is a branch of artificial intelligence which uses computer simulation. These simulations are agent-oriented, so there is some relationship to principal agent theory, but only in a tenuous way.

  
  

Bibliography

• Aidt, Toke S. (2003). Economic analysis of corruption: a survey. The Economic Journal 113(491): 632-652.
• Althaus, C. (1997). 'The application of agency theory to public sector management. In G. Davis, B. Sullivan & A. Yeatman, eds., The New Contractualism?, eds. Centre for Australian Public Sector Management, pp. 137–153.
• Ackere, A.(1993). The principal/agent paradigm:Its relevance to various functional fields. European Journal of Operational Research, Vol. 70: 83-103.
• Arrow, Kenneth, (1985). The economics of agency. Pp. 37-51 in J. Pratt and R. Zeckhauser, eds., Principals and agents: The Structure of business, Boston: Harvard University Press.
• Boston, J., Martin, J., Pallot, J. and Walsh, P. (1996). Public management: The New Zealand model. NY: Oxford University Press.
• Brehm, John & Gates, Scott (1993). Donut shops and speed traps: Evaluating models of supervision on police behavior. American Journal of Political Science 37" 555-581.
• DiIulio, J. (1994). Principle agents: The cultural bases of behavior in a federal government bureaucracy. Journal of Public Administration Research and Theory 4: 277-318.
• Frederickson, G. H. & Smith, K. B. (2003). The public administration theory primer. Boulder, CO: Westview Press.
• Goggin, M., Bowman, A., Lester, J., & O'Toole, L. (1990). Implementation theory and practice: Toward a third generation. Glenview, IL: Foresman.
• Grossman, Sanford J., and Oliver D. Hart (1983). An analysis of the principal-agent problem. Econometrica, Vol. 51: 7-46.
• Harris, M. and A. Raviv (1978). Some results on incentive contacts with applications to education and employment, health insurance, and law enforcement. American Economic Review 68: 20-30.
• Horn, M. J. (1995). The political economy of public administration: Institutional choice in the public sector, Cambridge, UK: Cambridge University Press.
• Jensen, Michael C., and William H. Meckling, (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, Vol. 3: 303-360.
• Jones, B. D. (2003). Bounded rationality and political science: Lessons from public administration and public policy. Journal of Public Administration Research and Theory 13(4), 395-412. (Critical of principal agent theory).
• Laffin, Martin (1997). Understanding minister-bureaucratic relations: Applying multitheoretic approaches in public management. Australian Journal of Public Administration Vol 56, No. 1.
• Leff, N. (1964). Economic development through bureaucratic corruption. American Behavioral Scientist 8(3): 8–14.
• Mahler, J. G. and Regan, P. M. (2005). Agency Internets and the changing dynamics of congressional oversight. International Journal of Public Administration, 28(7/8), 553-565.
• Marx, Wilhelm (2006). "Misinformation intern: My summer as a military propagandist in Iraq. Harper's Magazine 313(1876):
• Pratt, J., and R. Zeckhauser (1985). Principals and agents: The structure of business. Boston: Harvard Business School Press.
• Proust, E. (1997). Implementing the contract state. Australian Journal of Public Administration, Vol. 56, No. 3: 132–134.
• Ress, R. (1985). The theory of principal and agent. Bulletin of Economic Research,Vol. 37,No.1.
• Robins, J. A. (1987). Organizational economics: Note on the use of transaction-cost theory in the study of organisations. Administrative Science Quarterly, Vol. 32: 68–86.
• Ross, Steven, (1973). The economic theory of agency: The principal's problem. American Economic Review, 63(2): 134-139.
• Sappington, D. (1991). Incentives in principal agent relationships. Journal of Economic Perspectives 3(2): 45-66.
• Schick, Alan (1998) Why most developing countries should not try New Zealand's reforms', The World Bank Research Observer, Vol. 13, No. 1: 123–31.
• Shleifer, A. & Vishny, R. W. (1998). The grabbing hand, Government pathologies and their cures. Cambridge, MA: Harvard University Press.
• Simonsen, B. & Hill, L. (1998). Municipal bond issuance: Is there evidence of a principal-agent problem? Public Budgeting & Finance 18(4), 71-100.
• Smith, R. & Bertozzi, M. (1998). Principals and agents: An explanatory model for public budgeting. Journal of Public Budgeting, Accounting, & Financial Management 10(3): 325-352.
• Spence, M. and R. Zeckhauser (1971). Insurance, information and individual action. American Economic Review 61(2): 380-387.
• Stiglitz, Joseph (1987). Principal and agent. In John Eatwell, Murray Milgate, and Peter Newman, eds., The new Palgrave: A dictionary of economics. London: The Macmillan Press Limited.
• Thompson, Fred (1998). Public economics and public administration. Pp. 995-1063 in Rabin, Hildreth, & Miller, eds., Handbook of public administration. (Second edition) . New York: Marcel Dekker, Inc.Available online at http://www.willamette.edu/~fthompso/pubfin/ECON&PA.html.
• Waterman, R. W. & Meier, K. J. (1998). Principal-agent models: An expansion? Journal of Public Administration Research and Theory. 8(2): 173-202.
• White, William D. (1992). Information and the control of agents. Journal of Economic Behavior and Organization, Vol. 18: 111-117.
• Wood, D. & Waterman, R. (1994). Bureaucratic dynamics: The role of bureaucracy in a democracy/ Boulder, CO: Westview Press.