Principal Chooses Research Articles

Evidence Games: Truth and Commitment

An evidence game is a strategic disclosure game in which an informed agent who has some pieces of verifiable evidence decides which ones to disclose to an uninformed principal who chooses a reward. The agent, regardless of his information, prefers the reward to be as high as possible. We compare the setup in which the principal chooses the reward after the evidence is disclosed to the mechanism-design setup where he can commit in advance to a reward policy, and show that under natural conditions related to the evidence structure and the inherent prominence of truth, the two setups yield the same outcome. (JEL C72, D82, D83, K41)

Commitment vs. Flexibility with Costly Verification

We introduce costly verification into a general delegation framework. A principal faces an agent who is better informed about the efficient action but biased towards higher actions. An audit verifies the agent’s information, but is costly. The principal chooses a permissible action set as a function of the audit decision and result. We show that if the audit cost is small enough, a threshold with an escape clause (TEC) is optimal: the agent can select any action up to a threshold, or request audit and the efficient action if the threshold is sufficiently binding. For higher audit costs, the principal may instead prefer auditing only intermediate actions. However, if the principal cannot commit to inefficient allocations following the audit decision and result, TEC is always optimal. Our results provide a theoretical foundation for the use of TEC in practice, including in capital budgeting in organizations, fiscal policy, and consumption-savings problems.

Rewards and punishments: Informal contracting through social preferences

This paper develops a positive model of informal contracting in which rewards and punishments are not determined by an ex ante optimal plan but instead express the ex post moral sentiments of the arbitrating party. We consider a subjective performance evaluation problem in which a principal can privately assess the contribution of an agent to the welfare of a broader group. In the absence of formal contingent contracts, the principal chooses ex post transfers that maximize her social preferences. We characterize the incentives induced by the principal's preferences, contrast them with ex ante optimal contracts, and derive novel testable predictions about the way externalities are internalized in informal settings.

On the Reliability of Trusting

This paper presents a model of trust in which a principal chooses either to trust or monitor an agent who, in turn, chooses either to honor or exploit that trust. The principal's decision of whether to trust or monitor is based on the relative temptation an agent faces to exploit the principal's trust, which comprises two elements - the environmental incentives the agent faces and the personal characteristics of the agent. The model is used to develop a reliability condition that the principal uses to assess the likelihood that trust placed in an agent will be honored.

Transparency and Performance Evaluation in Sequential Agency

This study investigates the effects of transparency in a sequential moral hazard problem. The principal chooses the transparent or opaque organization, by which we mean that the action of the leader is observable to the follower or not. Compared to the opaque organization, the transparent organization imposes an additional incentive constraint on the follower but may relax the incentive constraint on the leader. Informativeness of the signal crucially affects the optimal transparency as well as the optimal incentive scheme. Transparent (resp. opaque) organization tends to be preferred when the optimal incentive contract exhibits joint (resp. relative) performance evaluation.

Enabling versus Controlling

Revenue sharing between principals and agents is commonly used to balance double-sided moral hazard. We provide a theory of how, when such revenue-sharing is optimal, a principal allocates control rights over decisions that either party could make. We show that the principal either keeps control over all such decisions, or gives up control entirely, and that this choice is aligned with whether the principal chooses to keep more or less than 50% of variable revenues. We explore how moral hazard, contractibility, and spillovers affect this choice. The theory helps explain whether professionals operate as employees or as independent contractors.The appendices for this paper are available at the following URL: http://ssrn.com/abstract=2628128

Principal‐agent and Peer Relationships in Tournaments

Effort provision in tournaments may depend on both social preferences towards the competitor and social preferences towards the principal who designed the tournament. In a laboratory experiment, we analyze agents' behavior in different tournament settings that vary the distribution of the prize between agents. The principal chooses between a low and a high guaranteed payment. Standard economic theory predicts the same effort provision in all situations. Instead, our results show that both the generosity of the principal and the strict separation between winner and loser increase the effort level in tournaments. Moreover, agents focus their preferences either on the principal or on the agent. Copyright © 2014 John Wiley & Sons, Ltd.

The Importance of Bureaucratic Hierarchy: Conflicting Preferences, Incomplete Control, and Policy Outcomes

Existing theories of legislative delegation feature spatial models in which a principal chooses a single agent to whom to delegate authority. In the canonical model, the ally principle holds – the principal picks the ideologically closest agent. However, elected politicians typically decide whether to delegate not to an individual but to an institution, which consists of many individuals with differing preferences. To improve on existing work, we model delegation with bureaucratic hierarchy. Our results show that hierarchy is sufficient to undercut the ally principle. Indeed, capturing the logic of delegation and its results requires incorporating agency structure, especially the costs of an agency head controlling her subordinates resulting in incomplete control, as different structures are associated with different policy outcomes even when the same people constitute a bureaucracy. We also demonstrate that integrating hierarchy has important implications for a wide range of considerations, such as agency structural choice and the measurement of agency ideal points.

Administrative Policymaking, Judicial Review, and the Optimal Level of Agency Bias

We develop a theory of policymaking between a principal, an agent, and an overseer. The agent increases the overall quality of policy outcomes through costly capacity investments. Oversight impacts agent investment incentives, but only if the policy bias of the agent does as well. The principal chooses whether or not to authorize the agent to make policy on her behalf in light of oversight dynamics. We show that anytime oversight is not an effective means of political control, the principal never benefits from a biased agent relative to an ally. When oversight is effective, the principal almost always benefits from agent bias. Based on the dynamics of policymaking oversight the ally principle fails for instrumental reasons. The principal benefits from trading off agent bias for higher capacity investments. The results have implications for bureaucratic personnel politics, political appointments, and the efficacy of managerial strategies.

Minimum payments and induced effort in moral hazard problems

Consider a moral hazard problem in which there is a constraint to pay the agent no less than some amount m. This paper studies the effect of changes in m on the effort that the principal chooses to induce from the agent. We present sufficient conditions on the informativeness of the signal observed by the principal and on the agentʼs utility under which when m increases, induced effort (and hence productivity) falls. We also study how the cost minimizing contract for any given effort level varies in m. We present an efficient algorithm for numerically calculating optimal contracts for given parameters and show that induced effort falls when m is increased in many cases even when our sufficient conditions fail.

Experience Benefits and Firm Organization

Abstract A principal needs a worker for the production of a good. The worker can be hired as an internal agent, or an external agent under a contract. These two organizational modes correspond to in-house production and outsourcing, respectively. In each case, the agent earns experience benefits: future monetary returns from managing production, reputation, and enjoyment. The principal would like to extract experience benefits, and can do so when production is outsourced. However, the external agent earns information rent from private information about production costs. The principal cannot fully extract experience benefits when production is in-house because the internal agent must be provided with a minimum income, although the principal has full information on production costs. Our theory proposes a new trade-off, one between information rent under outsourcing and experience rent under in-house production. The principal chooses outsourcing when experience benefits are high, but her organizational choice may be socially inefficient.

Multidimensional Procurement Auctions with Unknown Weights

This paper studies the consequences of holding a procurement auction when the principal chooses not to show its preferences. My paper extends the procurement auction model of Che (1993) to a situation where both the principal and the agents have private information. Thus, unknown parameters of both the principal and the agents leads to unclear reaction strategies. I show that an unknown weight on the principal’s valuation of quality leads to the production of to much quality and to high informational rent. A problem that can be reduced using a revelation mechanism. Having an unknown weight on quality gives rise to an analysis of a principal that can not fully commit to the outcome induced by the scoring rule. Therefore, my result apply to contract theory and it’s problems with imperfect commitment.

Authority and communication in the laboratory

We report findings from experiments on two delegation–communication games. An uninformed principal chooses whether to fully delegate her decision-making authority to an informed agent or to retain the authority and communicate with the agent via cheap talk to obtain decision-relevant information. In the game in which the delegation outcome is payoff-dominated by both the truthful and the babbling communication outcomes, we find that principal-subjects almost always retain their authority and agent-subjects communicate truthfully. Significantly more choices of delegation than of communication are observed in another game in which the delegation outcome payoff-dominates the unique babbling communication outcome; yet there is a non-negligible fraction of principal-subjects who holds on to their authority and agent-subjects who transmits some information. A level-k analysis of the game indicates that a principal-subject “under-delegates” due to the belief that her less-than-fully-strategic opponent will provide information; such belief is in turn consistent with the actual play.

Pandering to Persuade

A principal chooses one of n>=2 projects or an outside option. An agent is privately informed about the projects' benefits and shares the principal's preferences except for not internalizing her value from the outside option. We show that strategic communication is characterized by pandering: the agent biases his recommendation toward good-looking projects -- those with appealing observable attributes -- even when both parties would be better off with some other project. Projects become more acceptable when pitched against a stronger slate of alternatives. We study organizational responses to the pandering distortion, such as delegation and choosing to be less informed.

Why Doesn't the Principal Help the Agent?

This paper endogenizes the strategic relationship between a principal (a player with authority) and an agent (a player without authority), and shows that a submissive relationship is the unique equilibrium outcome. That is, the agent voluntarily chooses a task that will increase the principal's payoffs, but the principal does not help the agent in the same way. Instead, the principal chooses a task that will decrease the agent's payoffs. This paper also endogenizes the (Stackelberg) leadership, and shows that delegation of leadership with a submissive relationship is optimal.

Corruption as Betrayal: Experimental Evidence on Corruption Under Delegation

We consider corruption behavior in a three-players game : Principal, Agent, Corrupter. When the Principal chooses a fair wage, the Agent faces con°icting interests to reciprocate. This delegation effect is expected to lower the level of corruption as compared to what arises in two-players settings. We set up two experiments varying in the exogeneity of the delegation relationship. The experimental evidence supports the delegation effect. This, in turn, could account for the deterrence effect of wage on corruption even in the absence of detection.

SUBJECTIVE EVALUATION IN AGENCY CONTRACTS*

We study how a principal uses her subjective evaluation of the agent’s performance in an incentive contract. It is shown that the subjective evaluation can be used either 1) when there is no other information about the agent’s performance and the principal is able to discard money, or 2) when the principal chooses between wage payment based on subjective evaluation by foregoing objective evaluation, and the one based only on the objective evaluation and when the subjective evaluation is sufficiently accurate. The principal pays a high fixed wage when her rating at the subjective evaluation is above a certain level. On the other hand, when it is below that level, she either pays a low fixed wage or obtains objective evaluation and pays based on its outcome.

Delegation versus an approval process and the demand for talent

In many organizations, a principal uses simple decision rules or decision processes, as opposed to complete contingent monetary contracts, to regulate a more informed agent's behavior. This paper models two such rules: delegation and an approval process. We determine the principal's preferable rule under different conditions and characterize the optimal demand for talent. In addition, we describe the agent's expected welfare under these rules and characterize how different talent might sort into the rules. If the decision process is not a choice or if the principal chooses approval for private gain, the demand for talent can be tied to the magnitude of loss in the worst state, rather than the degree of discretion given to the manager. However, if the principal has discretion over the decision process and proper incentives, then more talent is demanded under delegation.

Collusion and Organization Design

This paper investigates how the possibility of collusion between members of an organization affects the choice of organizational design. We consider a moral hazard environment in which a principal chooses the organizational structure before designing contracts. The principal has the choice between a principal–agent organization, where it monitors the output produced by the agent itself, and a principal–supervisor–agent hierarchy, where monitoring is delegated. We identify the conditions under which the possibility of collusion in a principal–supervisor–agent hierarchy entails the superiority of a principal–agent organization. The results suggest that the possibility of collusion may set a limit to the size of organizations.

Corruption as Betrayal: Experimental Evidence on Corruption under Delegation

We consider corruption behavior in a three-players game : Principal, Agent, Corrupter. When the Principal chooses a fair wage, the Agent faces con°icting interests to reciprocate. This delegation effect is expected to lower the level of corruption as compared to what arises in two-players settings. We set up two experiments varying in the exogeneity of the delegation relationship. The experimental evidence supports the delegation effect. This, in turn, could account for the deterrence effect of wage on corruption even in the absence of detection.