Optimal Information Asymmetry

Abstract

At the heart of decentralization lies the notion that tasks are delegated by owners to managers who possess superior local information. The extent of this information asymmetry is often an endogenous construct, as it is influenced by the owner's choice of internal accounting systems and the manager's investment in acquiring local expertise. In this paper, we explore how varying levels of pre-contract, asymmetric information affect the owner-manager relationship. We provide three main sets of insights. First, we find that the owner's payoffs are initially decreasing, and strictly convex everywhere, in the quality of the manager's private information. The owner thus prefers to deal with either a perfectly informed or a perfectly uninformed manager, and we characterize conditions for either to be the preferred choice. Second, in contrast to recent work, we demonstrate that when information can be communicated internally, the optimal strength of managerial incentives unambiguously decreases as the manager becomes better informed. Third, we derive the surprising result that a self-interested manager does not always prefer to maximize his informational advantage. Our work has implications for the optimal design of organizations, and for internal accounting and control systems in particular.

Does a Manufacturer Benefit from Selling to a Better-Forecasting Retailer?

“…Lewis and Sappington (1991) and Rajan and Saouma (2006) consider a principalagent model where the agent privately exerts effort that influences the output. The agent has private information about his cost of effort.…”

Optimal pricing contracts and level of information asymmetry in a supply chain

“…For example, in a principal‐agency setting, Lewis and Sappington () investigate the principal's preference on the information accuracy available to her agent by extending the standard agency model to allow the principal to choose the probability P with which the agent receives perfect private state information. They show that the principal prefers “ P = 0” or “ P = 1,” that is, “the agent does not know anything about the private state information, or the agent knows all about the private state information.” Although in different settings, Rajan and Saouma (), Taylor and Xiao (), and Chen and Xiao () derive similar results as those in Lewis and Sappington (). All these studies focus on the principal's preference on the agent's forecasting capacity to learn his own private state information, while our study mainly focuses on the principal's (i.e., the supplier) preference on the level of information asymmetry between the principal and the agent given P = 1.…”

“…Essentially, when the principal prefers “ P = 1,” the agent perfectly knows the private state information and thus he can make more “efficient” decision (planning) that is beneficial to the whole system and the principal, though the principal has to pay a “higher information rents” to the agent; however, when the principal prefers “ P = 0,” although the agent's imperfect private state information is “not” particularly valuable for planning purposes, the principal does not have to pay “significant” rents to the agent. This explanation can also be applied to the results in Rajan and Saouma (), Taylor and Xiao (), and Chen and Xiao (). Second, these studies do not take the cutoff level policy into consideration.…”

Incentive Contracts, Market Risk, and Cost of Capital

“…By stewardship, I mean that the information is useful to resolve an agency problem between owners and one or more agents whose actions and information might not be contractible; see, for examples, Liang (), Christensen, Feltham, and Şabac (), Rajan and Saouma (), or Nan ().…”