The paper by Bengt Holmstrom, published in *The Bell Journal of Economics* in 1979, explores the role of imperfect information in principal-agent relationships subject to moral hazard. Holmstrom derives a necessary and sufficient condition for imperfect information to improve on contracts based solely on payoffs and characterizes the optimal use of such information. The paper begins by discussing the common problem of moral hazard, where individuals' actions affect the outcome probability, leading to a need for incentives to ensure proper behavior. Holmstrom argues that while complete monitoring is often impractical or costly, imperfect information can be used to mitigate moral hazard. He presents a model where the principal observes only the outcome, and the agent's action is not directly observable. The optimal sharing rule between the principal and agent is derived, showing that any additional imperfect information about the agent's action can improve welfare. The paper also addresses the value of imperfect information, proving that a signal is valuable if it provides new information about the agent's action, which can be used to design better contracts. Holmstrom concludes with a discussion on the implications of his findings for practical applications, such as insurance and labor contracting.The paper by Bengt Holmstrom, published in *The Bell Journal of Economics* in 1979, explores the role of imperfect information in principal-agent relationships subject to moral hazard. Holmstrom derives a necessary and sufficient condition for imperfect information to improve on contracts based solely on payoffs and characterizes the optimal use of such information. The paper begins by discussing the common problem of moral hazard, where individuals' actions affect the outcome probability, leading to a need for incentives to ensure proper behavior. Holmstrom argues that while complete monitoring is often impractical or costly, imperfect information can be used to mitigate moral hazard. He presents a model where the principal observes only the outcome, and the agent's action is not directly observable. The optimal sharing rule between the principal and agent is derived, showing that any additional imperfect information about the agent's action can improve welfare. The paper also addresses the value of imperfect information, proving that a signal is valuable if it provides new information about the agent's action, which can be used to design better contracts. Holmstrom concludes with a discussion on the implications of his findings for practical applications, such as insurance and labor contracting.