Moral hazard problem:

As  apparent  from  the above discussion, we  speak of moral hazard when  the agent  takes  an  action that affects her utility  and  the utility  of  the principal.

Remember  the  driver's insurance example presented in the preceding unit. After being insured, a driver may  drive recklessly as damage  in  the event of an accident will be borne by  the  insurance company. The net  result is change in  the utility of  both  the  parties  to  insurance. Mark  the  information imperfection in  such a behaviour. The principal does not observe the agent's action but only the outcome, which is an imperfect signal of this action. The action chosen by the agent is not Pareto optimal.

The  principal will  try  to  influence  the action of  the  agent by  conditioning agent's utility on  the only variable, which  is  observable, viz., the  outcome. You  can add  a number  of  examples describing this  kind of behaviour apart from the one cited in case of driver's insurance.  

Examples

• Insurance companies cannot observe the self-protection efforts of the insured;
• Patients cannot ascertain the effort put in by a doctor to find out the causes of a disease.
• A  landowner cannot observe the efforts of a tenant while cultivating the land.
• In a firm the manager cannot observe the efforts of the workers.
• The shareholders are uncertain about the manager's efforts in  profit maximisation.

If we try for a first-best (i.e., Pareto optimal) solution of moral hazard problem and follow the principal-agent framework, we may assume that the principal can observe agent's action. In that formulation, the principal orders the agent to take the optimal action  and will choose the payment, which achieves the optimal risk sharing.

To appreciate the above option, let  us  consider the employer-employee example. We will follow the presentation module, sequence and  notation of Meunier,  Valerie 2005 for this part of the discussion.