The principal-agent problem is a common problem that arises whenever there is a contractual relationship between two parties – the principal, who in accordance with agreed upon terms, assigns a task to an agent, who then executes the task for the principal. It is expected that the agent will work on the behalf of the principal. Likewise, it is expected that the principal will be honest with the agent. However, this is not always the case, ergo a problem is created of which there are two main types.
The first is an ex-post hazard which occurs after the closing of a contract. Without constant monitoring, it is possible for the agent, who with the right motivation and ability, can covertly work against the interests of the principal. A moral hazard results from this when the socially beneficial consequences are removed as a result of a lack of essential information needed to spur good behavior or discourage bad behavior. In other words, a party in the contract who is protected from some form of risk engages in behavior that they would otherwise not engage in, e.g. engaging in risky behavior by insurance holders or apathetic teachers who have received tenure. Solutions to such a problem can take the form of, for example, more sophisticated contracts or institutional arrangements; supplemental contractual components, e.g. fixed payments or performance pay; trust building; and third party involvement, e.g. insurance. However, it is important to note that the agent will likely always have information rent, i.e they will always have an informational advantage.
The second type of hazard is ex-ante and occurs before the closing of a contract. In this type of situation, the principal has the option to structure the agreement in their favor by keeping relevant and potentially damaging information from the agent. This results in adverse selection which creates an imbalance of power in favor of the party with more information. Such a problem can be rectified using, for example, a menu of contracts; additional contractual components, e.g. buy-back options; trust development; the involvement of third parties, e.g. labeling or licensing; and signaling.
In more simple terms, the principal-agent problem represents the fact that ownership ≠ control. Accordingly, institutions can need to be used in order to establish as equitable of a relationship between agent and principal as possible.