Applying agency theory to public administration (government)

Reading E. Fivat’s dissertation proposal made me think about the appropriateness of applying agency theory to public administration.

Developed for for-profit corporations in mind (Ross, 1973; Meckling & Jensen, 1976; Fama, 1980), agency or principal-agent theory argues that the interests of principals and agents are intrinsically misaligned, due to the theory’s rationality assumption, i.e. individuals are opportunistic in that they pursue their self-interest sometimes at the expense of others. It is also assumed that principals – the owners of corporate assets –  are risk-neutral whereas agents – who are employees of the corporation- are depicted as risk-averse. Principals being the corporate owners they are considered to be the legitimate residual claimants, i.e. the legitimate claimants of the rents which remain after all agents and other actors (suppliers) are paid.

Thus, in for-profit corporations, whether publicly-traded or private , due to their position as residual claimants shareholders (principals) in general assembly elect their representatives in the board of directors – sometimes some of themselves, when main or block shareholders exist such as in L’Oréal, Bouygues or LVMH. Thus, directors are agents of the shareholders. Then, directors appoint the Chief Executive Officer (CEO), the key executive agent, to whom the board delegates executive powers.

Given principals’ (and directors’) lack of knowledge, they need to delegate task execution to agents; the act which, given the opportunistic assumption, creates the agency problem: how to guarantee that agents will perform the task with the principals’ interest in mind. Agency theory makes two propositions:

1. to align the interests of principals and agents through a variable remuneration scheme for the agent which takes into account the principals’ objectives (e.g. profitability).

2. in order to deal with situations in which interest alignment might not be sufficient, to exert potent monitoring of the agent (namely the CEO) through a strong, independent board of directors. The theory argues and the empirical evidence shows to some extent that ownership fragmentation prevents effective monitoring (e.g. Amihud & Lev, 1981).

Now, how would agency theory apply to public administration?

In democratic systems, the (voting) population is sovereign and thus the principal (the people), which elects individuals to represent it as well as directly or indirectly lead each public administration. Thus, elected politicians who govern public administrations (prime ministers, presidents, and mayors) cannot be considered principals as shareholders are in private, for profit corporations but agents of the (voting) population.

Further, in contrast to shareholders, elected politicians are not the residual claimants of the “rents” generated by their organizations because they are not the owners of these organizations and their assets. By virtue of democratic elections, elected politicians are the agents of the (voting) population which is the sovereign owner of the assets of the collectivity. Approval of the use of administration’s assets (budget) requires parliamentary vote, that is, again the validation by the elected representatives of the population.

In fact, elected politicians are like directors in for-profit corporations, the delegates or agents of the owners. Politicians are elected on the basis of an electoral program which they are expected to implement if elected, and which is a kind of contract between the population and the elected politician.

But how does the population as principal deal with the agency problem? Elected politicians can also be opportunistic, as history and current events demonstrate in many latitudes.

How does the population control its agents (elected politicians)? Through elections? A General Accounting Office or an Ombudsman (“Defender of the People” in some countries)? But these organisms tend to be parliamentary and thus their heads are appointed by elected representatives.

How does the population (principal) try to align its interests with those of its agents, elected representatives and governors? Variable compensation doesn’t seem to be much used.

Also, is the parallel of ownership fragmentation – fragmentation of representatives across many political parties – also a problem for effective monitoring of governments? Of course, there is also Montesquieu’s separation of powers – judicial, parliamentary and governmental – which is supposed to generate a system of checks and balances. But if Supreme or Constitutional judges are appointed by parliaments then it seems that elected representatives have direct control on the other two powers. It would also be interesting to examine whether the separation of powers could be reproduced for corporations and non-profit organizations.

E. Fivat. (2011). Le rôle de la confiance et son impact sur l’autonomie et le contrôle dans différents modes de gouvernance. IDHEAP, Lausanne.