In this chapter we return to electoral competition, but we shift focus from elections as a mechanism for aggregating individual preferences to elections as a means of holding officeholders accountable to voters. The literature on political agency that we survey here has its roots in the economic theory of contracts, where the central question is how contracts govern the relationship between principals (e.g., employers or shareholders) and their agents (e.g., employees or CEOs). In models of political agency, voters play the role of principals; their agent is a politician elected to govern on their behalf.

In general, voters face two agency problems. First, elected politicians may not bear the full cost or receive the full benefit of their decisions, which can give them an incentive to take actions that are not in voters' best interest. Elections can help voters to alleviate this problem of moral hazard by providing them the opportunity to sanction politicians who have behaved poorly. Second, politicians may differ in their competence, preferences, or other characteristics. Elections can encourage the selection of politicians who are inclined to act in voters' best interest.

Although elections provide the means by which voters can address these agency problems, two obstacles stand in the way of full accountability. First, voters may not be able to directly observe the actions of elected officials, politicians' characteristics, or the mapping from policy to outcomes.