There are many ways that governments can fail to achieve good policy outcomes. Some interest groups
may have too much power or want the wrong things, or leaders may pursue personal goals that do not
have broad public support. Government agencies may also pursue parochial goals or fail to coordinate
with each other.
Most of the government failures that we might overcome, however, are arguably due to a lack of
information about the consequences of policy choices. Policymakers may have mistaken beliefs about the
direct effects of a policy. They may fail to consider an indirect effect or an alternative policy. Or we
might fail at a higher level to choose the right institutions to set the context for such judgments and
choices. In fact, much of what economists have been doing for the past half century is showing how most
social problems are at root caused by such information failures. For example, there is almost no end to the
troubles caused by the principal-agent problem—that is, the fact that experts know more than their
clients.
We rely on doctors, lawyers, teachers, and chief executive officers because they can know more
than we do. Government policy similarly relies heavily on expert knowledge, expressed directly via civil
servants, contractors, and official boards and commissions, and indirectly via mass media and academic
reports and analyses.
We hope that such experts acquire and use their knowledge for our benefit, but we fear that they
instead give us the advice that benefits them. Stereo salesmen, for example, may suggest overly expensive
stereos, and surgeons may recommend surgery over drug treatments. We cannot trust such experts
because of “professional ethics"; if they are trustworthy, it is because their institutional and contractual
incentives make them so.
Unfortunately, agency problems in the government-citizen relationship are among the worst,
because the chains of delegation are so long, the topic coverage is so broad, and citizens have so little
reason to pay attention. Each election, voters, with little expertise or chance of personally influencing the
outcome, form judgments about how the policy outcomes they favor relate to the candidates before them.
Representatives, anticipating those judgments and any other voter preferences, oversee the agencies that
2
Convention