Free Political Order and Political Decay by Francis Fukuyama Page A

extract taxes from its population as a capacity measure, the same used to measure scope. The reason for this is that taxes, particularly direct ones like an income tax, are hard to collect and also represent resources at the government’s disposal. However, the ability of an organization to perform its functions is never simply a matter of measurable resources. Organizational culture also matters—the degree to which the individuals who make up the organization can function cooperatively, engender trust, take risks, innovate, and the like. A Weberian bureaucracy defined only by formal procedures may or may not have the intangible qualities necessary to make it function effectively.
    A different approach to measuring the quality of government would be to look not at what the government is, but rather at what it does. The purpose of government, after all, is not to follow procedures but to provide the population with basic services including education, defense, public safety, and legal access; an output measure like the degree to which children were being educated by the public school system would be more informative than data about teacher numbers, recruitment, or training. Lant Pritchett, Michael Woolcock, and Matt Andrews have argued that one of the big problems with developing countries’ governments is that they engage in what they term “isomorphic mimicry,” that is, copying the outward forms of developed countries’ governments, while being unable to reproduce the kinds of outputs, like education and health, that the latter achieve. 10 Measuring what the government actually does rather than how it does it would avoid this problem.
    Appealing as output measures are, however, they can be misleading. Good outcomes, like quality public education, are a complex mixture of inputs provided by governments (teachers, curriculum, classrooms, etc.) as well as characteristics of the population being served, including their income, social habits, and culture (that is, the degree to which learning is valued in the home). A classic study of educational outcomes in the United States was the 1966 Coleman Report, whose statistical analysis showed that quality education was much more a reflection of a student’s friends and family than of the inputs being supplied by the government. 11 In any event, measuring outcomes is often difficult for the kinds of complex services offered by modern governments. For example, how does one measure the quality of a judicial system? Clearly a gross output measure of the number of cases closed or the number of convictions is meaningless in the absence of qualitative measures of whether the courts were adjudicating cases fairly, or using torture to extract confessions. Absent such judgments, police states will always seem to perform better than those that adhere to a strict rule of law.
    In addition to considering procedural and output functions, there is a final dimension of government quality that is relevant when assessing the functioning of a state: the degree of autonomy that a government enjoys. All governments serve a political master, whether a democratic public or an authoritarian ruler, but they can be granted more or less autonomy in their ability to carry out their tasks. The most basic form of autonomy concerns the right to control the government’s own staff and hire personnel based on professional rather than political grounds. But autonomy is important also for implementation, since highly complex or contradictory mandates seldom produce good results. Too much autonomy, on the other hand, can also lead to disaster, either in terms of corruption or bureaucracies that set their own agendas beyond any type of political control.
    Good procedures, capacity, outputs, and bureaucratic autonomy are thus all possible ways of defining where a state lies on the vertical axis of Figure 4. It would be nice if there were scholarly agreement on a standardized way of