Legibility refers to the breadth and depth of the state’s information about its citizens and their activities. It requires that the state collect information about its population and organize this information in ways that are administratively understandable to the state and its officials (Scott 1998).

We argue that state capacity depends on “legibility,” or the breadth and depth of the state’s information about its citizens and their activities. We demonstrate that legibility allows the state to effectively monitor private behavior and enforce rules and regulations — core functions central to the concept of state capacity.

We operationalize legibility by estimating the accuracy of age data reported in national population censuses. The census is one of the state’s primary sources of information about its population. Nearly all countries in the contemporary world conduct population censuses, and outside of the West, censuses are conducted via household interviews.

Importantly, systematic errors in age data arise for two reasons intimately related to state capacity. First, census interviewers may skip areas that are physically inaccessible or unsafe due to lack of infrastructure or violent conditions. Second, respondents may not know their ages because the state has not given meaning to precise quantitative age via rights or regulations. In both cases, interviewers and respondents estimate their ages, but typically do so in non-random ways detectable in aggregated data. These non-random errors often take the form of an excess number of population reporting ages ending in 0 or 5.

We estimate these errors using a demographic technique called the Myers Index (Myers 1940). Myers developed this technique to estimate age accuracy in censuses. We are the first to connect age accuracy to legibility, and thereby state capacity. Note that higher Myers Scores indicate more age inaccuracy, implying greater state weakness.