Measuring institutions: what we do not know

The last two decades have seen a significant increase in the number and quality of cross-national measures of institutions, which are also covering longer and longer time spans. However, despite the considerable progress that has been made, important measurement issues linger. This chapter argues that the key future challenge for the measurement of institutions is how to square objective or directly observable sources of information with subjective or expert perceptions that only measure institutions indirectly. The chapter discusses this issue in relation to two examples: the literature on measuring democracy (a formal institution), and the one on measuring corruption (an informal institution). The conundrum in both these literatures is that whereas objective measures are typically preferred in order to avoid endogeneity bias (the possibility that either the putative cause or consequence of the institution affect the measure itself), subjective measures are as a rule better at capturing the complex reality of most institutions. The way forward for the measurement of institutions proposed here is therefore to explicitly model the data-generating process underlying expert perceptions.