How to measure the rule of law

I argue that the rule of law consists of many dimensions and that much information is lost when variables proxying for these dimensions are simply aggregated. I draw on the most important innovations from various legal traditions to propose a concept of the rule of law likely to find general support. To make the concept measurable, an ideal approach is contrasted with a pragmatic one. The pragmatic approach consists of eight different dimensions. I show that the bivariate correlations between them are usually very low, evidence that more fine-grained indicators of the rule of law, rather than a single hard-to-interpret one, are necessary for its measurement. The paper presents a list of desirable variables that could improve the measurement of various aspects of the rule of law.