Bringing direct democracy back in: toward a three-dimensional measure of democracy

Most procedural definitions and measurements of democracy are missing one crucial component: direct popular decision-making. This is an important gap that does not allow users of data to ascertain some important variation among democracies. Thus, I propose a new measure that is strongly anchored in a procedural definition of democracy but includes this missing dimension. The proposed measure is well rooted in the literature and introduces a dimension whereby citizens may become the masters of their political fate at any time and without the consent of elected authorities, while avoiding the inclusion of extraneous attributes that are not highlighted in democratic theory. Tests of the validity of the new indicator, using Latin American cases, show that there is enough room for its inclusion without the typical collinearity problems this literature faces. This indicator is not only sound, but it is empirically appealing as it performs better than others when testing relevant hypotheses.