Re-examining thresholds of continuous democracy measures

Abstract

Scholars frequently dichotomize continuous measures of democracy by setting a regime cut-off. However, such cut-offs often lack theoretical or empirical justifications, making the resulting classifications difficult to interpret conceptually. We investigate this challenge involving three major continuous democracy measures: the Freedom House score (FH), the Polity score, and the Regime of the World (RoW) that is based on the V-Dem's Electoral Democracy Index (EDI). We develop a framework to empirically derive thresholds using categorical democracy measures as benchmarks. Our analyses find that the cut-offs that yield the highest consistency with the classifications of BMR, CGV, and GWF are 3.5 for FH, 5 for Polity and 0.39 for EDI/RoW. These levels are lower than the conventional cut-offs, implying less demanding democratic standards. Consequently, the conventional cut-offs (2.5 for FH, 6 for Polity and 0.5 for EDI/RoW) endeavour to reflect more stringent standards of democracy than what these dichotomous measures employ.