The paper first scrutinises the two indices of central bank independen
ce (CBI) most commonly used in the empirical literature. It defines an
d discovers an impressive interpretation spread, a major criteria spre
ad but a negligible weighting spread in those indices. Second, it exam
ines the robustness of the empirical 'common knowledge' on the benefit
s of CBI. It finds that, when rankings produced by various CBI indices
are regressed with, among others, average inflation, 87.5% of the reg
ression coefficients are not statistically significant. Third, followi
ng recent theoretical developments, it suggests an alternative approac
h to the measurement of a central bank's operational status.