The analysis of policy convergence, or: how to chase a black cat in a dark room

Political science research on policy convergence has largely remained inconclusive. While many studies found support for the convergence hypothesis, an almost equally large number of studies rejected it. Convergence thus could be a less general phenomenon than many theorists believe. This article identifies a second possible explanation. The variance approach, which dominates political science research on policy convergence, is likely to lead to wrong inferences. Analysing various artificially generated convergence processes, we find that neither the variance approach nor the coefficient of variation detects convergence when it is conditional or when theoretically unidentified convergence clubs exist. Our analysis suggests that researchers should estimate rather than measure convergence. By estimating convergence researchers may (a) test the causal relationship, (b) account for conditional convergence, (c) control for the existence of convergence clubs, and (d) examine convergence to an equilibrium level of a policy.