Measuring central bank independence: A latent variables approach

We argue that central bank independence (CBI) is a latent variable of which the various existing quantifications are imperfect indicators. We show how factor analysis techniques can be employed to assess the quality of the various indicators, and how an optimal weighting of the indicators can be obtained that gives the best approximation of CBI. We also show how these results can be utilized in models in which CBI is an explanatory variable. In contrast to the well-known study of Campillo and Miron (1997), we find that our CBI indicator is significantly related to inflation, also when various control variables are included.