We use several numerical tests in order to receive answers to our three questions. First, this paper aims to indicate, which measure of central bank independence explains economic changes the most accurately, and hence gives the most exact guidance onto institutional design of monetary authorities. Second, our aim is to prove that differences in legal proxies matter as much as institutional development of countries. Finally, we show that results are vulnerable to data modification. This experiment is performed by an empirical verification of the quality of CBI indices, comparing several widely used measures for around 100 countries, using a panel data approach. After a brief description of imprecision in CBI measures methodology and their definitions, a comparison using OLS method is made. Additional tests of TSLS, PCA and stepwise selection are used, as well. In the final conclusion we are able to point the “winner” of this experiment but also we indicate that a minor modification of data can change the result.