We revisit the relevance of the World Bank's Country Policy and Institutional Assessment (CPIA)against growth performance, with attention to possible biases arising from qualitative andquantitative changes in the CPIA methodology. The CPIA, introduced in the late 1970s, had aseries of extensive revisions in the late 1980s and the early 1990s, placing more emphasis oninstitutional capacity and social policies. We reexamine a claim by previous studies that the CPIAis weakly relevant for economic performance, by running cross-country growth regressions with apanel dataset covering 146 countries between 1995 and 2015, a period over which the CPIA iscomparable. By addressing the possible biases arising from the methodological changes, we showthat the CPIA is a good predictor for future growth.