5
corrupt actions. Third, despite the different sources of the surveys, Shang Jin Wei has
found that the pairwise correlations among the indices are very high. For example, the
correlation between the Business International and Transparency International indices
and that between Business International and Global Corruption Report indices are 0.88
and 0.77, respectively. These high correlations suggest that statistical inference on
corruption is not very sensitive to the choice of corruption index.
4
However, to further increase the accuracy of these measures, I will supplement the
statistics with expert analyses from academics specializing in China and Indonesia as
well as internal measures of the number of corruption cases prosecuted by the
government. These analyses support the trends described in the World Bank governance
data.
4
Shang Jin Wei, “Corruption in Economic Development: Beneficial Grease, Minor Annoyance, or Major
Obstacle?” working paper, 1998.