We derive the necessary and sufficient conditions under which the inverse relation between network centrality and optimal weights of a variance minimizing portfolio allocation holds. Our analytical and computational results show that certain simple and natural covariance structures satisfy these conditions. In addition to extensive simulation results, we validate our theoretical insights using return data from the constituents of the S&P 500 index.