“…Whde the weighted regression above shows convergence, the unweighted version shows divergence. When the empirical growth convergence literature has failed to find empirical support for simple 'unconditional convergence' of the type examined in Table 1 (i.e., without 'conditioning' for other explanatory variables) (see, for example, Sachs and Warner, 1995;Ben-David, 2000), it may thus be because they gave all countries'the same weights. In 'conditional' growth regressions, however, the stagnation in small and poor countries in Africa is captured by variables reflecting education level, political stability etc., and the growth regressions now support the convergence hypothesis (Sachs and Warner, 1995;Barro, 1997).…”