Pay and grade differentials at the World Bank

Large international organizations such as the World Bank pursue many objectives in hiring policies, including reduced costs, cultural diversity, and the avoidance of discrimination. There can be sharp tradeoffs between these objectives. Diversity is enhanced by recruiting from an international labor market, for example, but international organizations face unusually large differences in reservation wages for staff capable of doing the same work. One way to reduce costs would be to pay employees their reservation wages, which implies unequal pay for equal work, or discrimination. The authors show how these tradeoffs are resolved in the World Bank's hiring processes. They estimate disparities in salary and grades between men and women and by country of origin that cannot be attributed to differences in the productive characteristics of workers. The results indicate that about half the salary and grade differentials between men and women and staff from high- and low-income countries are attributable to differences in worker characteristics. They explore a number of alternative explanations for the rest of the salary and grade differentials, including omitted-variable bias, quotas imposed to ensure diversity, and discrimination in hiring and promoting. They argue that neither omitted-variable bias nor quotas are compelling explanations for disparities, and that discrimination probably exists, although certainly less than would be implied by a cost-minimizing hiring policy. A shift seems to be occurring in the hiring process of the Bank, possibly because 1) the application pool, including women and Part II nationals (from developing countries) has significantly improved in quality; 2) information gathering during hiring has intensified, decreasing guesswork; 3) there is more incentive to staff from minority groups; and 4) the Bank's increasing diversity in terms of gender and nationality groups is more conducive to high performance by the people against whom there may previously have been bias.