This chapter reviews different ways of comparing the efficiency frontiers for subgroups within a data set, specifically program efficiency, the metatechnology (or technology gap) ratio and the global frontier difference index. The latter is subsequently used to define a global Malmquist index, as well as in an alternative decomposition of the traditional Malmquist index which also considers socalled favourability and favourability change components, indicating whether individual observations are located in favourable positions in the production space based on the extent of frontier shifts they observe. The various approaches are illustrated in an empirical case of Ghanaian banks.