What has happened to inequality between and within countries since 1990? In this paper we explore who have been the winners and losers from global growth since 1990. We find that falls in total global inequality in the last 30 years are predominantly attributable to rising prosperity in China. We also identify a persistent global structure of two relatively homogeneous clusters (the poor/insecure and secure/prosperous). We detect the emergence of a 'new global middle' but question whether this implies the end of the historical two-cluster world rather than merely a transition as some people move from the poor/insecure cluster into the secure/prosperous cluster. Nevertheless, we do identify five different stylised patterns of national growth: pro-poor growth (e.g. Ethiopia); pro-middle growth (e.g. Brazil); anti-poor growth (e.g. Nigeria); anti-middle growth (e.g. Zambia) and equitable growth (e.g. Vietnam). We also find that 15 per cent of growth from 1990 to 2010 went to the world's richest 1 per cent, while just a modest amount of redistribution would have ended $2 poverty. If the share of global growth between 1990 and 2010 flowing to those who were living on under $2/day in 2010 had increased from 5 per cent to just 12 per cent, this would have been sufficient to end $2 poverty today. Persistence of global poverty, it seems, is not due to insufficient global growth but to a reluctance among the secure/prosperous cluster to forego a small share of their benefits from global growth in favour of fairly modest redistribution to the global poor.