By their calculations, 10% of the population went hungry, 74% had income 1-1.5 times basic subsistence, and 14% had a 'respectable' income between 1.7 and 10 times basic subsistence. The top 1.5% controlled 15-25% of total income, the next 10% another 15-25%, while the bottom 90% split the remaining half or so of all income. This works out to a Gini coefficient (a measure of income inequality, with 1 being perfect inequality and 0 being perfect equality) of 0.42-0.44. By this measure the Roman Empire was actually less unequal than some other pre-modern societies, like 18th-century Britain or France (0.52-0.59) - but only because these societies were richer overall. (You need larger surpluses to foster larger inequality.)
Tim de Chant at Per Square Mile (a fantastic geography blog) places this research in context, noting that the Gini coefficient of the United States is now 0.45 and rising: more unequal than a pre-modern empire famous for its oligarchs and mass enslavement. He concludes on a sobering note:
Schiedel and Friesen aren’t passing judgement on the ancient Romans, nor are they on modern day Americans. Theirs is an academic study, one used to further scholarship on one of the great ancient civilizations. But buried at the end, they make a point that’s difficult to parse, yet provocative. They point out that the majority of extant Roman ruins resulted from the economic activities of the top 10 percent. “Yet the disproportionate visibility of this ‘fortunate decile’ must not let us forget the vast but—to us—inconspicuous majority that failed even to begin to share in the moderate amount of economic growth associated with large-scale formation in the ancient Mediterranean and its hinterlands.”Let us never forget: most ruins are ruins of the 1%. An Occupy Archaeology movement would have to include field survey and rural settlement studies in its call to arms.