Reflecting current concerns about economic inequality, scholars who study the pre-modern past are increasingly addressing this issue. The obstacles to measuring the distribution of income or wealth in the ancient Roman world are formidable. Only a few highly localized datasets are available. Any appraisal of conditions in the Roman empire as a whole therefore requires parametric modeling. Building on earlier work by Scheidel and Friesen (2009), this paper explores new ways of establishing plausible parameters for a probabilistic reconstruction of the total size of Roman wealth and the share held by the top tier of society.