The authors depart from John Roemer's theory of equality of opportunities. They seek to determine what part of observed outcome inequality may be attributed to differences in observed "circumstances," including family background, and what part is due to "personal efforts." The authors use a microeconometric technique to simulate what the distribution of outcomes would look like if circumstances were the same for everybody. They apply this technique to Brazilian data from the 1996 household survey, both for earnings and for household incomes. The authors show that observed circumstances are a major source of outcome inequality in Brazil, probably more so than in other countries for which information is available. Nevertheless, the level of inequality after observed circumstances are equalized remains very high in Brazil.