[Excerpt] Before the appearance of my 1977 paper in this Review, it was widely thought that the income distribution worsened during the economic growth which took place in Brazil during the 1960's. My paper demonstrated that the familiar data, when analyzed from an absolute perspective, could show that the poor had benefited from growth. I found that the entire income distribution shifted, benefiting every income class; that the proportion of the economically active population with incomes below the poverty level (as defined by Brazilian standards) declined during the decade; that those who remained poor were less poor than before in absolute terms; and that the rate of growth of income among the poor was at least as great as the rate of growth among the nonpoor. These results came as a surprise to me, and so I did not expect that my conclusion – that Brazil seemed to do better on the income distribution front than many observers had originally thought-would be received uncritically by others.
The preceding comments paint a less rosy picture. Previously used data are shown to be deficient in important respects and new evidence is offered which contradicts the old. Because of this new and more critical evidence, I myself am less certain of what actually happened. But as I shall show, this latest re-examination also confirms some of the more positive aspects of the Brazilian experience. It is fair to say that neither the most favorable nor the most unfavorable position can be sustained unambiguously.