This paper explores poverty trends in Romania from 1994 to 1997. It finds no reduction in poverty. Regional and other types of poverty decompositions are
presented. The poverty profile relies on tests of stochastic dominance. This approach avoids the problem that poverty comparisons may not be robust to the subjective choice of a poverty line. In addition the use of tests of stochastic dominance avoids the potential that small movements across thresholds may have large impacts on poverty indexes. I compare the results of tests of stochastic dominance with the more traditional headcount indicator to determine the extent to which conclusions differ. I also examine the extent to which the findings are sensitive to the choice of equivalence scales.