This paper focuses on why metropolitan areas vary in their capacity to translate generally high employment rates into economic opportunity for the disadvantaged. Data come from the Urban Institute's Urban Underclass Database, which includes poverty and employment data for 1980 and 1990 for the 100 largest metropolitan areas down to the Census tract level. Despite unprecedented economic growth since 1993, large segments of the population remain poor, and many cities have unacceptably high rates of poverty and economic disadvantage. The first section of this paper illustrates the wide divergence among metro areas along two dimensions: (1) the proportion of residents employed, compared with the proportion who are poor and (2) the relationship between 1980-90 employment change in a metro area and the 1980-90 change in that area's poverty rate. The second section uses multiple regression to distill the systematic influence of structural factors on the ability of a metro area to convert employment and employment growth into poverty reduction. Among these factors are the quality of education, industrial structure, and forms of business ownership in a given metro area. The third section identifies metro areas whose experiences are exceptions to the patterns predicted by those structural factors. The final section discusses potential implications of the findings and suggests next steps for research. (SM)