Scholars continue to debate the degree to which electoral institutions matter for representation. The literature predicts that minorities benefit from districts while women benefit from at-large elections. The mechanisms by which institutions affect the ability of traditionally underrepresented groups to win seats have been understudied. Using an analysis of over 7,000 cities and interviews with city councilors, we find that compared to at-large systems, district systems can increase diversity only when underrepresented groups are highly concentrated and compose a substantial portion of the population. In addition, we find that the electoral system has a significant effect on representation only for African American male and white female councilors; the proportion of African American women and Latina councilors is not affected by the use of either district or at-large systems.E xtensive research has been devoted to understanding the continuing underrepresentation of women and people of color in legislatures. At the city level scholars have found mixed results for the effect of singlemember district elections in increasing descriptive representation. Particularly in places where citywide elections were implemented to dilute the vote strength of racial groups, districts have been seen as a key factor in increasing racial and ethnic diversity. Alternatively for women, districts have been found to be detrimental to the election of female councilors. Scholars have proposed numerous, contradictory explanations for these findings. For minorities the focus has been on residential segregation and size of the group, while women are said to benefit from the multicandidate setting of at-large elections. For women of color these explanations are in direct conflict. This article contributes to this large literature by exploring the mechanisms by which institutions affect the representation of different groups, concurrently testing the segregation and group size hypotheses and taking into account the joint relationship between race and gender. for extremely helpful comments on drafts of the article, and Gretchen Kafoury for her very helpful advice on securing interviews with city council members. Benjamin Tagoe provided excellent research assistance. of the effect of different demographic contexts in varied institutional environments.As the Supreme Court anticipated in the landmark case Thornburg v. Gingles (1986), we demonstrate that compared to citywide elections, districts increase representation when a group is geographically concentrated and moderately sized. Further, we find districts only benefit black men. That is, the positive effect of districts is conditional on the context. Districts can increase opportunities for representation, but in some cases districts are not helpful. Only rarely do districts have a substantial impact. Taking advantage of variation among city institutional structures, council composition, and demographics, we use quantitative and qualitative methods to study these relationships. We analyze dat...