| This dissertation asks three fundamental questions about representation in American cities. Who serves as mayor? How do voters select mayors? And, do mayors shape policy? Responsible for funding and providing essential services, municipal governments have a huge impact on the public's safety and quality of life. As chief elected officials, mayors are unquestionably important but also understudied political actors. A number of rich and detailed case studies provide valuable insights on individual mayors and their influence, but quantitative cross-city studies have yielded mixed findings on mayors' abilities to affect outcomes. To date, efforts to comprehensively and systematically study mayors have been hampered by a lack of data.;To overcome these data limitations, I amassed an original dataset that includes detailed background information on more than 3,200 mayoral candidates, covering nearly 300 U.S. cities over the last 60 years. My data reveal that mayors, like politicians at higher levels of government, are not very representative of their constituents---they are much more likely to be white and male, with prior political experience and white-collar careers. Business owners and executives are especially well represented in American city halls, accounting for about 32% candidates and mayors.;This study provides compelling new evidence that mayors can and do influence policy outcomes. Using a regression discontinuity design, I find that business executive mayors shape spending priorities, leading to significantly lower levels of spending on redistributive programs and greater investment in infrastructure. Perhaps counterintuitively, electing a business executive mayor appears to have little effect on the overall size of government. However, suggestive evidence indicates that they may increase local revenue, but in the form of fees and charges rather than taxes. My findings suggest that business executives preside over policy changes with implications for the distribution of both costs and benefits of local government.;In another component of the dissertation, I employ a conjoint survey experiment to investigate why voters so often elect business executives. The experimental results suggest that a candidate's experience as a business owner or executive is likely to influence voters preferences and evaluations. These findings are consistent with longstanding claims that voters rely on candidate characteristics, such as race, ethnicity, or incumbency, as information shortcuts in the absence of party cues. Notably, the cues they use may vary with party identification. In nonpartisan contests, political experience has an even stronger influence on the preferences of Democratic respondents, while Republicans give more weight to occupation.;Overall, my experimental results suggest that electoral institutions may interact with voters' preferences to shape descriptive representation. At the same time, my analyses of new observational data on mayoral candidates document striking deficits of descriptive representation in America's cities and suggest that who serves in office has meaningful policy consequences. |
