Files
Abstract
This dissertation uses computational text analysis to study gubernatorial policy agendas and analyze how they translate into state and city fiscal outcomes. Governors are commonly perceived as state leaders responsible for setting state policy agenda. State chief executives, however, are generally disadvantaged in the legislative process vis--vis legislators. Extant literature suggests that governors can use several formal and informal tools to overcome this disadvantage and achieve their policy goals. Empirical evidence on the power of American governors in state policymaking, however, remains inconclusive. It is also unclear from the current literature to what extent, if any, city officials may respond to gubernatorial policy priorities. This dissertation employs unsupervised machine learning algorithms to extract the main policy themes from the transcripts of gubernatorial speeches. Next, the dissertation explores the relationship between gubernatorial policy agenda and state and city fiscal outcomes. Empirical findings indicate that gubernatorial policy agenda may be a powerful force in some state policy areas, but not all. The results also suggest that in some policy areas cities align their allocation of fiscal resources with gubernatorial policy goals.