This dissertation examines the politics of contemporary welfare provision in the United States through the lens of the social impact bond (SIB, also called “Pay for Success” (PFS)), a novel market-based policy tool that harnesses private dollars to fund local public programs. In a SIB agreement, the government borrows funds from private banks or philanthropies to pay for program implementation, and government borrowers repay the loan, with interest, only if the program achieves certain desired outcomes (e.g., a 10% reduction in recidivism). The SIB provides much-needed funding for resource-strapped governments, but it also infuses financial logics into processes of social service provision. SIBs ask local governments to meet the needs of vulnerable individuals, reduce social ills and save money on program costs as a result, and produce a program evaluation as evidence that said outcomes have been reached. In other words, SIBs ask governments not only to meet citizen needs but also to create capital in doing so, be it in terms of cost savings or knowledge gained. The rise of the SIB inspires three interrelated questions: What happens when financial logics organize activity within the social service sector? How do financial logics fare in processes of local service provision? And why do these logics fail to take hold fully in this space, even though they appear to find solid footing as organizing discourses in many other social domains? To answer these questions, I conducted a comparative case study of SIB projects in Cuyahoga County, Ohio; Massachusetts; and Washington, D.C. My data stems from 47 in-depth interviews with actors party to the SIB process, supplemented by public documents and records detailing the parameters of my three cases, the characteristics of other SIB projects, and the structure of the SIB field. The dissertation presents three broad findings. First, I argue that SIBs catalyzed the creativity and skill of local policymakers, who made adaptations to entrenched processes and responded nimbly to unexpected obstacles, first in applying the SIB model to their specific service provision contexts and later in pushing back against the financial logic of the SIB when it compromised their mandate to do no harm to the communities they served. Second, I demonstrate that the SIB movement is a rare case in which a specific financial logic moved into a space, failed to gain secure footing, and ultimately retreated. This occurred, I argue, because the range of hazards that arise in the implementation, measurement, and evaluation of social service interventions is exceedingly difficult to anticipate. Financial models hinge on successful risk management, and risk management is very hard in this setting. Third, I show that a rigid focus on social outcomes, like that which is built into the SIB model, masks meaningful differences in the work that goes into these projects: the politics of implementation, the variance in expertise and flexibility among actors within this space, and the difficulty of evaluating the efficacy of social interventions in a causally rigorous way. Taken together, these findings demonstrate that though the SIB failed as a financial tool, it succeeded as a political project. The movement’s trajectory demonstrates that novel tools of governance can spark creative problem solving among local political actors and disrupt institutionalized practices, setting into relief the power of political agents to embody, navigate, and recombine competing logics of action.