Conference Agenda

Public service provision is often delegated to street-level bureaucrats (SLBs), who make discretionary decisions with large potential consequences for citizens such as separating children from their parents. While previous literature has focused on the role of systematic biases in street-level decision-making, unsystematic variation and its’ consequences has received less attention. This paper examines two related issues: How much variation exists between individual SLBs, and what are the consequences of this variation? I study discretionary decisions made by social caseworkers in their assessments of whether to place children and adolescents outside of their homes. By analyzing 3,667 child investigations conducted between 2016 and 2023 in a larger Danish municipality, I examine the consequences of out-of-home placement on salient long-term outcomes, including academic achievement, well-being, mental health, and criminal charges. Leveraging as-if-random assignment of caseworkers to child investigations, I employ an instrumental-variables design that uses variation in placement explained by differences in individual caseworkers’ placement tendencies. This strategy identifies causal effects of placements for children on the margin of placement i.e. cases where variation in caseworkers’ placement tendencies explain placement. Preliminary results show some interesting patterns. First, caseworkers’ placement tendencies significantly explain the likelihood of being placed, even after controlling for relevant case characteristics. Second, the results indicate that out-of-home placements have adverse effects among children on the margin of placement. Overall, the results suggest that there is considerable variation in placement decisions between SLBs which potentially has large consequences on the lives of citizens. These findings emphasize the importance of individual SLBs in public service provision and imply that citizens are treated differently depending on the assignment of the individual SLB.

While disparities in educational inputs across social classes are widely documented, their institutional origins remain poorly understood. In this study we propose that microdecisions by schools, teachers, and parents contribute to the macroinequality in access to educational resources and examine which of these actors most significantly contribute to class-based disparities in educational inputs. We conduct a large-scale field experiment with more than 14,000 children and their parents, in which we distribute free materials including children’s books and access to an online digital reading platform. Combining data on school, teacher, and parents’ decisions to obtain the educational materials we provide with rich administrative data on household inequality, we track a wide range of behaviors across schools, teachers, and parents. Our findings are striking: neither schools nor teachers generate inequality—indeed, teachers tend to equalize access. Instead, disparities emerge overwhelmingly at the level of the household. Parents from higher social strata are substantially more likely to sign up for the benefit, access the digital platform, and read free e-books on our online platform. Furthermore, we experimentally demonstrate how especially parents from higher educational and income strata borrow more books at the library for their children when randomly drawn to participate in the study. These results suggest that efforts to reduce educational inequality should prioritize supporting families, particularly in navigating and accessing resources, rather than focusing reforms solely on schools or educators.

Coproduction is essential for public agencies because of its significance in improving organizational performance, restoring declining public trust, and creating more citizen-centered public values. To promote citizens’ coproduction, this study proposes reduction in administrative burdens as a potential strategy. In particular, this study explores whether there are different levels of administrative burdens imposed on citizens from different racial groups (Jakobsen and Andersen 2013), and whether reduction in administrative burdens have different effects across racial groups (Christensen et al. 2020). The study adopts a vignette-based factorial experiment with random assignments to white and non-white US residents. The experiment consists of a 2x3 factorial design. The first part is the randomization of two different policy domains where the first domain is mainly serving racial minorities, and the second serving general citizens. The second manipulation is about different levels of administrative burdens: 1) control group, 2) low administrative burdens, and 3) high administrative burdens. After presenting the vignette and one of the treatment groups, this study measures willingness to coproduce and administrative burdens with multiple scale items. As one of the first attempts to test the effect of administrative burdens on coproduction, this study contributes to the literature on administrative burdens, coproduction, as well as social equity by investigating ways to reduce racial gaps in coproduction by reducing administrative burdens.