Reform for all? The limits of expertise in producing equitable outcomes

Administrative reforms often seek to improve performance by changing bureaucratic or professional expertise. Yet technical skills alone may not disrupt the forces that generate unequal outcomes across social groups. We argue that systemic bias shapes professionalization and discretionary judgment, limiting expertise-substitution reforms’ capacity to reduce racial disparities even when they improve average outcomes. We test this using a 2018 Dallas reform that deployed co-response teams pairing police officers with clinicians and social workers on mental health calls. A difference-in-­differences analysis across divisions finds significant reductions in psychiatric holds and arrests, but the gains concentrate among White male citizens; arrests of Black males increase significantly. A complementary incident-level analysis of the conditional probability of a coercive outcome points to a different pattern of heterogeneous effects: White males show a higher post-reform probability of psychiatric holds while Black females and Hispanic males experience lower ones. Both sets of findings shed new light on the scholarship of administrative reform and hold practical implications for both the impact of such programs and the ways in which they are evaluated.