Most efforts to evaluate algorithmic fairness apply a narrow approach that fails to represent the complex real-world settings in which decisions are made and implemented. In contrast, Dai et al. present a broader, systems-level approach that models how the policies adopted by multiple decision-makers interact. They simulate a scenario wherein employers leverage fairness conscious or fairness agnostic hiring policies to evaluate potential applicants. Their proposed model allows them to study the impact of partial compliance (i.e. a situation where only a subset of decision makers adopt fairness-conscious policies) in allocating decisions. They find that partial compliance can reduce progress towards full compliance outcomes (thereby limiting the potential benefits of the policy) and can even induce segregation (i.e. when coupled with incentive effects).