New queueing model aims to fix SNAP call center access crisis

A team of researchers, including Andrew Daw, Chloe Pache, and Angela Zhou, have published a paper on arXiv (2605.15165) proposing a queueing framework to address systemic access failures in SNAP (Supplemental Nutrition Assistance Program) call centers. The work is motivated by the Holmes v. Knodell lawsuit in Missouri, where a judge ruled that congested call centers violated applicants' procedural due process, with nearly half of all application denials being procedural rather than based on eligibility. The authors argue that such failures are distinct from typical algorithmic bias—they arise from system-level dynamics and demand congestion, not explicit rules.

The proposed model extends the classic Erlang-A queueing framework to include social-service-specific phenomena like redials and abandonment, which create endogenous congestion. The researchers use a fluid approximation to derive steady-state metrics and show that standard staffing guidelines understaff significantly in these settings. They fit model parameters to call-center data disclosed in court documents, providing a method for ex-ante evaluation of bundled staffing and service delivery changes. The framework can help policymakers design access systems that ensure applicants have a meaningful opportunity to be served at scale, addressing a critical gap in social safety net administration.