Promising Practices in SNAP PER Reduction: A State Case Study

By Sarah Esty (Aspen Institute Financial Security Program)

Phase 1: Preauthorization with simple rules, targeted reviews

One state has been able to achieve a PER close to 6% through use of a preauthorization review (before cases are finalized) for all new applications, reinstatements, or cases adding a person. They have historically picked cases to review using simple selection criteria: all households with 3+ members and $1,500 or more in gross income. Lead caseworkers review the earned and unearned income (budgeting method, frequency, calculated income amount, verification, and documentation – see their simple review checklist), spending under 5 minutes per case. If issues are identified, the case is returned to the worker to correct.

 This state has found that this high-volume, light touch review has prevented many of the most common agency-caused errors, including calculation issues, pay frequency (monthly vs. biweekly), categorization of income, prospective budgeting, and “reasonably anticipated” issues. The highly experienced lead caseworkers are often also able to catch client-caused errors, seeing patterns that often indicate unreported or underreported income. And their greater confidence in their intuitions leads them to be more comfortable returning cases to request additional verification when necessary.

Within the state’s workflow management system, the preauthorization request is triggered when a caseworker clicks “finalize” on a case that meets the above criteria. Then that case is automatically added to the leads’ review queue. The lead reviews the file, completes the checklist, and sends the file back to the caseworker to finalize or fix – often within minutes, and typically within less than a day (avoiding timeliness issues). 

Phase 2: Advanced analytics-driven preauthorization rules, expanded (targeted) reviews

Building on the success of their past operations, this state is now expanding their preauthorization review to include recertifications, while also being more data-driven in targeting those reviews. Through more advanced regression analysis of their own data and testing variables developed by leading academic centers using national data, the state has settled on a new 3-part inclusion criteria:

• All households with 3 or more people, AND income over $1500 OR self employment income over $750
• All households with earned income/household member >$300 AND benefit to maximum ratio > 60% AND shelter expense/household member <$300
• All households over $800 of benefits
  

Then they are removing all households with under $100 in issuances (and therefore low likelihood of having >$58 of errors that are material to PER). This exclusion criteria has cut the files to be reviewed by over 25%, while only reducing the percent of errors caught by ~15%. The refined targeting criteria plus this exclusion rule has allowed them to add recertifications while maintaining a manageable review workload for their staff.

The state is also considering expanding their reviews to cover additional elements beyond just income. They are taking a data-driven approach to this decision, measuring the additional caseworker time to review more elements (and therefore reduction in total number of cases reviewed) against the increased number of types of errors caught.

 To inform this decision, five leads timed themselves reviewing additional eligibility elements: Citizenship, Student, and ABAWD scenarios; Household Composition; Shelter expense/utilities deduction; and living above means. They found that review times would increase from under 5 minutes to 7.5 minutes – effectively reducing by >50% the number of cases that the team can review, but also expanding the share of errors the team could catch. The state is still evaluating the cost/benefit of adding some of these fields based on these results. This evaluation process is one that we recommend all states undertake to be most efficient with their review time. 

Using historical data and factoring in the share of errors that can be corrected in preauthorization review, the state was able to estimate that the new targeting criteria would cut their error rate by about 25%, or enough to be comfortably under 6% without imposing new burdens on their staff or clients.