In 2014, USDA identified concerns with data quality issues in the SNAP quality control (QC) process and immediately began a thorough review of quality control systems in all 53 state agencies. Due to the data quality issues uncovered in 42 of 53 state agencies during the reviews, the state-reported error rates derived from that data cannot be validated. Since that data cannot be validated, USDA was unable to calculate a national error rate for FY15.
USDA released state error rates for the 11 states whose quality control data could be validated and provided performance bonuses to 10 of those states in accordance with federal policy including: South Dakota, Kansas, Virgin Islands, North Dakota, New York, Arizona, Wyoming, Massachusetts, West Virginia, and Minnesota.
Error rates are not fraud rates. Error rates measure whether or not benefits provided to SNAP participants are accurate—not too high or too low. Errors can result from data entry mistakes with household information, or a lack of documentation in the files to support changes to household information. Most errors stem from unintentional mistakes on the part of the state agency or the household, such as mistyping of information, not intentional actions to violate program rules.
The issues identified with data quality are a measurement problem and are being resolved through activities at both the federal and state level. To address the data quality concerns USDA:
- Updated guidance and the "FNS Handbook 310."
- Provided training to state and federal staff.
- Worked with states to update their procedures to ensure consistency with federal guidelines.
These actions will improve the integrity of SNAP and ensure proper stewardship of taxpayer funds. Also, when USDA uncovered problematic measurement and reporting procedures during the reviews, states were required to commit to a robust corrective action plan. USDA is monitoring to ensure that states with corrective action plans are making the necessary changes on a timely basis.