Within a sample of 160 pending Supplemental Security Income underpayments, the Social Security Administration’s Office of the Inspector General found that employees did not follow agency policy to ensure 69 recipients, 43 percent of the sample, received money they were owed. Auditors drew that sample from roughly 2.5 million SSI underpayments pending as of September 2024 and examined each case individually rather than relying on SSA’s own tracking. The finding is not that SSA lacks a policy for resolving these cases; it is that employees, working the cases by hand, did not consistently follow the one already on the books.
What the 43 percent actually contains
Of the 160 sampled underpayments, auditors found SSA had correctly paid 80 recipients, half the sample. The remaining 80 split into two distinct problems: 69 recipients, the 43 percent named in the audit, were owed money that employees failed to follow policy in releasing, while 11 more cases, 7 percent of the sample, were erroneous amounts that staff had never removed from the record at all. Both groups sat on SSA’s books wrongly, but for opposite reasons: one group was owed money nobody paid, the other was never owed the recorded amount in the first place.
Inside the 69-recipient group, the audit traces the failure to specific record conditions rather than a single universal mistake. Forty of the underpayments involved recipients whose SSI record had been terminated even though they remained eligible for the money, 26 involved a recipient who had since died, and three involved recipients whose payments were suspended despite meeting SSA’s eligibility rules. None of the 69 involved a live, currently paying record; every one required an employee to intervene in a record that had already left SSA’s ordinary payment track.
The underlying benefit explains why that distinction matters as much as it does. Supplemental Security Income pays people who have little or no income, little or no resources, and are 65 or older, blind or disabled, a population with limited room to absorb a payment that quietly stalls once a record changes status. The 43 percent finding does not describe recipients SSA failed to notice were owed money; it describes recipients the agency’s own sample shows it already knew were owed money, on records employees were specifically responsible for working.
The step the file skipped: Sampled failures clustered in specific record types, a terminated file or a suspended payment, rather than one universal cause SSA could fix with a single system change. See the 2026 payment calendar in The Social Security Check Protection Kit.
Why the record types that failed share one trait
SSA’s underpayment policy allows the agency to release smaller amounts automatically: an underpayment under a set threshold, currently $15,000, can clear without an employee manually approving each payment. Auditors told the agency that its own system generates limited automatic notifications once a record leaves active, currently-eligible status, which means SSA employees must manually track down and resolve underpayments tied to terminated, suspended or deceased records rather than relying on the same automated release built for ordinary accounts.
That distinction matters because it points to where the 43 percent figure actually comes from. An underpayment on an active record is the case SSA’s systems are built to catch and pay automatically. An underpayment on a record that has been terminated, suspended or closed by death depends on a person noticing it, and the audit’s sample shows that noticing did not happen consistently even though written policy required it. The automation gap is not a defect in the rule; it is a gap in the record types the rule was ever designed to reach.
A parallel sample of underpayments SSA had already paid found a related but smaller compliance problem. Reviewing a separate group of roughly 2.4 million underpayments the agency paid between July 2021 and July 2024, auditors found employees skipped a required pre-payment review in 12 percent of a 100-case sample, evidence that the manual-review gap identified in the pending cases is not confined to underpayments still waiting to be resolved.
Taken together, the two samples describe the same weak point from opposite directions. In the pending population, the risk is that a record leaves the automated track and nobody manually finishes the job. In the already-paid population, the risk is that an employee releases money on a record that was supposed to get a documented second look first and does not get one. Both problems trace back to the same manual step, applied inconsistently on either side of the payment.
The recommendation aimed at the 43 percent problem
The inspector general made three recommendations, and SSA agreed with all of them. One directed the agency to take action on the specific underpayments the audit had already identified. A second told SSA to determine how to resolve the far larger backlog of roughly 1.1 million outstanding underpayments the same pending population implies. The third, aimed squarely at the 43 percent finding, told SSA to identify the factors behind employee errors and put corrective action in place so erroneous underpayments get removed and required reviews get documented going forward.
SSA’s response, as the report records it, commits the agency to identifying the causes behind these errors and documenting required reviews going forward, with no completion date attached and no specific change yet described to how terminated, suspended or deceased records get worked. The 43 percent figure itself measures only the 160-case sample; the pending population it was drawn from runs to roughly 2.5 million underpayments, which means the same record-type gap likely reaches far more recipients than the audit’s working sample was ever built to count one by one.
The record types automation never touches
The audit found that Social Security Administration employees missed underpayments concentrated in terminated, suspended and deceased-recipient records, the same record types that fall outside the agency’s automated payment release. Recognizing when a record has left that automated track, and confirming an underpayment on it was actually resolved, is the practical gap the finding leaves open.
The Social Security Check Protection Kit includes the three SSA forms that can pause or stop collection during a dispute and a first-24-hours plan for a payment that has not arrived, organized around exactly that kind of stalled case.
Look up the 2026 payment calendar in The Social Security Check Protection Kit.
This article was researched and drafted with the assistance of AI and reviewed by The Money Overview editorial team.
