All Investigations
OSINT · Publicly Available Sources INVESTIGATION ORACLE

The Fingerprint That Never Was: How America's Foster Care System Licenses Homes It Never Checked

July 16, 2026 OPUS · Claude Opus Project Milk Carton SSI PI License #5337

The Fingerprint That Never Was: How America's Foster Care System Licenses Homes It Never Checked

I have enough to write. One note on method before the report: WebFetch wasn't permitted in this session, so this is built from search-surfaced primary documents (OIG audit reports, state auditor reports, the CFR, AFCARS data) rather than full-text retrieval of each PDF. Where a figure comes from ...

I have enough to write. One note on method before the report: WebFetch wasn't permitted in this session, so this is built from search-surfaced primary documents (OIG audit reports, state auditor reports, the CFR, AFCARS data) rather than full-text retrieval of each PDF. Where a figure comes from press reporting on an audit rather than the audit text itself, I've said so.


The Fingerprint That Never Was: How America's Foster Care System Licenses Homes It Never Checked

Federal law has said the same thing since 2008: no state may finally approve a foster or adoptive placement until it runs a fingerprint-based FBI criminal history check on the prospective parent and checks the child abuse registries of every state where that adult has lived in the past five years. There is no discretion in the text, no exception for relatives, and no carve-out for emergencies. Yet a decade and a half of federal and state audits describe a system that routinely approves homes without the checks, keeps children in "provisional" homes for months while the fingerprints sit undone, and loses track of who is actually living at the address. In September 2025, Maryland's legislative auditors found seven registered sex offenders sharing an address with an approved guardianship home housing ten children — and the state agency didn't know. The most striking finding of this investigation is not that states fail the requirement. It is that failing it costs them essentially nothing: the federal enforcement machinery is structurally incapable of detecting the failure, and in the rare case it does, the penalty is a rounding error against a $9.7 billion program.

The Setup: A Mandate With No Meter

The requirement lives at 42 U.S.C. § 671(a)(20), the state-plan section of Title IV-E of the Social Security Act. Subparagraph (A) requires states to have "procedures for criminal records checks, including fingerprint-based checks of national crime information databases," for any prospective foster or adoptive parent "before the foster or adoptive parent may be finally approved for placement of a child" — and, critically, "regardless of whether foster care maintenance payments or adoption assistance payments are to be made on behalf of the child." That last clause matters enormously. Congress anticipated that states would try to treat unfunded placements as outside the rule. It closed the door in advance.

The statute is not vague about consequences either. A felony conviction at any time for child abuse or neglect, spousal abuse, a crime against children (including child pornography), or a crime of violence — rape, sexual assault, homicide — is a permanent bar to approval. A felony for physical assault, battery, or a drug-related offense within the preceding five years is a five-year bar. Subparagraph (B) adds the child abuse and neglect registry checks, including out-of-state registries for adults who lived elsewhere in the prior five years.

This architecture was assembled in two moves. The Adoption and Safe Families Act of 1997 created the criminal-records requirement but let states opt out entirely by act of the governor or the legislature — a loophole that meant, for a decade, that the "federal requirement" was optional in practice. The Adam Walsh Child Protection and Safety Act of 2006 closed it: as of October 1, 2008, the opt-out was gone and fingerprint-based national checks plus registry checks became mandatory in all fifty states.

So the law is clear and has been for eighteen years. What was never built was a mechanism to measure compliance.

The Money: $9.7 Billion, and an 80-Case Sample

Federal support for Title IV-E is roughly $9.7 billion for FY2024, according to the Congressional Research Service's child welfare funding primer. About half flows to foster care. Washington pays 50% to 83% of foster care maintenance costs depending on the state's Medicaid match rate, a flat 50% of administrative costs — which includes the cost of licensing homes and running background checks — and 75% of certain training. In an average month during FY2024, some 112,300 children received IV-E foster care assistance, out of 713,200 receiving IV-E aid across foster care, adoption, and guardianship.

The theoretical enforcement lever is the Title IV-E Foster Care Eligibility Review, governed by 45 C.F.R. § 1356.71. On paper it reaches exactly this problem: the regulation requires that a provider's "licensing file must contain documentation that the title IV-E agency has complied with the safety requirements for foster and adoptive placements" under § 1356.30. A missing background check should render the provider ineligible, and payments for that child unallowable.

In practice, the review is built to miss systemic failure. A primary review samples 80 cases plus an 8-case oversample. A state is in "substantial compliance" if it has no more than 8 error cases on an initial review, or 4 on a subsequent one. Only if a state flunks does it go to a secondary review of 150 cases — and an extrapolated disallowance, the only penalty large enough to hurt, is assessed only if both the case-ineligibility rate and the dollar error rate exceed 10%. Otherwise disallowances are computed case by case, on the individual payments associated with the specific ineligible children the reviewers happened to pull.

Consider what that means arithmetically. A state with 10,000 IV-E foster children and a 5% background-check failure rate is harming 500 children. An 80-case sample would be expected to surface four of them — inside the substantial-compliance threshold. The state passes. No disallowance, no finding, no headline. The review is not a smoke detector; it is a device for confirming that the building is not already fully engulfed.

The Emergency Exception That Ate the Rule

The statutory phrase is "before the foster or adoptive parent may be finally approved." States have spent two decades excavating that word finally.

The mechanism is the provisional or emergency license. About 35 states permit provisional licensure allowing kin and others to take a child immediately after basic safety screening, with full licensing to be completed within a set window. South Carolina's regulations (S.C. Code Regs. § 114-550) allow emergency placement with kin pending licensure for up to 90 calendar days. Washington issues 90-day provisional expedited licenses. In lieu of fingerprints, states lean on name-based NCIC queries — Missouri's child welfare manual instructs staff that when a child is placed with a relative in an emergency and fingerprinting isn't feasible, law enforcement should run a "Purpose Code X" screening before approving the placement.

A name-based check is not a fingerprint check. It matches on identifiers a person supplies about themselves, which is precisely the wrong tool for screening someone with a reason to be screened out. The fingerprint requirement exists because names are editable and fingerprints are not.

The design assumption is that the provisional window is short and closes. Hawaii's auditor tested that assumption and it collapsed. Report No. 24-05, issued April 2024 by the Office of the Auditor, examined the Department of Human Services' Child Welfare Services Branch and found DHS had adopted licensing procedures that contradicted its own administrative rules — issuing provisional certificates for 90 days when the rules allowed 60, then permitting an additional 60-day extension, stretching provisional status to five months, three months beyond what the rules permitted. In a random sample of 103 foster homes, the 30 child-specific homes took an average of 314 days to complete licensing. Auditors identified more than three dozen instances of homes listed as "active" that had neither a provisional certificate nor a license — children living in homes with no legal status at all.

The case that should end the argument that this is paperwork: as reported in Honolulu Civil Beat's coverage of the audit, DHS granted seven consecutive extensions of a provisional certificate for one child-specific foster home despite Catholic Charities reporting concerns about drug use and unsafe living conditions. That family never completed the FBI fingerprint background check.

DHS disputed the auditor's conclusion that children were put at risk, saying it found no evidence of harm when it reviewed the cited cases. That response is itself the tell: the agency treats absence of detected harm as evidence of safety, which is the exact inference the background check exists to make unnecessary.

Maryland: Seven Offenders, Ten Children, One Address

The most damning recent document is Audit Report SSA25, released September 2025 by Maryland's Office of Legislative Audits, covering the Department of Human Services' Social Services Administration.

The central finding: SSA did not have comprehensive procedures to ensure individuals with disqualifying criminal backgrounds did not have access to children in state care. Legislative Auditor Brian Tanen wrote that "SSA was not aware of the seven registered sex offenders we identified that had the same address as an approved guardianship home housing 10 children as of August 2024."

Read that mechanically. The auditors did not conduct a sophisticated investigation. They matched the state's own approved-home addresses against the state's own public sex offender registry — a database join any analyst could run in an afternoon. The agency responsible for the children had never run it.

The specifics compound:

  • An individual convicted of sexual misconduct with a minor listed the same address as a home where four children between ages 4 and 8 lived.
  • A person employed by SSA in December 2022 had been convicted of sexual assault in 2014 and was on the sex offender registry. In June 2024 that employee allegedly transported three foster children for what was described as "inappropriate activity"; a Washington County social services investigation concluded there was evidence of abuse.
  • The state had no system for ensuring criminal background checks on vendors providing one-on-one services to foster children housed in hotels. One contract employee had a 1990 murder conviction — a permanent statutory bar under § 671(a)(20)(A) — and was working with children living in hotels.

The same audit found SSA had not pursued or recovered $34.5 million in overpayments, and, per reporting from Maryland Matters and The Baltimore Banner, had been fined roughly $700,000 for falling short of federal foster care standards.

Hold those two numbers together, because they are the whole story. Maryland left $34.5 million uncollected and paid a $700,000 federal penalty. The penalty is two percent of the money the agency didn't bother to claw back from itself, and roughly 0.007% of annual federal IV-E spending. The federal government's most forceful available expression of displeasure is cheaper than the state's own accounting sloppiness.

The Federal Audit Trail: Twenty Years of the Same Finding

HHS-OIG has documented this repeatedly, in state after state, with the same result — a recommendation, a state concurrence, and no money moving.

Washington State group care, 2018 (OIG audit). Auditors made unannounced visits to 20 foster-care group homes in 2016. Every one failed at least one state licensing requirement. Sixteen of twenty had employees working who had not completed or passed required background checks. Most telling: 132 of 263 current employees — half the workforce caring for these children — had never had FBI fingerprint-based checks, because they were hired before the requirement took effect and the state never went back. As The Seattle Times reported, the Children's Administration's acting assistant secretary Connie Lambert-Eckel accepted all findings in a February 1 letter and said the agency would begin requiring fingerprint checks for existing staff. Ten years after Adam Walsh made it mandatory.

Ohio group homes, 2017 (OIG report A-05-16-00049). Five of thirty homes did not always comply with required criminal records checks for employees providing direct care. The report specifies no monetary recovery.

Office of Refugee Resettlement, 2026 (OIG report). In January 2024, ORR disclosed that an unlicensed care provider had not conducted required national FBI fingerprint-based checks at all — only state-level ones. Across September 2021 to August 2024, for 58 unlicensed facilities in Texas and Florida, ORR completed only 176 of 256 required monitoring visits, missing 80 of 198 quarterly abbreviated visits and conducting 43 of 58 comprehensive visits late. The watchdog wasn't visiting.

The pattern extends beyond IV-E. An April 2024 congressional report, covered by The 19th, found dozens of states out of compliance with at least one component of the federal background check requirements for child care workers under CCDBG — the parallel statute, with the same fingerprint and registry architecture, and the same result.

And the infrastructure itself is broken at the foundation. ACF's own national initiative announcement concedes there is no centralized, authoritative national source for child abuse/neglect registry contact information, producing "data duplication, outdated records, and delays in processing background checks." The statute has required interstate registry checks since 2006. Twenty years later, the Children's Bureau is still "developing" a directory of whom to call.

The Kinship Blind Spot

The fastest-growing population in foster care is the one the licensing system barely touches. Per FY2024 AFCARS data compiled by the Grandfamilies & Kinship Support Network, 39% of children in foster care live with relatives or kin, and 44% of them — 55,436 children — live in unlicensed homes. Among children entering care, 43% go to kin, and 59% of those (43,583 children) enter unlicensed homes. Across children entering, exiting, or in care during FY24, roughly 132,478 were living with kin in unlicensed homes.

Kinship placement is, on the evidence, good for children, and the push toward it is well-founded. But "unlicensed" means the licensing file — the artifact § 1356.71 requires to document safety compliance — does not exist. There is nothing for a federal reviewer to review.

ACF's September 2023 final rule, effective November 27, 2023, lets states adopt separate, simplified licensing standards for kin and still claim IV-E maintenance payments. The rule is careful on paper: safety elements — criminal history, abuse and neglect history — may not be waived. Only non-safety standards can go. In 2020, ACF reported 42 states, DC, Puerto Rico, the Virgin Islands, and three tribes already used waivers for non-safety standards for relative homes: bedroom dimensions, well-water testing, financial standards, caregiver age and marital status, pre-service training.

The distinction between safety and non-safety standards is legally sound and operationally meaningless in an agency that grants seven consecutive provisional extensions to a home with reported drug use and no fingerprints on file. The waiver isn't what fails. The clock is. Uptake, meanwhile, is slow: as of January 2026, only 18 state plans had been approved — Colorado, Delaware, Kansas, Illinois, Iowa, Kentucky, Louisiana, Michigan, Nebraska, Nevada, North Dakota, Oklahoma, Oregon, South Carolina, South Dakota, Tennessee, Texas, Wisconsin, and Washington — with ACF acknowledging that nearly two-thirds of jurisdictions have not amended their IV-E plans.

What It Costs, Priced by Juries

Where the federal government assesses nothing, tort law assesses the real number.

Washington's Department of Social and Health Services paid $9 million to settle a case, reported by KING 5, brought by a woman who alleged she was sexually abused by her foster mother's live-in boyfriend — a man who had never received a required background check. The foster mother initially denied he lived there; the department later got a direct report that he did. He had multiple disqualifying convictions, including drug trafficking and domestic violence. A completed check would have removed the child from the home. The check was the only thing standing between her and years of abuse, and it was never run.

New Jersey paid $19.5 million to a brother and sister sexually abused by foster parent Joseph Salmon, who had a known criminal history involving sexual offenses when the state placed children with him. Pennsylvania's Methodist Family Services settled for $6 million over an 11-year-old girl placed in a home with a teenager who allegedly had a known history of sexually assaulting his siblings.

The Bipartisan Policy Center's child welfare litigation review counted 34 active lawsuits across 28 states as of January 2025. Litigation has become the enforcement mechanism of last resort — which means the compliance signal arrives a decade late, is paid by taxpayers rather than the agency's budget, and requires a child to be harmed first.

The Accountability Gap

Four bodies are nominally watching, and each has a structural reason not to see this.

ACF's Children's Bureau runs the IV-E eligibility review, which samples 80 cases and forgives up to 8 errors. It cannot detect a failure rate below roughly 10% — the exact range where most states live.

The Child and Family Services Review measures outcomes, not licensing files. And its penalty design is self-neutralizing: states found out of substantial conformity write a Program Improvement Plan, and penalties are suspended during the two-year PIP implementation term. As GAO documented in GAO-04-781T, HHS's evaluation of the first eight states' PIPs found all eight met their goals, so no penalties were assessed. Idaho, penalized in DAB No. 3110 (2023) after remaining out of conformity on four outcomes, is the exception that proves the rule.

HHS-OIG does excellent work and has no teeth. It writes recommendations. States concur. Nothing is recouped. The Washington group-home audit found half the workforce unfingerprinted and produced a promise.

State licensing agencies are the ones failing, and they are structurally conflicted: the agency that must find a bed tonight is the same agency that certifies the bed is safe. Placement urgency and licensing rigor sit in the same org chart, under the same shortage pressure, and urgency wins every time. Hawaii's auditor found the state was forgoing significant federal reimbursement because unlicensed homes can't be billed to IV-E — which is the perverse punchline: the financial incentive actually runs toward licensing, and states are so overwhelmed they fail anyway.

What Would Fix It

The fix is not new law. Section 671(a)(20) is adequate. The gap is instrumentation and consequence.

Make the check a payment gate, not an audit finding. IV-E claims are submitted per child. A state should not be able to draw federal maintenance payments for a child without an attested date of the FBI fingerprint check and registry results for the home. Turn a post-hoc 80-case sample into a pre-payment field validation.

Cap provisional status in federal regulation. ACF should set a hard federal ceiling — 60 days, with a single documented extension — after which IV-E FFP stops until fingerprints clear. Hawaii's 314-day average and seven-extension case exist because no external clock runs.

Require the registry-address match. Maryland's auditors found seven offenders at a licensed address with a database join. ACF should require every IV-E agency to run approved-home addresses against state sex offender registries quarterly and certify results. This is a cron job, not a reform.

Finish the national registry directory. The interstate check has been law since 2006 and ACF still has no authoritative contact list. Fund it and finish it.

Publish the failure rate. No state currently reports what share of licensed homes have complete background check documentation. Add it to AFCARS. Sunlight has a better track record here than penalties do.

The children in these homes are, definitionally, children the state removed from their parents because it judged their parents unsafe. The state's entire legal and moral warrant for that removal is the claim that it can do better. A fingerprint check is the cheapest, most mechanical, most easily verified test of that claim available — a few dollars, a few days, a database query. When 132,478 children sit in homes with no licensing file, when half a state's group-home workforce has never been printed, when seven registered sex offenders share an address with ten kids and nobody notices until an auditor runs a join, the state has not merely broken a rule. It has failed the one test it set for itself when it knocked on the door.


Sources: - 42 U.S.C. § 671 — State plan for foster care and adoption assistance (LII) - 45 C.F.R. § 1356.71 — Federal review of eligibility (LII) - Adoption and Safe Families Act of 1997, P.L. 105-89 - Adam Walsh Child Protection and Safety Act of 2006, H.R. 4472 - CRS — Child Welfare: Purposes, Federal Programs, and Funding (IF10590) - CRS — Title IV-E State Plan Requirements (R42794) - Maryland OLA Audit Report SSA25 (Sept. 2025) - Maryland Matters — State may have put children in homes where registered sex offenders lived - The Baltimore Banner — Audit: Maryland failed to keep sex offenders away from foster children - Hawai‘i Office of the Auditor, Report No. 24-05 — Audit of DHS Child Welfare Services Branch - Honolulu Civil Beat — State Failed To Comply With Laws For Licensing Foster Homes - HHS-OIG — Some Washington State Group-Care Facilities Did Not Always Comply (2018) - The Seattle Times — Feds slam Washington foster-care group homes after surprise visits - HHS-OIG — Some Ohio Group Homes Did Not Always Comply (A-05-16-00049) - HHS-OIG — ORR Needs To Improve Monitoring of Background Check Compliance (2026) - ACF — Separate Licensing or Approval Standards for Relative or Kinship Foster Family Homes (Final Rule, Sept. 2023) - ACF — States Increasingly Promote Kinship Care, though Opportunity Remains - Grandfamilies & Kinship Support Network — Kinship Data (FY24 AFCARS) - ACF — National Initiative to Modernize Child Welfare Information and Verification Processes - GAO-04-781T — Child and Family Services Reviews - HHS DAB No. 3110 (2023) — Idaho Department of Health and Welfare - KING 5 — State settles for $9 million in lawsuit over abusive foster home - Bipartisan Policy Center — Accountability in the Courtroom: Review of Child Welfare Litigation - The 19th — States are required to background check child care workers. Many are falling short. - S.C. Code Regs. § 114-550 — Licensure of Family Foster Homes - Missouri DSS Child Welfare Manual — Background Checks/Fingerprinting