Paid to Finalize, Never Paid to Follow: The Federal Adoption Bounty and the Children Who Vanish After the Judge Signs

The United States runs the only large-scale child-placement program in the developed world that pays a cash bonus for a legal event and then stops watching. Under 42 U.S.C. § 673b, Washington wires states $5,000 to $10,000 for every foster-care adoption finalized above a moving baseline — $713 million since 1998 — while Title IV-E adoption assistance sends another $4.7 billion a year in per-child subsidies to adoptive parents whom no caseworker will ever visit again. The Adoption and Safe Families Act's 15-of-22-month termination clock feeds the pipeline from the front end, manufacturing legally free children on a statutory schedule. But the federal performance system that grades states on permanency explicitly excludes adoptions from its re-entry measure, HHS refused in its most recent AFCARS rulemaking to collect the reasons adoptions fail, and the largest bonus-earning state in the country — Texas, at $94 million — told researchers in 2021 that it could not produce the number of children re-entering its custody after an adoption. The result is a federally financed cohort of roughly 554,000 subsidized adopted children, five times the number of children in federally funded foster care, sitting entirely outside the child-welfare accountability system — and when one of those placements collapses, the child is not counted as a disruption, not reported as missing, and not screened as a trafficking risk.

The Setup: A Clock on One End, a Bounty on the Other

The Adoption and Safe Families Act of 1997 (P.L. 105-89) did two things simultaneously, and the interaction is the whole story.

First, it created the termination clock. Codified at 42 U.S.C. § 675(5)(E), ASFA requires a state to file a petition to terminate parental rights once a child has been in foster care for 15 of the most recent 22 months, subject to narrow exceptions: the child is placed with a relative, the state documents a "compelling reason" that termination is not in the child's best interests, or the agency failed to provide the reunification services required by the case plan. In practice, the exceptions are discretionary and the deadline is not. Agencies and courts treat the clock as a mandate, which — as the American Bar Association's Children's Rights Litigation committee has documented — drives case planning, pressures relatives who would prefer guardianship into adoption, and narrows judicial consideration of alternatives.

Second, ASFA created the bonus. The Adoption Incentives program began in FY1998, originally paying $4,000 per foster-child adoption above a state-specific base and $2,000 more per special-needs adoption, on a $20 million annual authorization. The Preventing Sex Trafficking and Strengthening Families Act of 2014 (P.L. 113-183) renamed it the Adoption and Legal Guardianship Incentive Payments Program and raised the stakes: $5,000 per foster-child adoption above the base rate, $7,500 per pre-adolescent (age 9–13) adoption or guardianship, and $10,000 per older-youth (14+) adoption, with awards payable in the fiscal year immediately following the year the adoptions were finalized.

Read the trigger language carefully, because it is the design flaw: finalized. The Children's Bureau's own program instruction to states is explicit that incentive funds "will only be paid to the State agency for finalized adoptions." Nothing in § 673b conditions payment on the adoption still existing one year later, three years later, or ever. There is no clawback. There is no durability requirement. The federal government buys a court order, not a childhood.

The Money

The bonus pool. Cumulative incentive earnings through FY2019 totaled roughly $713 million nationally. The largest lifetime earners are Texas (~$94 million), California (~$62 million), Florida (~$50 million), Arizona (~$49 million), Oklahoma (~$29 million), and Indiana (~$24 million). Congress level-funded the program at $75 million for FY2024, matching the President's request. In the FY2021 earning year, 47 states plus the District of Columbia and Puerto Rico earned awards ranging from $79,000 to $11,752,500, averaging $1,133,622.

Demand routinely outstrips the appropriation. In FY2015, states earned an estimated $45–46 million but only about $18 million was available — roughly 38 cents on the earned dollar — and HHS has since managed the structural shortfall by using each year's appropriation to pay off the prior year's obligations first, prorating whatever is left. States are, in other words, competing for a chronically oversubscribed pot, which sharpens rather than softens the incentive to produce finalizations.

The per-child subsidy. This is the bigger number by two orders of magnitude. Title IV-E adoption assistance paid $4.706 billion in FY2024, up from $4.123 billion in FY2023, out of roughly $9.7 billion in total Title IV-E spending. In FY2024 an estimated 712,000 children received IV-E assistance each month: 554,000 through adoption assistance, 112,000 through foster care, and 46,000 through guardianship assistance.

Sit with that ratio. The federal government writes monthly checks for nearly five times as many adopted children as foster children — and for the foster children it funds, it also funds caseworkers, court reviews, licensing inspections, placement-stability metrics, maltreatment-in-care reporting, and a federal audit regime. For the 554,000, it funds a check. Once the decree is entered, the adoptive family is a private family with a public subsidy and no public oversight. The Hart case in Washington State is the canonical illustration: Texas kept paying the Harts roughly $1,900 a month — about $277,000 between 2009 and 2018 for six children adopted in Colorado and Harris Counties — and, as the Washington Post reconstruction found, once the adoptions were finalized nobody was monitoring the children even though the state was still paying for their care. All six children and both adoptive parents died in March 2018 when the family SUV went off a Northern California cliff, days after a neighbor's abuse report.

The "savings" that were supposed to buy support. The Fostering Connections to Success and Increasing Adoptions Act of 2008 de-linked IV-E adoption assistance from the old AFDC income test, shifting cost from states to the federal government and generating state "adoption savings" that must, by law, be reinvested in child welfare — with at least 30 percent going to post-adoption services, post-guardianship services, and services to prevent foster care entry, and at least two-thirds of that 30 percent (i.e., 20 percent of savings) going specifically to post-permanency services.

GAO tested that promise and found it hollow. In GAO-22-6 (October 4, 2021), auditors found states accrued $843 million in adoption savings from FY2015 through FY2019 and reinvested only $516 million — leaving roughly $327 million unspent. More damning for accountability purposes: GAO found the annual state data the Children's Bureau collects "does not always allow it to definitively determine states' compliance" with the 30 percent post-permanency and prevention requirement. The federal agency that pays the bonus cannot verify that the money statutorily earmarked to hold adoptions together was actually spent holding adoptions together. Twenty-three of 52 reporting jurisdictions cited significant obstacles to reinvestment, most commonly confusion about the rules and competing budget priorities.

The Blind Spot Is Engineered, Not Accidental

Three separate federal instruments could count failed adoptions. None of them does.

AFCARS. The Adoption and Foster Care Analysis and Reporting System is the sole national source of foster care and adoption data. It carries a "prior adoption" flag and a field for the child's age at prior adoption. During the rulemaking that produced the current collection, multiple organizations asked HHS to require states to report why the prior adoption ended and whether the adoptive parents were still receiving federal adoption assistance payments. HHS declined, citing an "overwhelming response" from states that expanding the collection on failed adoptions would "be burdensome and outweigh its utility." The Child Welfare League of America has separately told HHS that AFCARS needs data showing that an adoption has disrupted or dissolved, with clear definitions of each term. It still doesn't have them. And the existing flag undercounts by construction: it only fires when a previously adopted child re-enters state custody. A child who runs, is handed to a relative, or is transferred to a stranger never enters the denominator.

The CFSR. The Child and Family Services Reviews are the federal government's actual enforcement lever — poor performance triggers program improvement plans and potential penalties. The Round 3 and Round 4 statewide data indicator for "Re-entry to Foster Care" measures children discharged to reunification, to a relative, or to guardianship — and excludes adoptions. The reasoning is circular: adoption is defined as permanent, therefore adoption cannot fail, therefore no measure is needed. The single federal metric designed to detect permanency that didn't hold is structurally incapable of registering the outcome the bonus program pays to produce.

OPRE's research arm. HHS's own Office of Planning, Research, and Evaluation funded the Understanding Post Adoption and Guardianship Instability (PAGI) project and published a Post Adoption and Guardianship Instability Tracking (PAGIT) Toolkit, developed by RTI International with Case Western Reserve and East Carolina University. The toolkit's premise concedes the entire problem: instability affects between 5 and 20 percent of children who exit foster care to adoption or guardianship; it can be "formal" (re-entry into care) or "informal" (the child quietly goes to live with someone else); and agencies frequently are no longer in contact with the family by the time it happens. The toolkit is voluntary. It includes a sample letter agencies can mail to adoptive families asking whether the child still lives there. That is the state of the art: an optional workbook and a form letter.

The peer-reviewed literature fills the vacuum with wide, uncomfortable ranges — 10 to 25 percent of placements disrupt before finalization; 1 to 10 percent dissolve after; a 2014 Donaldson Adoption Institute estimate that about 10 percent of youth adopted from foster care re-enter care at some point. Sattler and Font's Texas entry-cohort study found over 2 percent of adoptive placements and 7 percent of guardianship placements dissolved. And when researchers asked the Texas Department of Family and Protective Services for the number of children re-entering conservatorship following an adoption, DFPS could not provide it. The state that has collected more federal adoption bonus money than any other in the country cannot say how many of those adoptions held.

What Happens to the Children Nobody Counts

When an adoption collapses and the state isn't looking, the child goes somewhere. The documented destinations are the reason this is a missing-children story and not merely a data story.

The private transfer market. Reuters' 2013 five-part series The Child Exchange, reported by Megan Twohey over 18 months, analyzed more than 5,000 posts over five years on a single internet message board and found that, on average, a child was advertised for private "rehoming" about once a week — adoptive parents handing children to strangers via Yahoo and Facebook groups, typically with nothing more than a notarized power of attorney, no home study, no background check, no court, no caseworker. The series was a 2014 Pulitzer finalist. Reuters documented children transferred to adults with histories of abuse.

GAO confirmed the hole and then it stayed open. In GAO-15-733 (released September 16, 2015, at the request of Reps. Lloyd Doggett and Jim Langevin), GAO interviewed federal officials, 19 child welfare and adoption organizations, 15 adoption agencies, and seven states. Its findings: families turn to unregulated custody transfers because of family crisis combined with inability to access post-adoption services; and no one knows how often it happens because no agency collects it. As of July 2015, 15 states were pursuing legislation and seven had enacted it. A decade later, roughly 17 states prohibit unregulated custody transfers. California's SB 1040 makes it a misdemeanor punishable by up to a year and a $1,000 fine and criminalizes advertising children for transfer; Texas codified the offense at Penal Code § 25.081 effective September 1, 2017. In more than half the country, handing an adopted child to a stranger you met online remains, at worst, a civil ambiguity.

Congress has been offered the fix for twelve years and has not passed it. The Safe Home Act — which would define unregulated custody transfer in federal law, treat it as child abuse, and require HHS to issue guidance and report to Congress on prevalence — was first introduced by Rep. Langevin and Sen. Klobuchar in late 2013. It died in committee. It was reintroduced by Klobuchar and Blunt in 2021. It died. The current version, S. 604, the Safe Home Act of 2025, was introduced by Sens. Klobuchar and Cramer on February 13, 2025, referred to the Senate HELP Committee, and has not moved. The companion Protecting Adopted Children Act (H.R. 2068) would require collection and analysis of data on state post-adoption services and extend supports to age 21. It has never been enacted either.

The trafficking exposure is measurable — for the children the system still claims. NCMEC recorded more than 29,000 missing-child reports in 2024, of which over 27,000 were endangered runaways; roughly 1 in 7 were assessed as likely child sex trafficking victims, and among children who ran from child welfare care specifically, that figure rose to 18 percent. The Senate Permanent Subcommittee on Investigations probe led by Sens. Jon Ossoff and Marsha Blackburn found that between 2018 and 2022, NCMEC received over 2,400 reports of children missing from Georgia DFCS care involving 1,790 individual children, 410 of whom were identified as likely child sex trafficking victims — a figure Georgia DFCS publicly disputed as based on national projections rather than case-specific evidence.

Now apply the definitional trap. Every one of those Georgia children counted because they were in state custody when they went missing. A 15-year-old whose adoption dissolved is not in state custody. There is no agency of record, no placement to run from, no missing-from-care report, no NCMEC entry triggered by a caseworker, and no line in AFCARS. The most trafficking-vulnerable population in the child welfare pipeline — older, previously institutionalized, trauma-exposed, adopted at 9 or 13 or 15 under an explicit federal bonus tier that pays more for exactly those ages — becomes statistically invisible at the precise moment it becomes most at risk.

The Cases That Made It Into the Record

Arkansas, 2013–2015. State Rep. Justin Harris and his wife Marsha adopted two young sisters through Arkansas DHS, then transferred them to Eric Cameron Francis — a former employee of Harris's state-subsidized preschool, Growing God's Kingdom — outside any court or agency process. Francis raped the six-year-old within months and is serving 40 years. The Arkansas Times broke the story in March 2015. Arkansas then criminalized rehoming; Harris voted for the bill that would have made his own conduct a felony.

West Virginia, 2023–2025. Donald Lantz and Jeanne Whitefeather adopted five Black siblings while living in Minnesota, moved to Washington State, then to a Sissonville, West Virginia farm in May 2023 with children ranging from 5 to 16. Neighbors saw Lantz padlock two teenagers in a shed with no water or bathroom and drive away. On January 29, 2025, a jury convicted the couple of forced labor, human trafficking of a minor, and child abuse; Whitefeather was sentenced to 215 years and Lantz to 160. Three states approved, finalized, and subsidized that family. None of the three was looking when it turned into a trafficking case.

California, 2020–2025. The Imprint found that over five years at least 676 former foster youth adopted from care were sent out of state at taxpayer expense — through the Adoption Assistance Program — to residential facilities in Utah, Montana, and North Carolina, with at least 112 still placed out of state and monthly costs running as high as $17,000 per child. Destinations included Sequel Youth and Family Services programs — Iowa's Clarinda Academy and Utah's Red Rock Canyon School, both shuttered after abuse scandals — plus Utah facilities disciplined for improper restraints and excessive punishment. A child welfare expert called it "state-sponsored abandonment": placements California does not permit for children in its own custody, financed for children it had adopted out. Only after reporters started requesting county data did the Legislature insert a prohibition into a budget trailer bill, effective July 1, 2025.

The civil courts. Where families sue, the theory is wrongful adoption — agency failure to disclose a child's medical, psychiatric, or abuse history. A Jacksonville family won $13.5 million against Jewish Family and Community Services on exactly that theory. These verdicts are the closest thing to an audit the system has, and they arrive years late, one family at a time, and only where a family has counsel.

The Accountability Gap

Every actor in this chain has a reason not to look. The state IV-E agency earns the incentive payment, self-reports the AFCARS data, and would be the entity disclosing its own failure rate. The Children's Bureau administers both the bounty and the data system that would expose it, and told commenters that collecting dissolution reasons was too burdensome. Congress appropriates $75 million a year with no outcome condition attached. Private child-placing agencies operating under placement- and finalization-based contracts are paid on the same event. Post-adoption providers are funded out of a "savings" stream GAO says cannot be verified. As the National Coalition for Child Protection Reform has argued for years about child welfare financing generally: you get what you pay for. We pay for decrees.

What Would Actually Fix It

1. Make § 673b pay for durability, not paperwork. Hold back a portion of every incentive payment — 40 percent is a defensible starting point — releasable only after 24 or 36 months of intact placement verified against the IV-E payment record. The federal government already knows how to structure holdbacks and recoupments; it simply has never applied one here.
2. Close the AFCARS hole HHS reopened. Require, for every child entering care with a prior adoption: date of finalization, finalizing state, reason the adoptive placement ended, and whether IV-E adoption assistance was still being paid on the day the child left the home. HHS rejected these as burdensome; they are four fields.
3. Put adoption inside the CFSR re-entry indicator. Extend the follow-up window to 36 and 60 months and fold performance into the existing penalty-and-PIP structure. Nothing changes in state behavior until a failed adoption costs a state something.
4. Use the subsidy as the tripwire. IV-E adoption assistance is the only continuous federal touchpoint with 554,000 adopted children. Require an annual physical-custody attestation as a condition of payment. A parent who cannot attest that the child lives in the home triggers a welfare check, not a check.
5. Pass the Safe Home Act. Define unregulated custody transfer federally, make it a mandated-reporter event under CAPTA, and require a missing-child entry with NCIC and NCMEC when an adopted child leaves the adoptive home without a court order.
6. Audit the 30 percent. Implement GAO-22-6 with service-delivery counts — children served, services delivered, waitlists — not budget-line attestations, and publish state-by-state post-adoption service capacity.
7. Stop funding post-adoption services out of a residual. GAO found states left $327 million of adoption savings unspent while GAO-15-733 found inability to access post-adoption services is the single most cited driver of families giving children away. Fund it as a dedicated, claimable service under Title IV-E and IV-B.

The children at the center of this are not abstractions and they are not gone. They are living somewhere, with someone, under an arrangement no court approved and no agency recorded, while a federal payment file still lists them as a permanency success. The government paid to make them someone's child. It has never once paid to find out whether they still are.

Sources: - 42 U.S.C. § 673b — Adoption and legal guardianship incentive payments (U.S. House Office of Law Revision Counsel) - Social Security Act § 473A (SSA) - Adoption and Safe Families Act of 1997, P.L. 105-89 (Congress.gov) - Preventing Sex Trafficking and Strengthening Families Act, P.L. 113-183 (Congress.gov) - CRS — Child Welfare: The Adoption Incentive Program and Its Reauthorization (R43025) - CRS — Child Welfare: The Adoption Incentives Program (RL32296) - ACF — Adoption and Legal Guardianship Incentive Awards History - ACF — Title IV-E Programs Expenditure and Caseload Data 2024 - CRS — Child Welfare: Purposes, Federal Programs, and Funding (IF10590) - Voice for Adoption — Adoption and Permanency Program Funding in the Final 2024 Spending Bill - CWLA — Adoption Incentive Funding Growing Shortfall - GAO-22-6 — Better Data and Guidance Could Help States Reinvest Adoption Savings and Improve Federal Oversight - GAO-15-733 — Steps Have Been Taken to Address Unregulated Custody Transfers of Adopted Children - The Imprint — New AFCARS Data Collection: What to Know - CWLA — Comments to HHS on AFCARS - Child Welfare Capacity Building Collaborative — CFSR Round 4 Statewide Data Indicator: Reentry to Foster Care - ACF/OPRE — Post Adoption and Guardianship Instability Tracking (PAGIT) Toolkit - Sattler & Font — Predictors of Adoption and Guardianship Dissolution (PMC) - National Council For Adoption — Foster Care and Adoption Statistics (AFCARS update) - TACFS — Post-Adoption Services in Texas - Reuters "The Child Exchange" — Testimony of Megan Twohey, U.S. Senate HELP Committee - Washington Post — How adoption and support programs failed the Hart siblings - Fox News — Harts likely received $270K from Texas after adopting kids from Houston area - Arkansas Democrat-Gazette — 'Re-homed' by legislator, girl, 6, sexually assaulted - NBC News — Couple sentenced to hundreds of years for forcing adopted Black children to work as 'slaves' - The Imprint — 'State-sponsored Abandonment' No More: California Stops Paying for Adoptees Sent to Out-of-State Treatment Centers - Minnesota Journal of Law & Inequality — Adopters' Remorse: The Unregulated Use of Social Media to Rehome Children - S. 604 — Safe Home Act of 2025 (Congress.gov) - Rep. Langevin — Bipartisan Bill Will Protect Adopted Children from Rehoming - NCMEC — 2024 in Numbers - Sen. Ossoff — NCMEC Finds 410 Children Reported Missing from DFCS Care Were Likely Victims of Child Sex Trafficking - Georgia DFCS Response to the Ossoff Report - First Coast News — Jacksonville lawsuit nets family $13.5 million in adoption case - NCCPR — Financial Incentives: You Get What You Pay For