I have a thoroughly triangulated evidence base. Here is the investigation.

The Ghost Visit: How States Bill Washington for Foster-Care Checkups That Never Happened

Every month, a federal law requires that a caseworker physically lay eyes on each of America's roughly 350,000 foster children — in the home where the child sleeps, not over the phone, not in a parking lot, not in a case note copied from last month. Congress set the bar at 95 percent compliance and attached grant money to help states hit it. Yet across the country, HHS Inspector General audits, state auditors, and criminal prosecutors keep finding the same thing: visits that exist only in a database, "drive-by" checks that never assess whether a child is safe, and caseworkers who type notes for meetings that never occurred — sometimes on children who were, at that moment, being starved, beaten, or already dead. The states self-report these numbers to Washington with no independent verification, collect billions in open-ended Title IV-E reimbursement regardless of whether the visits happened, and — as of the fiscal year that began October 1, 2025 — face the elimination of the only federal financial penalty that ever existed for failing the standard. This is the anatomy of a compliance system that measures paperwork instead of children.

The Setup: A Mandate Built on Self-Reported Honor

The monthly-visit requirement is not a suggestion. It is codified at 42 U.S.C. § 624(f), enacted through the Child and Family Services Improvement and Innovation Act of 2011 (Public Law 112-34). The statute directs each state to ensure that the total number of monthly caseworker visits to children in foster care is "not less than 90 percent" — raised permanently to 95 percent for FY2015 and every year thereafter — of the visits that would occur if every child were seen once a month. A companion provision requires that at least 50 percent of those visits take place in the child's residence, precisely because a visit in a school hallway or an agency office cannot reveal the condition of the placement where a child actually lives.

The purpose is obvious to anyone who has worked a child-protection case: the monthly home visit is the single most important safeguard in the entire foster-care apparatus. It is the moment a trained adult is supposed to see the bruises, the empty refrigerator, the child flinching from a caregiver, the bedroom with no bed. Every other protection — the licensing file, the service plan, the court review — is downstream of whether that visit happened and whether the person conducting it actually looked.

Here is the structural flaw at the center of the entire scheme: states report their own compliance to the federal government in the aggregate, and no federal entity independently verifies that the underlying visits occurred. A state child-welfare agency compiles numbers out of its own case-management system — the same system its own caseworkers populate — and submits a percentage to the Administration for Children and Families (ACF). ACF accepts the number. There is no requirement that a federal reviewer pull a random sample of case files and confirm that a human being was in a home on the date claimed. The measurement of the safeguard is delegated to the very institution the safeguard is supposed to police. When the auditor and the audited are the same office, the incentive to inflate is baked in from the first keystroke.

The Money: Ten Billion Dollars That Doesn't Depend on the Visit

Follow the funding and the perverse incentive becomes clear. The dominant federal foster-care funding stream is Title IV-E of the Social Security Act, an open-ended entitlement. States are reimbursed roughly 50 percent of every dollar they spend on eligible foster-care maintenance — room, board, and administration — with no ceiling and no cap. In federal fiscal year 2023, total Title IV-E spending reached $9.5 billion, of which about $5.1 billion reimbursed states for foster-care costs (the remainder went to adoption and guardianship assistance and prevention services). Federal support for IV-E is projected at roughly $10.1 billion for FY2026, and it is authorized on a mandatory, open-ended, and permanent basis.

The critical fact: Title IV-E reimbursement does not hinge on whether a child received their monthly visit. A state draws the same federal dollars for a child who is genuinely monitored and a child who is a ghost in the system. The money attaches to the placement — the bed, the board payment, the administrative overhead — not to the protection. A caseworker who fabricates a visit and a caseworker who drives two hours to sit with a frightened teenager produce identical entries on the reimbursement ledger. The federal government keeps paying either way.

Separately, Congress created a small dedicated funding line — the Monthly Caseworker Visit formula grants under Title IV-B — explicitly to help states improve the frequency and quality of visits. It is a modest pool relative to the multibillion-dollar IV-E entitlement it is meant to support. So the architecture is this: a large, uncapped stream that pays for beds regardless of oversight, and a small stream that nominally funds the oversight — with the compliance figure that connects them supplied by the states themselves. The financial gravity points entirely toward maintaining placements and reporting good numbers, and not at all toward the expensive, labor-intensive work of actually seeing every child every month.

The Players and the Incentive Structure: The Math That Defeats the Mandate

No conspiracy is required to produce mass non-compliance; the incentive structure does it automatically, and the front-line arithmetic makes fabrication almost predictable. The Child Welfare League of America has long recommended caseloads in the range of 12 to 15 children per worker. Real caseloads routinely run to 30 or more. Consider what the mandate demands at that level. A genuine, substantive monthly home visit — 45 to 60 minutes with the child, plus travel across a county, plus preparation and honest documentation — consumes real hours. At 30 cases, a worker attempting to do it properly would need on the order of 60 to 90 hours a month on visits alone, which is 38 to 56 percent of total available working time before a single court appearance, birth-parent contact, crisis call, mandatory training, or administrative task. The math does not close. Something has to give, and the three things that give are always the same: visits shrink to drive-bys, documentation becomes superficial or fabricated, or the worker burns out and quits — after which the case is handed to the next overloaded worker, and the cycle repeats.

The U.S. Government Accountability Office has documented how this pressure corrupts the underlying data. In its reviews of state child-welfare reporting, GAO found that "insufficient training for caseworkers and inaccurate and incomplete data entry affect the quality of the data" states send to federal systems. Most tellingly, GAO reported that supervisors said caseworkers "tend to initially limit data entry to information necessary to ensure timely payment to foster care providers." Read that again: the data that gets entered reliably is the data that triggers a payment. The data about whether a child was actually seen is, by the system's own workers' account, secondary. The institution's nervous system responds to money, not to safety.

The incentive runs all the way up. A state agency that reported honest, sub-95-percent numbers would invite federal scrutiny, negative press, and — historically — a financial penalty. A state that reports 96 percent invites nothing. Where the metric is self-generated and unverified, the rational institutional behavior is to generate a passing metric. The children who fall into the gap between the reported number and the real one are, by design, invisible in the data — because the data is the thing being falsified.

The Cases: Ghosts With Names

The abstraction becomes concrete in the criminal and civil records.

Colorado — Larimer County. In December 2023, former caseworker Sandra Spraker, 45, of Wellington, was arrested on 99 criminal counts, including 29 counts of forgery of government documents and 10 counts of attempting to influence a public servant, along with embezzlement, official misconduct, and abuse of public records. Authorities said Spraker lied about interviewing at least 10 families and had not checked on children who might have been in danger, and that she filed for mileage reimbursement for local travel that never happened. She had completed caseworker training only in October 2022; by summer 2023 her supervisors were reviewing her files over concerns about her work. She was placed on administrative leave on September 8, 2023, and resigned a week later. In a plea deal, the 99 counts collapsed to two — attempt to influence a public servant (a Class 4 felony) and official misconduct — with the rest dismissed. The case prompted a broader reckoning: the Colorado Sun reported in 2024 that bad caseworkers can hop from county to county in Colorado undetected, moving on before misconduct catches up with them, and that new tracking rules were being drafted in response. Spraker was one worker who got caught. The structural finding — that a worker could fabricate visits on ten families for months before internal review noticed — is the part that should alarm every parent whose child is in state care.

Illinois — the death of Mackenzi Felmlee. In April 2020, Illinois DCFS placed Mackenzi Felmlee in a specialized foster home for children with medical and behavioral needs. Prosecutors say her foster caregiver made her wear diapers tied to her legs with plastic bags to keep waste from leaking; a physician opined that those bindings, combined with her deteriorated condition, likely caused the fatal blood clot that traveled to her lungs. In the criminal investigation into her death, investigators found that a caseworker had been "copying and pasting notes for home visits, which does not appear to have correctly documented the visits." Reporting later established that ten different caseworkers cycled through Mackenzi's case — one for only three weeks — and that a caseworker assigned to monitor her had a troubling history including an arrest for a violent crime and multiple orders of protection. The visit notes existed. The oversight did not. A child died inside a documented, "monitored," reimbursed placement.

The settlements — the price of the gap. When the fabrication and the failed monitoring reach a courtroom, taxpayers pay again, this time in damages: - Oregon, December 2023: the state Department of Human Services agreed to a $40 million settlement with four former foster children who suffered sexual and physical abuse, in a case alleging caseworkers repeatedly ignored signs of abuse and tried to cover up one child's injuries — seven broken bones — even as a criminal prosecution proceeded. - Iowa: after 16-year-old Sabrina Ray was starved to death in a foster home, the state paid $10 million to her siblings; the suit alleged the agency ignored repeated warnings and obvious signs of abuse. - California — Tulare County (J.G. v. County of Tulare): a $32 million settlement for an infant allowed to nearly starve to death through the negligence of a county child-welfare agency.

Each of these is a case where the paper said the system was working and the child's body said otherwise.

Texas — M.D. v. Abbott. The systemic version of the same failure has been under federal court supervision for more than a decade. The class action M.D. v. Abbott (2:11-cv-00084, S.D. Tex.), brought by Children's Rights, placed the Texas foster system under court-appointed monitors after findings of constitutional-level harm. Successive monitoring reports have concluded that children in Texas foster care "remain at serious risk," documenting persistent harm, delayed responses to abuse, and children who could not reliably reach a caseworker or even a phone. A federal court, with independent monitors and subpoena power, has spent years unable to force a single large state into reliable compliance — which is a measure of how weak the ordinary, self-reported federal mechanism is by comparison.

The Accountability Gap: The Watchdog Was Just Sent Home

Who is supposed to be watching? On paper, three layers: the states' own quality-assurance units, the HHS Office of Inspector General and state auditors, and the Children's Bureau's federal review process. In practice, each layer confirms the failure rather than fixing it.

The HHS-OIG and state auditors keep finding non-compliance wherever they look — but they look rarely and after the fact. An HHS-OIG audit of Kentucky, released in early 2024, found the state failed to record all required monthly caseworker visits, citing insufficient staffing during the pandemic. A 2024 Maine OIG audit found 94 percent of screened reports non-compliant with screening, risk-assessment, and investigation timelines. A California State Auditor report (2024-108) found that from July 2019 through March 2025, Alameda County did not consistently initiate or complete investigations within required timeframes, leaving children in potentially unsafe situations. The New York City Comptroller audited the Administration for Children's Services' monitoring of the safety of children in foster care and found gaps. These are point-in-time snapshots of individual jurisdictions; there is no continuous, nationwide, independent audit of whether the monthly visits states claim actually happened.

The Children's Bureau's Child and Family Services Reviews (CFSR) — the federal government's flagship performance review — reveal how deep the rot goes. Across four rounds of reviews, no state has ever achieved substantial conformity with all seven performance outcomes and all seven systemic factors. Not once. Not any state. And performance is getting worse: the average number of the 14 measures a state met fell from 5.4 in Round 1 to 4.4 in Round 2 to 2.8 in Round 3 to 2.6 in Round 4. The federal review that exists specifically to catch failing child-welfare systems has caught all of them, for two decades, and the aggregate response has been improvement plans rather than consequences. Critically, the CFSR itself begins with a state self-assessment — again placing the state's own account at the foundation of the federal judgment.

Then came the move that turned a weak accountability structure into a hollow one. On January 4, 2025, President signed the Supporting America's Children and Families Act (P.L. 118-258, H.R. 9076), which had passed the House 405-10 and cleared the Senate by unanimous consent. Buried in a broadly popular reauthorization was a provision that eliminated the only federal financial penalty that had ever existed for failing the 95-percent monthly-visit standard — the requirement that HHS determine and impose reductions in a state's federal Child Welfare Services cost-sharing when it fell short. Effective with FY2026, which began October 1, 2025, that penalty is gone. The same law also directs HHS to develop standards for "virtual caseworker visits" of older youth — formally blessing, for at least part of the population, the phone-and-video "visit" that OIG auditors have repeatedly flagged as inadequate for actually seeing a child's living conditions. Congress looked at a mandate that states were already failing on a self-reported honor system, and removed the last financial reason to comply while codifying a remote substitute for the in-person visit. The watchdog was toothless; now it has been sent home.

Why It Matters, and What Would Fix It

The monthly home visit is not bureaucratic box-checking; it is the point of contact between an abused child and the government that took custody of them precisely because someone was supposed to keep them safe. When that visit is fabricated, phoned in, or skipped, a specific, foreseeable thing happens: a Mackenzi Felmlee dies in a placement that the file says was being monitored; a Sabrina Ray starves while the system draws its reimbursement; an infant in Tulare County nearly wastes away; four children in Oregon are abused while caseworkers, a court found, look away and cover up. These are not the system's edge cases. They are what the accountability gap is for — the space into which children disappear while the paperwork reports success and the federal dollars keep flowing.

The fixes are not mysterious, and none of them requires reinventing child welfare:

1. Independent verification. Federal funding tied to the visit standard should require what every other high-stakes federal program requires — random, independent sampling of case files by auditors who do not work for the reporting agency, cross-checked against corroborating evidence (GPS or timestamp metadata, foster-parent and child confirmation, court records). A self-reported number with no verification is not a measurement; it is a request to be trusted.

2. Restore and strengthen — not repeal — the penalty. Congress should reverse the FY2026 elimination of the visit-compliance penalty and make consequences meaningful and graduated, so that a state gains nothing financially from reporting a fictitious 96 percent.

3. Attack the caseload math. No verification regime survives a 30-case caseload; the mandate is mathematically impossible at that level, which is what converts good-faith workers into copy-paste documentation. Federal dollars should be conditioned on, and directed toward, enforceable caseload ceilings near the 12-15 range professionals have recommended for decades.

4. Treat fabrication as the crime it is — consistently. Larimer County prosecuted Sandra Spraker. Most fabricated-visit cases never reach a courtroom; they end in a quiet resignation and, as the Colorado Sun documented, a new job in the next county. A national de-licensing and cross-jurisdiction tracking standard would stop workers who falsify records on children from simply moving down the road.

5. Publish the real numbers. Every state's verified visit-compliance rate, and every OIG and state-auditor finding, should be public, current, and searchable — so that families, courts, and journalists can see the gap between what a state claims and what an independent audit finds.

America takes children from their parents on the promise that the state can keep them safer than the home it removed them from. A monthly visit is the minimum expression of that promise. When Washington pays out $10 billion a year while accepting unverified compliance numbers — and then quietly deletes the only penalty for lying about them — it is not merely failing to keep the promise. It is financing the fiction that the promise is being kept, and billing the taxpayer for the ghosts.

Sources: 42 U.S.C. § 624(f) and P.L. 112-34, the Child and Family Services Improvement and Innovation Act (Congress.gov); P.L. 118-258, the Supporting America's Children and Families Act (Congress.gov; CRS R48503); Title IV-E funding data (ASPE; Bipartisan Policy Center); ASPE brief, "No States Have Ever Passed the CFSR" (November 2025); GAO child-welfare data-quality reports (GAO); HHS-OIG Kentucky foster-care audit (OIG); Larimer County / Sandra Spraker prosecution (Colorado Sun; Denver7; Larimer County); Colorado caseworker cross-county mobility (Colorado Sun); Mackenzi Felmlee / Illinois DCFS (Capitol News Illinois); Oregon $40M, Iowa Sabrina Ray, and Tulare J.G. settlements (Oregon Capital Chronicle; litigation summaries); M.D. v. Abbott (Children's Rights); California State Auditor 2024-108 (auditor.ca.gov); NYC Comptroller ACS audit (comptroller.nyc.gov).