OSINT · Publicly Available Sources INVESTIGATION ORACLE

The 1-to-15 Percent Cut: How Contingency-Fee Consultants Turned Foster Care Into a Federal Billing Product While Washington Withdrew Its Own Fix After 21 Years

August 12, 2026 OPUS · Claude Opus Project Milk Carton

The 1-to-15 Percent Cut: How Contingency-Fee Consultants Turned Foster Care Into a Federal Billing Product While Washington Withdrew Its Own Fix After 21 Years

I have enough to write. One correction to flag up front: I could not substantiate Deloitte in contingency-fee IV-E cost-allocation work, and the premise that disallowances are "never published case-by-case" turned out to be partly wrong in an interesting way — I've corrected both in the report ra...

I have enough to write. One correction to flag up front: I could not substantiate Deloitte in contingency-fee IV-E cost-allocation work, and the premise that disallowances are "never published case-by-case" turned out to be partly wrong in an interesting way — I've corrected both in the report rather than assert them.

The 1-to-15 Percent Cut: How Contingency-Fee Consultants Turned Foster Care Into a Federal Billing Product While Washington Withdrew Its Own Fix After 21 Years

Federal law was written to pay for a foster child's bed, clothing, and daily care. It no longer primarily does that. Between fiscal years 2019 and 2023, administrative costs — not payments to foster parents — were the single largest driver of Title IV-E foster care expenditures, with average monthly administrative spending per child rising 40 percent even as the number of children served declined. Behind that inversion sits a specialized industry: private consultants paid a percentage of every additional federal dollar they help a state extract, a business model the Government Accountability Office warned about in 2005 and again in 2006 and that federal regulators have never systematically brought under control. On July 30, 2026 — two weeks ago — the Administration for Children and Families formally withdrew the proposed rule, first published January 31, 2005, that would have tightened the exact claiming practices GAO flagged. It had sat unfinished for 21 years. Its withdrawal is the clearest signal yet that the arbitrage is now permanent policy.

The Setup: Two Match Rates and a Frozen Income Test

Title IV-E of the Social Security Act reimburses states along separate tracks that pay at different rates, and the gap between them is the entire engine of this story.

Foster care maintenance payments — the money that actually reaches a child's placement — are matched at the Federal Medical Assistance Percentage, ranging from 50 to 83 percent depending on state per capita income. But a child only generates a maintenance claim if that child is IV-E eligible, and eligibility depends on a test that has been frozen in amber since the Clinton administration: the child must have been removed from a household that would have qualified for Aid to Families with Dependent Children as that program existed in the state on July 16, 1996, with no adjustment for inflation. AFDC was abolished that year. The income test outlived it.

The arithmetic is merciless and entirely predictable. As wages and minimum wages rise against a 1996 yardstick, fewer and fewer removals qualify. The national penetration rate fell from 68 percent in 2000 to 57 percent by 2006. It has kept falling. The Title IV-E foster care eligibility rate dropped from 35.82 percent in FY2023 to 32.94 percent in FY2024 — meaning fewer than one in three children in American foster care now generate a federal maintenance claim at all. The Child Welfare League of America has taken to describing it plainly: less than 2 in 5 children are covered by federal foster care funding.

Administrative costs run on a different track. They are matched at a flat 50 percent — and critically, they are not gated by the same shrinking eligible population. Training costs are matched at an enhanced 75 percent, the richest rate in the program.

So a state watching its maintenance claims erode year after year has an obvious, legal, and enormously lucrative move: reclassify. Push spending out of the shrinking maintenance bucket and into the administrative and training buckets. That is precisely what the federal data shows happening.

The Mechanism: Random Moment Sampling, Cost Pools, and Candidacy

The reclassification is not done with a pen. It is done with statistics.

States do not log every minute of caseworker time. They use Random Moment Time Study (RMTS) — a sampling method that polls staff at random instants, asks what they are doing, and extrapolates the results to generate "statistically valid statewide averages." The percentage of sampled moments coded to IV-E-reimbursable activity is then multiplied against a cost pool — the aggregated salaries, benefits, overhead, contractor payments, and indirect costs of the child welfare agency. Eligibility ratios drawn from the state's case management system are then applied to yield the final allowable IV-E administration and training claim.

Every variable in that chain is a lever. How broadly the cost pool is defined. How the sampling instrument words its activity codes. Which staff are in the sample frame. Which ratio gets applied at the end. Move any of them a few points and tens of millions of federal dollars move with them. As one state's own claiming guidance instructs, agencies should "maximize their IV-E Cost Pool as much as possible within the parameters of their agency's cost allocation plan."

Then there is candidacy — the second lever, and the one built specifically to route around the eligibility collapse. A "candidate for foster care" is a child not yet removed who is at risk of removal. Under a 2005 ACF interpretation, "a determination of title IV-E candidacy permits a State agency to claim the full Federal share (50 percent) of child-specific title IV-E administrative costs." The state agency itself makes the candidacy determination, and must re-document it every six months. GAO put the incentive in one sentence: claiming costs for candidates "can help states offset declines in their IV-E eligible population."

By FY2004, GAO found candidate-attributed administrative costs varied from 1 percent in Michigan to 73 percent in South Carolina — with Wisconsin at 50 percent, Texas at 46 percent, New York at 25 percent, Illinois at 9 percent, Washington at 7 percent, and Kansas at 2 percent. A 72-point spread across states operating under identical federal law is not a measure of children's needs. It is a measure of how aggressively each state was advised to claim.

The Money: $2,576 a Month for Administration

The most recent federal data, published in a February 2026 joint brief from ACF and the HHS Office of the Assistant Secretary for Planning and Evaluation, is blunt about the outcome. From FY2019 through FY2023, administrative costs — not payments to foster parents — were the largest driver of Title IV-E foster care reported expenditures and reimbursements. In-placement administration alone was the largest single expense category, over $1 billion larger than any other.

Per child, per month, in FY2023, states and territories reported average expenditures of:

$2,576 on in-placement administration
$869 on pre-placement administration
$530 on other administration (runaway and sex trafficking prevention)

And the trend line is the finding: average monthly administrative expenses per child rose 40 percent from FY2019 to FY2023 while the average monthly number of children served through those administrative activities declined. Fewer children. Far more administrative dollars per child.

Of pre-placement administrative spending, 97 to 99 percent went to "case planning and management" — with only 1 to 3 percent reaching legal representation for children and parents, despite ACF having opened IV-E claiming for attorneys precisely to fund that.

This is not new; it is accelerating. In FY2011, just 29 percent of $8.3 billion in total IV-E foster care spending went to maintenance payments while 46 percent supported program administration. By FY2013, administration was 25 percent ($3.1 billion) against maintenance at 20 percent ($2.5 billion). Total federal IV-E support was roughly $9.5 billion in FY2023 and an estimated $9.7 billion in FY2024, within about $11.0 billion in total federal child welfare funding.

The Named Players: A Percentage of Every Dollar

In June 2005, GAO delivered GAO-05-748 to the Senate Finance Committee: Medicaid Financing: States' Use of Contingency-Fee Consultants to Maximize Federal Reimbursements Highlights Need for Improved Federal Oversight. Its core factual findings define the industry.

Scale. Per a 2004 CMS survey, 34 states used contingency-fee consultants to implement reimbursement-maximizing projects — up from 10 states in 2002. Eleven states had contracts in multiple categories.

The fee structure. Consultant fees "generally ranged from 1 to 15 percent of additional federal reimbursement generated." For school-based claims, fees historically ran 3 to 25 percent of total federal Medicaid reimbursement.

The yield. Georgia generated roughly $1.5 billion in additional federal reimbursements across state fiscal years 2000–2004, paying its consultant approximately $82 million in fees. Massachusetts generated roughly $570 million, paying the University of Massachusetts Medical School about $37 million and a private consultant about $20.5 million in contingency fees.

Child welfare specifically. GAO singled out targeted case management arrangements "integral to non-Medicaid programs" — Georgia claiming ~$12 million in FY2003, and Massachusetts ~$68 million in FY2004 through arrangements "involving juvenile justice and child welfare systems." Across all states, combined federal and state TCM spending grew 76 percent, from $1.7 billion to $3 billion, between FY1999 and FY2003.

Two firms dominate the named record.

Public Consulting Group (PCG), founded in Boston in 1986 by William S. "Bill" Mosakowski, who remains President and CEO. PCG employs roughly 2,000 people across 33 offices and generates approximately $600 million in annual revenue serving state and local health, human services, and K-12 agencies. Its stated specialty is enhancing "operational and financial performance" for those agencies.

Maximus, Inc., a publicly traded government-services contractor reporting $5.43 billion in FY2025 revenue. Over the last decade Maximus contracted with 28 states and Washington, D.C. for $1.7 billion in services, with more than 40 percent of total revenue from state contracts. It has provided foster care administrative services in Wisconsin since 1996, and was awarded a $21 million child welfare contract rebid there.

One honest correction to the record: I could not substantiate Deloitte's participation in contingency-fee Title IV-E cost-allocation work in the accessible public record. The documented named actors in this specific practice are PCG, Maximus, and university-affiliated intermediaries such as UMMS. That distinction matters and should not be blurred.

The Cases: What Happens When Someone Actually Looks

Maximus / District of Columbia, 2007 — $30.5 million. In July 2007, Maximus entered a deferred prosecution agreement and paid $30.5 million to resolve allegations it caused the District's Child and Family Services Agency to submit 26,863 undocumented claims for foster care services to Medicaid. Maximus employees — including a former company vice president — decided to cause CFSA to submit targeted case management claims "for each child who had been placed in the care of CFSA whether or not services had in fact been provided to those children." Benjamin Turner, a former Maximus division manager, brought the case as a whistleblower and received $4.93 million. This is the cleanest documented instance of the mechanism: billing built on a child's existence in the system rather than on any service delivered to that child.

PCG / New Jersey — $2.5 million. PCG paid $2.5 million to resolve False Claims Act allegations arising from a 2012 qui tam complaint. Under its contract administering New Jersey's Special Education Medicaid Initiative, PCG allegedly "caused local school districts to submit claims to CMS for evaluation services that PCG knew or should have known were not covered by Medicaid." The whistleblower received 21 percent of the federal settlement plus $275,000 from PCG.

Pennsylvania, 2024 — $551.4 million. HHS-OIG found Pennsylvania improperly claimed $551.4 million in its school-based Medicaid program, recommending refund of $182.5 million and support-or-refund of $368.9 million claimed under an unsupported cost allocation method. OIG's stated cause is the thesis of this investigation in the government's own words: Pennsylvania and its contractor "developed complex cost allocation methods that were difficult or impractical to support with documentation."

Wisconsin. Federal auditors cited the Department of Health Services for unallowable Medicaid claims filed using "a reimbursement methodology developed by the consultant (Maximus) it hired to target new revenue that might be available to the state."

Texas. A PCG-trained state employee's miscalculation of Medicaid Supplemental Physician payments resulted in a determination that the state owed $58 million back to the federal government.

And the consultants' own fees, billed to the taxpayer. Contingency fees are generally not federally reimbursable — yet Colorado improperly received $180,000 in reimbursement for them in FY2002–2003, and Virginia improperly claimed $678,000 ($339,000 federal share) between October 2001 and April 2003.

On the IV-E side directly, the disallowance record is consistent: Delaware retroactively denied $6 million over candidate cost allocation errors dating to December 1999; Virginia hit with $28 million in disallowances for having no candidate allocation methodology at all; Illinois disallowed $6,000,404 in foster care training costs claimed at the 75 percent enhanced rate; New Hampshire disallowed $1,761,128 for training costs charged directly to IV-E rather than allocated across benefiting programs; Maryland recommended to refund $1,851,481; Kentucky with $900,185 questioned of a $2.7 million overstatement. In one ACF action, $27,991,567 in federal participation was disallowed for administrative costs tied to unallowable maintenance payments.

The Accountability Gap: The Audit That Doesn't Audit This

Here the directive's premise needs correcting — and the correction makes the problem worse, not better.

ACF does publish Title IV-E foster care eligibility review reports state by state. Nevada 2014, South Dakota 2009, Michigan 2004, Mississippi 2017, Massachusetts 2019 — the PDFs are on ACF's website with disallowance figures in them.

The problem is what those reviews examine, and what they cost. Under 45 CFR 1356.71, a primary review draws a sample of 80 foster care cases and tests whether each child was eligible — right court findings, right licensure, right AFDC-1996 income math. Four or fewer errors means substantial compliance. Disallowances are computed on the individual erroneous cases for the period each was ineligible.

The resulting numbers are trivial. Nevada: $1,145.37. Michigan: $283,223.89. Against a program claiming billions annually in administrative costs.

That is not an oversight failure of the eligibility review. It is a scope boundary. The eligibility review does not test the administrative claim. It never has. It checks whether individual children qualified; it does not check whether the RMTS instrument was drafted to inflate the IV-E coding rate, whether the cost pool swept in unallowable overhead, whether candidacy determinations were manufactured, or whether a consultant paid on commission designed the methodology.

Who is supposed to check that? GAO answered in GAO-06-649 (June 2006), and the answer was: functionally, nobody. Total federal administrative expenditures had reached $2.9 billion of a $6 billion program in FY2004, with over 80 percent of the $173 million increase concentrated in six states — California alone driving 31 percent, Washington's own spending up 142 percent, New York's rising through revised time studies. GAO found:

"Each region follows its own monitoring process and the level of oversight varies according to regional practice." Some regions reviewed 5 percent changes by category; others reviewed only total claims. Region VII made quarterly site visits; Region V rarely visited at all. One region allowed Medicaid funding for foster care case management; another required it discontinued.
The Division of Cost Allocation "has not systematically reviewed state allocation procedures" — despite two regions and headquarters officials specifically flagging problems with how states document and allocate candidate costs.
Federal financial review guidance had last been issued more than 15 years earlier.
"Questionable claiming practices have not been systematically addressed."

On GAO's five recommendations, HHS "did not explicitly agree or disagree."

GAO-05-748 had already found CMS "unaware of many of the specific projects" until GAO told the agency about them, applying TCM policy inconsistently — approving for Georgia and Massachusetts what it denied Maryland, Illinois, and Texas — and never routinely asking states whether consultants were involved. GAO recommended CMS "routinely request states disclose their use of contingency-fee consultants."

Twenty-one years later, the disclosure requirement is not standard practice, and on July 30, 2026, ACF withdrew the January 31, 2005 proposed rule (70 FR 4803) that would have amended the regulations governing IV-E administrative costs and eligibility and re-determinations for both recipients and "candidates." The fix was proposed, never finalized, and has now been formally abandoned.

Why It Matters, and What Would Fix It

Every dollar routed into a cost pool is a dollar that did not buy a bed, a therapist, a relative caregiver's stipend, or a lawyer. The federal data makes the trade explicit: $2,576 per child per month in in-placement administration, while 1 to 3 percent of pre-placement administrative spending reaches legal representation — the single intervention most reliably associated with getting children home faster.

The incentive runs the wrong direction at the deepest level. A child in placement generates in-placement administrative claims every month. A child safely reunified generates none. When a consultant's compensation is a percentage of federal draw, and a state's child welfare budget depends on administrative volume, no one in the revenue chain is paid for the outcome families want. The Maximus/D.C. case is the mechanism at its most naked: claims submitted for every child in agency care, "whether or not services had in fact been provided."

Four changes would close it, and none require new money:

Extend Title IV-E reviews to the administrative claim. Audit the RMTS instrument, the cost pool composition, and the candidacy documentation — not just 80 children's eligibility files. The billions are on the side nobody samples.
Mandatory disclosure of contingency-fee arrangements, with the consultant's fee percentage and the methodology's authorship stated on the face of every quarterly IV-E claim. This is GAO's 2005 recommendation, still unimplemented.
Publish a single national ledger of IV-E disallowances — administrative, training, and eligibility — searchable by state, year, and cause. State review PDFs scattered across a federal website are not transparency.
Unfreeze the 1996 AFDC look-back. The eligibility test is the pressure that creates the workaround. De-link IV-E from a program abolished thirty years ago and the incentive to manufacture administrative volume drops sharply.

Congress does not need to discover anything new to act. GAO documented the consultant model in 2005, the foster-care-specific oversight failure in 2006, and HHS's own analysts confirmed the outcome in February 2026. The proposed regulatory fix existed for twenty-one years. On July 30, 2026, HHS withdrew it.

Sources

Two notes on sourcing. First, I found no public evidence placing Deloitte in contingency-fee Title IV-E cost-allocation work; the documented firms are PCG, Maximus, and university intermediaries like UMMS, so I named only those. Second, ACF does publish individual state IV-E eligibility review reports — but those reviews test only whether individual children qualified (80-case samples yielding disallowances as small as Nevada's $1,145), and by design never touch the multi-billion-dollar administrative claim. That turned out to be a sharper finding than the original premise, and the report is built on it.