No government publishes a national account of opioid settlement spending — more than 10,500 examples of 2024 spending were assembled from public records requests, as of 2025-11-03

What is happening?

Opioid manufacturers, distributors and retailers have agreed to settlements that will total 58.58 billion dollars once a pending 774 million dollar settlement is finalized, a figure recorded as of 2026-04-14. State and local governments are set to receive more than 50 billion dollars over almost two decades, as reported on 2026-08-07. The money is already moving: governments received about 6 billion dollars across 2022 and 2023 and more than 6.5 billion dollars in 2024.

What does not exist is a public record of what the money bought. A nonprofit newsroom found that of the money that arrived in 2022 and 2023, about one third was spent or committed, another third was set aside, and the final third was untrackable through public records. For 2024 the same newsroom counted nearly 2.7 billion dollars spent or committed and reported that about 20 percent of the money is untrackable through public records.

The national picture that does exist was assembled by journalists rather than by any government. Reporters filed public records requests and scoured government websites to build a database of more than 10,500 examples of how the money was used or not used in 2024.

Whose problem is this?

The party that decides whether to publish is the same party whose spending would be exposed by publishing. Nobody outside that party can compel the record into existence, so the reconstruction falls to newsrooms.

Where does this problem end?

The boundary matters here because the money is not missing and the spending is not necessarily wrong. What is absent is the record, and a record that is absent cannot be audited, ranked or compared.

What is the state now, and what should it be?

Now

Needs a new measurementThe target state: no source we opened states an official national standard for what share of settlement spending must be publicly reported, in what format, or on what schedule. The nearest thing to a stated norm is a voluntary principle from a university initiative calling for a fair and transparent process for deciding where to spend the funding, endorsed by more than 60 organizations and used by more than 25 states as of 2026. A voluntary principle from a non-governmental body is not an official target, and we did not confirm that any government has adopted a numeric reporting requirement.

How big is it?

Needs a new measurementevery source we opened counts dollars, spending records and reporting formats. None counts people. The number that would size this problem is the count of jurisdictions with no centralized public reporting together with the residents living in them, and that count is itself unpublished — which is the problem restated rather than a gap in our search. The population file records this as not-derivable rather than forcing a chain from dollars to headcount.

Under what conditions does it arise?

1. Control is fragmented by design. The money is controlled by different entities in each state, and the receiving units run from statehouses down to individual towns. There is no single ledger because there is no single holder. 2. Reporting quality is a ladder, not a switch. A tracking project that grades each share of funds needs ten categories, ordered from an interactive public dashboard at the top down to no centralized reporting at the bottom. Several rungs in the middle — appropriations statutes, press releases, legislative summaries — technically disclose something while leaving actual expenditures unreadable. 3. The lowest rung is occupied. One of the ten categories exists specifically for shares where journalists and other third parties stepped in because the state published nothing. A taxonomy needs that category only if jurisdictions are sitting in it. 4. Absence produces no event. Money that is set aside, or spent without a published report, generates no incident, no service interruption and no headline. The deficit shows up only when somebody goes looking, and the looking is voluntary.

No single body is required to publish a national accounting, so the absence of one is not the failure of any single participant. The condition is stable because every participant is behaving lawfully.

What has been tried?

Every attempt in this table is an outside reconstruction. None of them is a government publishing its own complete account, and none of them can compel a jurisdiction that declines to answer.

What was found?

Two of these findings sit in tension and both are worth holding. The trackable spending is heavily weighted toward treatment, medication and housing, which is what the settlements were meant to fund, while the untrackable remainder is large enough that the weighting cannot be trusted as a description of the whole.

Why is it still unsolved?

Measurement absent — there is no instrument that reads the whole. Not because the reading is hard, but because no party is assigned to take it and no party is required to answer.

The money moves through 51 separate state systems and thousands of local governments, and each one decides for itself what to publish and in what form. A jurisdiction that publishes an interactive dashboard and a jurisdiction that publishes nothing are both in compliance, because no source we opened identified a national reporting standard for either to fall short of. Reporting and transparency requirements vary drastically state by state, and that variation is the settled condition rather than a transitional one.

The gap is being filled from outside, and that is precisely what keeps it stable. When a newsroom publishes a database of more than 10,500 spending examples, the visible national picture improves without any government changing what it does. The measured quantity then tracks the effort of the reporters rather than the openness of the jurisdictions, and a figure that moves for that reason cannot be used to hold anyone to account. Meanwhile the underlying condition is invisible in the ordinary way: unreported spending interrupts no service, triggers no alarm and leaves no record of its own absence.

What observation would mean it is solved?

Candidates — (a) every state and every receiving locality publishes itemized expenditures on a fixed schedule (b) the untrackable share falls to near zero (c) an official body rather than a newsroom publishes the national total of what was spent and on what.

(a) is satisfied too cheaply. Publishing an appropriations statute counts as disclosure and sits eighth of ten on the reporting ladder. A budget line records what was authorized, not what was bought.

(b) can improve with no change in behavior by any government. The untrackable share falls whenever journalists file more records requests, so the number would record the intensity of outside reporting rather than the openness of the jurisdictions being reported on.

(c) is the only one that moves the obligation, and it still says nothing about quality. The District of Columbia case shows that fully disclosed spending can replace an existing budget line rather than add to it. Read the three together, or the measure rewards whoever is best at producing documents.

What is it connected to?

Fills with researchoverdose mortality trends · Medicaid and state behavioral health budgets · the supplantation question of whether settlement money adds to or replaces existing spending · tobacco master settlement spending as a historical precedent. Relation type and evidence grade not confirmed.

What these sources do not say

• How many jurisdictions publish nothing. No source we opened gives the count of states, counties or cities sitting in the no centralized reporting category, nor the population living in them. The tracking project defines that category but its populated values were not readable in the page we fetched.
• Whether the two untrackable figures are comparable. The final third for 2022 and 2023 and the figure of about 20 percent reported in 2025-11 come from the same newsroom in different rounds. Neither source states that the denominators or the methods match, and one of the two articles warns that methodology changes prevent direct year-to-year comparison.
• What the set-aside third is doing. Roughly a third of the 2022 and 2023 receipts had not been earmarked for use. No source we opened says where that money is held, what it earns, or by when it must be committed.
• Whether any jurisdiction has faced consequences for not reporting. No enforcement action, withheld payment or penalty for failure to publish appears in any source we opened.
• Whether reported figures are verified. The databases are assembled from public records and government websites. No source we opened states that reported expenditures were reconciled against actual disbursements or audited.
• Cumulative dollars received to date. The sources give a total settlement value, receipts for 2022 and 2023 combined, and receipts for 2024, but no running total of what has actually landed in government accounts as of 2026.

See the evidence

No original settlement agreement, state expenditure report or audit was read directly. Every figure above comes from a tracking project or a newsroom that assembled it from public records, which is why the evidence tier is secondary. The headline totals disagree across sources and we did not confirm the reason: the tracking project records 58.58 billion dollars in settlements as of 2026-04-14, the newsroom describes more than 54 billion dollars in restitution from manufacturers and distributors and separately more than 50 billion dollars set to reach state and local governments, and the university initiative states 50 billion dollars awarded. These are plausibly three different denominators — all settlements, the manufacturer and distributor subset, and the state and local share — but no source we opened says so. Two figures were confirmed in more than one place: the more than 50 billion dollar state and local total appears in both newsroom articles of 2025-11-03 and 2026-08-07, and the roughly one third of receipts spent or committed for 2022 and 2023 appears in both the newsroom database article of 2024-12-16 and the national dashboard page of that university initiative. This is a Path A output, so observation_refs is empty and provenance_mode: press-derived.

People affected

Estimated range Not derivable

The reason and what is missing are listed under “What is missing” below