Almost half of federally managed U.S. fish stocks have no overfished determination — 256 of 522, and the share has not moved in thirteen quarters

What is happening?

As of 2026-06-30, NOAA Fisheries manages 522 federally managed stocks and stock complexes, and 256 of them, 49.0 percent, have no overfished determination. A further 151, 28.9 percent, have no overfishing determination. Both totals are addition over the two tables the agency publishes each quarter. NOAA prints the parts and does not print the sums.

The quarterly tables split the roster in two. The Fish Stock Sustainability Index holds 175 stocks, of which 23 had an unknown overfishing status and 18 an unknown overfished status at 2026-06-30. The remaining 347 non-index stocks carried 128 unknown overfishing and 238 unknown overfished on the same date. The three most recent quarterly summaries opened this round — 2024-12-31, 2025-12-31 and 2026-06-30 — give the non-index unknown-overfishing count as 128 on every one of the three, and the unknown-overfished count as 239, then 238, then 238.

The share is older than that. At 2023-03-31 the tables carried 175 index and 317 non-index stocks, with 243 of the resulting 492 lacking an overfished determination — 49.4 percent. Thirteen quarters later the roster has grown to 522 and the figure reads 49.0 percent. Over the same span the published index score rose from 781 at 2024-12-31 to 790.5 at 2025-12-31 to 795 at 2026-06-30. The number that reports performance improved. The share of the roster nobody has measured did not.

The statutory annual report is where the absence is most visible. Status of Stocks 2024, the report to Congress required under 16 U.S.C. 1854, carries a cover date of July 2026 for calendar year 2024. It states that 372 of 522 stocks have a known overfishing status and 263 have a known overfished status. It never states the complements, 150 and 259. The word unknown appears zero times in the document.

Whose problem is this?

The body that would have to produce the missing determinations is the same body that publishes the number reporting how well the system is doing, and that number is scoped so that producing them would not move it.

Where does this problem end?

The boundary matters here because the machinery for acting on a bad result already exists and works. What is missing sits one step earlier, at the point where a result would be produced at all.

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

Now

Needs a new measurementthe target state: no source opened here names how many stocks should carry a determination, by when, or at what cost. The one national target-setting document, NOAA technical memorandum SPO-242 of May 2023, sets a target assessment frequency and target data input levels for each stock and sets nothing at all for status determinations; for the 225 stocks never run through prioritization it records that baseline targets could not be developed. The index has a 1,000-point maximum and no stated target score. The statute at 16 U.S.C. 1854 starts a two-year clock only after a determination exists, and sets no clock at all for producing one.

How big is it?

Between 122,000 and 816,041 jobs, as of data year 2023. The low bound counts commercial harvester jobs alone, the 122,000 reported in Table 2 of Fisheries Economics of the United States 2023. The high bound adds the 694,041 jobs the same report attributes to recreational saltwater fishing expenditures, giving 816,041 against a background of 204.1 million saltwater recreational fishing trips in 2023.

The width is not uncertainty about the world. It is the distance between two different questions, and no published source picks one. The low bound asks who makes a living catching these stocks. The high bound asks whose employment exists because Americans fish in salt water at all.

The headline figure NOAA publishes for the seafood industry, about 1.7 million jobs, is deliberately not used here. Table 2 of the same report decomposes that total into 246,000 importer jobs and 513,000 retail jobs against only 122,000 harvesters. Importer and retail employment tracks seafood consumption, most of it imported, and moves with trade rather than with whether a domestic stock carries a status determination.

What this band cannot count is the number of human beings involved. No source opened here publishes a headcount of United States saltwater anglers. Fisheries Economics of the United States 2023 counts trips and jobs across roughly 300 pages and never counts anglers, and the NOAA angler brochure says only that there are millions. Nothing published allocates participants, vessels or permits to individual stocks either, so the 49.0 percent figure and the population band live in different spaces and cannot be multiplied together.

Under what conditions does it arise?

1. The statutory duty sits downstream of the measurement. Section 1854(e)(3) of Title 16 starts a two-year clock for a rebuilding plan, but only after an identification under paragraph 1. Identification requires a determination and a determination requires an assessment. A stock with no assessment cannot be identified, so no clock starts, no plan is owed, and nothing is out of compliance. Section 1853(a)(15) meanwhile requires an annual catch limit for every managed fishery, with measures to ensure accountability, and carves out nothing for data-poor stocks. The stock is fished under a number, unmeasured, and lawfully so.

2. Assessment effort is allocated by fishery importance, and importance is measured by catch. The prioritization framework scores commercial fishery importance, recreational fishery importance, importance to subsistence, rebuilding status, constituent demand, non-catch value and ecosystem role. That is a defensible rule, and it is self-confirming at the bottom of the distribution. A stock nobody has assessed generates little recorded catch, scores low on importance and stays unassessed. One input makes the loop explicit. Rebuilding status is itself a scoring factor, and rebuilding status can exist only for a stock already declared overfished, which requires an assessment.

3. The queue that decides what gets measured does not contain 225 of the 554 stocks. SPO-242 records that those stocks have never undergone assessment prioritization, that importance scores could not be reliably developed for them, and that where stocks had not been previously prioritized baseline targets could not be developed. These stocks are not low in the queue. They are outside it. There is no position from which they could advance, and no target against which their shortfall could be expressed.

4. The reported measure is scoped to the stocks that are already measured. The index covers 175 of 522 stocks and pays half a point for knowing overfished status and half a point for knowing overfishing status, out of four points per stock. That is a real incentive, and it operates only inside the index. Resolving an unknown on any of the 347 non-index stocks changes the published score by exactly zero.

5. Survey capacity binds throughput. GAO records that trawl surveys are challenging and costly to conduct over large geographic areas, and names resource challenges as one reason many stocks may have unknown status. For fiscal year 2026 the House proposal for NOAA Fisheries was 686 million dollars against a Senate figure of 1.1 billion, including a 42 percent reduction for fisheries science covering surveys and stock assessments. Money governs the rate at which the pile shrinks. It does not explain why the size of the pile is not a reported quantity.

What has been tried?

Two directions have been tried at once. Set targets for the inputs, and report a score for the outputs. Neither direction produced a target for the quantity that is actually short.

What was found?

Why is it still unsolved?

Measurement absent — the quantity that would make this visible as a shortfall is not a quantity anyone publishes, and no rule anywhere requires it to exist.

The first movement is that the number reporting on the roster does not cover the roster. The Fish Stock Sustainability Index pays one point out of four for the act of knowing, which is a genuine incentive, and it pays it only inside 175 of 522 stocks. Resolving an unknown on any of the other 347 changes the published score by zero. So the one figure released every quarter is structurally incapable of registering the problem, and it behaved accordingly: it rose from 781 to 795 across three opened quarters in which the non-index unknown-overfishing count read 128 every time.

The second movement is that nearly half the roster never entered the process that would rank it. This is what turns a resourcing story into a measurement story. If the 225 unprioritized stocks were simply at the back of a queue, the gap would be a budget question with a known answer. Instead SPO-242 records that importance scores could not be reliably developed for them and that baseline targets could not be developed where stocks had not been previously prioritized. There is no queue position to advance from and no baseline to fall short of. The absence of measurement is itself unmeasured.

The third movement is that the report which would name the gap performs the addition and declines the subtraction. Status of Stocks 2024 states 372 known overfishing statuses and 263 known overfished statuses out of 522, and never states 150 and 259. It reports the index subset in percentages and the full roster in raw counts, so the two are never in the same unit on the same page. The reader has to do the arithmetic the responsible body did not do. And the word unknown does not appear in the document at all.

One correction is worth stating, because the obvious reading is wrong. The index does not divert assessment effort toward itself. NOAA describes index stocks as prioritized for the most advanced analysis techniques, which is a statement about technique tier and not about who gets assessed, and in FY2025 the science centers completed 90 non-index assessments against 83 index ones. Index membership and assessment priority are siblings, both descended from fishery importance, rather than parent and child. The index does not cause the neglect. It defines the frame in which results are reported, and the frame excludes the failure.

Two narrower mechanisms sit underneath and should not be allowed to stand in for the main one. Thirty-eight Pacific salmon rows are exempted by fishery management plan rule from having the criteria applied at all, which is an exemption rather than a gap, and they are nonetheless tallied into the unknown bucket. And survey cost is real, with GAO naming resource challenges directly and a 42 percent cut to fisheries science on the table for fiscal year 2026. Cost explains the rate at which the pile shrinks. It does not explain why the pile has no target, no owner and no line in the number that gets reported.

What observation would mean it is solved?

Candidates — (a) the share of federally managed stocks with a known overfished status rises above a stated threshold and stays there across several consecutive quarters (b) the annual report to Congress prints the count of stocks without a determination, alongside a target and a date for reducing it (c) the 225 stocks outside the prioritization process are ranked and given baseline assessment targets.

(a) alone is weak because no threshold exists to clear. No source opened here states what share should carry a determination. Worse, the share can move without anything being assessed: the non-index roster grew from 317 at 2023-03-31 to 347 at 2025-12-31 while the unknown-overfished count moved only from 223 to 238, so both the numerator and the denominator drift for reasons the published tables do not decompose.

(b) alone counts paper. A published count is the precondition for accountability and not accountability itself. The agency already publishes every input to that count each quarter; the arithmetic is two additions. A report that prints the sum has removed the excuse for not knowing, and it has not moved a single survey vessel.

(c) alone is the narrowest of the three and still the most load-bearing. Ranking the 225 creates the baseline that (a) and (b) both need in order to mean anything. But a target that can be suspended is a report rather than a constraint, and the record shows how easily that happens: the exemption regime that removes 38 salmon stocks from the criteria entirely sits inside a fishery management plan and is invisible in the summary tally.

The honest composite is (c) first, so that a baseline exists, then (b) so that the shortfall against it is printed where Congress reads it, and only then (a) as the outcome measure. Read alone, (a) is uninterpretable, because the denominator it needs has never been published.

What is it connected to?

Fills with researchthe Marine Recreational Information Program and how recreational catch estimates feed assessments, the Species Information System database that carried both GAO recommendations, state and interstate management of stocks outside federal plans, seafood import traceability, and stock assessment coverage in other national fishery regimes. Relation type and evidence grade were not confirmed in this round.

What these sources do not say

• The annual report to Congress never uses the word unknown. Zero occurrences across the full extracted text of Status of Stocks 2024, machine-counted. The report gives 372 of 522 with a known overfishing status and 263 with a known overfished status and never states the complements. The one word that would name the subject of this document is absent from the document that is statutorily required to describe it.
• The index subset and the full roster are never expressed in the same unit on the same page. The report gives 87 percent and 89 percent for the most-targeted stocks and raw counts of 372 and 263 for all 522, so 87 percent against 71.3 percent and 89 percent against 50.4 percent never appear side by side. Index coverage is stated only as more than 80 percent of total U.S. fishing catch. The count-weighted figure, 175 of 522, is stated nowhere.
• No target, no date and no cost for closing the gap appears in any source opened. SPO-242 sets per-stock data input levels and assessment frequencies for 554 stocks and sets nothing for status determinations. The index has a maximum and no stated target. There is no published answer to how many stocks should have determinations, by when, or at what price.
• No source states that the exempt salmon stocks are counted as unknown. Thirty-eight non-index rows display an exception rather than a status, and a row-level parse plus the published marginals shows all 38 inside the unknown bucket on both axes. Neither the 49-page table nor the summary says so. The public tally silently merges not yet measured with structurally outside the measurement regime.
• NOAA publishes at least three current totals for the same quantity and reconciles none of them. 522 in the annual report and the 2026-06-30 tables, 554 in SPO-242, and 506 on the live commercial fishing status page, with GAO giving about 460. The 506 page carries no as-of date and simultaneously reports 52 rebuilt stocks against 51 in the July 2026 report, so one figure on it is newer than the annual report and another is older. A reader cannot date either.
• The landing page for the performance measure is six quarters stale and says nothing about it. It still displays 781, the score at 2024-12-31, while the quarterly series has run to 795 at 2026-06-30. Neither page points to the currency of the other.
• Nothing published allocates fishery participants to stocks. So no source can say how many people fish the 256 stocks without an overfished determination, and none attempts it. The same silence removes any possibility of a harm rate, because no source links an unknown status to a measured outcome for any fishery.
• The statute names no deadline on either side, and no source remarks on it. Section 1854(e) starts its clock only after identification and sets no date for producing one, while paragraph 1 requires an annual report with no date certain and no remedy. Status of Stocks 2024 arrived eighteen months after the year it covers, no 2025 report exists as of 2026-08-08, and nothing opened here explains the lag or treats a stock lawfully fished forever without a determination as a gap.

See the evidence

Almost everything here was read directly, and almost none of it is press. Eighteen sources were opened this round, of which fifteen are federal documents read end to end — four quarterly status releases, the 49-page status table, the annual report to Congress, a 77-page NOAA technical memorandum, a GAO report, a 300-page economics report, three NOAA program pages and two sections of the U.S. Code. The remaining three are advocacy or trade sources used narrowly: Ocean Conservancy for the fiscal year 2026 appropriation figures only, Oceana solely to show the framing existed outside the agency, and no press source at all for stock status. Three rows could not be opened and are kept blank with the reason, because the alternative was to substitute a mirror or a parent page and let the reader believe a check had been made. Where the sources disagree the disagreement is left visible rather than resolved: the total number of managed stocks is published as 522, 554, 506 and about 460, and the page carrying 506 also carries a rebuilt count of 52 against 51 in the annual report while giving no date for either. The most important distinction in this document is between what NOAA states and what is arithmetic on top of it. NOAA states 175, 347, 23, 18, 128, 238, 795, 522, 372, 263, 225 of 554, and the assessment counts. The totals 256 and 151, the shares 49.0 percent and 49.4 percent and 33.5 percent, and the count of 117 stocks displaying neither determination are additions and a row-level parse performed here, labelled as such in block 8 and reconciling exactly against every published marginal. This is a Path A output built from research rather than from a linked observation, so observation_refs is empty and provenance_mode: press-derived, notwithstanding that the dominant sources are federal documents opened directly rather than press coverage.

People affected

Estimated range 122,000–816,041 As of 2023

Sensitivity The width of the band is not uncertainty about the world. It is the distance between two questions that no published source chooses between: who makes a living catching these stocks, and whose employment exists because Americans fish in salt water at all. Anchor on the low bound, because 122,000 is a single published cell in a single NOAA table, while every step upward adds categories whose dependence on United States stock status weakens as they move further from the water. What the band cannot count is the number of human beings involved. No source opened in this round publishes a headcount of United States saltwater anglers: the economics report counts trips and jobs across roughly 300 pages and never counts anglers, and the NOAA angler brochure says only that there are millions. The 694,041 figure counts employment in tackle retail, boat manufacturing and related sectors rather than people who fish, so if a defensible angler headcount is later found the high bound should be rebuilt from it rather than adjusted. The band also cannot be intersected with the problem it describes: nothing published allocates participants, vessels or permits to individual stocks, and a single trip routinely catches many stocks, so the finding that 256 of 522 federally managed stocks have no overfished determination as of 2026-06-30 cannot be applied as a fraction to this population. The two numbers live in different spaces. Missing in the opposite direction are subsistence, ceremonial and tribal fishing, which the annual report to Congress names as a category and never counts; coastal communities whose economies depend on fisheries without appearing in either job series; and future participants, whose fisheries a stock depleted while unmeasured would foreclose. That last omission is the largest, because a status determination is a prospective instrument and everyone counted here is someone a past catch already supported.

Regional breakdown No source opened in this round gives a state-level or regional count of fishery participants tied to stock status. The economics report publishes national job totals, and the quarterly status tables publish stock counts with no participant attached to any stock. The one regional fact available is about assessment effort rather than population: GAO records that the Southeast science center assessed about 10 percent of its 153 stocks each year between 2011 and 2020 against about 78 percent of 64 stocks at the Alaska center. That disparity describes where measurement is thin, not where the affected people live, and the two cannot be substituted for one another. Splitting the national job figures by state population would be proportional allocation and would be wrong in both directions at once, because fishing employment concentrates on coastlines rather than distributing with residents.