Through nine months of fiscal 2026 FEMA declared 82 disasters and posted $20.7 billion in net outlays — more disasters than two years earlier, less money, and 58% below last year

Two federal records measure the same agency from opposite ends. The Treasury's Monthly Treasury Statement reports FEMA's outlays. FEMA's own OpenFEMA log counts the disasters it declared. We pulled both this run and joined them. They do not move together. Net outlays by the Federal Emergency Management Agency, fiscal-year-to-date through June 30 (the first nine months of each fiscal year), from Monthly Treasury Statement Table 5 [verified, retrieved 2026-08-10]: FY2023, $31.5 billion; FY2024, $26.5 billion; FY2025, $49.8 billion; FY2026, $20.7 billion ($20,715,400,345). Net is not the literal cash paid out. It is gross outlays minus money that came back — offsetting collections and recoveries. Gross disbursement in the FY2026 window was $22.7 billion ($22,749,593,329). Net subtracts about $2.0 billion of that. FY2026 net is the lowest nine-month total of the four years — below even FY2024 — and 58% below the $49.8 billion net of the same window a year earlier, a drop of $29.1 billion. On a gross basis the fall is similar: $51.3 billion to $22.7 billion, a 56% fall. FEMA disaster declarations, counted as unique disaster numbers with a declaration date in the same October 1–June 30 window, from OpenFEMA's DisasterDeclarationsSummaries (v2), deduplicated from the county-level rows the API returns [verified, retrieved 2026-08-10]: FY2024, 73 declarations; FY2025, 110; FY2026, 82. By the OpenFEMA declaration-type field in FY2026: 35 major disaster declarations (DR), 18 emergency declarations (EM), 29 fire-management declarations (FM); the FY2025 split was 47 / 6 / 57; FY2024 was 54 / 4 / 15. Two different fields count fire in this data, and they do not match. Declaration type FM (fire-management) counts declarations issued under FEMA's fire-management grant authority. Incident type Fire counts declarations whose recorded cause was a fire, no matter their declaration type. In FY2025 the two diverge: 57 FM declarations, but 64 declarations with incident type Fire — because some fire-caused disasters were declared DR or EM rather than FM. The article uses incident type Fire wherever it compares fire counts across years. Put beside each other, within the same nine-month windows: from FY2025 to FY2026, declarations fell 25% (110 to 82) and major declarations fell 25% (47 to 35), while net outlays fell 58%. And FY2026's 82 declarations exceed FY2024's 73 by nine. Yet FY2026's nine-month net outlays were $20.7 billion, against FY2024's $26.5 billion — 22% less than 2024, for more declared disasters.

The following is analysis, not fact. The two numbers count different things, and that is the whole point. FEMA stamps a declaration the day it grants a governor's request. An outlay is cash leaving the Treasury — mostly to pay obligations from disasters declared in earlier years. The Disaster Relief Fund pays out over a multi-year tail — public-assistance projects reimburse as work is completed and documented, sometimes half a decade after the event. So a single fiscal year's dollars mostly pay the long-running bills of a few past catastrophes, not the disasters declared that year. That mechanism predicts exactly the decoupling in the record. Rank the four years two ways — by nine-month major-disaster count, then by nine-month net outlays — and the orders disagree. FY2024's nine months had the most major declarations of the recent set (54) and only middling net outlays ($26.5 billion). FY2025's nine months had fewer major declarations than FY2024 (47), yet posted nearly double the net outlays ($49.8 billion). Count went down while cash went up. If declarations drove dollars, that cannot happen; because payout lags declarations, it is ordinary. (All outlay figures here are net, nine-month fiscal-year-to-date through June 30.) The composition inside each year points the same way. FY2025's $49.8 billion net sits in a year with 64 declarations of incident type Fire — against 30 in FY2026 and 17 in FY2024. It also falls in the pay-out window of the fall-2024 hurricane season and the January 2025 California wildfires — events whose reimbursements were still flowing months later. FY2026's 82 declarations skew toward lower-cost incident types: 21 winter storms and 30 fires, with a single hurricane. More paperwork, smaller checks. This is a pattern within the record, not proof: we do not have the per-disaster obligation-and-outlay detail that would show which events drove which year's dollars ^sspeculative. None of this is a statement about whether FEMA should be spending more or less; it is a statement about which year's disasters a given year's money is likely paying for.

The strongest counter is that a declaration count is a crude severity proxy and net outlays are a crude spending proxy, so calling them "decoupled" partly restates their definitions. A fire-management declaration and a major hurricane each count as one declaration, but differ in cost by orders of magnitude. Net outlays subtract recoveries and reimbursements that can swing a month's figure. On this reading the mismatch is expected accounting, not a finding — and the honest claim is narrow: you cannot read FEMA's spending off its disaster count, which we show, rather than that spending has been cut, which we do not assert. A second caution is attribution. We can prove the two series diverge; we cannot, from these two datasets alone, prove which specific events drove which year's dollars — that would require per-disaster obligation-and-outlay detail the Monthly Treasury Statement does not break out. The fall-2024 hurricanes and the January 2025 wildfires are named here as documented context for FY2025's heavy fire count and spending window, not as an arithmetic decomposition of the $49.8 billion. Third, these are nine-month windows: a catastrophe declared late in a fiscal year lands most of its outlays in the next one, which is itself the lag the piece describes. And FY2026's window runs through June 30; the remaining fiscal-year months are not in these figures.

From the vantage of a reader watching the disaster map rather than the ledger, the intuitive equation is "more disasters, more federal money." The joined record breaks it in both directions at once, within matched nine-month windows: FY2026's first nine months declared more disasters than FY2024's but posted lower net outlays, and FY2025's nine months declared fewer major disasters than FY2024's but spent nearly twice as much. The number that makes the news — how many disasters were declared in a window — says almost nothing about the number that moves the budget: FEMA's net outlays. The money is chasing a different year's events. The two facts live in two databases, and only side by side does the gap between them appear.

A reader tracks the disaster news and assumes the federal money follows the disasters. Reading FEMA's cash and its declaration count side by side, over four years, breaks the assumption in a way no single year or single database shows. FEMA declared more disasters in the first nine months of FY2026 (82) than in the same window of FY2024 (73) and posted 22% less in net outlays doing it — $20.7 billion against $26.5 billion. And between FY2024 and FY2025 the agency declared fewer major disasters yet spent nearly double. The Disaster Relief Fund is documented to pay out over a multi-year tail; the inference that FY2025's dollars were largely paying for prior years' catastrophes fits that structure but is not proven from these two datasets alone ^sspeculative. The count fell 25% this year; the net cash fell 58% — net dollars dropping more than twice as fast as disasters. Neither series alone reveals it: the declaration log looks like an ordinary year, the outlay line looks like a 58% drop, and only the join shows that the two were never measuring the same thing.