How long does FEMA take to approve a disaster declaration?

What the data shows

• The instrument decides the speed, not the disaster. Median time from request to decision was 0 days for a fire management assistance grant (63 of them), 1 day for an emergency declaration (25), and 42 days for a major disaster declaration (52).
• Thirteen states and territories received both kinds in the same year, which makes the comparison like-for-like. Kentucky's emergency request was answered in one day; its two major disaster requests took 113 and 40. Montana: 3 days against 82, 87, 106 and 107. Washington: 1 and 2 against 76. Texas: 1 against 46. South Carolina: same day against 42.
• Major disaster declarations are not issued as they are decided. All 52 landed on just eleven dates, and three of those days — 7 April, 30 June and 1 September — carry 31 of the 52. The eleven signed on 30 June and the eleven signed on 1 September are the two largest single-day batches in FEMA’s register, which runs back to 1953.
• The spread is enormous. Fastest: Mississippi and Louisiana, 3 days. Slowest: New Mexico’s DR-4918, requested 22 October 2025 and granted 251 days later on 30 June 2026. Half of all requests fell between 21 and 67.5 days.
• The waiting is mostly federal. States take a median of 33 days after an incident begins to file a major disaster request, because a joint damage assessment has to be done first — but the federal side is still the larger share of the total for 33 of the 52, and a median 55% of the whole incident-to-declaration clock.
• The record batch has now been matched. On 1 September FEMA signed eleven major disasters at once, equalling the eleven of 30 June. Those requests had waited a median of 35 days, from 9 for Colorado’s wildfires, floods and mudslides to 57 for North Dakota’s June storms, and only two of the eleven, South Dakota’s and Hawaii’s, include help for households.

The data

Download CSV 140 rows · 12 columns · free to reuse with credit

How this was built

OpenFEMA publishes two disaster-declaration datasets and the well-known one is the wrong one for this question. DisasterDeclarationsSummaries (v2) is what most people find; it is one row per county and it does not carry a request date, so differencing it can only ever produce an incident-to-declaration figure that blends the state’s own preparation time with the federal decision. The older FemaWebDisasterDeclarations (v1) is one row per declaration and it does carry declarationRequestDate. That single field is what makes this table possible.

We queried every record with declarationDate ge '2025-10-01', which returns 140 declarations running from 22 October 2025 to 15 September 2026. All 140 carried a request date; none needed to be dropped. For each one we computed two whole-day differences: days from incident to request (incidentBeginDate to declarationRequestDate), which is the state’s clock, and days from request to decision (declarationRequestDate to declarationDate), which is Washington’s. Quartiles are linear-interpolated. No record was excluded, smoothed or weighted.

The API response carries its own rundate, which is how we can say that a gap between declarations is a real gap in decisions and not a lag in publishing. This table was rebuilt against a rundate of 24 September 2026, and the rebuild reproduced every one of the previous version’s 123 rows exactly before adding the 17 declarations made since 22 August. The 25 August rebuild showed why the distinction matters: two of the five grants it added, Oklahoma’s Train Trestle Fire and Montana’s Sand Creek Fire, were decided on 13 and 14 August — inside the window the previous version of this table already covered — but were not in the API response when it was built on 16 August. An earlier rebuild corrected two errors in our own first version of it: the fire management grant for Washington’s Old Trails Fire was recorded here under disaster number 5662, which in FEMA’s register belongs to the Gann Fire in California, and the Gann Fire itself was missing. Old Trails is FM-5656. No summary figure in this table changed as a result. This refresh also withdraws a line from the previous version, which described the 34 days after 30 June as the year’s longest stretch without a major disaster declaration. It was not: the fiscal year’s longest was 58 days, from 22 October to 19 December 2025, and there have been six stretches of 30 days or more.

What this does not show

• This is declarations, not requests. OpenFEMA publishes what was granted; there is no public feed of what is pending or what was turned down. A state that appears nowhere in this table may have asked and be waiting, or may never have asked at all — West Virginia’s expedited request of 25 July was invisible here until the day it was granted, on 3 August. An absence here is never evidence that nobody asked.
• One fiscal year, 140 records, and the year is not over: fiscal 2026 ends on 30 September. The medians are descriptive of FY2026 and should not be read as a long-run service standard.
• A declaration can be amended later to add counties or programmes; this table measures the first decision only.
• declarationRequestDate is the date FEMA records the request as received, which need not be the date a governor announced it.
• Fire management assistance grants are a genuinely different instrument — they are decided while a fire is burning and are meant to be same-day — so their 0-day median is a design feature, not a benchmark the other two should be judged against. The comparison that carries weight is emergency against major disaster within the same state.

Use it

This table is published under a Creative Commons Attribution 4.0 licence. Republish it, chart it, or build on it — commercially or not — as long as you credit The Fold and link back to this page.

The Fold (2026). "FEMA disaster declaration decision times, FY2026." https://fold-news.com/data/fema-disaster-declaration-decision-times-fy2026 (updated September 24, 2026).