How Long Federal Civil Cases Take, by District: 2025 Edition

Why the district a case is filed in matters to anyone pricing time

A funder who advances against a federal civil matter is buying time as much as outcome. The same multiple returned in two years and in five years are very different investments, as WP-003 works through. A general counsel setting a reserve needs to know when it will be released; outside counsel advising on strategy needs to know how long the court will take to reach the decisions that matter.

The district is the first thing that determines that clock. Every federal district works under the same rules, yet they process civil cases at very different speeds, because they differ in caseload, bench strength, vacancy and the kinds of cases they carry. This report reads all 94 districts from the public U.S. Courts statistics for the year ending September 30, 2025, and sets out what a reader pricing time should take from them.

The national picture

The most important line in the table is the gap between the national median and the typical district. The national median time from filing to disposition was 15.6 months. The median district disposed of its civil cases in 8.0 months, about half as long. 14 districts had a median under six months and only 8 had one over twelve.

The same gap appears in aged inventory. Nationally, 28.3% of pending civil cases had been pending more than three years.² In the median district the share was 9.4%, and 50 of 94 districts were under 10%. Five districts together held 69.6% of every aged civil case in the federal system.

The national figure describes a handful of very large dockets. The typical district is much faster.

The reason is concentration. A small number of districts host very large multidistrict litigation dockets, in which thousands of individual actions are consolidated before one judge and resolved together, often years after filing.³ Those actions dominate national counts. A reader who applies the national median to a matter in an ordinary district will overestimate its duration; a reader who applies the typical district figure to a matter inside a large consolidated docket will underestimate it badly.

Districts ranked by median time to disposition

At the fast end, D. Vt. reported a median of 2.5 months and S.D. Fla. 3.4. Several large metropolitan districts sit near the fast end as well, which is a reminder that volume alone does not make a court slow: a district with a full bench and a docket dominated by quickly resolving case types can move a high volume fast.

At the slow end, the ranking needs careful reading. E.D. La. (79.9 months) and N.D. Fla. (47.7 months) top the list because each closed a large number of old consolidated cases during the year. A median of dispositions describes the cases that closed, so a year in which a court resolves a backlog of five-year-old cases produces a very long median even if new cases move quickly. That is why the next section treats aged inventory as the better guide to what a newly filed matter faces.

Aged inventory is the better forward signal

A median time to disposition looks backward: it measures cases that finished. The share of pending cases older than three years looks forward: it measures the queue a new filing joins. For a reader pricing the time a new or recent matter will take, the queue is what matters.

The two measures disagree in instructive ways. N.D. Fla. reported a 47.7-month median but only 3.8% of its pending cases were over three years old: it had just cleared a backlog, and a new filing there faces a short queue. D.N.J. shows the opposite: a median of 8.1 months, which looks ordinary, alongside 57.0% of pending cases over three years old, 52,402 cases in all. S.D. Ohio is similar, at 9.0 months and 65.8%.

The practical rule follows. Use the median to describe how a district has been moving cases, and use the aged share to judge the tail a new matter carries. Where they disagree, find out why before relying on either: the answer is almost always a consolidated docket that is either building or being resolved.

Outliers, stated plainly

Districts carrying large consolidated dockets

The largest concentrations of pending MDL actions sit in D.N.J. (84,917), S.D. Ohio (23,323) and a handful of others.³ Those districts dominate the national aged count. For a matter that is part of such a docket, the district’s figures describe the docket, and the relevant clock is the consolidated proceeding’s own schedule. For a matter filed in the same district outside it, the district’s overall figures overstate the wait. The MDL tracker lists every pending docket.

Small and territorial courts

D. Guam reported a 77.3-month median and 69.0% aged share on 284 pending cases and one authorized judgeship. In courts this small, a few long-running matters move every statistic, and figures should be read as descriptions of a handful of cases rather than as rates. D.N.H., at 33.0 months and 50.3% aged on 1,229 pending cases with three judgeships, is a similar case.

Vacancy

Judgeship-months vacant measure lost bench capacity over the year. The federal system recorded 408 vacant judgeship-months in the period. Vacancy is a leading indicator: a district carrying long vacancies is likely to slow before its median shows it. Where a district’s vacancy is high relative to its bench, a reader should lengthen the tail assumption even if its median is ordinary today.

Circuits compared

Measured by the typical district, the circuits are close: every circuit’s median district disposes of civil cases in roughly seven to nine months. Measured by aged inventory, they are far apart, from 6.2% in the Eleventh Circuit to 50.6% in the Third Circuit. The difference is almost entirely consolidated litigation, which is why circuit averages should not be used to price an individual matter without looking at its district and whether it sits inside a consolidated docket.

Time to trial

Few federal civil cases reach trial, and for those that do the wait is long. Among the 51 districts that report a median time from filing to civil trial,⁴ the median district’s figure was 34.6 months. The range ran from 15.6 months in E.D. Va. to 66.8 months in E.D. Cal..

For a funder, the trial median is the clock that applies if a matter does not resolve earlier. It belongs in the tail of a duration assumption, not the middle. A matter whose economics only work if it settles well before trial should be stressed against this figure for its district.

What a funder should do with this

1. 1.Start from the district, not the national figure. The national median is roughly double the typical district.
2. 2.Check the aged share. Where it is high, carry a longer tail even if the median looks ordinary.
3. 3.Ask whether the matter sits inside a consolidated docket. If it does, the docket’s own schedule is the clock.
4. 4.Discount small-court statistics. Where pending cases number in the hundreds, read the figures as a few cases, not a rate.
5. 5.Watch vacancy. Long vacancies slow a court before its median moves.
6. 6.Stress against the trial median for the district, and size so the position survives it.
7. 7.Then read the matter itself. A district describes the environment; a matter-level band, by case type and stage, describes the matter.

The interactive companion to this report, the Duration Explorer, lets a reader filter and compare districts directly: federal duration explorer. Each district also has its own page under federal jurisdictions, and the duration metrics used to price the tail are defined in WP-001.

Full table and sources

All 94 districts, grouped by circuit. Filings and pending: 12-month period ending March 31, 2025. All other court figures: 12-month period ending September 30, 2025. Pending MDL actions: report dated September 1, 2026. A dash means the figure is not published for that district in the period.

Sources

• –U.S. Courts, Federal Court Management Statistics, U.S. District Courts, District Profiles (12-month periods ending September 30, 2020 through 2025), 12-month period ending September 30, 2025.
• –U.S. Courts, Table C-1, U.S. District Courts, Civil Cases Commenced, Terminated, and Pending, 12-month period ending March 31, 2025.
• –Judicial Panel on Multidistrict Litigation, Pending MDL Dockets by District, report dated September 1, 2026.

The compiled dataset is published under the site terms and may be cited as DR-003. Next edition: after U.S. Courts publishes the statistics for the year ending September 30, 2026.

The question, as the reader has it

A funder sizing a position, a general counsel setting a reserve release date and a litigation team planning a budget all need one thing from a federal docket before anything about the matter itself: how long does a civil case wait here? The decision is a duration assumption, and a single national number is a poor way to state one. It averages over districts that behave very differently, and it averages over a few very large consolidated dockets that do not describe the typical filing. A point estimate for the nation answers a question no one is asking.

The reader’s problem is to separate the environment a matter is filed into from the matter. This report describes the environment, district by district, in the court’s own published terms, so that a matter-level duration band can then be read against it.

What is measured

Inclusion is the court’s own: a case enters a median when its record shows a termination in the period, so the median describes cases that closed. A case still pending is counted only in the pending and aged-inventory measures. That difference is the reason the report reads the two together: a median of terminated cases looks backward, and the pending queue looks forward.

How a number travels from the record to the table

1. 1.Record to published statistic. The court system compiles its own case records into district profiles and tables. This report starts from those published statistics, not from case-level data.
2. 2.Statistic to table. Each district’s published values are placed side by side. National, circuit and district-median figures are aggregated on one stated basis: counts are summed, and shares are re-derived as total aged cases over the total implied by each district’s own percentage.
3. 3.Table to matter-level band. The district figures feed a matter-level duration estimate as context. That estimate belongs to a different class of method: survival analysis in the Kaplan-Meier and proportional-hazards tradition, which uses matters that are still open, with the district entered as a group in a hierarchical model so a thin district borrows strength from its circuit in proportion to how little data it has.

No step in this report produces a model output. The figures are published statistics or counts and shares derived directly from them. The third step is named so that a reader can see where a district median stops and a matter-level band begins.

Validation design

A descriptive report is validated by reconciliation. What is frozen is the published source and its period. What is compared is each aggregate against the figure the source publishes for the same quantity: the national share pending over three years computed from the district rows is checked against the published national rate, and the two agree. What prevents error is that no figure is typed by hand: every number on the page is computed from the same data file that produces the table.

The claim that aged inventory is the better forward signal is a testable one, and a reader can test it with the series the report cites: take a district’s aged share at the end of one year, take the median time to disposition in the next, and compare. That is an out-of-time design, with the earlier year frozen and the later year held out. The metric family is rank agreement between the two measures across districts. Figure 1 is evidence of the spread among districts on the backward-looking measure. Figure 2 is evidence that the backward-looking and forward-looking measures disagree where consolidated dockets are building or clearing. Figure 3 is evidence that the circuit differences are concentrated in aged inventory.

Scope and precision

A district median describes all civil cases the district closed, across every case type. It states the environment; a single matter’s duration depends on its case type, stage and facts, and a matter-level band is the estimate for that. The report states durations and counts. It does not state outcomes, values or who prevails.

• –Data lag: the statistics cover periods that have closed, and the filings and pending counts come from a period that overlaps but is not identical to the others.
• –Censoring: pending cases are not terminated cases, so the medians describe what closed and the aged share describes what is still waiting.
• –Small courts: where a district has few pending cases, a few matters move every statistic and the figures describe cases rather than rates.

What a reader can reproduce

Everything in this report can be reproduced from public sources. The three source series are cited with their periods and addresses in section 9. The full table of all 94 districts is printed on this page, and the national, circuit and median-of-district figures can be recomputed from it with a spreadsheet using the aggregation basis stated in the notes. The compiled dataset is published under the site terms and may be cited as DR-003. A matter-level duration band for a particular filing is the one thing here that requires the platform, because it depends on the case type, stage and record version stated with it.