Tribunal Intelligence · Yerty Findings
How often do employment tribunals extend time limits?
Analysis of published Employment Tribunal decisions from 2000 to 2026 finds that extensions of time are granted in roughly one case in seven, that which statutory test applies matters more than how late the claim was, and that the observed grant rate has fallen from 17.3% to 13.4% since 2019.
Everyone writing about employment tribunal time limits says the same three things: the limit is three months less one day, extensions are possible, and they are rare. Almost nobody says how rare, at what length of delay, or under which test.
This report sets out what published tribunal decisions actually show about late claims. It is the analytical companion to our guide on what to do if you have missed the tribunal deadline, which covers the law and the practical steps. This piece covers the numbers, the method, and the limits of both.
How we produced these figures
Methodology & coverage
- Source
- Published Employment Tribunal decisions, 2000 to 2026, held in Yerty's tribunal database.
- How cases were identified
- Cases were identified using the structured field recording a tribunal determination on jurisdiction where limitation was in issue, whether the claimant succeeded or failed on the time point. Selection is on the question being decided, not on its outcome. This matters: a corpus selected on refusals could not produce a meaningful grant rate. Using the structured field rather than a keyword search also separates limitation from the other reasons a tribunal may find it lacks jurisdiction, such as employment status, qualifying service or territorial scope. Earlier analyses using a broader no-jurisdiction category conflated these; this one does not.
- Unit of analysis
- The source corpus is organised by case, but limitation outcomes are analysed at claim/determination level, because different causes of action within the same proceedings can receive different limitation decisions. A single case may therefore contribute more than one observation to some analyses, and the observations within a case are not independent of one another. Where a figure counts cases rather than determinations it is labelled as such.
- How lateness was measured
- The length of delay is not held as a structured field. Of the out-of-time judgments, 923 state the delay in the text, and those are the cases in the lateness table below. Cases where the tribunal did not state how late the claim was are absent, so that table describes judgments that quantified the delay rather than all late claims.
- What "granted" means
- That the tribunal exercised its discretion to allow the claim to proceed out of time. It says nothing about what happened at the merits hearing afterwards, which is analysed separately below.
- Treat all counts as floors
- Extraction from published judgments recovers what the tribunal chose to record. Where a figure is a rate rather than a count, the rate is the more reliable number.
- What this cannot tell you
- Published judgments are not a sample of everyone who misses a deadline. Claims that settle, are withdrawn, or are never brought at all produce no judgment and are invisible here. There is no reason to assume those absent cases resemble the visible ones.
Finding 1: how late were the claims that got through?
| How late | Cases | Extension granted | Just and equitable | Not reasonably practicable |
|---|---|---|---|---|
| 1–7 days | 99 | 22.2% | 44.1% | 11.1% |
| 8–30 days | 192 | 22.4% | 39.7% | 11.1% |
| 1–3 months | 161 | 18.0% | 26.9% | 9.8% |
| 3–6 months | 125 | 12.8% | 20.9% | 3.8% |
| 6–12 months | 155 | 14.2% | 23.3% | 3.1% |
| Over a year | 191 | 8.9% | 12.4% | 0.0% |
| Total | 923 | |||
The distribution is flatter than almost anyone expects. At one to seven days late, 78% of applications were still refused. A claim a week late and a claim a month late are close to indistinguishable, 22.2% against 22.4%. The curve does decline with delay, but gradually, and it does not fall to zero: around a quarter of all granted extensions were more than six months late, and one was granted five years out of time.
Two practical consequences follow. Being marginally late is not the near-certain rescue that the phrase "only a few days" implies. And being very late is not the automatic bar that most guides imply either.
Finding 2: the test matters more than the delay
The largest observed difference in the data is not how late the claim was. It is which statutory test applied to it.
Unfair dismissal and most Employment Rights Act claims are governed by section 111(2)(b) ERA 1996, which permits a late claim only where it was "not reasonably practicable" to claim in time. Discrimination claims are governed by section 123(1)(b) EqA 2010, which permits "such other period as the employment tribunal thinks just and equitable".
At one to seven days late, the just and equitable route was granted 44.1% of the time. The not reasonably practicable route, at the identical delay, was granted 11.1%. That is a four-fold difference at the same degree of lateness, and the gap persists in every band. In the over-a-year band, the not reasonably practicable route succeeded in none of the cases in our sample.
Put another way: the difference between the two tests is larger than the difference between a claim one week late and a claim more than a year late.
The hardest cut in the data. Of late unfair dismissal claims brought with no discrimination claim alongside them, around 5.5% were allowed to proceed, roughly one in eighteen. That is the most common claim type in the tribunal system, under the strictest test, with nothing to soften it.
Finding 3: the grant rate has fallen
| Period | Observed grant rate |
|---|---|
| 2019–2021 | 17.3% |
| 2023–2026 | 13.4% |
A fall of more than a fifth. Over the same period the proportion of decided claims lost on time limits stayed broadly flat, so this is not a story about more people missing deadlines. It is a change in what happens to those who ask.
We would not present this as proof that tribunals have become stricter. The mix of cases reaching a published judgment can shift over time, and a compositional change would produce the same pattern. But the direction is consistent across the period rather than driven by a single year, and the timing is worth noting: it arrives immediately before the limitation period itself doubles from three months to six on 1 October 2026.
Finding 4: the escape route nobody counts
Extensions are the well-known way past a time bar. They are not the only one, and the alternative is far larger than the commentary suggests.
Where discrimination amounted to conduct extending over a period rather than a one-off act, section 123(3)(a) EqA treats it as done at the end of that period, so time runs from the last point rather than the first. Establishing that does not ask the tribunal for a favour. It relocates the start date as of right.
Tribunals considered a continuing act argument in 2,335 cases:
| Tribunal found | Claims in time | Claims out of time |
|---|---|---|
| Continuing act established | 485 | 143 |
| Continuing act rejected | 19 | 1,681 |
| Other or not classified | 7 | — |
Where the argument was accepted, 77% of those claims were in time. Where it was rejected, 99% were out. The argument succeeds in around 27% of the cases where it is run, and around 91% of the cases running it involve a discrimination claim, section 123(3) being an Equality Act provision.
Across cases where a claim escaped a time bar, roughly a third got through by showing they were never late rather than by being forgiven for it: 548 cases against 990.
Note the 143 in the first row. Establishing a continuing act does not by itself put a claimant in time. It moves the start date to the end of the conduct, and the claim must still have been brought within the limit running from that point. Those 143 won the argument and were late regardless.
Finding 5: what happens after the time point is won
This is the question no published source appears to answer, and it is the most encouraging figure in the dataset.
| Route in | Succeeded at merits | n |
|---|---|---|
| Rescued by an extension | 69.6% | 588 |
| In time via a continuing act | 72.0% | 279 |
Median award in those cases £12,856, with the middle half between £4,386 and £35,643 and the top tenth above £84,567.
What survives the caveat is the shape of the thing: for these claimants the time limit functioned as a gate rather than as a verdict on the case. Passing it was hard. Having passed it, they won roughly seven times in ten.
What we looked at and chose not to publish
Reporting the negative results matters as much as the positive ones.
- Regional variation
- Grant rates across tribunal regions spread from 11.0% to 19.4%. We are not publishing this as a finding. Two earlier rounds of analysis produced regional differences that dissolved into case-mix artefacts once claim type and test were controlled for, and we have no reason to think this one is different. A reader choosing where to bring a claim on the strength of a regional number would be acting on noise.
- Judge-level variation
- Computable from the deciding authority field. We consider publishing it editorially wrong and have not done so.
- Seasonality
- Cannot be asked cleanly without a reliable presentation date across the corpus.
- Representation effects
- Measurable, but confounded in a way we cannot currently separate: claimants who obtain representation differ systematically from those who do not, and Adedeji is a reminder that access to advice and acting on it are different things.
Ask this dataset your own question
These figures come from 133,285 Employment Tribunal cases. The Intelligence Hub lets you cut them by claim type, outcome, respondent and award.
Request early access →What this means if you are the one who is late
The full practical guidance is in our guide to missing the tribunal deadline. In short, three things follow from the data above.
Check whether you are actually late before you concede that you are, because the continuing-act route is both stronger and more common than the literature suggests. Establish which test applies to you, because it is the largest single factor in the data. And do not treat a short delay as safe or a long one as hopeless, because the distribution supports neither.
Frequently asked questions
How often do employment tribunals extend time limits?
Across the published decisions we analysed, extensions were granted in roughly one case in seven. The rate varies sharply by test: about 21% under the just and equitable test and about 9% under not reasonably practicable.
Does being only a few days late help?
Less than most people assume. At one to seven days late, around 78% of applications in our sample were still refused. The grant rate at a week late, 22.2%, is almost identical to the rate at a month late, 22.4%.
Which claims get extended most often?
Discrimination claims under the Equality Act, which use the broader just and equitable test. At one to seven days late they were extended 44.1% of the time against 11.1% for unfair dismissal.
What is the latest a claim has been allowed in?
In our sample, five years out of time. Around a quarter of granted extensions were more than six months late, so no delay is automatically fatal, though the rate does decline with length.
Do late claims that get in actually win?
In this dataset, roughly seven in ten succeeded at merits, with a median award of £12,856. That figure describes a heavily selected group: merit is weighed in some admission decisions, only published cases are visible, and claims that settle or withdraw after winning jurisdiction never reach a merits judgment.
Are extensions getting harder to obtain?
The observed grant rate fell from 17.3% in 2019–21 to 13.4% in 2023–26. We would not call that proof tribunals have hardened, as the mix of published cases can shift, but the direction is consistent across the period.
Where does this data come from?
Published Employment Tribunal decisions from 2000 to 2026, identified using a structured field recording determinations on jurisdiction where limitation was in issue, regardless of which way the time point went. The lateness table covers the 923 judgments that stated the length of delay.
Sources
- Yerty Intelligence Hub — analysis of published employment tribunal judgments, 2000 to 2026
- Employment Rights Act 1996, s.111 (legislation.gov.uk)
- Equality Act 2010, s.123 (legislation.gov.uk)
- Employment Rights Act 1996, s.207B (legislation.gov.uk)
- Employment Rights Act 2025 (ACAS)
- Hendricks v Commissioner of Police for the Metropolis [2002] EWCA Civ 1686, Court of Appeal
- Adedeji v University Hospitals Birmingham NHS Foundation Trust [2021] EWCA Civ 23, Court of Appeal
Related: missed tribunal deadline guide · ACAS early conciliation · tribunal success rates · discrimination success rates · award amounts
How to cite this
Yerty (2026). How often do employment tribunals extend time limits?. Yerty Intelligence Hub. https://yerty.co.uk/tribunal-data/findings/employment-tribunal-time-limit-extensions (last updated August 2026).Figures reflect the analytical subset used for this question; see our methodology for corpus coverage. Free to cite for non-commercial use with attribution to Yerty. For commercial use, bulk or API access, or AI training, see our Data Use & Citation policy.
Figures derived from Employment Tribunal decisions published by HM Courts & Tribunals Service on GOV.UK, licensed under the Open Government Licence v3.0. Analysis © Yerty. Re-use is subject to our Data Use & Citation policy.
Yerty provides information, not legal advice. These figures describe past decided cases. They cannot predict the outcome of any individual application. For advice on your circumstances, consider speaking to a qualified employment solicitor.