# Communications offences: the reports, the charges, and the arrests nobody publishes

Police in England and Wales recorded 91,987 malicious communications offences in 2025/26, down from 288,939 at the 2021/22 peak, a fall the statisticians attribute mainly to a change in counting rules rather than fewer reports. Three in every hundred closed investigations ended in a charge. The courts convicted 2,487 people under the four communications offences in 2025. The number arrested in between is published by no government department, which is why a campaign group's Freedom of Information tally of 62,199 arrests over five years could run across the press unchallenged and unconfirmed.

Source page: https://britishresilienceindex.com/deep-dive/communications-offences
Machine readable data: https://britishresilienceindex.com/data/deep-dive/communications-offences.json

## How to read this

Three official series that do not join up, shown side by side. Recorded offences are reports the police logged under Home Office code 8R, before any suspect was identified, by financial year. Outcomes are the closing outcome of each investigation, by financial year. Prosecutions and convictions are the Ministry of Justice's Court Proceedings Database, by calendar year, counted on the principal offence, summed across section 127 of the Communications Act 2003, section 1 of the Malicious Communications Act 1988 and the two Online Safety Act 2023 offences that replaced parts of them in January 2024. Arrests by offence are not published: the only arrest figures on this page are one force's own Freedom of Information disclosure and, clearly labelled, a campaign group's reported total. England and Wales only.

## Four offences, two of them new, none of them written for social media

"Communications offences" is shorthand for a small family of crimes about sending a message, not about what happened afterwards. Two date from before the web. Two arrived on 31 January 2024, when the Online Safety Act 2023 repealed the false message and threat limbs of the old offences and replaced them.

| name | detail |
|---|---|
| Communications Act 2003, section 127 | Sending, by a public electronic communications network, a message that is grossly offensive or of an indecent, obscene or menacing character; and, until 31 January 2024, sending a false message to cause annoyance, inconvenience or needless anxiety. Summary only, tried in the magistrates' court. Ministry of Justice offence code 19607. |
| Malicious Communications Act 1988, section 1 | Sending a letter, electronic communication or article that is indecent or grossly offensive, or that conveys a threat or false information, with the purpose of causing distress or anxiety. Made triable either way in 2015 with a maximum of two years. The threat and false information limbs were repealed on 31 January 2024. Ministry of Justice offence code 00872. |
| Online Safety Act 2023, section 181 | Threatening communications: sending a message that conveys a threat of death or serious harm, intending or being reckless as to whether the recipient fears it will be carried out. Triable either way, up to five years. Ministry of Justice offence code 00528, filed under endangering life. |
| Online Safety Act 2023, section 179 | False communications: sending information the sender knows to be false, intending to cause non trivial psychological or physical harm to a likely audience, without reasonable excuse. Summary only. Ministry of Justice offence code 13852. |

- **Note**: The police record all of these under one Home Office code, 8R Malicious Communications, so the recorded series cannot be split by statute. The courts data can.

## 91,987 offences recorded in 2025/26, 68 per cent below the peak, mostly on paper

Recorded malicious communications rose every year from 2017/18 to a peak of 288,939 in 2021/22, then fell by 38 per cent in 2023/24 alone. The Office for National Statistics attributes that fall to a change in how the police count, not to a change in what the public reports.

**Series**

| fy | n |
|---|---|
| 2017/18 | 159,826 |
| 2018/19 | 208,177 |
| 2019/20 | 228,031 |
| 2020/21 | 274,307 |
| 2021/22 | 288,939 |
| 2022/23 | 274,777 |
| 2023/24 | 170,370 |
| 2024/25 | 98,889 |
| 2025/26 | 91,987 |

**Break quote**: text The changes in recording rules for conduct crimes (stalking, harassment and coercive and controlling behaviour) in May 2023 removed the requirement to record two crimes when one of them was a conduct crime. This has led to a reduction in those crimes often associated with conduct crimes, including malicious communications., cite Office for National Statistics, Crime in England and Wales: year ending March 2024, section 10

- **Note**: Before May 2023 a victim of harassment who was also sent abusive messages generated two recorded crimes; since then, one. The series is therefore not comparable across that line, and the page draws no conclusion about demand from it. The financial year 2023/24 straddles the change.

## Recorded offences per 1,000 residents vary more than fourfold between forces

England and Wales recorded 1.49 malicious communications offences per 1,000 residents in 2025/26. The spread between forces is wide, and on this evidence it is a spread in recording practice as much as in behaviour: the same counting rule change landed at different times and to different degrees.

- **Ew rate**: 1.49
**Top**

| force | n | pop | rate |
|---|---|---|---|
| Metropolitan Police | 25,095 | 9,074,625 | 2.77 |
| West Midlands | 6,955 | 3,036,605 | 2.29 |
| Merseyside | 3,182 | 1,475,541 | 2.16 |
| Gwent | 1,217 | 601,686 | 2.02 |
| South Wales | 2,380 | 1,363,561 | 1.75 |
| Greater Manchester | 5,173 | 3,009,664 | 1.72 |
| Avon and Somerset | 2,952 | 1,813,665 | 1.63 |
| Durham | 1,051 | 650,500 | 1.62 |

**Bottom**

| force | n | pop | rate |
|---|---|---|---|
| Humberside | 419 | 962,532 | 0.44 |
| South Yorkshire | 852 | 1,430,623 | 0.6 |
| North Yorkshire | 604 | 844,571 | 0.72 |
| Lincolnshire | 639 | 789,502 | 0.81 |
| Surrey | 1,019 | 1,248,649 | 0.82 |
| Staffordshire | 964 | 1,177,578 | 0.82 |
| Wiltshire | 637 | 767,575 | 0.83 |
| Suffolk | 664 | 786,231 | 0.84 |

- **Note**: Home Office recorded crime open data, offence code 8R, financial year 2025/26, against ONS mid 2024 population. British Transport Police is excluded as it has no resident population, and the City of London because its 15,000 residents are a fraction of its daytime population.

## Three in a hundred investigations end in a charge; four in ten end because the victim withdraws

Of 92,032 malicious communications investigations closed in 2025/26, 2,750 ended in a charge or summons, 3.0 per cent. The single largest outcome, 42.3 per cent, is the victim not supporting further action.

**Mix**

| label | n | pct |
|---|---|---|
| Victim did not support further action | 38,918 | 42.3 |
| No suspect identified | 24,684 | 26.8 |
| Suspect identified, evidential difficulties | 20,186 | 21.9 |
| Charged or summonsed | 2,750 | 3 |
| Out of court, informal | 2,281 | 2.5 |
| Prosecution prevented or not in the public interest | 689 | 0.7 |
| Out of court, formal | 667 | 0.7 |
| Other outcomes | 1,857 | 2 |

**Charge series**

| fy | closed | charged | chargePct |
|---|---|---|---|
| 2018/19 | 199,530 | 4,929 | 2.5 |
| 2023/24 | 188,039 | 4,538 | 2.4 |
| 2024/25 | 105,737 | 2,790 | 2.6 |
| 2025/26 | 92,032 | 2,750 | 3 |

- **Met note**: The Metropolitan Police closed 24,271 malicious communications investigations in 2025/26, more than a quarter of the national total, with a charge rate of 1.5 per cent.
- **Note**: Home Office crime outcomes open data, offence code 8R, outcomes assigned to investigations closed in the year, all forces. The charge rate is charges as a share of closed investigations. Forces with fewer than 500 closed investigations are excluded from the force comparison.

## 2,487 convictions in 2025, about the same as 2017 once the new offences are added back

The Ministry of Justice counts 1,207 convictions under section 127 in 2025, 541 under the Malicious Communications Act, and 739 under the two Online Safety Act offences that replaced its threat and false message limbs in January 2024. Summed, the four offences have produced between 2,402 and 2,859 convictions a year since 2017.

**Series**

**Year**: 2017 · **S127 pros**: 1,505 · **S127 conv**: 1,336 · **Mca pros**: 1,720 · **Mca conv**: 1,457 · **Osa threat pros**: 0 · **Osa threat conv**: 0 · **Osa false pros**: 0 · **Osa false conv**: 0 · **Conv total**: 2,793 · **Pros total**: 3,225

**Year**: 2018 · **S127 pros**: 1,317 · **S127 conv**: 1,133 · **Mca pros**: 1,842 · **Mca conv**: 1,587 · **Osa threat pros**: 0 · **Osa threat conv**: 0 · **Osa false pros**: 0 · **Osa false conv**: 0 · **Conv total**: 2,720 · **Pros total**: 3,159

**Year**: 2019 · **S127 pros**: 1,148 · **S127 conv**: 1,007 · **Mca pros**: 1,732 · **Mca conv**: 1,523 · **Osa threat pros**: 0 · **Osa threat conv**: 0 · **Osa false pros**: 0 · **Osa false conv**: 0 · **Conv total**: 2,530 · **Pros total**: 2,880

**Year**: 2020 · **S127 pros**: 969 · **S127 conv**: 853 · **Mca pros**: 1,772 · **Mca conv**: 1,549 · **Osa threat pros**: 0 · **Osa threat conv**: 0 · **Osa false pros**: 0 · **Osa false conv**: 0 · **Conv total**: 2,402 · **Pros total**: 2,741

**Year**: 2021 · **S127 pros**: 1,187 · **S127 conv**: 1,036 · **Mca pros**: 2,124 · **Mca conv**: 1,823 · **Osa threat pros**: 0 · **Osa threat conv**: 0 · **Osa false pros**: 0 · **Osa false conv**: 0 · **Conv total**: 2,859 · **Pros total**: 3,311

**Year**: 2022 · **S127 pros**: 1,142 · **S127 conv**: 989 · **Mca pros**: 1,992 · **Mca conv**: 1,764 · **Osa threat pros**: 0 · **Osa threat conv**: 0 · **Osa false pros**: 0 · **Osa false conv**: 0 · **Conv total**: 2,753 · **Pros total**: 3,134

**Year**: 2023 · **S127 pros**: 1,257 · **S127 conv**: 1,090 · **Mca pros**: 1,816 · **Mca conv**: 1,627 · **Osa threat pros**: 0 · **Osa threat conv**: 0 · **Osa false pros**: 0 · **Osa false conv**: 0 · **Conv total**: 2,717 · **Pros total**: 3,073

**Year**: 2024 · **S127 pros**: 1,325 · **S127 conv**: 1,157 · **Mca pros**: 1,162 · **Mca conv**: 1,096 · **Osa threat pros**: 444 · **Osa threat conv**: 292 · **Osa false pros**: 15 · **Osa false conv**: 14 · **Conv total**: 2,559 · **Pros total**: 2,946

**Year**: 2025 · **S127 pros**: 1,403 · **S127 conv**: 1,207 · **Mca pros**: 602 · **Mca conv**: 541 · **Osa threat pros**: 1,033 · **Osa threat conv**: 722 · **Osa false pros**: 26 · **Osa false conv**: 17 · **Conv total**: 2,487 · **Pros total**: 3,064


- **Switch note**: Malicious Communications Act convictions fell from 1,627 in 2023 to 541 in 2025 while threatening communications convictions rose from nothing to 722. That is the statute book moving, not the courts: section 189 of the Online Safety Act 2023 repealed the threat limb of the 1988 Act on 31 January 2024.
**Who**: male pct 83, children 56, young adults 109, note Of those convicted in 2025, 83 per cent were male, 56 were children under 18 and 109 were young adults aged 18 to 20.

**Top forces**

| force | n |
|---|---|
| Metropolitan | 260 |
| Greater Manchester | 162 |
| South Wales | 111 |
| West Midlands | 109 |
| Cheshire | 99 |
| Merseyside | 97 |
| Essex | 93 |
| Northumbria | 88 |

- **Note**: Ministry of Justice outcomes by offence, Court Proceedings Database, England and Wales, calendar years, principal offence basis. The Ministry advises against dividing convictions by prosecutions as a rate, because the two are counted at different stages and can fall in different years. Figures from 2017 are on a revised processing basis and are not comparable with earlier years.

## Arrests sit between the report and the charge, and no department publishes them

The Home Office publishes arrests by broad offence group, not by offence, so there is no official count of arrests for communications offences. Each force holds its own. The page can show every stage of the pipeline except that one.

**National**

| stage | n | basis |
|---|---|---|
| Offences recorded by police, 2025/26 | 91,987 | Home Office recorded crime, code 8R |
| Investigations closed with a charge or summons, 2025/26 | 2,750 | Home Office crime outcomes, code 8R |
| Prosecutions, calendar year 2025 | 3,064 | Ministry of Justice, four offence codes, principal offence |
| Convictions, calendar year 2025 | 2,487 | Ministry of Justice, four offence codes, principal offence |

**Reported**: figure 62,199, text On 6 September 2026 GB News reported, from a forthcoming Big Brother Watch report, that at least 62,199 people were arrested for communications offences between 2021 and 2026, of whom 12,292 were convicted. The figures are a compilation of Freedom of Information responses from individual forces. The report was not published at the time of writing and the method could not be checked; two outlets already give different figures for the number charged. The arrest total is shown here as reported and is not used in any calculation., cross check One partial check is possible. The Ministry of Justice counts 13,375 convictions under the four offences in the five calendar years 2021 to 2025, which is the same order of magnitude as the reported 12,292, on a different period and a principal offence basis.

**West midlands**: note West Midlands Police, Freedom of Information disclosure 1732A/24 (December 2024): arrests by calendar year under section 127 and section 1 of the Malicious Communications Act, summed. Recorded offences are the financial year beginning in April of that year, Home Office code 8R. Prosecutions and convictions are the Ministry of Justice count for the West Midlands police force area, the four offence codes summed. The force warns its figures were collated for an ad hoc request from several systems and should be read with caution., reading In 2023 the force recorded 8,867 malicious communications offences, made 644 arrests under the two statutes, and the courts convicted 89 people from its area. Roughly one recorded offence in 14 led to an arrest, and one arrest in 7 to a conviction.

## The register of things this page cannot show

| question | status | detail |
|---|---|---|
| How many people are arrested for communications offences each year? | not published | Home Office arrest statistics are broken down by offence group (violence against the person, public order and so on), never by offence. Every national figure in circulation is a Freedom of Information compilation with uneven coverage and definitions. |
| What happened to the people who were arrested but not charged? | not published | The outcomes data is per investigation, not per arrest. An arrest that ends in release with no further action, a caution or a community resolution is visible only in the force's own systems. |
| How many recorded offences are the new Online Safety Act crimes? | not separable | The police record every communications offence under one Home Office code, 8R. Only the courts data can tell the 2003 Act, the 1988 Act and the 2023 Act apart. |
| How many reports were about social media, as against text messages, letters or calls? | not published | Neither the recorded crime nor the courts data records the medium. The phrase "social media arrests" has no official denominator. |
| What about Scotland and Northern Ireland? | separate systems | Section 127 applies across the UK but recording and prosecution are devolved. Scotland's recorded crime and the Crown Office publish on their own bases and are not included here. |

## What the numbers leave open

- If the Home Office can publish outcomes for every recorded offence, why can it not publish arrests by offence, when every force already holds the figure and answers Freedom of Information requests about it?
- Four in ten malicious communications investigations close because the victim withdraws. Is that victims deciding the process is not worth it, or the police recording a report the victim never wanted pursued?
- The charge rate has moved from 2.5 to 3.0 per cent in seven years while recorded offences fell by more than half. What does a three per cent charge rate on a report based offence mean for the 97 per cent?
- The counting rule change of May 2023 coincided with roughly 100,000 fewer recorded offences the following year. Should a series with a break that size be published without a marked discontinuity in every table?
- Convictions under the four offences have held at roughly 2,400 to 2,900 a year since 2017. If arrests have risen as the campaign figures suggest, is the pipeline filling at the front and emptying in the middle, and where?

## Sources

- Police recorded crime open data tables, police force area, year ending March 2026 (Home Office, published July 2026) (https://www.gov.uk/government/statistical-data-sets/police-recorded-crime-and-outcomes-open-data-tables) · Open Government Licence v3.0
- Crime outcomes in England and Wales, year ending March 2026 (Home Office) (https://www.gov.uk/government/statistics/crime-outcomes-in-england-and-wales-2025-to-2026) · Open Government Licence v3.0
- Criminal Justice Statistics Quarterly: December 2025, outcomes by offence data (Ministry of Justice, published April 2026) (https://www.gov.uk/government/statistics/criminal-justice-statistics-quarterly-december-2025) · Open Government Licence v3.0
- Crime in England and Wales: year ending March 2024, section 10, stalking and harassment (Office for National Statistics) (https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/bulletins/crimeinenglandandwales/yearendingmarch2024) · Open Government Licence v3.0
- Population estimates for police force areas, mid 2024 (Office for National Statistics) (https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationestimates/datasets/populationestimatesforpoliceforceareasinenglandandwales) · Open Government Licence v3.0
- Online Safety Act 2023, section 189, repeals in connection with the new communications offences (https://www.legislation.gov.uk/ukpga/2023/50/section/189) · Open Government Licence v3.0
- West Midlands Police, Freedom of Information disclosure 1732A/24, malicious communications arrests 2017 to 2024 (December 2024) (https://www.westmidlands.police.uk/foi-ai/west-midlands-police/disclosure-log/2024/december/malicious-communications-foi-ref-1732a24/) · Freedom of Information Act 2000 disclosure, published by the force
- GB News, 6 September 2026, reporting the Big Brother Watch arrest figures (attribution only, not used in any calculation) (https://www.gbnews.com/news/free-speech-row-more-than-60000-arrested-communications-offences) · Press report, quoted under attribution

## Caveats and data notes

- The recorded offence series breaks in May 2023. The Home Office removed the requirement to record a second crime alongside a stalking, harassment or controlling behaviour offence, and ONS attributes the 39 per cent fall in malicious communications in 2023/24 to that change. No figure on this page treats the fall as a fall in reports.
- Recorded offences, outcomes and court proceedings are three different counts of three different things: a report, the closure of an investigation, and a defendant's principal offence. They are shown as stages of one pipeline for orientation only; a case can move between years and between offence codes on the way.
- The Ministry of Justice counts on a principal offence basis, so a defendant convicted of a more serious offence alongside a communications offence is not counted here. It advises against calculating a conviction rate from its prosecution and conviction columns, and this page does not.
- The Online Safety Act 2023 repealed parts of both older offences from 31 January 2024. Convictions are therefore summed across four offence codes; any single line, read alone, shows the statute book moving.
- Arrests by offence are not published nationally. The West Midlands figures are the force's own Freedom of Information disclosure, collated for an ad hoc request and carrying the force's caution. The 62,199 figure is a campaign group's compilation reported by the press, not verified here and not used in any calculation.
- Recorded crime open data is revised between editions. The year ending March 2024 total here is within two per cent of the figure ONS published in its bulletin at the time, and the build asserts that.

---

British Resilience Index, a non-partisan UK civic data project. Every figure is
reconciled to its primary official publisher before it is shown. Free to cite with
attribution: https://britishresilienceindex.com