Crime · South Africa · Updated 2026-10-26

What is South Africa's murder rate?

SAPS recorded 24,692 murders in South Africa from April 2024 to March 2025, 39.84 per 100,000 people; Stats SA's death records for 2024 give assault as the cause of 7,486 deaths (12.08 per 100,000), with 78.5% of non-natural deaths coded as other accidental injury or undetermined intent.

Every figure in this summary is calculated from the sources listed on this page.

Murder rate in South Africa, 2024/25

Murder victims (unlawful and intentional killings) per 100,000 people. South Africa, national, April 2024 to March 2025 (SAPS financial year 2024/25); death records for calendar year 2024. Unit: per 100,000 people.

39.84per 100,000 people, 2024

60Trust scoreModerate certainty

Quality34 / 40: How well the sources are made
Agreement0 / 30: How close the sources are to each other
Independence16 / 20: How many separate ways the number was measured
Freshness10 / 10: How recent the newest data is

How much each source counts

SAPS63%
Stats SA38%

Shown but not counted: UNODC (republishes SAPS police recorded crime statistics, annual figures); WHO (republishes Stats SA, Mortality and causes of death: findings from death notification (P0309.3)).

Sources

Source	Value	Definition	Quality	Weight
SAPS police recorded crime statistics, annual figuresSouth African Police Service (SAPS) · Police and justice recordsRetrieved 2026-10-26	39.842024	Murders recorded by SAPS at all police stations from April 2024 to March 2025 (financial year 2024/25): unlawful and intentional killings, one count per victim, including farm murders and murders of police officers. A crime is counted when it is registered at a police station, whatever date it was committed. Divided by Stats SA's estimate of the population on 30 June 2024 (61,983,505; Mid-year population estimates 2026, P0302). The same Stats SA population series is used for every year and for the SAPS, Stats SA and WHO lines. Counts from the station-level sheet of SAPS's 2024/25 annual Excel file; the national row equals the sum of station rows. SAPS's own release gives 39.1 per 100,000, dividing by 63,115,905 people (https://www.saps.gov.za/services/downloads/2024-2025%20_Annual_Financial%20year_WEB.pdf, p. 18). Murders fell 10.6% from 2023/24.	5 of 5 checks Yes: Method and definitions published. SAPS publishes its methodology, counting rules and crime definitions in each release. Yes: Official statistics or peer-reviewed. SAPS states that its crime statistics were granted official-statistics status by the Statistician-General after an independent quality assessment in 2018/19. Yes: Full count or representative sample. Counts every murder recorded at all police stations, not a sample. Yes: Raw data downloadable. Murder counts by station and financial year are a free Excel download from the SAPS website. Yes: No known break or quality problem. No break in the murder series 2015/16-2024/25. In 2025 SAPS withdrew its breakdown of victims who were women and children after finding duplicate counts; the murder total was not withdrawn.	63%
UNODC Data Portal, victims of intentional homicideUnited Nations Office on Drugs and Crime (UNODC) · Police and justice recordsRetrieved 2026-10-26	39.722024	Victims of intentional homicide, total for both sexes and all ages, rate per 100,000 from UNODC's own population figures (UN World Population Prospects 2024). UNODC source code: NP/CTS/NP Adjusted. UNODC's counts for 2015-2021 equal SAPS's April-March totals for 2015/16-2021/22; for 2024 it gives 25,423 victims against SAPS's 24,692 for 2024/25. Raw file: data_cts_intentional_homicide.xlsx (UNODC, updated 12 July 2026).	4 of 5 checks Yes: Method and definitions published. UNODC publishes the ICCS definition and a metadata note for the dataset. Yes: Official statistics or peer-reviewed. National figures submitted by governments through the UN crime trends survey (UN-CTS) or official channels. Yes: Full count or representative sample. A count of all victims recorded by the authorities, not a sample. Yes: Raw data downloadable. The whole dataset is a free Excel download from the UNODC data portal. No: No known break or quality problem. UNODC marks this value "NP Adjusted", so it is not a plain national count.	Not counted
Stats SA, Mortality and causes of death: findings from death notification (P0309.3)Statistics South Africa (Stats SA) · Death and health registersRetrieved 2026-10-26	12.082024	Deaths in 2024 whose underlying cause on the death notification form is assault (ICD-10 X85-Y09), from the civil registration system of the Department of Home Affairs, as processed by Stats SA for its 2026 release. Divided by Stats SA's estimate of the population on 30 June 2024 (61,983,505; Mid-year population estimates 2026, P0302). The same Stats SA population series is used for every year and for the SAPS, Stats SA and WHO lines. Table 4.11 (p. 49) of the 2024 release (https://www.statssa.gov.za/publications/P03093/P030932024.pdf) and sheet App.K of its appendix tables (this line's link). Of 62,871 non-natural deaths, 47,098 were coded as other external causes of accidental injury (W00-X59) and 2,265 as undetermined intent (Y10-Y34).	3 of 5 checks Yes: Method and definitions published. Stats SA publishes its method, ICD-10 coding rules and data notes in the release. Yes: Official statistics or peer-reviewed. Official statistics from Statistics South Africa, based on civil registration. Yes: Full count or representative sample. Covers every death registered with Home Affairs and processed by Stats SA, not a sample. No: Raw data downloadable. The 2024 release has aggregated tables only; record-level death data on DataFirst so far reach 2023. No: No known break or quality problem. Stats SA says over three-quarters of non-natural deaths were not adequately classified and that this may partly explain the low assault count; late registrations will also add deaths for this year.	38%
WHO Mortality Database, deaths by assaultWorld Health Organization (WHO) · Death and health registersRetrieved 2026-10-26	9.562020	Registered deaths with underlying cause assault (ICD-10 X85-Y09), all ages and both sexes, in 2020, divided by Stats SA's estimate of the population on 30 June 2020 (59,284,850; same series as the other lines). WHO's 2020 counts (5,668 assault, 2,603 undetermined intent) equal those in Stats SA's 2020 release, so this is Stats SA's death register passed on; WHO has no South African data after 2020. Raw files: morticd10_part1-6.zip.	4 of 5 checks Yes: Method and definitions published. WHO publishes the file layout, ICD-10 code lists and data notes in its documentation. Yes: Official statistics or peer-reviewed. Official cause-of-death statistics sent to WHO by Stats SA. Yes: Full count or representative sample. Civil registration counts every registered death, not a sample. Yes: Raw data downloadable. Raw death and population files are free downloads. No: No known break or quality problem. The same coding problem as Stats SA: in 2020, 2,603 deaths had undetermined intent and many violent deaths are coded as accidental.	Not counted

Murder rate in South Africa, 2024/25, by source · per 100,000 people

SAPS police recorded crime statistics, annual figures
UNODC Data Portal, victims of intentional homicide
WHO Mortality Database, deaths by assault

Show the data as a table

Year	SAPS police recorded crime statistics, annual figures	UNODC Data Portal, victims of intentional homicide	WHO Mortality Database, deaths by assault
2015	33.84	32.92	13.05
2016	33.98	33.21	14.36
2017	35.85	35.28	14.13
2018	36.52	35.87	13.19
2019	36.49	35.79	13
2020	33.69	32.98	9.56
2021	42.05	40.94	–
2022	45.43	43.72	–
2023	45.11	43.38	–
2024	39.84	39.72	–

Murders recorded by SAPS, 2024/25

Murder victims recorded by the police. South Africa, national, April 2024 to March 2025 (financial year 2024/25). Unit: murders.

24,692murders, 2024

71Trust scoreModerate certainty

Only one source – not confirmed by others. Agreement can't be measured, so the score can't reach High.

Quality40 / 40: How well the sources are made
Agreement15 / 30: Unknown with one source, so set to half
Independence6 / 20: How many separate ways the number was measured
Freshness10 / 10: How recent the newest data is

How much each source counts

SAPS100%

Shown but not counted: UNODC (republishes SAPS police recorded crime statistics, annual figures).

Sources

Source	Value	Definition	Quality	Weight
SAPS police recorded crime statistics, annual figuresSouth African Police Service (SAPS) · Police and justice recordsRetrieved 2026-10-26	24,6922024	Murders recorded by SAPS at all police stations from April 2024 to March 2025 (financial year 2024/25): unlawful and intentional killings, one count per victim, including farm murders and murders of police officers. A crime is counted when it is registered at a police station, whatever date it was committed. Counts from the station-level sheet of SAPS's 2024/25 annual Excel file; the national row equals the sum of station rows. SAPS's own release gives 39.1 per 100,000, dividing by 63,115,905 people (https://www.saps.gov.za/services/downloads/2024-2025%20_Annual_Financial%20year_WEB.pdf, p. 18). Murders fell 10.6% from 2023/24.	5 of 5 checks Yes: Method and definitions published. SAPS publishes its methodology, counting rules and crime definitions in each release. Yes: Official statistics or peer-reviewed. SAPS states that its crime statistics were granted official-statistics status by the Statistician-General after an independent quality assessment in 2018/19. Yes: Full count or representative sample. Counts every murder recorded at all police stations, not a sample. Yes: Raw data downloadable. Murder counts by station and financial year are a free Excel download from the SAPS website. Yes: No known break or quality problem. No break in the murder series 2015/16-2024/25. In 2025 SAPS withdrew its breakdown of victims who were women and children after finding duplicate counts; the murder total was not withdrawn.	100%
UNODC Data Portal, victims of intentional homicideUnited Nations Office on Drugs and Crime (UNODC) · Police and justice recordsRetrieved 2026-10-26	25,4232024	Victims of intentional homicide, total for both sexes and all ages, rate per 100,000 from UNODC's own population figures (UN World Population Prospects 2024). UNODC source code: NP/CTS/NP Adjusted. UNODC's counts for 2015-2021 equal SAPS's April-March totals for 2015/16-2021/22; for 2024 it gives 25,423 victims against SAPS's 24,692 for 2024/25. Raw file: data_cts_intentional_homicide.xlsx (UNODC, updated 12 July 2026).	4 of 5 checks Yes: Method and definitions published. UNODC publishes the ICCS definition and a metadata note for the dataset. Yes: Official statistics or peer-reviewed. National figures submitted by governments through the UN crime trends survey (UN-CTS) or official channels. Yes: Full count or representative sample. A count of all victims recorded by the authorities, not a sample. Yes: Raw data downloadable. The whole dataset is a free Excel download from the UNODC data portal. No: No known break or quality problem. UNODC marks this value "NP Adjusted", so it is not a plain national count.	Not counted

Murders recorded by SAPS, 2024/25, by source · murders

SAPS police recorded crime statistics, annual figures
UNODC Data Portal, victims of intentional homicide

Show the data as a table

Year	SAPS police recorded crime statistics, annual figures	UNODC Data Portal, victims of intentional homicide
2015	18,673	18,673
2016	19,016	19,016
2017	20,336	20,336
2018	21,022	21,022
2019	21,325	21,325
2020	19,972	19,972
2021	25,181	25,181
2022	27,494	27,272
2023	27,621	27,419
2024	24,692	25,423

Households that experienced a murder, 2025/26

Share of households that say they experienced a murder in the 12 months before the interview. South Africa, private households, Survey year 2025/26 (April 2025 to March 2026). Unit: % of households.

0.1% of households, 2025

55Trust scoreLow certainty

Only one source – not confirmed by others. Agreement can't be measured, so the score can't reach High.

Quality24 / 40: How well the sources are made
Agreement15 / 30: Unknown with one source, so set to half
Independence6 / 20: How many separate ways the number was measured
Freshness10 / 10: How recent the newest data is

How much each source counts

GPSJS100%

Sources

Source	Value	Definition	Quality	Weight
Stats SA, Governance, Public Safety and Justice Survey (Victims of Crime, P0341)Statistics South Africa (Stats SA) · Victim surveyRetrieved 2026-10-26	0.12025	Households answering yes to "Have you or your household experienced murder during the past 12 months?", as a share of all households, unspecified answers excluded. Stats SA gives the reference period as April 2025 to March 2026. The survey's "murder" includes culpable homicide. Covers private households and workers' hostels, not institutions. Published to one decimal; Stats SA estimates 17,465 of about 20.3 million households. Sheet "Table 5 household crime 12mths" of the key indicators file and Table 5 and Table 17 of the report (https://www.statssa.gov.za/publications/P0341/P03412026.pdf). Stats SA estimates 19,000 murder incidents and says 94.0% of these households reported them to the police.	3 of 5 checks Yes: Method and definitions published. The report's technical notes give the sample design, questionnaire, weighting and estimation. Yes: Official statistics or peer-reviewed. Official household survey by Statistics South Africa. Yes: Full count or representative sample. Stratified two-stage probability sample from Stats SA's Master Sample; national response rate 85.57%. No: Raw data downloadable. Key indicator tables are an Excel download; record-level 2025/26 data are not yet on DataFirst, whose latest GPSJS file is 2024/25. No: No known break or quality problem. Few sample households reported a murder: the 95% interval runs from about 6,900 to 28,000 households, and Stats SA could not publish breakdowns of acceptable quality.	100%

Households that experienced a murder, 2025/26, by source · % of households

Show the data as a table

Year	Stats SA, Governance, Public Safety and Justice Survey (Victims of Crime, P0341)
2018	10
2019	14
2020	11
2021	20
2022	16
2023	22
2024	14
2025	17

Why the numbers differ

SAPS counts murder victims recorded in police case dockets, classified under criminal law; Stats SA codes the underlying cause written on death notification forms, so a death counts as assault only if the form shows that another person caused it.
SAPS reports by financial year (April to March) and counts a crime when it is registered; Stats SA and WHO count deaths by calendar year of death. UNODC lists SAPS's April-March totals under the year they start up to 2021 and marks later values "NP Adjusted", so its 2024 count (25,423) is not SAPS's 2024/25 total (24,692). (source, retrieved 2026-10-26)
Stats SA codes non-natural deaths whose form gives no manner of death as "other external causes of accidental injury" or "undetermined intent": in 2024 these held 49,363 deaths against 7,486 coded as assault. They include 20,774 deaths from an unspecified factor (X58-X59) and 12,377 from inanimate mechanical forces (W20-W49), the group with unspecified firearm discharge and contact with a knife (Stats SA, Tables 4.11 and 4.12). (source, retrieved 2026-10-26)
SAPS's own release gives 39.1 per 100,000 for 2024/25, dividing by 63,115,905 people; this page divides the SAPS, Stats SA and WHO counts by Stats SA's newer estimate of 61,983,505 for 30 June 2024, which gives 39.84. (source, retrieved 2026-10-26)
The household survey counts households, not victims, includes culpable homicide and rests on few sample households (95% interval about 6,900 to 28,000 households in 2025/26); Stats SA says 94.0% of these households reported the murder to the police. (source, retrieved 2026-10-26)

Common mistakes

Reading Stats SA's assault deaths as the number of murders: Stats SA warns that most non-natural deaths in its data have no adequately classified cause. (source, retrieved 2026-10-26)
Comparing SAPS's April-March figures with calendar-year figures from other countries without noting the different period.

Context

Considered, not used

SAPS quarterly releases for 2025/26: SAPS had not published annual figures for April 2025 to March 2026 by 26 October 2026, so the page uses 2024/25. Its fourth-quarter release shows 5,181 murders in January-March 2026, 9.5% fewer than the 5,727 a year earlier. (source, retrieved 2026-10-26)
SAMRC Injury Mortality Survey as a data line: Its latest national survey covers 2017, so it is shown as context rather than as a current source. (source, retrieved 2026-10-26)
GPSJS five-year measure of household murder: 0.3% of households experienced murder in the five years before the 2025/26 survey; the page uses the 12-month measure because it matches a one-year rate. (source, retrieved 2026-10-26)

Use this data

Download every value on this page (CSV). Scores useformula version 1.0.

Cite as: RealStats, “What is South Africa's murder rate?”, updated 2026-10-26, with the original sources listed on the page.

Report an error on this page

What is South Africa's murder rate?

Murder rate in South Africa, 2024/25

Score breakdown

How much each source counts

Sources

Murders recorded by SAPS, 2024/25

Score breakdown

How much each source counts

Sources

Households that experienced a murder, 2025/26

Score breakdown

How much each source counts

Sources

Why the numbers differ

Common mistakes