Skip to content

Gun share of homicides in South Africa

–

% of homicides committed with a gun

No usable source No source gives a usable figure, so there is no score.

Each source by year: % of homicides committed with a gun
00.511.52201020122014201620182020
Show the data as a table
YearWHO Mortality Database, deaths by assault with a firearm
20100.7
20110.9
20121.4
20131.8
20141.9
20151.3
20160.9
20170.9
20180.4
20190.1
20200.1

All countriesAll of South Africa's figuresCompare

What this number counts
Counts
Share of intentional homicide victims killed with a firearm
Where
South Africa
When
2020 (one year, not an average)

No source gives a usable figure.

Where the number comes from (1 source)
SourceFigure and yearDefinitionQualityCounts for
WHO Mortality Database, deaths by assault with a firearmWorld Health Organization (WHO) · Death and health registersRetrieved 2026-10-25
0.12020
Registered deaths from assault by firearm discharge (ICD-10 X93-X95: 8) divided by all deaths from assault (X85-Y09: 5668), all ages and both sexes.

Codes used: ICD-10 3-character codes. Assault by unspecified means (Y09) stays in the denominator: 8.7% for this year. Years with fewer than 20 assault deaths, and years a country reports only the condensed list without a firearm split, are skipped. National rows only. Year rule: only WHO has data for this country, so its latest year is used.

Raw files: morticd10_part1-6.zip. The same file records 0 deaths from firearm discharge of undetermined intent (Y22-Y24) and 8019 from accidental firearm discharge (W32-W34) for this year; they are in neither count. Directness 0, not counted: SAMRC's 2nd Injury Mortality Survey (printed page 49) reports that Statistics South Africa recodes deaths of unknown intent to an unintentional category, which artificially inflates unintentional injury deaths, and notes the effect of this misclassification on gunshot wounds specifically, so the firearm share of registered assault deaths does not measure the firearm share of homicides (source: https://www.samrc.ac.za/sites/default/files/attachments/2022-08/The%202nd%20Injury%20Mortality%20Survey%20Report_Final.pdf, retrieved 2026-10-26).

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 national authorities.
  • Yes: Full count or representative sample. Civil registration counts every registered, medically certified death, not a sample.
  • Yes: Raw data downloadable. Raw death and population files are free downloads.
  • No: No known break or quality problem. Registered deaths from all causes for 2020 are 83% of the UN World Population Prospects estimate of all deaths; not counted, because SAMRC's 2nd Injury Mortality Survey (printed page 49) reports that Statistics South Africa recodes deaths of unknown intent to an unintentional category, which artificially inflates unintentional injury deaths, and notes the effect of this misclassification on gunshot wounds specifically, so the firearm share of registered assault deaths does not measure the firearm share of homicides (source: https://www.samrc.ac.za/sites/default/files/attachments/2022-08/The%202nd%20Injury%20Mortality%20Survey%20Report_Final.pdf, retrieved 2026-10-26).
Not counted
Who reports it, and since when

Which kinds of source measure this number, which years each one published, and where the reporting stops or has holes.

No source of these kinds: police and justice records, independent count.

    Years each source has a figure here
    '10'15'20WHO

    figure no figure here, inside its range no later figure recorded change

    Why the numbers differ
    • UNODC compiles police and criminal justice records of intentional homicide by mechanism, while the WHO Mortality Database counts death certificates coded as assault by firearm discharge (ICD-10 X93-X95) out of all assault deaths (X85-Y09).
    • The share depends on how many killings have no recorded mechanism. UNODC "unspecified means" and WHO code Y09 (assault by unspecified means) stay in the denominator, so where they are large the share is a lower bound and the two sources can differ for that reason alone.
    • UNODC splits "firearms or explosives" into firearms and other parts for most countries; where only the combined count exists, the page uses it and marks the line as not exactly the defined number.
    • Deaths from firearm discharge of undetermined intent (ICD-10 Y22-Y24) are not counted as assault in death records, though police may record some of them as homicides.
    • Police counts can include victims who died abroad or cases later reclassified, and death records follow the place of registration, so the two counts of victims are not the same set of people.
    Common mistakes
    • The share of homicides committed with a gun is not the same as the share of gun deaths that are homicides. In many countries most firearm deaths are suicides, which this page does not count.
    • A high share does not by itself mean a high homicide rate: a country with few homicides can have a high firearm share, and the reverse. See the homicide rate page for the level.
    Download, cite and report an error

    From the table What share of murders are committed with a gun? Updated 2026-10-27.

    Download every value in this table (CSV)

    Cite as: RealStats, “Gun share of homicides in South Africa”, updated 2026-10-27, with the original sources listed on the page.

    Report an error on this page

    Questions about South Africa