Publications
Partners
This excerpt from A Citizen’s Guide to Crime Trends in South Africa has been published with permission from Jonathan Ball publishers.
About the book:
A Citizen’s Guide to Crime Trends in South Africa provides a basis on which to understand the statistics in a manner that is accessible to the general public. Each chapter challenges a set of oft-repeated assumptions about how bad crime is, where it occurs, and who its victims are.
It also demonstrates how and why crime statistics need to be matched with other forms of research, including criminal justice data, in order to produce a fuller account of what we are faced with.
The quality of the SAPS murder statistics
Police everywhere in the world have been accused of manipulating their crime figures, of ‘cooking the books’, for as long as they have collected and disseminated them. The SAPS is regularly accused of trying to make itself and the crime situation look better by wherever possible squeezing the numbers into a better shape.
In a lot of research and thinking about crime, most or even all the analytical weight is placed on the recorded figures of just one crime type: murder. There are a number of reasons why murder numbers are less subject than most crimes to such squeezing.
Although there is no legal requirement that ordinary people report most suspected crimes to the police, South Africa has a strict death registration system.
The Births and Deaths Registration Act requires that every death in the country and of a South African abroad be reported to an authorised party, and if such a party has reasonable doubt about whether the death was due to natural causes, the police must be informed.
Having been informed of an apparently unnatural death, the SAPS must record and proceed to investigate it under one of three categories: murder, which is intentional killing; culpable homicide, which is killing that is unlawful but unintentional or negligent (i.e. accidents); or inquest, which applies to deaths that require some investigation but are believed not to be due to another person (chiefly suicides).
The police decision as to what kind of docket to open is provisional. The final decision may only be made in court many years later, and – as the Oscar Pistorius case has driven into the heads of the world – even this may not clarify matters satisfactorily.
There have been isolated accusations that police may sometimes opt to record as culpable homicides or inquests what should really be murders.
The police may also intentionally or unintentionally fail to manage the dockets properly and to register them on the centralised system.
Luckily, we have a couple of ways in which we can check up on how many murders are slipping between the cracks and failing to make it into the SAPS figures.
The biggest single factor keeping the murder figures honest is, to put it crudely, the corpse. There are certainly circumstances under which dead bodies may not be found, reported to any official authority or correctly identified as murders. The SAPS website currently lists about 200 missing people. It may be that a considerable number of these and the many other unreported missing people have in fact been murdered, but they will be considered missing until more evidence is found or they are legally presumed dead.
Repeat killers in particular may be good at disguising their crimes, such that the only way to get information on their numbers is by asking the killers themselves.
Interviews with serial killers and contract killers are not unheard of, but they are rare and are constrained by issues especially of self-selection and reliability.
Although pinning down definitions is an ongoing problem, it has been said that South Africa has a large number of serial killers for its size.
Defined as occurring ‘when a suspect(s) murder two or more victims on at least two separate occasions and the motive for the homicides are not primarily for material gain nor to eliminate a witness in another matter’, the SAPS has identified 131 homicide series between 1953 and 2007, of which 74 percent were solved.
This represents a tiny proportion of the total known murders in this time. It may well be that a fair number of such murders are never discovered, but at least by something like the above definition, serial murders are generally understood to be rare, comprising less than 1 per cent of murders.
But as long as there is someone who knows about and reports a death, there should somewhere be a record of it.
Murder involves not just a criminal act but the end of a legal subject (official personhood) and the new presence of a distinctive object (the dead body), with the result that besides the police there are at least in theory two other relatively official sources on the number of murders: vital registration systems (which try to keep some official record of all the people within a country) and medical facilities.
Few low- and middle-income countries have either such forms of data of reasonable quality or at all, and those in South Africa prior to 1994 are considered extremely compromised for many of the same reasons as are those from criminal justice sources. They have become increasingly useful since then.
The quality of mortality data from vital registration systems depends on their coverage and completeness, the precision of their cause-of-death recording and the reliability of that attribution.
We’ve said that South Africa has quite a strict death registration system, which requires reporting of all deaths to the Department of Home Affairs or someone otherwise authorised to receive such reports.
The coverage and quality of the National Population Register has improved over the years, and it has most recently been estimated to be about 94 per cent complete (for adult deaths) as compared to census data.
Statistics South Africa, which collects all the completed death notification forms from the Department of Home Affairs, and cleans, categorises and analyses the data, has indicated that in 2014 deaths due to non-natural causes, or what are ‘external causes of morbidity and mortality’, comprised 10.5 per cent or 47 761 of the 453 360 total registered deaths.
Of these, about 11 per cent, or 5 314, were ascribed to assault, while the rest included accidents of various kinds (55 per cent), events of ‘undetermined intent’ (17 per cent), transport accidents (12 per cent) and intentional self-harm (1.2 per cent). Rather than rely on the official registration documents, it is also possible to go directly to mortuaries (or ‘medico-legal laboratories’) and get their data from death registers, which include basic information on all deaths, or from postmortem reports, which are required for all non-natural deaths.
Mortuary-based approaches have not yet overcome problems of funding, tangled institutional responsibility and the sheer volume of data spread across over a hundred locations countrywide to obtain full and continuous national coverage of the roughly half a million deaths in the country each year, but instead base their national estimates on samples and complicated statistical weightings.
These samples are seldom random, and may introduce their own problems for generalisability.
READ MORE: What if there were no whites in South Africa?
Follow Women24 on Twitter and like us on Facebook.
