In 1998, the Centre for the Study of Violence and Reconciliation (CSVR) was commissioned to conduct research into the following aspects:

• The use and abuse of force by members of the SAPS

• The skill of policing and improving the management of the use of force in the SAPS

• An examination of deaths in police custody

This research was funded by the Royal Danish Government.

This memorandum has been compiled with the assistance of the CSVR and is intended to make known some of the outcomes of the research as well as suggest a strategy / strategies to take the process forward.

2.  INTRODUCTION

Since the ICD started operating in April 1997 statistics have for the first time become available on the real number of police custody and police action deaths (referred to in ss53(2)(b) of the South African Police Services Act as "death in police custody or as a result of police action"). On average over the first two years till March 1999 the ICD recorded 730 of these deaths per year (a total of 1459). In the further 10-month period, April 1999 to the end of January 2000 there were a further 578 of these deaths.

This means that overall in the 30 month period from April 1997 to the end of January 2000 there have been 2071 of these deaths**. In effect on average every month in South Africa during this period there have been 60 (59.93) of these deaths (approximately 20 in police custody and 40 police action deaths).

We recognise that the issue of high levels of crime and particularly of high levels of violent crime represents something of a crisis for this country and fully support the government in prioritising this issue and emphasising the urgency with which it needs to be addressed. We realise that there are a range of issues which need to be prioritised by government at the present stage in improving the effectiveness of the criminal justice system overall and particularly that of the SAPS. We are of the opinion that measures to reduce the high number of police custody and action deaths do not contradict this objective and need to be recognised as one of these priorities.

A strategy for reducing police custody and police action deaths should be combined with a programme which emphasises the safety of police officers, as well as to reduce the number of deaths of police officers. In fact we believe that the two issues are to some extent inter-related and some components of a strategy to prevent police custody and action deaths would also reinforce measures to reduce the number of deaths of police officers.

At the centre, we believe, of the values which are upheld in the Constitution, are the values of ubuntu, the idea of value for and respect for human life.

Support of a strategy to reduce the number of police custody and police action deaths, must form part of efforts to rebuild South Africa and to establish it as a country where human dignity and human life is truly valued. The South African Police Service itself, through improving its practice in preventing police custody and action deaths, can enhance its own credibility and itself play a role in building a South Africa in which human life is valued, respected and protected.

In this memorandum we therefore:

• Clarify the issues;
• Look at trends in custody and action deaths over the 34 month period April 1997 – January 1999;
• Examine the causes of these deaths;
• Look at the motivation for and value of a focus on prevention;
• Propose a way of taking the issue forward.

There is some overlap between the concepts of police custody deaths (deaths in police custody) and police action deaths (deaths as a result of police action) which sometimes leads to confusion when issues relating to these deaths are being discussed. In distinguishing between them the ICD therefore adopts the following approach:

• Police action deaths are deaths which occur as a result of the actions of the police prior to a person being taken into custody. The use of force (particularly shootings) is the major cause of police action deaths. Note that the police action which is the direct cause of deaths takes place prior to the person being taken into custody but in some cases the person may die much later, effectively while s/he is in custody;

• Police custody deaths are deaths which occur when a person is in police custody other than when their injuries were caused by the police prior to their being taken into custody. Some deaths in custody are the result of use of force by the police in custody but this is not the major cause of deaths in custody

The use of force by the police is the cause of over 90% of police action deaths and also of roughly 12% of deaths in custody. The use of force by the police, whether inside or outside of custody, may be lawful or unlawful.

Current evidence is that most police custody and police action deaths are not the result of unlawful actions by the police. In particular most custody deaths are not directly the result of actions by the police but include deaths which are the result of suicide, assaults by members of the public prior to custody and deaths by natural causes. The issue of the prevention of deaths however does not relate to unlawful deaths only. Potentially a significant number of both police custody and action deaths which are not linked to unlawful actions by the police as well as those which are linked to unlawful actions could be prevented by improvements in police practice.

Table 1: Number of custody and action deaths recorded by ICD, April 1997 – January 2000.

As reflected in the above table during the period April 1997 to January 2000 - the first two years and 10 months of operation of the ICD - there have been a total of 2071 police custody and action deaths. Of these:

• 28% were deaths in custody – the number of these deaths declined significantly in the second years of operation of the ICD. The figure of 174 deaths for the final 10 month period indicate a slight upward trend in comparison to the period April 1998 – March 1999 (where there were 95 every 6 months)

• 72% were deaths as a result of police action – these have fluctuated with the highest number being recorded during the period April 1998 – March 1999. The figures for the final 10 month period (404) are lower than those for any of the previous 6 month periods.

• During the first two years of operation of the ICD the overall number of these deaths did not change significantly (increase of 1%). The figure for the April – January 2000 period (578) represents a decline on the average ten-months figure for the previous two-year period (617). However this appears primarily related to relatively low overall figures for deaths during the months of April and May 1999. Since then the rate has returned to a monthly average of 61 suggesting that there may not be an overall downward trend in the number of deaths.

Deaths in custody account for roughly 28% of the total number of deaths. The major causes of custody deaths are the following:

• Police action during custody accounts for roughly 12% of deaths in custody – these deaths include apparently lawful actions – such as cases where a suspect is lawfully shot while trying to escape from custody - and obviously unlawful cases of police violence against persons in custody.

• More than 30% of custody deaths are caused by another person’s (non-police) actions pre-custody – these are people who are fatally injured by members of the public and are then taken into custody by the police. These people include some killed by members of the public acting lawfully (e.g. in self defence) against criminals but also include a significant number which are the result of unlawful vigilante actions.

• More than 30% of custody deaths are apparent cases of suicide - it appears that most of these are actual cases of suicide. There may be cases where people die as a result of assaults by the police with the death then being made to appear to be a suicide but, if there are such cases they do not constitute the majority of suicides.

• Roughly another 6% are also the result of the deceased persons own actions in the form of substance abuse. Most of these are people who die from alcohol related causes.

• Roughly another 16% are the result of the deceased’s medical condition/natural causes.

• Other causes of death include, for instance, persons killed by other persons in custody.

Deaths are sometimes the result of a combination of factors such as substance abuse in addition to the deceased’s medical condition, or a use of force complicated by the effects of substance abuse.

Deaths as a result of police action constitute roughly 72% of the total number of deaths. As is apparent from the following table virtually 90% of deaths as a result of police action are caused by shootings by the police. The remaining 11% also include a significant number which are the result of uses of force by the police (physical force not involving firearms). Effectively therefore uses of force account for in the region of 95% of deaths as a result of police action. Other causes include, for instance, vehicle accidents.

On the basis that 28% of deaths are in custody and 72% are police action deaths the causes of custody and action deaths relative to each 100 custody and action deaths are summarised in the right hand column of the following table.

Table 3: Causes of death as percentages of combined total of police custody and police action deaths (April 1997 – September 1999).

Effectively therefore of all police custody and action deaths:

• Over 70 % of custody and action deaths (71% ) are caused by uses of force by the police, making this by far the biggest cause of these deaths. These are primarily deaths as a result of police action prior to custody but also include some which are caused by uses of force in custody.

• Roughly 9% are deaths in custody as a result of injuries sustained prior to custody at the hands of members of the public.

• Roughly 11 are the result of the deceased ’s own actions-self harm (9% suicide, 2% substance abuse)

• Roughly 4% are the result of the deceased’s medical condition or natural causes.

• The remaining 5% or so are caused by a variety of factors including vehicle accidents and the actions of another person in custody.

There are some measures which support the prevention of these deaths such as present regulations or laws which regulate the treatment of persons in custody or the use of force by the police. Also the requirement that the ICD investigate all of these deaths serves to some extent to discourage unlawful actions by the police. While there are some measures in place, there is scope for the number of these deaths to be significantly reduced through improved police practice.

Presently the prevention of the deaths is not regarded as a priority issue by the SAPS or government. Instead there is a strong tendency for the issue of these deaths to be seen as the "ICD’s business" and not as a SAPS responsibility. Thus for instance it appears that the SAPS itself does not see it as necessary to monitor these deaths but sees this as being an ICD concern. Preferably the issue of prevention of these deaths should be seen as a joint responsibility of the Ministry, the SAPS and the ICD.

The prevention of police custody and police action deaths should be given greater priority by both government and the SAPS for the following reasons:

• There are currently a large number of these deaths. At the current rate by the end of March 2000 – i.e the end of the 3rd year of operation of the ICD - there will have been over 2000 of these deaths. We recognise that the number of these deaths is in some ways linked to high rates of crime and is partly a symptom of more generalised violence in our society, violence which members of the public and the police are often victim of. However a focus on the prevention of police custody and action deaths is not inconsistent with a concern for these other issues. In fact some measures which are intended to promote the prevention of these deaths will also contribute to addressing issues of the victimisation of police officers and violence in our society more generally.

• The possibility of deliberate and unlawful police actions exists in relation to many of these deaths and there is therefore an ongoing need for ICD involvement in the investigation of these deaths. However it appears that most police custody and police action deaths are not the result of deliberate criminality by the police. Nevertheless some deaths in all categories could be prevented if the issue is given greater priority and there is an improvement in SAPS practise relating to prevention.

• Potentially some components of a strategy for prevention of police custody and police action deaths involves simply giving greater emphasis to regulations which are already in place and a slight tightening up of some regulations – such as in relation to the issue of persons who are taken into custody. Improved prevention of custody and actions deaths can therefore potentially be largely using measures which are readily implementable rather than major policy changes.

• Addressing the issue of the prevention of deaths includes addressing the issue of the use of force by the police. This includes developing an appropriate legislative framework for the control of the use of force and improving the effectiveness of the police in relation to their use of force. This will mean assisting members of the police service to use force in an effective manner where it is necessary as well as providing training in how to avoid the use of force where it is not necessary. The framework which we are putting forward for improving the use of force by the police therefore recognises that police need to continue to have recourse to the capacity to use force and in fact should be assisted to use force more effectively in situations where this is called for including situations where there own lives are in jeopardy.

• A concern with reducing deaths in police custody and police action deaths is consistent with the principles of the Constitution including the right to life and the idea that persons are presumed innocent until proven guilty. While a concern with the reduction of deaths in custody is not inconsistent with the idea that uses of force which result in death are sometimes justified, the concern is also consistent with a recognition that people, including police officers, are fallible, and that even where they do not deliberately break the law, they may sometimes make mistakes which lead to unnecessary loss of life.

• While the causes of crime and violence in South Africa are complex one facet of the problems is a deeply rooted disregard for the value of human life. Part of the process of rebuilding the social fabric of South Africa as a nation involves building a greater respect and value for human life. These values need to be built within state institutions, including the SAPS, as well as in society more broadly. A central component of a focus on the reduction of police custody and police action deaths involves a focus on building the values of the SAPS, in relation to respect for life.

• The concern with the prevention of deaths in custody is based on the assumption that we should protect human life in so far as we can.

• Even if there is no criminality on the part of the police, there is often suspicion against the police when these deaths occur. Through emphasising that it is concerned with reducing these deaths and by demonstrating its concern in this regard through improved practice, the SAPS can minimise suspicion and thereby enhance its credibility

In developing the framework particular consideration is being given to the identification of measures which are readily implementable, which build on existing mechanisms and procedures and policies, and which are consistent with effective policing. The framework for prevention will therefore hopefully contribute to enhancing the professionalism of the SAPS.

In particular it is also hoped that it will be possible to integrate measures to improve the prevention of custody and action deaths with measures intended to improve the safety of police officers.

The proposed framework will be discussed in consultation with representatives of the SAPS, the Ministry for Safety and Security and civil society at a workshop to be held in April 2000.