A WATCHDOG WITHOUT TEETH ? The Independent Complaints Directorate

When members of the ICD approached the Institute for Security Studies (ISS) during 2006 they were particularly concerned that resource shortages, deficiencies within its legal mandate and procedural guidelines, as well as poor ICD/police relations in some areas were having a negative impact on the ICD’s ability to achieve its constitutional objective. This objective is to ensure that complaints in respect of offences and misconduct by members of the police are investigated in an effective and efficient manner. The researchers accordingly formulated the following objectives for this project: • To develop a clear understanding of the relationship between the ICD and the police • To examine the processes and procedures used by the ICD to present its recommendations relating to disciplinary or administrative action, as well as recommendations pertaining to criminal matters dealt with by the Directorate of Public Prosecutions • To establish the general level of compliance by the police with recommendations by the ICD • To evaluate the impact of recommendations by the ICD • To identify possible reasons for non-compliance • To make recommendations that would address areas of non-compliance


T
he Independent Complaints Directorate (ICD) is an independent institution established in terms of section 53 of the South African Police Service Act 1995(Act No 68 of 1995).In an effort to ensure that those police officers that do not uphold the rule of law are held accountable for their actions, the ICD makes recommendations as to how these officers should be dealt with, both departmentally and criminally.
When members of the ICD approached the Institute for Security Studies (ISS) during 2006 they were particularly concerned that resource shortages, deficiencies within its legal mandate and procedural guidelines, as well as poor ICD/police relations in some areas were having a negative impact on the ICD's ability to achieve its constitutional objective.This objective is to ensure that complaints in respect of offences and misconduct by members of the police are investigated in an effective and efficient manner.The researchers accordingly formulated the following objectives for this project: • To develop a clear understanding of the relationship between the ICD and the police • To examine the processes and procedures used by the ICD to present its recommendations relating to disciplinary or administrative action, as well as recommendations pertaining to criminal matters dealt with by the Directorate of Public Prosecutions • To establish the general level of compliance by the police with recommendations by the ICD • To evaluate the impact of recommendations by the ICD • To identify possible reasons for non-compliance • To make recommendations that would address areas of non-compliance The rest of this article will take a look at methods used in the study, the legislative mandate of the ICD, the findings and recommendations of the study and, finally, important lessons from international experience.

Methodology
The researchers adopted a combination of quantitative and qualitative research methods for this study.The quantitative component of the study consisted of an analysis of 573 case dockets of substantiated and closed cases in all nine provinces.These are for complaints lodged with the ICD since its inception in 1997.The research team focused only on CLASS I (death in police custody or as a result of police action), CLASS III (criminal offences committed by members of the police), and CLASS IV cases (less serious misconduct committed by members of the police).
The qualitative component of the study consisted primarily of structured interviews with the ICD provincial heads and SAPS provincial commissioners (or senior officials delegated by them).In total, eight ICD provincial heads and five SAPS provincial commissioners constituted the qualitative sample for the study.In addition to the structured interviews, the researchers conducted limited research to determine lessons learnt from international experience such as provided by the report of the Commonwealth Human Rights Initiative (CHRI) in 2005.

Findings and recommendations of the study
For purposes of clarity and conciseness the findings of this study are discussed under sub-headings in line with the research objectives.

Relationship between the ICD and the police
The relationship between the ICD and police management in the provinces is generally good, although the same can not be said about middle and lower level staff in the two organisations.There is evidence of poor personal relationships in some areas, mutual distrust, perceptions of an attitude of superiority by the police vis-à-vis the ICD, and a perceived reluctance by some police managers to act against members, notwithstanding ICD recommendations.The ICD appears to distrust the police to the extent that only about 30 per cent of the cases that are reported to it, are referred to the police for investigation or further investigation.These include cases of misconduct.It would probably have made more sense, especially in view of the growing numbers of the police and the resource limitations of the ICD, to refer more investigations to the police, closely monitored by the ICD.
In addition to these factors relations between the ICD and the police are also negatively influenced by the following conditions: • Based on some of the views expressed during the structured interviews, it would seem that the independence and credibility of the ICD is compromised by its location within the Department of Safety and Security, and the fact that it has to report to the Minister who is also the Minister responsible for the police (viz.conflict of interest).The concern here is that the Minister, in his endeavours to protect the image of the police, might not always be objective when he is required to take decisions on recommendations by the ICD that may negatively reflect on the police.• The lack of legal authority for the ICD to compel the police to report back on whether or not the ICD recommendations were implemented, or to provide documented reasons for non-implementation.• The general practice by the police not to respond to written communication from the ICD or to respond irregularly.• The absence of a prescribed coordinating mechanism between the ICD and the police, which leaves liaison between the two organisations to personal arrangements.• The ICD is frustrated by the police practice of waiting for the conclusion of a criminal trial against an accused member before they consider the possibility of departmental steps against that member.It is also not always certain who in the police is responsible for such decisions or for implementing ICD recommendations.

Processes and procedures used by the ICD
A number of shortcomings were found in the way that dockets were kept at most of the provincial offices of the ICD.In most of the dockets there was incomplete or missing information, and many case files were closed without any indication of the outcome of criminal or disciplinary proceedings.Part of the reason for this is that the criteria used for the closing of substantiated cases is not always clear and it appears that there is no uniform way in which this is done in the provinces.

ICD resources
The ICD is severely understaffed, especially when considering the rapidly increasing numbers of police members and the fact that the ICD investigates, on average, just over 5 000 cases per year.In addition, the effectiveness of investigators is marred by insufficient resources such as vehicles.A work-study report done by the Department of Public Service and Administration (DPSA) in 1997 allocated 535 posts to the ICD, including 339 posts for investigators.By 2007, ten years later, the ICD's staff complement stands at 247, less than half the recommended number.Some progress has been made with the setting up of two satellite offices in KwaZulu-Natal and the Eastern Cape, but even with these offices, the ICD simply does not have enough resources to carry out its mandate effectively.

The level of police compliance
The level of police compliance with ICD recommendations relating to Class I cases (deaths in police custody or as a result of police action) appears to be relatively high.This is probably the result of the legal obligation that the police have to inform the ICD of deaths in these instances.Furthermore, these deaths are matters for the courts and the police have virtually no influence on the outcome of these cases.
However, Class IV cases (misconduct) seem to have the highest frequency of non-compliance.Even in cases where the police do comply with ICD recommendations to take departmental action, a disciplinary panel consisting exclusively of police officials decides the guilt or innocence of the accused.This raises serious questions about the objectivity of these hearings.
Out of the 573 cases studied, 331 were referred by the ICD to the police for specific action.The majority of these (187) were referred to the police for investigations or for further investigations monitored by the ICD.In 131 referred cases specific recommendations were made in terms of departmental steps (disciplinary action).In some cases a single referral was for both criminal and departmental prosecutions.The remaining cases were either referred to the Director of Public Prosecutions (DPP) or, in spite of being closed as substantiated, were closed by the ICD because there were no real prospects of a successful prosecution.
In the process of communicating with the police relating to the 331 referrals, the ICD wrote 761 letters and received only 439 in reply (i.e.approximately 58 per cent).In 143 cases (referrals) there were no written replies from the police.On top of all the difficulties the police inaction creates for the ICD, it also makes it impossible for ICD staff to conclusively close their files in the affected cases.Based on this data it can only be concluded that the police largely ignore the ICD.This is an untenable situation and makes a farce of the oversight function of the ICD.

Possible reasons for non-compliance
There are a number of possible reasons for the non-compliance by the police.The primary reason seems to be in the wording of section 53 of the South African Police Service Act (1995), and in particular the absence of an obligation to report to the ICD in terms of referrals (particularly recommendations).In this regard the wording of section 18 of the Domestic Violence Act, 1998, serves as an example of how the police can be obligated to comply with recommendations.
Compliance does not mean that the recommendation should be acted upon in the same way as a directive, but that the police should at least report the outcome of their actions to the ICD.Should they decide to either take no action or to take alternative action, they should be compelled to inform the ICD both of the steps they have taken, as well as the reasons for their decision.
Other possible reasons have to do with the absence of a prescribed coordinating mechanism and other forms of formal interaction between the ICD and the police, e.g.formally appointed liaison officials.Formal contact and coordination should also help to solve many of the problems that relate to distrust between members of the ICD and the police.

Impact of ICD recommendations
For reasons discussed above it is difficult to assess the impact of the ICD in terms of its recommendations to the police.When there is only a 58 per cent response to the ICD's letters relating to specific referrals, and when another 143 referrals elicit no response at all from the police, it would appear that the impact is relatively low.

Lessons learnt from international experience
With the disparate range of civilian oversight mechanisms that exist for police agencies across the world, it is a challenge to find best practices that can be regarded as universally acceptable.In the United States, for example, there is an ongoing debate as to whether, and to what extent, the discretion of police chiefs in relation to disciplinary action against police officials should be subject to civilian oversight (Stone & Bobb 2002:1) should be empowered to report its findings and recommendations to the public, and to follow up on actions taken by the police chief in response to its recommendations.It should also be able to draw Parliament's attention to instances where police take no action.
From the research findings listed previously, it is obvious that civilian oversight of the police in South Africa falls visibly short of the above guidelines.

Recommendations
Implementing the following recommendations would, in our view, allow the ICD to perform its constitutional mandate effectively: • For purposes of independence and credibility the ICD should report directly to parliament (a special parliamentary committee or, alternatively, the Portfolio Committee for Safety and Security).Alternatively, the ICD should report to a minister who is not also the minister responsible for the police.

Conclusion
It can be concluded from this study that the ICD is not completely toothless -consider, for instance, the legal obligation on the police to report Class I incidents to the ICD, and the referral of criminal offences by the ICD to the Director of Public Prosecutions for further action.However, if it is to perform its constitutional mandate effectively it certainly needs more and sharper teeth, especially with regard to disciplinary action against offending police officials.
Provision should be made, either in the Act or in the regulations, for regular ICD inspections at police offices to determine the outcome and impact of ICD recommendations.In this regard a form of cooperation with the police's internal inspectorate should also be considered.•Care should be taken not to make ICD recommendations obligatory, but to ensure that the police provide acceptable reasons for noncompliance or alternative action if recommendations from the ICD are not implemented.• ICD policy should set out the conditions and circumstances for both taking over investigations from the police, and referring investigations to the police.This should include monitoring and reporting on such investigations.This may relieve the ICD's case load.• The ICD should provide for in-service training at police stations that should include how information is shared about the purpose, structure and functioning of the ICD.In addition, the ICD should be involved in the development of training curricula for the police and the provision of relevant formal training at police training colleges.