Submission to:
THE PORTFOLIO COMMITTEE FOR JUSTICE AND CONSTITUTIONAL DEVELOPMENT

In Response to: The Criminal Law (Sexual Offences) Amendment Bill
Government Gazette No. 25282 30 July 2003

Prepared by:
Lillian Artz
Institute of Criminology, Faculty of Law
University of Cape Town

Adv. Bronwyn Pithey
The Sexual Offences and Community Affairs Unit
National Directorate of Public Prosecutions

PART 1

SECTION 22:
National Director of Public Prosecutions and Police Investigations

Authors:
LILLIAN ARTZ
Criminologist, Institute of Criminology, Faculty of Law, University of Cape Town
[The author would like to acknowledge the contribution of Dee Smythe, particularly in relation to the sections relating to non-compliance and codes of good practice.]
BRONWYN PITHEY
Senior State Advocate, Sexual Offences and Community Affairs Unit, National Directorate of Public Prosecutions

The writers would like to commend the South African Law Reform Commission [SALRC] for introducing section 21 into the Sexual Offences Bill. The decision to include a provision in the Bill that allows for the decision as whether an investigation of a complaint of a sexual offence should be discontinued should rest with the National Director of Public Prosecutions is welcomed.

We also support the specific recommendations by the SALRC in their Final Report on Sexual Offences that removing the discretion of the police as to whether or not to proceed with an investigation, even when requested not to proceed is by the complainant, should also lie with the prosecuting authority. We also support the recommendation that victims should be informed victim of her/his right to ask the DPP to review decision not to proceed with investigation.

This submission will provide the Justice Portfolio Committee with empirical research to support the inclusion of this provision and will attempt to address the concerns of the Committee regarding the constitutionality of such a provision. It is our understanding that the Committee, during the briefing by the South African Law Reform Commission on 6 August 2003, expressed its concern that section 21 could be regarded as unconstitutional as crime investigation is the preserve of the police.

We do not believe that Section 21 is unconstitutional with regard to crime investigation being the preserve of the police. It is submitted that it is imperative that the prosecution is involved at all stages of an investigation in relation to decisions that need to be made regarding prosecution. The criterion that is used to decide on whether to close a docket or to continue with an investigation, and possibly a prosecution, is whether there is a reasonable prospect for a successful prosecution. If this criterion is to be put into effect, prosecutors must be placed in a position where they have knowledge of the contents of a docket, even where no arrest has been affected and the docket has not been referred to court.

Research conducted by the Institute of Criminology at the University of Cape Town, commissioned by the SOCA Unit of the NDPP, found that a significant proportion of dockets (17%), where no arrest was affected, were filed by the police. In the ‘normal’ course of events, these dockets would never have been seen by the prosecution prior to being filed. Upon examination of these dockets it was determined, by the prosecuting authority, that there were a number of these cases that would yield a reasonable prospect of successful prosecution if further investigations were pursued.

Common practice in criminal courts is the continued interaction between prosecution and investigating officers, where an arrest has been affected. Instructions are given on a daily basis by prosecutors to police for further investigation. At this juncture, no docket may be closed by a police officer and this decision lies with the prosecutor. This system is important as it focuses and limits the investigation to evidence that is necessary for a successful prosecution as opposed to peripheral, non-essential evidence. The partnership between police and prosecution at this stage of the investigation, in theory at least, provides for a useful model inter-department co-operation and improves the ‘prosecutability’ of sexual offences cases by combining investigative and prosecutorial skills and knowledge.

The Thuthuzela Project, a multi-disciplinary, prosecutor-guided model for criminal justice – which will be described in more detail later in this submission – has identified large numbers of reported sexual offences cases that were not resulting in arrest. In response to this, a system was put in place where all non-arrest dockets had to be brought to the prosecutors for a decision whether to continue with the investigation or not. This system has yielded interesting results. Firstly, it has forged effective working relationships between investigating officers and prosecutors in that both parties consider the decision as whether to continue investigations a joint decision. Prosecutorial interest and guidance in the investigation of sexual offences has also encouraged investigating officers to pursue cases, resulting in increased arrests and subsequent prosecutions, in cases where this would not have previously occurred. Investigating officers have also indicated that joint responsibility, together with the prosecution, of not proceeding with an investigation has both improved their confidence in decision-making and has allowed them to dispose of a number of dockets under investigation that require no further investigation. This system has been embraced by police management who believe that it assists them in ensuring that their officers are accountable to other parts of the system apart from the direct managing structures of the SAPS.

It is submitted that this system does not in any way infringe upon the powers, duties and functions of either the SAPS or the NPA. It is acknowledged that both structures have specific and separate roles but, belonging to the criminal justice cluster, maintain the same objectives. By working together, from the outset, this can only enhance the quality of investigation and prosecution, rather than undermine the independence of both. The functions of both the police and the prosecution should not be seen as mutually exclusive. It should also be noted that section 24(1)(c) of the National Prosecution Authority Act (32 of 1998) provides that Directors and Deputy Directors of Public Prosecution has the power to … supervise, direct, and coordinate specific investigations. It is submitted here that the supervision, direction and coordination of investigations may include police investigations.

The Department of Justice statistics on sexual offences reveal that of more than 54,000 cases of rape reported in 1998, less than 7% were prosecuted. In 2002 an article was published in the Mail & Guardian that furnished the findings of a report commissioned by the National Directorate of Public Prosecutions. The study revealed that a large numbers of cases are still being withdrawn (43%) at the both pre-trail (investigation) and trial stages, despite the fact that police investigators are under instructions not to do so.

Another, well-documented study on attrition rates in rape cases was CietAfrica’s study (cited in s.8.3.3 of the South African Law Reform Commissions Discussion Paper 102) on the attrition rates between reported rapes and ‘founded’ (recorded) cases. It produced alarming results and served as sufficient empirical evidence to support proposals attempting to close the gap between reported and founded (investigated) cases. It found that for every 394 women raped, 272 (69%) are reported to the police. Of the 272 cases reported, 17 (6%) become cases, where the docket is opened for investigation. Of these 17 cases, 1 docket "gets lost", 5 are referred to court by the police and only 1 gets convicted. This means that only 1 in 272 reported rape cases (or just over 3%) result in conviction.

The study also found that the key attrition points included the victim’s decision to report the rape, the likelihood of arresting the accused, the scope of the investigation, the dismissal of the case by the prosecutor, and acquitted at trial. While the new definition of rape may reduce attrition rates to some extent, the proposed Bill must reduce police discretion in discontinuing the investigation, in more substantive terms.

It appears from these studies that a high proportion of sexual offence cases are getting ‘lost’ at the early stages of the criminal investigation. This may be due to the police designating cases as ‘false reports’ or as ‘withdrawals by the victim’. Internationally – and increasingly in South Africa – research is also beginning to reveal that the two most important factors influencing outcome of a rape trial are the evidence of physical injury and admission to offence by the perpetrator. Adler’s (1987) research shows that there are six primary factors to predict successful prosecutions: whether the victim was sexually ‘inexperienced’; whether she was ‘respectable’; the absence of any consensual contact before the assault; the presence of resistance and injury; early complaint; and the assailant being a stranger (in Kelley 2002).

If rape law reform is to be effective it must be vigorous in addressing the basic procedural issues such as the case processing and police investigations. It is imperative to ensure that adequate procedures and practices are put in place to govern these aspects. We believe that this is best achieved through the imposition by the legislature of positive duties. The Domestic Violence Act provides both a model and precedent for this approach.

Our experience with monitoring the Domestic Violence Act has also shown that the promise of good legislation is often not borne out in practice. Over a three-year period, our monitoring research of the Domestic Violence Act continually found that a major blockage in terms of delivery lies with the SAPS. A number of factors were implicated, including low resource levels, heavy caseloads, poor training and the tenacity of myths around violence within the home. Prosecutors were implicated to a lesser extent, because domestic violence matters only become criminal when protection orders are breached. Because of its criminal nature, rape cases will involve both these players from the outset, with resource and attitudinal issues remaining an acute problem.

The study found that the police based their decision on whether to arrest respondents (and investigate cases) on the following –

1. Only if there was obvious physical harm to the complainant or if the complainant was in imminent danger of further harm.
2. One the basis of personal assessment of the situation, including "fairness" to the accused where there appears to be a lack of evidence. Gauging the extent to which the investigation of the case would impact on "caseloads" also featured prominently.
3. Because of the legal ramifications of not arresting or investigating the case (i.e. non-compliance with duties imposed by the Act). The fear of an investigation by the ICD and disciplinary action was a major impetus to pursue the arrest and/or investigation of a case of domestic violence.

Both investigating officers and prosecutors inevitably approach a rape complaint from a cost–benefit perspective that is ultimately focused on the ‘convictability’ of the case. That is, given the resources to hand will the time, energy and money spent on investigation and preparation for trial result in a realistic possibility of conviction. The content of the substantive law and procedural rules and the way in which this is interpreted by judicial officers must necessarily inform this exercise. The focus on the "convictability" of a case is understandable, considering the number of reported offences and the limited resources and capacity of the police fully investigate a matter. Convictability, however, has been used as the bedrock for decisions surrounding the investigation of cases, instead of a discretionary factor.

As a result of the traditional division of tasks between the role players involved in the management of rape cases, the services provided have been largely fragmented and inconsistent. In an attempt to improve the investigation and prosecution of rape cases as well as provide better services to rape victims, a multidisciplinary centre has been established in the Western Cape, the Thuthuzela Care Centre.

The Thuthuzela Care Centre model demonstrates the use of a centralised, multi-disciplinary team comprised of police investigators, medical personnel, social workers, prosecutors and community volunteers. The aims of the project include the facilitation of skilled professional delivery of services to rape victims by police, medical personnel and prosecutors, joint investigation of cases by police and prosecutors, the development of accurate data collection tools, and the building of better cooperation and communication between rape victims and the justice system.

The project aims to demonstrate that a team approach to the investigation of rape cases can improve the treatment of rape survivors. This approach also streamlines and accelerates investigations and prosecutions, and provides concrete, reliable information to communities about this serious crime. The project hopes to inspire confidence in the criminal justice system among victims, as well as those structures working on their behalf.

The team managing the project, headed by the Office of the National Director of Public Prosecutions, with the assistance of the Bureau of Justice Assistance, focuses on the improved investigation of rape cases, the care of rape victims, and the creation of an accurate system and database designed to monitor and evaluate data in order to develop integrated strategies. The team works simultaneously on improving the following six components: (1) initial reception of victims at the police station and transport to the Centre; (2) medical examination, collection and storage of medical forensic evidence; (3) police statement taking from complainants; (4) better care of and communication with victims; (5) joint investigations between police and prosecutors, and (6) improved data collection and analysis.

1. Number of Cases Reported

For the period of January 2001 to the end of May 2003 a total of 1334 cases were recorded on the Thuthuzela database. Of these 1259 (94%) were drawn from the three participating police stations, Khayelitsha, Mannenberg and Gugulethu. Thirty-two cases (2.5%) were CPU cases and 43 (3.5%) were drawn from other police stations.

During 2002 the NPA conducted an audit of cases that had been withdrawn ('filed') by the SAPS without there having been prosecutorial participation in the decision. Approximately 300 cases were assessed by the NPA prosecutors to be potentially 'prosecutable' and these were returned to the relevant police stations with instructions for further investigation.

The Thuthuzela Project introduced the practice of involving prosecutors in management of rape cases from an early stage. It is hoped that through this involvement prosecutors will be able to guide the investigation of the rape complaint in a way that increases the ultimate likelihood of a successful conviction. While the prosecutors interviewed in this study approved of this innovation, they were vary wary of the terminology used, emphasising that such involvement should not be seen as amounting to a prosecutor-led or prosecutor-driven investigation. In other words, they saw themselves as partnering the police in, and in this way adding value to, the investigation.

However, there is a fine line between 'prosecutor-guided investigations' and 'prosecutor-driven investigations', which essentially amount to 'prosecutorial investigations'. Prosecutors suggested that there seemed to be an increasing tendency amongst certain investigation officers not to do their investigations thoroughly, as they knew that they could rely on the prosecutor to issue further instructions where more detail was required. That said, some investigating officers appear not to follow these instructions at all, taking the attitude that 'once the complainant's statement is taken their job is over'. This manifests, particularly, in the investigating officers being uncontactable for long periods of time, which means that the court staff spend inordinate amounts of time trying to establish communication, waiting for files to be returned and issuing further instructions. With some investigating officers very detailed instructions were required, while others, it was acknowledged, were strong investigators and used their own initiative. Feeding into the communication difficulties was the fact that there seems to be a high turn-over of staff at the participating police stations. Prosecutors are not informed, as a matter of course, when a Thuthuzela investigating officer is transferred or on leave and formal introductions are not done when a new investigating officer is appointed. This results in a troubling lack of continuity.

In general, however, the sense was that better investigations were being done, despite communication problems and variability amongst investigating officers. Prosecutors also felt that their early involvement had been a factor in reducing delays and postponements and that they were better prepared for trial because they had been on top of the case from an early stage.

In order for a case to be categorised as 'undetected', the SAPS standing order on closing of dockets (SO(G)325) specifies the investigation should have failed to disclose the identity of the offender, although an offence had been committed. According to the Thuthuzela database the percentages of cases filed as undetected, per police station, are as follows:

• Gugulethu: 33%
• Khayelitsha: 31%
• Mannenberg: 4%

Similarly high levels of cases filed as 'undetected' at police level are reflected in the 2002 Thuthuzela audit, which records that at Gugulethu out of 53 cases withdrawn at this level 50 were withdrawn because the suspect was undetected and only 3 at the complainant's request. At Khayelitsha out of 163 cases withdrawn by the SAPS 99 resulted from a failure to detect the perpetrator, while 49 were at the request of the complainant. In 21 of the 163 cases the complainant had 'disappeared' and in only 4 cases was there insufficient evidence. Only at Manenberg are the figures of a reasonable level with 1 out of 6 cases being filed in this manner. This is however balanced by the relatively high number of cases nollied at this station, meaning that rather than withdraw the matter at station level, the investigating officers are consulting with the prosecution prior to closing a case.

The SAPS standing orders for closing of dockets allows for a docket to be closed as 'undetected – complainant not traced' where a complainant cannot be found after reporting the matter. According to the standing orders this manner of closing should always be substantiated by witnesses (e.g. neighbours). There is a sense that cases are also erroneously filed under this category where the case is withdrawn at the request of the complainant. In terms of SO 325 an investigating officer may only withdraw a case of 'no consequence' upon an affidavit from the complainant requesting withdrawal. Given that rape is a serious offence it is arguable that no cases may be withdrawn at station level on this basis. Serious cases require prosecutorial approval before withdrawal in this manner.

The Thuthuzela database shows that dockets were closed on this basis, during the period January 2001 to May 2003, in the following percentage of cases:

• Gugulethu: 16%
• Khayelitsha: 3%
• Mannenberg: 7%

It is unclear from statistics provided by the SAPS how many cases were withdrawn at station level because the complainant could not be found or chose not to continue with the matter. However, as indicated above, the 2002 Thuthuzela audit reflects that only at Khayelitsha did this occur in a substantial number of the cases: that is, 21 out of 163 (13%) cases withdrawn because the complainant could not be found and 49 out of 163 (30%) cases withdrawn at the complainant's request.

A prosecutor may decline to prosecute an alleged offence where he/she does not believe that there is a reasonable chance of instituting a successful prosecution. It is a central tenet of the Thuthuzela Project that the SOCA prosecutors should be involved in any decision as to whether to continue with or 'file' a matter. As such, we would expect to see a relatively high number of cases 'nollied' (with a concomitantly lower number of cases withdrawn at station level). This is only true for Manenberg, where we see a remarkable 61% of all cases being nollied. At this station a total of 11% of cases were 'filed' (either as 'undetected' or 'complainant'). In contrast, at Khayelitsha 10% of cases were nollied and 34% filed for the same period. At Gugulethu 16% of cases were nollied, with 49% of cases finalised at that station actually reported as having been filed. The Manenberg statistics are consistent with reports from the SOCA prosecutors that they are being consulted before a case from that station is closed and should be read with the low number of cases being closed at station level. From the available data reasons for nollied cases reflected in the 2002 Thuthuzela audit are as follows:

Bail applications are heard in the District Court and, for the most part, cases are withdrawn at this level for further investigation. For the period from January 2001 to May 2003 this occurred in 16% of the Gugulethu cases and 8% of those from Manenberg. However, at Khayelitsha 32% of cases were withdrawn at District Court level. This is perhaps reflective of the fact that the SOCA prosecutors have less input into the Khayelitsha cases because of the jurisdictional problems relating to that station. The fact that Khayelitsha falls within a different magisterial district means that bail hearings are conducted at Khayelitsha District Court and that the SOCA prosecutors have little control over the cases at that stage. The process of transferring dockets from Khayelitsha District Court to WSOC has proved difficult to streamline. This means that cases that police statistics show as having been 'referred to court' have often in fact been withdrawn, for one reason or another, without notification to the Thuthuzela administrator. Difficulties in setting up adequate communication systems with Khayelitsha District Court has also meant that the Thuthuzela administrator is not apprised of other developments (such as finalisation of bail) and able to facilitate the transfer of the matter to WSOC. Respondents were all of the opinion that Khayelitsha should either be excluded from the Project or that provision needs to be made for all Thuthuzela bail hearings to occur at Wynberg Court.

Over the period from January 2001 to May 2003 12% of the cases reported to Gugulethu Police Station were finalised in the Regional Court. From Manenberg station, 17% of cases were finalised and in Khayelitsha 8% were finalised. In terms of withdrawn cases, is unclear how many cases were in fact withdrawn in terms of s6(a) of the CPA and whether any of the cases recorded under this category were cases of stopped prosecutions under s 6(b). It would appear that the cases from the Thuthuzela feeder stations have a greater likelihood of being finalised (and in a shorter period of time) than cases reported at other stations.

9. Quality of Investigations By The South African Police Service

The respondents of this study also had a considerable amount to say about the quality of investigations by the police. All respondents were particularly concerned about the accessibility of the police at all stages of the criminal justice process. The consensus of the respondents was that the police are very difficult to get a hold of – they either had no cell phones, their phones were switched off or they were on leave and did not provide the names and contact details of the investigating officers charged with taking over their cases. There was an understanding by the respondents that investigating officers are over-worked and have high case loads, but that some sort of system should be established to ensure that prosecutors, case managers and co-oordinators could more easily track investigating officers down when necessary. Concerns were also raised surrounding the accessibility of police when the accused was arrested and when Thuthuzela staff needed information regarding the status of case after instructions for "further investigation" had been issued to the police. The Khayelitsha police station was repeatedly indicted for their inaccessibility at every stage of the Thuthuzela process.

Statement-taking by the police was also regarded as problematic – it was argued by the respondents that the general quality of statements taken by police were very poor in that:

• the contents of the statements were limited;
• some police do not know the proscribed legal elements of what constitutes a sexual offence and therefore do not know what to look for/ask for from a rape complainant;
• basic information is lacking in the docket about the offence and the complainant (such as where the complainant lives) and as a result the case manager sometimes has to track this information down;
• the statements contain information and details that are irrelevant; and
• the language used in the statements in poor (for instance, the poor translation from Xhosa to English and the inappropriate use of terms such as "he raped me" when the offence was an indecent assault).

The combination of these factors means that complainants may be signing statements that they do not understand or that a statement may appear "contradictory" to what a complainant testifies to in court. It was argued that the first statement taken by the police (the reporting officer) is essential and should at least contain the following details:

• The complainants name and residential address.
• The place where the incident took place, if this can be determined.
• When the incident took place.
• A brief description of the events (if the complainant is ready to discuss this).
• The name or a description of the suspect.
• The name or description of any witnesses.

In practice, the initial statement taken from the complainant must be followed up by a more in-depth statement after the medical examination of the complainant. However, in some cases, the name of the complainant isn’t even taken down by the reporting officer. It was argued that the first statement should contain sufficient evidence for arrest and constitutes the "framework" for the case. In terms of the follow-up statements by the investigating officers, the National Police Guidelines for Victims of Sexual Offences are clear and instructive in this regard. Prosecutors naturally have to instruct investigating officers for specific clarification from the original statement and sometimes get "additional" rather than "supplementary" statements from complainants. The practice was considered problematic by prosecutors because it left the door open for defence attorneys to attack the complainant in cross-examination with regard to the different between to two statements. It was therefore recommended that the police be ‘re-trained’ with regard to what to ask the complainant, how to ask it and what to observe/look for.

It was also reported that:

• Reservists continue to take statements from the complainant at the charge office.
• J88’s are sometimes not found in the dockets.
• There are sometimes no returns of service in the dockets (resulting in numerous postponements, part-heards or the case being struck off the roll).
• Prosecutors often end up doing the "follow-up" work with complainants after arrests.
• Vehicles are not designated to the police, so follow-up work is difficult for them.
• When phoned in connection with a specific docket, the police often say they do not have the docket (although it is in their possession or floating around at district court).
• Investigating officers sometimes do not arrange consultations as requested or say that the complainant does not want to come to a consultation with the prosecutor.
• Investigating officers sometimes do not pick complainants up from the TCC or send the complainant to the TCC with no crime kit.

The Thuthuzela concept creates a specialist enclave of prosecutors, who are, in theory, meant to guide the investigation of cases. Prosecutors argued that the model of prosecutor-guided investigations is ‘a good one’, but in practice, the model is only as strong as the people working in it. It was maintained, by both magistrates and prosecutors, that the decision to contract experienced prosecutors for the TCC Project has improved the overall quality of prosecutions of sexual offence cases in Wynberg Court. Magistrates, in particular, emphasised that complainants appear generally more confident when testifying in court and that the quality of investigations (contents of dockets) and the presentation of evidence has also improved since the inception of the Project.

Lower case loads mean that prosecutors have more time to consult with witnesses and to follow-up on cases where the police have not followed instructions for further information. Prosecutors were generally pleased with the protocols and systems put in place for the management of sexual offence cases, but made mention that ‘basic resources’ still hampered the effectiveness of these systems. When questioned about their relationship with the police, the responses were consistent with other all the other Thuthuzela staff: some police are dedicated, accessible and take initiative to ensure that cases are properly investigated, while others are difficult to get a hold of and consistently fail to provide the prosecution with good quality investigations.

Before the Thuthuzela ‘J’ Court came into operation, only 4 cases reported at the Thuthuzela Care Centre were finalised, in an 18-month period. Since the Thuthuzela J court was instituted (August 2002), 96 cases have been finalised. The 4 Wynberg Sexual Offences Courts have finalised 137 cases since January 2003. Additionally:

The writers reviewed 5 studies from the U.K. on attrition rates in sexual assault cases that were conducted between 1978 and 2002. The studies revealed that:

• Circulars from the Home Office for "no-criming" (filing) were not followed by the police.

• An average 20% of cases are nollied by prosecutors.
• 50% of cases that the prosecution proceeds with result in conviction. It was argued that this figure should be treated with caution as the number of perpetrators is less than the number of convictions (multiple convictions).

• 25% of sexual offences cases result in no-criming and 31% result in no further action by the police.
• 56% of reported cases fall out at police level.
• Of the reported cases that are "crimed" 6-11% result in conviction "of some kind" (only 5-6% result in a conviction of rape).

In comparison, the Thuthuzela statistics are as follows:

• There were 405 cases reported at Thuthuzela between Aug 02 and May 03.
• There were 71 convictions at the Thuthuzela court during the same period.
• The report-conviction rate of this court is therefore 17.5%

National Instructions are, by definition, intended to establish and maintain uniform standards of policing (s25(1)(b) of the South African Police Service Act 68 of 1995). As such they set out policy guidelines for processing and management of certain offences (see, for example, National Instruction 22 of 1998, Sexual Offences: Support to Victims and Crucial Aspects of the Investigation). We believe that making the SAPS policy explicit will ultimately serve to inculcate a systemic change in the approach taken at individual member and at station level to dealing with specific offences. They cannot serve to force or enforce immediate changes in practice. This is not least because the sanction for non-adherence to National Instructions is limited to internal disciplinary action. The South African Law Reform Commission’s [SALRC] assessment of compliance with NI 22/1998 is also reflective of poor compliance levels in practice. By imposing positive duties on the SAPS through legislation, the State’s expectations of the quality of service to be delivered by the SAPS is made explicit. Likewise the threat of sanction for non-compliance. Furthermore, legislation being far more accessible than an internal police document, citizens are better apprised of their rights in respect of police conduct. The seriousness with which individual members of the SAPS view such legislatively imposed positive duties is apparent from our research on the Implementation of the Domestic Violence Act. It is our view that National Instructions, Standing Orders and other regulations should act to clarify and expressly implement duties imposed by legislation.

The National Instructions for police (no 22/1998) on sexual offences ("Support to Victims and Crucial Aspects of the Investigation") has been in place since 1998. It claims to provide step-by-step guidelines on how to manage a victim and an investigation and to ensure that police follow these instructions. These instructions, however, have not been implemented adequately.

The SALRC also recognized that the National Instructions were not distributed widely to SAPS members or stations and even that "very few members of SAPS comply with or seem to be aware of its existence" (Discussion Paper 102 at. 3.2.6.1). The Commission also acknowledges the "unwritten rules" in the management and investigation of rape cases (s. 3.2.6.6). The instructions also do not contain guidelines relating to information sharing, consultation with and collection of evidence for the trial between the investigating officers and the prosecuting authority prior to the trial. Most importantly, it does not provide any guidance as to what constitutes an "unfounded" case or what to do when a victim wishes to withdrawal her case (s. 8.4.4).

The SALRC also acknowledges that despite the development of a number of national strategies and programmes, as well as provincial and regional protocols (s.2.5.1) no official uniform standardised procedures, guidelines or management protocols exist for dealing with victims of sexual offences or with sexual offenders (s.2.5.1). By the SALRC’s own account existing services are mostly fragmented and under-resourced (s.2.5.1). The Commission correctly concludes that ‘(t)here is therefore no guarantee that a victim of a sexual offence entering the system will be dealt with in terms of acceptable procedures or be protected from further harm’ (s. 2.5.1).

The recommendation by the Commission that the National Instruction No22/1998 be revisited and amended where necessary (s. 3.2.9.2) is therefore not sufficient to ensure the integrity of police interventions in rape cases.

We strongly support the views of the Commission (at 2.5.7) that ‘a national strategy which includes legislation that supports and enforces all aspects of a national framework on sexual offences, and which is accompanied by specific accountability and measuring mechanisms for enforcement, is vital to the effective management of sexual offence cases.’ (our emphasis)

Experience with the National Policy Guidelines for Victims of Sexual Offences shows clearly the need for more than good intentions as a basis for proper implementation and application of the necessary rules and procedures. We need to reconcile ourselves to the reality that critical duties to which we are strongly committed must be embodied in legislation and accompanied by strong enforcement mechanisms. Internal protocols and regulations can act to effect structural and systemic changes, where they are well conceived, by inculcating a specific approach to the work in question. We believe that if implemented over a sustained period such protocols have the promise of bringing about a change in work ethic. However, certain duties are so critical to the immediate management of and investigation of sexual offences cases and the crisis that we face in addressing this issue is so acute, that these duties demand the full weight of legislative backing and enforcement.

It is, we believe, important to consider whether the radical changes brought by the new Sexual Offences Act will impact on the nature of police investigations and the exercise of prosecutorial discretion. Despite changes to the law, will these role-players continue to filter out complainants on a basis which approximates the myths underpinning the very laws being reformed, for example, in requiring corroboration or in making subjective assessments about the sexual history of the complainant and, by extension, her credibility. The proposed Bill should therefore include clear and specific provisions in the Bill to improve the quality of case processing and the treatment of rape survivors.

In light of the above, it is strongly recommended that one of the central objectives of the Bill should be to develop clear and distinct procedures to shift the current methods of pre-trial processing of reported rape complaints. International studies have largely focused on the number of pre- and post-reform reported cases, arrests, prosecutions, convictions and sentencing, making the assumption that ‘effectiveness’ of legislation can (or should) be measured by an increase in numbers alone. Gender activists in South Africa know all to well that the effectiveness of the criminal justice system in processing sexual assault cases goes beyond the numbers game.

The inclusion of duties to assist the victim in the body of legislation is not unusual, and has been effective in ensuring that the police perform certain primary functions in the Domestic Violence Act. Other jurisdictions, such as Michigan State and New York have prescribed police duties to victims within the body of legislation. Imposing positive duties on the police should not place any additional burden on the South African Police Service, as these duties should form part of their current line function.

Imposing specific duties on police will not only reduce discretion but will improve the evidentiary standards of investigations and instil ‘quality control’ of the information imparted to the prosecutor.

We therefore recommend that that the following guidelines from the instructions be included as provisions within the Act, or at least, be set out in the Regulations of the Act:

1. Where a complainant reports a case of rape, the first reporting member of the SAPS must, except where compelling reasons exist:

We further recommend that, in order to clarify the issue of jurisdiction, a clause be included to the following effect:

It is further recommended that the Bill also include the following provisions, as embodied in the Domestic Violence Act (116 of 1998):

We also support the Commissions recommendation that the police should review procedures for recording and following up unfounded cases and cases in which the complainant requests to withdraw the matter.

It is recommended that the state develop criteria for the withdrawal of cases by complainants, or at least provide guidelines to establish why cases are withdrawn (i.e. threats, duress, socio-economic implications). The criteria should balance the rights of the complainant with the duty of the state to protect the complainant and the community at large. These criteria should be made a matter of public record.

International studies have found that rape law reforms undertaken in - what has been referred to - as a "piecemeal" fashion or incrementally over a period of time, are ineffective changing the reality of women’s experiences with the criminal justice system. Feminist legal scholars such as Henderson (1993) have also argued that the problem with rape law reform is the ‘law in action’. Where legislation is not clear and prescriptive, the impact of statutory reforms becomes largely symbolic, rather than an instrument for law enforcement. Futter et al (2001) concluded that by removing many of the barriers to rape case processing, rape law reform would lead to less unfounded cases and in an increase of arrests. Griffiths (1999) also argues that it is unlikely that legal reform will address attrition rates between reporting and prosecution if the investigation of rape cases is not addressed directly. The risk of poor case processing (as CIETafrica have shown) is that fewer "rapes" are investigated, and more cases of ‘less serious sexual assault’ are reported because of the lower evidentiary burdens attached to them.