Poverty as a Ground of Indirect Discrimination in the Allocation of Police Resources – A Discussion of Social Justice Coalition v Minister of Police 2019 4 SA 82 (WCC)

The Promotion of Equality and Prevention of Unfair Discrimination Act 4 of 2000 (PEPUDA) prohibits indirect and direct unfair discrimination in terms of the grounds listed in the act (such as race, sex, and sexual orientation) as well as unlisted grounds (which are to be alleged and proven by an applicant). South African courts had also grappled with the specific issue of indirect unfair discrimination prior to the enactment of PEPUDA, where applicants could rely on the Constitution of the Republic of South Africa, 1996 directly. This is evident in cases such as Pretoria City Council v Walker 1998 2 SA 363 (CC) and S v Jordan 2002 6 SA 642 (CC). This contribution is an analysis of the pioneering judgment in Social Justice Coalition v Minister of Police 2019 4 SA 82 (WCC) (SCJ case), wherein a South African court for the first time recognised poverty as a ground of indirect discrimination under PEPUDA. This conclusion flows from the court's finding, based on expert evidence that the formula used to allocate police resources in the Western Cape unfairly discriminates against poor and Black people in an indirect manner. The analysis of the SCJ judgment will take place against the backdrop of the antidiscrimination framework under PEPUDA and direct constitutional litigation that predates PEPUDA. The underlying theme of intersectionality will also be discussed, as it was apparent from a reading of the SCJ case that grounds of discrimination often intersect with one another and disproportionately affect certain groups of people.


Introduction
The status of South Africa as a country marred by violence is uncontested. Cape Town, which is situated in the Western Cape and is the focus of this discussion, has recently been identified as having the eleventh highest murder rate in the world. 1 Overall, inhabitants of the country as a whole also do not feel safe. Research confirms that only 31% of South Africans feel safe, ranking South Africa amongst post-conflict areas such as Afghanistan (20%) and Liberia (40%). 2 One of the key factors affecting both the presence of and resistance against crime is the police force. The lack of an adequately staffed police force inevitably leads to an increase in crime. 3 In 2015 it was reported that up to 85% of police stations in the Western Cape were understaffed, 4 while in 2018 the National Commissioner of Police noted a national deficit of 62 000 police officers. 5 The Commission of Inquiry into Allegations of Police Inefficiency and a Breakdown in Relations between SAPS and the Community of Khayelitsha (the Khayelitsha Commission) also highlighted this issue in their 2014 report. 6 It is noteworthy that areas where the majority https://www.westerncape.gov.za/news/police%E2%80%99s-gangstrategy-not-stopping-scourge-gang-and-drug-related-crimes. 5 Pitt 2018 https://www.news24.com/SouthAfrica/News/crime-stats-call-for-morepolice-for-the-poor-as-murder-rate-climbs-20180911. 6 Khayelitsha Commission 2014 https://www.saferspaces.org.za/resources/entry/ towards-a-safer-khayelitsha-report-of-the-khayeitsha-commission-of-inquiry (Khayelitsha Commission Towards a Safer Khayelitsha). The Constitutional Court in Minister of Police v Premier of the Western Cape 2014 1 SA 1 (CC) para 4 summarised the complaints that led to the establishment of the Commission of Inquiry as follow: "The complaint contained statistics showing high and escalating crime rates, with particular concern over figures relating to homicides, assaults and sexual crimes. Various and serious inefficiencies in policing were claimed, including insufficient visible policing in the community, lack of witness protection, lack of coordination between the police and prosecuting services and poor treatment of victims of crimes. The complaint described the routine violation of the rights of the residents of Khayelitsha and highlighted the impact of high crime rates on residents, including children and people vulnerable to discrimination. It added that 'the [Khayelitsha] community has lost confidence in the ability of the police to protect them from crime, and to investigate crimes once they have occurred'. The civil society organisations concerned proposed that the premier appoint a commission of inquiry into the Police Service and Metro Police operating in Khayelitsha." [footnotes omitted] Also see Social Justice Coalition v Minister of Police 2019 4 SA 82 (WCC) fn 5.
DC VAN DER LINDE PER / PELJ 2020 (23) 3 of the inhabitants were living in poverty had a much weaker police presence than more affluent areas in Cape Town. The population of the latter had the benefits of a stronger economy and the areas had lower crime rates as well as disproportionately better staffed police stations. The Commission noted in this regard that … the residents of the poorest areas of Cape Town that bore the brunt of apartheid are still woefully under-policed twenty years into our new democracy and are often the police stations with the highest levels of serious contact crime. This pattern needs to change as a matter of urgency. 7 And also: One of the questions that has most troubled the Commission is how a system of human resource allocation that appears to be systematically biased against poor black communities could have survived twenty years into our postapartheid democracy. 8 The topic at hand is the litigation that ensued subsequent to the findings of the Khayelitsha Commissionspecifically the case of Social Justice Coalition v Minister of Police (hereafter the SCJ case). 9 This article will explore the finding of the Western Cape High Court (acting as an Equality Court) that the theoretical human resources requirement (THRR) formula utilised to determine the police allocation nationally and provincially unfairly indirectly discriminated against poor and Black people. This is the first time a South African court has explicitly found that poverty is a ground of unfair discrimination. In part 2 of this article, I contextualise the law relating to unfair discrimination in South Africa, and then engage in a discussion, in part 3, of the most pertinent facts and findings of the Equality Court. Subsequently, in parts 4 and 5, I endeavour to identify and comment on the concept of intersectionality due to the overlapping categories of race and poverty. Part 6 concludes this article.

Generally
The applicants in the SCJ case sought an order inter alia declaring that the "allocation of police human resources in the Western Cape unfairly discriminates against Black and poor people on the basis of race and poverty" and further Discrimination is prohibited both under the Constitution of the Republic of South Africa, 1996 (hereafter the Constitution) and in terms of legislation. The Constitution holds that: The state may not unfairly discriminate directly or indirectly against anyone on one or more grounds, including race, gender, sex, pregnancy, marital status, ethnic or social origin, colour, sexual orientation, age, disability, religion, conscience, belief, culture, language and birth. 11 These grounds are known as the so-called "listed grounds". The Constitution further enjoins the Legislature to enact legislation to prevent or prohibited unfair discrimination. 12 This legislation was enacted in the form of the Promotion of Equality and Prevention of Unfair Discrimination Act 4 of 2000 (PEPUDA) and places a general prohibition on unfair discrimination, stating that: "[n]either the State nor any person may unfairly discriminate against any person." 13 The same "listed grounds" are also contained in PEPUDA and are referred to as "prohibited grounds". 14 It is possible under PEPUDA (and previously under the Constitution directly) for an applicant to allege that unfair discrimination took place on any ground including grounds other than the listed or prohibited groundsa fact which will be discussed below.

Indirect discrimination
Although neither the Constitution nor PEPUDA specifically defines direct or indirect discrimination, both make provision for them. Direct discrimination requires no further elucidation, but indirect discrimination does. Indirect discrimination occurs when the differential treatment due to conduct, practices, or in this case, the application of legislation, seems innocent or neutral, but the impact of the differential treatment is discriminatory. 15  case contended that the different methods used to calculate monthly dues for municipal services (water and electricity) were discriminatory on the basis of race. Two methods were used: actual consumption and a uniform rate per household. 17 Consumption meters were used to determine actual consumption and were installed in the suburbs of "old Pretoria" (such as Constantia Park), where the respondents resided. It was common cause that the residents of old Pretoria were predominantly White. Due to the lack of consumption meters, the townships of Mamelodi and Atteridgeville were billed on the uniform ratewhich was a flat rate and therefore generally lower, as it was fixed regardless of usage. The population of these townships was Black. 18 The Constitutional Court ultimately agreed with the respondents and found that, although not directly or overtly discriminatory, the differentiation was indeed unfairly discriminatory, albeit indirectly. Langa DP (as he then was) held that: It is sufficient for the purposes of this judgment to say that this conduct which differentiated between the treatment of residents of townships which were historically black areas and whose residents are still overwhelmingly black, and residents in municipalities which were historically white areas and whose residents are still overwhelmingly white constituted indirect discrimination on the grounds of race. The fact that the differential treatment was made applicable to geographical areas rather than to persons of a particular race may mean that the discrimination was not direct, but it does not in my view alter the fact that in the circumstances of the present case it constituted discrimination, albeit indirect, on the grounds of race. It would be artificial to make a comparison between an area known to be overwhelmingly a 'black area' and another known to be overwhelmingly a 'white area', on the grounds of geography alone. The effect of apartheid laws was that race and geography were inextricably linked and the application of a geographical standard, although seemingly neutral, may in fact be racially discriminatory. In this case, its impact was clearly one which differentiated in substance between black residents and white residents. The fact that there may have been a few black residents in old Pretoria does not detract from this. 19 He made it quite clear that the differential rates were geographically related. However, racial groups were (and still are often) clustered in geographical areas due to the lingering impact of apartheid, which enforced racial segregation. 20 The differential rates therefore disfavoured a particular racial group, albeit through the application of neutral geographical differentiation.

18
Walker case paras 4-5. 19 Walker case para 32. 20 In particular, the Group Areas Act 41 of 1950. The geographic, racial divide was also glaringly apparent in the SCJ case which will be discussed further below. 21 An allegation of indirect discrimination was also made in S v Jordan. 22 This case concerned a multi-pronged challenge to the criminalisation of sex work, 23 but for the purposes of this article, only the aspect relating to unfair and indirect gender discrimination will be discussed. Before the promulgation of the See Part 3.2 below. Also see SCJ case para 90, where the Court stated that: "25 years into our democracy people, Black people in particular, still live under conditions which existed during the apartheid system of government. The dawn of democracy has not changed the lot of the people of Khayelitsha. They continue to live in informal settlements where the provisions of services are non-existent or at a minimum. This is more glaring where a comparison is made with the more affluent areas, mainly occupied by the privileged minority. Such a comparison brings to the fore the stark reality of abject poverty. The unfortunate reality is that the residences of Khayelitsha, who are predominantly Black, continue to receive inferior services, including services from the SAPS. The SAPS discriminates against this impoverished community by using a system of human resources allocation."

22
S v Jordan 2002 6 SA 642 (CC) (the Jordan case). 23 The terms "sex worker" and "sex work" shall be used in this contribution instead of "prostitute" and "prostitution" because the latter has been described as humiliating, degrading and hurtful. See Tenga and Wasserman 2017 https://genderjustice.org.za/publication/health-safety-guide-clients-sex-workers/; HRW 2019 https://www.hrw.org/sites/default/files/report_pdf/southafrica0819_ web_0.pdf. gender-neutral and was therefore applicable to both men and women, and the legislative intent was to outlaw commercial sex. Since the sex worker is the one engaged in the business of commercial sex and not the client, the prohibition is aimed at the sex worker (who is "likely to be a repeat offender") and not the client ("[who] may or may not be a repeat offender"). 28 It was further held that the client is subject to the same punishment as the principal offender under the RAA and that there is, in fact, no distinction; the client and the sex worker are equally liable, and both are subject to the same punishment. 29 The finding that there is no distinction between the client's punishment and that of the sex worker is patently false, as is evident from the fact that there were no cases against the clients at all. In practice, a conspirator (in this case the client) usually receives a more lenient sentence than the perpetrator (the sex worker). 30 This is due to the fact that the conduct of the conspirator is considered less harmful or reprehensible because his conduct has not manifested into a complete crime. 31 There is no prohibition against being charged with both conspiracy and the substantive crime, but one cannot be convicted of both. 32 Considering that the client can at most be found guilty of conspiracy, he could never receive equal punishment for what is essentially equally criminal conduct. This was a distasteful scenario because the client and sex worker engaged in the same conduct, but for the payment/receipt of a fee.
The majority in Jordan also rejected submissions regarding the prosecutorial and police practices where only the sex worker is prosecuted and held that that was not the issue under consideration. The issue was the potential unconstitutionality of the vexed provision of the Sexual Offences Act and that the prosecutorial practices were an issue pertaining to the application of the law and not to the law itself. 33 It was consequently found that the vexed provision was not unconstitutional. 34 This was an unfortunate outcome as the scenario in Jordan was a textbook instance of indirect discrimination. When considering the constitutionality of an impugned provision, it cannot be viewed in a legal vacuum devoid of its real-world application. The truth of the matter is that most sex workers are female (and, as will be illustrated below, women are most often the victims of poverty) 35 and, although the conspiratorial conduct of the male client is punishable under the RAA, such prosecution, on the facts before the court, had never occurred. This application therefore has a disproportionate and unfairly discriminatory effect. Whether or not the sex worker is the "supplier", as per Ngcobo's analogy, this does not ameliorate the fact that at the time of a particular transaction of sex work, both parties are engaging in the exact same conduct, save for the fact that one receives money and the other pays it. The event is not analogous to drug dealing, where there are wholly different acts involved in the unlawful conduct.
The minority held differently: In the present case, the stigma is prejudicial to women and runs along the fault lines of archetypal presuppositions about male and female behaviour, thereby fostering gender inequality. To the extent therefore that prostitutes are directly criminally liable in terms of s 20(1)(aA) while customers, if liable at all, are only indirectly criminally liable as accomplices or co-conspirators, the harmful social prejudices against women are reflected and reinforced.
Although the difference may on its face appear to be a difference of form, it is in our view a difference of substance that stems from and perpetuates gender stereotypes in a manner which causes discrimination. The inference is that the primary cause of the problem is not the man who creates the demand but the woman who responds to it: she is fallen, he is at best virile, at worst weak. Such discrimination, therefore, has the potential to impair the fundamental human dignity and personhood of women. 36 Both the minority and the majority referred to the differing societal views of the participants in sex work, where there is greater stigmaand most often harsher criminal sanctionon the sex worker. 37 This affronts the principle of fair labelling. The principle of fair labelling dictates that the sanction imposed on an offender must adequately reflect what he or she has done. 38 No fair labelling can be said to have been present under the old regime of the Sexual Offences Act in conjunction with the RAA.
Both these cases illustrate the intersectionality of poverty with other categories of deprivation. grappling with poverty. The intersectionality between poverty and other categories of discrimination will be discussed below. 39

Unlisted or unspecified prohibited grounds
As mentioned above, discrimination can occur on one or more of the listed or prohibited grounds. Under the Constitution, when dealing with a listed ground of discrimination, the unfairness is presumed and the onus would then shift to the respondent to show that the conduct, practice, law and the like was fair. 40 Similarly, under PEPUDA, if the complainant proved a prima facie case of discrimination, the respondent must either prove that the discrimination did not take place as the claimant has alleged 41 or that the vexed conduct is not based on one or more of the prohibited grounds. 42 If a court finds that the discrimination based on a prohibited ground did in fact take place, however, a presumption of unfairness arises and the onus shifts to the respondent to prove that the discrimination is fair. 43 If it is not a listed (or prohibited) ground, then it will be automatically unfair if one or more of the following conditions are met: 44 any other ground where discrimination based on that other ground-(i) causes or perpetuates systemic disadvantage; (ii) undermines human dignity; or (iii) adversely affects the equal enjoyment of a person's rights and freedoms in a serious manner that is comparable to discrimination on a ground in paragraph (a) […] 45 The discrimination will also not qualify as unfair if the respondent disproves it. 46 The following factors must be taken into account when a court determines whether a respondent has proven that the discrimination is fair: Section 1 of PEPUDA. Paragraph (a) here refers to the listed grounds as quoted above. 46 Section 13(2)(b)(ii) of PEPUDA.
PER / PELJ 2020 (23)  10 (c) whether the discrimination reasonably and justifiably differentiates between persons according to objectively determinable criteria, intrinsic to the activity concerned. 47 The nine factors mentioned in paragraph (b) are whether the discriminatory conduct either impairs or is likely to impair human dignity; 48 the impact or likely impact on the complainant; 49 whether the complainant suffers from patterns of disadvantage or belongs to such a group (his or her position in society); 50 the nature and extent of the discriminatory conduct; 51 whether it is of a systematic nature; 52 whether the discriminatory conduct has a legitimate purpose; 53 the availability of less restrictive or less disadvantageous means to achieve the same purpose; 54 whether and to what extent the alleged discrimination has achieved its purpose; 55 and whether (if at all) and to what extent the respondent has taken reasonable steps to address the disadvantages that arose due to one or several of the prohibited grounds 56 and to accommodate diversity. 57 The subsequent sections will deal with the most pertinent facts of the SCJ case, including the most relevant role players involved, the formula used to determine police distribution in South Africa, and the court's engagement with the issue of unfair discrimination.

Relevant role players
The applicants were public interest groups. The first applicant was the Social Justice Coalitiona non-profit and nongovernmental organisation. The second applicant, Equal Education, was also a nongovernmental organisation. The national and provincial police leadership is structured in a manner one could only describe as a bureaucratic matrix. The respondents were all governmental functionaries. 59 The first respondent was the Minister of Police. He was joined as a respondent as he is enjoined through the Constitution 60 to oversee policing throughout the jurisdiction of the Republic. The National Commissioner of Police was cited as the second respondent as he was enjoined through both legislation 61 and the Constitution 62 for "controlling and managing the police service".

Overview of the formula used to determine police resources
SAPS engages in two phases of human resource allocation. The first phase is a theoretical allocation and denotes the calculation "of the number of posts per level required to perform the duties associated with police stations." 66 This was said to indicate the ideal distribution of human resources, had there been no budgetary constraints. 67 Firstly, crime statistics from the police stations across the country are gathered. Without going into specificity, SAPS indicated that this information includes "a wide range of determinants". It includes, most significantly, the average incidence of all reported crime at every police station over a period of four yearswith the most recent year weighing the most in the formula. The subsequent formulas, weighting, and ratios are applied in order to determine the number of police officers required in order to satisfy the theoretical crime prevention requirements. 68 The evidence submitted to the court indicated that one police officer post is allocated for an average of twenty reported contact crimes per month and one post for every 25 property-related crimes. One post is further allocated 59 See the SCJ case paras 6-9. This baseline figure is then utilised in a demographic analysis comprising of 79 demographic determinants. These determinants are not crime statistics but are demographical determinants that may influence the incidence of crime in a particular area. SAPS further alleged that greater weighting is given to these factors in underdeveloped areas than in developed or affluent areas. 70 The population and geographical size of the area serviced by a particular police station, the unemployment rate and the size of the informal population are amongst the first determinants considered. Further determinants, and these are important for the purpose of this discussion, include the presence of commuters, and the presence of venues and facilities used for hosting sports, religious, and other recreational events, as well as whether there is a seasonal influx of people present in the particular area. These would primarily be holiday visitors. Geographical factors are also considered, such as mountains and rivers that may hamper the execution of the duties of the police. 71 Another important factor in the calculation is ostensible socio-economic considerations such as the lack of roads and streetlights and/or telecommunication and social degradation. The presence of informal housing is also considered, as this is often associated with a lack of access routes and street names. Lastly, the number of identified gangs is included under socio-economic considerations. Although there was no specific mention of gangs in this case, it was considered in the initial Report by the July 2019 in an attempt to stabilise gang hot spots due to more than 900 gang-related murders being committed in the first half of 2019 alone. 74 Another cluster of determinants which can also be described broadly as socio-economic factors are the areas where people assemble. These include malls, airports, overnight accommodation, places were liquor is sold and consumed (either lawfully or unlawfully), educational facilities, and whether there are outlets that are permitted to sell firearms in the area. 75 Lastly listed by SAPS were areas that the police were under a statutory obligation to police, such as farms and smallholdings, as well as abattoirs. 76 Experts are then consulted to ascertain how many investigations of a particular crime a theoretical detective would be able to conduct each month. This result subsequently leads to a theoretical detective requirement, where ratios are applied to particular forms of crime. For example, one detective is theoretically allocated for every murder and one for every ten robberies. 77 With the theoretical numbers in mind, underdeveloped areas are then favoured, according to SAPS. Factors such as the distances the police need to travel to reach collateral service in the investigative process are considered (for example, the distance between hospitals and forensic laboratories). 78 A contingency allowance is also factored in for police officers. This allowance includes meetings, procurement, interaction with other officers, and reporting for and being off duty, as well as recovery from fatigue and compulsory leave. 79 As stated above, the result of this morass of determinants, factors and considerations is the ideal allocation where budgetary constraints are not a consideration. The second phase, naturally, would involve the actual allocation of police posts, which is informed by the available budget. It was alleged that disadvantaged areas would, considering the determinants and 74 Haysom considerations, on average receive one post for every 2 500 members of the community instead of 1 per 5 000, as with non-disadvantaged areas. 80 The consideration of the THRR for all police stations in the country provided the national requirement for both the number of posts as well as the ranks. 81 After the national allocation, the provinces bear the responsibility of distributing the funded posts. SAPS indicated that THRR and other factors such as crime trends and patterns are given due consideration, but that the distribution is "a dynamic and flexible process". 82 From a reading of the case as a whole combined with the description of SAPS, particularly of their resource allocation, it follows that the underlying and explicit supposition is that the allocation, save for the purposeful weighing in favour of underprivileged areas, is entirely neutral and done with reference to objective and neutral determinants. As will be shown in the following sections, the determinants seem on the surface to be quite innocuous, but their impact on the allocation of police resources is anything but neutral.

The expert evidence by Jean Redpath 83
The expert witness initially testified at the Khayelitsha Commission about the skewed allocation of police resources in Khayelitsha and other areas predominantly occupied by Black people as a result of using the THRR. 84 Redpath made the crucial observation of the unfortunate cycle in areas such as Khayelitsha. The cycle starts with a significant underreporting of crimes in these areas, which would, in turn, lead to less police resources being allocated. In other words, areas where there are low levels of police resources would have low levels of reporting and therefore less additional resources would be allocated to them. Areas that already had an oversupply of resources would have better reporting and therefore would receive more resources due to the better reporting rates. 85 This is an inescapable cycle that benefits certain areas more than others as a result of the operation of the THRR. Redpath's calculations were represented as the number of police staff per 100 000 members of the population. The average number of police members was 283 for every 100 000 peoplewith the lowest incidence being in Harare (111 police members per 100 000 people) and the highest in Table Bay Harbour (2 636 police members per 100 000 people). Khayelitsha had an incidence of 190 police members for every 100 000 and Lingelethu-West 275 police members per 100 000 people. Areas like Khayelitsha, which also contained informal settlements and were subject to serious and violent crime, were also noted to be understaffed. 86 It was also pointed out that the budget could not accommodate the THRR requirements for police stationsonly 68% could be satisfied at every police station.
The THRR factors and determinants, albeit prima facie objectives, had anything but neutral outcomes. Redpath indicated that unlike formal areas (predominantly populated by non-Black and more affluent persons), only 15 of the factors weighed in favour of informal areas. Formal areas have a potential additional weighting of up to 205%, while informal areas have a potential additional weighting of 75%, according to Redpath. 87 For example, informal areas will have a lower incidence of incoming commuters, malls, airports, overnight accommodation, and a seasonal influx of people (holidaymakers). These factors are more frequently found in affluent areas and therefore skew police allocation in favour of those areas. Further, poor Black people do not inhabit these areas.
Poverty and informal housing were also considered to bolster the finding that police resourcing is disproportionate. 88 Redpath found, unsurprisingly, that a high incidence of piped water and electricity correlated with formal housing while a low incidence of these resources correlated with informal housing. She further submitted that historically, Black people tend to live in informal housing which, as was pointed out, had a lower level of service provision. And further, that lower levels of service provision also correlated significantly with lower levels of police resources. 89 The court was satisfied that Redpath's research proved the skewed allocation of police resources based on the grounds given.

How the court in SCJ dealt with the issue of unfair discrimination
The applicants relied on two grounds of discrimination, namely race and poverty. 91 Race, of course, is a prohibited ground under PEPUDA and unfairness was consequently presumed. Poverty, however, is an unspecified or unlisted ground. There had also never been an instance where a court has recognised poverty directly as a ground of discrimination. 92 The applicants therefore had to satisfy the court, as described above, that the unspecified ground of poverty has the effect of causing or perpetuating systemic disadvantage; that poverty undermines human dignity; or that poverty has the effect of negatively impacting on someone's "rights and freedoms in a serious manner that is comparable to discrimination on a ground in paragraph (a)." 93 The applicants submitted that, although there had never been an instance where poverty was expressly recognised as a ground of discrimination, there had been several instances where the Constitutional Court (before the promulgation of PEPUDA) had by relying on analogous grounds found in favour of applicants that unfair discrimination was present. 94 This alludes to the intersectional nature of discrimination which was pointed out in Harksen v Lane (hereafter the Harksen case): 95 There is often a complex relationship between these grounds. In some cases they relate to immutable biological attributes or characteristics, in some to the associational life of humans, in some to the intellectual, expressive and religious dimensions of humanity and in some cases to a combination of one or more of these features. The temptation to force them into neatly selfcontained categories should be resisted. 96 In this passage Goldstone J warns against an artificial categorisation of discrimination and points out that these grounds are often interrelated and can overlap. An unlisted category, therefore, should not be denied merely because of its own disadvantageous status as an unlisted ground.
Although there is no explicit ground of poverty in PEPUDA, section 34(1) specifically contains a directive regarding "HIV/AIDS, nationality, socioeconomic status and family responsibility status". The court consequently agreed with the applicants and for the first time in South Africa accepted poverty as a ground of unfair discrimination. 107

The state's response
The state's main challenges were aimed at the expert witness. It was contended that she was not a policing expert and that she had relied on outdated statistics without taking due cognisance of applicable variables, including the unworkable complexities that are synonymous with policing. Redpath was also accused of not considering budgetary constraints properly. The state submitted that in its process of allocating resources, SAPS cannot rely solely on population and crime statistics. 108 It further contended that the THRR in fact did not discriminate against poor and Black people as it was skewed in the favour of these categories of personsand 97 Section 1 of PEPUDA. None of these complaints, however, addressed the empirically proven discriminatory impact of the THRR. The actual validity of Redpath's analyses was not contested. Stated differently, the state did not present any empirical evidence showing that areas occupied by poor and Black people were not disadvantaged by the THRR but in fact had been benefitted. For the sake of the argument, even if the figures were outdated and even if Redpath was not the most apt expert, the state did not even attempt to prove that persons living in the relevant areas were receiving sufficient or aboveaverage police allocation.
The court rejected these superficial contentions and ultimately held that the THRR, though ostensibly operating on a neutral basis (founded in objective facts), had a discriminatory effect. 111 The state had therefore ultimately failed to discharge its onus to prove that there was either no discrimination at all or that the discrimination was fair. 112

Discussion of the SCJ judgment: the underlying theme of intersectionality
The court's decision in the SCJ case cannot be faulted. The applicants are especially commended for their vigorous pursuit of justice through a strong body of evidence. The jurisprudential impact that the case will have is undeniable. The matter of poverty as an indirect and intersectional form of unfair discrimination is deserving of some further discussion.

Grounds of discrimination often overlap. The Constitutional Court in
Harksen held that "[t]here is often a complex relationship between these grounds … and in some cases … a combination of one or more of these features." The court also warned not "to force them into neatly selfcontained categories …". 113 The effect of this complex relationship and overlap is that of an intersectional group of persons having a unique or particular disadvantage. 109 In terms of s 12(3) of the SAPS Act. American legal scholar Kimberlé Crenshaw brought the concept of intersectionality to prominence in 1989. 114 The central thesis of Crenshaw's paper and subsequent seminal works was that US antidiscrimination law tends to "treat race and gender as mutually exclusive categories of experience and analysis" 115 and the status quo actually erases Black women from the antidiscrimination dialogue "by limiting inquiry to the experiences of otherwise-privileged members of the group". 116 Stated simply, "people's disadvantage is [often] composed of multiple and interlocking systems of power." 117 American courts failed to recognise Black women as a separate and distinct category of a protected class, often holding that either Black men (of the protected class of race) or White women (of the protected class gender) were not discriminated againstand therefore they -Black womenwere not being discriminated against.
Crenshaw points to the judgment in DeGraffenreid v General Motors (hereafter the DeGraffenreid case) as an example of intersectionality. 118 The recession caused an economic downturn that necessitated certain retrenchments to be actioned. This in turn resulted in all Black women hired post-1970 being fired, based on the "last hired, first fired" approach (or "lastin-first-out" as it is known in South Africa). General Motors had started to hire Black women only after 1964. The effect was of course that this system indirectly impacted on Black women, who were already underrepresented in the senior work corps of the company. The court rejected the notion that Black women constituted a particular protected class and held that the plaintiffs should not be allowed "to create a new 'super-remedy' which would give them relief beyond what the drafters of the relevant statutes intended." 119 And further "[that] this lawsuit must be examined to see if it states a cause of action for race discrimination, sex discrimination, or alternatively either, but not a combination of both." 120 The court pointed out that General Motors had hired women prior to the promulgation of the Civil Rights Act of 1964 but neglected to take into account that no women of colour were hired in the period preceding the Civil Rights Act. 121  It was also suggested by the court that this case be joined to a case specifically dealing with race discrimination, but the plaintiffs rejected this suggestion since, as Crenshaw correctly points out, it would defeat the object of instituting proceedings based both on race and sex and denying their particular form of discrimination. 122 According to the Court: The legislative history surrounding Title VII does not indicate that the goal of the statute was to create a new classification of 'black women' who would have greater standing than, for example, a black male. The prospect of the creation of new classes of protected minorities, governed only by the mathematical principles of permutation and combination, clearly raises the prospect of opening the hackneyed Pandora's box. 123 The court was therefore reluctant to entertain any arguments relating to Black women as a class of persons specifically discriminated against. It even seemed to find it incomprehensible that Black women as a class of persons suffered a specific plight different from that of their colleagues of the same gender (or race).
In the Canadian case of Sparks v Dartmouth/Halifax County Regional Housing Authority (hereafter the Sparks case), 124 the Nova Scotia Court of Appeal had to consider whether differentiating between public and private sector housing legislation 125 constituted discrimination for the purposes of section 15(1) of the Canadian Charter of Rights and Freedoms. 126 Tenants who lived in private sector housing had "security of tenure". If they occupied premises for more than five years, they could not have their lease agreements terminated on short noticesuch as a period of one monthas was the case of the appellant. 127 A judicial officer had to be satisfied that a tenant in private housing was defaulting on one or more of his/her statutory obligations before short-term notice would be allowed. 128  case for persons living in private housing, however. In such cases the legislative protections could be circumvented.
The appellant, who lived in public housing, was a single, Black mother who had two children and received welfare. 129 Black women on welfare comprised a disproportionally large percentage of the tenants in public housing and on the waiting list for public housing. 130 The court found that the legislation had a discriminatory effect, albeit "neutral on its face". 131 However, discrimination here was not based on any of the listed categories under section 15(1) of the Charter. The court emphasised the fact that the legislation was neutral, but its impact was discriminatory on the basis of race and sex and therefore affected a large portion of the inhabitants of public housing. It was also pointed out that poverty was also experienced by these categories of persons and was present here, because low income was a prerequisite to qualify for public housing. Like the Court in SCJ, this court found that: The public housing tenants group as a whole is historically disadvantaged as a result of the combined effect of several personal characteristics listed in s. 15(1). As a result, they are a group analogous to those persons or groups specifically referred to by the characteristics set out in s. 15(1) of the Charter being characteristics that are most commonly the subject of discrimination. 132 Intersectionality is once again patent. To be clear, it is not my submission here that the applicants in SCJ argued a case of intersectionality in the narrow sense (in other words, litigated their case based on a unique category of poor, Black people, as in DeGraffenreid, which concerned Black women). Intersectionality is rather used as a theory to describe the overlapping categories of discrimination, where one often impacts on the other or where, primarily due to historical reasons, one increases the likelihood of the other, and gives rise to a particular form of suffering or injustice. This is clear from Sparks and SCJ.
Race and poverty were the two grounds relied on by the applicants to show that the THRR had an indirect discriminatory effect. Although it was not one of the listed grounds of discrimination, geography, just as in Walker, had an impact on resource allocation. In SCJ there was a seemingly neutral and objective allocation of resources, which was determined largely by geographical location. However, its effects were indirectly discriminatory due to the fact that it affected mainly one race group and disadvantaged a particular socio-economic group. Due to South Africa's past racial segregation and the exclusion of this group from the economy, race, poverty, and geography are closely interlinked.
The relationship between race and poverty in South Africa is wellestablished and supported by empirical evidence. It is for that reason that it was somewhat peculiar for the applicants to rely on case law to illustrate systemic poverty, especially in the context of a case that turned on statistical analysis. In the Foreword to a Report by The World Bank, 133 Dr Nkosazana Dlamini-Zuma 134 pointed out that although significant strides have been made since 1994 (the overall poverty rate is generally lower), the economy is still not producing sufficient jobs. The unemployment rate was 27,7% during the third quarter of 2017, with the youth unemployment rate being 38,6%. 135 The Report notes that these statistics are "a reminder of the reality that the country's socio-economic challenges are deep, structural and longterm." 136 The intersectional nature of certain categories of discrimination alluded to above is also patent in the findings. It was pointed out, for example, that in 2006 female-headed households saw a higher incidence of poverty (63,4%). 137 Further, that Black South Africans consistently had the highest recorded poverty rates, with 47% of Black-headed households living in poverty in 2015. 138 It is trite that a lack of access to education hampers economic prosperity. It was then also shown, unsurprisingly, that the levels of poverty tended to decline as the levels of education rose. 139 Furthermore, young children are most impacted upon by poverty, especially in multiple-children households. 140 It is therefore undeniable that poverty has an intersectional, systematic, and pervasive presence in South Africa. Any type of superficial analysis of these statistics reveals that your age, your gender, your level of education, and especially your race have a potential socio-economic impact in this country especially if these characteristics overlap. It was therefore not far-fetched for the applicants in the SCJ case to contend that poverty is analogous to 133 In these aforementioned listed grounds of discrimination. Particular care should therefore be taken in discrimination cases involving an intersectional class of persons.
In contrast to the decision in SCJ, the majority in Jordan seemed unreceptive to the socio-economic realities of sex workers, the majority of whom are women (and statistically likely to be Black), and deemed the gender-neutral criminal sanction a mere consequence of their willingly engaging in illicit behaviour. 141 It was therefore pleasing to see the Court in SCJ (like the Court in Sparks) focussing on the impact rather than just blindly looking at the neutral wording of the THRR.

Conclusion
The first part of this contribution detailed South Africa's antidiscrimination legislation under PEPUDA as it relates to listed and unlisted grounds of discrimination. Further, case law decided directly under the Constitution (in other words, predating PEPUDA) was also discussed. This legislative and constitutional framework was then utilised as a background to reviewing the SCJ judgment. The facts and judgment in SCJ, in turn, highlighted the underlying theme of intersectionality present in cases dealing with indirect discrimination, in particular race, sex and poverty.
South African courts have shown a general willingness to engage with the concept of indirect unfair discrimination. They have also, save for the majority in Jordan, displayed an underlying understanding of and willingness to engage with the overlapping and intersectional nature of deep-rooted categories of discrimination. Our Constitution, along with PEPUDA, has recognised this intersectionality by providing for unlisted or unspecified grounds of discrimination, as it is impossible to legislate the broad spectrum of inequalities that may exist or develop in our society at any particular point in time.
The judgment in SCJ is therefore welcomed.

Case law
DeGraffenreid v  Sparks v Dartmouth/Halifax County Regional Housing Authority 1993 CanLII 3176 (NS CA)

Constitution of the Republic of South Africa 200 of 1993
Constitution of the Republic of South Africa, 1996