The Cost of “Too Late”

Reacting to Autopsies, Not Cries for Help.

---

The deadly gatekeeper: the illusion of a safe haven

The most dangerous place for a victim of Gender-Based Violence (GBV) is not always the home; sometimes, it is the front desk of a local police station. When an individual finally gathers the courage required to report an intimate partner, they are taking their survival into their own hands. It is a moment of extreme vulnerability, a literal life-or-death plea for state intervention. Yet the institutional response to this act of desperation is too often reduced to a single, dismissive phrase: “It’s a family matter. Go home and resolve it among yourselves.”

By treating intimate partner violence as a private dispute rather than a public crime, law enforcement acts as a gatekeeper. Officers close the doors of the legal system before its statutory protections can be applied, leaving the victim exposed to retaliation. South Africa is a country that Parliament’s own police committee describes as struggling with alarming levels of GBV and femicide, yet the front line of defence continues to fall short of its duties. The Portfolio Committee on Police has described the level of non-compliance with the Domestic Violence Act as shocking, noting that 59 cases of non-compliance were reported between October 2023 and March 2024 (Parliament of the Republic of South Africa, 2024b).

When the front line views domestic criminality through the lens of interpersonal mediation, it compromises its constitutional obligations. The law on paper may promise protection, but in practice it is only as robust as the officer sitting behind the charge office desk.

Shifting the blame: the privatisation of public violence

Why does the state struggle to treat domestic abuse with the same gravity as street-level crime? Part of the answer lies in a long-standing culture of treating violence in the home as a private matter. Society has long drawn an imaginary boundary around the household, treating the front door as a barrier the state should not cross. When a police officer advises a bloodied or terrified complainant to “talk things out,” they are not engaging in conflict resolution; they are exposing the complainant to further violence.

This dynamic rests on a flawed assumption: that a victim and a perpetrator have equal bargaining power. In reality, intimate partner violence thrives on coercive control, psychological intimidation, and economic dependency. Forcing a victim to negotiate with their abuser treats an asymmetric criminal dynamic as a simple relationship disagreement. It trivialises the assault and tells the victim that the state does not regard their suffering as a public concern.

Furthermore, when officers demand visible, severe physical injury before a complaint is taken seriously, they ignore the psychological reality of living in fear. The problem is not new. In 2018, the Dullah Omar Institute’s ParlyBeat reported that, across three consecutive reporting periods up to March 2018, not a single police station visited by the Civilian Secretariat for Police Service had fully complied with the Domestic Violence Act (October, 2018). Six years later, Parliament’s police committee was still reporting station-level non-compliance (Parliament of the Republic of South Africa, 2024b). Such indifference reinforces the abuser’s ultimate leverage: the belief that the victim is truly alone, and that no one is coming to help.

The structural blind spot: law versus enforcement bureaucracy

Even when a victim gets past the dismissive barrier at the front desk, they enter a second maze of vulnerability: administrative delay. The legal apparatus is slow, built on procedural checks, paperwork, and court rolls. An abuser acting in a fit of rage does not wait for a court date.

Research on intimate partner femicide identifies estrangement, particularly from a controlling partner, as a significant risk factor (Campbell et al., 2003). The period immediately after a victim seeks help or tries to leave the relationship is therefore one of heightened danger. An abuser may escalate precisely because the victim has tried to reclaim control by involving the state. Yet the legal process treats this volatile period with a staggering lack of urgency.

Resource allocation tells a similar story. Specialised units such as the Family Violence, Child Protection and Sexual Offences (FCS) units depend on adequate funding, and in 2024 the Portfolio Committee on Police noted that SAPS had identified additional funds for these units as a critical need (Parliament of the Republic of South Africa, 2024a). Meanwhile, the committee’s chairperson has described the roughly R2 billion spent on VIP protection as disproportionate (Ndenze, 2025).

Backlogs in the wider justice system compound the problem. In March 2025, the police committee heard that the DNA backlog had again exceeded 140,000 cases, which its chairperson called a direct threat to justice for victims of GBV (TimesLIVE, 2025). When evidence takes months or years to process and protection orders go unserved, bureaucratic delay becomes a physical threat. The system treats time as an administrative metric, whereas the victim experiences it as a countdown to the next assault.

The cost of “too late”: reacting to autopsies, not cries for help

This structural failure reveals a grim reality: the legal system is fundamentally reactive, designed to respond to visible, undeniable harm rather than to intercept escalating danger. The state demands a high burden of physical proof, meaning it often validates a victim’s terror only when it arrives as a medical report or a corpse.

This creates a tragic double standard. If a stranger assaults someone on a public street, the state is expected to intervene with flashing lights, forensic collection, and handcuffs. But if an intimate partner inflicts the same terror behind closed doors, the system hesitates. It tells the victim that their fear is not currency the court accepts; only their blood is.

Consequently, the criminal justice system overlooks the documented breadcrumbs of escalating danger. Dismissed police reports, unserved protection orders, and ignored emergency calls are treated as minor procedural omissions. Yet when the inevitable escalation occurs, the state suddenly finds an abundance of resources.

Blue lights arrive not to protect a living person, but to cordon off a crime scene. Detectives are assigned not to prevent a tragedy, but to process one. Forensic units are dispatched to collect evidence from a body that spent months pleading for the state to take notice while it was still alive. By refusing to act when the victim cries for help, the state becomes an accidental accomplice, validating the abuser’s sense of impunity until the moment the violence becomes fatal.

The Constitutional Court has made clear that police failures carry legal consequences. In AK v Minister of Police (2022), the majority of the Court found that the police breached their statutory and constitutional duties to a woman who was held captive and repeatedly raped, because of significant mistakes in the search for her and in the investigation that followed. That case did not involve intimate partner violence, but it confirms that police inaction in GBV cases is not merely poor service; it can be unlawful. True justice cannot look like a well-managed homicide investigation; it must look like a proactive, immediate shield.

Beyond the front desk: rethinking training, deployment and protection orders

Holding individual officers accountable is necessary, but it may not be enough. Perhaps another layer is to revisit the training that our police receive, because we may be expecting something of them that they have not been equipped to provide. Research on SAPS training found that although most members interviewed had a basic understanding of the Domestic Violence Act, their ability to apply it in practical problem-solving scenarios was often limited, and it recommended more practice-oriented training (Combrinck & Wakefield, 2010). Training and deployment sit at the core of service delivery. If crime takes place not only in the streets but also in the home, then policing models built around public spaces must shift with it, and home-based policing, with officers trained and deployed to engage with domestic settings, may be necessary.

A second element is the burden of proof required for an interim protection order. The Domestic Violence Act directs the court to issue an interim order, even without notice to the respondent, if it is satisfied that there is prima facie evidence that the respondent is committing or has committed an act of domestic violence, and that the complainant may suffer undue hardship if an order is not issued immediately (Domestic Violence Act 116 of 1998, s. 5(2)). The statutory test is therefore not proof. Yet there remains an expectation that an applicant must show at least some form of violence, and even proof of it, to qualify. A threshold applied this way risks shutting out people whose danger is real but not yet visible, which is precisely the escalation the system should be intercepting.

The third element is the prejudice that a respondent, or alleged perpetrator, may suffer. A protection order restricts a person’s rights, and those interests are legitimate. The Act does contain safeguards: an interim order is temporary, and a final order may be issued only if the court finds, on a balance of probabilities, that the respondent has committed or is committing an act of domestic violence (Domestic Violence Act 116 of 1998, s. 6(4)). Even so, when balancing these competing rights, courts should be given a clear guideline of inquiry for deciding whether to grant an order. The test should not be stringent, but it must properly weigh the interests of all parties. A well-defined standard serves everyone: victims are not turned away for want of visible injuries, and respondents are not made subject to orders without a fair basis.

Conclusion: dismantling the gate

For family law and criminal justice systems to truly protect the vulnerable, they must undergo a profound cultural shift, supported by better training, policing that reaches into the home, and a fair but accessible threshold for protection orders. True protection does not start in a high court or a forensic laboratory; it starts at the police station front desk. Law enforcement officers must be stripped of their self-appointed roles as marital mediators.

Police stations must enforce zero-tolerance accountability for officers who fail to log domestic complaints. The need is sharpened by evidence that officers are sometimes themselves implicated: the Independent Police Investigative Directorate (IPID) registered 5,259 cases of police misconduct in 2025/26, of which 332 were linked to gender-based violence and femicide, including 122 referrals that were mainly domestic violence (TimesLIVE, 2026).

Until domestic and gender-based violence is treated with the same tactical urgency as an armed robbery in progress, the legal system will remain structurally complicit in harm to victims. The metric of success for a justice system should never be how many abusers it convicts after a murder; it must be how many victims it keeps alive. The state must stop answering autopsies and start answering cries for help.

---

References

---

AK v Minister of Police (CCT 94/20) [2022] ZACC 14 (5 April 2022).
https://lawlibrary.org.za/akn/za/judgment/zacc/2022/14

Campbell, J. C., Webster, D., Koziol-McLain, J., Block, C., Campbell, D., Curry, M. A., Gary, F., Glass, N., McFarlane, J., Sachs, C., Sharps, P., Ulrich, Y., Wilt, S. A., Manganello, J., Xu, X., & Schollenberger, J. (2003). Risk factors for femicide in abusive relationships: Results from a multisite case control study. American Journal of Public Health, 93(7), 1089–1097. https://doi.org/10.2105/ajph.93.7.1089

Combrinck, H., & Wakefield, L. (2010). Going the extra mile: Police training on domestic violence. South African Crime Quarterly, 31, 27–34. https://doi.org/10.17159/2413-3108/2010/i31a890

Domestic Violence Act 116 of 1998 (S. Afr.).

Ndenze, B. (2025, January 21). Time to reallocate non-essential functions of VIP Protection Unit – Police committee. EWN. https://www.ewn.co.za/2025/01/21/time-to-reallocate-non-essential-functions-of-vip-protection-unit-police-committee

October, A. (2018, September 11). Police failing domestic abuse victims? Counting the costs of SAPS non-compliance. ParlyBeat. Dullah Omar Institute, University of the Western Cape. https://dullahomarinstitute.org.za/our-work/women-and-democracy/parlybeat/police-failing-domestic-abuse-victims

Parliament of the Republic of South Africa. (2024a, July 11). Media statement: Police committee concerned by continued capacity shortages within crime intelligence and detectives. https://www.parliament.gov.za/press-releases/media-statement-police-committee-concerned-continued-capacity-shortages-within-crime-intelligence-and-detectives

Parliament of the Republic of South Africa. (2024b, October 30). Media statement: Shocking non-compliance with provisions of Domestic Violence Act at police station-level. https://www.parliament.gov.za/press-releases/media-statement-shocking-non-compliance-provisions-domestic-violence-act-police-station-level

TimesLIVE. (2025, March 12). DNA backlogs exceed 140,000 cases, prompting call for AG to intervene. TimesLIVE. https://www.timeslive.co.za/news/south-africa/2025-03-12-dna-backlogs-exceed-140000-cases-prompting-call-for-ag-to-intervene/

TimesLIVE. (2026, October 1). Police officers linked to 332 GBVF cases in 2025/26, says IPID. TimesLIVE. https://www.timeslive.co.za/news/south-africa/2026-10-01-police-officers-linked-to-332-gbvf-cases-in-202526-says-ipid/

---

Back to Articles