In his budget speech last month, Police Minister Nathi Nhleko promised to demilitarise the police, as proposed by the National Development Plan, improve the police’s crowd control skills and equip them with less lethal crowd-control equipment. He made these promises to reduce police violence against protests, which had led to several protester deaths.

These initiatives are much needed, but are they enough to arrest the authoritarian drift in protest policing? Unfortunately not. One reason for this is because the militarisation concept is understood very superficially in public debate, and because the government, journalists and many analysts have equated militarisation largely with the reintroduction of the military ranking system to the police.

The dominant narrative suggests that the main task at hand is to reverse the ranking system that has apparently brainwashed the police into believing that they are a “force” rather than a “service”, and to retrain them in more civilianised crowd control skills. In fact, the argument that police violence is as a result of a lack of police training has been elevated to the level of common sense.

This narrative portrays police violence as a huge mistake, an aberration, but it masks the deliberate political decisions that have moved South Africa towards becoming a more repressive state. These arguments are simply too easy, too convenient.

When South Africa shifted from apartheid to democracy, the political leadership endorsed community policing, but never really gave it an opportunity to succeed. In fact, it is difficult not to conclude that they set it up to fail, so they could make the case for more authoritarian policing models that had become ascendant elsewhere, including militarised policing and intelligence-led policing.

Much has been made of the fact that when Jackie Selebi became police commissioner he dismantled many of the specialist policing units, ostensibly to democratise the skills they contained.

It has been argued that this led to a loss of vital crowd-control skills and hence an escalation of police violence against protests. Independent Police Investigative Directorate statistics show there is an element of truth to this, but it does not provide a complete explanation.

Tellingly, the one area of specialist policing that continued to expand was paramilitary policing – the number of paramilitary units has trebled over the past decade. So a skills vacuum was created in the more civilianised sections of the police.

A superficial understanding of the militarisation concept serves the police leadership. In fact, the militarisation problem is much more deeply rooted that the military ranking system – and is likely to be more intractable than is being made out.

Possibly the foremost academic on police militarisation, Professor Peter Kraska, has defined militarisation as an ideology that stresses the use of force and threats of violence to solve social problems, and the use of military power as a problem-solving tool.

Based on his ethnographic work in the United States police, Kraska developed a continuum to measure levels of police militarisation. He identified four main indicators: material indicators (the extent of martial weaponry), cultural indicators (the extent of martial language), organisational indicators (the extent of martial arrangements) and operational indicators (the extent of operational patterns modelled on the military).

Drawing on Kraska’s work, the American Civil Liberties Union launched a campaign in 2013 to plot the extent of police militarisation in the US. The nongovernmental organisation’s affiliates filed more than 260 public records requests to determine the extent to which federal funding and support has fuelled police militarisation.

There has been no similar attempt made in South Africa.

Nhleko’s speech focused on the cultural indicators of militarisation – and to a lesser extent on material indicators. Yet there are signs that militarisation has extended beyond these indicators. The danger of not having quantified the extent of militarisation is that the police may remove its superficial manifestations to appease public outrage, while leaving intact its substance. The resulting change is likely to be merely cosmetic.

Of organisational indicators, the most obvious in the police is the proliferation of paramilitary units. There has been no systematic attempt to track when and why they are deployed, which should be of concern because there are signs that their interventions are being normalised in increasing areas of policing, including public order policing.

In Mpumalanga, for instance, the paramilitarised Tactical Response Team appears to have acted as the provincial government’s personal security guards to quell protests. All three paramilitary units were deployed to Marikana, and the head of the most highly trained of the paramilitary units, the Special Task Force, designed the operational plan for Marikana.

This unit has been trained by, among others, the US Special Operations Command, a military unit that co-ordinates the Special Operations Forces. These forces conduct anti-terrorism operations and have prosecuted the US’s “war on terror”, including by way of “renditions” and extrajudicial killings.

This means that the police unit at least partly responsible for the single biggest act of police violence in the post-apartheid period is also extremely skilled, exploding the myth that police violence is a result of a lack of training. Some of the most highly trained policing units have been among the most violent.

It could be argued that the re-establishment of the public order police will make their deployment in protest situations less necessary. But these police have themselves embraced a more paramilitarised model – the French model of public order policing – which involves saturation policing of crowds at close distances.

In the run-up to the 2010 World Cup in South Africa, thousands of police officers were trained in this model, which has been controversial in France for fuelling rather than de-escalating social conflicts in French public housing estates.

The joint deployment of the military and the police are operational indicators: civilianised police forces generally keep the military out of policing and keep the roles of the police, military and intelligence separate. The post-apartheid white paper on defence rejected a domestic role for the military on the grounds that this was a hallmark of repressive regimes, but the country’s current political leadership appears intent on reviewing this decision. Yet no public analysis has been conducted of the frequency of these deployments, or their necessity.

Centralisation is another indicator of militarised policing, because it allows for greater political control to ensure rapid deployment. A good example of how extraordinary security measures become normalised is that a “security legacy” of the 2010 World Cup, the National Joint Intelligence and Operations Structure (Natjoints), appears to have become a permanent feature of the policing landscape.

This body brings together the police, the intelligence agencies and various government departments, co-ordinating the joint deployment of the police and the military, and has grown provincial legs as well. In an eerie echo of the past, in September 2013, Natjoints announced that its newly created “stability committee” was monitoring and fast-tracking prosecutions, which suggests a more centralised approach to the protests, and an escalation of disruptive protests to the level of priority crimes.

The establishment of these structures also suggests that the security cluster is seeing the police, the military and the intelligence services increasingly as one organism – or, in the words of Natjoints, “joint security forces”. Mandate and function creep are likely to intensify, and these roles do not appear to be converging along civilian lines.

It is, however, likely that the police leadership will rein in the more overt forms of police violence, because the political costs are escalating. Yet the drift towards more authoritarian policing will probably continue. There has been little public debate about the shift towards intelligence-led policing, with the growing power of the crime intelligence division being its most visible manifestation.

This form of policing puts intelligence-gathering at the heart of police work, ostensibly allowing the police to take much more strategic decisions about where and how to deploy their resources. In his speech, Nhleko indicated that the police will build the capacity of crime intelligence throughout the country, ­noting that “its significance cannot be overemphasised”.

Intelligence-led policing has been controversial elsewhere. It does not stop the police from engaging in human rights abuses; it merely makes these abuses less visible. Police work also becomes more secretive, which is not good news for South Africans. Already, more activists are complaining about crime intelligence interfering with legitimate advocacy.

The police are not politically neutral; policing choices are shaped by the social forces in which they operate.

In expansionary economic periods, the capitalist system can afford to grant the police semi-autonomy to experiment with “touchy-feely” forms of policing such as community policing. In recessionary periods, by contrast, the police become more visible as instruments of class power. Marikana was the most graphic illustration of this shift.

The political tasks related to police violence are determined by how the problem is understood. Journalists, academics and activists need to become much more serious about documenting police authoritarianism, quantifying militarisation and debating policing models that have been widely criticised elsewhere; these bodies of knowledge are necessary to undertake informed advocacy for police reforms.

But the tasks must not stop there. Growing social polarisation implies that, once again, community self-defence is becoming an important political task, and that the police cannot and must not be relied on to “serve and protect”.

Jane Duncan is a professor in the department of journalism, film and television at the University of Johannesburg.