Conventional wisdom holds that the threat of harsh punishment acts as a deterrent against crime. This view is based on the assumption that criminals weigh the potential benefit of committing a crime against the potential cost of being caught and punished.
This theory of criminal behaviour has largely been discredited, because in the real world criminals often tend not to think that they will get caught. Nevertheless, the view persists.
Evidence from a large body of research, conducted primarily in the United States, suggests that long sentences and harsh punishment on their own have very little deterrent effect.
The research also shows that the prison experience can even make matters worse. Someone who has been imprisoned is just as likely, if not more likely, to commit a crime than someone who has not gone to prison.
However, there is some evidence that increasing the certainty of being caught by having an effective and visible police service may help. Criminologists such as Steven Durlauf (University of Wisconsin) and Daniel Nagan (Carnegie Mellon University) argue that it is the certainty of being caught and punished (even if that punishment is not a lengthy prison sentence) rather than the length of the sentence that is important in deterring crime.
The approach in South Africa has been to increase the length of sentences for serious crimes. So while the country's prison population has declined over the seven years from 1995 to 2012, the proportion of prisoners serving long sentences increased between 1999 and 2010, declining slightly in 2011 and 2012.
However, the effect of this approach cannot be properly assessed as it is currently not possible to determine how many of the people who are arrested end up in court and receive sentences.
Being able to track individuals and cases through the criminal justice system (consisting of the police, courts and prisons) is essential if we want to know how many people who are arrested make their way to court, and how many are convicted and sentenced. This was identified as a priority in the late 1990s, and substantial investments were made in the necessary technology.
However, South Africa still seems no closer to getting the data that will allow us to keep track of people and cases as they move through the system. Not knowing what happens to these individuals means that we cannot assess whether laws such as those mandating long prison sentences have the desired effect.
Currently, analysts have to rely on annual reports by criminal justice agencies in an attempt to understand the extent to which the system is working. However, it is not possible to compare data presented in the reports of the South African Police Service (SAPS), the National Prosecuting Authority (NPA) and the Department of Correctional Services (DCS).
One reason for this is because there are delays throughout the process from arrest to incarceration. Someone arrested in 2011 may be held as a remand detainee (waiting for trial) for between two and five years before the case against him/her is finalised.
A single case will therefore contribute to the statistics of different annual reports in different years. This does complicate the process of tracking cases, but the three departments do have basic information that could make it possible to track cases.
The SAPS' Annual Report for 2012/13 tells us that the police arrested 1 682 763 people that year (slightly up from the previous year). Of these, 806 298 (47,9%) were for 'serious' crimes (ranging from shoplifting to murder) and 876 465 (52,1%) were for crimes less serious than shoplifting. Most of the cases for crimes less serious than shoplifting would be unlikely to result in a court case, let alone a prison sentence, so those can be set aside for the moment.
The only indication of what might happen to those people arrested for serious crimes is that the police report having 369 204 cases ready for court. However, here we encounter the first difficulty. In cases where more than one person is arrested (for example, when five people are arrested for a house robbery), only one docket is prepared.
This means that we cannot say that almost half of the people who are arrested fall out of the system even before they get to court. To know this, the police would need to indicate the number of suspects arrested in these 369 204 cases. This would allow analysts to assess how many arrests result in cases before the courts.
We run into another difficulty as soon as we turn to the NPA report. We need to know how many of the 369 204 cases that the SAPS says are court-ready are accepted and processed by the NPA.
What we do know is that 143 410 cases were finalised by the NPA through alternative dispute resolution, and that there were verdicts in 323 309 cases.
This means that the NPA finalised 512 614 cases in 2012/13, more than the SAPS sent to it, because of cases already sent the previous year. It would seem that if one adds up the number of convictions in all the courts, the NPA had a total of 290 834 convictions in the year. We don't know if this number represents cases or individuals.
When it comes to prisons, we need to know how many people were sent to prison by the courts. While this information is not contained in the DCS's annual report, it is available on request.
According to the DCS, 291 578 sentenced offenders were admitted to prison during 2012/13. Thus, 744 more people were sent to prison than were convicted.
That can be explained by the fact that the NPA reports on convictions by case, rather than by individuals (i.e. a single case may result in more than one person being incarcerated). This brings us no closer to understanding the throughput of the criminal justice system.
The existing annual reports by criminal justice departments are important and useful tools for assessing the departments and for holding them to account. However, they don't enable analysts to make the kind of assessment necessary for establishing whether South Africa's approaches to crime and criminal justice are working.
A reporting system that encompasses the whole criminal justice system would help us to determine how many people who are arrested end up in court, and how many of those are sent to prison.
The addition of a shared report providing some basic information, which each of the departments already have, would go a long way towards helping us assess whether the country's criminal justice policies are indeed effective.
Chandre Gould, Senior Research Fellow, Governance, Crime and Justice Division, ISS Pretoria