# ANALYSIS: Crime rates worse than South African police calculated

South Africa’s annual crime statistics provide important, useful information. But a calculation error in this year’s stats has skewed crime rates, and how much they have increased or decreased.

 NOTE: Following the publication of this analysis, the South African Police Service revised their official crime rates for 2017/18. You can view the corrected crime rates in our factsheet.

After South Africa’s crime statistics for 2017/18 were released, Africa Check quickly discovered that something didn’t quite check out.

To understand what went wrong, you first need to wrap your head around the difference between absolute crime numbers and crime rates.

• Absolute numbers refer to the number of crimes reported in a given year. For example, in 2017/18 there were 20,336 murders reported.
• Crime rates present the numbers of crimes that occurred per 100,000 people in a country, for example, 35.8 murders per 100,000 people. This allows for population growth.

The problem Africa Check discovered is in the crime rate – specifically the way the South African Police Service calculated it for 2017/18.

## Population statistics used inconsistently

In 2016/17, the police used population estimates for the end of September of each year to calculate the crime rates. This is the middle of the financial year, which is the period the crime statistics cover.

Police spokesperson Col Athlenda Mathe told Africa Check that they used Statistics South Africa’s 2018 mid-year population estimates to put together the 2017/18 crime rates.

There are a number of problems with this.

First, the 2017/18 crime statistics cover the period 1 April 2017 to 31 March 2018. The population estimate the police used is not from this period – it’s from June 2018.

Second, by using a later and larger population estimate, it appeared that South Africa’s population increased from 55,843,011 people in 2016/17 to 57,725,600 in 2017/18 – a difference of more than 1.8 million people.

“There is no way that the national population increased by 3.4% in that year,” Anine Kriegler, researcher and doctoral candidate at the Centre of Criminology at the University of Cape Town, told Africa Check. (She’s right. Based on Statistics South Africa’s estimates, the country’s population increased 1.6% between September 2016 and September 2017.)

“The police need to be consistent in the population statistics they use.”

## Smaller increases and bigger decreases

The higher population estimate may seem like a small issue, but it skews changes in the crime rates between 2016/17 and 2017/18.

“It has the effect of making the increases in crime rates look smaller and the decreases look larger,” Kriegler explained.

For example, the police’s calculations suggest that the murder rate increased by 1.1 from 34.1 per 100,000 in 2016/17 to 35.2 in 2017/18. Using the correct population estimate shows that the murder rate actually increased by 1.7 from 34.1 per 100,000 to 35.8.

Kriegler said this increase was “a lot larger and a lot more worrying”.

 Comparison of change in murder rate 2016/17 2017/18 Change Police murder rate 34.1 35.2 +1.1 Corrected murder rate 34.1 35.8 +1.7

Decreases in crime rates were amplified by using the higher population estimate.

The police’s statistics suggest that assault with the intention to do grievous bodily harm decreased by 15.6 from 305.5 per 100,000 people to 289.9. Using the correct population estimate shows that the rate only decreased by 10.8.

 Comparison of change in assault with the intention to do grievous bodily harm rate 2016/17 2017/18 Change Police assault GBH rate 305.5 289.9 -15.6 Corrected assault GBH rate 305.5 294.9 -10.6

## Accuracy and transparency needed

This is not the first time the police have been criticised for using incorrect population estimates to calculate crime rates. Going forward, the service must commit to using correct and consistent population estimates – and they need to be transparent about it.

“Faith and trust in the police can only take a knock every time they do something like this,” said Kriegler.

Don’t toss all the crime statistics in the bin because of this, though. They contain important and useful information.

The head of the Institute for Security Studies’ justice and violence prevention programme, Gareth Newham, pointed to the detailed information released on murder, in particular.

“The police provided a lot more data on murder and who was being murdered and in what circumstance this year than ever before,” Newham told a briefing in Cape Town.

Africa Check has included the corrected crime rates in its 2017/18 crime statistics factsheet. You can find it here.

 CORRECTION: A previous version of this analysis referred to “percentage point” changes in the crime rates. This was incorrect and has been removed. We apologise for the error.

© Copyright Africa Check 2020. Read our republishing guidelines. You may reproduce this piece or content from it for the purpose of reporting and/or discussing news and current events. This is subject to: Crediting Africa Check in the byline, keeping all hyperlinks to the sources used and adding this sentence at the end of your publication: “This report was written by Africa Check, a non-partisan fact-checking organisation. View the original piece on their website", with a link back to this page.

## Comment on this report

1. By Neil Higgs

Well picked up but I question your use of the term “percentage point”. This normally refers to a change such as 34% to 35%. That is a one percentage point change. A change from 33.9 per 100 000 to 35.6 per 100 000 is not 1.7 percentage points but 1.7 per 100 000

vote