/ 6 November 2013

ISS: ‘Misleading’ crime statistics need inquiry

National commissioner Riah Phiyegah after the release of the 2012/2013 crime statistics.
National commissioner Riah Phiyegah after the release of the 2012/2013 crime statistics. (Paul Botes, M&G)

Police significantly over-reported successes in crime fighting in the latest set of crime statistics, the Institute for Security Studies (ISS) said on Wednesday, and under-reported increases in serious crime – and that should be investigated in a formal and independent inquiry.

Yes, it is an election year, ISS researcher Gareth Newham conceded on Wednesday. And yes, that may have played a role. But is that why the South African Police Service (SAPS) significantly mis-reported crime statistics?

"We don't know why it happened. We just know that it has happened, and that there should be an inquiry so it doesn't happen again."

That call is unlikely to be welcomed by national police commissioner Riah Phiyega (who stands accused of undermining an investigation into one of her subordinates), or by the ruling party, which has claimed part responsibility for what has been painted as a continuation in the last year of a long-standing trend of a reduction in crime.

But such claims are based on an amateurish calculation error, the ISS said during a seminar on the national crime statistics for the 2012 to 2013 year, released just over a month ago. Instead of a reduction in crime, the ISS said, the last reported year showed the first uptick in crime generally, and in serious crimes especially, in many years. And that is cause for concern.

Difference in numbers
The results of the error vary between crime categories and areas. On a provincial level, it is most pronounced in the Western Cape, where the difference between police-reported numbers and those calculated by the ISS are greater than 9%. In that province SAPS numbers showed year-on-year reductions in common assault, residential burglary and motor theft. In reality, the ISS said, those crimes actually increased.

"For example, the official crime statistics say that murder in the Western Cape only increased by 0.5%," the ISS said in a statement. "In fact, murder in that province has increased by 10.1%. Similarly, the official statistics say that attempted murder in Gauteng went down by 5.5% when in fact this crime increased by 1.8%."

The police dismissed the ISS recalculations as a mere difference in interpretation. But Newham said the ISS calculations matched very closely with statistics prepared by US analysts and published as international informaiton firm IHS Crime Index, while SAPS numbers were at variance with the IHS statistics. Last week, Police Minister Nathi Mthethwa trumpeted the IHS numbers in a statement.

"It is gratifying that an internationally recognised, credible, and globally respected institution has now affirmed the fact that crime is decreasing in South Africa," Mthethwa said.

That leaves the government in a position where it can either claim independent verification of the long-term trend of crime reduction, or a continuing reduction in the last year of reporting – but can not plausibly claim both.

While the ISS does not dispute the total numbers of crimes reported by the police, it says the calculation of the ratio of crimes per citizen went badly awry.

Types of crimes
Such ratios are crucial in measuring the risk of certain types of crimes, including rape and murder, in a country with a constantly growing population. Should the number of murders increase more slowly than the population grows, for example, the likelihood of being murdered decreases even when more murders take place.

In reporting such ratios for 2012/13, the ISS said, the SAPS used population numbers as determined by the national census conducted in 2011. Because the census is considered a definitive count, it also affects population estimates for previous years. But the police did not re-calculate the ratios for the 2011/12 reporting period based on the more accurate population numbers. Instead, it effectively used two very different sets of population figures to calculate the numbers of murders, rapes and other crimes per 100 000 South Africans.

"If you want to use the ratios, which is correct, then you have to use the same basis for comparison," said Chris de Kock, a retired police major general who oversaw statistics within the SAPS for many years. "You can't just jump between [population estimates] and build in a benefit of 2% nationally and 9% in the Western Cape."

De Kock would not be drawn on whether he suspected a simple – if enormous – mistake in the preparation of data, or whether there may have been a political motive behind a conspiracy to manipulate the numbers. But the response, or lack thereof, from the police was concerning, he said.

"The day after the release [of the annual statistics], which was seven weeks ago, both the ISS and myself reacted on that. On that weekend, there was quite a lot of reporting on it in newspapers … I can tell you if I was there and read that in the newspapers and realised I had made a mistake, it would really have been immediately rectified … But the reaction from the police so far is not to rectify. The reaction so far is 'we have a way to work out the ratios, and a so-called analyst has his way; what is the problem?'. There are not two ways, there is one right way."