Analyst casts doubt on job figures

The latest unemployment numbers, which reached a record high of 26.7%, could be unreliable because of the new master sample Statistics South Africa is using, according to a policy brief issued last week by Neva Makgetla, the programme manager of trade and industry at the nonprofit organisation Trade and Industrial Policy Strategies.

“Year on year, employment reportedly rose by 204 000, or 1.3%, which is essentially in line with previous years and with expected GDP [gross domestic product] growth. The data likely reflect an effective correction to overestimates of job creation in the previous four quarters rather than a sharp contraction in the real world,” she wrote.

Although Makgetla and other labour economists have questioned the figures, which they believe could have been increased because of changes in the master sample and which Statistics SA says has increased in size “to improve precision”, the statistician general, Pali Lehohla, denied the possibility.

“That claim is false and runs against sampling theory, and especially within stratum sample rotation. Changes in the areas from which people would be sampled should not play a role when observing population in the same strata,” he said.

The master sample is designed to be representative at provincial level and within provinces at metro and nonmetro levels. In the metros, the sample is further broken up by geographical type. And it is further divided into urban, tribal and farms, implying that the sample is representative of the different geography types in the metro.

“Stats SA said themselves, in their January publication, that they were worried that the new sampling frame might introduce some inconsistency,” Makgetla said.

She noted that, though Stats SA have run a parallel survey process to crosscheck recent results, they did not crosscheck the data with other economic trends or consider whether their employment finding makes sense.

Makgetla said that Stats SA doesn’t check these numbers against what is going on in the economy. “It doesn’t make sense to have a crash now,” she said, noting it is hard to tell what is going on in the face of statistical problems.

“For virtually every quarter in 2015, the QLFS [quarterly labour force survey] found higher growth than in the previous four years, despite slower economic growth. In the first three quarters, in particular the figures for 2015, it diverged strongly from the averages for the previous five years. It seems likely that the reported jobs crash reflects a more realistic estimate of employment levels following four quarters of exaggerated figures.”

It seems unlikely that this kind of jobs bloodbath could have occurred without a significantly more harmful effect on people, Makgetla said. —