NATIONAL ASSEMBLY
QUESTIONS FOR WRITTEN REPLY
QUESTION NUMBER 1946
DATE OF PUBLICATION: 2 AUGUST 2013
1946. Mr T D Harris (DA) to ask the Minister in the Presidency: National Planning Commission:
(1) Whether there was any statistically significant difference between the preliminary population estimates of the 2011 Post Enumeration Survey presented to the SA Statistics Council on 2 August 2012 and the final results published in October 2012; if not, what was the justification for the additional amounts spent on (a) consultants, (b) fieldworkers and (c) other services between August 2012 and October 2012; if so, what was the difference;
(2) what amounts were spent on (a) consultants, (b) fieldworkers and (c) other services between August 2012 and October 2012?
NW2295E
Reply:
(1) There was a statistically significant difference between the preliminary population estimates of the 2011 Post Enumeration Survey (PES) as presented in August 2012 and the one published in October 2012. The difference arose because the August result did not execute the PES methodology thoroughly, thus generating a total population with significant inherent biases at the level of sub-groups.
The second significant difference was at the level of provinces arising out of both clerical omissions and methodological errors in the computation of the PES. The final results of the PES showed that the undercount in Census 2011 was 14.6% instead of the 18.3% presented earlier to the Statistician-General in July 2012. The table below illustrates the point.
|
Province |
Reopened undercount Rate |
Initial undercount rate |
Difference in undercount estimates |
Ranking |
|
Western Cape |
18.5 |
22.0 |
-3.5 |
(1) (2) |
|
Eastern Cape |
12.9 |
15.7 |
-2.8 |
(7) (7) |
|
Northern Cape |
13.4 |
25.9 |
-12.5 |
(6) (1) |
|
Free State |
10.1 |
13.0 |
-2.9 |
(8) (8) |
|
Kwa-Zulu Natal |
16.7 |
21.1 |
-4.4 |
(2) (3) |
|
North West |
14.9 |
18.4 |
-3.5 |
(4) (5) |
|
Gauteng |
14.7 |
19.0 |
-4.3 |
(5) (4) |
|
Mpumalanga |
15.5 |
17.8 |
-2.3 |
(3) (6) |
|
Limpopo |
10.0 |
12.2 |
-2.2 |
(9) (9) |
|
RSA |
14.6 |
18.3 |
-3.3 |
|
The table above shows the level of undercount as well as the shift in ranking between the initial undercount estimate and the final undercount estimate. For instance Gauteng had an estimated undercount of 19.0% initially and the reopened level settled Gauteng at 14.7%. A 4.3 percentage point difference in undercount for Gauteng would have overestimated Gauteng's population by 1.3 million and subsequently potentially increasing the claim of Gauteng to the provincial allocation by up to R 12 billion, thus reducing the claim to provincial allocations by other provinces. The Northern Cape, which is small in population size, ranked first in undercount, but was, in fact, sixth.
(2) The costs of fieldwork and processing of the reopened PES amounted to R9 million. The cost of consultants was R2 million and associated services were R700 000 for the work performed from July to October 2012.
Issued by Parliament, August 12 2013
Click here to sign up to receive our free daily headline email newsletter
