This paper investigates the difference of over 4 million business owners found in the
estimates produced by the QLFS 2010:2 and the Finscope (2010) South Africa Small
Business Survey. The paper discusses a number of possible reasons for the
discrepancy and finds that the QLFS 2010:2 understates the number of small business
owners, with the converse being true for the Finscope (2010) South Africa Small
Business Survey. Even after accounting for differences in the inclusion criteria of the
two surveys as well as the use of divergent sampling weights, a large part of the
difference in the estimates produced is still unaccounted for. The paper also calls into
question the reliability of the estimates produced in Finscope survey as well as the
validity of the negative binomial sampling methodology used to conduct the survey.
Access to additional data from both Statistics South Africa and Finscope, will further
unravel this mystery and provide a better understanding of the sources of the
differences in the estimates produced.