‘Two Million Net New Jobs’: A Reconsideration of
the Rise in Employment in South Africa, 1995-2003
Daniela Casale
Division of Economics
University of Kwa-Zulu Natal
casaled@nu.ac.za
Colette Muller
Division of Economics
University of Kwa-Zulu Natal
mullerc2@nu.ac.za
Dorrit Posel
Division of Economics
University of Kwa-Zulu Natal
posel@ukzn.ac.za
Development Policy Research Unit August 2005
Working Paper 05/97 ISBN: 1-920055-14-2
In this paper we investigate labour market trends in South Africa between October 1995
and March 2003. In particular, we evaluate the South African government’s claim that over
this period, the economy created two million net new jobs. Using the same household
survey data as that used to generate offi cial employment estimates, we also fi nd an
almost two million net increase in employment. However, we show that this increase is
likely to have been infl ated by changes in data capture and defi nitions of employment
over the years, and that the real increase may be considerably less, with a lower bound
of approximately 1.4 million jobs. We argue further that the rise in employment over the
period must be evaluated in the context of a dramatically larger growth in labour supply
and therefore rising rates of unemployment, declining real earnings, and an increase in
the number of the working poor, particularly among Africans.
This paper has been published in The South African Journal of Economics,Vol. 72:5
Desember/December 2004.
Development Policy
Research Unit
Tel: +27 21 650 5705
Fax: +27 21 650 5711
Information about our Working Papers and other
published titles are available on our website at:
http://www.commerce.uct.ac.za/dpru/
Abstract:
Acknowledgements:
1. Introduction....................................................................................1
2. Data...............................................................................................2
3. Employment Trends, 1995 - 2003..................................................3
(a) A two million net increase in employment?...................3
(b) Where has employment increased?..............................7
(c ) The increase in labour supply......................................8
4. Implications of the Increase in Employment.................................10
5. Conclusion...................................................................................18
6. References...................................................................................20
Table of Contents
‘Two million net new jobs’: A reconsideration of the rise in employment in South Africa, 1995-2003
1
There has been much debate in South Africa over the past decade concerning national
employment and unemployment statistics (see, for example, Standing et al, 1996;
Bhorat, 1999; Klasen and Woolard, 1999; Schlemmer and Levitz, 1999; Nattrass, 2000).
This followed the introduction in the 1990s of household surveys that for the fi rst time in
the country captured detailed information on individual employment status. The debate
about statistics intensifi ed more recently in the run-up to the 2004 national election. The
statement by the ANC government that “the economy created two million net new jobs
since 1995” based on these surveys, in particular provoked considerable discussion in
the media as to whether these trends are credible.
1
With the growing recognition of very high (and increasing) rates of unemployment in the
country, and reports of fi rms engaging in large-scale retrenchments, it is perhaps not sur-
prising that statistics, which suggest an increase in employment of this magnitude, are
met with controversy and disbelief. Furthermore, in 2000 a new household survey (the
Labour Force Survey) was introduced which was designed to capture all forms of work
with greater effi ciency. More specifi cally, increased emphasis was placed on classifying
as employed those engaged in informal activities and small-scale agriculture, even if for
only one hour in the previous week. This has confounded the debate because it could be
argued that the increase in employment is not real, but rather is an artefact of changing
defi nitions and improved data collection.
Our objective in this paper is to evaluate the claim that two million net new jobs were
created between 1995 and 2003, drawing on the same data sources as those used to
generate offi cial estimates on employment. We begin by assessing the validity of this
fi nding, given the problems with measuring employment status consistently across the
years using national household survey data. We show that, taken at face value, recorded
employment did indeed increase by close to two million jobs over this eight-year period.
Even if a sizeable part of this increase is real, however, we explain why it is very likely
that some (not inconsiderable) portion is the result of changes in defi nitions and data
capture.
1 See “Employment I: We need a people’s contract to create work” and “Employment II: The myth of jobless
growth” (ANC Today, Volume 4, No. 9, 5-11 March 2004); “The debate on the President’s State of the Nation Address” (Speech
by Minister of Labour MMS Mdladlana in Parliament, 10 February 2004, Cape Town); and “Doomsayers take liberties with
facts on employment” (Minister of Trade and Industry Alec Erwin, Sunday Independent, Business Report, 15 February
2004). And in response: “Unions and researchers slam government over claims of jobs growth” (Terry Bell, Published on
the Business Report website, 15 February 2004); “Erwin digs in his heels on job increase data” (Quentin Wray, Published
on Business Report website on 23 February 2004); and “DA rubbishes ANC jobs claims” (Nic Dawes, This Day, 8 March 2004).
We argue further that in evaluating the government’s claim of job creation there are other
factors, beyond the verifi cation of statistics, which need to be highlighted for a more
complete picture of labour market trends in South Africa. This paper adds to the debate
by exploring three of these: the types of employment that have increased; the magnitude
of the employment increase in relation to the growth in labour supply; and the changes in
earnings that have accompanied the rise in employment.
2. Data
Labour market analysts are generally in agreement that the most accurate way of
measuring employment status in South Africa is through the national household surveys
conducted by Statistics South Africa (SSA) – the October Household Survey (OHS), an
annual survey conducted from 1993 to 1999, and the Labour Force Survey (LFS), a
biannual survey introduced in 2000. The two other sources from which labour market
statistics have been drawn are the population census and the Survey of Employment and
Earnings (SEE). Both of these latter sources however are not suitable for a comprehen-
sive analysis of the labour market over time.
The population census, which covered all households in South Africa in a representative
manner in 1996 and 2001, does not include detailed questioning on an individual’s labour
market status. Also, the few questions asked changed quite substantially between 1996
and 2001, making comparisons over the period very diffi cult. For example, a distinction
between formal and informal employment is only drawn in the 2001 census.
The SEE, conducted since 1998
2
, attempts to capture employment from the fi rm side.
Before 2002, however, an outdated sample of fi rms was covered such that employment
growth in rapidly expanding sectors (information technology, for example) was not picked
up. In 2002 the sample of fi rms was redrawn to take into account the changing structure
of the economy (SSA, 2000; 2002). Another key downfall of the SEE is that only formal
(registered) non-agricultural businesses are sampled and therefore no information on
informal, agricultural and domestic work is captured. As we show in Section 2 below,
these types of employment constitute a substantial part of total employment in South
Africa, and the SEE consequently cannot be used to identify trends in aggregate employ-
ment.
2 The SEE was piloted in the third quarter of 1997 and was “in operation from the first quarter in 1998” (SSA, 2004:4).
2
DPRU WP 05/97 Daniela Casale, Colette Muller & Dorrit Posel