Skip Ribbon Commands
Skip to main content



    Interrogating Methodologies on Labour Statistics

    Venue: South Africa

     April 02 2012



    A context for the discussion:
    Statistics South Africa (Stats SA) has since 1994 endeavoured to build a credible National Statistics System that captures metadata on various indices of human development.
    Among these indices, official data on economic activity and inactivity – the number and proportions of employers, employees and the self-employed as well as those unemployed or economically inactive – have been central in determining progress in improving people’s quality of life. Indeed, the rate and quality of employment constitutes one of the most critical measures of a country’s economic performance and the extent of people’s well-being.
    In undertaking this work, Stats SA has had to develop a new base of information that takes into account the totality of South Africa’s geography and population, from the fragmented system that existed before 1994. Working in partnership with experts in this field in academia, the research community, the private sector and international agencies, Stats SA has sought over the years to improve both the methodology and quality of labour statistics.
    This has included the improvement and, in 2005, termination of the Survey on Employment and Earnings and its replacement by the Quarterly Employment Statistics (Quarterly Labour Force Survey) as well as a more comprehensive and less regular publication that captures additional aspects of the labour market on informal employment, underemployment and underutilised labour. This is buttressed by the Community Survey and the Census.   
    The Problem
    Despite the progress made in improving labour statistics, debate on the accuracy of the data has continued to rage. This has intensified in the past few years, ranging from the questioning of the utility of the ‘narrow’ versus ‘broad’ definitions of employment, to the methodology and the very accuracy of the data.
    Debate on employment statistics – as with any other data on human development – is welcome, as it can help improve the rigour of data collection and analysis. However, taken to extremes, it can lead to uncertainty on the actual state of affairs and, ultimately, paralysis in policy development, implementation, and monitoring and evaluation.
    In the words of Stats SA:
    “The inadequacy of statistics in 1994 created a desperate need in the new government for information to plan and to assess the performance of its own policies, programmes and projects. [The old Central Statistics Service] CSS was not in a position to cope with the prevailing demand. Accordingly a free-for-all situation developed resulting in a plethora of peddlers of statistics in competition with the CSS. This competition, quite often unhealthy, continues to the present day and constrains the work of Statistics South Africa, the statutory organ of state responsible for official statistics.” (Statistical Reform in SA, Stats SA, 2005?)
    On the other hand, experts located in other institutions such as universities, research institutes and private agencies have argued that their work can complement the endeavours of Stats SA. They cite instances such as on inflation figures where research-based interrogation of the official statistics pointed to weaknesses that had to be rectified.
    It is in this context that Adcorp has posed questions about the accuracy of the QLF, and argued:
    “Due to small sample sizes, many of the survey results are unreliable. For example, estimates of total employment have a confidence value of just 34%, and four out of nine provinces have confidence values below 40%. Additionally, Statistics SA data are not adequately representative of non-traditional employment types, such as contract and temporary employment amounting to approximately 0.9 million employees” (
    While acknowledging that its data are “is not fully representative of the South African economy, [the] data is readily available … and reliable (using population measures rather than sample survey methodologies)”. Adcorp also punts its monthly frequency of calculation, fine-grained as distinct from survey data, and its location in relation to atypical employment as attributes that undergird its reliability.
    Adcorp’s assertion that the unemployment rate in South Africa is closer to 8% as distinct from the official 25% during the middle of 2011, drew the ire of the Statistician-General who said Adcorp’s claims were “spurious and barren on methodology and science of design, collection, processing and dissemination of labour market statistics” (Business Times, 12/07/2011). Others have weighed in accusing Adcorp of pursuing a selfish interest, given its involvement in labour broking, and labelled its figures “part of the company’s ideological warfare against organized workers in this country” (COSATU, 13/09/2012).
    Objective and outcomes
    Given all the above, it is critical that South Africa should have an informed debate on methodologies utilised in gathering, collating and analysing data on the levels of employment and unemployment. This requires calm and rational discourse among researchers and practitioners in the field, shorn of sensationalism, finger-pointing and denialism.
    The MISTRA Roundtable was aimed at bringing together experts on labour statistics to compare notes on methodologies and the disputed data, and more systematically identify areas of divergence and convergence. Besides  assisting in enlightening public discourse, the discussion could lay the basis for more comprehensive research that may, hopefully, lay the matter to rest.

    Related Photos

    in this event?

    Leave us your details and we’ll keep you posted on what’s happening with the event.