Quarterly Labour Force Survey: Statistics South Africa briefing

This premium content has been made freely available

Finance Standing Committee

17 September 2008
Chairperson: Mr N Nene (ANC)
Share this page:

Meeting Summary

Statistics South Africa (Stats SA) formally introduced its new Quarterly Labour Force Survey (QLFS) to the Committee. The Committee was provided with an update of the re-engineering of the collection of labour market statistics after an evaluation and review of the Labour Force Survey (LFS) that had been informed by consultations with key stakeholders such as the International Labour Organization (ILO). Stats SA submitted that the survey had improved from the previous surveys that had been carried out bi-annually. The survey results were now released within a month of the last quarter. These shortened timelines had enhanced the capability of the survey to inform policy-making. Delays in information delivery to policy-makers were now a thing of the past. The QLFS had also brought in new ways of defining labour market concepts such as employment and unemployment.

Some of the key reporting changes introduced by the QLFS included greater international comparability, the exclusion of non-market activities from the definition of employment, and greater recognition of the informal sector in terms of employers, own-account workers and persons who worked unpaid in their household business. The summary highlighted key labour market indicators such as employment and unemployment rates. Quarterly changes in these rates revealed that employment had increased by 0.8% quarter to quarter whilst unemployment had declined by 8.3%. The statistics also showed gender disparities in labour force participation and employment absorption rates that reflected a higher rate of unemployment for women compared to men. There were many jobs that had been created in the community and social services sector, whilst trade accounted for the highest number of jobs, being 22.6% of total employment.

The Committee discussed with Stats SA whether the new methodology for collecting labour market data would enhance policy-making. Questions were raised about the efficacy of different labour market concepts, including unemployment and market related activity, whether seasonal work was recognized and the effects of this and other sporadic employment on unemployment and employment rates. Members questioned the broad and narrow definitions of unemployment, with some feeling that  a narrower definition evaded government’s liability towards alleviating high unemployment. This was compared to the international situation, and it was determined that unemployment had a structural basis for certain population groups, and that the challenge was for policy-makers to resolve the social inequalities that contributed to unequal labour market participation. The gap between skills and demand in the economy was also identified as demanding the urgent attention of policy-makers. Members also considered the likely effect of inter-provincial migration on the accuracy of provincial labour market data, and debated on the status of women’s unremunerated labour value in activity such as domestic work and home-making. Stats SA summarised that its work provided a picture that only revealed broad trends, and greater research was required to adequately inform policy-makers.

Meeting report

Quarterly Labour Force Survey (QLFS): Briefing by Statistics South Africa (Stats SA)
Mr Pali Lehohla, Statistician General, said that the Quarterly Labour Force Survey (QLFS) was a part of Stats SA’s re-engineering process. It was an indication of the progressive improvement that Stats SA was achieving over time to improve the delivery of information for policy-making. The organisation had previously suffered negative publicity regarding the accuracy of its statistics pertaining to its community survey, after irresponsible reporting by a certain journalist from the Star had misrepresented facts to the public, and after allegations that the results of the survey could not be relied upon. However, the matter had been set straight, the Star had made an apology and the journalist concerned had retracted the misleading story. It was unfortunate however that this was only done after a period of eight or nine months when the damage had already been done.

Ms Kefiloe Masiteng, Deputy Director General, Population and Social Statistics, provided the Committee with an update of the progress of the re-engineering of the labour force survey. The QLFS results had been launched on 28 August and a nationwide education and awareness campaign was under way to ensure that they were disseminated and understood and would be used to inform policy-making. Her presentation provided a brief historical background behind the initiative to re-engineer the collection of labour statistics due to several concerns about the efficacy of the Labour Force Survey (LFS). She submitted that concerns had been raised about the reliability of data; the conduct of the survey on a six monthly basis; its level of timeliness and coverage. These concerns had prompted Stats SA to undertake certain evaluations that then led to the decision to re-engineer the LFS in June 2005. A team had been set up within the organisation to ensure continuity and the sustainability of this process. Since the inception of the re-engineering drive, there had been consultations with the International Labour Organization (ILO) and several other international consultants. Consultations had also taken place with SADC and East African countries. These included issues around the labour market and statistics, and how other countries practiced, and how to use the information generated by Stats SA. Meetings had also been convened with key stakeholders in March, June and December 2008 to update them on the progress of re-engineering the LFS.

Ms Masiteng also provided an overview of the large-scale field tests that had been undertaken to determine the feasibility of the QLFS during its formative stages. Field tests had been conducted in different parts of the country to test new core questionnaires and the use of GPS satellite technology. Tests had also been conducted in Limpopo, Gauteng, Mpumalanga and Eastern Cape to test the revised core questionnaire and Real Time Management Systems (RTMS). A report had been compiled by the International Monetary Fund (IMF) on the outcome of the field tests and was available at the organisation’s website. The key feature of the QLFS was that it was a quarterly continuous household survey as opposed to the bi-annual LFS. It measured three components of the labour market, namely those who were employed; unemployed; and not economically active. The QLFS also attempted to disaggregate data on the basis of different aspects of the labour market including categories such as industry, occupation, sector, hours of work and discouraged work-seekers. The survey relied on a sample of 30 000 dwelling units across the country, where information was collected from over 60 000 persons of between the ages of 16 and 64 years, using face-to-face interviews. The structure of the questionnaire used was provided in the guide attached to their submission. It had a tracking system using a reference number. The questionnaire contained different sections, which Ms Masiteng explained to the Committee in terms of their objectives and the nature of information that would be captured. For instance the first section of the questionnaire captured basic demographic information such as name, age, contact details and education. Labour market related data was captured in subsequent sections, which dealt with specific issues such as employment status. The questionnaire defined certain important labour market concepts. The criteria for “unemployed”, for instance, included consideration of whether a person was available to work and had taken steps towards looking for employment. Those deemed to be “not economically active” included individuals who had indicated that they were not available to work or those that had stated that they were available but had not taken any active steps towards seeking employment. This category also included discouraged work seekers. The unemployment rate was determined by the proportion of those people who were not employed as a percentage of the labour force.

In terms of the changes implemented as a result of process re-engineering, decisions had been taken regarding certain ways of disaggregating the data. For instance the working age population group had been shifted from 16 to 65 down to 16 to 64, to conform to international standards. In the definition of employment, the survey would only consider issues that were related to market production activities. Non-market production activities, although forming a part of the statistics gathered, were not a part of the employed. Discouraged work-seekers comprised only those whose reason for not seeking work was either that they had lost hope of finding work, they did not have relevant skills or qualifications or would have said that there was no work available in the area. The informal sector would now comprise of both employers and employees. It was recognized that within this sector, there were employers in the sense of own-account workers and persons working unpaid in their household business. These household businesses would not however be registered for VAT and income tax. Employees were those people who worked in enterprises of less than five employees and were not registered by their employers for income tax. Ms Masiteng also provided details of the criteria for determining discouraged work seekers and the procedure of recording this information in the questionnaire during the face-to-face interview.

The QLFS had begun on 14 January 2008. Data had been collected for the first quarter between January and March, and for the second quarter between April and June of 2008. Quarterly changes in employment figures revealed that 106 000 new jobs had been created in the labour market reflecting an increase of 0.8%. The number of unemployed persons had gone down by 77 000 or 1.8%, whilst those in the category of ‘discouraged work seekers’ had declined by 98 000 or 8.3%. Looking at labour market indicators, the average unemployment rate in the second quarter was estimated at 23.1% of the working population of whom 26.8% were women and 19.9% were men. The employment population ratio, which was the absorption rate, had shown that there was higher absorption for men than there was for women. The labour force participation rate for women had been estimated at 51.2% in the second quarter, showing a 0.1% decline from the first quarter estimate of 51.1%.

Ms Masiteng also took the Committee through the survey results pertaining to employment by industry. She highlighted the fact that more than half of the additional jobs were in community and social services. Trade accounted for the highest number of jobs created, namely 3.1 million or 22.6% of total employment. The results pertaining to employment by occupation indicated that elementary jobs accounted for the largest share of total employment with figures of 3.1 million or 22.8% of total employment. The increase in elementary jobs accounted for 75 000 out of 106 000 additional jobs. Employment in the formal and informal sector was characterised by quarterly employment gains of 106 000, which were largest in the formal sector, with figures indicating an increase to 73 000 or 0.8%. The informal sector constituted 17% of total employment. Comparisons of the informal and formal sector figures in the second quarter revealed that trade was the most important industry accounting for 46.1% of employment in the informal sector. The survey’s results on employment revealed relatively higher unemployment for women than men. The quarterly fall in unemployment was largely among men. In terms of long-term unemployment, the results showed that the quarterly decline in unemployment was due to a reduction in short-term by 112 000 or 6.1%, and yet persons in long-term unemployment accounted for over half of all unemployed persons (58.5%).

Unemployment statistics were also disaggregated according to province. The largest quarterly difference in the unemployment rate was in the Eastern Cape with figures showing a decline of 3.3%. The national unemployment rate had decreased by 0.4%. Unemployment rates were also classified according to population group as well as gender. This revealed that African women had the highest rate of unemployment with a figure of 31.1%, when compared to other population groups. The results also provided information pertaining to persons who were not economically active. Students accounted for the largest number of not economically active persons, with figures of 5.7 million in the second quarter, or 44.1%. The discouraged work-seekers had the largest quarterly decline of 98 000 or 8.3%. The survey had considered the involvement of respondents in non-market production activities. This had revealed that fetching water or collecting wood or dung was the most important non-market activity, followed by subsistence farming. The vast majority of persons in non-market activity were not economically active.

Ms Masiteng also informed the Committee of the historical revisions to the LFS based on the overlap between the LFS of March 2008 and the QLFS (First Quarter) of January-March 2008. This had been done to ensure that the series would not be interrupted. The revised series revealed higher employment than originally published, although the actual figures showed that this was not a substantial difference. This was the same with unemployment, which was lower in the revised series than originally published. This trend continued with rates for economic inactivity. Discussion had taken place within the organisation over the sources of these differences. Some of the reasons were that non-market activities had been excluded from employment in the QLFS or there had been different sequences (persons in the market were not asked about their non-market activities) or different instruments (shorter core and less respondent fatigue). The QLFS questionnaire had been shortened to remove certain variables that were not considered essential elements of the labour market. Other reasons included that the QLFS survey collected data in a continuous cycle as opposed to the LFS point-in-time technique whereby data was collected for two weeks in March, June and September of every year, individual questionnaires were contained within one book, and permanent field-staff were employed for the QLFS as opposed to contract field-staff in the LFS.

The additional aspects of the QLFS results that would be introduced were:
-Time-related underemployment
-Measures of underutilisation of labour
-The characteristics of the unemployed, being divided into new entrants, job-losers, job-leavers, re-entrants and informal employment.

All these would be posted onto the website once they had been done.

Mr B Mnguni (ANC) asked why it had taken long to conform to internationally recognised standards.

Mr Lehohla responded that it had indeed taken long to conform to international standards. However, the review of the collection of statistics had begun in 2003 and it had not been possible to pursue change on all fronts. Although there had been an eagerness to do so, the reality had been that it would not be possible as the workload was simply too much and other areas demanded higher priority. 

Mr Mnguni asked what the effect would be on the statistics in terms of the quality of the jobs, considering that some employment could be short-term and would lapse after the period covered by the survey.

Ms Masiteng responded that the information received by Stats SA was subjective. The issue around the quality of jobs was a policy consideration. Stats SA was restricted in its mandate to counting the number of jobs and disaggregating them according to different classifications. There was a slow rate of development in ensuring that required skills were made available in the labour market. There were structural issues that had resulted in a skills-gap, and as South Africa’s economy developed this continued to outpace the available skills.

Mr Mnguni asked if the survey had a category that captured information related to skilled persons who were unemployed.

Ms Yandiswa Mpetsheni, Executive Manager, LFS Re-engineering Programme, Stats SA, responded that a breakdown could be done in terms of the skills of the unemployed.

Mr Mnguni asked why skilled people should be referred to as unemployed. There were people without formal qualifications who nonetheless had the skills required for certain jobs.

Ms Mpetsheni responded that this was a question that policy-makers were better equipped to answer.

Dr D George (DA) asked whether the level of skills of not economically active people, especially the discouraged work seekers, was known.

Dr George observed that there was also a category for “other” and he wanted to know who would fall under it.

Ms Mpetsheni responded that there were about sixteen reasons provided for in the questionnaire and some other reasons in the release. In the release, the answers for non-employment included a student; home-maker; people who said they were ill or disabled; people who were too old or too young to work; people undergoing training; or people without transport. Since the numbers had been too small, they had been lumped together under the category “other”. There were a whole lot of categories under “other” and these could be found in the guide to the questionnaire.

Mr S Marais (DA) asked clarity on the submission regarding consultation with key stakeholders. He wanted to be given an indication of who the stakeholders were.

Ms Mpetsheni responded that Stats SA consulted the Department of Labour, The Reserve Bank, National Treasury and the Presidency. They had been consulted on a series of issues related to labour market concepts such as the definition of employment and market-related activities. Other stakeholders had included researchers, government employees and both local and provincial government.

Mr Marais asked whether only government stakeholders had been consulted and not civil society labour organisations, agriculture or business or any such organisations that could have shed light on these issues.

Ms Mpetsheni responded that Stats SA had consulted with organised labour and agriculture. She had not provided a full list of the stakeholders except to highlight the key ones. There had been a whole variety of stakeholders, to cover everyone who could have an interest in the collection of labour market statistics.

Mr Marais asked a question relating to what he referred to as the narrow and broad definition of employment. The narrow definition did not seem to adequately address government’s obligation to deal with unemployment. It seemed that the survey was steering away from the broad definition of unemployment, which placed a liability on government.

Mr Lehohla responded that the narrow definition was adopted for international comparability and hence was the official definition for unemployment. The broad definition had not been hidden because it was included in the organisation’s flow chart, in terms of discouraged work seekers.

Ms Masiteng added that South Africa was no exception to the rest of the world as it was recognised that the definition of unemployment was both narrow and expanded, but that the narrow definition was the official one for the publication of statistics.

Mr Marais (DA) asked for clarity on what was meant by “improvements”, asking how these were compared to the targets and the objectives. For instances, he asked how these did compare with other African countries and South Africa’s trading partners.

Mr Lehohla responded that there were two levels at which this question could be addressed. Firstly, the differences between the LFS and the QLFS were very minimal in terms of the rates of employment and unemployment. In terms of the Millennium Development Goals, and the objective to halve unemployment by 2015, Stats SA would measure and see if this had been done, but it was up to policy-makers to answer whether this would happen. Secondly, he addressed how South Africa compared with other African countries on the continent and elsewhere, and with its trading partners. There were differences in employment compared to Brazil. The proportion of the unemployed was the same in all sectors except agriculture. In Brazil a good percentage of the employed were in the agricultural sector whilst in South Africa this number was negligible. Compared with the rest of the continent, there was a huge informal sector and agriculture on the continent. However there was no comparable breakdown in terms of industrial sectors such as mining with South Africa on the continent.

Ms Masiteng responded that South Africa was no exception to the high levels of unemployment that were characteristic of most developing countries. However, a lot of work had been done internationally to improve the collection labour market statistics, to enhance an understanding of unemployment. The organisation would compile a report that would shed more light on the nature of employment.

Ms J Fubbs (ANC) asked whether the questionnaire would reveal those instances where respondents were inclined to identify themselves as unemployed if they felt that their earnings were not remarkable or substantial. She asked whether there was a way that the questionnaire determined if there was any income at all in such households.

Ms Mpetsheni responded that there was no information on earnings collected in the questionnaire. There was no cut-off point in terms of income levels in order to determine that a respondent was unemployed. The determination of whether one was unemployed was arrived at after a series of questions such as “Did you do any work last week?” If the answer was no, then it was asked whether the person had looked for work in the last four weeks. If the answer was yes, then the person would be asked if he was available to work in the last week, and if the answer to that was “yes”, then the person was deemed to be currently unemployed. With regard to any other sources of income, there was a question that asked respondents to state how they supported themselves. A module would be introduced in 2009 to enable more information to be collected on earnings.

Ms Fubbs requested clarity on whether the questionnaire envisaged certain situations that were not clearly seasonal employment. For example, a person might prefer to work during peak periods of employment, such as during a fair or show, or at any other times that certain activities or events took place at which they could sell their labour.

Ms Mpetsheni responded that the questionnaire adequately dealt with such scenarios through questions that determined if respondents were economically active or inactive. Respondents were subjected to the same criteria, despite what they might conceive of as peak employment periods or seasonal work. She responded further that there were guidelines for the classification of people as unemployed. For example a question would be asked to determine the reason for absence from employment in the past week. Many reasons could be given for this, and many were contemplated in the questionnaire, and did include seasonal work. As soon as a respondent indicated seasonal work, he or she would be referred to the questions on unemployment, since there was no guarantee that the work would be available in future. However, the questionnaire still enabled statistics on the number of people who did not work because they were waiting for seasonal work to be captured.

Ms Fubbs asked a question pertaining to the retail sector. She observed that according to media reports, the retail sector was under serious pressure with respect to jobs. She wanted to know if there was any time in the survey that it was possible to focus on the high unemployment that would arise as a result of external factors such as this.

Ms Mpetsheni responded that because there was now a continuous quarterly survey, it was possible throughout to track changes between industries throughout the year. This would assist in identifying individual trends.

Mr M Johnson (ANC) said that owing to the great amount of work that was put into generating reports and statistics of this nature, it was important to assured of the reliability of the outputs. He expressed the concern that during the data-collection process, field workers had a fair amount of discretion to determine what would go into the survey. There was also a fair amount of suspicion about the reliability of survey methods, such as dropping questionnaires in people’s post boxes.

Mr Lehohla responded that the organisation believed in self-reporting despite its limitations. It appeared that self-reporting was a consistent methodology. The issue of what was objectivity and what was truth was a philosophical one. The organisation believed and trusted in the goodness of human beings to rely on their self-reported data.

Mr Mnguni asked whether statistics could be distorted by migration between provinces.

Mr Lehohla responded that such a phenomenon of exported unemployment could be possible. For instance a drop in unemployment in the Eastern Cape could cause an increase in the Western Cape. It would be helpful to factor migration in the questionnaire.

Ms Fubbs asked whether the unemployment rates could be disaggregated in terms of structural, cyclic or frictional unemployment according to industry sectors.

Ms Mpetsheni responded that the retail sector had contributed to the loss of jobs in the Western Cape as well as agriculture. The release provided by the organisation contained a geographical spread of employment by industry and region.

Mr Lehohla added that this was possible. The relationship between the structural unemployment and the sectors where it existed could be established. This relationship could, however, be indirect and could be tenuous in certain sectors.

Mr Johnson asked about gender equality and the recognition of unremunerated labour in terms of domestic work and home-making. He wanted to know how this information would be represented in statistical information. He was particularly interested in the unemployment rates of African women and whether these considerations would have an impact.

Ms Masiteng responded to the issue around trends amongst African women in terms of the high rate of unemployment and low labour force participation and absorption. She pointed out that these rates were high as a result of a number of factors, including the high population of African women relative to other groups and this created a higher denominator. This contributed to the higher rates for unemployment. For instance the total number of unemployed women in South Africa was estimated at 2.2 million, or 26% of the total unemployed working population. Out of the 2.2 million, 31% were African women. She was of the opinion this could have been as a result of structural inequality that had affected their livelihood and restricted their opportunities for labour market participation. The absorption of African women into the labour force was moving at a slow rate and this demanded policy intervention.

Ms Mpetsheni responded that the organization had the information on unremunerated labour value. Women who were doing housework were not considered as employed. The organisation had what it referred to as satellite accounts, which would enable a survey on what people did with their time. There was a need for further study to be able to determine the labour value of such activities as housework to determine how much should be paid for services such as child care.

The meeting was adjourned.


  • We don't have attendance info for this committee meeting
Share this page: