Address by Trevor Andrew Manuel, Minister in The Presidency: National Planning Commission to Parliament on the Budget Vote of Statistics South Africa - Vote 13

15 May 2013

Chairperson

Honourable Members

Ladies and Gentlemen

Last year when we participated in the Statistics SA (Stats SA) budget vote debate, we thanked everyone in the country for having participated in Census 2011 the previous October and at that point, we were awaiting the results. The Census 2011 results have since been released on 30 October 2012 in both the standard formats as well as in the ground-breaking iPad tablet format that some members here may have seen. Whilst the purpose of the debate here today is to discuss Budget Vote 13 - Statistics South Africa, it is crucial to ensure that the discussions here have context. So, the appeal to Parliament in the course of this debate is to ask a series of questions about the resources available to Stats SA and about the quality of these and other outputs being published.

For the outputs of Stats SA to have relevance for development, Parliament needs to advance the discussion about the manner in which available statistics are being utilised for evidence-based policy-making. I want to emphasise that the responsibility of Stats SA is solely to ensure that the necessary statistics are available, verifiable and that they meet internationally set standards. Stats SA supported by the Stats Council and the Executive Authority must ensure that the manner in which the data is presented is above reproach.

One of the significant breakthroughs from the publication of Census 2011 last year is the form and style of data availability. Last week, members of the Standing Committee on Finance were provided with a demonstration of the availability of detailed statistics at ward level using the My Ward, My Councillor section of the Stats SA tablet application. We should point out that the availability of this huge wealth of data in this format is without precedent and places Stats SA in a position where it is without peer. The same application is now being used for new releases and so the quarterly GDP release was made available in this format, the Mortality and Causes of Death information was made available in this format, last week the first Quarterly

Labour Force Survey of 2013 was availed in this format and so was the Mid-Year Estimates release yesterday. We want to submit that in doing so Stats SA is discharging its obligation to ensure the timeous dissemination of statistics as prescribed in section 3 of the Statistics Act (Act 6 of 1999). However, despite the ease of use and the level of detail, availing all of this data in this contemporary format appears insufficient to spark curiosity or interest from legislators, policy-makers or the public. It is this that Parliament must take a view of because once the data is available it is no longer a Stats SA responsibility but is over all of us who care.

Before I get into the substance of this debate, it may be important for me to reiterate the explanation provided to the Standing Committee about the definition of employment used by the Quarterly Labour Force Survey. It would appear that, despite our explanations at the meeting last week, there is still some misunderstanding about the information and figures that Stats SA releases.

Stats SA reinstated the expanded definition of unemployment when the decision was made to expand the labour force survey to a quarterly one in 2008. However, the headline figure that is released as the official figure is based on the International Labour Organisation (ILO) definition where someone is considered unemployed if they meet the following criteria:

1) ‘Without work' i.e. were not in paid employment or self-employment as defined by international definition of employment

2) ‘Seeking work' i.e. have taken active steps in a specified recent period to seek paid employment or self-employment

3) ‘Currently available for work'

This adherence to the ILO definition is to ensure that the current labour market situation is measured objectively and that the measurement can be used to assess South Africa in relation to other countries. We need to bear in mind that this definition works well under circumstances where the working population engages in paid employment and where the channels for exchange of labour exist and are widely used.

However, the ILO recognises that the definition does not take into account instances where a large portion of the population is engaged in subsistence farming, the labour force is largely self-employed or the labour market is largely unorganised or of limited scope. Stats SA recognises that, there could be impediments to searching for employment for some sections of the population.

The cost of looking for a job is often so exhorbitant, relatively speaking, resulting in increasing numbers of discouraged work-seekers. Thus, it reports on both the strict unemployment rate and the relaxed (expanded) unemployment rate. So, if any member chooses to look at the information available either on the website or by using the app, they will have access to a complete breakdown of unemployment in terms of race, gender, geographical location, age-group, industry, etc.

The notion of having more than one unemployment rate is not unique to South Africa; for example, the United States has six measures of unemployment with only two being official numbers. We need to recognise that the two unemployment rates in South Africa are not in conflict but rather that they provide indicators for different components of the economy.

Chairperson, while it may seem that I have digressed somewhat from the Budget Vote under discussion today, it is imperative that we understand the importance of maintaining public trust in the outputs of Stats SA. If this budget is to be approved by Parliament, we must address not only the issue of resources but also the quality of the outputs. I want to remind the House of the mandate of Stats SA as outlined in the Stats Act, specifically section 3 (1) that states:

(1) The purpose of official statistics is to assist organs of state, businesses, other organisations or the public in

(a) planning;

(b) decision-making or other actions;

(c) monitoring or assessment of policies, decision-making or other actions.

The definitions used by Stats SA therefore have wider implications for planning and assessing the state of our economy as a barometer within the country and internationally. While the QLFS measures unemployment using a quarterly household survey of 30 000 households, it is not the only measure. Stats SA releases Quarterly Employment Statistics (QES), that measures the level of employment in all non-agricultural sectors.

It does so by undertaking quarterly surveys of about 20 000 (20 208 to be exact in the latest figures released on 19 March 2013) vat-registered private and public enterprises. In an ideal situation, the results of these two datasets should tend to be similar and which is why definitions or in the case of the QLFS, self-definition becomes a factor. The strategic focus of Stats SA is to ensure that the work of these two areas is more closely integrated in future to provide for an improved understanding of employment.

In the current financial year, there will be a continued focus on improving the output of a range of equally important datasets that concentrate on the changing demographic profile and health of the population, poverty levels, and income and expenditure trends. These are managed through the Poverty and Inequality Statistics, Health and Vital Statistics and Demographic Analysis sub-programmes.

In addition to these regular series relating to social statistics, both government and the private sector rely on key economic data published by Stats SA. Chairperson, during my speech last year I referred to the implementation of significant changes in economic statistics, notably in the form of a reweighted and rebased consumer price index (CPI). In line with its program of continuous quality improvement, the CPI was updated to reflect more accurately changes of prices that affect the lives of households. This refinement assists in ensuring to ensure that the South African Reserve Bank has accurate information at its disposal when it makes interest rate decisions. In addition, the Producer Price Index (PPI) was completely overhauled to align it with international benchmarks. The new suite of five PPI replaced the single index to allow analysts to provide a better understanding of the transmission of prices through the economy. The new 2012 base ensures that the PPI reflects the dynamics of the economy more accurately. In line with the announcement last year, Stats SA implemented those changes in February 2013. In addition, other improvements on further releases such as improved seasonal adjustment techniques were implemented during the year. They will continue to monitor the impact of these changes to evaluate its efficacy.

For South African statistics to remain valid in a changing global context, the next challenge for economic statistics will depend on their ability to ensure that their classification systems remain relevant and current. In this regard, Stats SA will continue to explore the implementation of two international standards during the course of the current financial year. The first is the 2008 system of national accounts (SNA). At present, South Africa's national accounts and therefore its economic statistics are estimated according to the 1993 SNA, the international framework. The new SNA was developed by IMF, OECD, World Bank and adopted by the United Nations Statistics Commission to ensure that economic statistics keep track of new economic phenomena that have become important in economies in the last 20 years. An example of this is the need to quantify the value of research and development and include it in estimates of value added in the economy.

The second set of international standards is called the International Standard Industrial Classification (ISIC) used by statisticians to divide the economy into different industries such as agriculture, mining and transport according to an agreed standard. The ISIC was updated to version 4 in order to identify and emphasize new aspects of economic activity that have become important, especially in the services sector.

Ideally, these two standards should be implemented simultaneously and Stats SA is faced with the challenge to ensure that they continue to measure the dynamics of the economy accurately while ensuring that our ability to compare with our international counterparts in OECD, BRICS and other trading partners does not become compromised.

The responsibility of ensuring that official statistics are published in terms of distinct standards extends beyond those statistics compiled by Stats SA. Section 14 of the Statistics Act provides Stats SA with the responsibility of co-ordinating statistics among organs of state. This includes advising departments about improving the quality of statistics, enhancing the comparability of statistics and minimising any possible overlapping or duplication with the collection or publication of statistics.

Chairperson, in the past I have raised concerns in this House about the effect of our poor educational outcomes. The reality is that a scarcity of maths and science skills in this context must be understood in terms of its impact not only on the human resource challenges faced by Stats SA directly but more significantly in terms of the impact on statistical collection and processing by other organs of state. Ideally, the skills set would exist across departments with the system for managing the quality and standards being with Stats SA. This unfortunate reality, however, does not absolve Stats SA of its statutory responsibilities and in this regard, a more hands-on approach is often required. An example of such intervention is a census of schools in Limpopo and the Eastern Cape Province being undertaken at the request of the Department of Basic Education and the National Treasury. The objectives are to verify the total number of schools within the provinces, establish the number of school learners and educators per school, establish the education stream and subject offerings of educators and learners and to establish the qualifications of teachers in relation to the subjects that they are teaching. It is important that we understand that while Stats SA is responsible for statistical co-ordination, the capacity to compile data and use statistics to influence policy actually rests with, in this case, the provincial departments of Education. The challenge of a shortage of relevant skills and the importance of statistical training cannot be emphasised enough.

The approach to the lack of statistical skills cannot simply be a resigned fatalistic one, and in this regard, Stats SA has established strategic partnerships with various universities including the University of Stellenbosch, University of Cape Town and the University of KwaZulu-Natal.

Internally, Stats SA has adopted an approach to addressing the skills shortage by implementing a talent management model that includes recruiting twice as many interns as required. Their skills are developed and fostered within Stats SA for a period of two years, including providing them with workplace experience by placing the interns with mentors and direct supervisors. Stats SA piloted the internship programme during the 2005/06 financial year when fifteen graduates from different South African universities were recruited into the programme. In the 2013/14 financial year, the programme has grown to 53 interns. This approach recognises that while the institution will ultimately not employ many of those being provided with these valuable skills, many will also be taken up by the private sector where the skills are in demand.

Stats SA remains committed to build statistical skills throughout our country by developing a statistical curriculum that has been incorporated into the maths curriculum at schools. In addition, they have implements a programme that builds statistical capacity across government.

The value of developing skills in order for Stats SA to meet its strategic mandate can only be measured in the quality of the datasets being published and the public trust placed in Stats SA. It is precisely because of this focus on the issue of public trust and reliability of data that we need to raise the issue of the single number registration under the Business Register Reform. This is an issue that has been raised both by the Standing Committee on Finance and in this House over the past few years. As mentioned on numerous occasions, the launch and management of such a system of registration requires co-ordination and correlation of three disparate systems of records managed by the South African Revenue Services (Sars), the Companies and Intellectual Properties Commission (CIPC) and Stats SA respectively. Members of the House will also be aware that the CIPC has undergone several changes and challenges in recent times including the implementation of new provisions of the Companies Act.

Notwithstanding the challenges, the first phase of the system of a single Business Register has been initiated as from 4 May 2013. This phase deals with the registration of new businesses by the CIPC. The re-registration process for existing companies is expected to take place by October 2013 when companies will file details with both Sars and the CIPC during the tax filing season. The new forms have been formulated using International Standard of Industrial Classification (ISIC), which will provide Stats SA with the basis to begin to verify classification and disaggregate multiple classifications where it exists. As the verification process takes place, Stats SA will establish feedback loops to test the veracity of the information being captured. In addition, Stats SA intends running an ISIC survey to act as a further test of accuracy. Taking into account these requirements to verify the precision of the system and the levels of classification, Stats SA has indicated that they will require a period of 12 months to test whether the data being collected is fit for statistical use in terms of the South African Statistical Quality Assessment Framework. A comprehensive business register is an imperative for quality economic statistics as it forms the sampling frame for all economic surveys.

While the Statistics Act clearly designates Stats SA and in particular, the Statistician-General as responsible for the collection, production and dissemination of official and other statistics, it is important that we make the point about the necessity for other sources of information and the need for correlation. In addition to the requirement that data produced by Stats SA is accurate, verifiable and timeous, it is crucial that the data is internally verifiable as well. I have mentioned the correlation between the QLFS and the Quarterly Employment Statistics in this regard. Equally, if there is no functioning Business Register or system of reporting, the task of sampling is more challenging and the QES is constrained. When there are such conditions, it merely affords naysayers the opportunity to question the veracity of the data. We must ask Parliament to join us in ensuring that the releases by Stats SA become and remain the only reliable sources of statistics in South Africa. It should not just be the preferred but the only reasonable, trusted and verifiable statistics. It must be able to allow every South African to say, ‘The South Africa I know, the home I understand'

While South Africa is the home we have to understand, we are part of a wider community on the continent and globally. Stats SA has established partnerships with statistics institutions that are part of the African Union Commission, the United Nations Economic Commission for Africa, the African Development Bank, and Regional Economic Communities. Stats SA plays a leading role in initiatives such as the Africa Symposia on Statistical Development (ASSD), the Statistics Commission for Africa (Statcom-Africa), and the African Statistics Committee and the Africa Group on Statistical Harmonisation (AGSHA). One of the significant strides on the continent has been the successful mobilisation of all African countries to undertake population censuses in the 2010 Round of Population and Housing Censuses. We can only assess development and growth on the continent if we have evidence to measure changes and it is crucial that as a continent we have appropriate, trustworthy statistics based on the same standards. So building international partnerships becomes important to ensure that there is adherence to benchmarks and that skills are exchanged.

Finally, Chairperson, allow me to comment very briefly on the negative audit outcomes of the 2011/12 budget i.e. the budget of the census year. As the Members of this House are aware, the qualified audit was largely on the basis of accruals due to non-payment of service providers that then flowed into the following year.

While I am fully cognisant of the impending Scopa process, I raise this issue principally in the context of any future planning. It is important to understand the reality of the magnitude of the payments that were being processed and the limits of the government ICT systems to manage such numbers. Government systems are not designed to cope with such a large numbers of invoices and are not integrated to account for intergovernmental transactions. We call on Members of this House to discuss these challenges with a view to future similar large surveys. This call is made particularly in light of recent appeals to Stats SA to conduct a census every five years as required by the Stats Act.

In conclusion, I would like to take this opportunity to thank the Standing Committee on Finance ably chaired by the Honourable Thaba Mufamadi for their keen engagement with the issues and for their unswerving support of the work of Stats SA.

In addition, it is important that I take this opportunity to give a special thanks to the outgoing Stats Council for the sterling work done of safeguarding the quality of our official statistics. We need to recognise the significant role they played in steering Stats SA through the immense task of conducting Census 2011 as well as the support they provided to address the challenges faced with the release and the verification of the data.

I thank you.

Issued by The Presidency, May 15 2013

Click here to sign up to receive our free daily headline email newsletter