Statistics South Africa Annual Report: briefing

Meeting report

FINANCE PORTFOLIO AND SELECT COMMITTEES: JOINT MEETING
21 May 2003
STATISTICS SOUTH AFRICA ANNUAL REPORT: BRIEFING

Chairperson: Ms Q Mahlangu (ANC) [NCOP]
Co-chairperson: Ms B Hogan (ANC) [NA]

Documents handed out
Statistics South Africa Presentation

Relevant news articles
City Press article 11 May 2003: Stats SA bungling cost public R200m Leaked report `unfortunate'
City Press article 18 May 2003: That leaked report was merely a draft
Business Day article 7 May 2003: Rentals weighting may have distorted inflation details
Business Day article 21 May 2003: Markets await Stats SA's next move

SUMMARY
The presentation focused on the progress made and the challenges facing Statistics South Africa in producing economic, social and population statistics, and the challenges in governance and investment. The Statistics South Africa budget 2003/4 was also outlined.

Members raised the following matters: clarity on the measures put in place to record the activities and trends of the new economy, how Stats SA will ensure access to economic information from local and provincial spheres of government, the reasons for the delay in the release of the CPIX and Census 2001 figures and the policy implications of these delays, the rationale behind discontinuing the annual October household survey, the age groups and provinces most affected by the HIV/AIDS virus, the extent of Stats SA's consultation with relevant government departments to receive accurate information on the actual causes of HIV/AIDS related deaths. Stats SA was also asked to comment on a recent news article about a leaked report.

MINUTES
Presentation by Statistics South Africa
Mr P Lehohla, Stats SA Statistician-General, conducted the presentation (document attached), which focused on the progress and challenges facing Statistics South Africa with reference to economic, social and population statistics, problems and challenges in statistical processes, governance, investment in the future as well as the total Statistics South Africa budget for 2003/2004.

Discussion
Mr K Moloto (ANC) stated that the presentation indicated that Statistics South Africa is having problems with integrating its databases. What measures does Statistics South Africa intend putting in place to ensure that activities in the new economy are recorded, or that trends are followed in the new economy?

Mr Lehohla responded that Stats SA has made progress in that its current database contains about 700 000 businesses. On this basis a whole range of issues of governance and Black Economic Empowerment (BEE) can be dealt with, as access can be provided to track directors of companies and thus deal with BEE. More importantly then is how this is appropriated adequately between Stats SA and the South African Revenue Service (SARS), who takes responsibility of it, and who ensures that the updates occur. This is where Stats SA is experiencing a challenge. The Minister will soon be calling for a meeting of all Ministers involved, so that "a big push forward" can be put on this issue.

There are challenges in recording the new economy, and Stats SA itself is weak in understanding this new economy. There are also challenges posed by infrastructure and access to the new economy as well.

Ms A Myburgh, Stats SA Executive Manager: Programme Co-ordination and Management Information Systems, added that Stats SA is presently using a Value Added Tax (VAT) register as its main source for drawing samples for business statistics. Stats SA has been working pretty closely with SARS on this, and the sample business frame is updated. The main concern around the sample is classification, especially the difference between a legal and statistical classification. This is due to the fact that Stats SA needs results by industries, whereas this is not a major concern for the Receiver of Revenue. Stats SA has thus been working very closely with SARS this year on improving on the classification of businesses.

A major exercise which still has to be conducted is "profiling". In this exercise Stats SA "beefs up" its knowledge of the large complex businesses that cut across the various industries in South Africa. Many personal visits are conducted to classify those businesses into the various industries in which they function, as and when changes occur. Stats SA has to ensure that this happens on a regular and ongoing basis. It is currently taking place, but the process has to be sped up. So far Stats SA has drawn samples from the new business frame starting from July 2002, and Stats SA plans to redraw samples in July 2003.

In 2002 Stats SA was running a series of parallel surveys, and it was thus comparing its results from the older frames it was using with the new database. These new comparisons have not yet been published, but it is preparing to do so. The hope is that the parallel sample can be dropped once this has been done, so that Stats SA can then move on to just using the new business frame. A number of parallel comparisons have been conducted, and it does seem as Stats SA is able to pick up the newer industries much better through the new business frame rather than the old business frame.

The VAT registration still poses a problem, because Stats SA is missing out on whole businesses that are smaller but that are not VAT payers. This includes the informal sector, such as subsistence agriculture. The only way in which Stats SA has been able to reach those businesses to date is through household surveys. The labour force survey has been Stats SA's main tool for reaching those businesses in the informal sector. A survey of agriculture covering the entire country has been conducted for the Department of Agriculture, and these small businesses were included. This survey probably has to be repeated, if and when funding becomes available.

A household-based survey was also conducted in which the labour force survey was used to identify households from which small businesses are run, whether they be in the formal or informal sector, which were not registered for VAT. The distinction used by Stats SA was non-Vat registered businesses. This was the first survey of its kind that was conducted in South Africa. It was a very beneficial exercise and has to be repeated on a regular basis.

The methodology used contains three components: the first is the labour force survey, which identifies the businesses. The second phase is to go back to these businesses and get much more detail on them, and to find out exactly how they function through the households. The third phase is a consumer survey, in which Stats SA finds out what people buy from where and why they buy from that businesses. This has not yet been done, and Stats SA has not planned to do so at this stage. But when this is done, it would really provide a complete circle. Yet this time the information would have to be received on the business owner side, rather than from the consumer side. This is how Stats SA is trying to address the gaps in the register.

Stats SA will soon come out with a survey covering the economy-wide quarterly financial statistics, which should be released by the end of May 2003. This would be the first survey based on the new register, which covers the various sectors.

As far as employment is concerned, the Employment in Earning survey is not complete in its coverage. The sample that has been drawn is better in coverage, and any VAT-registered business would have the probability of being drawn into that sample. Stats SA plans to release the results of those comparisons for the labour force in the formal sector in the near future.

Mr Moloto asked Stats SA to explain the measures it would put in place to ensure that people can benefit from information on the economic activities in the local and provincial spheres of government.

Mr Lehohla replied that Stats SA really needs to increase its sample sizes, and more frequent surveys are needed. A better plan for the collection of statistics is needed. The frequency, availability and timeliness (FAT) principle of these is also important. A more regularised cycle is needed that would grant certainty in these processes.

Ms Myburgh added that Stats SA can, with sufficient confidence, disaggregate on its present sample to the level of provinces. But at a lower level one has to look at whether people want disaggregated data, by district municipalities for example. The survey sizes are not large enough because many of the more detailed surveys employ smaller cells, and some of the cells are likely to be empty. Thus is this is to be done Stats SA has to think through what kind of sample sizes are needed, and the methodologies it should be using.

The interesting thing arising from the statistics are the seasonal variations in work patterns. For example, in the agriculture sector there are times when unemployment rates are higher or people are not being as economically active. This is due to the cycles in agriculture. The extent to which Stats SA is covering these patterns is not sufficient, as it is not currently understanding this adequately. The informal sector is being covered by the non-Vat registered household survey, which was mentioned earlier, but Stats SA needs to get some idea of when people can and do not go into informal businesses. This is important because it is very easy for them to move in and out of those businesses.

Ms S Nqodi (ANC) stated that there seems to be much confusion from the South African Reserve Bank (SARB) on the determination of the targeted inflation measure, CPIX (consumer inflation less mortgages = Consumer Price Index. Stats SA has now also contended that it will be postponing the release of the result of Census 2001. What are the policy implications or concerns of this delay?

Mr Lehohla responded that the Census results have not been postponed. The "constipated system" had to be dealt with, which resulted in a three month delay. It is thus not a postponement.

As far as the policy implications are concerned, certainly it would have benefited the Stats SA budget and the like if the Census results were published earlier. The process will probably be completed just in time for the provincial allocations, and it will most likely be completed by July 2003 in time for the Cabinet Lekgotla.

There has been a postponement of the CPIX because of an internal review that Stats SA is currently conducting, which is aimed at highlighting the issues and problems. Stats SA is still poised to deliver this report by 30 May 2003.

Ms Myburgh added that Stats SA was made aware that it may be over-estimating the CPIX because of rents. Stats SA was running a survey amongst rental agencies in 1997, but that survey was dropped in the same year. The questions were added to the October 1997 household survey. The household survey in turn was discontinued in 1999 and some of the questions were not brought forward into the labour force survey, so that employment could be measured more accurately. Stats SA was not aware of the relevance of the questions on what people were paying for rent on the labour force survey.

Stats SA was thus using the data from the October 1998/1999 household survey for the CPIX, and its use initially had very little impact. The main assumption was that the rate of increase in rent has been constant over the time covered by the survey. This assumption has been queried, with the result that the CPIX requires careful review to gauge the impact of the old data collection.

A Member of the Stats SA delegation explained the methodology used in the past by stating that a household survey was conducted in October 1997, and the information became available at the end of 1998. It was thus applied to the CPIX as of 1 January 1999. The information of the household survey of October 1998 became available at the end of 1999, and was implemented in the CPIX from 1 January 2000. The information on the October household survey of 1999.became available at the end of 2000, and was implemented from 1 January 2001. The problem thus really is that from 1 January 2002 Stats SA had no information on the matter, because the October 2000 household survey did not take place.

The Co-Chair stated that she had been informed in the past that the October household survey was discontinued due to lack of funds. Could a more detailed reason be provided? The discontinuation seems to be quite a serious error because it has left a gap in the statistics for certain years.

Mr Lehohla responded that the October household survey was discontinued due to lack of funding. A whole range of data items fell through the cracks with the October household survey, one of which was the issue of the rentals.

Ms Myburgh replied that Stats SA had set priorities that at that stage getting better information on the labour market was a higher priority than getting the general information from the October household survey. Thus the October household survey was replaced with the labour force survey. This has caused numerous problems, and in 2002 Treasury did provide money to Stats SA to conduct a general household survey. Unfortunately in this general household survey the residential rents question was not asked, but Stats SA is doing a general household survey again this year. Stats SA has therefore restarted the general annual household survey, which will not be conducted in October but in June/July. It was decided that the best time to have these surveys would be during the school holidays. Stats SA does have the data from this general household survey, but it wants to adjust it in terms of the population count from Census 2001 before the final version is published.

As far as collecting rents from the general household survey is concerned, Stats SA is a bit worried about this. The prime reason again is the sample size. For example, in the urban areas that are approximately half of the sample, only 15-20% of people are renting properties. This proportion of people renting properties can then be divided into the various types of properties rented, and can be even further divided into the size of the rooms etc. But Stats SA does not have a very good database on which to base these rents. The thinking is thus that it might be better to reintroduce the survey that was dropped in 1997, which involved the collection of data from estate agents on a much larger sample than the October household survey.

Ms Nqodi asked what age groups are most affected by the HIV/AIDS epidemic, and which provinces are mostly affected?

Mr Lehohla replied that Stats SA is currently working towards capturing the full data sets. This information will then be played against the geography to identify which provinces are most affected. The data is there, it is just the capturing that now has to take place. It is in fact the capturing that can become a bit expensive, because Stats SA's data collection methods are a bit archaic. When any attempts are made to photocopy a death certificate from its microfilm, the quality of the copy is so degenerated that it is illegible. A photograph of the microfilm then has to be taken, but taking a photograph of data that has already been captured is an expensive exercise. Stats SA needs to look at its methods of collecting data from the Department of Home Affairs. The microfilm route has to be done away with, as it is unnecessary and expensive.

Dr G Gule, Stats SA DDG: Population Statistics, added that the reports published by Stats SA during 2002 focused on a 12% sample of deaths between 1997 and 2001. As a result, Stats SA is now focusing on completing the rest of the data, the 86%. The major problem is that Stats SA is not moving as fast as it would like to with this process, due to the limited funding. Stats SA would need about R3-4m to move as fast as was envisaged during 2002, in order to fast-track that project. Yet the allocation for 2003 is only about R700 000. Stats SA is thus using very limited resources in an attempt to address the backlog, but things will not move as fast.

The problem with this is that if Stats SA relies solely on the sample data, it would not be able to process the disaggregation. For example, if one were to look into the people that are most affected by HIV/AIDS, the study indicates that they would primarily be women aged 15-34. But Stats SA is not able to look at the extent to which children under the age of five are affected for example, because of the small numbers. Thus the only way to get a better understanding of what is happening in these different patterns of the age of mortality is to look at the bigger data set.

Stats SA is commencing this process by currently looking at 2001. People have been employed to start the coding of the causes of death, which is really the most challenging aspect of the project. Stats SA is collaborating with Medical Research Council (MRC) and the Department of Health to ensure that Stats SA has the correct classification in terms of the ICD 10 programme.

The study did find that amongst the leading causes of death, apart from HIV/AIDS related deaths, also included influenza, tuberculosis and pneumonia. These were actually related to HIV/AIDS. Unfortunately because of the way the death certificates are formatted by the Department of Health, Stats SA is not able to determine the underlying cause beyond the information provided.

The other challenge is that Stats SA relies on certificates which it does not really administer itself. The certificates are received when they have already been completed, and it is thus really dependent on the medical practitioner who has completed the certificate. This is a problem especially in the rural areas, because sometimes there is no way of knowing which medical practitioner has completed the certificate. There are thus circumstances in which there is very little that Stats SA can do to verify the cause of death, except perhaps working with the Department of Health to revisit the completion of the certificate.

The Co-Chair asked Stats SA explain its request that it be awarded R3-4m for the compilation of a causes of death database. Why has it only been allocated R700 000 for this important project? It appears that it is another important project that might be discontinued due to lack of fund, such as the October household survey.

Mr Lehohla replied that Stats SA's work on this has been largely slow. An additional R10m was granted to handle just the sample. The coding also takes up a large amount of time, as well as the taking of photographs. Stats SA is certain that it has to provide the denominators, and the population census, the geography and the registration of businesses are the critical denominators that are needed. Rapid change has occurred in recent times with regard to statistics, and these have to be reviewed.

Dr Gule added that one of the challenges facing government is the principle of inter-departmental collaboration does not progress to the next step of having integrated budgets. These will facilitate the carrying out of government's various priorities. For example, during 2002 a government department had promised that it would bring funds to the table for Stats SA to look into the causes of death, in order to address the backlogs. At some point Stats SA is also limited by the MTEF processes, because plans are made three years before the time. But if a matter suddenly arises and has to be dealt with, government departments do not always have the necessary funds to cover the costs of the survey.

Mr D Hanekom (ANC) thanked the delegation for the detailed presentation and responses. Statistics on the causes of death are very important for deciding on a health policy. The deaths due to tuberculosis and pneumonia for example, cannot surely be coincidental and have a definite effect on this. The relationship between those deaths and HIV/AIDS also has to be considered, as well as the role of other factors such as the resistance to treatment. To what extent does Stats SA co-ordinate with the relevant government departments in order to get the most reliable information available on the actual causes, and what are the difficulties encountered here?

Mr Lehohla responded that Stats SA's current data on this is scanty and it needs coherent in acquiring this data.

Dr Gula added that Stats SA is looking at hospital statistics, because these have not really been focused on in the past. The Department of Health also collects much information from the surveillance side, and this also has to be collected and analysed. There is also the upcoming Demographic and Health Survey which is being conducted by the Department of Health this year, in collaboration with Stats SA. All these efforts will allow Stats SA to have a better of picture of the health of South African citizens.

Ms Myburgh added further that the key tool here is the national statistics system, as there is a need to look at different ways in which all departments collect statistics. Environmental methodology has become part of the national accounting system. Stats SA started with air pollution account, but the methodology is still being learnt. It is very expensive to do, but Stats SA is moving in that direction. Thus Stats SA is trying to cover this via the satellite accounts on environmental accounting.

It is very difficult for Stats SA to conduct pro-active surveys when funds are allocated in a certain way, and to then change track for the year. Thus the budget processes do make it a bit difficult, even though Stats SA may want to do this.

Mr Lehohla added further that it is difficult to be pro-active, because the major concern is always who will pay for the survey. In fact, government departments have even refused to pay for the commissioned study. This has hardened Stats SA, which may not be a good thing.

Dr Gula added that it is very difficult to be pro-active. Stats SA might have good intentions here, but the problem is that it does not have the necessary money to carry it out. At times Stats SA's eyes are too big for its stomach, and perhaps it should focus on the tasks currently before it.

As a result Stats SA is not able at this point to provide an answer as to exactly how many of the tuberculosis deaths were attributable to HIV/AIDS, because it does not currently have the instruments to determine this. The competence of the medical practitioner is also important here, because they may not check for all the symptoms. It would not be wise to rely on assumptions here, but instead a modeling framework has to be devised.

Stats SA is working with the National Skills Fund (NSF) on the issue of co-ordination between government departments, and all efforts are being made to arrive at a common set of viable standards to be employed. This forms part of the Stats SA's whole improvement of quality drive.

Mr Hanekom asked whether Stats SA, additional to its very in-depth research, also conducts "rapid surveys" in response to urgent public issues or policy issues.

Mr Lehohla replied by stating that there are pressing issues of public policy that arise from time to time. When Departments do have funds to pursue this they do commission Stats SA for a specific project. The funding could be the biggest issue. The quality assessment framework standards for the work conducted have to be decided at the outset.

Dr W Odendaal (NNP) asked what the accuracy rate is for the population, unemployment and immigration rates.

Mr Lehohla replied that the total of 45 million South Africans relates to the media estimates, which are based on a set of assumptions on the number of people born, die, emigrate and immigrate to South Africa. The information on illegal immigrants is not easy to ascertain. It could be assembled from the Department of Home Affairs and SAPS, but even then the methodology to be used is difficult to define. As far as the population census is concerned, Stats SA can only indicate the number of South Africans born and the number of foreign nationals.

Ms Myburgh added that Stats SA tries to ensure as much accuracy in its sampling methodology as it can. It does regularly calculate confidence limits and standard areas, which are published with each of the labour force surveys. Within 1,5% either way Stats SA is confident that, at the 95% confidence limit, that its unemployment rates are accurate. Unemployment statistics for the informal sector are much harder to define, because its very nature people move in and out of the sector very quickly. Stats SA needs to get a better understanding of the relationship between the formal and informal sectors.

Dr P Rabie (DA) stated that he had asked Stats SA during the teabreak whether most of the social statistics were obtained via household surveys. He stated that Dr R Hirschowitz, Stats SA DDG: Quality and Integration, explained to him that this in the international norm.

Dr G Koornhof (ANC) asked whether Stats SA has already responded to the recommendations made by in the Financial Fiscal Commission (FFC) Report to Stats SA regarding the measurement of poverty.

Mr Lehohla replied that the World Bank Mission will be at the offices of Stats SA for the following four weeks to look at statistical audits on this. Ms Myburgh added that Stats SA certainly will take account of their recommendations. Stats SA has been looking at various measures of poverty not only expenditure poverty, and the hope is that Stats SA will be able to do extra poverty mapping when the Census results come.

Ms R Joemat (ANC) asked whether Stats SA is satisfied with the state of financial management.

The Chair stated that she had read the article in the City Press dealing with human resources and the lack of proper management in the Stats SA. It exposes a study that was done by Stats SA but which was fraught with problems. Could Stats SA comment on these issues?

Mr Lehohla responded to these two questions by stating that the Auditor-General's Report which indicates a disclaimer of R122m has to be looked at, over and above the over-expenditure. Stats SA conducted a census in October 2001, and employed 100 000 throughout the country. Stats SA spent about R293m for paying those employees and it could have sat back and not declared that money during the year of expenditure, which would probably have violated the PFMA. It's alternative was to declare the amount that had been used and then subsequently deal with the reconciliation. Stats SA was thus sitting between a rock and a hard place. Not declaring the money was a lesser evil, but Stats SA was morally obliged to inform the State of the monies it had. Yet this then triggers an audit, which would conclude that Stats SA was unable to reconcile its data. This is in fact what happened in relation to the disclaimer received from the Office of the Auditor-General.

The Auditor-General was correct in issuing the disclaimer because Stats SA failed to reconcile its accounts, but it is not a disclaimer which indicates that Stats SA does not know what it is doing. Stats SA's response to this issue was thus that it had employed 100 000 people, it broke down the figures in terms of the exact payment for each category. Thus in the broad sense Stats SA knew exactly how much it had paid, it was just that the reconciliation was not done properly. In the next two months Stats SA would have fully reconciled its work, which will then be re-submitted to the Office of the Auditor-General.

The report leaked to the City Press is an unfortunate situation. The decision to look at the human resource records was based largely on the fact that the pensions were are at times not paid on time. The investigation missed out on the transformation process that was put in place, which introduced an organisational developmental scheme which will lead how Stats SA arrives at that pensions structure. In December 2002 or January 2003 the Public Service Commission (PSC) was consulted on the steps to be taken in the restructuring process, and it has issues Stats SA with a letter allowing it to continue with the process. The Minister also approved the restructuring and Stats SA has since began to appoint people.

The Gobodo Report stated that Stats SA "set the structure before strategy". This was not the case, because Stats SA had been engaged in a six-year long process which bore the seeds for the strategy that was finally adopted. Thus the premise of the Gobodo is false and the facts that surround it, including the allegation that Stats SA employs more consultants than core employees is "absolutely incomprehensible". Stats SA has about 3000 consultants. It is true that their total number is three times the size of the organisation, but they are employed probably for a period of one or two months. In the Census process about 1000 people were employed for approximately one year, and it is these people that the Gobodo Report refers to as consultants.

The Gobodo Report also suggests that Stats SA has promoted people unduly. There are people within Stats SA that have been fully trained in the statistical field and have been sponsored by Swedish missions. Yet despite all this training they earned only
R36 000 per annum. They were then graded as being Grade 8, and it was decided that there is no reason to keep them back because they are producing good work. Yet the Gobodo Report states that Stats SA has "over promoted". Mr Lehohla stated that he is thus not sure what the statistics in the Gobodo Report refers to. It really was an unfortunate report.

Mr Lehohla stated that when he received this Godobo Report he instructed the Human Resource division to look into this matter, because the statistics in the Report are not correct. Unfortunately while this verification process was being conducted, the Report with all its inaccuracies was leaked to the media. This is the story of the mismanagement in Stats SA. Mr Lehohla stated that he is not suggesting that Stats SA is that clean, as there are one or two employees who are currently involved in legal actions.

Ms Myburgh added that Stats SA does not have permanent people to do the surveys, not even at a regional level. Stats SA does have a database, and when there's work it goes around and asks people to help with the surveys. They are given the best training possible and Stats SA works as best as it can with them. But as household surveys increase Stats SA has to start employing permanent people to do the Census and for surveys, especially with the change that the regional offices could be introduced. This will ensure much greater accountability.

People are employed during a survey on a temporary basis, but they were never employed on a full time-basis. The best people who worked with Stats SA on the data processing during the Census were then referred to SARS when the data processing had been completed. Some of them have subsequently been employed by SARS. Stats SA thus tries to find alternative employment for people when it lets them go. The reality of the matter is that Stats SA is always in the situation in which it will at one time need many people, then fewer at another stage and then many at a future stage. This is the nature of the work. Other statistics agencies also have this kind of semi-permanent force. It is a problem, but perhaps Stats SA should start to look at having this as a permanent structure.

The Chair stated that she understands that employing 100 000 people over a short period is no joke. Would it not be possible for Stats SA to establish some kind of relationship with the Independent Electoral Commission's (IEC), to work on a common systems basis for the 2004 elections? Surely the IEC would have similar problems with paying volunteer staff at large numbers during the elections. This would enable both to draw on a common pool of managerial and systems issues.

Mr Lehohla responded that Stats SA did collaborate with the IEC quite extensively in 2000 on how to manage these operations. The IEC recommended a group of consultants who left some quite residual value with the IEC, and Stats SA engaged the same consultants to its benefit. This brings some predictability and continuity in both areas of work, as people can then arrange at what point they exit one activity and enter into another. This ties into the volunteerism principle and the work suggested by the President in his State of the Nation Address. Even though they earn comparatively smaller wages, they can still receive the necessary skills and enter the labour market in a much more organised way.

Mr B Mnguni (ANC) asked whether the increase in the Stats SA's personnel expenditure is related to the training given for skills development, the use of consultants etc.

Mr Lehohla responded that this is in part largely due to these consultants employed by Stats SA, which ranges from 1000 to 3000 at times. This increases Stats SA's wage bill by about 464%. The questions then is whether they in fact bring the value bargained for. Sadly at the end of this process Stats SA then has to inform them that it no longer requires their services. Others, however, are absorbed into Stats SA. Stats SA is currently looking into a strategy, in consultation with the Trade Unions, in which it recruits people from this pool into the permanent structure. Thus the increase in personnel in that year is as a result of the personnel that were employed to do the Census.

In fact, Stats SA only has about a total of ten consultants. The other enumerators employed on a temporary basis have been mistakenly referred to by the sensational Gobodo Report as consultants.

The Chair asked whether SARB was consulted about the delay in the release of the CPIX figures, because it has to be informed of this information. To what extent is Stats SA working with SARB to ensure the release of these figures from time to time? How does Stats SA work with other agencies as well?

Ms Myburgh replied that Stats SA did discuss this with one of the deputy governors of SARB, and they also met before the publication was delayed. Stats SA also met with Treasury and the Statistics Council. Thus all the role-players were informed of the decision to postpone the CPIX release this month.

Stats SA does work with SARB. South Africa has signed the International Monetary Fund Special Data Dissemination Standard, which means that nobody can get the information before the time ahead of anybody else. Stats SA has been pretty vigilant in adhering to that. The actual data on the CPI is released to the Minister and the press at the same time, but the discussion on the methodology is ongoing. A balance has to be found between working with other agencies and independence when announcing the results, so that they cannot be seen to be influenced by anybody.

The Chair stated that no mention has been made of the Stats SA's time use survey which looks at how women use their time, as compared to men. How has this contributed to the understanding of the role that women play in the economy?

Ms Myburgh responded that the main contribution of this survey is that Stats SA could calculate a satellite account based on certain assumptions, which gave some indication of the value of unpaid workwomen, and their contribution to the overall economy. The study has to be repeated and monitored over a period of time. It showed that at least 30% of the total GDP may be underestimated because the unpaid contribution of women is not taken into account. It was funded by a Norwegian agency.

Ms L Mabe (ANC) stated that she would like to correct the misperception that most emigrants are only from other African States. But this is not the case. How is this reflected in the immigration statistics?

Mr Lehohla replied that this is a difficult issue, but Stats SA would probably be able to analyse employees by country of origin in its analysis. The systems of registers for those entering the country would also be important here. This would enable one to establish the country of origin, but verifying their legality would be a more difficult exercise. This would have to be dealt with by the agencies that specialise in this, and Stats SA could then consider their data.

In conclusion, the Chair emphasised that the release of the Census 2001 is really important. This is so especially for the "trimming up of figures" in the budget process, because the formula to be used to determine the division of nationally raised revenue will be informed largely by the product of the Census 2001. Stats SA has grown, but as the presentation indicates it is still experiencing many challenges. Yet by working with Parliament, government departments and respective agencies, its leadership will be able to help Stats SA grow from strength to strength.

The meeting was adjourned.