Presentation to the ICAS IV 2007 conference delegates

1. Presentation to the ICAS IV 2007 conference delegates Lessons learned from the 2002 Census of Commercial Agriculture Moses Mnyaka Statistics South Africa 22 October 2007

2. Contents Introduction Needs analysis Frame Tool/instrument used and data collection Data processing and imputation Lessons learned from the census Conclusion

3. Introduction • History of agricultural censuses in South Africa • Importance of the agricultural census • Factors affecting the success of the census • How did the census link to the national agriculture • statistics systems and the national statistical • system at large

4. Needs analysis • Decision to conduct a census was based on a • request to national Department of Agriculture • Other stakeholders, e.g. Department of Water • Affairs and Forestry; Department of • Environmental Affairs and Department of Trade • and Industry were not consulted • Government departments and other producers of • agricultural statistics work in silos • Suppliers (farmers) were not involved from the • beginning • No sufficient time was given for planning • The information collected was based on size of • the land, production, financial statistics, • employment statistics and personal detailsIn • South Africa

5. Frame • The frame was drawn from a list of enterprises • registered for Value Added Tax (VAT) • Measure of size was the annual turnover of businesses • Covered mostly formal businesses as informal had to • register voluntarily • Frame not integrated with other administrative to give a • full coverage of the sector • Approximately, 60% of the businesses registered onto • the frame were registered by bookkeepers, auditors • or accountants using their contact details • The frame was not maintained and updated on • businesses that have been liquidated, closed, sold or • merged

6. Agricultural businesses units registered for VAT

7. Breakdown of units that were not usable during the census

8. Tool/instrument used and data collection • Questionnaire designed for a postal census and • keyboard capturing • Questionnaire contained questions regarding all types of • farming activity • Input received from stakeholders through an advisory • committee meeting • Political instability, crime, lack of sufficient publicity led to • low response rate • Response rate was fairly good on large enterprises

9. Response rate according to stratum

10. Data processing and imputation • Data checked, edited and captured as each • individual questionnaire was received • Warning and inconsistency error systems were • built with in the capturing system • Non-responding units were dealt with by using • their annual turnover from the Business Frame • Imputation was done for the financial statistics • variables by using the ration imputation • method • Data about the size of land and production • was published without imputation

11. Lessons learned from the census • Planning for the census is vital as early as the • previous census is conducted • Availability of resources, financial and human (skilled • expertise) is crucial • Political climate and crime have an influence in • conducting the census • Other similar projects/surveys conducted over the • same population have an impact on the census • Overloading of the questionnaire affects the response • rate • Need for non-commercial farming data was reflected • by the non availability of data for GDP estimation of • forestry and fishing sub sectors • Clear concept and definitions not properly covered • affects are important for good participation by • respondents

12. Lessons learned from the census • Uncertainty regarding funding • Inclusion of questions related to small scale farming • Political climate and crime have an influence in • conducting the census • Other similar projects/surveys conducted over the • same population have an impact on the census • Overloading of the questionnaire affects the response • rate • Need for non-commercial farming data was reflected • by the non availability of data for GDP estimation of • forestry and fishing sub sectors • Clear concept and definitions not properly covered • affects are important for good participation by • respondents

13. Conclusion • Designing a strategy on national agricultural • system in conjunction with other role players • in the sector; • Using standard concepts and definitions • nationally; • Focusing on collection of financial and • employment statistics when conducting annual • surveys; • Strategy on the coverage of the secondary • economy; • Measuring of the sector to withstand bio-fuel • needs against food security; • Formulation of an Advisory Committee on a • permanent basis;

14. Conclusion • Coverage of businesses operating farming activity • under the non-agriculture sectors • Conducting a national Workshop on Agricultural • Statistics; • Keeping track of international best practice; • Advance planning and research work; • Bring on board users, suppliers and other data • producers of agricultural statistics

15. Conclusion These have been shown by recent requests to Stats SA, - by organised agriculture, that Stats SA co- ordinate all data collection survey in the farming communities, - by National Accounts for a new survey on environmental studies, - by Dept of Water Affairs and Forestry on water usage, - by Dept of Land Affairs for a survey on eviction of farm dwellers, - by conducting for the first time a census of agricultural services parallel to censuses of forestry and fishing

16. Moses Mnyaka Statistics South Africa, Agricultural Statistics 170 Andries Street Pretoria 0001, South Africa mosesmn@statssa.gov.za