Small area estimation of unemployment for South African labour market statistics
| dc.contributor.author | Hakizimana, Jean-Marie Vianney | |
| dc.date.accessioned | 2012-02-23T12:53:30Z | |
| dc.date.available | 2012-02-23T12:53:30Z | |
| dc.date.issued | 2012-02-23 | |
| dc.description | M.Sc., Faculty of Science, University of the Witwatersrand, 2011 | en_US |
| dc.description.abstract | The need for Official Statistics to assist in the planning and monitoring of development projects is becoming more intense, as the country shifts toward better service delivery by local government. It is evident that the demand for statistics at small area level (municipal rather than provincial) is high. However, the statistics with respect to employment status at municipal level is limited by the poor estimation of unemployment in 2001 Census and by changes in boundaries in local government areas. Estimates are judged to be reliable only at provincial level (Stats SA, 2003) The aim of this study is to investigate possible methods to resolve the problem of the misclassification of employment status in Census 2001 by readjusting the data with respect to the classification of people as employed, unemployed or economically inactive, to that of Labour Force Survey of September 2001. This report gives an overview of the different methods of small area estimation proposed in the literature, and investigates the use of these methods to provide better estimates of employment status at a small area (municipal) level. The application of the small area estimation methods to employment status shows that the choice of the method used is dependent on the available data as well as the specification of the required domain of estimation. This study uses a two-stage small area model to give estimates of unemployment at different small areas of estimation across the geographical hierarchy (i.e. District Council and Municipality). Even though plausible estimates of the unemployment rate were calculated for each local municipality, the study points out some limitations, one of which is the poor statistical representation (very few people) living in some specific municipalities (e.g. District Management Areas used for national parks). Another issue is the poor classification of employment status in rural areas due to poor data with respect to economic activities, mostly with respect to family businesses, and the non-availability of additional auxiliary data at municipal level, for the validation of the results. The inability to incorporate the time difference factors in the small area estimation model is also a problem. In spite those limitations, the small area estimation of unemployment in South Africa gives the reference estimates of unemployment at municipality level for targeted policy intervention when looking at reducing the gap between those who have jobs and those who do not. Hence, the outcome of the small area estimation investigation should assist policy makers in their decision-making. In addition, the methodological approach used in this report constitutes a technical contribution to the knowledge of using Small Area Estimation techniques for South African Employment statistics. | en_US |
| dc.identifier.uri | http://hdl.handle.net/10539/11343 | |
| dc.language.iso | en | en_US |
| dc.subject | Labor market (South Africa) | en_US |
| dc.subject | Unemployment (South Africa) | en_US |
| dc.subject | Unemployment (South Africa, statistics) | en_US |
| dc.subject | Social sciences | en_US |
| dc.title | Small area estimation of unemployment for South African labour market statistics | en_US |
| dc.type | Thesis | en_US |