Unemployment and abject poverty in South Africa are widespread, persistent and disproportionately high. These problems are aggravated by inadequate capacity at all tiers of government and huge infrastructure backlogs in rural South Africa. Employment-intensive means of infrastructure delivery have been successfully implemented in several sub-Saharan African countries and elsewhere to generate employment and reduce poverty. It should be possible to replicate similar large-scale national programmes in South Africa. But South Africa has failed. Both prior to and since 1994, several supposedly employment-intensive programmes have been implemented in South Africa with poor results. In an endeavour to contribute to future South African policy and good practices regarding employment, this thesis describes and analyses in detail five major Sub-Saharan programmes and reaches conclusions regarding their achievements and shortcomings. Lessons derived bridge the knowledge gap between the large-scale programmes embarked upon in the sub-Saharan countries in the mid-1980s and the year 2007; these lessons should be applied to future endeavours in South Africa to generate significant employment per unit of expenditure and contribute to poverty alleviation. A major conclusion reached was that the success of employment-intensive infrastructure development programmes depends to a large extent on fundamental factors such as appropriate and implementable policy, government commitment, adequate and sustainable funding, adequate capacity and good preparation. Specifically, the thesis demonstrated that prior to implementation a sufficient timeframe is required for programme preparation in order to make significant contribution towards poverty reduction. Equally, national programme expansion requires a strategic balance between centralisation and decentralisation. In particular, for programme extension and decentralisation, due regard must be given to training and capability building and available resources. Deriving from the thesis’ major conclusions, the author developed three crucially important frameworks for anti-poverty infrastructure development programmes, namely; a four-phased model for evaluating the chances of success of infrastructure programmes, a five-phased result-oriented guidelines for testing the workability of infrastructure development policies, and a practical guideline for monitoring and evaluating employment-creation programmes that maximises the benefits thereof and pre-empts institutional memory loss through systematic knowledge management.

Employment programmes, or on-the-job training programmes, are designed to increase the likelihood of unemployed workers receiving work opportunities by providing actual work experience. Whether the unemployed workers lack human resources, obtained a degree in a foreign country or simply need a boost towards obtaining a new job, these programmes give workers the opportunity to execute tasks at organised firms and therefore strengthen their connection to the labour market.  The focus of this paper is towards a unique programme, called Arbetspraktik. This specific programme is designed with the intention to increase the labour market outcomes of unemployed workers. Geographically, Växjö has been selected as the main priority. The local focus is due to two reasons; firstly, the internship at the Swedish Employment Service in Växjö provided local support, insight and experience and therefore estimating local effects matched the purpose of the internship. The result of this thesis may also be used as informative research for the Swedish Employment Service where I performed my internship. Secondly, according to previous work, specific regions within a country are not looked at on a high frequency in Sweden and therefore focusing on regional effects could be beneficial to already existing studies. Previous research shows that the evaluation of these programmes has been done in many countries. Data for this study is obtained through the Swedish Employment Service. With the use of Matching and Propensity Scores, the findings claim that participants in Arbetspraktik have a 28.3% higher probability of receiving better outcomes than the controlled counterpart in Växjö. The interviews, despite having low observations, indicate that positive outcomes for trainees are a possibility and may be connected to the outcomes of the model. For example, trainees are shown to become more independent and productivity across trainees increases over time. Lastly, the interviews claim that trainees perform the same tasks as full-time workers, which is discussed to be a positive thing. Self-critique and tips for future research are presented at the end.

The last 15 years have seen governments in a number of mature welfare states attempting to reintegrate people out of work for reasons of sickness and disability into employment, principally through changes to the value and conditions of incapacity benefits and the provision of active labour market programmes. Whilst the academic interest in these changes has been considerable, this thesis begins by arguing that these studies hitherto have been satisfied to categorise these emerging regimes according to a familiar Work-first v Human Capital Development activation typology (for example, Peck & Theodore, 2001), or a variation upon that, according to the presence or absence of different activation services. They largely do not apply the insights that the broader activation literature has provided in recent years, particularly those on the governance of activation. Instead, this thesis proposes that it is better to examine recent changes through the lens of institutionalisation: how well-embedded employment-related support for sick and disabled claimants has become in the structure and functioning of welfare-to-work regimes for sick and disabled benefit claimants. Though not a concept much used in academic analysis of Active Labour Market Policy (ALMP), a case is made for the value of looking at, firstly, how well activating sick and disabled claimants becomes a national government labour market policy priority and secondly, how well the organisation and governance of active labour market programmes for this group support this, in additional to analyses of the services themselves. Working from what is already known about the factors that can influence a workless benefit claimant's access to employment support, the contention of such a framework is that the successful embedding of an activation strategy for sick and disabled claimants into national Labour Market Policy (LMP) is a function of the interaction of a range of factors. Crucial here is the distinction between ALMP for these claimants, and for other activation target groups – there is good evidence to believe that the changes made to activation governance to promote active work-search for the unemployed may, however unintentionally, militate against a comprehensive system of support for 'non-employed' jobseekers considered to be further from the labour market, claimants of incapacity benefits included. Alongside this framework, a case is made for being much clearer and more precise in describing what measures apply to which parts of the incapacity benefit claimant pool. In most countries, this is a very diverse population with several distinct sub-sets with different levels of distance from the labour market, ranging from those with very severe disabilities or health conditions; others with multiple employment barriers not all stemming directly from their condition (outdated skills, for example), and those whose employability is high, their disability or health condition notwithstanding. As a small number of studies have pointed out (Evans, 2001, for example), activation regimes – defined in this study as the set of services that are provided to help nonemployed sick and disabled benefit claimants back to work; and how these are organised; delivered; targeted and financed – 'sort and select' claimants, applying different types or more or less intensive support for different categorisations of claimants. An activation regime for the claimant group can thus be very inclusive or rather narrow, depending on the extent to which these sub-pools are catered for. To demonstrate the value of this framework in reaching a more accurate understanding of the nature of these emerging regimes relative to extant approaches, a cross-national comparison of activation of sick and disabled claimants in Denmark and the United Kingdom is offered. Whilst they are considered to be very nearly diametrically opposed in a number of key ways – their approaches to activation; benefit generosity and broader welfare regime contexts – when looked at using the institutionalisation framework, they emerge as more similar than expected. Regardless of their quite different starting points, they experience many of the same challenges in creating a system in which the employment activation of the full extent of the claimant group is a priority and where a sick or disabled benefit claimant's right to back-to-work support is secure.