This paper investigates the details behind aggregate shifts in household electricity access in South Africa. More specifically, when viewed from a cross-sectional perspective, we note a significant (and surprising) decline in electricity access between 2008 and 2010, followed by a substantial improvement in access between 2010 and 2012. In order to further investigate these interesting dynamics and move beyond a limited cross-sectional analysis, we then set up the National Income Dynamics Study (NIDS) in a novel form that allows one to track household units in a longitudinal fashion. Using this data, we identify the initial drop in electricity access to have come as a result of a large number of household disconnections, as well as a significant degree of "misdirected" household formation (with people leaving household with access and setting up households in locations without access). We also identify the subsequent improvement in aggregate access to have come primarily as a result of a significant fall in the number of households that lose access over the period, an increase in the number of households that gain access, and favourable household formation processes (with people leaving households without access and moving into households with access). It is therefore vital that those involved in coordinating service delivery take into account that, if one's aim is to improve aggregate electricity access, preventing loss of access is just as important as expanding access. Policy makers should also take note of household formation and dissolution processes when considering service delivery expansion - to prevent government from needlessly chasing a moving target.

Includes abstract.~Includes bibliographical references (leaves 39-41). / The withdwal of the Maize Board in 1996 meant that farmers could no longer rely on their pre-planting price or "voorskat" for price discovery and price risk management. Some have claimed (UNCTAD, 2007) that the South African Futures Exchange (SAFEX) can provide these functions. We test this claim and analyse the impliacation of it.

The poverty trap hypothesis postulates that very low income individuals may be trapped in poverty because severe constraints give rise to behaviour that limits their ability to take up and benefit from new investments. Furthermore, the theory suggests that insurance and credit can serve as effective tools in counteracting the mechanisms that create this type of persistent poverty. In this study, using data obtained from two separate samples of farmers in the Western Cape, we explore the validity of this hypothesis in a South African context. One sample consists of organic farmers from Cape Town whilst the other comprises of conventional farmers from the Matzikama Municipality. We elicit behavioural traits; more specifically risk preferences, using lottery type experiments with real money at stake. Using a series of logit regressions we look at the relationship between these preferences i.e. risk aversion, loss aversion and nonlinear probability weighting and actual uptake of farm technology in both samples. Using the Matzikama sample, we then apply the estimates to uptake in an experimental setting where insurance and credit are provided. The experimental study allows us to test for both absolute and path dependent effects by examining both the levels, using a multinomial logit model, and the timing, using a cox proportional hazard model, of uptake. The results from the real life data show that the effects of the risk preferences tend to differ depending on the type of technology, and this is true for both samples. One consistent finding from both the real life and experimental uptake data is that the farmers who live in households that have below average relative income levels are less likely to take up technology; even with insurance and loans being made available in the experiment. This finding is unexpected given that all the farmers face the same objective risk levels and do not have their real life income at stake in the experiment. Our results show that the availability of insurance improves uptake in the overall sample and can serve as effective tools in reducing poverty. However, contrary to the poverty trap hypotheses, little evidence is found to suggest that the insurance contract in the study sufficiently serves as a device to counteract the risk preferences that are linked to low technology uptake. This finding is evident when considering both absolute uptake and the timing of uptake. Therefore, the results on the effects of insurance and credit on technology uptake, given risk preference and relative income position, may imply that low income farmers in South Africa are not only constrained by behaviour that is prompted by monetary or risk factors but also other behavioural or psychological components e.g. the feeling of hopelessness that stems from persisting conditions of poverty; however, this requires further investigation.