The extensive use of mercury, a persistent contaminant, in the artisanal and small-scale mining (ASM) sector is a growing environmental issue. Mounting concern over findings which point to the widespread use of mercury to amalgamate gold in ASM having significant environmental and health-related impacts has fuelled the launch of numerous donor-funded and government backed projects aimed at minimising its use. Most have taken the form of educational and technical assistance, implemented specifically to stem emissions from the industry and introduce miners to safer practices. These interventions, however, have had minimal impact. In an attempt to understand why the 'mercury pollution agenda' has yielded so little in the way of improvement, two divergent viewpoints have emerged. The first of these explanations, a view prevalent in donor and regulatory circles, is that miners ignore messages about the dangers of mercury and are reluctant to adopt more environmentally-friendly practices. The second body of opinion contests that approaches taken to educate and provide training on mercury management to communities are flawed, failing to take stock of the successful aspects of previous work and sufficiently eng~e target communities. There is insufficient empirical evidence to support either view, however. A broadened understanding of operators' perceptions toward the environmental impacts of amaJgamation is a necessary first step towards improving the policy context for mercury management in the ASM sector. This thesis helps to bridge this gap, focllsing on the case of Ghana, the location of one of the largest ASM industries in sub-Saharan Africa. It determines levels of awareness and prevaJent attitudes towards mercury use amongst small-scale gold miners, capturing a level of detail not yet undertaken. It also identifies barriers that prevent miners from adopting improved mercury management practices. To address these issues, a mixture of qualitative and quantitative analyses were undertaken: interviews with key institutional stakeholders; individual interviews, group discussions and feedback sessions with miners; and an environmental assessment. This research provided a comprehensive understanding of the dynamics of mercury use in Ghana's ASM industry.

In this dissertation the technical efficiency in gold mining is investigated. To the best available knowledge, this is the first such study on gold mining, whether on a localised (one country) or for a cross-section of countries. Since the work by Farrell (1957), much work has been done using nonparametric methods such as DEA. Although extensions in DEA technique, such as bootstrapping have been available for some time, their use has been limited in comparison with the number of overall DEA studies carried out. In this dissertation both DEA and bootstrap DEA are applied to two gold mining cross sectional samples, one on Zimbabwe consisting of thirty-four mines, and the an international one which also included some Zimbabwean mines which comprise fifty-nine observations. The main reason for carrying out the study is an interest in gold mining in general and its importance to Zimbabwe in particular. As will be noted in Chapter 2, the economic development of Zimbabwe has been linked, to a varying extent over the ages, to its growth of the gold mining sector. The results of the dissertation provide some useful insights into the relative performances of gold mines and also some characteristics of the Zimbabwean gold mining sector. The main results indicate that gold mining is characterised mainly by technical efficiency dominating scale efficiency. This is particular relevant when the Zimbabwean mines are compared with their international counterparts. Zimbabwean mines are found to be relatively technically efficient but less so when overall efficiency is considered. In fact they have the lowest overall efficiency scores in the international sample. The results also indicate that mines from the so-called developed mining economies, Australia, Canada, the US and South Africa are the benchmarks in terms of optimal operations. It is mines from these countries which define the overall efficiency frontier. The results of both the samples highlight potential shortcomings in applying DEA and bootstrap extension to gold mining, both for single country and for cross-country cases. Additionally, there are possibilities, with adequate data, of relating country-specific characteristics to differences in overall efficiency among countries. Finally there are indications that including mineralogical factors such as the recovery rate in the production technology has an effect on technical efficiency. Mines with low recovery rates tend to exhibit comparatively higher technical efficiency. The study does have some limitations, mainly because of lack of data. In particular, there were problems in coming with attributing the contribution of capital services to efficiency with the result that a different measure for the flow of capital services is used for each sample. In addition, the two samples are for different time periods. This limits comparative analysis.