Abstract
Protest, a vital signal of shared discontent, requires courage and coordination. Despite South Africa’s transition to democracy, it continues to experience numerous community and labour-related protests daily; by some claims, more than any other country. While state measures to curtail such gatherings are often fierce, a sharp rise in public protests has nonetheless been perceived since the early 2000s. Nevertheless, quantifying protests remains problematic, and the drivers behind protests – whether community-based, labour-related or some other type – are contested. Scholars hypothesise a range of factors in their explanations – some conflicting – for the apparent rise in protests. However, these commonly advanced theories lack quantitative rigour. Applying machine learning algorithms to the world’s largest publicly-available single-country protest event database, this thesis classifies over 150 000 police-identified incidents between 1997 and 2013, as protest or non-protest events, and the former by levels of “tumult”. Each of the 89 000 identified protest events is differentiated, considering count, count per capita, and “general propensity”, as well as “tumult” models; and then located within one of the country’s 234 municipalities. Combining that analysis with demographic, socio-economic, governance, and voting-behaviour official statistics, the study tests proxies for hypothesised factors associated with protest. It finds that “general propensity to protest”, representing the number of protesters per capita in the respective municipalities, is a more accurate measure to quantify protest than simply counting “protest occurrences” or protests per capita. Further, it finds that municipal-level Gini inequality, unemployment, voter turnout, and the dependency ratio, expressed as the proportion of people aged 17 to 64 in the population, offer the most explanatory power for protests in South Africa, a finding that is further disaggregated into community and labour-related protests...
D.Litt. et Phil. (Anthropology and Development)