Net neutrality rules have been implemented in many developed countries, often in response to concerns over ISP market power and potential blocking or throttling of content (Greenstein et al, 2016 and Easley et al, 2018). However, developing countries have significantly lower levels of internet penetration and usage, and ISPs charging content providers in violation of net neutrality may result in lower prices to consumers and greater adoption of internet services (Kramer et al, 2013 and Easley et al, 2018). Furthermore, the case for ex-ante net neutrality regulation in markets with competition among ISPs is not clear (Crocioni, 2011 and Kramer et al, 2013). Market power in respect of internet access looks quite different in developing countries, given that mobile is the predominant means of connection and there are often three or more mobile operators. In certain developing countries, including South Africa and Kenya, fibre to the premises is being rolled out, in some cases on an 'open-access' basis. A further consideration in developing countries like South Africa and many other African countries is that there is a quasi-monopoly in the paid broadcasting market which is not vertically integrated into internet services (the same satellite TV provider, Multichoice, broadcasts across much of the African continent). Competition from mobile operators in offering their own content and that of third parties (such as Netflix) at lower data prices may help to bring about more competition in markets for paid TV. This means that developing countries may have quite different policy objectives where net neutrality is concerned, depending on market circumstances. First, we describe the market structure and dynamics in ISP and content provider markets in South Africa. Next, we investigate the theories of harm in more detail. In terms of the ISP market we investigate whether conduct in violation of net neutrality such as throttling and blocking has been taking place and whether it is likely to in future. We explore data on the number of announced prefixes and peers and IP addresses originated in order to get a sense of the size and market power of ISPs and the relationships between them. We then assess competition at the content provision level including websites, social media over-the-top services, video on demand and broadcasting. We consider examples of bundling and zero-rating conduct by ISPs which appear to be the result of vibrant competition and beneficial for consumers. We provide conclusions in a final section.
