Anglia Ruskin University and Hughes Hall, Cambridge, UK
Artificial Neural Networks forecasts for Internet Market concentration In Tamil Nadu: Arguing for the need of Algorithmic Prioritisation
Emanuele Giovannetti and Sebastian Sigloch
Internet upstream market concentration is challenging to be measured, and its evolution difficult to be predicted, as it relies on complex networks of bilateral interconnection agreements. This paper compares the predictive performance of Artificial Neural Networks vs traditional regressions in capturing the emergence of bottlenecks in the upstream Internet connectivity networks. The comparison is performed on a large Mobile Internet connectivity data set collected by the authors in Tamil Nadu using Portolan a smartphone-dedicated application (Sigloch, 2018). The collected data give rise to large networks of complex connectivity interactions that can only be meaningfully analyzed using complex network metrics. These metrics, Centrality and Clustering, allow to capture the emergence of critically relevant operators that form bottlenecks in the upstream internet access, and transform these in market power (D’Ignazio and Giovannetti, 2016). Individual interconnection decisions, giving rise to the global Internet connectivity, are managed by dedicated interconnection algorithms such as TCP- IP or BGP that may adapt in real time. These protocols, while building the end to end nature of the Internet, also allow the probing and exploring of the Internet and can be used by competition authorities to build prioritisation policies based on algorithmic criteria. While the debate on algorithmic pricing and collusion is well established (See OECD, 2017, Ezrachi, and Stucke, 2016), this contribution focuses on algorithmic prediction of interconnection structures and its impact on market power suggesting a novel way for regulatory authorities to prioritise cases, based on Al
Emanuele Giovannetti (PhD, MPhil Cantab, Trinity College) is Professor of Economics at Anglia Ruskin University, Governing Body Fellow at Hughes Hall, University of Cambridge, Vice-Rapporteur for the Study Group 1 of the International Telecommunications Union, Developments Bureau and a full member of the Regent House. Prior to this, Emanuele was Associate Professor in Economics at the University of Cape Town and the University of Verona and Senior Research Associate at the Department of Applied Economics of the University of Cambridge where he led the FP7 project Cocombine (Competition contents and broadband for the Internet in Europe). His research focuses on Market power in the Internet, Mobile Internet access, ICT platforms, Digital divide, Crowdfunding, Diffusion of mobile social networking and Adoption of new technologies. He has advised governments, competition authorities and businesses in Europe, Africa and Asia on Internet access and competition economics, and has led multidisciplinary and multinational projects focussing on network competition, Internet infrastructure and mobile access in developing countries. Emanuele published on leading academic journals including International Economic Review, Economic Journal, Environment and Planning A, International Journal of Industrial Organization, International Journal of Production Economics, Journal of Industrial Economics, Technological Forecasting and Social Change, International Journal of Forecasting, Review of Network Economics, Information Economics and Policy, Spatial Economic Analysis, Journal of Economic Surveys. He edited a special issue of Telecommunications Policy on “Peering and Roaming in the Internet” and co-edited “The Internet Revolution: A Global Perspective” published by Cambridge University Press.