Conference Agenda

Extended Abstract

A fundamental debate is ongoing regarding EU competitiveness and its triggers. It can be promoted through either more cooperation (mergers and acquisitions) or competition. This highlights the potential of 'coopetition', such as network sharing, as a compromise between the two.

The general problem is that, although regulators acknowledge the social welfare benefits of coopetition over full-scale cooperation, they assess and clear it using a similar horizontal merger guideline approach. This means that countrywide common mobile network sharing agreements are often not approved. In high-density urban areas, regulators still favor parallel infrastructure-based competition over service-based competition on shared infrastructure.

We focus on the Hungarian 4G LTE 800 MHz network sharing that received no regulatory clearance for 8 years from 2015, but has been running in unchanged form. The incumbent market leader, Magyar Telekom entered into a network sharing agreement with Telenor Hungary, the second market player. The agreement is limited to the rollout of 4G LTE 800 MHz across the whole country, except in the capital, Budapest. This agreement is a Multi-Operator Core Network (MOCN) with shared passive, active and spectrum elements. The competition investigation concluded in 2023 without any findings or recommendations relating to the past or the future (GVH, 2023). This was 1.5 years after the European Commission (EC) had approved a similar network sharing agreement in the Czech Republic, including recommendation for 5G deployment (European Commission, 2022).

Our research question focuses on whether the potential harm to social welfare posed by the Hungarian 4G network sharing scheme justified the 8-year competition investigation that concluded without issuing a concrete recommendation and prevented the rollout of 5G.

This controversial case enabled us to conduct a quantitative, empirical analysis. Our research methodology is based on multivariate regression and applies a difference-in-differences fixed-effects panel model. Using this model, we compared the highest level of network sharing (MOCN) in Hungary with other global and European countries where operators have a similar MOCN or a lower level of sharing (Multiple Operator Radio Access Network, or MORAN, where only passive and active elements are shared), or no sharing at all. Through these peer group comparisons, we aim to demonstrate that the impact of the Hungarian network sharing case is in line with European trends, as evidenced Koutroumpis et al., (2023). We apply the extended dataset and set the time period of pre-2020 window, when the effects of the Hungarian sharing's economic and technological run rates are best measurable.

The systematic literature review covers a range of multidisciplinary topics, including transaction cost economics (TCE), coopetition, competition policy, and horizontal production agreements in the telecommunications sector. The review uses the TCE model presented by Hennart (1988) to powerfully explain coopetition. Motta structured the competition policy aspects; Motta and Tarantino (2021) developed the key horizontal production agreement theoretical model; Bourreau et al. (2020) and Pápai et al. (2020) collected the recommendations; and Földes (2023) prepared the qualitative case study for Central and Eastern Europe (CEE). Empirical quantitative models were performed for the Czech Republic by Maier-Rigaud et al. (2020) and for the EU by Koutroumpis et al. (2023) and Koutroumpis and Masselos (2024). These models establish the baseline of the literature for this study.

The novelty of our paper lies in its aim to provide empirical, quantitative, model-based evidence on the impact of Hungarian network sharing on prices, investment, 4G coverage and network quality in the context of the dynamic competition approach, which is welcomed by the National Competition Authority (GVH).

Our research findings suggest that Average Revenue Per User (ARPU) increased modestly, however in a data-traffic-driven market, rather negative correlation is seen between ARPU and price. Cost savings feeding profitability (EBITDA) and, in turn, investment (CAPEX) moved broadly in line with peers. The empirical center of gravity of the paper was that the efficiency gains from sharing were primarily passed on to consumers through investment in network quality (i.e. speed), enabled by shared spectrum of 800 MHz (MOCN). As a consequence of the rural band limited scope for sharing, improvements in 4G coverage and data traffic per connection lagged behind those of peers. Evidence of market concentration provides no support for theories of harm.

Our assessment indicates that the delay in regulatory closure on the rural band limited network sharing discouraged operators from enhancing network sharing for 5G and 2-4G lifecycle swap that could have enabled full grid consolidation for scale, which was subsequently rolled out without coopetition. The resulting foregone efficiency gains may, in turn, have reduced social welfare.