سرمایه گذاری فناوری و رژیم های نظارتی جایگزین با عدم اطمینان از تقاضا

A vertically integrated incumbent and an OLO (Other Licensed Operator) compete in the market for broadband access. The incumbent has the option to invest in building a Next Generation Network that covers all urban areas with similar demand structures. The investment return in terms of demand increase is uncertain. We compare the impact of different access regulation regimes – full regulation, partial regulation (only the copper network is regulated), risk sharing – on investment incentives and social welfare. We find that, when the alternative for the OLO is using the copper network rather than leaving the market entirely, exclusion of the OLO does not necessarily happen in equilibrium even when the incumbent is better in offering value-added services. Risk sharing emerges as the most preferable regime both from a consumer and a social welfare perspective for a large range of parameters.

Telecommunications markets are experiencing a period of drastic technological development. The possibility to build a so-called Next Generation Network (NGN) gives firms the chance to exploit extremely faster transmission and thereby enrich their offer with more interactive and sophisticated services. However, the actual existence and importance of a demand for NGN applications is often uncertain.1 The technology has been available for a while now, but given the high fixed costs needed to build the necessary infrastructure, and the risks associated with demand and regulatory uncertainty, the NGN deployment is progressing slowly in many countries. The vexing issue as to how to provide firms with enough investment incentives, while eventually retaining the benefits of network development for final consumers, is highly debated by industry actors, regulators and scholars. In particular, access regulation is widely argued about its potential discouraging effect on regulated firms’ investment. When obliged to share its network elements with facilities-free rivals at a regulated access price, the incumbent may feel reluctant to invest in NGN because of the spillover effect enjoyed by the Other Licensed Operators (OLOs). For these reasons, access regulation, mainly in the form of mandatory unbundling, may induce less or later incumbent’s investment compared to an unregulated scenario, but also compared to the socially desired level (Chang et al., 2003, Crandall and Singer, 2003, Ingraham and Sidak, 2003, Bourreau and Doğan, 2005, Pindyck, 2007, Grajek and Röller, forthcoming and Nardotto et al., 2012). The European Commission seems to acknowledge these concerns for future investments in NGN. In the Recommendation C (2010) 6223 on “Regulated Access to NGANs” (September 2010), the possibility of relaxing – if not eliminating – ex ante regulation when a risk sharing agreement backs up the deployment of NGN is openly considered. The issue of broadband investment and regulation has attracted and still attracts a lot of research attention.2 Our paper contributes to this strand of literature by addressing the issue of access price setting when the incumbent has the option to invest in NGN and investment returns in terms of demand increase are uncertain. Using a model, where a vertically integrated incumbent and an OLO dynamically compete in the market for broadband access, we analyse the effect of three different access regimes on the incentives to invest by the incumbent: full regulation (mandatory unbundling for NGN), partial regulation (no mandatory unbundling for NGN) and risk sharing. We then compare their impact on social welfare, balancing the effect of each regulatory regime on static and dynamic efficiency. In our paper, we follow the original set-up of broadband investment and access regulation developed by Foros (2004).3 We develop a model with two firms having different ability to offer value-added services, and analyse the impact of access price regulation on the incumbent’s investment incentive. Differently from Foros (2004), however, we adopt a dynamic model of technology adoption and we include demand uncertainty for value-added NGN services. Considering that NGN investment might fail to expand market demand, we also assume that the OLO can possibly switch back to the copper network if there is no demand for NGN applications and the access to copper is cheaper. We then conduct our analysis comparing the impact on investment of three alternative access regimes. In this respect, the paper closer to ours is Nitsche and Wiethaus (2011). The authors analyse a simple two-stage framework with identical firms, where the incumbent is the only firm entitled with investment option and investment success in terms of demand increase is uncertain. Their work compares the impact of different modes of regulation (access price based on costs, risk sharing and regulatory holiday) in terms of investment extent and consumer welfare outcomes. There are several differences between our work and Nitsche and Wiethaus’s (2011) one. Firstly, in their model, following Klumpp and Su (2010), the access charge is determined ex post from the equilibrium quantities and it includes a partial allocation of the fixed costs borne by the incumbent. In our model, the regulator establishes ex ante the level of access price, via first-order conditions. As a consequence, the benchmark case for access regulation in our model is a marginal cost-based rule, as in much of the literature in this field ( Foros, 2004 and Kotakorpi, 2006 for instance). Secondly, our setting is dynamic and we investigate the timing of investment in a context with demand uncertainty, rather than the extent of the investment. Moreover, we are able to carry out a complete welfare analysis, whereas Nitsche and Wiethaus’s (2011) work only gives an overview of the different modes of regulation’s implications in terms of consumer welfare. Lastly, our model includes quality differentiation á la Foros and considers its impact on equilibrium results, while, in Nitsche and Wiethaus’s (2011) model, firms are undifferentiated. The impact of uncertainty on the timing of telecommunications infrastructure development has also been analysed in several papers that feature dynamic race models between incumbent and entrant operators and focus on specific access pricing regimes, mainly regulatory holidays (Hori and Mizuno, 2006, Hori and Mizuno, 2009, Gans, 2001, Gans, 2007 and Vareda and Hoernig, 2010). In our model, by contrast, we consider uncertainty in a dynamic setting, but we focus on services-based competition, while taking into account different possible regulatory regimes. Since the co-existence of the legacy network and the new fibre network is highly inefficient and not sustainable over a long period, we also analyse the case in which the switch to NGN is compulsory and compare it with the case in which the switch can be postponed and depends only on the access conditions. We find that, in case of compulsory switch, the OLO gets relatively worse access condition absent regulation. Our paper also differs from a recent strand of studies that analyse an investment game where both the incumbent and entrants have the option to invest. Brito et al. (2012) examine the incentives of a vertically integrated firm (regulated at wholesale level) to invest and give access to a new (upgraded) wholesale technology, that is not subject to access regulation. Bourreau et al., forthcoming and Inderst and Peitz, 2012a analyse the incentives to migrate from an old technology to a new one, and examine how wholesale access conditions affect this migration. Finally, Manenti and Scialà (2011) study the impact of access regulation on entrant and incumbent’s investment and show that, in absence of regulation, the incumbent would choose an access charge to the new infrastructure that prevents resale-based entry, thereby overstimulating entrant’s investment which may eventually turn out to be inefficiently high. Our model reveals that the differences in ability to provide value-added services and their absolute values with respect to the overall level of demand highly affect the investment choice. Since the OLO has the alternative to switch back to the copper network instead of leaving the market entirely, we find, in contrast to Foros (2004), that there are cases in which exclusion does not happen even when the incumbent is better in providing value-added services than the OLO. In case of mandatory switch to the NGN, we find that the OLO remains active in the market if and only if its ability to provide value-added services is higher than the incumbent’s one. The equilibrium results show that the investment is always undertaken later than in the social optimum and that the presence of uncertainty has the effect of delaying the investment even further. Full regulation lowers the investment incentives due to a hold-up problem of the regulator who exploits the irreversibility of the incumbent’s investment ex post. Due to a combination of competitive intensity and investment incentives, we find that risk sharing is the most preferable regime from a consumer welfare perspective, but also from a total welfare perspective for a large range of parameters. The remainder of the paper is organised as follows. Section 2 introduces the model and the main findings under the three different regulatory regimes. Section 3 summarises the paper and concludes.

In this paper we model the competition between a vertically-integrated incumbent firm and a facilities-free OLO in the broadband market, where the former has the option to invest in building a NGN that allows firms to drastically increase the quality and variety of their services. Market success of the NGN in terms of demand increase is uncertain. Differently from other studies that assume demand uncertainty, the investment choice is analysed in a dynamic setting with differentiated products. The analysis is conducted under three different possible modes of regulation: full regulation (the NGN is regulated), partial regulation (the NGN is unregulated) and risk sharing (fixed investment costs are shared but there are no side payments between firms for the use of the NGN). Our analysis reveals that the investment is always undertaken later than in the social optimum in all regulatory regimes. The investment choice is affected by the OLO’s ability to offer value-added services. Such effect is positive with partial regulation and negative with full regulation, while with risk sharing the effect changes from positive for high values of the OLO’s ability, β2, to negative as the incumbent’s ability, β1, gets considerably bigger than β2, and vice versa. Partial regulation always yields the earliest investment compared to the other regulatory regimes, while risk sharing ensures the highest level of competitive intensity. Welfare outcomes reveal that risk sharing is the dominant regime in a consumer surplus perspective. Expected consumer surplus is always higher under risk sharing than under partial regulation, but also under full regulation for a large set of parameters. In particular, when both firms are active, full regulation’s consumer surplus outcome is the least preferable; only when the incumbent’s ability is so high that the regulated access price to the NGN is above marginal cost, the comparison of outcomes in terms of consumer surplus between full regulation and risk sharing becomes ambiguous. Furthermore, when the OLO is better in offering value-added services, risk sharing is the dominant regime also from a total welfare perspective. When the incumbent is better, instead, welfare comparisons between the three regulatory regimes become less clear. It is worth pointing out that these results are valid for the reduced form of risk sharing that we have considered in this paper. Such form of risk sharing implies a long term contract with no side payments for the use of the NGN, thereby excluding sources of inefficiency from the market and increasing the level of competitiveness downstream. More complicated contractual forms of risk sharing might arise in reality, which could well make the welfare comparison with full regulation less favourable to risk sharing. In terms of policy recommendations, we can state that, if risk sharing works smoothly, as in the model, then it allows for welfare improvements compared to full regulation. While, indeed, looking deeper into how risk sharing works is worth additional future research. Our analysis sheds some conceptual light on the debate about what is the socially preferable access regulation regime to prompt telecommunications network development. The difference in firms’ ability to provide value-added services is important in the context. It exerts influence on the investment choice and on the access pricing decisions, which in turn affect market competition and social welfare. We find that demand uncertainty requires a careful formulation of access regulation rules. A robust set of rules should take into account the potential for an investment failure and provide reasonable access conditions for the firms involved in all possible cases. Also, uncertainty plays the role of delaying the investment decision in all regimes. According to our analysis, risk sharing can be particularly beneficial for consumers and give fairly high investment incentives at the same time. At this stage, it would also be interesting to go further in the research to study how risk sharing agreement can be robust to the inclusion of late entrants, to avoid that the construction of the NGN could possibly become a new source of market power and thereof be unable to deploy all of its benefits. It would also be interesting to make the choice to engage in a risk sharing agreement endogenous. We leave these questions for future research.