Workshop SIEPI 2013 January, 24 th -25 th, 2013 New regulatory tools and investment in broadband networks Carlo Cambini Politecnico di Torino & Florence.

New regulatory tools and investment in broadband networks

Carlo CambiniPolitecnico di Torino & Florence School of Regulation – [email protected]

Roadmap Fixed broadband in Europe:

Facts and Figures Ultra-broadband networks (Next Generation

Access Networks – NGANs) deployment and financing

Governments initiatives

New regulatory issues for the NGANs

Some Economics of NGAN Regulation: technological transition, geographical markets and co-investment

Fixed Broadband in Europe

Investment in telecommunications infrastructure is believed to be a significant contributor to economic growth ever since a long time:

Röller and Waverman (2001, AER): An increase of 10% in the broadband penetration rate leads on average to an increase of 2.8% of GDP growth (21 OECD countries).

Koutroumpis (2009, JTPO): the average impact of broadband infrastructure on GDP is 0.63% (for the EU-15, in the period 2002–2007).

Czernich et al. (2011, EJ): a 10% increase in the broadband penetration rate results in 1-1.5% increase in annual GDP per-capita. Faster broadband = higher GDP growth.

Still, the EU is lagging behind in high speed broadband infrastructure investments.

4

An IntroductionAn Introduction

NGANs take up in Europe (Source: Digital Agenda Scorecard, June 2012)

Policy context: Digital Agenda for Europe

One of seven lead initiatives of the Europe 2020 strategy with the aim of overcoming the crisis

Fast and ultra-fast Internet access a key pillar. Targets:

2013: Broadband for all Europeans 2020: 30Mbps for all Europeans (50% of

subscriber lines ≥100Mbps)

EU Financing for Broadband Estimated cost: €38-58 billion for 30Mbps full

coverage by 2020

€181 – 268 billion to provide coverage so that 50% households are on 100Mbps services

EU financing for BB in 2007-2013 - €2.3 billion national and sub-national public funding

Revised Broadband State aid guidelines

State aids for NGAN deployment (Source: Digital Agenda Scorecard, June 2012)

Government Interventions All over Europe Governments generally intervene only in providing funds for

deployment NGANs. In some countries through Private-Public Partnerships (Greece), or deploying of

passive infrastructure (civil works, ducts ..; France and Germany); or through competitive procedures or tenders (Spain, Sweden); or through a private consortium (Thales in UK).

Direct intervention in NGN through a Government Business Enterprise (GBE) in Australia. The newco - National Broadband Network - will invest 43Aus$ billion in 8 years for covering 93% of customers. It will be privatized around 2020.

A mixed case in Italy: Metroweb (partially controlled by a government investment bank – Cassa Depositi

e Prestiti) controls a fibre network in Milan and will deploy its own NGA network (FTTH) in 30 big cities (4.5€ billions in 8 years) covering approx. the 20% of entire population. Deployment of 28.000 km of fibre optic cables.

Incentives – and Government support (?) - for a cooperation with Telecom Italia (the incumbent operator) in the deployment of a single (mixed owned) infrastructure

Regulatory Framework for NGANs

NGA Recommendation - Main features/1 • By the EU Recommendation on “Regulated Access to

NGANs (Sept. 2010)”, cost-orientation remains the rule but NRAs can grant a risk premium when setting regulated ex ante access prices

• It recognises co-investment schemes between operators in order to diversify the investment risk, which can lead to a relaxation of ex ante regulation (examples in Italy, Netherlands, France, Switzerland, Portugal)

• To enhance regulatory credibility:

• “Recommendation on costing methodologies for key wholesale access prices” problem of transition of technology between the “old” copper network to the “new” fibre one

NGA Recommendation - Main features/2 • Geographical Markets

• The EC framework allows NRAs to introduce differentiated remedies within a national market (Directive 2009/140/EC, ”Better Regulation Directive”)

• The NGA Recommendation also invites NRAs to examine the possibility of differentiated remedies

• Current NRAs Decisions on Market Definition for the (DSL) broadband markets:

• Geo wholesale markets and differentiated remedies for different (competitive vs not competitive) areas in UK, Portugal, Finland and Hungary.

• National markets (with uniform remedies) in France, Germany, Spain, and Italy.

NGANs remedies in selected EU countries

A focus on Italy

Incentives for cooperation already in place (agreement between Telecom Italia and Fastweb for co-sharing of civil facilities)

The Italian Authority has started several Consultations on: Cost methodology and risk premium market-based access prices to NGA with price test, geographical markets symmetric obligation to (in building) open access to fibre

Towards a return to the “classic” Rate of Return regulation to NGAN due to excessive demand risk

The broadband market in Italy today

Source: AGCOM, Quarterly Telecommunication Market Observatory, September 2012

The Economics of NGA Regulation

Starting points: Tension between regulatory goals: static efficiency and dynamic

efficiency (Laffont and Tirole, 2000) Impact of access regulation on network investment (see the survey by

Cambini and Jiang, 2009 JTPO): on entrants’ investment incentives: Bourreau and Doğan (2005 AER; and

2006 EER), Avenali et al. (2010, IJIO); on incumbents’ investment incentives: Gans and Williams (1999,EcRec),

Foros (2004 IJIO), Kotakorpi (2006 IJIO), Klump and Su (2010, AEJ: Micro); on incumbents’ and entrants’ investment incentives: Gans (2001 JRE), Hori

and Mizuno (2006, IJIO), Vareda and Hoernig (2010, BJEC); NGAN invesments and access rules: Nitsche and Wiethaus (2011 IJIO), Brito

et al. (2010 IJIO, 2012 IEP), Inderst and Peitz (2012 TELPOL, 2012 RNE)

Moreover, Telcos are highly leveraged, thus creating an indirect pressure on regulators not to decrease regulated – retail and wholesale – prices (Cambini and Rondi, 2012 ICC)

The Economics of NGA Regulation

With NGANs, additional issues emerge: regulatory interventions are becoming even more complex and articolate than before

New regulatory issues:

The transition from the “old” copper network to the “new” fibre network;

The presence of sub-national markets; The incentives for co-investment under demand

uncertainty.

Migration to NGAs

Coexistence between legacy networks and NGANs: the transition from copper to fiber will go slowly and this creates rooms for regulatory intervention (art. 40, EU NGN Recommendation)

The regulatory framework will therefore affect investment in NGANs through two different channels: Regulation of access to NGANs ... but also: regulation of access on the legacy “copper” network

Huge policy discussions in Europe (Plum, 2011, for ETNO; WIK, 2011, for ECTA)

In Bourreau, Cambini and Dogan (2012 IJIO) we analyzed:

The impact of access rules on “old” copper network on NGANs deployment

the “interplay” between access regulation on the “old” copper network and the “new” NGN network.

Economic incentives for migration

Three effects at work that affect investment incentives:

Replacement effect: if the access charge on old network is high (low), lower (higher) opportunity cost of investing, hence higher (lower) incentives to invest for entrant in the NGAN (=> Arrow, 1962)

Wholesale revenue effect: if the incumbent invests in a higher quality network, the entrant may invest in reaction, and the incumbent will then lose some wholesale profits; the higher wholesale profits are, the lower the incumbent’s incentives to switch

Retail migration effect: when the access price of the legacy network is low, the retail prices for the services on this network are low too; hence, in order to encourage customers to switch from the old network to the new network, operators should also offer low prices for ultra-broadband services too, reducing the profitability of the new technology infrastructure, and the incentives to invest in it

Main Results

These effects calls for different solutions: Replacement effect. Solution = access charge Wholesale revenue effect. Solution = access charge Retail migration effect. Solution = access charge

Bottom line: ambiguous effect of access charge on the legacy network on investment incentives For an entrant, a higher LLU price implies more

investment For an incumbent, it can lead to either more or less

investment

This analysis has been used by the European Commission for its recent recommendation on cost methodologies

Interplay between Old and New Infrastructures

Regulatory perspective: Interdependency of access rules When the incumbent has larger NGN coverage than the

entrant, positive correlation between access charges: the lower the access charge on the “old” network, the lower the access charge for NGN should be to have migration at the wholesale level.

The results are reversed when the entrant has larger NGN coverage than the incumbent, given the relative advantage the incumbent enjoys for controlling the old technology (negative correlation)

This calls for a sort of “asymmetric” access regulation

The Geographical Dimension of Investments

(I) Fibre networks likely to emerge in some areas, not all (e.g., big cities). Example of Orange’s fibre roll-out in the Paris area:

The Geographical Dimension of Investments

(II) Market structure will also differ across areas: some areas with a monopoly infrastructure, others with competing infrastructures.

For example, from a sample of 3,022 municipalities in France, listed on broadband operators’ websites (as of June, 2012):

Different market structures in different areas: should regulation be tailored at the local level? And how?

A Model (Bourreau, Cambini and Hoernig, 2012a CEPR DP)

The firms: Two incumbent operators (firms 1 and 2) invest in NGAN coverage A potential entrant, firm e, might ask for access but does not invest

Investment and access: Incumbents build infrastructure in different areas of a country of size z Firm e can enter a given area only via buying access to an incumbent

firm’s infrastructure An incumbent may ask for access to another incumbent

Retail competition: Firms can set different retail prices in each area They adjust their prices according to competitive conditions Main analysis based on general profit functions

Investment Cost The country is composed of different areas, ordered according to the cost of

deploying an NGN

Investment coverage decision

area z

cost of NGN in area

Full commitment

The two (symmetric) incumbent network operators decide, simultaneously, on the areas in which they will roll out an NGN. Then, they compete. The entrant may ask access (randomly) to one infrastructure operator.

The regulator sets different regulated access prices in duopoly and monopoly areas.

Firm i’s profit is given by:

Investment and access (Ex post asymmetric) equilibria:

Partial commitment: Alternative Regimes

We consider alternative access regimes, with more reasonable informational requirements:

Uniform Access Pricing: The same access charge everywhere Probably lowest informational requirements

Competition-Based Differentiated Remedies: The access charge in monopoly areas is regulated Light access regulation in duopoly areas: wholesale

prices set on commercial basis

Geographical segmented remedies Rationale for segmenting remedies according to geography

(e.g., access to “monopoly” NGANs only)

Where more than one infrastructure, we allow for competition at wholesale level for the provision of access ”softening effect”, the access provider becomes a less aggressive competitor

Bourreau, Cambini and Hoernig (2012a) propose to introduce “competition-based access remedies”: Regulation of access to monopoly infrastructure Dispute resolution ( avoiding foreclosure) in areas with

competing infrastructures (with price caps)

Co-investment in fibre networks Viewed by the European commission as a means of increasing fiber

coverage “Co-investment into NGA networks can reduce both the costs

and the risk (…), and can thus lead to more extensive deployment of FTTH.” (NGA recommendation, 2010).

Is it true? Operators will share costs, but also revenues (cf. Bourreau, Cambini and Hoernig, 2012b)

Only a significant demand expansion effect can make co-investment a “game changer” For example, if co-investors target very different market segments

Otherwise, total coverage would not be affected by joint agreements

Co-investment analysis Co-investment problem with access similar to joint R&D investment

(collusion at retail level) and patenting

Co-investment decisions and access prices to NGANs are also interrelated (in a complex way) An infrastructure builder will trade-off between building its own

network, co-investing with another operator… and asking for access to an NGAN

Co-investment also leads to higher deployment rate in presence of demand uncertainty (risk)

If voluntary access is provided, co-investment coverage increases as compared to regulated access, but might lead to lower welfare in case of low product differentiation

Future research Theory:

Dynamic model of investments may affect migration of technology.

The size of competitive and non-competitive areas also change over time, calling for a dynamic adjustment of access remedies.

In the co-investment analysis: the effects of asymmetry between co-investors and the choice of rules for sharing investments.

Role of Public-Private Partnerships in innovative services Empirics:

Still no sound empirical analysis on NGA adoption and the role of regulation

Existence of several cross-countries (extra-EU) studies, with too much heterogeneity ( countries with highly different regulatory framework and different Governments’ intervention)