35
Economic characteristics and policy challenges of NGANs Glenn A Woroch C.R.T.P. / UC Berkeley Policy for Next Generation Networks: European and US Perspectives MIT Communications Futures Program 27 March 2009

Economic characteristics and policy challenges of NGANs Glenn A Woroch C.R.T.P. / UC Berkeley Policy for Next Generation Networks: European and US Perspectives

Embed Size (px)

Citation preview

Economic characteristics and policy challenges of NGANsGlenn A Woroch

C.R.T.P. / UC Berkeley

Policy for Next Generation Networks: European and US Perspectives

MIT Communications Futures Program27 March 2009

Overview

1. Engineering & economic properties of NGANs

2. Competition & regulatory concerns

3. Policy options & economic criteria

4. Vertical separation as a remedy

5. Different approaches in the U.S. and E.U.

6. Open questions & policy challenges

1. What is the NGAN?

Consensus so far … Converged: IP transport of all kinds of content to

all kinds of devices Layered: OSI and other layered architectures Modular and Open: interoperability among layers,

components Backward compatible: interoperate with legacy

PSTN and data networks

What is the NGAN?

Not your father’s Internet Consolidation of ATM, SONET, Ethernet, MPLS More than best effort: carrier class, security

But still a work in progress Competing technologies: FTTx’s, xDSL’s, etc. Competing architectures: star, PON, etc. Open and proprietary technologies

Distinct economic features of NGAN Natural monopoly at the core

High fiber results in high fixed costs, low marginal costs

Impediments to facility sharing Open entry at the edges

Open interfaces, standard protocols Limited capacity of critical facilities

Converged devices at the ends Need to manage connections & traffic

2. Competitive & regulatory concerns How to respond to a fortified access

monopoly? How to maintain progress toward competition

in retail services? How to preserve incentives to invest in

facilities and to rollout innovative services? How to transition from current institutions?

Competitive concerns

Disadvantage downstream rivals Margin squeeze Quality sabotage Refusals to supply Tying and bundling Technical incompatibilities

Vertically integrated bottleneck monopoly

Consumers

V.I.M.

Retail rival

Regulatory concerns

What parts of NGAN are natural monopolies? Distinguishing new from legacy bottleneck

facilities How carve out potentially-competitive portions

of the NGAN? Engineering: layers Economics: markets

Regulatory concerns (cont’d)

What policy to impose on the boundaries of the NGAN? Nondiscriminatory access Full interoperability

How to transition from legacy network and ownership structure to NGAN? Maintain regulation on legacy network? “Regulatory holiday” for NGAN?

3. Policy options & economic criteria Objectives of policy intervention Economic criteria for evaluating policies Structural and behavioral strategies

What are the objectives of intervention? Curb monopoly excesses Prevent exclusionary behavior Promote efficient entry and competition Facilitate investment in infrastructure Preserve incentives to innovate

Economic guidelines for policy analysis Overarching criterion should be efficiency

Consumer welfare is paramount But measured by social welfare

Time is of the essence Not just short term or just long term Dynamic progress critical to network evolution

Policies need to be “incentive compatible” All agents will pursue their private interests

Policy options

Forbearance Promotion of platform competition Opening access network and unbundling Facility sharing Vertical separation Ex ante regulation Ex post competition policy

4. Vertical separation as remedy Kinds of vertical separation Vertical separations experiments in U.S., E.U.,

and elsewhere Economic benefits and costs of separation Obstacles to separation

Why the interest in separation now? Dissatisfaction with progress toward

competition Lack of entry under unbundling policies Slow development of platform competition

Fear that NGAN could reverse progress Opportunity to establish dominance in new

infrastructure market Such dominance could be extended to retail

Vertical separation

Consumers

AccessDivision

Retail rival

RetailDivision

Degrees of vertical separation/integration

Full integration Forward and backward integration

Quasi-integration Long term contracting

Nonstructural separation Accounting separation Functional separation Operational separation

Structural (ownership) separation Reciprocal partitioning Club ownership No vertical integration

Four degrees of separation

Accounting

Operational

Functional

Structural

Telecom separation around the world

accountingoperationalfunctionalstructural

Efficiencies of vertical separation

Benefits Eliminates of cross

subsidies Eliminates conflict of

interest Counteracts

discriminatory treatment of retail rivals

Costs Forgoes vertical scope

economies Causes double

marginalization Lowers reliability due to

quality externality Raises transaction costs Blunts incentive for

access investment

Obstacles to vertical separation Divorced facilities may not be sharable

Certain portions of a non-IP video market Separation may preclude certain bundled

services Access + retail services, especially for businesses Arms’ length contracting alternative can be costly

5. Divergence in U.S. and E.U. approaches Origins of institutions and property rights of

the two telecoms sector were vastly different Current network infrastructure and industry

structure varies considerably in EU, but not in the US

Subtle but fundamental differences in competition policy

Example of competition policy differences

U.S. E.U.

Quality degradation

Trinko (2005) IMS Health (2004) Microsoft (2004)

Price/margin squeeze

LinkLine (2008) Deutsche Telekom (2003) Telefonica (2007)

Separation policy in the U.S.

Anti-trust separations LOB restrictions on AT&T in 1956 Divestiture of AT&T in 1984

Regulatory separations FCC’s Computer I, II, III LOB restrictions on

RBOCs PSCNY’s Rochester Tel experiment in 1995 SBC’s separation of DSL service in 2001 Failure of state nonstructural separations (e.g.,

Pennsylvania)

Separation policy in Europe

Regulatory stance on separation in the EC OECD separation recommendation (2001) EC recommendations on accounting separation

(1998, 2004) Reding speech (2006)

OFCOM-BT’s undertakings in 2005 Separation of OpenReach and BT Retail Applies to 21CN products

But much earlier there was the French approach to vertical separation …

6. Open questions & policy challenges Implementing vertical separations Enforcing non-discrimination Measuring success

Implementing vertical separation Pre-conditions for separation policy

Activity prone to “enduring bottleneck” Complements are potentially competitive Bottleneck monopolist integrated into the complementary

markets Essential elements of vertical separation

Quarantine of bottleneck Isolate of essential facilities Impose line-of-business restrictions

Regulate the bottleneck Dictate range of wholesale access services Enforce price and quality nondiscrimination

Measuring success

Direct measures of success Entry of new firms upstream and downstream

Facilities based or service based Development of infrastructure/platform

competitors (e.g., wireless) Indirect measures of success

Lower retail prices, higher service quality Expanded retail service offerings

Measuring success (cont’d)

All impacts need to be evaluated incremental to the status quo Any unbundling regime Any access/interconnection pricing regulation Any form of vertical separation

Case of Italy Unbundling Interconnection pricing Operational separation of network services

Various non-discrimination principles OFCOM’s “equivalence of inputs”

Same products, sold at same prices, using same order processes as BT’s retail division

Applied to certain existing and future products FCC’s “parity principle” TCNZ’s “fair and equivalent” AGCOM’s “parity of internal-external

treatment” (152/02/CONS) Competitors can replicate TI’s retail services

Why non-discrimination may not be good Retailers value input quality differently

Some retailers prefer lower quality input if they receive a sufficient price break

Sabotage may impose cost on access supplier Greater costs to creating multiple qualities of an

input Greater sales to the advantaged affiliate retailer

offset by reduced sales to rival retailer

Performance measurement: the plan Map indicators of wholesale service into

monetary penalties Benchmark is service received by bottleneck’s

retail division Induce bottleneck monopoly to provide

equivalent service to retail rivals Discriminatory treatment becomes sufficiently

costly

Performance measurement: the reality Evolution of PMPs

Measures divide and multiply Penalties occur for uncontolled events Competitors learn to “game the system” Link between effort and penalty becomes weak

PMP distorts behavior of incumbent and its competitors