WIK conference on vertical separation in telecoms Brussels, 22-23 November 2010 ‘The evolving theory of separation

’ Martin Cave Department of Law London School of Economics [email protected]

From the general to the particular

The general: a discussion of the role and consequences of vertical integration and separation: the main protagonists Smith, Coase, Williamson, Klein, Tirole, Hart, Holmstrom etc The particular: a largely separate discussion of separation in telecoms.

The modern starting point-R H Coase and the nature of the firm (1937)

Asks the question: what determines the scope of activities carried out inside a firm, and the scope of activities outside the firm? (antithesis later described as ‘market versus hierarchy’) Answer lies in terms of relative efficiency, or internal complexity versus transactions costs Applies to horizontal and to vertical relationships.

A modern restatement: why firms integrate or separate. I - the moral hazard model

Consider integration between a manufacturer/network operator and a retailer. What arrangement ensures high effort/absence of shirking by the retailer?

Answer depends on -risk : integration provides retailer insurance -importance of retailer effort, and need for high-powered incentives -importance of upstream effort, and need for high powered incentives there incentives there -cost of monitoring

Why firms integrate or separate : II - the transactions cost model

Firms often make investments which are worth more inside than outside a relationship, because of geographic or functional asset specificity Such interdependence can be protected by complete contracts, but these are hard to write Hence separation can open up risks such as those of hold ups.

The hold-up problem

The upstream firm has to make a sunk investment; once it is made, the downstream firm can drive the wholesale price down to short run incremental cost; anticipating this, the upstream firm refrains from investing Can contracting solve this problem – such as long term or take or pay arrangements?

How serious is it in practice? See the GM-Fisher Body case of the 1920s (Coase JEMS, 2006).

Why firms integrate or separate. III- the market power model

Double marginalisation- Spengler 1950. Separation raise prices above monopoly level and introduces inefficiency in production (anti-separation) Foreclosure and raising rival’s costs, by price and non price methods (pro-separation ) Strategic delegation and collusion – upstream firms can soften competition by delegating the pricing decision to independent retailers (anti-separation) etc

What happens when they integrate or separate? -I

Lafontaine and Slade

: “.We did not have any particular conclusion in mind when we began to collect the evidence..We are therefore somewhat surprised at what the weight of evidence is telling us. It says that in most circumstances, profit maximising vertical integration decisions are efficient, not just from the firms’ but also from the consumers’ point of view. The vast majority of studies support this claim,..even in industries which are highly concentrated..”

(JEL, 2007, based on review of c200 empirical studies)

What happens when they integrate or separate? - II

Joskow

: “Overall I would argue that there is substantial support in the empirical literature for various efficiency motivations for vertical integration. There is minimal empirical support for anticompetitive foreclosure motivations. This suggests that there is little empirical support for antitrust law’s traditional suspicion of and hostility toward vertical integration and related non standard vertical contractual relationships except under extreme conditions where firms controlling bottleneck have the incentive and ability to exercise an anticompetitive foreclosure strategy”.

American Bar Association, 2008

But, Joskow favours separation in energy

(Energy Policy 2008)

Highlights of the telecoms separation debate

• • • • • • • • • 1985- the break up of Bell: local versus long distance- ‘a reckless but successful gamble’ 1999- regional separation of NTT in Japan 2001- OECD recommendation on structural separation in regulated industries and associated debates [2000s-divestment of telco cable operations in Germany, Portugal etc] 2005-functional separation of BT in UK 2000s-divestment of telco cable operations 2005-2008-further functional separations (Australia, Italy, Sweden, NZ) post 2008-the age of NGAs, public subsidy and their consequences (Singapore, solvent countries passim) 2010- further moves in Australia, NZ.

Telecoms exceptionalism1: the half-way house, or having your cake and eating it

Types of separation Ownership Legal Functional Virtual Accounting

Where does the benefit of separation lie? The sabotage argument in telecoms

The problem of non-price discrimination

• Accounting separation should make it relatively(?) simple to deal with margin squeezes.

• But the firm has the incentive to engage in non-price discrimination or ‘sabotage’ if: i) monopoly activities are regulated to cost ii) the retail market is homogeneous iii) the integrated firm is efficient downstream.

• • Absent adequate enforcement, this creates a case for stronger forms of separation.

BT confessed? Who else is guilty?

Where are we now in the EU? mandatory functional separation • Mandatory functional is now a recognised access remedy. • Because it is costly and intrusive it should be employed in the presence of evidence of failure of other remedies- ie only deployed if the goal of equivalence is not obtainable without it.

• It should take account of the prospects for end-to end competition (but cannot be geographically differentiated).

• Summary: it will be difficult to fulfil the regulatory conditions and to accomplish it in practice

Where are we now in Europe? ‘Voluntary’ functional separation • No obvious reason to prohibit separation.

• Openreach created over 5 years ago. Success claimed by BT and regulator.

• Telecom Italia has agreed arrangements with regulator, which do not require much re-engineering. • Voluntary separation achieved in Sweden, subsequently formalised and strengthened (but excluding fibre).

Telecoms exceptionalism II: NGNs A ‘once in a century’ reworking of the local loop Competitors need to invest to meet cable challenge; loop monopolists need a regulatory or other incentive to invest Copper unbundlers are separated companies with a choice of copper or fibre; does this make the NGN investment more risky? Would it be worse with no integration?

New contracting possibilities include quantity discounts and long term contracts; do these solve the transaction costs problems?

Telecoms exceptionalism III: the role of public investment • • • • • Central or local government part-ownership of fibre networks is now commonplace In this framework, separation can be enforced contractually as a matter of public policy (Singapore, elsewhere) Public bodies can inject limited funds and require networks to be open, but not necessarily separated (eg Portugal) Public/private partnerships or JVs in networks entail separation because of different ownership structures in different network and retail layers (New Zealand) In the limit, complete public ownership of a separated network, with assimilation or elimination of competitors (Australia)

Conclusions I - general

The theoretical argument can go both ways: moral hazard –pro separation; transactions costs –pro integration; market power –mixed/pro separation Strong empirical evidence in favour of integration puts a burden of proof on separation exponents.

Conclusions II – particular

• • • • Conventional wisdom that limited separation of copper networks is a second-best solution to deal with a recalcitrant incumbent may be right, subject to problems with the counterfactual The NGN challenge in non-competitive areas sets private incentives to invest against the goal of opening an separating networks Public funding can ‘solve’ this problem, but points of separation may be driven by investment rather than operational considerations With public funding, the incremental burden of separation can fall on the public purse.