jump to navigation

Europe: International roaming – ATKearney CRAI report September 30, 2006

Posted by Jasper in Mobile, Regulation.

On 21 September 2006 the GSM Association (GSMA), the global trade association for mobile operators, filed a complaint with the European Ombudsman setting out its concerns about the European Commission’s proposal for regulation of the international roaming market. The GSMA’s complaint relates to three aspects of the regulation and the associated adoption process (see press release):

1) The GSMA believes the European Commission failed to conduct a proper consultation process. While the Commission conducted two public consultations on regulation of the international roaming market, they both failed to follow the Commission’s own procedures for providing clear content and adequate time for responses. Moreover, following these ‘public consultations’, the Commission drastically changed its proposal for regulation without submitting the new proposal to additional public consultation.

2) The GSMA believes the European Commission failed to conduct a proper impact assessment. The published impact assessment, which didn’t follow the Commission’s internal guidelines, is incomplete, contains important methodological flaws and is based on unrealistic assumptions. For example, the Commission concluded that consumers would see a net gain of 3.78 billion euros from the proposed regulation. But a review of the Commission’s impact assessment, prepared for the GSMA by AT Kearney and CRA International (CRAI), found that the impact on consumer welfare would be marginal at best, and could even be negative.

3) The GSMA believes the European Commission’s proposed regulation violates the principles of proportionality, subsidiarity, legitimate expectations and non-discrimination. The holistic nature of the existing regulatory framework restricts the Commission’s ability to adopt an ad hoc regulation on roaming services.

What interests me is the report undertaken for the GSM Association by AT Kearney and CRAI. The report discusses the impact of the European Commission’s plans for international roaming and maintains that the Commission has used the wrong figure for the size of the retail roaming market. According to the report’s authors:

…their corrected welfare modelling indicates that the case for the Commission’s proposed regulation is only marginal at best and could even be negative, rather than significant, as the Commission has estimated.

Let me offer a few comments.

The ATKearney CRAI recommendation centres on a Cost Benefit Analysis (CBA) and on the their own evaluation of that CBA. The CBA is an assessment of a counterfactual scenario (“the future without regulation”) with a factual scenario (“the future with regulation”) – in fact there are two factual scenarios.

In principle, the approach adopted is more complete compared to that of the Commission. It attempts to model detriments to regulation which the Commission appear to have neglected (I have not read the Commission assessment). In this sense it provides a more complete analysis of the market for international roaming and the (net) consequences of regulation. Whether or not there are measurable detriments of the magnitude suggested by the authors is an issue I explore briefly below.

One crucial correction is the Commission’s estimate of the benefits which is based on the GSMA’s estimate of the aggregate EU roaming revenues of about €8.5 billion. However, according the authors the GSMA’s €8.5 billion figure includes revenues from both the retail and wholesale markets and is not an appropriate basis on which to calculate the ultimate saving for consumers. I do not have the data to evaluate the figure, but expect that ATKearney and CRAI are correct as the number has been sourced from their client.

I agree with ATKearney and CRAI that the Commission should take account of announcements by individual operators and any announced industry agreements in the factual scenario. Also I agree that the Commission has adopted some rather suspect elasticity assumptions in some of their scenarios.

The authors introduce two concepts: cost pass-through and the waterbed effect or a rebalancing effect.

Cost pass-through is degree to which reductions in wholesale prices are passed-through to retail prices, while the waterbed effect is the degree to which operators seek to recoup lost profits in the market for international roaming from other services or markets. The waterbed effect might also be called a pass-through rate, i.e. it is the extent to which any given loss in international roaming profits resulting from regulation is translated through into increases in other retail mobile prices. A decrease (increase) in this rate produces a smaller (greater) waterbed effect, and hence a smaller (larger) detriment. Setting the pass-through rate to zero is equivalent to assuming that there is no waterbed effect.

Both assumptions are critical to the author’s analysis, unfortunately very little documentation is provided to substantiate the values used. In particular I would have liked to have seen an analysis of historical cost pass-through. In their Medium case, the authors assume that rebalancing would operate so that 80 per cent of the loss in profits on retail roaming calls would be recovered in mobile subscription charges and mobile outgoing call prices. No evidence is provided to support this value – a value I believe is exaggerated.

The authors criticize the Commission for using an aggregate measure for total welfare for consumers and industry and not considering how the proposal will impact different Member States. However, the authors themselves do not attempt a disaggregated approach. I think it would desirable to disaggregate into at least three or four category of Member State.

Any offsetting effects of their new assumptions are disregarded. For example, as a result their assumed reductions in mobile subscribers there would a corresponding outward shift of the demand for fixed calls.

In addition, makes no attempt to discuss a number of issues that could have major impact the results (although to their defense I assume they adopt the same values as the Commission). For example:

  • Study Period: what is period over which the welfare gains/losses are calculated?
  • Discount rate: The choice of discount rate in a CBA is potentially a very contentious issue. What is the level of the discount rate used to calculated the net present value of net benefits?
  • VAT: VAT should be excluded for estimates of productive surpluses since the VAT is a transfer payment; however, it is not correct in the calculation of allocative surpluses. How is it treated here?
  • Shape of demand curve. Two alternatives are typically used: a linear curve and the constant elasticity of demand curve. Which one is used?

ATKearney CRAI assume that the impact of lower wholesale roaming charges on the profitability of operators maintaining cell sites in areas in which roaming revenues are significant leading to loss of coverage. I disagree. It seems highly implausible that an operator will remove (sunk) cell sites already deployed.

A CBA should be applied rigorously and transparently. The ATKearney CRAI report appears to be an improvement on the Commission’s analysis, but it is overly critical and hence biased towards a non-regulation scenario and does not provide the detail needed for a third party to do a proper evaluation.

Access the AT Kearney CRAI report “Review of the Commission’s Impact Assessment” here and the Executive Summary of the Review here.



1. John Middleton - October 4, 2006

Nice analysis – its a shame you don’t work for the Commission! 🙂 They need some people who actually take time to evaluate different factors and think through what they’re putting forward. There are far worse examples of this in their Television without Frontiers proposal…

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: