Ofcom response to the European Commission public consultation

Ofcom response to the
European Commission public
consultation on the review of the
2009 Termination Rates
Recommendation
June 2016
Submission date:
07/06/2016
1
Introduction
1.1
Ofcom welcomes the opportunity to respond to the European Commission’s
consultation of the evaluation of the 2009 Termination Rate Recommendation (“the
Recommendation”). Ofcom has fully implemented the Recommendation with MTRs
charge controlled at pure LRIC since 1 April 2013 and FTRs charge controlled at
pure LRIC since 1 January 2014. Ofcom supports the continued use of the pure LRIC
standard for setting termination rates.
1.2
We believe that the Recommendation has been broadly successful in achieving its
objectives by promoting competition and the interests of EU citizens. Since MTRs
and FTRs moved to pure LRIC in the UK, our evidence indicates outcomes of
increased competition and stable/falling prices for telecoms services. We have not
found evidence that moving to LRIC has negatively impacted investment, though the
available evidence is limited. We include further details of our economic assessment
of the Recommendation in Annex 1.
1.3
Ofcom has also contributed to the consultation response produced by BEREC and
supports the content of that response. The purpose of this document is to highlight
the UK’s experience of implementing the Recommendation and some suggestions
Ofcom has for future developments. We will seek to largely avoid issues already
discussed in the BEREC response.
Approach to modelling
1.4
The Recommendation recommends the use of a bottom-up model which may then
be calibrated using a top-down approach based on audited data. This is the
approach we have taken in both the 2013 Narrowband charge control (to set FTRs)
and the 2015 MCT charge control1 (to set MTRs). We believe this approach provides
the best combination of robustness and accuracy in line with the overall approach to
model the costs of a hypothetical efficient network providing a termination service.
Single European cost model for termination rates
1.5
Even if NRAs implement the recommended modelling approach, we would still
expect some difference in termination rates across Member States that reflect the
differences in costs between networks in different countries.
1.6
It has been suggested that a single EU-level cost model could be created to replace
the current system of each NRA developing its own model. We do not believe this to
be practicable without placing a burden on NRAs and risking unintended
consequences. We consider there to be significant risk of such a model failing to
account for all required specifics, or to be so complicated that it would be extremely
difficult and time-consuming to develop.
1.7
There are a wide range of differences between Member States that lead to
differences in the costs of providing wholesale call termination services, resulting in
valid differences between the estimates of efficient TRs of Member States. These
include differences in:

2
population distribution;

geography and network topology;

equipment, building and labour costs;

maturity of markets; and

operator WACC.
1.8
Even where we might expect model input parameters to be similar, it is not sufficient
to say there is a problem if parameters are different across models. Input parameters
are often changed as part of the model calibration process and when comparing
models a holistic approach should be taken.
1.9
A model which is unable to adequately account for the differences in parameters
across Member States runs the risk of setting a TR that is too high or too low,
running counter to the arguments set out in Annex 1 regarding the value of setting
TRs at LRIC. This risk would certainly be realised by using a model to generate a
single pan-European TR for each of mobile and fixed, as there will be different
efficient rates for each Member State based on national circumstances (and indeed
there is, as can be seen from the range of rates set by Member States that have fully
implemented the Recommendation).
1.10
Creating a model that does adequately account for all relevant differences in national
circumstances, though technically difficult, should be possible. Building such a model
would likely be a very time consuming task, involving extended consultation with
NRAs and operators in addition to the time required to build and sufficiently audit the
model. This would significantly reduce or negate the proposed benefits of reduced
ongoing regulatory burden for NRAs and operators, and could instead increase the
regulatory burden (particularly on Member States that have already developed
bottom-up LRIC models).
1.11
Instead, we believe that discretion should remain with individual NRAs to build cost
models that account for their own national circumstances. In order to reduce
regulatory burden on resource-constrained NRAs that may find it unfeasible to build a
bottom-up cost model, the EC could create a template cost model that NRAs can
choose to use, and could modify as needed based on national circumstances and the
specifics of the model. Any benefits of building such a model should be compared to
the existing benchmarking approach used by some NRAs with limited resources.
Guidance on modelling approach
1.12
The Recommendation already includes substantial guidance on its interpretation and
implementation. This guidance is helpful to NRAs, but we believe that some limited
further guidance could be included to reduce differences in the TRs calculated by
Member States, where these differ due to differences in interpretation of the
Recommendation rather than differences in national circumstances.
1.13
In particular, one significant cause of the divergence between the rates calculated by
Member States, particularly in FTRs, is the treatment of administrative and billing
costs. We have found that some of these costs do appear to be incremental to the
wholesale fixed call termination increment, mostly due to the need to maintain a
billing platform that can accurately account for calls on a far end hand-over system.
In our 2013 NBMR decision, this assumption in the model accounted for a significant
proportion of the final FTR.
3
1.14
We believe that it remains right that NRAs should continue to base their estimates on
the level of administrative and billing costs that are incremental to termination on the
best evidence available to them. However, further guidance would be helpful to
clarify the interpretation of the available data in order to prevent unrealistic levels of
cost associated with admin and billing being allocated to termination services.
Reducing the regulatory burden and the frequency of model
updates
1.15
We believe there may be benefits in reducing the burden on NRAs and industry
stakeholders when setting termination rates. The market reviews in which termination
rates are set currently operate on a 3-year cycle. In some cases, for example in
cases where volumes are relatively easily predictable and the efficient network
technology is well-established, updating the model every 3 years provides rather
limited benefit in terms of increased accuracy and robustness of the analysis when
compared to the burden placed on NRAs and industry stakeholders required to
perform such an update. For example, in the UK the “net termination revenues” (i.e.
net of mobile-to-mobile or fixed-to-fixed calls, as payments for these net off against
one another) are currently around £114m in mobile and £12m in fixed.2 Because
differences in modelling approach from one market review to the next will change
only a fraction of this amount, the resources involved in building and consulting on a
bottom-up model can quickly outstrip any benefit of such an exercise.
1.16
As discussed above, we do not believe that the creation of a single European cost
model is likely to be an effective way to reduce the burden placed on NRAs and CPs.
Rather, as we have previously argued in our response to the EC’s consultation on
the review of the regulatory Framework3, we would support NRAs having greater
flexibility over determining the appropriate frequency of market reviews. This would
allow NRAs to reduce the regulatory overhead in markets that are more stable and
for which a full update of a bottom-up cost model would be unduly onerous every
three years.
Impact of asymmetries of implementation among Member States
1.17
Asymmetry in the extent to which the Recommendation has been implemented by
different Member States has led to asymmetries of regulated termination rates. While
some differences are to be expected due to national circumstances and the data
available to NRAs, the largest cause of asymmetric rates is the application by some
Member States of a cost standard other than pure LRIC. Large differences in the
termination rates across Europe leads to cross-subsidisation effects from those
Member States with lower rates to those with higher rates.
1.18
These asymmetric rates may also be a barrier to the development of pan-European
tariffs and bundles due to the increased average marginal cost of intra-EU
international calls. Differences in implementation of the Recommendation, particularly
regarding the cost standard, may therefore create a barrier to the development of the
internal market. We believe that it is up to the EC to decide the best approach to
address this issue.
2
Assuming non-M2M off-net volumes of 16.72bn minutes in mobile and 36.4bn non-F2F off-net
volumes in fixed, and current termination rates of 0.68ppm in mobile and 0.034ppm in fixed.
Assumptions based on the 2015 MCT Statement and the 2013 NBMR Statement
3
See paragraphs 30 to 32 of Ofcom’s response, available online:
http://stakeholders.ofcom.org.uk/binaries/international/responses/Ofcom_response_Commission_con
sultation.pdf
4
5
Annex 1: Assessment of the Recommendation
A1.1
In this section we will briefly summarise the theoretical arguments supporting the
implementation of pure LRIC and any evidence we have found that supports or
contradicts these arguments.
Efficiency as a basis for setting termination rate at pure LRIC
A1.2
Ofcom has argued that in the absence of common costs, TRs set at pure LRIC
should be allocatively efficient as pure LRIC provides a reasonable proxy for the
marginal cost of the service. It is less clear what the allocative efficient outcome is
in the presence of common costs.
A1.3
Ofcom has considered the validity of Ramsey pricing arguments, which state that
under certain restrictive assumptions, having some mark-up for common costs
would be more efficient than having none. We concluded that actually calculating a
Ramsay-pricing mark-up is impractical, and there are good reasons to believe that
the efficient mark-up would be very small or zero. Any gains from trying to fine-tune
allocative efficiency in this way are therefore unlikely to be large and, because there
is so much uncertainty over the model and parameters that could be used to
determine the ‘optimal’ TR, any mark-ups risk being arbitrary and decreasing
allocative efficiency.4
A1.4
We also consider that rates at pure LRIC are unlikely to be less dynamically
efficient than rates at LRIC+. While rates set at LRIC+ might generate more
revenue and potentially more profit than pure LRIC for the fixed/mobile industry as a
whole, it is unlikely to lead to significantly greater investment or more efficient
investment overall. This is consistent with the empirical evidence we have found for
the mobile industry, which does not suggest that the reduction of MTRs down to
pure LRIC has been associated with lower investment (see Figure 1 which shows
aggregate capex by the four largest MCPs over the period in which MTRs dropped
from LRIC+ to pure LRIC). We also do not consider there to be a significant
asymmetry of risk in terms of dynamic efficiency from being above or below the outturn level of pure LRIC. This is because there is an opportunity to recover costs on
the retail-side of the market and the impact of any potential forecast error is unlikely
to be material in terms of CPs’ net revenues and/or incentives to invest.5
4
Further details available from paragraphs 6.24 to 6.56 of the 2015 MCT Statement, and from
paragraphs 8.25 to 8.70 of the 2013 NBMR Statement.
5
Further details available from paragraphs 6.57 to 6.92 of the 2015 MCT Statement.
6
Figure 1: Capex by the four largest MCPs 2009-2013 (£bn, 2012/13 prices)
MTRs began moving to pure LRIC from April
2011, reaching pure LRIC from April 2013
£2.5 bn
£2.0 bn
£2.0 bn
£1.8 bn
£1.8 bn
2009
2010
£2.1 bn
£2.0 bn
£1.5 bn
£1.0 bn
£0.5 bn
2011
2012
2013
Source: Figure 5 of the 2015 MCT Statement.
Competition effects
Between existing large CPs and new entrants / smaller CPs
A1.5
We have argued that TRs act as a floor to retail call prices and because of this, TRs
above incremental costs drive a wedge between the marginal costs of off-net and
on-net calls. The impact on an individual CP is dependent on its market share.
Other things being equal, CPs with a smaller share of subscribers will be expected
to have a higher proportion of calls going off-net. This can act as a disincentive to
lowering call prices, since lowering call prices is likely to stimulate more calls, and
for a CP with a small market share, the majority of these calls will go off-net (and
incur a TR above incremental cost).
A1.6
We also consider that TRs above pure LRIC, relative to TRs set at pure LRIC, could
reduce incentives for CPs to compete on retail call prices irrespective of market
share. The reason for this is that reducing retail call prices should lead to customers
making a larger number of outbound calls, and therefore would lead to the CP
making larger MTR outpayments, without leading to an increase in inpayments.
These outpayments will be larger the higher the TR. This means that high TRs
could dampen incentives for CPs to reduce retail call prices, irrespective of market
shares.
A1.7
As well as the relative balance of off-net and on-net calls (due to market share), the
competition effects also depend on the relative volume of inbound and outbound
traffic for different customer segments. Even though a small CP would be expected
to see a larger share of inbound traffic generated off-net (i.e. from domestic fixed
and mobile CPs and from overseas) and hence earn higher revenues from TRs at
LRIC+, such an offset will not occur in customer segments where there are more
outbound than inbound calls. While we recognise that for those segments where
outbound calls are proportionately lower than inbound calls, the effects are
reversed, our preference would be for there to be as much of a level playing field
across the market as possible – i.e. that particular CPs were not unduly constrained
7
in retail pricing by virtue of the combination of calling patterns, market share and
TRs above LRIC.
A1.8
Under pure LRIC, these distortions to not occur as the marginal costs of off-net and
on-net calls are the same. We would expect that these competition benefits from
TRs at pure LRIC will benefit consumers in the longer term.
Between mobile and fixed
A1.9
When TRs in one market (mobile or fixed) are set at LRIC while TRs in the other
are set at LRIC+, we would expect total revenues in the market with LRIC TRs to
fall due to reduced incoming payments from the other market and from international
callers. Asymmetry of cost standards between mobile and fixed (e.g. FTRs set at
LRIC, MTRs set at LRIC+) result in operators in one market contributing to the fixed
and common costs of the other market without reciprocal contributions. While we
acknowledge that competition between FCPs and MCPs in calls is not as intense as
it is among FCPs or among MCPs we still believe that it is a relevant consideration
that could distort the competition that does exist between the two markets.
A1.10
In 2013 we calculated that such an asymmetry of cost standards between mobile
and fixed would amount to a per-line cross-subsidy from mobile to fixed of around
£2.14 p.a. per fixed line (in 2012/13 prices). This is comparable to the equivalent
figure calculated in 2011 regarding the move from LRIC+ to LRIC rates
(approximately £2.95 p.a. per subscriber in 2012/13 prices). This is a relatively
small but far from insignificant difference.6
Waterbed effect
A1.11
We acknowledge that there may be waterbed effects on retail prices. The loss of
revenue from TRs generated by calls going between fixed and mobile networks and
calls coming from overseas may be counteracted by retail price increases. This
would allow CPs to recover the lost margin on the lower TRs from the retail-side of
the market. The more competitive the fixed and mobile markets are, the greater we
would expect the associated waterbed effect to be. However, as the waterbed is
only one of many contributing factors that we would expect to play a role in price
changes over time, it is difficult to assess the extent of any waterbed effect
empirically.
A1.12
As described above, we would expect competition to be increased by a change of
cost standard from LRIC+ to pure LRIC which would tend to mitigate this effect and
also have a positive impact on other factors such as innovation and service quality.
Evidence of impacts on rates and prices
A1.13
6
8
As discussed above in relation to the waterbed effect, it is difficult to isolate
empirical evidence to show a direct relationship between the implementation of the
Recommendation and market developments. There are many changes both within
termination markets (e.g. development and deployment of 4G in mobile) and in
related markets (e.g. landline services being increasingly bundled with broadband
and pay TV services) that all jointly affect the market outcomes. Therefore, we can
only point to evidence that is (or is not) consistent with our theoretical arguments
laid out above.
Further details available from paragraphs 8.30 to 8.32 of the 2013 NBMR Statement
A1.14
In mobile, data indicate that consumers were able to get significantly greater bundle
allowance in 2014 than they could in 2011. Prices tended to be lower for equivalent
bundles in 2014 than they were in 2011, and often also included higher data, SMS
and/or voice minute allowances. This was true across both high-end and mid-range
plans in both 12-month and 1-month contracts. The data were less clear for entrylevel plans, where bundles tended to have greater allowances but also greater
prices in 2014 than in 2011.7
A1.15
Furthermore, while out-of-bundle call charges on mobile plans have generally
increased, the overall impact on consumers when combined with the general trend
of increasing bundle allowances seems to be generally neutral in terms of overall
expenditure on mobile voice services.8 These developments are consistent with our
view that a reduction of MTRs to LRIC would increase competition to at least offset
an increase of prices due to a waterbed effect.
A1.16
In fixed, data indicate that consumers face increasing prices year-by-year for a
basket of residential fixed voice services (including line rental and calls). Prices
have been rising by around 1.2% on average over the past 5 years9, and this trend
does not seem to have been affected by the move to pure LRIC.
2
7
Further details available from paragraphs A5.37 to A5.57 of the 2015 MCT Statement.
See paragraphs A5.58 to A5.61 of the 2015 MCT Statement.
9
Figure 4.67 of Ofcom’s 2015 Communications Market Report, available online at:
http://stakeholders.ofcom.org.uk/market-data-research/market-data/communications-marketreports/cmr15/
8
9