Co-Investment Models for Broadband Infrastructure

August 13, 2018

The new European Communications Code (which I wrote about here)
will, in its latest draft form, introduce a mechanism allowing investments in
fibre networks made by operators with significant market power (SMP), in some
circumstances, to be excluded from the normal access rules that are usually
imposed by national regulatory authorities (NRAs). This article will discuss
this further and look at some possible models that could qualify for the
exemption before concluding with some comments critiquing this new approach on
the basis of its deviation from the well-respected (and broadly successful)
approach that would otherwise have applied.

For the reasons explained below the new rules could even act
as a disincentive to new investment over the next two or more years.

The Exemption – Commitments and the ‘cumulative conditions’

The rules on co-investment are to be found at Article 74 of
the Code. The latest (though not-necessarily final) draft says that:

‘Undertakings that have been designated as having SMP may
offer “commitments” to open the deployment of a new very high capacity network
(that consists of optical fibre elements up to the end-user premises or base
station) to co-investment.’

The first point to note is that this applies only to optical
fibre and would not apply to other technologies (such as satellite)
irrespective of their merits. This is a deviation from the normal principles of
technology-neutrality that usually govern EU telecoms regulation.

It is also important to note that the investment must be new
and must either be for fibre all the way to the user premises (FTTP – ‘fibre to
the premises’) or (for mobile networks) to the base station. Other types of
fibre investments, such as ‘fibre to the curb’ (FTTC) will not qualify.

The following are the ‘cumulative conditions’ that must be
met in respect of the co-investment:

  • it must be open at any point during the lifetime of the
    network to any provider of electronic communications networks or services on a
    non-discriminatory basis (though reasonable conditions concerning financial
    capacity may be included);
  • it must allow other co-investors who are providers of
    communications services or networks to compete effectively with the SMP
    operator in downstream markets on terms which are fair, reasonable and
    non-discriminatory as between the co-investors, and must allow them all access
    to the full capacity of the new network infrastructure;
  • it must also allow flexibility on the timing for when
    commitments are made, with the possibility to increase participation in the
  • it must be publicised in a timely manner, at least six
    months before the start of deployment (unless the SMP operator is
  • access seekers who decide not to participate must be able
    still to benefit from the same quality, speed, reach and conditions as were
    available before the deployment, complemented with a mechanism allowing adaptation,
    over time in light of developments in the retail market – such access seekers should
    be able to use the very high capacity elements of the network at a time, and on
    the basis of transparent and non-discriminatory terms which reflect
    appropriately the degree of risk inherent at different stages; and
  • it must comply with the minimum criteria set out in Annex IV
    of the Code. 

Annex IV adds that the full terms of the co-investment offer must:

  • be transparent (eg published on the SMP operator’s website);
  • must be designed to favour sustainable competition in the
    long term, meaning that all undertakings should be offered fair, reasonable and
    non-discriminatory terms for participation, though (importantly) this can allow
    for a premium to apply to those joining later to reflect the lower level of
    risk at that stage – all co-investors must be awarded the same level of
    protection both during the build phase and during the exploitation phase, eg by
    granting all participants the same rights of ownership (such as indefeasible
    rights of use – or IRUs);
  • allow for flexibility in terms of the value and timing of commitments
    from each co-investor, eg by an agreed and potentially increasing percentage of
    total end-user lines to which the co-investors must commit – these should be
    set at such a level as to allow smaller service providers still to participate;
  • allow co-investors to assign their rights to other
  • grant all co-investors reciprocal access on fair and
    reasonable terms and on transparent conditions, especially where co-investors
    are separately responsible for deployment of specific parts of the new network
    – if a new investment vehicle is created then it must offer access to all
    co-investors on an ‘equivalence of input’ basis;
  • ensure that any new network elements that contribute
    significantly to the deployment of very high capacity networks are used.

Article 76bis of the Code then sets out the procedure
for confirming these commitments, including through a public consultation.
Where the commitments have been adopted in this way then, provided that at
least one co-investor has entered-into a (compliant) co-investment agreement
with the SMP operator, the NRA will not impose on the SMP operator any of the
other ex ante access obligations in respect of the new infrastructure that
would normally apply. This means that the SMP operator will not have to supply
wholesale broadband products to other ISPs which use this new infrastructure.

Possible Models

Working within the complex criteria set out above, there are
a few high-level models that co-investors (consisting of the SMP operator plus
at least one other) could adopt in order to ensure that their new investment
will be exempt from access regulations. Here are some ideas for this.

Setting up an SPV

The co-investors could set up an entity (a special purpose
vehicle or SPV) that would itself build and own the new infrastructure. The SPV
could offer a standard set of wholesale telecoms services like ‘bitstream
access’ (on fair, reasonable and non-discriminatory terms) and over this new
infrastructure exclusively to its shareholders – refusing to provide services
to other operators. It would have to offer the same set of services to all
shareholder-customers. Importantly it would have to be possible for new
shareholders to join at any time on pre-set terms, without the possibility of
existing shareholders vetoing this (unless on a reasonable basis such as their
financial capacity). Each time a new investor joined there would be an increase
in share capital, or else a sale of some of the current owners’ existing shares
to the new joiner. Although all customers would pay the same price for the
services received from the SPV irrespective of the quantities purchased, the
price for buying a share in the SPV could increase for shareholders joining at
a later date (to reflect the reduced risk they would be taking). The financing
could be arranged either through the balance-sheets of each shareholder (ie the
SPV itself is funded entirely through equity) or else some sort of project
finance for the SPV itself would be an option. In either case each of the
shareholders would likely need to enter into commitments to purchase a minimum
level of services, perhaps at several stages over time.

Joint Build Consortium

Another model would be that each of the co-investors enter
into some sort of consortium agreement under which they each agree to build and
operate infrastructure to a defined specification in a defined geographic area
and then make services using that infrastructure available to other
co-investors on pre-set terms. In this case it may not be sufficient for the
services to be regular wholesale telecom services such as bitstream access
because this may not offer sufficient ‘protection’ to the customers. Instead it
may make sense for the infrastructure-developer to sell ‘IRUs’ to co-investor
customers, allowing very long-term rights (15 years plus) to specific amounts
of capacity, or to individual dark-fibre strands. The arrangement would have to
be open to new co-investors to join later, which could mean new joiners who
themselves commit to build infrastructure in a new geographical area, but might
also include new joiners who do not themselves build any new infrastructure,
but who instead commit to sufficient IRU purchases to qualify. The price of
IRUs taken at a later date can be greater than for IRUs which are committed-to
early on, but should be the same for all co-investors whether or not they are
original members of the consortium or later joiners.

SMP Operator as Manager

Another model might be similar to the Joint Build Consortium
just discussed, but with only the SMP operator building and owning all new
infrastructure in all regions. Other members of the consortium would still ‘own’
IRUs over that infrastructure, and (presumably) would have to commit in advance
to pay a material element of the cost of the new infrastructure to be built. The
price for an IRU would have to be agreed (and open to all co-investors on a
non-discriminatory basis) but – importantly – there is no specific requirement
that it be ‘cost-based’ so it could be at any level the parties agree is


These new rules on co-investment offer a (very complex) way
for new investments to be exempt from the wholesale access rules that would
otherwise apply. These are intended to incentivise additional investment in one
specific technology (fibre-optics) – and even then only when the new
infrastructure constitutes FTTP rather than alternatives like FTTC. When all of
the above conditions are met, the SMP operator will be able to refuse to
provide access to the new infrastructure, or the services which can be provided
over it, to competitors who are not co-investors.

This whole concept is contrary to many of the principles
that have made the EU model for telecoms regulation the gold standard of
international best practice, much emulated in other parts of the world.
Specifically this is not technology-neutral – it very clearly favours one type
of technology over others that might have better cost-benefit ratios, and it
offers a benefit for incumbents in the form of a derogation from the rules that
would otherwise apply, and thus makes it harder for smaller-scale internet
service providers to compete.

Because the new rules apply only to new infrastructure this
could also, perversely, act as a disincentive to any new investment during the
period from now until the new Code has been implemented (which will be two
years from when it is finalised).

By contrast, if these new rules were not put in place, then
the SMP operator would still be obliged to offer wholesale internet services
(such as bitstream access) over any new infrastructure at a regulated price.
Note, though, that under the Code the regulated price could already allow a ‘reasonable
rate of return on capital […] taking into account any risks specific to a
particular new investment network project’ (Article 72.1, para 2 of the Code). The
same paragraph also specifically states that in determining whether or not to
apply price controls in the first place the NRA should take into account the
long-term end-user interest in the deployment of very high capacity networks.

It seems to me that there would have been an option to leave
out these new rules on co-investment entirely, leaving national regulators to
determine, on the basis of Article 72, either that no access rules at all
should apply to certain infrastructure in certain locations, or else that the
price for access should be set at such a level as to incentivise the new
investment. After all, as BEREC itself acknowledged some years ago (before
these new rules were under consideration) in its April 2012, Report
on Co-Investment and SMP in NGA Networks
(BoR (12)41):

‘In those cases where LRIC [Long-Run Incremental
Cost] -based cost-orientation is imposed as an appropriate remedy on the
regulated market, an infrastructure rental solution, based on a per-line rental
charge (including in terms of investment risk taken by the infrastructure
operator, which can still charge the appropriate risk premium) should be
equivalent to building its own infrastructure where the operator would obtain
the totality of the revenues, and paid the totality of costs for a given area’ (section
1.3.1 page 14, emphasis added)

In other words the LRIC price for regulated access to
infrastructure should already reflect the risk taken and offer an appropriate
reward to the SMP operator making any new investment. Thus the regulated access
price ought anyway to offer appropriate incentives for new investment.

Either of these two options (not imposing access rules at
all, or ensuring that the regulated price incentivises new investment) might
have required that the NRA make some differentiation – perhaps segmenting
markets on the basis of a technical measure like minimum (symmetric) upload
speeds – and then allowing, for example a higher regulated access price to
apply for services which met the higher technical standard.

This would have been a much less interventionist approach,
with less risk of distortion of the market and would have had the significant
advantage of allowing investors to come up with whatever structure they wished
for their investment rather than being constrained to comply with rules set out
above. It would also have avoided the risk, inherent in the co-investment model,
of reducing competition by encouraging competitors to co-operate with one

Although this new regime may succeed in incentivising new
network investment, there might have been a simpler, and easier, way to achieve
this with less risk of working against the interests of consumers throughout
the EU.

Mike Conradi is a Partner at DLA Piper LLP, with a focus on
providing commercial and regulatory advice to businesses in the telecoms