Developing Government objectives for broadband - Vodacom
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Moving the debate forward The Vodafone Policy Paper Series • Number 10 • March 2010 Developing Government objectives for broadband
Welcome
I hope you enjoy our tenth Vodafone Policy Paper. Our aim in these papers
is to provide a platform for leading experts to write on issues in public
policy that are important to us at Vodafone. These are the people that we
listen to, even if we do not always agree with them. These are their views,
not ours. We think that they have important things to say that should be of
interest to anybody concerned with good public policy.
Vittorio Colao, Chief Executive, Vodafone Group
Contents
Welcome
– Vittorio Colao
Introduction
– Richard Feasey
Optimal Investment in Broadband : The Trade-Off Between Coverage and Network Capability
– Robert Kenny
Demand-side measures to stimulate Internet and broadband take-up
– David Lewin
Published by Vodafone Group Plc
Copyright © 2010 Vodafone Group Plc
ISBN 978-0-9552578-5-8
The Vodafone Policy Paper Series • Number 10 • March 2010 Developing Government objectives for broadband
1
Richard Feasey
Introduction
The recent global financial crisis has seen an increase in the public benefits, or externalities, associated with very
intervention by public authorities in many areas of high speed broadband networks to justify the level of
activity we had previously thought reserved for the public subsidy which is proposed, for example, by the
private sector. Telecommunications markets have been Australian Government today.
no exception. The past few years have seen
Given the difficulty of making a case for wide public
policymakers committing billions of euros (and billions
intervention to support fibre if there is already extensive
of dollars) of taxpayers' money - if not yet actually
basic broadband deployment, we were also interested in
spending it – towards the funding of new telecoms
how public resources had been used, or could be used,
infrastructure, particularly high speed broadband
networks. These funds are intended not only to deliver to boost demand over those networks that already exist.
short term stimulus and create jobs, but also to lay the Our hypothesis was that Governments can do a lot to
foundations for future competitiveness, sustainability stimulate demand amongst groups who have so far
and social cohesion. Such aims are shared by most proven immune to the marketing activities of private
policymakers (and by many in industry). Achieving them firms. If demand could be expanded in this way, then the
has meant that Governments are now more directly prospects for further investment on the supply side
engaged in telecommunications markets than at any might also improve.
time since privatisation swept the world almost twenty We asked Plum to assess which Europeans used the
years ago. internet today and what might be done to get more of
The initiatives taken by Governments around the world them using it in future. This kind of work has been done
have not, however, been informed by a common view by some agencies – most notably the FCC in their very
about how or where public funds should best be comprehensive Broadband Plan preparations – but not
allocated. Policymakers appeared to be setting a range by others. We also wondered whether previous
of different targets with very different aspirations in Government attempts to boost use of the internet could
terms of the coverage, speed and timing of the networks teach us anything about how public funds should be
they proposed to support. This implied either that each applied – or not - in future. The results here are worrying.
country had very different conditions which justified A great deal of public money has been spent on what we
very different goals or, more plausibly in our view, that might loosely call demand side initiatives in Europe but
there was really no analytical underpinning for the way Plum find that much of it appears to have been wasted.
in which most of these targets were being set. We Notable exceptions include Portugal's initiative to
therefore asked Ingenious Consulting to think about increase student adoption of the internet in schools and
what such an analytical framework might look like, and at home.
to assess some of the existing schemes against their
benchmarks. Plum are optimistic about the prospects for increased
internet adoption in Europe as innovation by the private
Ingenious find that there is a strong case for using public sector continues to break down many of the remaining
funds to extend basic broadband infrastructure – ADSL barriers to adoption. But we should be worried about
or possibly wireless – to as much of the population as anyone under 25 in Europe who does not use the
possible. This suggests that there may be a case for a
internet today. Almost half of these are to be found in
'Mobility Fund' to extend wireless coverage, as the FCC's
Italy. There are also over 30 million adults between 25
much anticipated National Broadband Plan is expected
and 55 who are more evenly distributed across Europe,
to propose or as the Irish Government have
some of whom will not use the internet until 2018
implemented. However, Ingenious find that the case for
without some form of additional assistance. Plum
public funding of higher bandwidth fibre to the home or
suggest how European Governments might use public
fibre to the curb is much less clear. Policymakers would
funds to bring forward this date.
need to make implausibly ambitious assumptions about
Developing Government objectives for broadband The Vodafone Policy Paper Series • Number 10 • March 2010
2
These papers do not provide country by country These papers present the key findings arising from the
recommendations for Governments. But they remind us work of Ingenious and Plum but do not include
that having a clear analytical framework and a good extensive annexes which accompany both documents.
understanding of what the existing data tells us is a These are available at
good basis from which to begin to set ambitious targets http://www.vodafone.com/start/misc/public_policy.html
or to decide how to allocate large sums of taxpayers'
money.
The Vodafone Policy Paper Series • Number 10 • March 2010 Developing Government objectives for broadband
3
Managing Director
Robert Kenny Ingenious Consulting
Rob has extensive telecoms experience, and is Managing Director of Ingenious Consulting.
He previously headed M&A for Level 3. Prior to this, Rob led sales, marketing, strategy and
M&A for REACH (the co-investment vehicle for Telstra and Hongkong Telecom’s
international businesses) and strategy and M&A for Hongkong Telecom. Rob began his
career in consultancy with the LEK partnership, and holds an honours degree in
Mathematics and Management Studies from Cambridge University.
Optimal Investment in Broadband : The Trade-Off
Between Coverage and Network Capability
Executive summary
Governments around the world are announcing and Our analysis focuses on three types of broadband
implementing substantial plans to support high speed technology: standard (up to 15 Mbps download)3, fast
broadband roll-out. However, in many countries there is (up to 50 Mbps download) and superfast (over 50 Mbps
little evidence that these plans are based on a thoughtful download). We consider the incremental costs and
consideration of the pros and cons of different potential benefits of each, acknowledging that the trade-offs are
market interventions, and certainly the plans are widely complex. For example, there are a range of local market
divergent in their scale and objectives. differences including variations in the 'counterfactual'
(the likely broadband infrastructure in a given country
Given the multi-billion Euro sums being spent on these
absent intervention), uncertainties exist over consumer
projects, we believe an analytical framework to support
demand and there are severe difficulties in modelling
decision making in this area could be highly valuable.
externalities.
This paper seeks to provide such a framework.
Our analysis allows us to consider the relative merits of a
The decisions are undoubtedly complex. While costs can
range of deployment strategies. For example, based on
be relatively accurately assessed, consumer demand for
assumptions for the UK, we can contrast sudsidising the
higher speed is far less certain, and the associated
deployment of standard broadband to the final group of
externalities are even harder to quantify (though many
households (and achieving 100% coverage), subsidising
government investment plans are based on the idea that
1 fast broadband to areas where BT and Virgin do not
they will be significant) . Moreover, given that most
already supply, and subsidising superfast broadband to
countries now have relatively wide availability of
the urban core. Figure 1 below illustrates the relative
standard broadband, any rationale for high speed
effectiveness of each approach in terms of the value of
investment must consider the incremental benefits and
consumer benefit realised per € of subsidy.
costs, not the absolute benefits and costs.
Based on our assumptions, the most effective approach
Nonetheless, we believe these decisions can be usefully
(before considering externalities) is to extend the
supported by quantitative analysis. Core to our work is a
coverage of standard broadband to the final 3% of
flexible model allowing for assessment of the
households. For each €1 of subsidy, €2.25 of incremental
incremental benefits of broadband investment, by
consumer value is created.
technology, country and region.2
Developing Government objectives for broadband The Vodafone Policy Paper Series • Number 10 • March 2010
4
Figure 1 : Relative effectiveness of each € of subsidy for a range of
deployment options based on expected UK infrastructure 4 5
€
surplus per € of subsidy 2.50 2.25
2.00
Incremental consumer
1.55
1.50
1.11
1.00 0.72
0.50 0.34
0.02 0.01
0
Standard Fast bb to Fast b b to Fast b b to Superfast Su perfast Supe rfast
broadband to the most the n ext the next bb to the bb to the bb to th e
the most urban area most urban most urban most urban next mo st next most
remo te h hs not already area not area not area urban area u rb an area
se rved already already
served served
To 100% of hhs To 64% To 74% To 92% To 16 % To 38% To 64%
Standard Fast Superfast
Given the existing provision of fast broadband services A fundamental issue when assessing broadband policy is
in the urban core, the case for investing in superfast therefore the value of the incremental externality
broadband services in these regions is very weak. resulting from network deployment.
Competition from existing services reduces the number
By considering different levels of incremental externality,
of customers for superfast broadband, and moreover
we can estimate the potential loss from some of the
reduces the incremental consumer value for those
more aggressive broadband policies. For instance, if you
customers. Therefore, any remaining subsidy after
believe the incremental externality of superfast
supporting standard broadband in the most remote
broadband is €10 per connected household per month,
areas would be more effectively employed in
then France's proposed roll-out of fibre to 70% of the
encouraging deployment of fast broadband services in
population could lead to annualised loss of over €3bn,
areas not already served (starting with the most urban,
compared to a plan focused on regions where the
to 64% coverage). 6
benefits exceeded the costs.
Of course, this analysis ignores the impact of
externalities. The question of which approach has the Note that we are not suggesting that the policies of
greatest overall societal benefit therefore depends on countries such as France are in aggregate value
your perception of the value of externalities under each destructive, only that the extent of the proposed roll-out
option. is such that in the more rural areas covered, the cost is
likely to far exceed the benefits, and thus a more limited
Our analysis also explores the scale of the externalities roll-out would be much better. In more rural areas, a
required to justify current broadband deployment plans government must believe in extremely high incremental
in a range of illustrative countries, ranging from externality benefits to justify current plans.
Australia's commitment to 90% superfast deployment to
Germany's subsidised roll-out of standard broadband to Overall, our analysis suggests that a range of general
100% coverage. We focus on the remotest region lessons can be drawn:
planned to be covered, since the greater expense here • There is a strong case for subsidising the roll-out of
will require the most optimistic view of externalities. standard speed broadband to all households, and
Figure 2 illustrates these results. generally this should be the first priority for
Policies for ubiquitous standard broadband in Italy, the governments (subject to any market specific issues)
UK and Germany can be justified based on the increased • If funds are still available thereafter, there is also a
consumer surplus alone (which more than offsets the 7
case for subsidising fast FTTC or cable broadband
producer deficit). At the other extreme, Australia's (in those areas where the market is not already
ambitions for 90% coverage of superfast FTTH providing). However, in areas with lower population
broadband means that the incremental externalities of density the case becomes highly dependent on the
superfast broadband would need to be around €90 per incremental externalities of fast over standard
connected household per month to justify a roll out this broadband
extensive.
The Vodafone Policy Paper Series • Number 10 • March 2010 Developing Government objectives for broadband
5
Figure 2 : National broadband plans – incremental externalities per month per connected
household required to justify proposed investment in remotest region covered
Possible over-investment Possible under-investment
100
80
h
t 60
n
o
m
r
e
p
€ 40
20
0
Australia Fr ance Sweden Por tugal Germany Italy UK Germany
superfast super fast superfast superfast fast standard standar d standard
• The case for subsidising superfast (FTTH or FTTB8) unless aggressive assumptions are made about the
broadband is weaker. To believe it can create greater value of externalities. Despite this, regional
societal value than fast broadband requires an targeting is, at best, peripheral in many centralised
aggressive assumption about incremental broadband policies. We suggest it should play a
externalities of superfast over fast broadband, but greater role.
even then the societal benefits will be much less
evenly distributed We recognise that this paper is only a small first step
towards a more rigorous framework for decision making,
• Geography is an important consideration in and we would welcome your comments.
broadband policy. In some regions, the market is
likely to deliver without intervention. In other areas, We would like to thank Vodafone for their funding of this
there are clear arguments for government subsidy. work. However, the views and opinions expressed in this
In many of the most rural locations, the case for study are solely those of the authors and do not
subsidy of superfast broadband deployment is weak necessarily reflect the views and opinions of Vodafone.
Developing Government objectives for broadband The Vodafone Policy Paper Series • Number 10 • March 2010
6
Introduction
Governments around the world have been announcing framework that policy makers can use to inform the
ambitious plans to support broadband investment. debate.
However, there is no consensus on the focus of these
At the heart of our analysis is a quantitative model which
plans. Some governments have emphasised high
estimates the value created for consumers and providers
capacity connectivity. Others are more concerned with
of broadband services in a range of scenarios. We do not
assuring the availability of basic broadband to the
aim to provide a definitive answer as to the 'right' form
greatest number. Some countries have announced twin
of broadband subsidy and the manner in which
targets of both: increasing network capability and
infrastructure should be deployed. Rather, we seek to
broadband access.
explore the trade-offs between different broadband
The expenditure involved in deploying broadband investment approaches in a quantitative manner.
networks is significant and therefore even the wealthiest
countries must make trade-offs between depth of Specifically, we have sought to develop a framework
coverage (the proportion of the population with access) which will allow us to understand:
and network capabilities (access speed, technology, • The trade-offs between depth of coverage and
latency, etc.). However, to date the process by which network capabilities, including speed;
governments have made these trade-offs might
generously be described as opaque. There is often little • How these trade-offs are affected by country-
or no discussion as to why a particular broadband plan specific variables;
has been chosen over the almost endless range of • The appropriateness of current broadband policy;
alternatives. Indeed in some cases policy makers have and
actively rejected applying cost-benefit analyses.9
• The questions that should be asked by
There is no question that the issues involved are
governments, regulators and investors when
complex, and that there are gaps in relevant data (for
developing a coherent and socially beneficial
example, the incremental benefits to society of higher
strategy for broadband deployment.
speed broadband). However, the sums being put at risk
by broadband are far too large to be spent without In the report we note the importance of different
rigorous consideration of the alternatives. Therefore the geographic regions, and make reference to different
10
ambition of this report is to provide an analytical 'geotypes' .
The Vodafone Policy Paper Series • Number 10 • March 2010 Developing Government objectives for broadband
7
Broadband and government policy
Many governments have stated their intent to stimulate While our discussions of broadband networks primarily
the provision, or directly provide, fast and superfast related to wireline networks, wireless technologies
broadband networks. However, the details of these plans (mobile, fixed wireless and satellite) are increasingly
vary significantly between countries. prevalent means of broadband delivery.13 For example,
the Irish government has awarded a contract to
Hutchison 3G to provide broadband to the final 10% of
Types of broadband infrastructure population through a hybrid wireless/satellite
14
One question for governments is which type of approach.
broadband technology they wish to support.
Governments frequently articulate this in terms of a Policy objectives
11
particular speed. However, given significant
discrepancies between headline and actual speeds, and Some governments focus on supplying high capacity
differences in upload and download characteristics, superfast broadband for a proportion of the population,
reference to speed alone can be ambiguous. whereas others stress the importance of ubiquitous
broadband at lower speeds. For example, Germany
In practical terms, the decision is to invest in a particular
intends to reach its entire territory with a 1Mbps service
technology rather than a specific speed. Therefore, in this
and 75% coverage of the country with a 50Mbps service.
report we refer to the type of technology, and the speed
The United Kingdom has set a target of 2Mbps for
and characteristics of that technology, rather than simply
ubiquitous access and expects a 50Mbps services to be
the headline download speed. We consider three
deployed to around 40% of the country. Australia has
categories, 'standard', 'fast' and 'superfast' broadband:
stated its ambition to provide high speed 100Mbps
? Standard broadband is capable of achieving access services to 90% of the country.
speeds of up to 15 Mbps download and 1.5 Mbps As might be expected given the different objectives, the
upload. It includes both wireless (e.g. 3G, 4G) and level of planned government spend also varies
wireline technologies, the most notable fixed significantly. At one extreme, the government of
technology being asymmetric digital subscriber line Australia has announced plans for a superfast
(ADSL), currently the most widespread form of broadband network costing A$43bn/€28bn (with the
broadband. Although ADSL connections can government to provide at least A$4.7bn), estimated to
theoretically achieve higher download speeds of up take more than eight years to build and requiring
to 24Mbps, actual speeds are generally considerably roughly 25,000 full-time workers. Conversely Germany,
lower than this.12 with a population roughly four times as large as
? Fast broadband is capable of achieving download Australia's, is planning to spend €150m, or roughly 5% of
speeds of up to 50Mbps and upload speeds of up to Australia's minimum subsidy.
10Mbps. Key technologies includes fibre to the below illustrates the disparity in policy objectives (and
cabinet (FTTC) and cable. FTTC involves laying fibre- plans of commercial operators). Further detail on
optic cables to street cabinets typically located broadband policy by country (and sources) is provided in
within a few hundred metres of the customer appendix A.
premises. Households are then connected from the
cabinet by copper lines. Cable networks often have
a similar architecture, with fibre to the cabinet and Manner of intervention
coax cable from there to the home. FTTC and cable The manner of government intervention also varies. In
speeds are higher than ADSL, but are often not fully some countries, governments are providing direct
symmetric and are determined, in part, by a financial assistance. In others, intervention focuses on
household's distance from the cabinet. encouraging consumer demand. Elsewhere, more
? Superfast broadband connections can achieve market led approaches have been adopted, facilitated by
upload and download speed of over 50Mbps. Main a regulatory framework which seeks to develop
technologies include fibre to the home (FTTH) and competition, encourage efficient investment in
fibre to the building (FTTB), which involve laying infrastructure and ultimately let market dynamics
fibre-optic cables directly to the customer premises, decide.
either through a gigabit passive optical network In Europe EU restrictions on state aid (put in place
(GPON) or point-to-point fibre (PTP). FTTH and FTTB originally to prevent national governments from using
connections typically allow the highest speeds, their funds to aid local industries in contravention of the
lowest latency, greatest reliability and truly single market) has constrained intervention. There has
symmetric connections when contrasted against been an emphasis either on underserved populations or
FTTC and ADSL. on company- and technology-neutral public tenders.
Developing Government objectives for broadband The Vodafone Policy Paper Series • Number 10 • March 2010 8
The Vodafone Policy Paper Series • Number 10 • March 2010 Developing Government objectives for broadband
9
Pricing regulation is another important aspect of considerations including the existing fixed infrastructure,
intervention. While examples of geographically de- the likelihood of commercial provision, consumer
averaged prices are rare, in Finland regulation around demand for fast and superfast broadband technologies,
price discrimination has been relaxed as a method of topography, laissez-faire or interventionist government
stimulating roll-out. philosophy and so on.
However, the wide variation in policies suggests that
The lack of a clear decision making there may a further reason: a lack of a structured
approach for making policy decisions. In the remainder
framework of this report, we introduce such an approach and assess
There are a number of reasons why we would expect various national policies through this prism.
government broadband policy to vary: local marketDeveloping Government objectives for broadband The Vodafone Policy Paper Series • Number 10 • March 2010
10
A framework for assessing broadband policy
Investment trade-offs Overview of the modelling
Broadband investment covers a number of dimensions approach
and even the most affluent of nations are likely to need To develop a practical framework for assessing
to make trade-offs between them. These dimensions broadband trade-offs, we have considered the value of
include: broadband against the classical economic concepts of
? Coverage, with costs per household passed consumer value, producer value and externalities. These
generally increasing with roll-out are illustrated below.
? Speed, driven by both the underlying Consumer and producer value are the most direct
technology (standard, fast, superfast) and measures of economic benefit from the consumption of
network characteristics (network architecture, broadband. The (limited) set of literature exists which
15
distance from the exchange, etc.) measures these types of value forms the basis of our
analysis of consumer and producer value.
? Take-up, often achieved through demand side
stimuli (training, awareness, pricing subsidies, It is generally believed that broadband has significant
etc.) positive externalities, and indeed this is a critical
underpinning assumption for the consensus that
? Mobility, with wireless networks increasingly
government intervention to support broadband may be
viable as a means of broadband delivery
justified. Positive externalities are represented
There is little evidence that broadband policy is being (illustratively) by the green shaded area above the
based on a thoughtful consideration of the trade-offs broadband demand curve. Positive externalities brought
between these investment alternatives. Given the multi- about by different types of broadband may include the
billion Euro sums being spent by governments on following:
broadband projects, we believe an analytical framework
However, it is worth noting that not all externalities
is needed to support decision making in this area. We
associated with high speed broadband are necessarily
have therefore developed a quantitative model which
positive. Some have pointed to the increased carbon
focuses, in particular, on the first two of the above
emissions likely to result from deployment, and others
dimensions: coverage and speed (proxied by network
have posited that high speed networks will increase
type).
digital content piracy. Plum for BSG (2008) also note
Figure 3 : Illustrative value created by broadband
Price (€/month)
Illustrative value of
Consumer value externalities
Broadband market
price assuming no
Producer value price discrimination
p1
Broadband
demand curve
q (% of
q1 (broadband take-up) households)The Vodafone Policy Paper Series • Number 10 • March 2010 Developing Government objectives for broadband
11
negative externalities associated with intervention itself: prediction of the value of externalities associated with
“if public funds rather than voluntary user payments are particular broadband coverage and network capability
used to fund next generation broadband, then an will be subject to a large degree of uncertainty.
additional cost is incurred in terms of the economic cost of
Our modelling approach is therefore to:
raising taxes”.
• Quantitatively focus on estimating consumer and
Externalities associated with broadband are hard to
producer value / surplus, where a more consistent
measure and there is no quantitatively rigorous,
body of quantitative literature exists
comprehensive estimation of the value of the
externalities from broadband, particularly when • Discuss the scale of externalities that would be
considering the incremental value of externalities required to materially change the conclusions,
relating to fast and superfast broadband. Thus any point- drawing on existing research to assess the
likelihood of this outcome
Figure 4 : Sample externalities
Figure 5 : Overview of modelling approach adopted
€
Fully loaded
costs producer
including:
Producer Value • Depreciation
Externalities
• Cost of capital
• Operating
costs
Consumer Net value exc
Surplus externalities
€0
Quantitatively modelled point prediction Quantitatively testedDeveloping Government objectives for broadband The Vodafone Policy Paper Series • Number 10 • March 2010
12
This approach is illustrated in below. If the net value is negative, a subsidy may still be
justified, but a government would need to believe firmly
The model performs the above calculation for each of 8
in the value of sufficient externalities to offset the
geographic regions (geotypes) in the country in
negative net value.
question. Broadly speaking, if for a given geotype the
producer value is greater than the costs (that is, the
producer surplus is positive), then that geotype will be
served by commercial players without the need for
Scope of the modelling approach
intervention. We do not seek to provide a definitive answer to the
If however the producer surplus is negative (i.e. producer value of broadband and the manner in which
value is less than costs) but the net value is positive, then infrastructure should be deployed; rather, we aim to
a subsidy may be needed to support roll-out, but that provide a framework to inform policy debate. A full
subsidy can be justified purely on the basis of private discussion of the approach is provided in the appendices
value. This is the case illustrated above, where total value to this report.
is greater than costs, but producer costs are greater than The model estimates the incremental value created for
producer value. consumers and providers of broadband services under a
Note that we do not imply that as a general rule range of scenarios relating to coverage and technology.
governments should intervene purely to create The model also allows us to explore the relationships
consumer surpluses; rather, we believe that the risk of between other variables, in particular country-specific
intervention is much less when its cost is exceeded by factors such as pricing, penetration and geographic
such surpluses, before bringing into account profile. The costs and benefits of broadband roll-out in a
externalities. country will depend on such variables, and our model
takes these into account where possible.The Vodafone Policy Paper Series • Number 10 • March 2010 Developing Government objectives for broadband
13
Analysis and findings
In this section we explore the case for any government higher speed broadband investment fall rapidly outside
intervention, how intervention should be targetted urban areas.
(particularly in terms of higher speeds vs wider
For instance, based on an Australian profile of household
coverage) and how our analysis compares to actual
mix by geography, there is unlikely to be a significant
government plans.
commercial motivation for a new infrastructure provider
to invest in widespread roll-out of fast or superfast
The case for government broadband (note that we discuss the comparative
incentive for an incumbent provider in the following
intervention section). This is illustrated below.
Much of the discussion of the value of higher speed A positive producer surplus (expected revenue from the
broadband compares total costs and benefits. However, sale of broadband access services less costs) exists only
the critical question for a given government intervention in first three geographic areas (or, geotypes 1-3), which
is whether the incremental gains from the investment represent approximately 30% of households. For the
(the value derived from the upgrade to the base case remaining 70% of households (geotypes 4-8), the
network in the 'counterfactual') exceed the associated producer surplus is negative.
incremental costs. Put another way, even if the total
benefits (as measured by aggregate consumer and In European countries with less population living in
producer value) exceed total costs, this says nothing dense urban areas, such as Sweden and France, the case
Figure 6 : Market incentives to provide high speed broadband for a
16
new monopolist infrastructure provider in Australia (2020)
€m 600 Consumer surplus
400 Producer surplus
200
0
-200
-400
-600
-800
G1 G2 G3 G4 G5 G6 G7 G8
-1,000
Positive producer surplus, creating Negative producer surplus which is
incentives for infrastructure rollout unlikely to stimulate investment based
without intervention solely on expected commercial returns
about whether society gains, as the project's incremental for extended roll-out is similarly weak. In these countries,
benefits (over the counterfactual) might be less than its a direct commercial incentive exists for less than 20% of
incremental costs. This therefore requires us to develop a households.
robust understanding of what the market will deliver by
itself. We note that in Sweden and France the availability of
fast and superfast broadband is already higher than that
A new infrastructure provider is unlikely to predicted to be delivered by the model in 2020
(currently 21% and 16% respectively). However, this has
deliver widespread high speed broadband
been driven by a combination of government
without intervention intervention, historical artifact and non-financial drivers,
The cost of deploying broadband varies significantly rather than the existence of direct commercial
within a country. More remote and less dense areas will incentives. In France, for example, superfast (FTTH) roll-
be more expensive to serve than urban, highly out by non-incumbent operators such as Iliad and
populated regions. Given that broadband prices are NeufCegetel has been fuelled by the bundling of higher
generally flat nationwide, this means that returns for value IPTV services with broadband access in urbanDeveloping Government objectives for broadband The Vodafone Policy Paper Series • Number 10 • March 2010
14
Figure 7 : Market incentives to provide high speed broadband for a new monopolist
infrastructure provider in various EU countries – percent of population (2020)17
26%
38%
45%
10%
71%
81% Need externalities
48% 36%
64%
15%
7% Net value positive
14% 19% 14% Producer surplus
12% positive
UK France Belgium Sweden Italy
areas. In Sweden, innovative municipality-sponsored This is illustrated in figure 8 below. Based on a
roll-out schemes have subsidised open access superfast Portuguese geographic profile and infrastructure, an
networks in towns such as Västerås. incumbent will realise a producer surplus of around
€472m per annum through its standard broadband
We should note that this result is dependent on certain
network. Given costs of deployment and cannibalisation
assumptions that may be optimistic:
of revenues (either wholesale or retail), providing a fast
• That the new entrant has a broadly similar cost base broadband network will erode this surplus, even at a
to the incumbent, and in particular has access to price premium.
ducts on favourable terms
For example, if the incumbent were to deploy a fast FTTC
• That the new entrant can rely on no competitive network in geotype 1, producer surplus would fall by
response from the incumbent (duplicated high around €18m per annum. If an incumbent deployed fast
speed networks would significantly reduce the new broadband to the whole country, the model suggests
entrant's returns in geotypes 2 to 3) that producer surplus would fall by €285m per annum,
or 60%.
• That the moderately positive returns available are
sufficient to justify the capital put at risk (though Contrast this to the results in , where a market incentive
18
note that a cost of capital has been included). for a new entrant exists to provide fast broadband to
26% of households. The incentives for a new entrant are
Thus overall it seems unlikely that, without intervention,
greater than for an incumbent as they will not be
roll-out would go beyond the first three geotypes, and
concerned with cannibalisation of standard broadband
indeed could be appreciably narrower. A new
revenues.
infrastructure provider is unlikely to deliver widespread
roll-out of fast or superfast broadband based purely on Indeed, the very presence of incentives for a new entrant
commercial incentives. may result in deployment of competitive high speed
broadband networks for the most urban regions. This
Incentives for investment are extremely weak has certainly been the case in countries such as the UK,
for incumbents where Virgin Media have deployed fast cable networks in
the first two geotypes, in part to cannibalise revenues
The incentives for widespread deployment of high speed from the incumbent BT. BT has attempted to counter the
broadband are weak for new infrastructure providers, threat by announcing its own plans to deploy a fibre
but are even weaker for incumbents who already network (to a broadly similar geographic footprint).
operate standard speed broadband networks. For these
incumbents, roll-out of high speed broadband services The benefits of copper switch off will help the
to areas already served by standard speed broadband
deployment of broadband, but initially only in
will result in cannibalisation of revenues, further eroding
urban areas
incentives to invest (unless, of course, a third party is
already threatening those standard broadband revenues Commentators have pointed to the benefits of copper
by building its own fibre network). switch off (CSO) as an incentive for upgradingThe Vodafone Policy Paper Series • Number 10 • March 2010 Developing Government objectives for broadband
15
Figure 8 : Producer surplus from fast FTTC broadband deployment,
versus base case in Portugal19
500
472
454
428 426
400 380
Producer surplus (€m)
300 298
243
200 192 187
100
0
ADSL G1 G1-2 G1-3 G1-4 G1-5 G1-6 G1-7 G1-8
base case
ADSL plus FTTC in…
broadband networks. Migrating all consumers to a high greater marketing). This is illustrated in below, based on
speed (fibre) network and switching off the standard Germany's geographic profile.
copper network would allow an incumbent to enjoy
Fast broadband roll-out generates a negative producer
reduced operating costs and release value from the
surplus for all geotypes even if there is only a single fibre
copper itself, land and buildings.
network, but that loss increases significantly if a second
Based on our analysis, a monopolistic incumbent in a network is added. Put another way, the necessary
country with, say, a Portuguese topography will have no subsidy to incentivise fast broadband roll-out would be
direct commercial incentives to invest in parallel build of much larger.
a superfast broadband network. This is illustrated in
Of course, in most circumstances it is axiomatic that
below, where the incremental producer surplus for
more competition will ultimately lead to a better
parallel build is negative for all geotypes.
outcome for consumers. However, if the effect of
If the benefits of CSO are taken into account, and competition is to create or increase a negative producer
consumers are migrated to the new high speed network surplus, then in this context it simply increases the
on a geotype by geotype basis, the commercial subsidy necessary to enable roll-out. Moreover,
incentives for the incumbent improve. However, the regulators with an eye to the long term should be
improvement is sufficient to flip the producer surplus seeking to maximise consumer and producer surplus,
positive only for geotype 1 (11% of households in not just the former.
Portugal). In this most dense region an incumbent may
In different ways, Australia and Singapore's broadband
be incentivised to roll-out superfast broadband and to
plans recognise the impact of competition on potential
transition customers onto a superfast network, but
fibre roll-out, essentially by creating (to a greater or
elsewhere the prospect of CSO is insufficient to turn the
lesser extent) de-facto monopoly providers of
fibre business case positive.
infrastructure, with retail providers riding on top.
Competition from multiple networks is likely to
Given the lack of clear market incentives,
adversely impact on total value
government subsidy may be required to
Multiple providers within a geographic region are likely stimulate deployment
to erode aggregate producer surplus, since network
Government policies broadly fall into the two main
duplication provides additional cost without direct
categories: supply side and demand side policies. Our
additional value for the providers. (Note that we have
focus in this report is on the supply side - where
not sought to quantify the impact of competition
governments invest in infrastructure or tailor their
leading to greater adoption through, for instance,
regulatory action so as to improve provision.22Developing Government objectives for broadband The Vodafone Policy Paper Series • Number 10 • March 2010
16
Figure 9 : Change in producer surplus with and without the benefits of CSO (Portugal)20
Parallel Build With CSO
100
50
Difference in Producer Surplus
vs ADSL Base Case (€m)
0
-50
-100
-150
-200
-250
-300
-350
-400
G1 G2 G3 G4 G5 G6 G7 G8
Figure 10 : Incremental producer surplus over standard broadband only, one or two
fast broadband networks, based on German household geotype mix21
1 network 2 networks
-
-200
-400
-600
-800
€m
-1,000
-1,200
-1,400
-1,600
-1,800
G1 G2 G3 G4 G5 G6 G7 G8
Given that in most geotypes deployment of higher relative to the producer surplus, as coverage of superfast
speed broadband infrastructure results in a producer broadband increases.
deficit, particularly for the incumbent, a supply side
Given that the case for subsidy varies by region,
subsidy may be required to offset the net loss.
government intervention through subsidy should
therefore, at the very least, be targeted to those regions
The case for subsidy varies by region where the case is strongest:
In many geotypes, deployment of high speed • In regions where the consumer surplus exceeds the
broadband results in a net producer deficit. To provide a producer deficit, there is a case for subsidy based on
commercial stimulus to infrastructure providers for these consumer surplus alone. In other words, if only
regions, a subsidy may be required to offset these private value (consumer and producer surplus) is
deficits. The per-household subsidy requirement will considered, society would still benefit from
increase for less dense populations. government subsidy. In the example above (based
This is illustrated in Figure 11 below, where the annual on a Belgian geotype profile in 2020) this private
subsidy required to offset the producer deficit increases, value subsidy case applies for the first two geotypesThe Vodafone Policy Paper Series • Number 10 • March 2010 Developing Government objectives for broadband
17
Figure 11 : Consumer surplus and subsidy requirement for
superfast broadband, Belgian geotype profile23
Incremental consumer surplus Required subsidy per year
500
400
300
€m
200
100
0
G1 G2 G3 G4 G5 G6 G7 G8
only. Thereafter the required subsidy is greater than which, if applied on a flat rate basis, would likely severely
the consumer value. damage the producer surplus. This points to the
importance of pricing flexibility or targeted consumer
• Beyond this point, consumer surplus alone does not
subsidies as tools for maximising societal value.
justify the subsidy investment. To justify further
subsidy to stimulate wider roll-out, a government In our model we assume a flat national price for
must believe there are additional benefits which are broadband. This assumption is consistent with actual
not captured in the private transaction – practice in most countries, with price discrimination very
externalities. From an aggregate societal rarely permitted by regulators. However, our analysis
perspective, the wider the deployment is, the implies that in rural areas where the market is unlikely to
greater the externality value that is required to provide on its own, it may be possible to offset negative
justify the subsidy in each region. In the example producer surplus through higher prices. In other words,
above, externalities per connected households in allowing higher retail prices in less densely populated
the final, most remote area would have to be €145 areas could act as a partial alternative to government
per month to justify subsidising roll-out. subsidy. This is supported by evidence from Finland,
where broadband providers will be expected to fund
Overall, our analysis illustrates the importance of
ubiquitous roll-out without government assistance, but
geography in broadband policy. Despite this, regional
will not be subject to the prohibition of geographic price
targeting is, at best, peripheral the broadband policies of
de-averaging that is prevalent elsewhere.
most central governments. The European Commission
refers to “white”, “black” and “grey” zones based on the Whether or not geographic de-averaging is likely to
number of existing broadband providers, but we believe improve market incentives to deploy broadband
a more geographically targeted approach should play a networks will depend, in part, on the consumer demand
much greater role. curve and whether rural users have higher valuations of
broadband. We believe that further research in this area
De-averaged prices may provide further would be beneficial.
investment incentives
Once a decision has been made to roll out to a particular Trade-offs between coverage and
area (either with or without subsidy), societal value will network capability
be maximised by signing up all households for whom
externalities plus consumer value is greater than the Given the costs of deploying broadband infrastructure,
marginal cost to serve. Given the low variable costs of trade-offs between breadth of coverage and network
telecoms, this may be virtually all customers. However, to capability typically need to be made. From a
persuade the tail of customers (those with low consumer government's perspective, an important question is
value) to sign up would require aggressive pricing, therefore what combination of roll-out and network
24
capability maximises value.Developing Government objectives for broadband The Vodafone Policy Paper Series • Number 10 • March 2010
18
There are benefits to ubiquitous rollout of However, whether universal standard speed broadband
standard broadband, but the case for deployment will be delivered by the market without
investment without intervention is unclear intervention is less clear.
In most EU countries there has been widespread In 2015, providing standard broadband to the final
deployment of standard (typically ADSL) broadband. tranche of the most remote households results in a net
However, there remains a material number of loss of €29m per year for a monopolistic supplier. By
households who do not have broadband coverage of 2020, producer value increases to a nominal €5m per
any speed, particularly in rural areas (those above the year thanks to decreased costs and increased demand,
blue shaded areas) but even in some urban areas (those but given the certain roll-out costs required (around
28
above the yellow shaded areas). €435m in total capital expenditure to serve the final 3%
under a fixed infrastructure) and uncertain demand, it is
Policy makers frequently stress the importance of questionable whether such an approach would be seen
universality of broadband access. Germany intends to as viable by an infrastructure provider.
reach its entire territory with a 1Mbps service. The United
Kingdom has set a target of 2Mbps for ubiquitous access. There is a subsidy case for universal roll-out of
Last year, Finland passed a law making access to standard speed broadband, irrespective of the
broadband a legal right for its citizens, guaranteeing
26 perceived value of externalities
every person access to a 1Mbps broadband connection.
The question this raises is what the relative cost and Given that the producer loss in 2015 is more than offset
benefits of fulfilling such universal service ambitions are. by the increase in consumer surplus, there is a case for
Based on a UK infrastructure where broadband is government subsidy in the final 3% based on private
available for 97% of households, we consider the value alone. Naturally the case would be even stronger if
consumer and producer benefits of ubiquitous (100%) externalities were factored in, and there may be felt to
deployment of standard broadband. This is illustrated in be particular societal value from enabling universal
Figure 13 above. availability of broadband (e.g. increased social inclusion).
Our analysis shows that considerable consumer surplus If the combination of consumer benefit and externality
is realised by roll-out to the final 3% of households value made a compelling case for government
(before considering externalities). The incremental intervention to support universality, there remains the
consumer value also increases over time, driven by question of how it would be most effectively achieved.
falling access prices, crystallisation of demand and
Given the significant costs in connecting the most
increased take-up. Between 2015 and 2020 the total
consumer surplus accruing from the final geotype remote households to a fixed broadband infrastructure,
increases from €64m to €70m. alternative wireless technologies may be a more viable
mechanism for reaching universality. The Irish
Figure 12 : Standard broadband coverage in rural areas by European country, 200825
DSL coverage in rural areas Gap with DSL national coverage
100%
90%
80%
70%
60%
50%
40%
30%
20%
Bulgaria n/a
Romania n/a
Malta n/a
10%
0%
Sweden
Ireland
Netherlands
Cyprus
Slovenia
Austria
Estonia
Lithuania
Latvia
Slovakia
Denmark
Norway
Germany
Italy
Belgium
Spain
Finland
Iceland
UK
Czech Republic
Portugal
France
Greece
Luxemburg
Hungary
PolandThe Vodafone Policy Paper Series • Number 10 • March 2010 Developing Government objectives for broadband
19
government, for example, has awarded a contract to Virgin Media will not deploy fast broadband
Hutchison 3G to provide broadband to the final 10% of services)
population. 3 are adopting a hybrid wireless/satellite • Superfast broadband to the three most urban
approach, rolling out HSDPA services to the majority of geotypes (1, 2 or 3)
this 10% and partnering with satellite provider Avanti
Communications for the remainder. Based on the analysis, we find that:
• The most effective approach is to extend the
Based on the UK infrastructure, providing coverage of standard broadband to the final 3% of
broadband to the final 3% will yield a higher households. For each €1 of subsidy, €2.25 of
return than extending fast and superfast incremental consumer value is created.
broadband coverage
• Given the existing provision of fast broadband
Although the benefits of rolling-out standard broadband services in the most urban areas (geotypes 1 and 2),
to the final group of households outweigh the costs, this the case for investing in superfast broadband
does not necessarily mean that subsidising basic services in these regions is very weak. Competition
broadband universality is the value maximizing from fast broadband reduces the number of
approach. To test this, it needs to be considered against a customers for superfast, and moreover reduces the
range of alternative policies, including further incremental consumer value for those customers
deployment of fast and superfast services. (who would otherwise receive the benefits of fast
We compare the required level of subsidy and the broadband).
corresponding incremental consumer surplus for a range • Any remaining subsidy after support for universal
of deployment options, based on a UK infrastructure standard broadband deployment would be most
profile in 2015. We assume the market has already effectively employed in encouraging deployment of
provided fast broadband to the first 38% of households fast broadband services to the most densely
(geotypes 1 and 2).29 populated areas in which it is not already available.
The results are illustrated in Figure 14 below for the Of course, this analysis ignores the impact of
following subsidy options : externalities. The question of which approach has the
• Standard broadband to the final 3% greatest overall societal benefit therefore depends on
your perception of the value of externalities under each
• Fast broadband to the areas where it is not already option. A government would need to believe that the
available, namely geotypes 3, 4 or 5 (where we externalities resulting from fast broadband are
assume that existing infrastructure providers BT and approximately €5/connected household per month
Figure 13 : Producer and consumer surplus per year from deployment of
27
standard broadband to the final geotype
Producer surplus Consumer surplus
2015 2020
€m 80
70
/ yr 64
60
40
20
5
0
-20
-29
-40Developing Government objectives for broadband The Vodafone Policy Paper Series • Number 10 • March 2010
20
Figure 14 : Relative effectiveness of each € of subsidy for a range of
deployment options based on expected UK infrastructure30 31
€ 2.50 2.25
2.00
surplus per € of subsidy
Incremental consumer
1.55
1.50
1.11
1.00 0.72
0.50 0.34
0.02 0.01
0
Standard Fast bb to Fast bb to Fast bb to Superfast Superfast Superfast
broadband to the m ost the next the next bb to the bb to the bb to the
the most urban area most urban m ost urban most urban next most next m ost
remote hhs not already area not area not area urban area urban area
(3.2%) served already already (15.8%) (21.9%) (26.1%)
(26.1%) served served
(10.4%) (17.4%)
Standard Fast Superfast
higher than from standard broadband in order to prefer incremental consumer benefit, whereas investing €231m
fast broadband to G3 rather than basic to G8. This figure per year in a superfast network generates less (c. €100m)
does not seem excessive, suggesting that while the UK's additional consumer surplus. Ignoring any discussion of
investment is likely to pay societal dividends there externalities, this suggests that if this government is
maybe other options that are at least as attractive. considering subsidising a single technology, investment
in fast broadband, rather than superfast broadband, is
Any analysis should also take into account the longer
most cost effective.
term benefits (and costs) of intervention. For example, it
is argued that superfast broadband provides a more However, what makes the case for fast rather than
'future proofed' solution than other options. However, superfast broadband even stronger is that the benefits
this, in itself, is not an argument for deployment of are much more evenly distributed:
superfast networks. Rather, the assumed future benefits
• The subsidy of fast broadband supports roll-out to
need to analysed and considered in reference to a cost- geotypes 1-6.
benefit framework such as that previously discussed.
• The superfast broadband subsidy only reaches
Distribution of value varies based on the geotypes 1-4.
technology adopted • Moreover, superfast broadband is used by a smaller
group within covered areas – those with the highest
In the example above, subsidising the deployment of willingness-to-pay.
superfast broadband to the first geotype releases less
consumer surplus, € for €, than deploying fast • The net result is that the benefit of the fast
broadband to the next underserved regions. However, in broadband subsidy is shared by 1.7m users, as
addition to considering the absolute value of the opposed to superfast broadband, which is confined
consumer surplus, governments may well be interested to 1m users.
in how evenly and fairly a given consumer surplus is Thus in order to prefer a superfast investment, our
distributed. hypothetical Portuguese policy maker would have to
For example, Figure 15 illustrates the choice for the believe that the externalities per superfast line were
sufficiently greater than the fast externalities to
Portuguese government facing the choice between
compensate for:
spending approximately €225m on either superfast or
fast broadband. • The fact that those externalities will be received
from only 1.0m superfast lines, as opposed to the
A government investing a subsidy of €217m per year in
potential 1.7m fast lines
fast broadband would generate over €150m ofYou can also read
