SOCIOECONOMIC EFFECTS OF BROADBAND SPEED - Research by Ericsson, Arthur D. Little and Chalmers University of Technology
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
3/221 01-FGB 101 0003 | September 2013 SOCIOECONOMIC EFFECTS OF BROADBAND SPEED Research by Ericsson, Arthur D. Little and Chalmers University of Technology In collaboration with:
Executive summary
Interest in the economic impact of Information and PART 2 – SOCIOECONOMIC EFFECTS OF BROADBAND
Communication Technology (ICT) investments is growing SPEED: A MICROECONOMIC INVESTIGATION
worldwide as governments seek new ways to spur growth. The second study in this report is a microeconomic (household)
Moreover, investment decisions must undergo extra scrutiny level investigation which assesses the impacts of broadband
in the current tough economic conditions. access and upgrades on household income.
In 2010, Ericsson and Arthur D. Little conducted an in-depth The microeconomic study analyzes data from eight OECD
study titled “Socioeconomic Impact of Broadband Network countries, as well as Brazil, India and China (BIC), investigating
Investments,” which measured the economic effects of the similarities and differences between them. The study
broadband penetration. It explored the latest understanding used data from Ericsson ConsumerLab’s web-based survey
of the impact of fixed and mobile broadband investments. conducted in 2010 with over 22,000 respondents.
The study was based on more than 120 academic reports
and business papers published between 2001 and 2010, and Delving deeper than its macroeconomic counterpart, it
results showed that broadband penetration has positive effects measures the impact of broadband speed on household
on society. income, by analyzing whether leveraging the benefits of faster
broadband can improve competitiveness in the labor market
In 2011, Ericsson initiated a follow-up research project,
“Socioeconomic Effects of Broadband Speed,” comprising The main findings of the microeconomic investigation can be
two studies: the first on a macroeconomic (national) level, and summarized as follows:
the second on a microeconomic (household) level. This report
focuses on the results of these two studies. >> Broadband access affects development:
PART 1 – SOCIOECONOMIC EFFECTS OF BROADBAND –– In OECD countries, gaining 4 Mbps of broadband
SPEED: A MACROECONOMIC INVESTIGATION increases household income by USD 2,100 per year
The first study featured in this report is on a macroeconomic –– In BIC countries, introducing a 0.5 Mbps
(national) level, and investigates the impact of broadband broadband connection increases household
speed on macroeconomic growth. Using panel data income by USD 800 per year
econometric methods, Ericsson, in co-operation with Arthur D. >> Broadband speed upgrades affect development:
Little and Chalmers University of Technology, investigated the
–– In OECD countries, upgrading from 0.5 Mbps to 4 Mbps
impact of broadband speed on the economy.
increases income by around USD 322 per month
This study is a world first in quantifying the economic impact –– In BIC countries, upgrading from 0.5 to 4 Mbps
of broadband speed upgrades using a comprehensive, best- increases income by USD 46 per month
practice, scientific method based on empirical data. Data has
been collected for 33 Organization for Economic Co-operation Several key drivers of household income are believed to be
and Development (OECD) member countries, using publicly linked to internet access speed:
available sources. Results show high statistical significance,
are robust to model changes and pass rigorous statistical >> A faster broadband speed boosts personal productivity and
testing regarding the direction of causality. allows for more flexible work arrangements
The study was first presented at the Broadband World Forum in >> A higher speed also opens up possibilities for more
Paris on September 27, 2011. It was significant in the debate on advanced home-based businesses as a replacement, or
the benefits of broadband, being cited in more than 150 press complement to, an ordinary job
articles worldwide as well as at several industry conferences. >> Broadband speed enables people to be more informed,
The study is also published as an academic paper in the better educated and socially and culturally enriched –
International Journal of Management and Network Economics ultimately leading to a faster career path
(Rohman and Bohlin, 2012).
The results from both the macroeconomic and microeconomic
The main result of the study can be summarized as follows:
studies support the theory that both broadband availability
and speed drive growth in an economy. Further investigation is
>> Doubling broadband speeds for an economy can add
needed concerning the relationship between broadband speed
0.3 percent to GDP growth, in a simulation relative to the
and services used.
base year 2008
>> The benefits of faster broadband can be categorized as: To date, the full interplay between broadband coverage, speed,
–– Economic effects, including increased benefits and the value of externalities has yet to be modeled.
innovation and productivity in business Nevertheless, this research project into the socioeconomic
effects of broadband scientifically confirms that speed is
–– Social effects, including better access to a highly important factor in spurring economic growth, for
services and improved healthcare individuals and the overall economy.
–– Environmental effects, including more
efficient energy consumption 2CONTENTS
EXECUTIVE SUMMARY 2
INTRODUCTION
>> Background 5
>> Who should read this report 5
>> The concept of broadband 5
–– Defining speed 5
–– Measuring speed 6
PART 1 – SOCIOECONOMIC EFFECTS OF BROADBAND SPEED:
A MACROECONOMIC INVESTIGATION
>> Exploring the impact of broadband on GDP 10
>> Key sources 10
>> Why speed matters – a macroeconomic investigation 10
–– Economic effects 11
–– Social effects 15
–– Environmental effects 17
–– Empirical modeling for measuring the impact of broadband speed 18
>> Data used for analyzing the effect of broadband speed 19
–– OECD and the World Bank 19
–– Ookla 19
–– Data treatment 19
>> Regression analysis findings: Economic effects of increased broadband speed 20
PART 2 – SOCIOECONOMIC EFFECTS OF BROADBAND SPEED:
A MICROECONOMIC INVESTIGATION
>> Continuing to explore the socioeconomic impact of ICT 21
>> Why speed matters – a microeconomic investigation 21
>> Investigating the impact of broadband speed on the household 21
>> Using Ericsson ConsumerLab global survey data 23
>> The impacts of broadband access and speed on household income 25
>> The escalator effect of broadband investments 26
CONCLUSION
>> Implications for key stakeholders 28
3APPENDICES
Macroeconomic appendices
>> Appendix A: Measuring the economic effects of broadband investments 29
–– Variables in measuring the economic effects of broadband speed investments 29
–– Selection method 29
>> Input-output method 30
>> Regression analysis 31
>> Cost-benefit analysis 32
>> Eclectic method 32
>> Choice of method: Regression analysis 33
>> Appendix B: Regression analysis framework for measuring
the economic effects of broadband speed 34
–– The concept of regression analysis 34
–– On the concept of causality 36
–– Common challenges when conducting regression analysis 37
–– Empirical modeling for measuring the impact of broadband speed 38
>> Appendix C: Definitions 39
>> Appendix D: Statistics glossary 41
>> Appendix E: Literature study overview 42
>> Appendix F: Case studies 46
>> Appendix G: Descriptive statistics 47
>> Appendix H: Regression analysis output 53
>> Appendix I: Statistical tests used 54
Microeconomic appendices
>> Appendix J: References 55
>> Appendix K: About Ericsson ConsumerLab 55
>> Appendix L: Data treatment 55
4INTRODUCTION
BACKGROUND
In 2010, Ericsson, in co-operation with Arthur D. Little, A link may also be dedicated, which normally is the case for a
published the study “Socioeconomic Impact of Broadband wired link between a home computer and the first aggregation
Network Investments,” which measured the economic node. However, if the computer is connected to the internet in
effects of broadband penetration. Following its success, they a wireless manner, even the first link is typically shared.
launched a new research project, “Socioeconomic Effects of
Broadband Speed,” in spring 2011 comprising two studies: Broadband connections use telephone lines, optical fibers,
macroeconomic and microeconomic. The purpose of the the cable TV network, power lines or mobile networks to
project was to isolate and quantify the economic effect of reach the internet. A typical broadband connection and its
broadband capacity upgrades. The project aimed to explore links are presented in Figure 1.
the latest global understanding of the economic effects of
increased broadband speed.
It also aimed to provide statistical
analyses on the relationship The connection between
between broadband speed and the computer and the
economic values using proven modem may be wired
scientific methods. or wireless
This report represents unique
findings, as the impact of
broadband speed upgrades Home computer Broadband access Broadband
has previously not been network provider The rest of
investigated and published network the internet
on an empirical basis.
Home modem Aggregation node for Internet gateway
broadband provider
WHO SHOULD READ (DSLAM, cable node, etc.)
THIS REPORT?
The results of this report are likely Home network Access network Aggregation and core network
to be of interest to a number of
industry stakeholders, including Links of the broadband connection
governments contemplating
investments in broadband
Figure 1: Schematic of a typical broadband connection
infrastructure, the European Source: Ericsson, Bauer, Clark and Lehr, 2010
Union and Commission, NGOs
and industry organizations, operators and regulators.
Defining speed
THE CONCEPT OF BROADBAND SPEED The transmission rate of a data communications system is
The report examines the effects of broadband speed, commonly referred to as its “speed.” Clarity regarding the
which is a concept that must be clarified. However before meaning of this is of great importance, as many different
investigating the concept of broadband speed, a few other definitions exist. The transmission rate depends on the
terms should be explained. characteristics of the dedicated links and how the shared
resources are allocated and loaded. Broadband speed
Broadband internet access, often referred to as “broadband,” depends on:
is defined as a high data transmission, which is always
connected to the internet, typically contrasted with temporary, >> The physical characteristics of the connection links,
lower-rate, dial-up access. The internet consists of shared for example distance and bandwidth
resources, to which anyone can have access through the use >> Policies set for allocation of shared resources,
of a link – a data communications system – that is connected for example priorities and queuing
to shared resources. The data communications system
>> The behavior of other users loading the shared resources
transports information formatted as binary digits called “bits.”
These bits are grouped into delimited information packets
that are transported via communications links. Where there Hence, the achieved broadband speed is not equal to the
is no information transfer, no information packets are being bandwidth, as many parameters play a role here.
shipped over the link. In other words, the link lays idle. Most Traditionally, dial-up connections were used to access
communication systems take advantage of the fact that not the internet. These are still being used in many emerging
all users use the links at the same time. Links are shared economies. A dial-up connection uses the telephone network
between many users. A user occupies transfer capacity when to link the computer to the internet, and has a limited but
that particular user sends or receives information. predictable access capacity, normally restricted to 56 Kbps.
5A broadband connection is – in Occupied bandwidth
contrast to dial-up – capable of Available (unused) bandwidth
always being on. Traditionally,
internet access speeds of
256 Kbps and above have been
considered broadband (OECD),
but recent definitions of basic/
Bandwidth capacity
functional broadband set the
lower limit at 2–4 Mbps (EU, FCC).
The exact speed threshold that
defines broadband is not highly
important to this study, instead it is
the comparison between different The link with least amount
internet access speeds that serves 6F6F6E of unused bandwidth
a purpose here. along the path determines
the end-to-end available
A speed bottleneck can occur bandwidth — the “speed”
anywhere along the broadband Links
connection. The link with the least 1. 2. 3.
amount of unused bandwidth along Home Broadband Broadband
the path determines the end-to-end network access provider network
available bandwidth, see Figure 2. network
Here the difference between Data transfer
capacity and the maximum
possible speed becomes clear. Figure 2: Capacity and available bandwidth of a network path
The capacity is the total bandwidth Source: Arthur D Little analysis, Bauer et al. 2010
of a link, whereas the maximum
possible speed for a user is the
amount of available bandwidth not
Current status of technology
currently used. Both capacity and
available bandwidth are measured Technology potential
in bits per second.
FTTH
FTTB
A link’s capacity is affected
TELCO VDSL/2
by the technology used. Most
technologies today normally do ADSL/2+
not exceed 100 Megabits per DIAL-UP
second per connection, but there EURODOCSIS 2.0
is potential to reach 1 Gigabit CABLE
EURODOCSIS 3.0
per second, particularly with
optical fiber technology. Figure 3 LTE
presents a snapshot of different MOBILE HSPA+
technologies’ capacities. HSPA
Mbps
1000 200 100 25 20 15 10 5 5 10 15 20 25 100 200 1000
Uplink Downlink
Typical net bit rate by technology
Figure 3: Typical net bit rate by technology
Source: Arthur D. Little analysis
Measuring speed
When discussing broadband speed, it is important to define >> A header which contains the sender’s and the recipient’s
which bits are actually included. There are several ways of IP addresses, the protocol governing the information, etc.
counting bits. Either gross bits or net bits are considered. >> The original information – the body
Gross bits are the total number of physically transferred
bits, including user data as well as control signals. Net bits >> A trailer which informs the receiver that the data packet
exclude basic control bits, but include some control signaling is complete
(the protocol) and user data. Net bits form a net data packet,
which consists of three main parts:
6Transfer direction
Speed
Gross Gross bit
bits rate
Trailer Body Header
Net bits T I H Net bit rate
“User data” I Data transfer rate
Figure 4: Breakdown of transfer rates
Source: Ericsson, Arthur D. Little analysis
Gross and net bits relate to speed measurements, as seen in In addition to these measured speeds, one may consider the
Figure 4. Lastly, data transfer rate can be referred to, which advertised speed communicated by the operators. Ideally,
is the achieved average net bit rate that is delivered to the when analyzing the effects of broadband speed, achieved
applications, excluding all protocol overhead and potential speed is preferable over advertised speed, since the latter
retransmissions. This is the actual useful data rate. For irregularly corresponds to the real speed experienced
instance, when downloading a file from the internet, the data by the user (Figure 5). Even disregarding outliers such
transfer rate is the file size (measured in bits) divided by the as South Korea, Japan and France, these differences
file transfer time. are striking.
Average achieved download speed
Average advertised download speed
Download speed (Mbps)
South Korea
Sweden
Japan
Netherlands
France
Denmark
Germany
Switzerland
Finland
Norway
Portugal
Belgium
Slovak Republic
United States
Czech Republic
Austria
United Kingdom
Canada
Australia
Spain
Greece
Hungary
Luxembourg
Ireland
New Zealand
Iceland
Italy
Poland
Turkey
Mexico
Higher achieved speed Lower achieved speed
Figure 5: Comparing advertised speed and achieved speed in OECD countries (2009)
Source: OECD, Ookla speedtest.net
7Term Definition Pros Cons
Advertised speed Typically the theoretically highest > If correctly cited, the advertised > Not real measured data
possible downlink net bit rate speed gives an indication of the
that may be obtainable on a theoretically highest possible
subscribed to connection downlink speed
in a broadband access network
Achieved speed The average (downlink) net bit > Measured data > Few reliable sources
rate (or data transfer rate) that is
obtained over a connection in a > Takes into account the entire > There are several ways to
broadband access network, connection line – end-to-end measure speed over a
caused by capacity limitations in connection with different
the total end-to-end broadband > It is a speed measurement methods giving different results
network. This is the data rate that is easily interpreted by (Akamai uses mirror servers,
typically measured by users the general public which may – but are not aimed
performing self-initiated to – bias the outcome)
speed tests
Peak speed The average highest achieved > Measured data > Depends on the user's internet
downlink net bit rate for a behavior, and does not always
population of broadband users > Takes into account the entire correspond to the maximum
measured from the server end in connection line – end-to-end possible speed
a broadband network over a
defined time span > Few reliable sources
> There are several ways to
measure speed over a
connection, different methods
giving different results
Capacity The average amount of > Measured data > The monthly data rate is not of
consumption rate downloaded bytes over a high relevance when trying to
broadband connection during > Provides a measurement of measure broadband speed,
one month. Measured in the actual volume (bytes) seconds matter to
Mbyte/month or Gbyte/month used per month internet users
Figure 6: Comparing measurements of speed
Taking into account the concerns expressed for the different
ways of measuring broadband speed, the pros and cons of
each “speed term” can be summarized and compared, as
seen in Figure 6. While advertised speed is readily available
and gives an indication of the theoretical maximum, which
may play a role in technology investment decisions, it
offers poor insights into the actual user experience and
risks biasing the data. Achieved speed on the other hand,
is based on actual data from speed testing institutes
(such as Ookla).
8Akamai OECD Ookla
Top 3 countries with highest Top 3 countries with highest Top 3 countries with highest
Term
achieved broadband speed Definition Pros
achieved broadband speed achieved broadband speed
(Mbps) (Mbps) (Mbps)
No.1 No.1 No.1
South Korea 13.7 South Korea 21 South Korea 30.9
No.2 No.2 No.2
Hong Kong 9.4 Hong Kong 15 Hong Kong 26.9
No.3 No.3 No.3
Japan 8.3 Japan 13 Japan 26.6
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
Akamai has servers placed worldwide in The statistics are based on data from The statistics are based on millions of
order to speed up downloads from the speedtest.net, which is a service provided recent test results from Speedtest.net.
internet.They also measure the by Ookla. The measurements were done Top 10 countries require at least 100,000
download speed from these servers, in 2009. The figures are indicative – unique IP addresses for a given country.
which forms the data on which these OECD does not provide exact numbers. Speed measured between Nov 3, 2008
figures are based. and May 5, 2011.
Figure 7: Measurements of achieved speed from three different sources
Source: Akamai (2010), Benkler (2009), Ookla’s netindex.con, accessed on May 6, 2011
As seen in Figure 7, three different, well respected
organizations have estimated different results in terms of top
three achieved broadband speed rates (Mbps). Results are
comparable in the ranking between the countries, yet differ
considerably between institutes for measuring the same
country. This illustrates the immaturity of broadband speed
measurement techniques and concepts.
Peak speed is also based on real data but is likely to be
closer to advertised speed since it counts the highest
achieved speed. This is heavily dependent on the user’s
internet behavior and is not representative of the speed
they experience over time. Finally, capacity consumption
rate takes into account not only the flow rate but the actual
volume of the flow. While this measurement adds an
additional dimension, it does not reflect the speed but
rather the consumption per time unit.
Keeping these considerations in mind, achieved speed
was the chosen measurement. This is due to its relatively
high accuracy in capturing the concept of speed and
its acceptable level of data availability compared to,
for example, peak speed measures.
9PART 1
SOCIOECONOMIC EFFECTS OF BROADBAND SPEED:
A MACROECONOMIC INVESTIGATION
EXPLORING THE IMPACT
OF BROADBAND ON GDP Source Type Data
Previous 2010 research concluded
that increased broadband OECD Dependent variable GDP per capita – USD Purchasing Power Parity (PPP)
penetration adds to economic
output (GDP). The study outlined Ookla Investigated variable Average achieved broadband speed
in this part of the report was
conducted as a follow-up Fixed broadband penetration rate
investigation. It analyzes if, and to
what extent, increased broadband Average broadband subscription price
speeds impact economic activity OECD Control variable
on a national level. Education level
KEY SOURCES Telecommunication revenue in USD (millions)
There are several key sources
World Bank Population density
to this study, of which statistical
data form the major part. For
Figure 8: Main sources used in the macroeconomic study
the analysis on the impact
on GDP, large data sets have
been obtained from well-
renowned sources including
Increased
the Organization for Economic Faster communication
productivity Increased
Co-operation and Development New forms
GDP
(OECD), the World Bank and of sales New forms of
organizing
Reduced
costs Save time
Ookla. Figure 8 provides an Salary
development
overview of the statistics’ Increased New jobs
innovation Building new
data sources. New types of Broadband penetration infrastructure
Old products
computer programs become waste
and network services Incr eased br oadband Less daily
As a complement to the penetration
and speed
commuting Negative
environmental
quantitative analysis of the E-health
Broadband speed
More
working
New computer and
network products impact
Improved
economic impact of broadband welfare Increased online (physical)
automatization1) Less Energy Energy for
speed upgrades, qualitative Increased quality business during manufacturing
of distant Possibility to deal with travels usage
analysis has been conducted Traffic
monitoring communication larger digital online
content (dematerialize)
using secondary sources. The Increased
political
Reduced
environmental
E-learning
qualitative findings in this report transparency
Use many apps
impact
Less need
derive from a literature study Increased
knowledge
On-demand
TV/ audio Energy Consumer
simultaneously for paper
monitoring benefits Increased media
covering over 80 academic spreading quality online
reports and business papers
published between 2002 and
Note: 1) Increased automatization has a negative effect on number of jobs – however the net
2011, in which socioeconomic effect o
f increased broadband speed on number of jobs is positive, according to several studies.
effects of broadband upgrades Figure 9: Schematic of effects stemming from increased broadband speed
have been analyzed.1 Note: This map is a simplification – in reality there are even more factors and linkages.
WHY SPEED MATTERS –
A MACROECONOMIC INVESTIGATION
This chapter describes the impacts on society from increasing >> Environmental effects are impacts on the local or
the average broadband speed, as identified in previous global environment
research by Ericsson and Arthur D. Little. The effects are
classified as economic, social, and environmental. A detailed, yet incomplete, schematic of the effects
stemming from increased broadband speed is presented in
>> Economic effects are measurable impacts on the Figure 9. As seen, these elements form a complex web of
general economy interdependencies.
>> Social effects are behavioral changes among individuals,
groups or society
1
For details of reviewed reports, please consult Appendix E: Literature study overview 10Economic effects
It is widely documented that investments in infrastructure >> The ability to store large amounts of information online
are important for economic growth, and influential political means that more people can work from home more often,
economists have incorporated infrastructure investments which saves commuting time. The amount of business
in theories of growth.2 In other words, broadband upgrade trips can also be reduced, thanks to high-quality video
investments cause certain economic effects, especially when conferences (Ezell, Castro & Atkinson, 2009). The time saved
evolving from one technology group to another, for example can then be used for more productive tasks.
from dial-up to fiber. >> With increased broadband speed come new types of
computer programs and network services, which may
A large-scale infrastructure investment significantly increases reduce firms’ hardware and software costs substantially.
the economic activity in the specific region/city/country The money saved may be used for other investments
as a result of increased employment and purchase of raw leading to higher demands for goods and services, and
material. These are referred to as direct effects. In the long thereby an increase in GDP.
term, improved infrastructure, such as better roads, better
flight connections, and higher broadband speed, stimulates
productivity which is beneficial for the greater economy. This It can also be argued that increased broadband speed will
latter consequence is referred to as an indirect effect, and the lead to more jobs. The main reason is the building of new
more structural changes that take place (often in other sectors), infrastructure. More jobs will be needed in the construction,
are referred to as induced effects. This is a useful framework telecommunications, and electronics sectors. In the long run,
for discussing the impact of broadband investments, however it there are indirect effects that create new jobs, triggered by,
is not used by all studies investigated. for example, enhanced innovation resulting in new services.
Despite the economic benefits coming from increased
The economic effects of infrastructure investments are normally broadband speed, the trade-off between increased penetration
measured by the amount of increased economic activity, which and increased speed must be discussed.
applies to certain areas of the studies concerning broadband
investments. Direct effects of the investment will add to GDP as The Ingenious Consulting Network examined what is more
a result of increased employment and purchase of raw material advantageous – providing broadband to those without it, or
when installing the upgrade of the broadband infrastructure. increasing speed to those who already have it. The findings show
that a penetration increase gives a greater return on investment.
Higher electronic communication speeds are relevant in Figure 10 presents an example from the study. The numbers
knowledge-based communities. This is because it spurs exclude externalities that may arise from the investment, which
productivity, and hence increases competitiveness, which means the rate of return is probably higher than that stated here.3
gives GDP the ability to grow. The reasons for increased The British example suggests there is an upgrade investment
productivity include: opportunity, as 64 percent of British households have “fast”
broadband. This provides a return of EUR 1.55 for every EUR
>> Increased broadband speed means that larger packets of 1 invested. The study argues that authorities should aim for
information can be transferred digitally. This saves time and 100 percent penetration before investing in higher speeds.
forms the basis for greater productivity.
>> Larger amounts of better quality information can be
gathered and distributed than was previously possible,
which opens up new creative and commercial opportunities,
for example being able to reach customers worldwide.
2.5 Standard broadband
2.25
Fast broadband
Consumer surplus per EUR 1
2.0 €23 €37 €64 €220
Incremental externalities per Super-fast broadband
1.55
1.5 household and month are required to
bridge the gap between the consumer
1.11
surplus coming from 100 percent
1.0 penetration of standard broadband
0.72
0.5 0.34 Type of broadband
investment and its
0.02 0.01
0.0 household penetration
To To To To To To To
100% 64% 74% 92% 16% 38% 64%
Standard Fast Super-fast
2–15 Mbps 15–50 Mbps >50 Mbps
Figure 10: Incremental consumer surplus per EUR of subsidy in UK 2015.
Source: Meek, Kenny, Broughton, Thanki and Com, 2010 Note: It is assumed that 38 percent of households already deploy fast broadband
2
For example: Arrow & Kurz (1970), Weitzman (1970) and Barro (1991).
3
Externalities are positive effects in employment and economic output resulting from enhanced productivity, innovation and value chain decomposition. 11However, if the externality effects are big enough (each Crandall & Singer (2010) analyzed the economic impact of
household spending an extra EUR 23 per month), then an broadband investments (both penetration and upgrades)
investment in fast broadband of up to 64 percent of British on consumer welfare, job creation, and economic impact.
households can be motivated, because the rate of return Furthermore, they compared their prior predictions with the
would be as high as for maximum penetration. It must be actual US broadband experience during the past decade.
pointed out that the study finds a positive economic return in They concluded those US investments between 2003 and
investing in higher speeds, even though the rate of return is 2009 led to significant numbers of new jobs, as well as
lower than for penetration (again excluding externalities). There increased GDP. The numbers are specified in Figure 11, and
are also other studies showing that broadband penetration are compared to the estimates made by the authors in 2003.
should be prioritized (Kenny & Kenny, 2011). The study shows that previous forecasts are likely to have
underestimated the economic impacts of the US broadband
investment. Estimating effects of new technologies is
difficult, since one cannot know beforehand how the
technology is going to influence users and society.
Jobs created 600 526
500 495
from broadband 475
500 431
investment
(thousands) 400 329
282
300
164 181
200 155 149
121 129
93
2003 forecast 100
Current estimates 0
2003 2004 2005 2006 2007 2008 2009
93 97
90 87
Economic output 100
81
from broadband
investment 80
61
(USD billions) 60
53
28
40 24 23
19 20 21
2003 forecast 15
20
Current estimates 0
2003 2004 2005 2006 2007 2008 2009
Figure 11: Previous forecasts of economic effects from broadband investments in the US
Source: Crandall et al. 2010
Annual investments New jobs created annually Figure 12 shows how authors
believe that upgrading to second
USD 12.5 billion generation technologies will result
Wireline broadband 247,000 in an average of more than half
incl. FTTH and FTTN a million US jobs sustained from
2010 to 2015, compared to a
USD 14.0 billion
205,000 world without such investments.
Wireless technologies
USD 3.6 billion
Cable broadband incl. 52,000
DOCSIS 3.0
USD 0.3 billion 4,000
Satellite broadband
USD 30.4 billion – total 508,000
Figure 12: Estimated employment effects of broadband investments in the US 2010-2015
Source: Crandall et al. 2010
12GDP Innovation
Scenario 1 4.9 percent increase in GDP 18.8 percent increase in product and
in 2020, of which: process innovation in 2013, of which:
2013 2010
9 Mbps > 4.1 percent applies to the effect > 12.3 percent applies to the effect
of speed increase of speed increase
> 0.8 percent applies to the effect > 6.5 percent applies to the effect
42 Mbps
of new broadband users of new broadband users
Broadband speed increase based on In addition, the study concludes that
Denmark’s High-Speed Committee’s vision Equivalent to DKK 114 billion organization and sales innovation will
also take place.
7.3 percent increased GDP in 2020, The study does not cover innovation
Scenario 2 of which: effects of this particular speed.
However, even further broadband
> 6.3 percent applies to the effect
speed increases should lead to at
2013 2010
9 Mbps of speed increase
least the effects presented above.
> 1.0 percent applies to the effect
92 Mbps of new broadband users
Broadband speed increase based on Equivalent to DKK 170 billion
Sweden’s vision for 2020
Figure 13: Estimated gains in GDP and innovation from increasing the average broadband speed in Denmark
Source: Jespersen and Hansen, 2010
The effect of speed is difficult to examine in detail. Most The New Zealand study “The Need for Speed” by
studies, such as Crandall et al. 2010, discuss upgrades Motu (2009) is a good example of where there have
in technologies, instead of precisely debating speed been productivity gains from faster internet connection
differences. Deployments of new broadband technologies speeds, in which empirical data has been used (Grimes,
often correspond to upgrades in speeds however these are Ren & Stevens, 2009). Two scenarios are examined:
often not well-defined. A study by Copenhagen Economics, upgrading from no broadband (dial-up or no internet
published in 2010, estimates the impact on GDP by connection) to slow broadband (256 Kbps to some Mbps),
increasing the entire economy’s average broadband speed. and from slow broadband to fast broadband (represents
It suggests that an increase in broadband speed from 5 to speeds accomplished through fiber-optic cable network, for
10 Mbps leads to an approximate GDP gain of 1.9 percent. example 10 Mbps). According to survey data from 2006, only
the first scenario showed an increase in firm productivity.
Furthermore, increasing from 25 to 30 Mbps corresponds However, at the time, applications needing faster speeds
to a GDP gain of approximately 0.5 percent according were not used as much (for example videoconferencing),
to the study. Those gains are not thought to be realized which led to the conclusion that only an upgrade from very
immediately, but require around 10 years to come to fruition. low to low speed proved beneficial. Today, the results might
Moreover, the study examines two scenarios applied to the be different. The results from the study are summarized
Danish economy. The first one concerns an upgrade of the in Figure 14.
infrastructure so that the average Dane reaches 42 Mbps,
instead of the 9 Mbps it is today. Innovations in products and processes have been
mentioned as effects of increased broadband speed. This
The second scenario is more extreme – an increase of the mainly stems from the capability of dealing with larger
average speed to 92 Mbps. Both scenarios are thought to amounts of digital content online. It opens up possibilities
cause substantial gains in GDP. The study also discusses for new creative commercial opportunities such as revising
gains in innovation from the first scenario, see Figure 13. traditional processes, for example monitoring a factory
from another part of the world. Product/process innovation
Therefore, the study suggests there are economic benefits also enables new kinds of products, for example the
both in upgrading from 9 to 42 Mbps, and from 42 to development of online games, which has emerged as a huge
90 Mbps. However, the gain is smaller when the original market in Japan and South Korea, with 3D graphics, virtual
speed is higher. Hence, the Copenhagen Economics study reality features, multi-player platforms, etc. Such products
implies that the marginal effect decreases as the speed and services are not possible without high communication
increases. This also conforms to other studies, for example, speeds, which are available in South Korea and Japan.
Meek et al. (2010).
13Conclusions Rationale
Scenario 1 > Increased productivity > 2,688 firms that upgraded from
having no internet or having dial-up
– Indicates a productivity gain
to having slow broadband were
of 11.1 percent
compared to 681 firms that did
– The figure seems to be similar not upgrade
to that of adoption of any
> The conclusion is statistically
128 Kbps 1 Mbps broadband type in general
significant on a level of five percent
Upgrade from no broadband or dial-up
connection to slow broadband
> Indicates no impact on productivity > 369 firms that upgraded from
Scenario 2 slow broadband to fast
– In other words, these
broadband were compared to
estimates imply no average
2,685 firms that did not upgrade
firm productivity improvement
as a result of a shift from > A “raw” productivity increase of
an ADSL connection to a 4-5 percent was found, but is not
cable connection statistically significant
1 Mbps 10 Mbps
> Further calculations (controlling
Upgrade from slow broadband to fast broadband for firm characteristics) suggest
that the non-difference comes
from more productive firms
adopting cable rather than a
productivity-enhancing effect
arising from cable adoption
Figure 14: Changes in firm productivity stemming from internet connection upgrades
in New Zealand. The study is partly based on surveys conducted in 2006.
Source: Grimes et al. 2009
Innovation within sales and organizations can also be Indeed, analyzing socioeconomic effects is a challenging
enhanced. Faster digital communication leads to more endeavor due to the complex web of interrelations that exist
innovative sales procedures and the ability to sell to between society and the economy (Figure 9). Therefore, this
anyone connected to the internet worldwide. Sales can study has been conducted based on actual reported data
also be made through new channels, for example via from reliable sources using advanced regression analysis.
online social networks, or by providing more exciting sales It is clearly stated that the findings presented below are
material to attract customers. An example of organizational based on data within a speed interval of about 2-20 Mbps,
innovation could be to discard the physical office, and calculated over 12 quarters from 2008 to 2010 using
develop an online office as a substitute, thanks to increased measurements of achieved speed. The speed data used
broadband speed. from Ookla for this study includes cable and wireless,
as well as mobile devices.
While a majority of the research on broadband penetration
and speed finds a positive effect on the economy, many As is evident, the effects on broadband investments in
of these studies are forward looking, based on forecasts, general, and of broadband penetration rate in particular,
scenarios and business cases. Thus, there may be an element have been investigated previously by several leading
of inherent optimism regarding the magnitude of this impact. research institutes. However, previous studies investigating
As highlighted by the report “Superfast: is it Really Worth broadband speed using actual data to analyze a large
a Subsidy?” (Kenny & Kenny, 2011), there are challenges to set of nations have not been found. In this regard, the
measuring the impact of broadband on wider society such macroeconomic study in this report plays an important
as the economy. role in providing empirically founded results.
For example, the authors point to the problem of
generalizing from findings on basic broadband to superfast
broadband, given their different cost profiles.4 Moreover,
some previous studies have not incorporated cable and
wireless connections for investigating the impact of speed,
and several studies do not consider the incremental benefits
from higher speeds. Additionally, conclusions based on
methodologies that do not isolate the one-way impact of
speed are certainly open to interpretation with regards to the
direction of causality and the influence of other variables not
accounted for.
4
Kenny et al. (2011) define basic broadband as around 2 Mbps. Although not explicitly defined, they mention 100 Mbps as one example of superfast broadband. 14Social effects
Social effects are behavioral changes among individuals, In healthcare, increased broadband speed makes many new
groups or the society as a result of an infrastructure services possible. The most intuitive one could be real-time
investment. Examples of social effects are: doctor-to-patient communication, which is possible thanks
to decreased latency (delay of the data packets traversing
>> Improved access to mobility for disadvantaged people a network). This is sometimes referred to as e-health. There
are additional advantages arising in the healthcare sector,
>> Better accessibility to basic services with some applications requiring especially high speeds of
(education, government services, etc.) up to 1 Gbps, as seen in Figure 15. File transfer is one of
>> Achievements in terms of safety in traffic and security the most basic of application technologies in healthcare,
in public space yet it can be extremely important. In emergency medicine,
the transmission speed is crucial if images need to be
>> Improved health conditions
analyzed by someone who is not present where the patient
is. The report “Bringing America up to Speed” (Maynard,
The social effects induced from increased broadband speed Cloke, Vock & Wilson, 2010) describes a case where a
can be classified into two groups: social benefits that apply woman was injured in a car crash, and the images could not
to the society as a whole, and social benefits applying be transferred for treatment diagnosis in time due to slow
to the individual. Previously, enhanced productivity was broadband speed, with tragic consequences.
discussed. All productivity-enhancing effects should also
be taking place in the welfare system. Therefore citizens can
benefit from better welfare services, for example healthcare
and education.
Application Broadband speed
Application
technology 1 Mbps 10 Mbps 100 Mbps 1 Gbps
High-quality, non-real-time File transfer High quality High quality High quality High quality
video imaging for diagnosis
Cardiology, neurology H.323 video Low/medium High quality High quality High quality
and emergency room quality1
consultations
Cineo-angiography H.323 video Low/medium High quality High quality High quality
and echocardiograms quality1
3D interactive brain SGI Vizserver Unsupportable Unsupportable Medium High quality
imaging quality
Clinical decision Web browsing High quality High quality High quality High quality
support systems
Advanced clinical decision Image transfer Unsupportable Low quality Medium High quality
support system quality
Professional tele-education MPEG 1 video Low quality2 High quality High quality High quality
Note: 1) At least 10 Mbps is normally recommended for the H.323 video.
Note: 2) MPEG1 video typically 1.5 Mbps without excessive quality loss.
Figure 15: Healthcare applications and corresponding qualities in relation to speed
Source: Erick and Reynolds (2009), Arthur D. Little analysis
15As stated previously, high
broadband speeds enable Study groups can meet
Increased with members remotely
enhanced quality of content
quality of
sent over the internet. For real-time Students who are absent due to
example, high-quality online online video illness can follow lectures, give
Increased Increased quality
communication enables extended presentations, etc.
broadband of real-time
distant e-learning for under- speed online video
served populations. Figure Parents can confer with teachers
16 sums up the educational Improved
benefits deriving from increased data transfer
Remote consulting with databases,
broadband speed. video libraries, computer simulations
Figure 16: Educational benefits of faster broadband speeds
Source: Arthur D. Little analysis, grimes et al (2009), Atkinson et al (2009)
The two main reasons for these benefits are increased quality In addition to improved welfare, there are individual benefits
of real-time online video, and improved data transfer. An arising from using faster broadband speeds. These consumer
example of an educational benefit is that students who are benefits are schematically provided in Figure 17. Enabling distant
absent due to illness can follow lectures, give presentations, communication, increased media quality, time savings, and cost
etc. The benefits add to the overall value of the education. savings are all valuable factors to the individual – and all stem
from increased broadband speed.
Moreover, local, regional and national databases can be
linked for administrative purposes or supervision, through
which information can be transferred easily and quickly.
The private sector also contributes to improved welfare,
by taking advantage of higher broadband speeds. Good
examples are new bibliographic libraries available to the
public, such as Google Scholar or CiteSeer, with seemingly
infinite amounts of digital content.
Consumers benefits
Distant communication Increased media quality Time savings Reduced costs
> Facilitates collaboration > Higher quality of online media > Increased downloading speed > Increased speeds enable
between people regardless content, including movies, often leads to time savings. telemedicine and
of distance, e.g. through music, images, etc. This is especially valuable for telecommuting, which can
social media media content, e.g. movies save consumers money in
> Faster download speeds transportation costs,
– Higher speeds allow these allow for streaming videos – At higher speeds it is particularly rural
groups to provide and audio – this is becoming possible to directly stream communities and the elderly
improved services, more common with the content, e.g. listening
e.g. video sharing on-demand episodes to the music directly from > Smart electrical grids can
> Higher speeds enable more a site without having to provide consumers with a
games to be played online, download it before vision of their electricity
including online communities consumption in real time,
where thousands of players allowing them to adjust
can play on the same platform consumption based on
at once price signals
> Higher speeds decrease > E-shopping may reduce
latency, improving interaction costs to consumers, both
over communication networks1 through lower prices, but
also from traveling to stores
Note: 1) Latency is a measure of the delay of data packets travelling a network.
Figure 17: Broadband speed benefits directly related to the consumer
Source: Crandall et al. (2010), Erick et al. (2009), Arthur D. Little analysis
16Environmental effects Higher broadband speeds can reduce our environmental
Environmental effects are impacts on the local footprint, but there are also negative environmental
or global environment, including: impacts. Figure 18 shows the environmental effects
of broadband speed.
>> Increasing/decreasing CO2 emissions as a result
of traveling and transportation
>> Change in waste production
>> Increased/decreased littering
>> Change in energy consumption
>> Reduced/increased need for fossil fuel
>> Change in traffic congestion leading to change
in air quality and living conditions
Environmental impacts
Positive environmental impacts2 Negative environmental impacts
New types of
Capability of dealing with larger amounts computer and Energy use Contribution
of digital content online (dematerialization) network services to waste
> Efficient smart > Faster broadband > New computer and
Enabling video Less need Working electric grids speeds lead to network products
conferences for paper from home monitor real-time new products, make the older
energy consumption. which require products obsolete.
They reduce energy during The Australian
> Video conferences > Emails will be used > People will be able unnecessary manufacturing, Bureau of Statistics
will replace more which will to work from home energy use and often causing confirmed a
meetings and save paper and reduce their give providers emissions of growing stockpile
reduce business use of transport the necessary greenhouse gases of domestic ICT
travel, decreasing information to equipment and
CO2 emissions ration demand > These products growth in e-waste1
through flexible also require energy
pricing and ensure during usage
efficient supply
> Smart electric grids
require small to
medium amounts of
data to be collected
with very low
latency from
multiple sources
and in a reliable way.
The capacity (bytes)
is of moderate
importance, but
high speed is vital
> Advanced real-time
traffic flow models
may inform drivers
of faster routes, and
reduce congestion
Note: 1) Based on a market research in 2005: Household Electrical and Electronic Waste Benchmark carried out by Australian Bureau of Statistics
Note: 2) If broadband upgrades imply the use of fiber optics instead of copper wires, it results in a ‘greener’ employment, since fiber dissipates
a lot less energy than copper based cables and it saves a lot of materials, according to Bouras et al 2009.
Figure 18: Environmental impacts of increased broadband speeds
Source: Jesperson et al. (2010), Erick et al (2009), Roberts (2009), Bouras et al. (2009), Ericsson and Arthur D. Little analysis
17Increased broadband speeds Model for estimating GDP effect
enhance digital use of content, so of broadband speed upgrades
there is less of a need for paper
and therefore less energy is used Model type Static panel data regression
and less waste is produced. It
also means there is a reduction Source of framework Shiu and Lam (2008), Lehr et. al. (2005), Kontrompis (2005)
in greenhouse gas emissions.
This is due to a reduced need Empirical framework Two-stage least squares
for traveling and physical
meetings, because people can Number of observations 33 OECD countries
work from home and use video
conferences. Also, new types Time series 3 years (12 quarters)
of computer programs and
network services can reduce Dependent variable GDP/capita (OECD)
environmental impacts. One
example is smart electric grids Quarterly data on average achieved broadband downlink speed (Ookla)
Independent variables
that can monitor real-time energy
consumption, and thereby reduce Control variables (independent) > GDP per capita (USD PPP)
unnecessary energy use. They
> Fixed broadband penetration rate
also give electricity providers the
> Average broadband subscription price (USD PPP)
necessary information to ration
demand through flexible pricing > Graduates from upper secondary education or higher
and to ensure efficient energy > Telecommunication revenue in millions of US dollars (PPP)
supply. High communication > Population density
speeds are crucial in smart
electric grids, because they Figure 19: Selected regression model
require small to medium amounts
of data to be collected with
very low latency from a multitude of sources and in a very Empirical modeling for measuring the impact
predictable manner. Hence the requirement of capacity of broadband speed
(bytes) is moderate, but the requirement on high speed The analysis conducted in this macroeconomic study is on a
(low latency) is significant. national, GDP level and provides a comprehensive view
on the economic impact of broadband speed, as well as
Also, increased communication speeds make it possible to novel insights in the field of research. The model used is a
use more advanced traffic flow models in real time. Such static multivariate panel regression analysis focusing on the
models may inform drivers of faster routes, and reduce macro level impacts of broadband speed. It investigates the
congestion – this is also beneficial for the environment economic impact of speed on economic output, measured
because of reduced greenhouse gas emissions. in GDP per capita. Figure 19 provides a concise conceptual
overview on the model analyzed.
Despite these favorable environmental effects as seen in
Figure 18, the negative impacts must be pointed out. First, Previous research (Koutroumpis, 2009; Lehr, Osorio, Gillett
increased broadband speeds lead to the development and & Sirbu, 2005, and Shiu & Lam, 2008), has been used in
production of new products (through innovation, technology combination with statistical testing to identify the one-way
requirements, etc.), which demand energy – often causing impact from broadband speed.
emission of greenhouse gases. In addition, these products
normally require energy during usage. Increased broadband Waverman et al. 2002, and Koutroumpis, 2009, have used a
speeds also require new computer and network programs, similar analysis to identify the impact of broadband penetration
which in many cases make the former products less useful. rate on GDP, assuming there is simultaneity problem. The
Most governments worldwide are aware of this challenge. authors employ a supply-demand model, similar to this study.
The Australian Bureau of Statistics recently confirmed a
growing stockpile of domestic ICT equipment – sometimes Lehr et al. 2005, estimated the effect of broadband
referred to as e-waste. penetration on a number of indicators of economic activity,
including employment, wages, and industry mix, using
In conclusion, increased broadband speeds cause both a cross-sectional panel data set of communities (by zip
positive and negative impacts on the environment. It is likely code) across the United States. They found support for
that the net impact will be positive, if society is careful about the conclusion that broadband positively affects economic
consumption and usage. activity in ways that are consistent with the qualitative
stories told by broadband advocates. The results presented
suggest that the assumed economic impacts of broadband
are both real and measurable.
Shiu et al. 2008, studied the relationships between
economic growth, telecommunications development and
productivity growth of the telecommunications sector in
different countries and regions of the world. They found that
that there is a bidirectional relationship between real GDP
and telecommunications development (as measured by
teledensity) for European and high-income countries.
18DATA USED FOR ANALYZING THE
EFFECT OF BROADBAND SPEED Source Type Data
In order to investigate the effects
of speed on the overall economy OECD Dependent variable GDP per capita (USD PPP)
using regression analysis,
appropriate high-quality panel
Ookla Investigated variable Average achieved broadband speed
data on OECD countries for the
dependent variable (GDP/net new
Fixed broadband penetration rate
jobs created) is required, as well
as for relevant control variables.
Average broadband subscription price
OECD Control variable
Data has been collected from
Education level
the following sources:
Telecommunication revenue in USD (millions)
OECD and the World Bank
Two of the most trusted authorities World Bank Population density
on macroeconomic data, the
Organization for Economic Figure 20: Main sources
Co-operation and Development
(OECD) and World Bank, provide
regularly updated figures for
numerous measurements
worldwide. Using these sources Variable Description Source Comments
allows for high transparency, with
clearly defined indicators that are GDP per capita OECD GDP per capita is obtained
GDP per capita in millions of
publicly available and possible to US dollars at constant prices as follows: GDP per capita =
track over time. and constant PPPs GDP in USD/population
The OECD is an international GDP (USD) GDP (expenditure approach) in OECD
economic organization of millions of US dollars at constant
34 countries founded in 1961 prices and constant PPPs
to stimulate economic progress
and world trade. The OECD Total employment Total employment by OECD France is not included in OECD
also provides a forum where number of people statistics. Therefore, the data for
governments can work together France is collected from ILO
to share their experience and Labor statistics
seek solutions to common
problems. The OECD works with Penetration rate Fixed broadband penetration OECD Only Q2 and Q4 data is available
governments to understand what rate by percentage in the data set. Missing data for
drives economic, social and Q1 and Q3 is calculated by taking
environmental change, measuring the average of values for the
previous and following quarters
productivity and global flows
of trade and investment. It also
produces significant amounts Broadband speed Broadband download speed DataFigure
Ookla 21: Data treatment
is available for the model
on a daily
of economic data analysis to by kilobit per second (kbps) basis. Monthly data is obtained
by taking the data at the end of
provide insights and forecasts.
the month. Quarterly data is
calculated by taking the average
The World Bank, established in 1944, is an international Ookla provides data on achieved
of speeds
every three formonths.
all countries
financial institution that provides financing to developing in the study based on the results from Speedtest.net. The
countries. It includes 187 member-countries, and provides low- site is the most established source online for speed testing
interest loans, interest-free credits and grants to developing and is available on numerous devices and platforms. The
countries for a wide array of purposes such as investments in data from Ookla includes both mobile and fixed broadband
education, health, public administration, infrastructure, financial speeds as blended data. Data is publicly available and
and private sector development, agriculture and environmental updated continuously.
and natural resource management. In addition, the World
Bank compiles statistical data on a wide array of development Data treatment
indicators including urban statistics. Observations have been converted to comparable units, for
example by converting GDP/capita to a single comparable
Ookla currency, adjusted to PPP. The resultant common unit is
There are a few public sources available for measuring provided in USD PPP.
broadband speed, but they only feature short time series.
As a consequence, the best available measurement was Additionally, converted time series have been adapted to
chosen. At the time of production the best publicly available equal intervals, in this case quarterly. Variables that were
data was from Ookla. Ookla is a global leader in broadband not provided at a quarterly level were linearly interpolated,
testing and web-based network diagnostic applications. and thus include the penetration rate where OECD data
Over three million people a day use Ookla software. Its sites was only available for Q2 and Q4 for each year investigated.
include Speedtest.net, Pingtest.net and Net Index. Quarterly speed data is calculated as the average per three
months from Ookla data. See overview of variables and data
treatment in Figure 21. 19You can also read
